60015 TOOLKIT PRACTITIONERS’ TOOLKIT FOR AGRICULTURE PUBLIC EXPENDITURE ANALYSIS MARCH 2011 REPORT NUMBER 60015-GLB TOOLKIT PRACTITIONERS’ TOOLKIT FOR AGRICULTURE PUBLIC EXPENDITURE ANALYSIS Report No. 60015-GLB © 2011 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW Washington, DC 20433 Telephone 202-473-1000 Internet www.worldbank.org/rural E-mail ard@worldbank.org All rights reserved. This volume is a product of the staff of the International Bank for Reconstruction and Development / The World Bank and the UK Department for International Development. The �ndings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, http:// www.copyright.com/. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Of�ce of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org. Cover photo: Quy-Toan Do / World Bank—Man in �eld sifting seeds along Red River in Northern Vietnam. CONTENTS III CONTENTS Acronyms and Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii Introduction and Structure of the Toolkit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1 Context. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Objectives of the Toolkit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.3 Primary Audience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.4 Approach. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.5 User Guide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Part I: Types of PERs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Part II: Preparation Phase of PERs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 2.1 Engaging with Government and Development Partner Counterparts . . . . . . . . . . . . . . . . . . . . . 6 2.2 De�ning the Objectives and Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 2.3 Identifying the Types and Sources of Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2.4 Preparing the Concept Note . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2.5 Estimating the Task Budget and Time Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Part III: Analysis and Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 3.1 Sector Objective and Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 3.2 Budget Allocation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 3.3 Budget Execution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 3.4 Monitoring and Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 3.5 Consolidating Recommendations of the PER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Part IV: Dissemination and Support for Implemntation . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 4.1 Dissemination: To Whom and in What Form?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 4.2 Implementation of Recommendations: Improving Public Spending . . . . . . . . . . . . . . . . . . . . . 24 Part V: Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS LI S T O F E X A MP L E S V LIST OF EXAMPLES Example 1: Annual and Periodic PERs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Example 2: Sector Scope and Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Example 3: Complementarities with Macro-Level PERs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Example 4: Terms of Reference—Levels and Composition of Spending . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Example 5: Terms of Reference—Budget Process and Performance. . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Example 6: Sector Objectives and International Performance Benchmarks . . . . . . . . . . . . . . . . . . . . . . . . 29 Example 7: SWOT Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Example 8: Declining Taxation of Agriculture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Example 9: Agriculture Spending as a Share of Total Spending and of GDP . . . . . . . . . . . . . . . . . . . . . . . . 31 Example 10: Agriculture Spending Per Capita . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Example 11: Donor Spending on Agriculture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 Example 12: Private Sector Investment in Agriculture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Example 13: Functional Allocations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Example 14: Capital vs. Recurrent. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 Example 15: Wage vs. Nonwage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Example 16: Public vs. Private. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Example 17: State-Owned Enterprises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Example 18: National vs. Subnational . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 Example 19: Geographic Allocations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 Example 20: Sampling Subnational Governments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 Example 21: Off-Budget Expenditures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 Example 22: Gender Disaggregation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 Example 23: Cost Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Example 24: Simple Inductive Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Example 25: Simple Congruence Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Example 26: Ef�ciency Gains from Switching from Private to Public Goods . . . . . . . . . . . . . . . . . . . . . . . . 49 Example 27: Marginal Returns to Public Spending across Sectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Example 28: Elasticity Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Example 29: Understanding the Political Economy of Public Spending . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Example 30: Agriculture Sector-Wide Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 Example 31: Institutional Coordination Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 Example 32: Speci�c Recommendations to Government Budget Guidelines . . . . . . . . . . . . . . . . . . . . . . . 53 Example 33: Incentive Mechanisms to Improve Budget Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Example 34: Budgeted vs. Actual . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 Example 35: Timing of Releases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 Example 36: Approaches to Gathering District-Level Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 Example 37: Public Expenditure Tracking Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS VI LIS T OF EX A M PLES Example 38: Cost Effectiveness. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 Example 39: Explaining Inef�ciencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 Example 40: Monitoring and Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 Example 41: Incidence Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 Example 42: Impact Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 Example 43: Public-Private Partnerships . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 Example 44: Matching Grants to Leverage Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 Example 45: Environmental Services and Climate Change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Example 46: Decoupling Agricultural Subsidies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 Example 47: The Case for and against Fertilizer Subsidies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 Example 48: Accountability Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 Example 49: Institutionalizing PERs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 Example 50: Dissemination Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 TOOLK IT LI S T O F TA B L E S V II LIST OF TABLES Table 1: Global Sources of Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Table 2: Generic Elements of a Concept Note for an Agriculture Public Expenditure Review . . . . . . . . . . . . . . . . 9 Table 3: Summary of PER Data Requirements, Types of Analysis, Areas of Recommendations, and Entry Points. . . . . 11 Table 4: Selected Indicators for Agricultural PERs and Guidance on Their Use . . . . . . . . . . . . . . . . . . . . . . . 12 Table 5: An Aggregate View as a Potential Starting Point—Some Examples . . . . . . . . . . . . . . . . . . . . . . . . 14 Table 6: Distortion Indicators for Ugandan Agriculture, Five-Year Averages . . . . . . . . . . . . . . . . . . . . . . . . . 31 Table 7: Agriculture Spending as a Share of Total Spending and of GDP: Nigeria vs. Selected Countries . . . . . . . . . 31 Table 8: Uganda Agriculture Sector Budget as a Share of National Budget and GDP (%), 2001/02–08/09 . . . . . . . . . 32 Table 9: International Comparison of Fiscal Transfers to Agriculture, Average for 2002–04. . . . . . . . . . . . . . . . . 32 Table 10: Correlation of Per Capita GDP and Per Capita Organizational Spending, Turkey, 2002 . . . . . . . . . . . . . . 33 Table 11: Agricultural Expenditure Per Capita (in US$ 1995 Per Rural Inhabitant), LAC . . . . . . . . . . . . . . . . . . . 34 Table 12: Trends in Investment in Agriculture and the Overall Economy . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Table 13: Composition of General Services Support Estimates in Selected Countries (%, avg. 2004–05) . . . . . . . . . 36 Table 14: Functional Composition of the Uganda Agriculture Budget, 2005/06–07/08 . . . . . . . . . . . . . . . . . . . 37 Table 15: Economic Composition of MAAIF’s Budget (Excluding Grants and Domestic Arrears) (Ush Billions) . . . . . . 37 Table 16: Economic Composition of Agricultural Sector Budget (Excluding Local Governments) in Tanzania, 2010/11 (Tsh Billions) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Table 17: Share of O&M in Recurrent and Total Expenditure in Agriculture . . . . . . . . . . . . . . . . . . . . . . . . . 38 Table 18: Decomposition of Wage vs. Nonwage MAAIF Budget, 2005/06–08/09 (Ush Billions) . . . . . . . . . . . . . . 39 Table 19: Classi�cation of Public Expenditure by Private, Public, and Equity and Poverty-Reduction Oriented . . . . . . . 40 Table 20: Sources and Channels of Funding for Agriculture and Local Government . . . . . . . . . . . . . . . . . . . . 41 Table 21: District Conditional Grants Transferred to Local Governments for Agriculture (Based on COFOG De�nition), 2001/02–05/06 (Ush Billions) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 Table 22: Evolution of Agriculture Expenditures, Lao PDR (in Constant 2000/01 KN Billion) . . . . . . . . . . . . . . . . 42 Table 23: Disbursements to Uganda’s Agricultural Sector by Two Off-Budget Donors, 2000/01–06/07 (Ush Billions) (“Agriculture� De�ned According to COFOG Parameters) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 Table 24: Seven Tools for Gender Budget Initiatives and Examples of Their Use . . . . . . . . . . . . . . . . . . . . . . 45 Table 25: Cost Recovery in Irrigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 Table 26: Comparison of Budgetary Allocation to Strategy Elements in Honduras, 2004 . . . . . . . . . . . . . . . . . . 49 Table 27: Marginal Returns to Public Spending in India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Table 28: Disbursements as a Share of Budget Allocations (%). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 Table 29: Cost of Extension Services Provided by MoFA in Selected Regions, 2006 . . . . . . . . . . . . . . . . . . . . 60 Table 30: Comparison of Cost Per Hectare of Irrigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS LI S T O F F I G U R E S IX LIST OF FIGURES Figure 1: Spectrum of Agriculture Public Expenditure Reviews . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Figure 2: Generic Budget Cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Figure 3: Public Spending Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Figure 4: Five-Percent Sustained Agricultural Growth Is an Ambitious Target . . . . . . . . . . . . . . . . . . . . . . . 29 Figure 5: Donor Shares in Total Donor Spending to SSA, Average 2006–08 (%) . . . . . . . . . . . . . . . . . . . . . . 35 Figure 6: State Share of National Agricultural GDP and Agriculture Public Expenditure (2005–2006). . . . . . . . . . . . 42 Figure 7: Map of Mexico Indicating the Share of Agriculture Public Expenditure Relative to Agriculture GDP . . . . . . . 43 Figure 8: NAADS (Districts) Budgeted and Actual Cash Flows, 2005/06 (Ush Millions). . . . . . . . . . . . . . . . . . . 56 Figure 9: Relationship between Inputs, Outputs, and Outcomes in Agriculture . . . . . . . . . . . . . . . . . . . . . . 58 Figure 10: Comparison of Distribution of PROCAMPO and Oportunidades Transfer for Rural Population Decile. . . . . . 63 Figure 11: Concentration Coef�cients of ARD Expenditures, Income, and Land (2006) . . . . . . . . . . . . . . . . . . 64 Figure 12: Concentration Curves of Major Agricultural Programs and Land (%) . . . . . . . . . . . . . . . . . . . . . . 64 Figure 13: Bene�t-Cost Analysis of Public Expenditures in Ghana . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 Figure 14: Agricultural Growth Rate in Latin America in Relation to Agricultural Public Expenditure . . . . . . . . . . . . 66 Figure 15: Average Annual Agricultural GDP and TFP Growth Rates and Distribution of APE/AGDP . . . . . . . . . . . . 67 Figure 16: Government of Rwanda Proposed Budget Calendar. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 Figure 17: Match Publication to Audience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS ACR O N Y MS A N D A BBRE VIAT IONS XI ACRONYMS AND ABBREVIATIONS AFDB African Development Bank GDRS General Directorate of Rural Services AFDF African Development Fund GIDA Ghana Irrigation Development Authority AGDP Agricultural Gross Domestic Product GLSS Ghana Living Standards Survey APE Agricultural Public Expenditure GNP Gross National Product APR Annual Progress Report GTZ Deutsche Gesellschaft für Technische ARD Agriculture and Rural Development Zusammenarbeit BCC Budget Call Circular HR Human Resources BFP Budget Framework Paper IDA International Development Association COFOG Classi�cation of the Functions of Government IDB Inter-American Development Bank (United Nations) IDRC International Development Research Center CAGD Controller and Accountant General’s Department IFAD International Fund for Agricultural Development CC Concentration Coef�cient IFPRI International Food Policy Research Institute CDD Community-Driven Development IMC Irrigation Management Company CIDRS Interministerial Commission for Sustainable ISSER Institute of Statistical, Social and Economic Rural Development Research CONGUA Comisión Nacional de Agua JBSR Joint Budget Support Review CSE Consumer Support Estimates JICA Japan International Cooperation Agency DAC Development Assistance Committee LAC Latin America and the Caribbean DADU District Agricultural Development Unit LGAs Local Government Authorities DFID Department for International Development LGCDG Local Government Capital Development Grant DIS Direct Income Support M&E Monitoring & Evaluation DRA Direct Rate of Assistance MAAIF Ministry of Agriculture, Animal Industry and DSIP Development Strategy and Investment Plan Fisheries EC European Commission MARA Ministry of Agriculture and Rural Affairs EDPRS Economic Development and Poverty Reduction MEF Ministry of Environment and Forestry Strategy MENA Middle East and North Africa ENIGH Encuesta Nacional de Ingresos y Gastos de los MTEF Medium Term Expenditure Framework Hogares MoFPED Ministry of Finance, Planning, and Economic ENNVIH Encuesta Nacional sobre los Niveles de Vida de Development los Hogares MoFA Ministry of Food and Agriculture EU European Union MOU Memorandum of Understanding FAO Food and Agriculture Organization MTEF Medium Term Expenditure Framework FASDEP Food and Agriculture Sector Development NAADS National Agricultural Advisory Service Policy NGO Nongovernmental Organization FDI Foreign Direct Investment NLPIP National Livestock Productivity Improvement FFV Fresh Fruit and Vegetable Project GBI Gender Budget Initiatives O&M Operations and Maintenance GDF General Directorate of Forestry OECD Organisation for Economic Co-operation and GDP Gross Domestic Product Development PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS XI I A CR ONY MS A ND A BB R EV IATIONS OED Operations Evaluation Department SBR Standard Business Reporting PADEP Participatory Agricultural Development and SIDA Swedish International Development Empowerment Project Cooperation Agency PAF Poverty Action Fund SOE State-Owned Enterprises PEAP Poverty Eradication Action Plan SPA Special Provincial Administrations PEC Programa Especial Concurrente para el SPO State Planning Organization Desarrollo Rural Sostenible SSA Sub-Saharan Africa PEIR Public Expenditure and Institutional Review SUE State Unitary Enterprises PER Public Expenditure Review SWAp Sector-Wide Approach PES Payment for Environmental Services SWG Sectoral Working Group PESA Programa Especial de Seguridad Alimentaría SWOT Strengths, Weaknesses, Opportunities, and PETS Public Expenditure Tracking Surveys Threats PFMC Public Financial Management and Control TEDAS Turkey Electricity Distribution Corporation PMA Plan for Modernization of Agriculture TFP Total Factor Productivity PROCAMPO Programa de Apoyos Directos al Campo UN United Nations PPP Public-Private Partnership UNDP United Nations Development Programme PSE Producer Support Estimates USAID United States Agency for International RADU Regional Agricultural Development Unit Development R&D Research and Development VAT Value-Added Tax RUTA Regional Unit for Technical Assistance WBI Women’s Budget Initiative SBFP Sector Budget Framework Paper WDR World Development Report TOOLK IT PR E FA C E X III PREFACE This toolkit for analyzing public expenditures in agriculture contributes to a broader effort to enhance the focus, quality, and appropriate scaling of public spending in the sector. More speci�cally, the toolkit has two goals: 1. To provide checklists for practitioners conducting various kinds of agriculture public expenditure analyses, and 2. To provide selected examples on aspects of the checklist to help guide analysis. The toolkit presents a diversity of approaches and describes experiences—both positive and negative—in conducting agri- cultural public spending analyses in different settings and with different objectives. It offers checklists of issues and options, rather than a minimum list of issues to be covered. Needs, existing work time, and budget constraints will likely drive the selection of the checklist topics to be covered in any given analysis of public expenditures. The toolkit is organized to facilitate this selectivity of topic, while maintaining a strategic perspective. The supporting examples draw on numerous analyses of public expenditures in agriculture. The toolkit was developed by the World Bank Agriculture and Rural Development Department with support from the UK Department for International Development (DFID) as part of the partnership on “Public Expenditures for Pro-Poor Agricultural Growth�.1 This effort has bene�ted from the experiences and feedback of numerous World Bank staff, including regional and central unit staff working on agricultural and rural development issues, and members of the Bank’s Rural Policies Thematic Group. The completion of the toolkit was led by Robert Townsend, with substantive inputs from Mona Sur, Richard Anson, Christopher Delgado, Limin Wang, Saswati Bora, Sergiy Zorya, Yurie Tanimichi, Iride Ceccacci, and Pauline Zwaans. An initial consultation on the draft outline of this toolkit was held in February 2009, followed by a second consultation in April 2009. A global consultative workshop in May 2009 in Addis Ababa involved technical of�cials from several donor organiza- tions and from governments in Africa, Asia, and Latin America.2 Staff from donor agencies, including the International Food Policy Research Institute (IFPRI), the Food and Agriculture Organization of the United Nations (FAO), the Inter-American Development Bank (IDB), the International Fund for Agricultural Development (IFAD), and the Regional Unit for Technical Assistance (RUTA) have also been involved in developing this toolkit and provided useful feedback. The toolkit is available in print and on the Internet. Subject to its demonstrated value and user feedback, it may be updated periodically, especially as an input for ongoing sectoral expenditure analyses and related training activities. 1 For commissioned background papers, conference materials, and other relevant information, see www.worldbank.org/agper. 2 For the Addis Ababa workshop agenda and corresponding presentations, see www.worldbank.org/agper. PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS I N T R O D U C T I O N A ND ST RUCT URE OF T HE T OOL KIT 1 INTRODUCTION AND STRUCTURE OF THE TOOLKIT 1.1: CONTEXT The United Kingdom Department for International Development This emphasis often becomes self-enforcing, result- (DFID) and the World Bank Agriculture and Rural Development ing in subsidy traps that are dif�cult to exit. These Department (ARD) jointly launched in June 2006 the “Public concerns have increased attention on public spending Expenditures for Pro-Poor Agricultural Growth� project. The in the transforming economies of Asia. project is composed of three modules. Module 1 produced ƒ Many governments are scaling up their support for �ve background papers that reviewed and analyzed trends agriculture, particularly in the poorest countries. With and impacts of public expenditure in agriculture. Module 2 pro- greater recognition of the importance of agriculture, duced country case studies on agricultural public expenditure ampli�ed by the global food price spike in 2008, reviews (Nigeria, Uganda, Lao PDR, Nepal, and Honduras). many countries are scaling up support for agricul- Finally, Module 3 consisted of a series of lesson-learning, ture. African countries have committed to increasing capacity-building activities and dissemination, which included their spending on agriculture to 10 percent of overall a two-day workshop in Addis Ababa, Ethiopia in May 2009, public spending but also recognize that the quality of the publication of a synthesis report and this agriculture public spending matters. The Africa Union’s Comprehensive expenditure review (PER) toolkit for practitioners. Africa Agriculture Development Programme is helping improve the composition of public spending, within The increased attention on enhancing the focus, quality and which public expenditure analysis is an important tool. appropriate scaling of public spending in agriculture is the ƒ Donors are also scaling up support for agriculture, often result of the following factors: with a shift to budget support, with more reliance on gov- ƒ The agricultural policy environment has generally ernment budget processes. In support of government ef- improved in the poorest countries. Macroeconomic forts to scale up investment in agriculture, development stability, lower taxation of agriculture, and reduced partners are also increasing their support. For example, conflict have improved the incentives of both farmers the World Bank Group projects an increase in its support and agribusiness entrepreneurs to invest and produce to agriculture from an average of US$ 4.1 billion annually more. Recent agricultural growth, particularly in Africa, in 2006–08 to between US$ 6.2 billion and US$ 8.3 billion has been driven by this less distortive policy environ- annually in 2010–2012. There is also an increasing shift ment, which has raised investment return. However, to sector-wide programs in some regions, as well as to as the space for future policy improvements declines, pooled funding and budget support, which all rely more future growth will need to rely more on public invest- on the government budget process for allocating, imple- ments for productivity growth. As a result, the levels menting, and accounting for agriculture public spending, and quality of agricultural public spending will become in line with the Paris Declaration and Accra Accord. an increasingly important contributor to growth, pov- ƒ A desire to improve the quality of public expenditure erty reduction, and food security, particularly in Africa, analyses. Many agriculture expenditure analyses have and hence deserves increased attention. been conducted, but they remain uneven in quality and ƒ Agricultural subsidies remain signi�cant in richer impact, especially in the implementation of key recom- developing countries. Rapidly rising rural-urban income mendations. The variability is often caused by data con- gaps in many richer developing countries have often straints, the size of the budget available for the analyses, induced spending on subsidy transfers to rural areas, insuf�cient focus, a lack of stakeholder buy-in, inadequate rather than on investments to raise earned incomes. commitment by counterparts, and limited capacity. PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 2 INTROD UC TION A ND S TR UCTURE OF THE TOOLK IT It should be noted that the agricultural sector is different from attention in PERs to issues such as the roles of the private vs. public many other sectors due to the dominancy of the private sector sector, timing of expenditure releases, breadth and depth of the (farmers), the seasonal nature of production, and the multiplicity coverage of analysis, and differentiation in expenditure tracking of subsectors. This unique nature of the sector requires more by subsector. These issues are addressed in this toolkit. 1.2: OBJECTIVES OF THE TOOLKIT This toolkit has two broad objectives: analyses and providing information to better inform public spending decisions 1. To provide checklists for practitioners conducting 2. To provide selected examples on aspects of the various kinds of agriculture public expenditure analy- checklists, drawn from the growing body of public ses, with the goal of strengthening public budget spending analyses, to help guide analysis 1.3: PRIMARY AUDIENCE The toolkit is directed at multiple audiences, including ƒ Government of�cials, especially at the technical level, who are responsible for planning and manag- ƒ Donor agency practitioners who are responsible for ing public expenditures in the agricultural sector— carrying out and/or supporting public expenditure particularly of�cials from Ministries of Agriculture, analysis of the agricultural sector related ministries and agencies, and Ministries of ƒ Expenditure analysts from academia, research institutes, Finance and Planning and the private sector (including farmer organizations) 1.4: APPROACH The toolkit reflects the approaches, experiences, and per- to broader multisectoral reviews, or as part of a more spectives of the many practitioners who participated in its frequent review process in support of a county’s an- development, including World Bank staff, staff of govern- nual budget cycle. ment agencies, and staff from other donor agencies. The ƒ Thematic reviews: Undertaken on a speci�c issue, toolkit offers checklists and a menu of options for conducting subsector, or program. These can take the form of public expenditure analyses. It offers guidance for conduct- either a comprehensive or rapid review. ing a range of analyses and reviews including Choice among these three common types of reviews de- ƒ Comprehensive reviews: Undertaken as a detailed, pends on the needs and circumstances of a given country, self-standing, sector-wide agriculture public expendi- the starting point for the analysis, and the resources avail- ture review. These are usually undertaken periodically able. The checklists and menu of options offered in this tool- with detailed analysis and can be programmatic, span- kit are intended to encourage a more systematic, strategic ning over several years. approach to selecting and applying the relevant tools over a ƒ Rapid reviews: Undertaken as a quick review to period of time. The ultimate goal is to improve the quality of deepen policy dialogue, frame strategic action plans, expenditure analysis and thus its role in enhancing sectoral guide agricultural project preparations, and contribute expenditure ef�ciencies and poverty reduction. 1.5: USER GUIDE The toolkit is organized around checklists (or sets of ques- 1.5.1 CHECKLISTS tions and themes) to help guide analysis. The checklist refer- The checklists are provided as a reminder to practitio- ences examples from previous analysis. ners of the main features that might be included in an TOOLK IT 1. 5 : U S E R G U I D E 3 agricultural public expenditure review. They can be used 1.5.2 EXAMPLES to guide the scope, the subsequent analysis and recom- The last part of the document provides selected examples mendations, and the dissemination and uptake of �ndings. on aspects of the checklists, drawn from the growing body The checklists highlight the major issues across the ele- of literature on public spending analyses (drawn from studies ments of expenditure analysis and are organized around: �nanced by the DFID-World Bank Partnership and other stud- (i) the types of public expenditure reviews; (ii) preparation ies carried out by the Bank and other organizations, including phase of public expenditure reviews, (iii) analysis and IFPRI, IFAD, FAO, and RUTA). The examples offer “how-to� recommendations, and (iv) dissemination and support for advice on the issues and questions raised in the analysis implementation. checklists. PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 4 PA RT I — TY PES OF PER S Part I: TYPES OF PER S At the outset, it’s important to highlight the spectrum of of allocations made to a particular subsector), and the Uganda agricultural PERs undertaken, which vary by both breadth study focused on both. (thematic and institutional coverage) and depth (flow of funds and impact) of analysis. PER examples within this spectrum ƒ Comprehensive: Undertaken as a detailed self- include comprehensive (extensive breadth and depth of anal- standing sector-wide agriculture PER. These are ysis, e.g., Uganda3), rapid review (extensive breadth but fairly usually undertaken periodically with detailed analy- limited depth, e.g., Lebanon4), and thematic (limited breadth, sis, often including an expenditure and institutional but extensive depth, e.g., Zambia5) (Figure 1). Choice within assessment. this spectrum determines the scope of possible allocative • Example: Uganda Public Expenditure Review 2010. and technical ef�ciency analyses6. For example, the Lebanon The review was undertaken in three phases: (i) an review focused on allocative ef�ciency (better aligning the al- analysis of trends and allocations against compara- location of spending to comparative advantage), the Zambia tors together with an analysis of budget execution; study focused on technical ef�ciency (better implementation (ii) a value-for-money analysis of related infrastruc- ture investments (rehabilitation of rural roads, bridges, livestock markets, and �sh ponds), an FIGURE 1: Spectrum of Agriculture Public Expenditure assessment of the quality of delivery mechanisms Reviews and strategic alignment of several agriculture ser- Greater scope for allocative efficiency analysis vices, and an incidence analysis of various govern- Breadth ment programs; and (iii) a detailed impact evalu- Less (thematic and institutional coverage) More ation of the largest agricultural service program. Less An 80-page synthesis report was then prepared, Lebanon (flow of funds and impact analysis) rapid integrating the outputs and �ndings of each phase. review The analysis was undertaken over a 2-year period Greater by a team composed of both government staff and scope for Depth development partners. technical efficiency ƒ Rapid: Undertaken as a quick review to deepen policy analysis dialogue, frame strategic action plans, guide agricul- Zambia Uganda thematic comprehensive tural project preparations, contribute to broader multi- review review sectoral reviews, or as part of a more frequent review More process in support of a country’s annual budget cycle. • Example: Lebanon Public Expenditure Review Source: Authors. 2009. As part of a rapid technical assistance to 3 World Bank (2010b). Lebanon, the review drew on previous work. It 4 World Bank (2010c). provided (i) an overview of the agriculture sector 5 World Bank (2010e). 6 Allocative ef�ciency refers to the degree to which resources are and a strategic vision, following an assessment allocated in conformity with government priorities and highest of agricultural strengths, weaknesses, opportuni- returns (see Section 3.2.2), while technical ef�ciency refers to ties, and threats; (ii) a brief assessment of institu- the capacity to use allocated resources at a cost that achieves ef�ciency gains and (to the extent appropriate) is competitive tional capacity to deliver the strategy; (iii) a broad with market prices (see Section 3.3.2). TOOLK IT PA RT I — T Y P E S O F P E R S 5 assessment of the alignment of public spending The analysis then presented a summary of the with the strategy; and (iv) recommendations for major �ndings and recommendation for program better alignment. The main report was 15 pages improvement. The �nal 80-page report integrated but with a detailed annex assessing the Tobacco these analyses. The work was undertaken over a Price Support Program (a large budget expendi- 2-year period by a team comprising both govern- ture), with a presentation of possible alternative ment staff and development partners. mechanisms of support. The review was complet- Annual vs. periodic: Another choice to be made is whether ed in several months. to conduct PERs annually or periodically. Annual PERs typi- ƒ Thematic: Undertaken on a speci�c issue, sub-sector cally have a primary focus on allocative ef�ciency—that is, to or program. These can take the form of either a what extent the (preliminary) budget allocation is aligned to comprehensive or rapid review of a particular thematic the national development strategy and the types of invest- area. ments with highest returns and how this has changed from • Example: Zambia Public Expenditure Review previous years. There is also typically some analysis of budget 2010 (with a focus on fertilizer subsidies). With execution with a focus on expenditures vs. allocations and the Zambia fertilizer subsidy program accounting the timeliness of these expenditures, with an assessment for about 40 percent of public spending in agricul- of how this has changed from previous years. More detailed ture, a focused assessment on the ef�ciency and analysis of technical ef�ciency (e.g., public expenditure track- effectiveness of the program was undertaken. ing surveys, cost-ef�ciency, incidence analysis, and impact The analysis (i) presented a brief overview of the evaluation) is not typically done annually but undertaken peri- fertilizer subsidy program objectives and design odically (see Example 1: Annual and Periodic PERs). features, (ii) assessed the relative size of the pro- gram through a general review of Zambia’s public Choice across the spectrum of possible PERs will de- expenditures in agriculture and focused on the im- pend on need, time frame for analysis, and available pact of the annual budget cycle on implementation budget−determined during the preparation phase of PERs. of the fertilizer subsidy program, and (iii) assessed Yet, teams are encouraged to do more inclusive, broader the quality of service delivery (targeting, timeli- agricultural PERs, which would allow for an overview of all ness, delivery) and impact (through survey data). spending given the fungible nature of public expenditure. PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 6 PA RT II — PREPA R ATION PH AS E OF PER S Part II: PREPARATION PHASE OF PER S 2.1: ENGAGING WITH GOVERNMENT AND DEVELOPMENT PARTNER COUNTERPARTS A checklist for practitioners: and end of the expenditure review process. The team will need to weigh the bene�ts and costs. ƒ Decide on the main government counterparts to engage, • Ensuring government counterparts are part of the reflecting the intended purpose of the public expendi- preparation team may help ensure easier access to ture analysis. Counterparts could include the ministries public expenditure data. of �nance, planning, agriculture, and local government, • Engaging senior government of�cials at various dependent on client demand, intended purpose, and stages of preparation may speed adoption of data requirements. If the issues to be addressed require recommendations. eventual action by a speci�c government ministry, then they should be engaged in some way in the public ex- • Greater collaboration with stakeholders may en- penditure review. For example: hance the effectiveness of the analysis but will add to the cost and time required to complete the task. • Issues related to spending levels, the overall The task team will need to assess these trade-offs. budget cycle, budget releases, and accountability usually require actions by the ministry of �nance. ƒ Decide on the expected role of other development partners. In many countries, development partners are • Issues related to decentralization, including �scal becoming increasingly interested in the performance transfers, usually require actions by the ministry of of the government budget process with a shift to local government. more sector-wide programs. Again, the team will need • Issues on the composition of spending usually to weigh the bene�ts and costs. Having joint develop- require actions by the ministry(ies) of agriculture. ment partner reviews can Ensuring early engagement of respective govern- • serve to pool �nancial and human resources for ment counterparts may increase the likelihood of preparation both getting access to expenditure data and for recommendations to be adopted. • avoid duplication of efforts ƒ Decide on the expected role of government counter- • help ensure a common voice on the �ndings and parts in the public expenditure analysis. Government recommendations, and counterparts can have varying roles, from being part • help ensure uptake of actions required by develop- of the preparation team, to being updated at various ment partners, including donor alignment. stages of progress, to being engaged at the beginning 2.2: DEFINING THE OBJECTIVES AND SCOPE ƒ General objectives. Public expenditure review objec- • Identifying a set of evidence-based recommenda- tives should reflect need, demand, and the intended tions to improve the ef�ciency and equity of public use of the public spending analysis (by the ministry of expenditures through analysis of the impact of �nance, ministry of agriculture, development partners, etc.).7 These include variants of the following two 7 The 1998 Operations Evaluation Department (OED) report aspects: (World Bank 1998) stressed that objectives also guide the level of Government and Development Partner assistance. “[I]f the TOOLK IT 2. 2 : D E F I N I N G T H E OBJ E CT IVE S AND SCOP E 7 past spending, the alignment of the levels and an important thematic focus (particularly in Africa) composition of spending to sectoral objectives and (see Example 2: Sector Scope and Analysis). to the areas with the highest impact on agricultural (ii) Institutional coverage: Decide on which public growth and food security (if there is an inconsisten- entities to include in the analysis. The spectrum cy between these areas and the sector objectives), and the performance of the budget process (based of public entities engaged in the agriculture on a backward-looking diagnostic assessment and sector is often large, including ministries, local a forward-looking and strategic perspective). governments, and state-owned enterprises. Each often has a separate parliamentary vote • Building capacity to undertake future public expen- in the budget. This often implies multiple data diture reviews through database development and sources and subsector budget processes. In work with counterpart teams. addition, decide on the extent to which the PER ƒ Specific objectives. Speci�c objectives could range will try to collect off-budget expenditures in the from identifying a speci�c set of policy recommenda- sector, such as donor- or NGO-funded activities, tions related to allocations across thematic areas to the state-owned enterprises, and implicit subsidies flow-of-fund and impacts within selected thematic ar- such as tax breaks for agricultural enterprises or eas, or they could be combined . These objectives relate discounts on input use such as water or fuel for to the breadth and depth of analysis to be undertaken. agricultural enterprises.8 ƒ Scope. Decide on the breadth (thematic and institu- tional coverage) and depth (level of detail proposed for • Depth: the analysis of flow of funds and impact) of analysis. (iii) Budget execution: Decide on the depth of • Breadth: analysis on budget execution. This could range (i) Thematic coverage: Decide on which subsectors from an assessment of simply out-turns (differ- to include. For a comprehensive sector-wide ences between budgeted, allocated, and actual expenditure review, it’s important to start by expenditures) to a more comprehensive assess- de�ning the parameters of the agriculture sector. ment of the level and timing of budget releases The United Nations Classi�cation of the Functions to individual local governments and tracking the of Government (COFOG) de�nes the agricultural flow of funds to end-users. sector to include crops, livestock, �shing, forestry, (iv) Impact evaluation: Decide on the extent of im- water-for-production, and issues related to agri- pact evaluation. This is often determined by the cultural land. The OECD Development Assistance existence of baseline data and usually takes the Committee uses the COFOG de�nition to form of impact evaluations of individual sector aggregate cross-country donor assistance to subprograms (e.g., extension, research, irriga- agriculture. Use of common de�nitions can allow tion, fertilizers). In addition, decide on whether for more accurate international comparisons of the aspect of environmental impact of agricul- public spending. In practice, many countries adopt tural programs (including carbon footprints) is their own de�nition of the sector, often aligned to warranted or of interest to government of�cials sector ministries or the composition of agricultural to include in the analysis. GDP. These country-speci�c de�nitions can also be used to guide the thematic scope. Narrower • Selectivity: Inclusion or exclusion of particular top- thematic expenditure reviews have often fo- ics will depend on numerous factors, including (i) cused on subsectors, types of spending (e.g., the policy priorities for agriculture, (ii) the speci�c fertilizer subsidies in Zambia and Philippines), operational requirements, (iii) the budget available or cross-cutting themes (e.g., natural resource to the task team to do the work, (iv) the timeframe management in Uganda). Gender could also be for the analysis, and (v) the data availability and analytical studies to support the work. PER objective is to provide a seal of good housekeeping or to re- assure external partners that their assistance is being put to good uses, the [development partner’s] . . . role is expected to be large and dominant. On the other hand, if the objective is to improve 8 Transfers from consumers to farmers arisen from border mea- service delivery, the client must lead and manage the process sures (e.g., import tariffs and other market support measures are unless a convincing case can be made otherwise.� not considered to be a part of budget and off-budget support). PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 8 PA RT II — PREPA R ATION PH AS E OF PER S ƒ Complementarities. Build on ongoing initiatives to colleagues (e.g., public sector management, local improve the ef�ciency and equity of public spend- government specialists). There are two main ques- ing. These include (i) broader cross-cutting efforts by tions to address: How can the agriculture PER con- the ministry of �nance to improve the overall public tribute to these broader efforts, and how can these budget alignment, (ii) investments in other sectors broader efforts address some of the public spending that are critical for agricultural development (e.g., issues in the agriculture sector? These questions rural roads and other infrastructure), (iii) �scal decen- can inform both the choice of the scope of the public tralization, (iv) mechanisms of �scal transfers to local expenditure analysis and the nature of the recom- governments, and (v) monitoring and accountability. mendations (see Example 3: Complementarities Dialogue on these issues is often led by other sector with Macro-Level PERs). 2.3: IDENTIFYING THE TYPES AND SOURCES OF DATA ƒ Be clear at the outset on data sources and availability moving organizationally in and out of the ministry of (on outputs, outcomes, and impacts) as they can con- agriculture and its public expenditure reporting and/or dition the scope and quality of analysis. It is helpful to when donor flows are off-budget. review some of the data while de�ning the objectives TABLE 1: Global Sources of Data and scope of the work. Together with government Public Spending http://www.imfstatistics.org/imf/ sources and other regional datasets, there are other IFSGover.htm sources of global data (Table 1). International Comparisons http://siteresources.worldbank.org/ ƒ Clarify the constraints to obtaining the data required, INTWDR2008/Resources/2795087- 1192112387976/WDR08_24_SWDI.pdf the options and likelihood for overcoming those con- Donor Spending http://www.oecd.org/dac/stats/ straints, and the subsequent impact on the choice of agriculture the methodological approach to data analysis. Sector Performance http://databank.worldbank.org/ddp ƒ Identify ways in which the study itself can help de- Prices http://www.fao.org/giews/pricetool/ velop an action plan to address the data constraints Agriculture Producer and http://www.oecd.org/agriculture/pse to allow for more in-depth future analysis. In some Customer Support Estimates cases, the expenditure analysis can help develop (for OECD and select non-OECD middle income countries) some of the data, such as a consistent time series of Agricultural Distortions Database, http://www.worldbank.org/ expenditures (and revenues, when relevant), when World Bank agdistortions subcomponents of the agricultural sector have been Household Surveys http://www.fao.org/es/ESA/riga 2.4: PREPARING THE CONCEPT NOTE ƒ The concept note for the PER will need to be tailored analysis is presented in Table 3 (see column on Types to the speci�c requirements, priorities, and type of ex- of Analysis). penditure analysis – reflective of the proposed breadth ƒ Terms of Reference will need to be developed for the and depth of analysis. Generic elements of a concept PER work contracted out, or for speci�c members note for a PER are reflected in Table 2. of a PER team focusing on various aspects of a PER ƒ The critical step at the concept stage is to identify the (see Example 4: Terms of Reference—Levels and right methodology for the given scope of the study. Composition of Spending and Example 5: Terms of The general range of methodologies and types of Reference—Budget Process and Performance). TOOLK IT 2. 5 : E S T I MAT I N G THE TASK BUDGE T AND T IME F RA ME 9 TABLE 2: Generic Elements of a Concept Note for an Agriculture Public Expenditure Review • Context • Resources/budget • Government request • Team composition (including non-agriculture public • Objectives and audience sector management specialists, country econo- • Synthesis of recent literature, link to past mists, and government counterparts) and ongoing activities, and value added by • Timetable the study • Peer reviewers • Scope and methodology of the study (the • Key annexes, such as a draft outline of the report, respective sections of the toolkit can help) with responsibilities; a brief overview of the agri- • Consultation plan with key stakeholders cultural sector (main trends and issues); a list of • Dissemination plan key references. 2.5: ESTIMATING THE TASK BUDGET AND TIME FRAME ƒ Make realistic estimates of the budget required with associated activities, such as training workshops. The the inevitable scaling of the scope of work to the bud- various instruments for supporting implementation of get available—ensure required quality standards and recommendations will likely involve other sources of take into account the availability and quality of data funding (e.g., from preparing an investment operation and cost-effective approaches. or budget support operation). Limited dissemination is • The cost of carrying out an agricultural expenditure one general area of weakness of previous PERs. analysis varies widely according to the type and ƒ Factor in sustainability considerations into the budget. scope of the analysis and the amount of data that If the agriculture PER is to become part of the institu- must be collected. A rapid or descriptive agri- tional processes (e.g., if it becomes a regular input to cultural expenditure review can cost from about the annual budgetary process), its design and costs US$50,000 to US$100,000. A more analytical should reflect the likely recurrent resources available agricultural public expenditure review could cost from the ministry of �nance or agriculture to conduct about US$250,000 or as much as US$350,000 for the expenditure analyses periodically. a comprehensive analytical and participatory ap- ƒ Explore options to augment resources, such as from proach (estimates are in 2010 dollars). other donors. The comprehensive PERs are usually undertaken with joint donor �nancing, while the more ƒ Consider options for improving cost effectiveness, rapid reviews are usually �nanced by a single donor. including the trade-offs of using international vs. local ƒ Make realistic estimates of the time required to consultants, sampling vs. more universal approaches undertake a public expenditure analysis. As with (particularly for decentralization issues and impact costs, there is a wide variation in the time taken to analysis), rigorous vs. more basic analytical methods, complete expenditure reviews that is dependent on and stand-alone vs. linked activities (e.g., linking policy the scope of the proposed work. The rapid review, dialogue with on-going projects or programs can im- as done in Lebanon, was completed over several prove cost effectiveness). months, while the more detailed analyses under- ƒ Don’t forget to factor in a dissemination budget, taken in Zambia and Uganda were completed over estimated to be about 15 percent to 20 percent of about a 2-year period. the overall cost. Given that the ultimate role of an ƒ Try to time the output of the public expenditure analy- agricultural expenditure analysis is to stimulate ac- sis with the budget cycle. This will hopefully allow tions to improve the focus, quality, and scaling of recommendations to directly feed into decisions in the public spending, it is highly desirable for the budget government’s annual planning and budgeting cycle. to include support for disseminating the report and for PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 10 PA RT III — A NA LY S IS A ND REC OMME ND ATIONS Part III: ANALYSIS AND RECOMMENDATIONS As recommendations from a PER are directed at improving 3. Budget Execution: After the budget vote and ap- the composition, scaling, and impact of public spending, a proval by parliament, funds are allocated to govern- useful framework for analysis is the budget cycle that led ment entities, and programs are implemented. to the current state of public spending on which improve- 4. Monitoring and evaluation: Financial and physi- ment is sought. Using a budget cycle framework helps focus cal reports are prepared by government entities recommendations on entry points to strengthen the budget receiving funds and submitted to the ministry of process and subsequent improvements to public spending. �nance. Relevance of sector objectives, ef�cacy A generic budget cycle includes four phases: (i) sector objec- of design and implementation, the ef�ciency of tive and strategy development, (ii) budget allocation, (iii) bud- resource use, and sustainability are evaluated. get execution, and (iv) monitoring and evaluation (Figure 2). Results inform future planning and budgetary processes. 1. Sector objectives and strategy: Includes setting of sector objectives and targets and formulation of as- Analysis of each phase of the budget cycle includes public- sociated strategies. The sector strategy is linked spending data collection, data analysis, understanding what to broader national development strategies. led to the current state of public spending, and recommenda- 2. Budget allocation: Budget allocations are made tions and potential entry points for improvement. This section against sector objectives and strategies within an of the toolkit provides guidance and checklists for the data overall public spending ceiling. Medium-term spend- requirements, types of analysis, areas of recommendations, ing projections are often made within a medium-term and potential entry points for each stage of the budget cycle expenditure framework. (Figure 3 and Tables 3 and 4). FIGURE 2: Generic Budget Cycle FIGURE 3: Public Spending Analysis Sector Data objectives requirements and strategy Types of analysis Monitoring Budget and allocation evaluation Areas of recommendations Budget Potential entry execution points Source: Authors. Source: Authors. TOOLK IT TABLE 3: Summary of PER Data Requirements, Types of Analysis, Areas of Recommendations, and Entry Points BUDGET DATA TYPES OF AREAS OF POTENTIAL CYCLE PUBLIC SPENDING ANALYSIS SCOPE / QUESTIONS REQUIREMENTS ANALYSIS RECOMMENDATION ENTRY POINTS INTENDED FOCUS OF PUBLIC SPENDING Sector Objectives Questions: What are the government’s national and sector Government Targets: Desk review Strategy Alignment Sector Reviews & Strategy objectives and strategies? Is the policy environment conducive (e.g.) Agricultural Growth, Performance Comparisons Potential Gains from Policy Strategy Updates to investment returns? Food Security, Improvements vs. Public Poverty Spending Reduction Strategy components Policies (e.g.) Net taxation PA RT I I I — A N A LY SIS AND RE COMME NDAT IONS ALLOCATIVE EFFICIENCY OF PUBLIC SPENDING Budget Allocation Scope: The broader the coverage, the greater the scope for Budget allocations: Simple Inductive Analysis Changes to Spending Ministry/Local Government analysis of allocative ef�ciency. Levels Allocations across Spending Budget Guidelines Questions: How can allocative ef�ciency of public spending be Functional Composition Categories Formulas Used in improved? (Is money being allocated to the right things?) What Development vs. Recurrent Simple Congruence Allocation Decisions aspects of the budget process have led to the current levels of Capital vs. Current Analysis Sector Review Links ef�ciency? Wage vs. Nonwage Public vs. Private Estimating Marginal National vs. Subnational Returns Across provinces/districts TECHNICAL EFFICIENCY OF PUBLIC SPENDING PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS Budget Execution Scope: The deeper the coverage, the greater the scope for Budget execution: Budgeted vs. Public Expenditure Tracking Budget Releases Ministry of Finance Budget analysis of technical ef�ciency. Allocated vs. Actual Cost Effectiveness Technical Design Management Reviews of Questions: How can technical ef�ciency of public spending be Timing of Releases Subprograms Procurement and improved? (Is implementation effective?) What aspects of the Indicators on: Quality of Services Ministry of Agriculture/ Fiduciary Capacity budget process have led to the current levels of ef�ciency? Local Government Guidelines for Project Appraisal Training and HR plans ACCOUNTABILITY OF PUBLIC SPENDING Monitoring & Questions: What indicators are used to measure performance? Coverage of Services Impact of Incidence Analysis Program Targeting Ministry/Local Government Evaluation Who is accountable for public spending outcomes, and to Services Impact Evaluation Implementation Budget Guidelines whom are they accountable? Formal Channels of Accountability Ministry M&E System Accountability Mechanisms to Stakeholders Source: Authors. 11 12 PA RT III — A NA LY S IS A ND REC OMME ND ATIONS TABLE 4: Selected Indicators for Agricultural PERs and Guidance on their Use KEY INDICATORS GUIDANCE ON THE USE OF INDICATOR RELATIVE SIZE OF AGRICULTURAL SPENDING Agricultural spending as a This indicator shows �scal costs of agricultural policy relative to the size of the overall economy. In most developing countries this share of GDP share averages 1–2 percent of GDP, similar to that in most OECD countries. Agricultural spending as a This indicator shows the dependence of agricultural GDP on public expenditure. The higher the share, the more dependent the sector share of agricultural GDP is on state support. An increase in this indicator over time may suggest a decline in the productivity of spending (expenditure growth is not matched by growth in agricultural value-added). In addition, the larger a country’s dependence on agriculture (e.g., lower-income countries), the less it can afford a high agriculture spending share of agriculture GDP relative to countries with a lower dependence on agriculture (e.g., developed economies). Thus, agricultural spending at 1 percent of GDP translates into higher effective support of farmers in developed countries compared to that in lower-income countries. Agricultural spending as a This indicator shows the �scal costs of agricultural policy. It is often used for comparison of relative allocations with other sectors, share of total budget reflecting the importance that governments give to agriculture relative to other sectors. Spending on agricultural This indicator is used for assessing an adequacy of allocations to agricultural research, which is vital for sustained productivity growth research as a share of and adjustments to climate change. The general target for low-income countries is to spend at least 1 percent of their agricultural GDP agricultural GDP on agricultural research, while for higher-income developing countries 3-4 percent is common. LEVELS OF AGRICULTURAL PUBLIC SPENDING Total sector budget, in nomi- It is important to consider both nominal and real indicators because in countries with high inflation, the growth of the nominal budget nal and real terms does not necessarily turn into growth of the real budget. Share of off-budget support in Off-budget expenditure generally refers to accounts of government transactions that are not included in budget totals or documents total sector budget and typically do not operate through normal budgetary execution procedures. Such transactions may, for example, be �nanced through foreign aid or earmarked revenues not included in the budget. Another example is agricultural tax exemptions. Share of decentralized sector This indicator shows the extent of decentralized spending, with implications for management, flows of funds, and program budget implementation. Share of foreign funds in total This indicator shows the dependence of the agriculture budget on donor �nance. A large share of donor funds, usually in the develop- sector budget ment budget, often indicates less predictability of budget levels if there are many small donor projects with different rules and/or the implementing agency has a weak capacity to comply with the donors’ �duciary requirements. Yet donor funds may bring many positive outcomes, including improved quality of investments; timely cash flows (particularly in the �rst months of the �nancial year when the government budget is not yet released but when funds are needed in agriculture); and strengthened capacity of ministries to implement government-�nanced programs, among many others. ECONOMIC COMPOSITION (ALLOCATIVE EFFICIENCY) Share of recurrent vs. devel- This indicator shows the composition of funds in the recurrent vs. development budget. The development budget in many develop- opment expenditure in total ing countries often comes from donors, while the recurrent budget is typically fully �nanced by the government. A large development sector budget budget share may indicate a priority toward productive investments vis-à-vis wages and operating expenses. Share of current vs. capital Not all development budgets are for capital expenditure as is often perceived. In many African countries, for example, a large share of expenditure in development the development budget is allocated to current spending such as staff allowances, inputs, maintenance, and operating expenses. This budget indicator therefore provides a true picture of the economic composition of the development budget. Share of current vs. capital This indicator shows a complete picture regarding the nature of the sector budget, distinguishing between current and capital spending, expenditure in total sector to be found in both recurrent and development budgets. budget Share of wages vs. nonwage This indicator shows the balance between wage and nonwage expenses. In many countries, the recurrent budget is often used almost expenditure in total sector entirely for staff wages, leaving too little to allow an effective execution of operational tasks of technical staff. budget Share of maintenance vs. There is no �xed rule on the ratio of maintenance to capital spending in capital-intensive expenditures such as rural roads and irriga- capital budget in total sector tion. It depends on the initial capital base. In Africa, which has severe underinvestment in infrastructure, the share of capital spending budget for new investments is expected to exceed that for maintenance. In Asia, where much capital investment in irrigation and roads has already been made, the dominant focus would likely be on maintaining the existing structures versus investing in new infrastructure. BUDGET EXECUTION Share of allocated funds to This indicator shows the share of released funds to an implementing ministry in comparison to what is approved in the budget at the approved funds beginning of the �nancial year. A widening gap may be the result of cuts across all ministries or of cuts that apply only to particular sec- tor ministries. Large deviations of fund releases weaken the link between planning and implementation and thus reduces the quality of the budget process. This indicator can be compared across different ministries to assess a relative effectiveness of agricultural sector ministries. Share of funds actually spent Not all released funds are actually spent. The changes in this indicator may demonstrate the absorption capacity of the ministry. relative to allocated funds However, if funds are only released in the last quarter of the �nancial year, even ministries with satisfactory absorption capacity may not necessarily be able to utilize all allocated funds in the shortened timeframe. Share of released funds in This indicator shows the timeliness of fund releases. In many developing countries, a small share of funds is released in the �rst two the �rst two quarters of the quarters, limiting the ability of ministries to provide meaningful support to critical seasonal activities such as crop planting. �nancial year Share of released funds in the This is an additional indicator to show the share of funds that are either highly likely to remain unspent or be wasted on unnecessary last quarter of the �nancial activities. year Source: Authors. TOOLK IT 3. 1 S E C T O R O B J E CT IVE AND ST RAT E GY 13 ƒ Data requirements: Includes agriculture sector per- ƒ Potential entry points: At speci�c points in the bud- formance and public-spending data. get cycle (e.g., revision of ministry budget guidelines, ƒ Types of analysis: Undertaken to assess allocative formulas used for allocations, etc.). and technical ef�ciency. ƒ Areas of recommendation: Agriculture sector objec- tives, budget allocation, program design, implementa- tion capacity, and accountability. 3.1: SECTOR OBJECTIVE AND STRATEGY Agriculture PERs should provide an assessment of (i) the to large farm sector accompanied by an extensive existing sector strategies (if any) (i.e., agriculture sector social safety net to compensate the losers and the strategy, rural development strategy, subsector strategies excluded, or earned incomes in a smallholder sec- such as water resources strategy, �sheries strategy, forestry tor that can compete in modern food markets and strategy, and also the link with higher level national develop- nontraditional exports. ment strategies) and (ii) the policy environment to see how conducive it is to yield returns on investments in the agricul- ƒ Assess the extent to which the strategies clearly ture sector. identify the roles of government in the sector, and provide a rationale for government involvement in the ƒ Determine whether the country objectives for the ag- sector (i.e., the types of things it will �nance and not ricultural sector are clearly stated. This often includes �nance). For example, this could include: speci�c sector growth rates, food security targets, • To address market failure—historically viewed as a employment generation potential, and expected con- situation where the pursuit of private interest does tributions to poverty reduction. Are these objectives not lead to an ef�cient use of society’s resources. reflected in both the national and sector strategy (are Market failures are often associated with asym- they consistent)? metric information, noncompetitive markets, high ƒ Assess the extent to which the national and sector transaction cost in coordinating activities with strategies are aligned to achieving the stated sector smallholders, negative externalities, and insuf- goals. What are the components of the agriculture �cient public goods. Public goods are the most strategy, and are they amendable to easy translation relevant example for public expenditure reviews; into public spending plans? What strategic choices given their characteristics being nonrival (con- (trade-offs) has the government made to achieve its sumption by one does not reduce the availability objectives? For example (from the WDR 2008), of consumption by another) or nonexcludable (no • For agriculture-based countries, the choice or poli- exclusion from using the good), these goods tend cy dilemma is often between addressing food se- to be underprovided by markets, leading to the curity directly by focusing on subsistence farming inferior social outcomes. through resilient farming systems and safety nets, • To improve equity—viewed as ensuring a fair distri- such as food aid, or by focusing on the more entre- bution of society’s goods to induce more equitable preneurial actors and favored areas that can secure and often pro-poor development. A good function- growth and deliver food security through cheaper ing market can facilitate higher production but food and better employment opportunities. does not always ensure equitable distribution. This • For transforming countries, the choice or policy di- provides justi�cation for many income redistribu- lemma is often the selection of the instruments to tion programs. address the rural-urban income disparity problem ƒ Determine the breadth of the sector strategy and the between subsidy transfers or investment to raise institutional arrangements for its implementation. Does rural households’ earned incomes. it include components beyond the ministry of agriculture • For urbanized economies, the choice or policy di- (e.g., rural roads, rural �nance, and education)? Which lemma is often between rapid growth in a medium government entities are involved in its implementation? PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 14 PA RT III — A NA LY S IS A ND REC OMME ND ATIONS TABLE 5: An Aggregate View as a Potential Starting Point—Some Examples AGRICULTURE PUBLIC AGRICULTURAL RATE OF SPENDING AS SHARE GROWTH RATE POVERTY OF TOTAL PUBLIC COUNTRY % (2002–06) REDUCTION SPENDING, % POTENTIAL AREA OF INVESTIGATION/ANALYSIS Nigeria 6.9 Low 6 High agricultural growth has not yet translated into higher rates of poverty reduc- tion. Areas of investigation could be (i) composition of spending (targeting), (ii) distributional impacts, and (iii) levels of spending. Uganda 4.5 High 4 Relatively high agricultural growth translated into high rates of poverty reduction. Areas of investigation could be (i) increasing spending on existing programs, and (ii) sources of potential ef�ciency gains. Zambia 1.9 Low 3 Current public spending patterns and policies have not translated into higher agricultural growth, and rates of poverty reduction remain low. Areas of investi- gation may need to be more comprehensive, including the policy environment and the composition and level of spending. Calculations derived from World Bank data. ƒ Review any documented implementation performance and poverty reduction against targets). The longer of the sector strategy over recent years. Has there the time period that implementation of the strategy been any progress in implementation? has been underway, the more comparisons can be ƒ Assess the extent to which the policy environment made. Compare sector performance against other is conducive to public investment returns in the country benchmarks. Often sector performance sector. may lag behind sector objective and strategy targets simply because the targets were overambitious, rather than because of poor relative performance (see 3.1.1 DATA REQUIREMENTS Example 6: Sector Objectives and International ƒ Recent sector performance data, including recent Performance Benchmarks). agriculture growth, food security, poverty reduction, ƒ If sector strategies do not exist, a rapid analysis of and exports. Sources of these data include national strengths, weaknesses, opportunities, and threats accounts, surveys, and Millennium Development Goal can be undertaken as a basis for assessing alignment data consolidated by the United Nations. of public spending in the PER. This is the approach ƒ Policy distortions data could include net taxation data taken in the Lebanon PER (see Example 7: SWOT (reflecting implicit and explicit taxation on the sec- Analysis). tor). For example, a source of these data is the recent ƒ An aggregate assessment can often provide guidance global update on policy distortions.9 In addition, the for the direction of subsequent analysis (Table 5), and rural sector performance scores developed by the some analysis of sector constraints can guide poten- International Fund for Agricultural Development (IFAD) tial areas of needed public investments. also provide indicators of sector policy.10 ƒ Assess the extent to which existing price policies are conducive to enhancing investment returns. A poor 3.1.2 TYPES OF ANALYSIS policy environment can erode the bene�ts of public ƒ Desk review of national and sector strategies. spending at a high cost to taxpayers. The assessment includes both macroeconomic policies (exchange ƒ Performance comparisons. Compare recent sector rates, interest rates, and trade policies) and sector performance against the sector objectives and strate- policies (taxes and subsidies). Much of this analysis gies (e.g., recent trends on growth, food security, will likely build on existing work (see Example 8: Declining Taxation of Agriculture). 9 Anderson, K (ed). 2009. Distortions to Agricultural Incentives: A Global Perspective, 1955 to 2007. London: Palgrave McMillian ƒ Assess the extent to which access to assets (es- and Washington DC, World Bank. The policy data base can be pecially land and water) and the protection of these found at: http://econ.worldbank.org/WBSITE/EXTERNAL/EXT- assets is conducive to investment returns. In Africa, of DEC/EXTRESEARCH/0,,contentMDK:21960058~pagePK:64214 825~piPK:64214943~theSitePK:469382,00.html. particular importance is the strength of land rights of 10 IFAD rural sector performance score: http://www.ifad.org/ women, which signi�cantly impacts on-farm invest- gbdocs/eb/92/e/EB-2007-92-R-46-Add-1.pdf. ment for food production. More broadly, essential is TOOLK IT 3. 2 : B U D G E T A L L OCAT ION 15 the design of water-right systems and those systems’ for its assessment of the ministry of agriculture’s impact on irrigation investment. All are important submission. for raising public investment returns. (Cross-country ƒ An implementable strategy. Sometimes agriculture indicators are provided by the components of IFAD’s strategies (due to their form and presentation) are rural policy score.) not easily translated into implementable programs ƒ Assess the effectiveness and technical objectivity (components and activities), often requiring ad- of systems the country has in place (if any) to carry ditional work and subsequent variants to get to out ex-ante feasibility evaluations of projects. How implementation. Recommendations on ensur- effective have these systems been at ensuring the ing an implementable strategy could help speed alignment of projects to the sector strategy and ob- implementation. jectives (i.e., criteria for selection)? Can the review ƒ Options to improve the macro and sector policy envi- process in place to screen investment proposals ronment. If the �ndings are that macro-policies (e.g., with line ministries against economic, technical, exchange rates, tariffs, interest rates) and/or sector social, environmental criteria be improved? Assess policy (e.g., taxes, land and water policies) severely the adequacy of these processes in ensuring inhibit returns to public investment, efforts should sound investment decisions and ways they can be perhaps focus on improving the policy environment improved. �rst before focusing on public spending. ƒ Appraisal processes of projects within the ministries of agriculture and local government: ways to improve 3.1.3 AREAS OF RECOMMENDATION the quality of the appraisal process. ƒ Realism of sector objectives. If current sector objec- tives are unrealistic compared to past performance, potential, capacity to implement, and available 3.1.4 POTENTIAL ENTRY POINTS resources (and any expected gains from ef�ciency and ƒ Sector reviews. A growing number of countries have policy improvements), it will be important for recom- annual sector reviews as part of the budget cycle to mendations to reflect these disconnects. review sector performance over the past year, impedi- ƒ Alignment of strategy components to achieve the ments to performance, and implications for public sector objectives. For example, some countries spending for the next year. This provides an important have an expressed agricultural growth objective with forum for debate and potential implementation of the ambitious growth targets, yet much of their strategy above recommendations. components and public spending are directed to sup- ƒ Strategy updates. These are done by countries period- porting subsistence livelihoods and safety nets. These ically (e.g., every �ve years or after new governments aggregate comparisons can provide useful insights on are elected). This may not coincide with the timing of (mis)alignment. an agriculture PER, but if it does, it offers another key ƒ Alignment of the sector and national strategy. These entry point for the recommendations. two strategies are sometimes not aligned, with the ƒ Ongoing policy dialogue. In many countries there is agriculture sector strategy being the basis for prepara- ongoing policy dialogue through other instruments, tion of the ministry of agriculture’s budget submission such as development-policy lending, which also pro- to the ministry of �nance, while the national strat- vides another important entry point for the recommen- egy is used by the ministry of �nance as the basis dation of policy and sector strategy. 3.2: BUDGET ALLOCATION The core part of PERs is on budget allocations and execution. scope), the greater the scope for analysis of its allocative ef- Analyses of allocations focus on how to improve the alloca- �ciency. This section highlights a checklist and range of ques- tive ef�ciency of public spending in the sector. The broader tions that could be helpful in guiding recommendations on the the coverage of an agriculture expenditure review (thematic scale and composition of public spending on agriculture. PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 16 PA RT III — A NA LY S IS A ND REC OMME ND ATIONS 3.2.1 DATA REQUIREMENTS include decomposition by (i) functional area (see Levels of Public Spending Example 13: Functional Allocations); (ii) economic area (e.g., development vs. recurrent)11, and within ƒ Determine how much is spent on agriculture by the the development budget, current vs. capital12 (see government as a share of its total spending and as a Example 14: Capital vs. Current), wage vs. nonwage percent of agricultural GDP. How have these shares (see Example 15: Wage vs. Nonwage), public vs. changed over the last 5 to 10 years? How do these private (see Example 16: Public vs. Private); and (iii) shares differ from other countries? How much of the to state-owned enterprises (see Example 17: State- spending on agriculture comes from outside of the Owned Enterprises). Reflect how spending compo- ministry of agriculture (e.g., special programs linked to sitions have changed over the last 5 to 10 years. It the president’s of�ce and tax exemptions)? Sources may be useful to cite some comparisons with other of data include the ministries of �nance and agricul- countries as possible benchmarks if they are available. ture (see Example 9: Agriculture Spending as a Sources of data include the ministries of �nance and Share of Total Spending and of GDP, Example 10: agriculture. Agriculture Spending Per Capita). ƒ Determine the functional and economic composition ƒ Determine how much is spent on agriculture by of public spending at the subnational level (govern- donors as a share of government spending and as ment and donors). Reflect how the spending composi- a share of overall donor spending in the country. tion has changed over the past 5 to 10 years. Sources How have these shares changed over the last 5 to of data include the ministry of �nance and the ministry 10 years? How do these shares differ from other of local government (or equivalent). countries? To what extent are these expenditures incorporated into the public budget and reflected in National and Subnational Composition of public spending numbers (see Example 11: Donor Public Spending Spending on Agriculture)? Sources of data include ƒ Estimate how much the government spends on the in-country donor group and the ministries of agriculture at international (e.g., on international com- �nance and agriculture. Also include NGO projects, mitments such as the WTO), regional (e.g., regional state-owned enterprises, and implicit subsidies. [supra-national] research), national, and subnational ƒ Estimate how much is spent by the private sector levels (see Example 18: National vs. Subnational). (very few PERs have estimated this; an attempt was How has this changed over the past 5 to 10 years? made in the Mozambique PER). One explicit objective What main factors have determined these changes? of public investment is to stimulate private invest- At the subnational level, what was the geographic ment as a means to achieving sector objectives. allocation of spending across provinces, districts, or Knowledge of the level of private investment would counties (see Example 19: Geographic Allocations, then be useful for assessing the ef�ciency of public Example 20: Sampling Subnational Governments, spending. Questions included are: how much private Example 21: Off-Budget Expenditures)? How are investment has there been in the agriculture sector these subnational allocations determined (e.g., based as a share of government spending, how have these on historical transfers or formula based)? Are there shares changed over the last 5 to 10 years, and how any off-budget expenditures? How large are these? do these shares differ from other countries. Various proxy measures will probably need to be used, given the general paucity and unreliability of data on private sector investment (see Example 12: Private Sector 11 Recurrent expenditure is spending on items that are consumed Investment in Agriculture). Sources of data include and only last a limited period of time. They are items that are used in provision of a good or service. In the case of the govern- foreign direct investment (FDI), enterprise surveys, ment, recurrent (and current when analyzing the development and household surveys (for on-farm investment). budget) expenditures include wages and salaries, operations and maintenance, and goods and services. Functional and Economic Composition of 12 Capital expenditure covers payments for the purchase or produc- tion of new or existing durable goods (i.e., goods with a life of Public Spending more than one year). This includes expenditure incurred for the ƒ Determine the composition of public spending at the acquisition of land and other physical assets, equipment, build- ings, and non�nancial assets with an expected lifetime of more national level (government and donors). Analysis could than one year. TOOLK IT 3. 2 : B U D G E T A L L O CAT ION 17 Sources of data include the ministries of �nance, agri- 3.2.2 TYPES OF ANALYSIS—ALLOCATIVE culture and local government (or equivalent). EFFICIENCY ƒ Estimate how much donors spend on agriculture at Once the level, composition, and �nancing of sectoral ex- national and subnational levels. What is the distribu- penditures have been identi�ed, the practitioner can begin tion of area-based programs (and the reasons for the to develop a better understanding of their performance and distribution)? How has this changed? How are these emerging impacts based on assessments of their ef�ciency determined? What is their relation to overall district as well as their impact on growth, food security, and poverty budgets? reduction. The focus of allocative ef�ciency analysis is on answering the question of whether public �nancing is being Gender and Other Compositions of Public Spending allocated for the right things. ƒ Estimate to the extent possible the share of public spending with explicit targeting (gender, poverty, crop Simple Inductive Analysis vs. livestock, subsistence vs. commercial, agronomic ƒ Identify any stand-out issues on allocative ef�ciency potential). Is there a difference in targeting between from a simple review and comparison of the data public and donor resources (see Example 22: Gender (e.g., having small capital budget allocations relative Disaggregation)? to recurrent expenditures are going to make it hard to Sources of Public Spending Finance recapitalize the sector). If there are no operation and A sound understanding of the sources of �nance and as- maintenance expenditures in nonwage recurrent spend- sociated funding and revenue-generating mechanisms, mo- ing, maintaining infrastructure (e.g., irrigation) will be dif- dalities, and processes is an important part of an analysis of �cult.13 Lack of integration of donor �nancing will make agricultural public expenditure. The following checklists and comprehensive spending trade-offs and planning more associated questions can help practitioners arrive at a better dif�cult (see Example 24: Simple Inductive Analysis). understanding of how public expenditures are �nanced and Simple Congruence Analysis of possible tax policies and trade-offs. ƒ To what extent have budget allocations been aligned ƒ Determine the �nancing sources of national and sub- to the sector objectives and strategy to achieve these national agricultural spending. For example, to what objectives (see Example 25: Simple Congruence extent do subnational governments rely on formula- Analysis)? For example, allocating most resources to based transfers from national governments versus subsistence livelihoods types of spending (from func- their own revenue generation? How has the �nancing tional analysis) will make achieving the sector growth source evolved over time? What is the scope for objective dif�cult. If the government’s objective is to subnational governments to generate their own increase household food security versus promoting resources (usually mandated in the Finance Act) to an overall agricultural growth, the congruence analysis ensure that decentralized services such as extension would need to consider measures for the increased are adequately �nanced? availability of food and improved access to food by ƒ Determine the extent to which service fees and insecure households, as well as measures to increase levies are used. Is there scope to increase their use the institutional coverage to ministries related to food for services, such as phyto-sanitary certi�cation or security and social safety nets. agricultural research, and/or to access key natural Estimating Marginal Returns resources, such as forests, water, and �sheries? Are ƒ Can the sector objectives be achieved more ef�cient- these fees too high, excluding access to services by ly by reallocating resources to other areas? More poor people (see Example 23: Cost Recovery)? What precise estimates to this question can be derived sectoral public entities (including SOEs) generate from estimates of marginal returns (see Example revenues? What are the amounts, trends, and relative 26: Efficiency Gains from Switching from Private importance of these revenues? What are the ministry of �nance’s rules for reporting, using, and transfer- ring these revenues? Are the rules followed? What 13 Maintenance is related to the performance of routine, preven- tive, predictive, scheduled, and unscheduled actions aimed at are the options for enhancing the revenue-generating preventing equipment failure or decline with the goal of increas- rules to bene�t both the treasury and the sector? ing ef�ciency, reliability, and safety. PRA C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 18 PA RT III — A NA LY S IS A ND REC OMME ND ATIONS to Public Goods, Example 27: Marginal Returns to ƒ Assess the extent of compliance with budgetary pro- Public Spending across Sectors). Estimates of margin- cesses. (i.e., do the in-country budgetary processes al returns to public spending across spending categories need improvements or just compliance to them?) are usually estimated across functional areas (Fan et al ƒ Assess the adequacy of institutional coordina- 2007) or across public-private categories. These types tion among government agencies involved in rural of marginal returns analyses typically require time series development and their impact on technical ef�ciency. data and are often done as stand-alone studies. Where ƒ Describe the political economy of public spend- such studies are not readily available for a given country, ing. What is the political and economic justi�cation the results from other countries and general knowledge for the main spending allocations? Who are the about what kinds of programs have the highest returns expected winners and losers of the current alloca- can be used for PERs. tions? What are the bargained compromises needed to improve allocative ef�ciency (see Example 29: Estimating How Much Public Spending Will Be Needed Understanding the Political Economy of Public ƒ Estimate how much spending is enough. Is the level of Spending)? sectoral budgetary allocations consistent with the gov- ƒ Assess the factors that have led to the lack of donor ernment’s stated priority for the agricultural sector, the funding integration into the budget process, the sector’s importance to the economy, and the govern- mechanisms in place to align donor support to the ment’s stated agricultural strategy (see Example 28: sector objectives and strategy, and how those factors Elasticity Approach)? What are the trade-offs with reduce the distortionary effects of the scale of donor other sectors if spending is increased? support to the subnational government relative to their ƒ Assess the (indicative) costs and bene�ts of raising subnational public budgets. public spending on agriculture. For example, what level of taxes, levies, and tariffs are generated by the sector relative to public spending? What are the trade-offs of 3.2.3 AREAS OF RECOMMENDATIONS these policy choices? (E.g., if agricultural households are ƒ Changes in annual allocations across spending catego- taxed, and these public �nances are spent on services ries and in the level of sector spending: this includes to these same households less the public administrative changes in functional, economic, and subnational cost to provide the service, is there a net bene�t? An allocations. example is the simultaneous use of import charges that ƒ Changes in medium-term allocations across spending raise fertilizer prices and subsequent fertilizer subsidies categories reflected in the medium-term expenditure to lower prices. Is this a better approach than simply framework (if used): in the absence of a formal medi- removing the import charges and subsidies?) um-term expenditure framework, recommendations Understand and Describe What Has Led to Current could include medium-term spending allocations for Levels of Allocative Ef�ciency the sector to foster strategic continuity in expendi- ƒ Describe the factors that led to the current spend- tures, formulated by the ministry of agriculture. ing patterns and recent changes at the national and ƒ Ways to improve consistency, predictability, subnational levels. and alignment of the source of public �nancing: ƒ Assess the strengths and weaknesses in budget Consistency—to ensure revenue generation efforts do preparation processes at the national and subnational not outweigh the bene�ts of subsequent investment levels, including approaches taken to ensure align- (i.e., efforts are consistent with long-term objectives ment of the budget with the sector objective and the of the sector). Predictability—to ensure medium-term associated strategy to achieve the objectives, the investment planning. Alignment—to ensure donor mechanism for assessing trade-offs, and the level of �nancing is focused on implementation of country transparency and participation. The budget planning programs (see Example 30: Agriculture Sector-Wide and formulation process could include establishment Approaches). of sector ceilings through medium-term expenditure ƒ Ways to improve institutional coordination for more frameworks, annual plans, midyear budget reviews, effective implementation of programs impacting and audits. It is useful to show a budgetary flow chart agriculture outcomes (see Example 31: Institutional and timetable describing the current system and iden- Coordination Mechanisms). tifying strengths and weaknesses. TOOLK IT 3. 3 : B U D G E T E X E C UT ION 19 3.2.4 POTENTIAL ENTRY POINTS Incentive Mechanisms to Improve Budget ƒ National ministry and subnational government budget Performance). guidelines: annual budget circular or guidelines to ƒ Medium-term expenditure frameworks: revised annu- be sent by the ministry of �nance to the ministry ally, medium-term expenditure frameworks, prepared of agriculture or to local governments. This could by the ministry of �nance, offer an important entry include an (i) assessment of trade-offs, (ii) linking point for recommendations to changes in the sector inputs to outputs to outcomes, and (iii) the way �scal spending ceilings. transfer ceilings are determined (see Example 32: ƒ Sector reviews: As with the previous phase (sector Specific Recommendations to Government Budget objectives and strategy), a growing number of coun- Guidelines). A common pitfall of annual work plans tries have annual sector reviews as part of the budget is that they are based on activities and fail to focus cycle to review sector performance over the past year, enough on strategic programs and measurable results impediments to performance, and implications for and outcomes, thus making it dif�cult to improve ex- public spending over the next year. This provides an penditure ef�ciencies and impacts (see Example 33: important forum for the above recommendations. 3.3: BUDGET EXECUTION Analyses of budget execution focus on how to improve the between recurrent and development budgets and technical ef�ciency of public spending in the sector (i.e., the between donor and government funds (see Example ef�ciency of budget system implementation and programs). 35: Timing of Releases). The deeper the coverage is of an agriculture expenditure review (of flow of funds and impact), the greater the scope Program Implementation for analysis of technical ef�ciency. This section highlights a ƒ Assess the institutional processes and implementa- checklist and range of questions and analytical approaches tion performance at national and subnational levels that could be helpful in guiding recommendation on technical (see Example 36: Approaches to Gathering District- ef�ciency. Level Data). 3.3.1 DATA REQUIREMENTS 3.3.2 TYPES OF ANALYSIS: TECHNICAL Budget vs. Actual Expenditures EFFICIENCY OF PUBLIC SPENDING ƒ Determine the extent of differences between actu- Public Expenditure Tracking Surveys ally spent, allocated and approved expenditures at the ƒ Follow the money. Use a public expenditure track- overall sectoral level and for major programs at both ing survey (PETS), which typically examines �nancial the national and subnational levels. How have these flows, facility characteristics, outputs, and accountabil- changed over the last 5 to 10 years? How do these ity arrangements. PETS data can have multiple uses, differences compare with other sectors (see Example ranging from a simple diagnostic tool for operations to 34: Budgeted vs. Actual)? Sources of data are primar- empirical research on capture and cost-ef�ciency. It ily from the ministry of �nance. examines resource use and leakages. A PETS mainly ƒ Estimate the extent of differences in planned dis- evaluates the proportion of public resources (�nancial, bursements and actual disbursements of donor proj- human, and in-kind) that reaches each level, in par- ects to the extent possible. The source of data is likely ticular frontline service providers. Analysis takes the to be donors, or in some cases the ministry of �nance. form of micro-level surveys (see Example 37: Public Expenditure Tracking Survey). Timing of Budget Releases ƒ Determine the timing of the budget release for ag- Cost Effectiveness riculture to the national ministries of agriculture and ƒ Estimate the cost effectiveness of programs. To what subnational governments. How predictable have these extent have �nanced programs been cost effective at budget releases been? Have they been aligned with delivering results? It may not be possible to assess the seasonal demand for public spending? Distinguish cost-effectiveness (or value for money) for all programs, PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 20 PA RT III — A NA LY S IS A ND REC OMME ND ATIONS so selecting major programs or projects can provide 3.3.3 AREAS OF RECOMMENDATIONS initial insights for a more comprehensive review in the Following on from the analysis, the areas of recommenda- future. What agricultural services does public spending tions will likely be related to buy? What are the unit costs of delivering key public services? Is there scope for reducing the unit costs of ƒ The budget process on ways to reduce the gap these services (see Example 38: Cost Effectiveness)? between budgeted and actual spending and on ways to ensure the timing of budget releases against Understand and Describe What Has Led to Current the seasonal demand for agricultural programs and Levels of Technical Ef�ciency services. ƒ If actual spending is lower than budgeted amounts, ƒ Technical design of programs to improve implementa- where did the money go? Were readjustments to al- tion and resource use ef�ciency. locations made after budget approval? Were budgeted ƒ Procurement and �duciary capacity, training needs for funds not released because of revenue shortfalls procurement and �nancial management, and perhaps (cash budgets)? Or was there simply a low use of staff recommendations at the national and local levels funds by ministries, so subsequent releases were to strengthen capacity. not made? Answering these will require understand- ƒ Ways to strengthen coordination mechanisms among ing the budget implementation process (budget sector ministries and across sectors, speci�cally when releases, reporting, cash budgeting) (see Example 39: coordinated action is needed to strengthen program Explaining Inefficiencies). outcomes. ƒ If the timing of budget releases was not as planned, why was that the case? Is cash budgeting the reason? Was it simply a delay in the fund transfer system? 3.3.4 POTENTIAL ENTRY POINTS ƒ Are any of these factors unique to the agriculture sec- ƒ Ministry of �nance budget management reviews tor, or are they prevalent across many sectors in the ƒ National ministry and subnational government project country? appraisal guidelines ƒ To what extent did the quality of �nancial manage- ƒ Ongoing public service reforms to ensure needed ment and procurement of implementing agencies staf�ng and training, both for improved technical impact technical ef�ciency? This could be assessed as quality and �nancial management and procurement part of a PETS or reviewing/drawing broader country ƒ Sector reviews. As with the previous phase (sector assessments of national and subnational capacities. objective and strategy), a growing number of coun- How timely and complete are �nancial manage- tries have annual sector reviews as part of the budget ment reports? How transparent is the procurement cycle to review sector performance, impediments to process? The assessments should identify ways to performance, and implications for public spending strengthen both the �nancial management and pro- over the next year. These provide important forums curement processes. for the above recommendations. 3.4: MONITORING AND EVALUATION Agriculture PERs should assess the coverage and impact of the extent possible. Data requirements for rigorous public spending (to the extent possible) and review the moni- impact evaluation are extensive and can be costly. toring and evaluation (M&E) systems in place. What indicators ƒ Performance-monitoring indicators used by govern- are being used to measure performance? Who is accountable ment at the national and subnational levels. for public spending outcomes, and to whom are they account- ƒ Formal channels of accountability to stakeholders. able (see Example 40: Monitoring and Evaluation)? 3.4.2 TYPES OF ANALYSIS 3.4.1 DATA REQUIREMENTS For analysis of incidence and impact of programs: ƒ Coverage and impact of agricultural investments and services. Rely on existing studies and surveys to TOOLK IT 3. 4 : M O N I T O R I N G AND E VAL UAT ION 21 Incidence Analysis accountability requirements and standards? Are there ƒ Measure incidence of public spending. Who’s being rewards for surpassing standards? reached by the agricultural service or investment? ƒ What managerial and institutional incentives are in Differentiate by income and gender. Has this changed place to generate better value for money, results, and over time? Analysis involves household surveys (see accountability? (Note that this issue also enters into Example 41: Incidence Analysis). the ef�ciency analysis, discussed later.) Do the as- sessments encompass the subnational level? How do Impact Evaluation they evaluate performance at the subnational level? ƒ Measure impact. What impact has public spending ƒ What mechanisms and sanctions exist to reduce cor- had on household income, on agricultural growth, and ruption? Special reference can be made, for example, on poverty reduction? This is the most data-intensive to the adequacy of procurement and �nancial manage- analysis given the controls needed to attribute impact ment procedures and associated checks and balances to speci�c government interventions rather than to (assessed in the previous section), various types of other factors (e.g., weather). Analysis often involves audits, the application and public transparency of the use of multiple sources of data (see Example 42: sound procurement processes and practices, the con- Impact Evaluation). certed enforcement of anticorruption measures and For analysis of the M&E system, which will typically take the sanctions, and corruption cases. form of systematically answering a series of questions: ƒ How are intra and interagency coordination arrange- ments, mechanisms, and incentives assessed to ƒ Is there a functional institutional or sectoral monitor- ensure coherent budgetary allocations, enhanced and ing and evaluation system that assesses key inputs, timely implementation of expenditures, avoidance of outputs, and intermediate outcomes of sectoral duplication of expenditures, alignment with national expenditures? Is there demand for an operational and sectoral strategies and policies, and clear, sound M&E system from senior management, such as the public investment priorities? ministers of �nance or agriculture? Can the current M&E system support a performance-based budgetary ƒ Does a functional, sound system monitor and evaluate allocation system? institutional and sectoral expenditures? To what extent is the resulting information used for decision making? ƒ Within national and subnational governments, are What underlying factors constrain managers’ demand managers of spending units (programs, project-man- for strong, effective M&E systems in the agricultural agement units) accountable for veri�able results and sector? An assessment will need to be done for each outputs (not just disbursements)? Are there systems relevant ministry, with a strong focus on the ministry to reward excellent performance? Are there sanc- of agriculture. tions for not meeting key standards and results? Does the legislature receive regular, timely reports on the performance of expenditures by the agricultural sector 3.4.3 AREAS OF RECOMMENDATIONS or institutions? Within the national system of account- On Improving Incidence and Impact ability to the legislature, how do the ministries in the ƒ Can the technical design of the underlying public agriculture sector compare to those in other sectors programs be improved to maximize impact per unit in complying with reporting requirements? What role of expenditure? Consideration includes decentraliza- does the legislature play in promoting more strategic tion, public-private partnerships (see Example 43: allocations, better implementation performance, and Public-Private Partnerships), matching grants to enhanced accountability and governance of public induce leverage (see Example 44: Matching Grants expenditures? to Leverage Resources), enhancing environmental ƒ Assess and apply where possible various transpar- services (see Example 45: Environmental Services ency and accountability mechanisms and tools inside and Climate Change), decoupled support (see and outside the government (such as M&E systems Example 46: Decoupling Agricultural Subsidies, and indicators, the �nancial management system, Example 47: The Case for and against Fertilizer timely annual accounts and audits, citizen-driven social Subsidies), and better targeting of underlying pro- audits and citizen report cards, and public expendi- grams, speci�cally those for women. ture tracking). Are there sanctions for failing to meet PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 22 PA RT III — A NA LY S IS A ND REC OMME ND ATIONS ƒ Ways to strengthen implementation. This may include ƒ More strategic and focused impact evaluation of recommendations for quarterly or semiannual port- public expenditures. The emphasis should be placed folio reviews for major and strategic programs and on improving monitoring of high-priority programs projects. It is useful to suggest simple templates that and projects, which will certainly contribute to more can focus the review on important strategic themes, ef�cient public expenditures. Even so, govern- key performance indicators, and progress in scaling ments generally neglect to improve evaluations of up strategic programs (potential or actual). Portfolio the impact of public expenditures, which require reviews can generate multiple bene�ts if they are data-intensive, analytical approaches. Many govern- managed effectively. For example, they can highlight ments avoid independent evaluations unless they the need for a functional M&E system as an instru- are required by donor agencies. The limited impact ment for enhancing the quality of the portfolio. evaluations that have been done also reflect the limited analytical capabilities within many ministries On the M&E System of agriculture, although this situation is beginning to ƒ Ways to ensure a functional, strategically oriented change. M&E system for the agricultural sector to (i) generate the disaggregated information required to ensure 3.4.4 POTENTIAL ENTRY POINTS more ef�cient and equitable sectoral allocations, (ii) improve the implementation of expenditure pro- ƒ Approaches currently used for targeting program grams, (iii) enhance governance and accountability services and investments (e.g., formula-based of expenditures (using various mechanisms) (see approaches), program mid-term implementation Example 48: Accountability Mechanisms), and reviews (iv) serve as a tool that facilitates learning-by-doing. ƒ Ministry and subnational guidelines to strengthen Expenditure data disaggregated by program, project, existing reporting requirements activity, or region are essential for more rigorous ƒ Sector reviews. Participation by a broader spectrum of expenditure analysis. A phased approach can be taken, stakeholders may improve voice and accountability. starting with the more strategic programs and projects. 3.5: CONSOLIDATING RECOMMENDATIONS OF THE PER ƒ Consolidate recommendations into one section. 1. Agriculture sector objectives and associated sector ƒ Recommendations should directly follow the analysis. strategy This sounds simple enough, but this has often not been 2. Budget allocation the case. Every agricultural public expenditure analysis, 3. Budget execution regardless of its scope, should offer useful inputs to 4. Monitoring and evaluation current and future decisions related to planning and budgetary management with the objective to improve ƒ Recommendations can stimulate their further formula- the quality and scaling of public spending in agriculture. tion and implementation by the relevant government ƒ Recommendations should (i) reflect country conditions, agencies, provided there is effective leadership by (ii) build on ongoing initiatives, (iii) separate short- and the ministry of agriculture. It is not the role of the long-term recommendations, (iv) separate recom- agriculture public expenditure analysis to complete mendations by institution or audience, and (v) ensure this process, but by providing clear direction, it appropriate sequencing. The spectrum of recommenda- can lend considerable impetus to the process. The tions follows primarily from the analysis of the budget essence of the recommendations should be promi- process and leads to the current levels of spending and nent in the action plan matrix or whatever format is ef�ciency. Following the structure of the earlier sections most likely to encourage the implementation of the of the toolkit, the recommendations would likely cover recommendations. TOOLK IT PA RT I V — D I S S E M INAT ION AND SUP P ORT F OR IMPLEMENTATION 23 Part IV: DISSEMINATION AND SUPPORT FOR IMPLEMENTATION ƒ The objective of the agricultural PER (including rapid, limited national budget are also heightening demand thematic, or comprehensive) is to contribute to im- for agriculture PERs and are another reason to use proved policies and implementation performance and their results. Efforts to institutionalize PERs as part of impacts in agricultural expenditures. The completion the planning and budgetary cycle will also help (see of a sound PER report or note is only the beginning of Example 49: Institutionalizing PERs). this process. All too often, the bene�ts of agriculture ƒ The following key questions and suggestions draw on PERs (and other expenditure review products) are good practices for realizing the full potential bene�ts not fully used after the �nal meeting or consultation of an agriculture PER. In disseminating the results of workshop has concluded. The result is an informative a PER and implementing its recommendations, the report (obtained at a relatively high cost) of underused approach taken and the coordination of activities will operational value. Increasingly, ministries of �nance be influenced by whether the sector PER is part of and donor agencies require ministries of agriculture a country-wide PER, by the quality and signi�cance and donor colleagues themselves to obtain a clearer of the analysis and recommendations, and by fac- assessment of strategic and expenditure priorities and tors speci�c to the country, sector, and individuals performance before allocating increased investments involved. These processes are infrequently document- (and donor assistance). This trend has heightened the ed. The suggestions that follow synthesize experi- demand for expenditure analyses and, more impor- ences from various practitioners in lieu of speci�c tantly, for fully using the results. Competing demands notes and examples. Currently there is more art than from many sectors for a share of an increasingly science in this phase of expenditure analysis. 4.1: DISSEMINATION: TO WHOM AND IN WHAT FORM? ƒ Who is the primary target audience? In other words, recommendations are available (see Example 50: who are the strategic champions of the agriculture Dissemination Strategy). PER report—the key actor or preferably several key ƒ Consistent with the participatory approach taken actors who are most committed and best placed from the outset, what are the most appropriate to implement the recommendations? If the target and effective approaches for disseminating and audience is not clearly identi�ed or apparent, what especially for discussing and debating the results strategies can be pursued to cultivate strategically of the expenditure analysis? What sequencing is placed champions in the course of the expenditure the most effective and appropriate for disseminat- work? For example, is there an entry point in the form ing the results? Several focus group discussions, of an expenditure issue that is important to a particu- followed by participatory workshops, can be very lar individual? As noted, it is vital to actively involve valuable with strategic stakeholders (at the local and the ministries of �nance and planning (or equivalent) regional level), including central and local govern- and the sectoral ministries (especially the ministry ment, the private sector, academia, donor agencies, of agriculture) from the outset and at key stages of and key civil society groups. It is important to ensure the study, especially when the initial �ndings and that the ministry of agriculture is the main technical PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 24 PA RT IV — D IS S EMINATION A ND SUPPORT FOR IMPLEM ENTATION counterpart and feels ownership of the consultation other agencies and ideally in close collaboration with process, as the ministry will have primary responsibil- the ministry of �nance). ity for implementing the key recommendations (with 4.2: IMPLEMENTATION OF RECOMMENDATIONS: IMPROVING PUBLIC SPENDING ƒ What are the implementation strategy and action plan ƒ What is the most effective capacity-building and for maximizing the operational bene�ts of the �nd- analytical support that donors can provide to enhance ings and recommendations, especially with respect the effectiveness and timeliness with which the ac- to government counterparts and other key stakehold- tion plan is implemented? Donors may be especially ers? What are the measurable indicators of success important in working closely with the agriculture PER in implementing key recommendations? It is desirable champions and possibly providing transitional funding to devote time and effort to develop the action plan to support some of the initiatives described earlier. in concert with the key government counterparts be- ƒ What cost-effective approaches and mechanisms cause their ownership is vital, especially when there can be used in periodically reviewing and discussing is a comprehensive agriculture PER. The action plan progress in implementing the action plan? Can these might warrant some technical assistance to support activities be undertaken, for example, as part of the implementation, including capacity building among key annual budgetary preparation and mid-year implemen- stakeholders. tation process, or as part of programmatic support ƒ What are the most appropriate implementation mech- to the sector, possibly supported by an agricultural anisms that can be worked out and used within the government or donor working group? country to promote and monitor the implementation ƒ How can the implementation phase be used strategi- of recommended actions? Build on existing intra and cally to lay the groundwork for an updated agriculture intersectoral coordination mechanisms and processes, PER, especially in the context of a sequential ap- and actively involve the ministry of �nance. It may be proach and the development of subsequent products? useful to integrate the action plan into annual planning How can the PER tasks be institutionalized as part of and budgeting, especially because these processes ongoing planning and budgetary processes? A change bring together staff from central and sectoral minis- in administration offers a timely opportunity for newly tries. Implementation mechanisms should focus on appointed of�cials to update the strategic and expen- the ministry of agriculture, given its pivotal role in sec- diture priorities of any existing agricultural strategy or tor development and coordination. expenditure review. ƒ How can other key stakeholders, especially broad- ƒ How can institutional memory be maintained? A based farmer groups or apex organizations, “ac- stronger institutional or sectoral M&E system and company� and support the implementation process appropriate incentives for evidenced-based decision and help provide accountability for implementing the making can help ensure a smooth transition when the agreed actions? Various mechanisms and tools, such government changes, and it can also encourage a cul- as PETS, social audits, and service quality expenditure ture of transparency, learning-by-doing, and continu- surveys, can facilitate their participation. ous innovation and improvement. TOOLK IT PA RT V: — E X A M P L E S 25 Part V: EXAMPLES Example 1: ANNUAL AND PERIODIC PER S An example of where a PER is undertaken annually is in tracking surveys like cost-ef�ciency, incidence analysis, and Tanzania. The annual PERs usually have twin objectives of impact evaluation). supporting the budget process and undertaking an exter- nal review of �scal developments (World Bank 2009e). A The annual PERs typically provide recommendations on Tanzania PER Working Group, comprising representatives (i) improving the overall alignment of spending with the from the Government of Tanzania, the World Bank, United national development strategy, (ii) functional and economic Nations agencies, other bilateral and multilateral donors, allocation (including capital versus recurrent), (iii) national research and academic institutions, and NGOs, determines versus local government allocations, and (iv) improving ele- the agenda (and focus) for the annual PER process, guides ments of the budget execution cycle (e.g., integrating donor and �nances the implementation of the agreed-upon work �nancing, stream-lining internal ministry procedures, and im- program, and reviews all outputs. The agriculture sector is proving cash management systems). They may also highlight included in this broader PER. areas which may need further, more detailed analysis (e.g., large expenditure items in ministry budgets such as strategic The annual PERs include a budget analysis (macro-�scal grains reserves and input programs). policy and macro-economic budget framework, budget al- location, budget execution) and a public �nance manage- The periodic PERs (e.g., World Bank 2010b) typically include ment assessment (budget processes and institutions). The a more detailed analysis on both allocative and technical ef- budget analysis primarily focuses on allocative ef�ciency �ciency (the latter often including public expenditure tracking by assessing budget alignment with Tanzania’s growth surveys, incidence analysis, and impact evaluation). These can and poverty reduction strategy, some assessment of the be useful in establishing clear baselines and needed directional functional allocations within sectors, and the allocation be- changes on key elements of budget alignment and implemen- tween central and local government. Analysis of technical tation. The key elements and directional changes can then be ef�ciency is usually less extensive, focusing on the time- tracked in the less detailed annual reviews. Recommendation liness and completeness of actual expenditures against of these periodic PERs often includes medium- to long- allocations (rather than more detailed public expenditure term strategic actions (rather than simply short-term annual action). Example 2: SECTOR SCOPE AND ANALYSIS A good starting point for de�ning and agreeing on the scope of agriculture. In many countries, the ministry of the envi- of the expenditure analysis is to determine what is meant by ronment and natural resources inevitably overlaps with the the agricultural sector in the particular context for the agricul- ministry of agriculture (as in watershed programs). Important ture public expenditure. The UN Classi�cation of Functions programs for agriculture may be managed by another min- of Government (COFOG) de�nes agriculture to include crops, istry or agency; for example, rural roads are handled by the livestock, �shing, forestry, water-for-production, and issues ministry of transport or local governments. Given the varying related to agricultural land. An increasing number of countries circumstance in each country, the precise scope of agricul- recognize the need to adjust their budgetary structures to ture expenditures that will be covered in the agriculture PER reflect broader de�nitions of the agricultural sector that may will ultimately need to be clari�ed, de�ned, and agreed upon extend beyond the mandated responsibilities of the ministry with the government. The analysis must also make an effort PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 26 PA RT V: — EX A M PLES to promote better coordination of policies and expenditure that the Bank include federal public expenditures on “ag- allocations among institutions. riculture and rural development� that coincided with the Ministry’s policy and budgetary structure and frameworks. Examples of agriculture PERs conducted in Uganda (World Thus the Agriculture PER also assessed nonagricultural Bank 2010b), Mexico (World Bank 2009a), and Nigeria (World public expenditures made in rural areas under the Special Bank 2008) illustrate some of the issues encountered when Concurrent Program for Sustainable Rural Development de�ning the scope of analysis. (Programa Especial Concurrente para el Desarrollo Rural The Uganda agriculture PER encountered three major issues Sostenible [PEC]). The program does not necessarily include when using the COFOG de�nition of agriculture: the budget for all activities pertaining to rural development, but this classi�cation was used for the agriculture PER be- ƒ Over the period covered by the analysis, forestry, cause the government had identi�ed it as the federal budget the use of water for agricultural production, and land for agriculture and rural development. When nonagricultural issues were under the purview of the Ministry of rural public expenditures are included in a particular agricul- Water, Lands, and the Environment. Land issues were ture PER, the conceptual and data challenges increase, and subsequently transferred to a new Ministry of Lands, comparisons with other countries are limited. Housing, and Urban Development, whereas forestry and water-for-production remained with the Ministry In the Nigeria agriculture PER, the �rst challenge faced by of Water and Environment. Identifying project expen- the study team was to de�ne agricultural spending with ditures on forestry and irrigation was not problematic; precision. After considering various de�nitions, the team the dif�culty lay in deciding what proportion of the decided to use a fairly restrictive de�nition: they would Ministries’ overheads to assign to agriculture. The is- review public spending in agriculture, as opposed to public sue was decided through an analysis of expenditures spending for agriculture. The more restrictive de�nition was under the most appropriate departments or programs, adopted for two reasons. First, there were time and funding namely land valuation and registration in the Ministry of constraints. Second, public spending in other sectors had Lands, Housing, and Urban Development and water-for- been reviewed recently through several other studies, so production in the Ministry of Water and Environment. it was important to avoid duplication of effort. The team ƒ Some agricultural projects fell under agencies other proceeded to de�ne the categories of public expenditure than the Ministry of Agriculture, such as the Of�ce of in agriculture based on three considerations: (i) de�nitions the Prime Minister; the National Planning Authority; of agricultural spending commonly found in the literature, the Ministry of Finance, Planning, and Economic (ii) the expenditure responsibilities of Nigeria’s Federal and Development; and the Ministry of Local Government. State Ministries of Agriculture, and (iii) the structure of the Some of these projects involved more than one sector Government of Nigeria’s expenditure and budget accounts. of the economy in the sense that they had nonagri- Based on these considerations, agriculture included the fol- cultural components (such as the building of feeder lowing expenditure categories: agricultural research, agricul- roads) and did not fall under the COFOG de�nition of tural extension and training, agricultural marketing, the sup- agriculture. ply and subsidization of agricultural inputs (seed, fertilizer, crop chemicals, and so forth), crop development, livestock ƒ A considerable amount of agricultural activity is carried development, �sheries, irrigation (to the extent undertaken out at the district level and lower in Uganda, some- by the Federal and State Ministries of Agriculture and lo- times through the partial use of conditional grants. This cal Departments of Agriculture), and food security. Forestry arrangement made it necessary to estimate the propor- and wildlife were initially considered but then eliminated tions of various grants used for agricultural purposes. because in Nigeria, these investments take place outside For the Mexico agriculture PER (2009), the Ministry of the Federal and State Ministries of Agriculture, with major Finance—the main source of demand for the study—requested implications for data collection. Example 3: COMPLEMENTARITIES WITH MACRO-LEVEL PERS Turkey provides an example of how a macro-level public productively in a later agricultural public expenditure task. The expenditure analysis and the resulting reforms were used resulting report, “Turkey: Policy and Investment Priorities for TOOLK IT EX A M P L E 4 : T E R M S OF RE F E RE NCE —L E VE L S AND C OMPOS ITION OF SPEND ING 27 Agricultural and Rural Development� (World Bank 2005a), il- classi�cation; (ii) enhancing the budget preparation process, lustrates how a sectoral study and the subsequent dialogue restoring its credibility with line ministries, and strengthen- can deepen the broader expenditure analysis already under- ing the capacity to formulate policies and budgets within a taken as part of a macro-level PER, sharpen strategic recom- medium-term perspective; and (iii) improving budget execu- mendations, and extend reforms in sectoral institutions. tion and �nancial accountability. In 2003, Turkey also enacted a Public Financial Management and Control (PFMC) Law. The The need for a macro-level PER in Turkey (Public Expenditure law provided the framework for structural and institutional and Institutional Review [World Bank 2001]) was reflected in reforms to upgrade public expenditure management. the severe �scal crisis in which public expenditures rose to about 3 percent of gross national product, well beyond the By 2005, several important initiatives in budgetary reform had target of 0.8 percent. The �ndings of the review highlighted been taken. Budget transparency was enhanced, facilitating major weaknesses in the management of public expendi- implementation; the linkages between policy, planning, and tures. Although the Government of Turkey had made prog- the budget were improved; changes were made in the role ress in improving budget systems, many de�ciencies identi- and size of public administration; and an initial rationalizing �ed were relevant to rural institutions for most of the period of public employment was undertaken. All of these macro- under review. Information on rural sector expenditures was level initiatives had positive effects on budget management fragmented and presented in a way that made it dif�cult to reforms and processes in the Ministry of Agriculture and analyze and interpret in a precise manner. other rural sector institutions, which probably would not have taken place without the reforms. The subsequent agriculture Following the macro-level PER, the government undertook expenditure analysis deepened the earlier macro-level analy- important steps toward (i) improving budget coverage and sis and sharpened the strategic recommendations. Example 4: TERMS OF REFERENCE—LEVELS AND COMPOSITION OF SPENDING Terms of Reference Used for the Uganda (World Bank ƒ Overall trends in allocations to the sector (going 2007a) Study on Levels and Composition of Spending back 20 years if feasible), in absolute terms and Context and objective: The study is part of a wider review of relative to other sectors, the overall budget, and public expenditure in the sector that will help identify which sector GDP types of expenditures are best for pro-poor growth. It will ƒ Functional and economic classi�cation of the agricul- provide a comprehensive assessment of public �nancing in tural sector budget and changes therein over time the agricultural sector, using the de�nition of the sector set ƒ The levels and trends of recurrent and development out in COFOG. The study will assist MAAIF in undertaking expenditures, clearly identifying the salary, operational further PER work in the sector and provide tools and meth- costs, and development components (which are ods to support this. often included in both the recurrent and development expenditure categories) Recipient: The Ministry of Agriculture, Animal Industry, and ƒ Subsectoral allocations against core functions as set Fisheries (MAAIF); Ministry of Finance, Economic Planning, out in MAAIF’s Development Strategy and Investment and Development (MFEPD); and the Plan for Modernization of Plan (DSIP) and the PMA core document and by com- Agriculture (PMA) Development Partners group. modity (crop, livestock, and �sheries) Scope: The study will look at current and historical patterns ƒ Allocations against policy priorities and principles set of allocation within the sector and the source of funding (gov- out in the Poverty Eradication Action Plan (PEAP), ernment and donor) and mode of �nancing (loans, grants, PMA, and other more recent policy statements, and so forth). About 37 percent of the annual budget goes including an assessment of expenditures from the directly to the districts, which under the decentralization pro- perspective of public and private sector roles and cess are responsible for providing most of the services. The functions in the delivery of goods and services analysis thus needs to be conducted both at the national or ƒ The scale and management of allocations to the sec- central level and at the district or local government levels. tor functions (from COFOG) made through central Among the issues to be reviewed are: government ministries and local government PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 28 PA RT V: — EX A M PLES ƒ Composition and performance of the agriculture bud- The consultants will work closely with MAAIF staff in the get by funding source: government versus external or Agricultural Planning Department to ensure that they donor, and external or donor �nancing by funding mo- dality (projects, sector budget support, general budget ƒ Become familiar with the tools, methods, and support, and so forth) approaches used in the study and are able to under- take similar analyses in the future ƒ Changes over time in the share of the agricultural bud- get spent at the center and transferred to districts ƒ Improve their capacity for setting sector priorities and managing the annual budget process. ƒ Estimation of the amount of off-budget expenditure going into the agricultural sector by activity (subsector) Example 5: TERMS OF REFERENCE—BUDGET PROCESS AND PERFORMANCE Terms of Reference Used for the Uganda (World Bank implicit variances and overall sector budget execution 2007a) Study on Levels and Composition of Spending by agency as an assessment of absorptive capacity Context and objective: The study is part of a wider review of and management issues public expenditure in the sector that will help identify which ƒ Allocations and expenditures against key policy priori- types of expenditure are best for pro-poor growth. It will pro- ties, along with a description of the process underly- vide a comprehensive assessment of public �nancing in the ing this prioritization. Are the linkages between objec- agricultural sector using the de�nition of the sector set out in tives, organizational functions, outputs, and available COFOG. The study will assist MAAIF in undertaking further resources adequate? PER work in the sector and provide tools and methods to ƒ Influence and impact or performance of differing fund- support this. ing sources (for example, projects versus earmarked budget support) on the allocation to agriculture and Recipient: The Ministry of Agriculture, Animal Industry and within agriculture and the alignment and execution of Fisheries (MAAIF); Ministry of Finance, Economic Planning, the budget against sector priorities and Development (MFEPD); and the Plan for Modernization of ƒ The capacity of the institutions in the sector to formu- Agriculture (PMA) Development Partners group. late and execute their budgets and make a convincing Scope: The study will critically review the formulation and exe- claim for scarce public resources cution of the budget by government agencies. It will assess the ƒ The links between and the impact of decentralization overall budgeting process and system, identify weaknesses, on the budget process, including the accountability of and suggest options and an implementation strategy for im- services provement, with a view to attaining long-term ef�ciency in pub- ƒ Political economic analysis of agricultural public expen- lic spending. Other speci�c issues to be addressed include: ditures and processes and an implementation strategy ƒ A critical assessment of the performance and effec- The consultants will work closely with MAAIF staff in the tiveness of the overall budget process—formulation; Agricultural Planning Department and with budget Vote hold- implementation; M&E of budgets, programs, outputs, ers to ensure that they and outcomes; identi�cation of potential value-for- money issues; and recommendations on how to ƒ Become familiar with the tools, methods, and ap- improve the process where necessary proaches used in the study and are able to undertake ƒ The agricultural sector budget execution and perfor- similar analyses in the future mance by item and agency, assessment of the utiliza- ƒ Improve their capacity for setting sector priorities and tion and operational ef�ciency of public funds, and managing the annual budget process reforms identi�cation of critical constraints to effective budget Method: The study will be undertaken by a team of con- execution sultants, based for the duration of the study in Kampala ƒ The trends in budget allocations to the sector, cor- and working under the overall guidance of the Agriculture responding releases and actual expenditures, and TOOLK IT EX A M P L E 6 : S E C T OR OBJ E CT IVE S AND INT E RNAT IONA L PER FORMA NCE BENCH MA R K S 29 Sector Working Group. Management of the work by MAAIF and the structure of the �nal report. The inception report will on a day-to-day basis will be undertaken by the Agricultural also highlight any design issues that need to be further dis- Planning Department in MAAIF. cussed by the Sector Working Group. A draft �nal report will be submitted to the Sector Working Group two days before Reporting: The consultants will report back to the Permanent the study team �nishes its work. Secretary, MAAIF. The consultants will provide an inception report to the Permanent Secretary within �ve days of the Time frame: The work will begin on March 19 and will run beginning of the work, setting out in detail the approach to for 11 weeks. be taken, key milestones in the delivery of the �nal report, Example 6: SECTOR OBJECTIVES AND INTERNATIONAL PERFORMANCE BENCHMARKS Comparators of sector objectives with other countries can and has remained there (Figure 4). Because some countries, help gauge realism. For example, some countries have set particularly in Africa, are starting from a relatively low base long-term annual growth targets for agricultural growth and can bene�t from more widespread adoption of existing at 10 percent. This is a high target when viewed from an technologies, growth of 5 percent annually seems achiev- historical and global perspective. For example, India has able with suf�cient and well-targeted public investment and rarely exceeded a 5-year average agricultural growth rate the maintenance of a supportive policy framework, including of over 5 percent. China achieved agricultural growth above measures aimed at increasing private sector investments in 5 percent following the 1978 reforms, but the rate subse- agriculture. However, sustained growth rates of 10 percent quently settled back to between 3 percent and 5 percent per annual have little-to-no historical precedent. FIGURE 4: Five-Percent Sustained Agricultural Growth Is an Ambitious Target Rolling 5-Year Agricultural GDP Growth (%) China India 11 11 9 9 7 7 5 5 3 3 1 1 -1 -1 1969 1972 1975 1978 1981 1984 1987 1990 1993 1996 1999 2002 2005 1969 1972 1975 1978 1981 1984 1987 1990 1993 1996 1999 2002 2005 Derived from World Development Indicators. Example 7: SWOT ANALYSIS The Lebanon Agriculture Public Expenditure Review (World Lebanon should be to (i) increase domestic market share and Bank 2010c) used an analysis of agriculture strengths, exports to Arab markets, (ii) consider exporting to Europe weaknesses, opportunities, and threats (SWOT) to identify as a secondary opportunity, and (iii) focus on food qual- comparative advantage of Lebanon agriculture to be used ity and safety. Public spending recommendations derived as the basis for recommendations for public expenditure from the analysis included options to reform the signi�cant alignments. The SWOT analysis (summarized below) sug- tobacco subsidy (as reflected in the complementary note gests the strategic objectives for the agricultural sector in on “Decoupling Income Support from Tobacco Production PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 30 PA RT V: — EX A M PLES in Lebanon: Challenges and Opportunities�), and for the agriculture through investments in water capacity. In the immediate term, (i) improve logistics by promoting private longer term it was suggested to consolidate agriculture ex- sector investment and by providing training; (ii) increase penditures into a single institution and create a directorate public investment in agricultural research and develop- that oversees and invests in food quality and safety. This ment to enhance productivity and to �nd innovative ways would reduce the current high levels of fragmentation. to improve food quality and safety, and (iii) expand irrigated A SWOT Analysis Indicates Lebanon Is Well-Suited for High-End Fresh Fruit and Vegetables (FFVs) and Agro-Processing Markets STRENGTHS WEAKNESSES � Close proximity to Gulf and European markets, which are net importers of fruits � Lacking food quality and safety standards and vegetables � High cost structure compared to other MENA countries � Large diaspora creates access points to new markets � Institutional fragmentation creates bottlenecks � Very positive name recognition in Gulf countries, particularly with “Lebanese � Poor logistics, particularly in timeliness of delivery Apples� � Water scarcity as a production constraint and a lack of adequate investments � Mediterranean climate allows for a long growing season and for crop diversity in irrigation infrastructure � FFVs are often handpicked and gathered in unpolluted areas and Lebanon’s juices � Weak marketing infrastructure are considered high quality, when made from 100 percent natural fruits and � High debt service vegetables � Lack of land-use planning and rapid urban encroachment over prime farm � Ready availability and accessibility to agricultural inputs, unlike in many other land developing countries � Highly fragmented land holdings and predominance of part-time farming � Access to cheap, seasonal agriculture labor from neighboring countries (mainly � Severe politicization of agricultural and rural development institutions, with Syria, but also from Egypt) agriculture programs and policies driven mainly by political considerations � Well-established food-canning industry with extensive markets in the Gulf, in the European Union, and in North America OPPORTUNITIES THREATS � Organic FFVs offer higher premiums than conventional products � Egypt, Syria, and Turkey compete in Lebanon’s major export markets; Jordan � High-end products offer more stable returns than low-end alternatives is an emerging threat � Food commodity prices have fallen considerably since the recent price shock, � Climate change resulting in lower food subsidies, which may enable realignment of public invest- � High vulnerability to future grain-price shocks ment in R&D and food quality and safety � Projected long-run increase in key input prices (i.e., petroleum and fertilizer) � Low penetration rate of MNA FFV exports to the European Union � Lack of political interest to push agriculture strategy forward � “Unfreezing� the agriculture free-trade agreement with Syria � Loss of most valuable coastal arable land due to urban encroachment and � Enhanced flow of goods across the Syrian border salinization � Production and packaging of high-value organic products (e.g., olive oil, vinegar, � Soil erosion in mountainous areas due to costly maintenance of terraces, soap) for the European Union haphazard construction, quarries, etc. � Integration of FFV with emerging eco-tourism � Loss of image of Lebanese FFV in Gulf markets due to uncontrolled problem � Investment in water resources, particularly in the poor areas of the south (e.g., of pesticide residues and more rigid import standards in the Gulf Hermel and Akkar) Source: World Bank 2010c. Example 8: DECLINING TAXATION OF AGRICULTURE The Uganda agriculture public expenditure (World Bank The Direct Rate of Assistance (DRA) to agriculture was 1 per- 2010b) review highlights the policy prerequisites for ef�cient cent during 2001–04, with zero support to exportable prod- public spending, indicating that Uganda has successfully ucts and 13 percent support to importable products through addressed these prerequisites by largely eliminating agricul- import tariffs (Table 6). By contrast, taxation of coffee and tural price distortions and that the current policy environment cotton was as high as 80 percent to 90 percent in the late is conducive to public investment returns. Using data from 1980s, but still 50 percent to 70 percent by the late 1990s, the recent global study on agricultural distortions, the PER inhibiting investment returns. shows that Uganda has shifted from agricultural taxation to agricultural support in recent years. TOOLK IT EX A M P L E 9 : A G R I CULT URE SP E NDING AS A SHARE OF TOTA L SPEND ING A ND OF GD P 31 TABLE 6: Distortion Indicators for Ugandan Agriculture, Five-Year Averages, % COMMODITY 1961–65 1966–70 1971–75 1976–80 1981–85 1986–90 1991–95 1996–2000 2001–04 Coffee −15 −39 −64 −89 −73 −74 −20 −2 −1 Cotton −13 −22 −53 −80 −47 −51 −4 0 0 Exportables DRA −11 −28 −59 −88 −66 −66 −8 −1 0 Rice 14 21 52 49 48 36 7 13 18 Importables DRA 16 22 52 81 48 54 14 14 13 CHANGING STATUS COMMODITIES Maize −1 8 10 17 0 –13 –4 7 0 Beans 11 –7 0 0 0 0 –4 4 0 DRA agriculture 0 −4 −8 −14 −7 −7 −1 2 1 DRA non-agric. 8 9 13 15 13 13 8 10 8 TRA agriculture −8 −13 −21 −29 −20 −19 −9 −9 −7 Source: Matthews et al. (2006). Example 9: AGRICULTURE SPENDING AS A SHARE OF TOTAL SPENDING AND OF GDP Examples of the presentation of these spending shares and as a share of GDP (Nigeria versus selected countries) include the Nigeria (World Bank 2008), Ukraine (Zorya 2006), (Table 7). In 2000, Nigeria’s agricultural expenditures ex- and Uganda (World Bank 2010b) PERs. They provide a histori- pressed as a share of total public spending was lower than cal snapshot of agricultural expenditures, normally in compari- that of all other African countries for which data were avail- son with other key variables (e.g., agricultural expenditures as able, and it was also substantially lower than the regional av- a percentage of total expenditures, as a percentage of agricul- erages for Asia and Latin America. In the same year, Nigeria’s tural GDP, and relative to other countries or comparators, and agricultural spending expressed as a share of agricultural wherever possible, disaggregated by tier of government). GDP was low, not only compared with the Latin American and Asian averages, but even when compared with the The Nigeria agriculture PER (World Bank 2008) presents Africa-wide average. The two sets of data, together, allow Nigeria’s agricultural public spending expressed as a share of making a cross-country comparison of agricultural spending total public spending (Nigeria versus selected comparators) expressed as a share of total public spending. TABLE 7: Agriculture Spending as a Share of Total Spending and of GDP: Nigeria vs. Selected Countries AGRICULTURE SPENDING AS A % OF TOTAL SPENDING AS A % OF GDP REGION/COUNTRY 1980 1990 2000 1980 1990 2000 Africa 6.42 5.15 4.05 7.40 5.44 5.71 Botswana 9.71 6.47 4.29 24.28 47.99 71.71 Burkina Faso 5.47 5.83 7.15 2.98 2.79 4.38 Cameroon 2.22 4.06 2.01 1.22 3.58 0.95 Côte d’Ivoire 3.40 2.97 1.46 4.17 2.24 1.08 Egypt 4.35 4.73 6.85 12.56 7.13 11.21 Ethiopia 8.38 4.91 6.55 1.49 4.05 6.59 Ghana 12.23 4.10 2.57 2.30 1.21 2.00 (Continued) PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 32 PA RT V: — EX A M PLES TABLE 7: Agriculture Spending as a Share of Total Spending and of GDP: Nigeria vs. Selected Countries (Continued) AGRICULTURE SPENDING AS A % OF TOTAL SPENDING AS A % OF GDP REGION/COUNTRY 1980 1990 2000 1980 1990 2000 Kenya 8.42 6.03 4.82 7.65 6.64 7.05 Malawi 10.17 11.1 4.30 8.97 7.34 4.53 Mali 8.31 2.33 4.84 3.70 1.69 3.93 Morocco 6.46 4.98 3.29 11.59 8.11 7.87 Nigeria 2.80 2.91 0.74 1.80 2.20 1.14 Togo 7.03 3.51 1.78 7.87 1.74 1.23 Tunisia 14.52 8.00 5.76 32.42 17.61 15.02 Uganda 32.55 3.91 4.00 2.80 0.86 2.38 Zambia 22.97 2.91 5.08 60.85 4.36 6.21 Zimbabwe 7.03 11.18 1.70 13.01 20.6 5.36 Asia 14.80 12.23 9.11 9.44 8.51 9.54 LAC 8.04 2.02 2.53 11.51 6.79 11.1 Total 11.25 7.90 6.95 10.76 8.04 9.34 Source: Database for Fan, Yu, and Saurkar (2008). Derived from Nigeria Agriculture PER (2008) (Tables 3 and 4). TABLE 8: Uganda Agriculture Sector Budget as a Share of National Budget and GDP (%), 2001/02–2008/09 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 Approved agriculture budget as share of 5.7 5.6 4.1 3.8 4.4 4.6 4.3 3.8 national budget Released agriculture budget as share of 8.2 5.5 3.8 3.3 4.8 n/a n/a n/a national budget Approved agriculture budget as share of GDP n/a n/a 1.5 1.4 1.5 1.7 1.6 1.6 Released agriculture budget as share of GDP n/a n/a 1.4 1.2 1.2 n/a n/a n/a Source: Ministry of Finance Planning and Economic Development; Ministry of Agriculture, Animal Industries and Fisheries. TABLE 9: International Comparison of Fiscal Transfers to Agriculture, Average for 2002–04 SHARE OF AGRICULTURAL SHARE OF AGRICULTURAL FISCAL EXPENDITURES IN AGRICULTURE AS A FISCAL EXPENDITURES IN GDP ADJUSTED TO THE SHARE OF GDP NATIONAL GDP SIZE OF AGRICULTURE REGION/COUNTRY A B B/A Ukraine (budget expenditures) 11.6% 1.3% 0.11 Ukraine (total �scal expenditures, including VAT expenditures) 11.6% 2.1% 0.18 HIGH-INCOME COUNTRIES Australia 3.0% 0.31% 0.10 Canada 2.3% 0.51% 0.22 European Union 2.3% 0.65% 0.28 United States of America 1.6% 0.73% 0.46 MIDDLE-INCOME COUNTRIES Turkey 13.0% 2.0% 0.15 Mexico 4.0% 0.7% 0.18 Venezuela 5.0% 0.5% 0.12 TOOLK IT EX A M P L E 1 0 : A G R ICULT URE SP E NDING P E R CAP ITA 33 TABLE 9: International Comparison of Fiscal Transfers to Agriculture, Average for 2002–04 (continued) SHARE OF AGRICULTURAL SHARE OF AGRICULTURAL FISCAL EXPENDITURES IN AGRICULTURE AS A FISCAL EXPENDITURES IN GDP ADJUSTED TO THE SHARE OF GDP NATIONAL GDP SIZE OF AGRICULTURE REGION/COUNTRY A B B/A China 15.0% 1.2% 0.08 Brazil 9.3% 0.7% 0.08 Russia 6.0% 0.95% 0.16 LOW-INCOME COUNTRIES Uganda 32% 1.5% 0.05 Tanzania 45% 1.2% 0.03 Ethiopia 44% 2.7% 0.06 Kenya 29% 1.3% 0.04 Source: OECD (2004), World Bank (2005a), Tangermann (2006), Zorya (2006), and World Bank (2006). Note: The reference year for Russia is 2003. The Uganda agricultural PER (World Bank 2010b) presents Ukraine’s Agriculture Fiscal Policy Analysis (Zorya 2006) historical trends of sector expenditures in the national bud- presents an informative international comparison of �scal get and GDP and shows that over time, the approved budget transfers to agriculture (Table 9). The comparison shows that for agriculture in relative terms has declined from a level that Ukraine’s �scal spending on agriculture is at the same level as was already quite moderate (Table 8). that of middle-income countries and even some high-income countries. This table was added to by the Uganda PER (World Bank 2010b). Example 10: AGRICULTURE SPENDING PER CAPITA Trends in per capita expenditure levels are useful for as- priorities (based on per capita income levels for each prov- sessing patterns and possible biases of agricultural spend- ince, as indicated in the Five-Year Plan) and shows a negative ing and evaluating whether expenditures are consistent correlation (−0.70) between per capita income and total per with stated policies. Examples from Turkey, Mexico, and capita spending by region, which means that more spending Guatemala analyze regional expenditure patterns on a per occurred in less developed regions (Table 10). The review capita basis and compare per capita expenditures across points out signi�cant organizational differences. There is a various countries. strong negative correlation between per capita spending for rural roads and per capita income; however, the Ministry of The Turkey Agriculture PER (World Bank 2005a) reports Agriculture and Rural Affairs’ per capita spending does not that total per capita spending is in line with planned regional establish such a strong negative correlation. On the other TABLE 10: Correlation of Per Capita GDP and Per Capita Organizational Spending, Turkey, 2002 CORRELATION BETWEEN PER CAPITA GDP AND PER CAPITA SPENDING 1 Per capita spending for Ministry of Agriculture and Rural Affairs services −0.30 2 Per capita spending for State Hydraulic Agency (irrigation services) −0.51 3 Per capita spending for General Directorate of Rural Services (rural roads services) −0.75 4 Per capita spending for General Directorate of Rural Services (drinking water services) −0.50 5 Per capita spending for GD (road services) −0.64 6 Per capita spending for Turkey Electricity Distribution Corporation (electricity services) 0.32 7 Per capita spending for Ministry of Environment and Forestry services 0.18 8 Per capita rural spending −0.70 Source: Turkey Agriculture PER (World Bank 2006e). PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 34 PA RT V: — EX A M PLES hand, the Ministry of Environment and Forestry and Turkey urban and rural areas, unlike many other Latin American Electricity Distribution Corporation per capita spending show countries, where an urban bias persists in the allocation of positive correlation between per capita GDP and per capita public expenditure. Mexico’s public expenditure in agriculture spending. is the highest in the Latin American and Caribbean Regions in terms of overall expenditure and of the size of the agricultural The Mexico Agriculture PER (World Bank 2009a) points out sector. that Mexico’s large Agriculture and Rural Development ex- penditure program represents a very signi�cant �scal effort A comparison of countries in Latin America in terms of the undertaken by the government in favor of the rural popula- historical average per capita agricultural expenditure (per rural tion. The review showed that the average public expenditure inhabitant, constant prices) is provided in Table 11. per capita (some US$700 per rural inhabitant) is similar in TABLE 11: Agricultural Expenditure Per Capita (in US$ Per Rural Inhabitant), LAC COUNTRY 1985–88 1989–92 1993–96 1997–01 AVERAGE Uruguay 82.07 162.83 450.15 622.71 346.69 Mexico 167.04 206.23 279.22 268.06 232.37 Venezuela 169.03 80.00 48.95 92.07 97.19 Brazil n/a 15.64 199.12 154.04 95.84 Costa Rica 89.41 93.01 102.90 71.57 88.19 Argentina 8.93 67.42 123.72 123.38 83.36 Chile 0.00 34.99 102.41 138.25 72.99 Panama 76.48 41.03 49.25 109.75 71.52 Dominican Republic 27.73 43.95 51.40 84.61 53.84 Nicaragua n/a 12.21 46.84 56.77 30.59 Guatemala 12.11 11.13 23.34 60.03 28.61 Peru n/a 18.23 46.73 44.18 28.28 Ecuador 9.38 10.16 25.36 32.88 20.24 Jamaica 15.56 12.31 17.81 23.68 17.71 Paraguay 6.90 11.76 24.71 24.23 17.33 Honduras 18.57 7.72 4.63 14.54 11.55 Bolivia 4.38 5.05 1.64 19.59 8.36 Average 40.45 49.04 94.01 114.14 76.75 Source: FAO LAC Regional Of�ce database from several countries (1985–2001) as referenced by R. Lopez (2004). Example 11: DONOR SPENDING ON AGRICULTURE Donor �nancing still accounts for a larger share of overall Most bilateral donors currently provide aid only in the form of spending in the agriculture sector, providing resources for grants. However, some larger donors also provide soft loans over half of the government budget in many countries. In ad- for agriculture, including Japan (US$708 million in 2008) and dition, countries often receive �nancing from multiple donors. France (US$102 million). Among multilateral agencies, the The OECD maintains a database of donor commitments by major loan provider was IDA (US$1 billion in 2008). Overall, donor, by country. These data show both levels and shares of 28 percent of aid to agriculture was in loan form in 2008. donor spending. Figure 5 shows the average shares of donor Development Assistance Committee (DAC) members’ loans spending in the period 2006-08 in Sub-Saharan Africa. went mainly to lower-middle-income countries (77 percent), TOOLK IT EX A M P L E 1 2 : P R I VAT E SE CT OR INVE ST ME NT IN AGRICULTUR E 35 FIGURE 5: Donor Shares in Total Donor Spending to whereas projects in least developed countries were practi- SSA, Average 2006–08 (%) cally all �nanced through grants (OECD 2010). IDA (28) However, when estimating donor �nancing at a country level, it is important to avoid double counting when aggregat- United States ing a donor’s budget and off-budget expenditures, especially (20) when a high proportion of the national budget is provided France (11) by donors. The Uganda agriculture PER (World Bank 2010b) EC (9) highlights the importance of gathering information also from Af DF (7) local institutions to take into account the increased role of Germany (7) nongovernmental organizations (NGOs) in speci�c areas of the country. It was estimated that the total value of the agri- Japan (6) cultural inputs supplied in the northern regions by NGOs was IFAD (5) equivalent to 10 percent of COFOG-based expenditures on Canada (4) programs and activities. Information on those programs was Norway (4) unavailable at the district or the country level. Derived from OECD data. Example 12: PRIVATE SECTOR INVESTMENT IN AGRICULTURE Private investment accounts for the majority of investment in in order to qualify for later repatriation of dividends the agricultural sector (in many countries, private investment and capital. makes up at least 70 percent of all investments, varying ac- 3. Commercial Bank Lending to Agriculture. A coun- cording to the prevailing policy regime and subsector). Very try’s central bank usually compiles regular data reports few countries compile and analyze data on private investments on commercial bank credit to the economy, with break- in the agricultural sector, especially considering that there are downs by sector, type of loan, and province. Net flow many factors involved (especially small-scale farmers, who are of lending can be calculated as the differences in credit the main private investors). The Mozambique agriculture PER outstanding at two points in time. Data is also often (World Bank 2009b) highlights six sources of data and discusses complied on gross new lending, repayments, and net several possible approaches to improving existing data systems lending by broad sector classi�cation and type of loan. to provide better information for monitoring how effectively 4. Private Investment in the National Accounts. public expenditures on agricultural development stimulate in- There are three ways of measuring GDP, one of vestment in agriculture (see the background paper “What are which includes estimating gross private capital the facts on private investment trends in agriculture?� prepared formation. However, most countries use the sum-of- for the Mozambique PER). The six data sources are: value-added method. 1. Authorized Investment Projects. Usually available 5. Agricultural Survey (Small- and Medium-Scale through investment promoting centers, they cover Producers). These surveys often provide a measure both foreign and national investments, often with a of on-farm investment (or capital stock), which can breakdown by sector, province, district, and country be tracked over time through repeated surveys. of origin, along with the proposed owner’s equity 6. Enterprise Surveys. Similarly, enterprise surveys can capital, the value of loans and supplementary capital, also be an important source of information on private and the expected number of jobs to be created. investment (capital formation) in agriculture, and input 2. Foreign Investment Inflows. Foreign capital inflows from the sector on questionnaire design is important. are usually compiled by a country’s central bank for The Vietnam PER (World Bank 2005b) provided estimates the balance of payment statistics. In the case of of private investments, showing that agriculture depended Mozambique, the data are obtained from documents heavily on the state budget (estimated to account for about �led by foreign investors, who must register inflows 57 percent of total investment in the sector between 1999 PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 36 PA RT V: — EX A M PLES and 2002), owing to the relatively small level of private en- There have also been efforts to measure private investment terprise (nonhousehold) investment (which was 40 percent in the agriculture sector in India, including a recent study of the total investment, on average). Foreign direct invest- by Golait and Lokare (2008) at the Reserve Bank of India. ment (FDI) in agriculture accounted for a small share— Note that the estimate in Table 12 is relative to public and only 3 percent of investment capital over the study period, private investment, the former not being the total public much smaller than the 17 percent for the whole economy. spending (excluding expenditures on items like subsidies Nonstate local enterprises have a similarly small presence. and maintenance). TABLE 12: Trends in Investment in Agriculture and the Overall Economy AVERAGE ANNUAL INVESTMENT AVERAGE ANNUAL INVESTMENT IN THE SHARE OF IN AGRICULTURE OVERALL ECONOMY AGRICULTURE IN TOTAL DECADE/YEAR PUBLIC PRIVATE TOTAL PUBLIC PRIVATE TOTAL INVESTMENT, % 1950s n/a n/a 4,370 n/a n/a 25,508 17.9 1960s 2,904 3,929 6,833 21,281 27,577 48,858 13.9 1970s 4,851 7,297 12,149 33,511 44,690 78,201 15.3 1980s 6,443 7,840 14,283 57,539 71,914 1,29,454 11.4 1990s 4,837 12,299 17,136 74,265 1,48,627 2,22,892 7.9 2000–06 5,237 17,184 22,387 85,327 2,27,216 3,12,543 7.4 2000–01 4,435 15,574 19,809 81,718 1,80,428 2,62,146 7.6 2001–02 5,488 14,872 20,360 82,824 1,68,840 2,51,664 8.1 2002–03 4,760 16,740 21,500 75,469 1,64,485 2,39,954 9.0 2003–04 5,699 18,487 24,186 82,998 2,04,946 2,87,944 8.4 2004–05 4,832 18,028 22,860 87,311 2,93,569 3,80,880 6.0 2005–06 6,206 19,400 25,606 1,01,640 3,51,028 4,52,668 5.7 Source: Adapted from Golait and Lokare (2008). Data computed from National Accounts Statistics and Agricultural Statistics at a Glance, Ministry of Agriculture, Government of India. The Golati and Lokare paper can be found at http://rbidocs.rbi.org.in/rdocs/Content/PDFs/cap_ad.pdf. Example 13: FUNCTIONAL ALLOCATIONS Functional composition of sector expenditure provides an addi- reduction. This �rst step is determining the functional com- tional means of understanding allocative ef�ciency in terms of position of spending. Valdes (2008a) provides estimates on priorities, level, and balance. Public funds should be allocated the composition of general services and supports estimates in to obtain the highest social payoffs for growth and poverty selected countries using various functional aggregates (Table 13). TABLE 13: Composition of General Services Support Estimates in Selected Countries (% avg. 2004–05) UNITED EUROPEAN NEW ITEM MEXICO BRAZIL CHILE STATES UNION ZEALAND SWITZERLAND Research and development 16.2 30.7 23.0 5.9 16.4 40.3 18.1 Agricultural schools 21.5 13.7 1.1 0.0 7.8 8.6 3.7 Inspection services 16.5 3.7 9.3 2.6 6.0 32.3 2.4 Infrastructure 13.8 44.5 55.2 14.0 42.4 18.4 18.1 Marketing and promotion 31.2 0.4 11.0 70.5 19.5 0.0 11.2 Public stockholding 0.0 6.9 0.0 0.4 7.3 0.0 8.6 Miscellaneous 0.7 0.0 0.8 6.6 0.6 0.2 37.9 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Source: Valdes (2008a) estimates, based on OECD PSE/CSE database, 2005 and 2007. TOOLK IT EX A M P L E 1 4 : C A P I TAL VS. CURRE NT 37 TABLE 14: Functional Composition of the Uganda Agriculture Budget (%, 2005/06–2007/08) BUDGET ALLOCATIONS AVERAGE OVER THE 2005/06 2006/07 2007/08 PERIOD Research 17 19 23 20 Advisory services 30 45 41 39 Livestock disease 9 7 4 7 Plant pests and diseases 1 1 0 1 Livestock and �sh regulatory services 2 2 2 2 Planning and policy 2 2 1 2 Institutional development 4 1 0 2 Water capacity building 3 4 4 4 Seed capacity development 9 3 3 5 Processing and marketing 7 2 2 3 Physical infrastructure 12 14 18 15 Promotion 3 1 1 2 Source: Uganda PER (World Bank 2010b). Another example is the functional composition in the Uganda payments based on (i) outputs, (ii) input use (variable input use, agricultural PER (World Bank 2010b). The PER presents a �xed capital, and on-farm services), (iii) current area planted or broad functional composition of sector expenditure and animals numbers (coupled support), (iv) noncurrent area planted changes in these over time (Table 14). (decoupled support), and (v) income support. The general sup- port services include (i) agricultural research, (ii) training and The public expenditure can also be grouped into the categories technical assistance, (iii) extension services, (iv) rural infrastruc- used by OECD for constructing its Producer Support Estimates. ture, (v) marketing, (vi) public stocks, and (vii) inspections and The expenditures are divided into payments to producers and veterinary services The database for all OECD and selected general sector support. The payments to producers include the non-OECD countries can be accessed at www.oecd.org. Example 14: CAPITAL VS. CURRENT Development expenditure is often falsely assumed to consist expenditure makes up about 80 percent of the agriculture of only or mainly capital spending. In many developing coun- budget in Uganda but is heavily oriented toward nonwage re- tries, especially in Africa, the development budget includes current expenditures rather than toward capital expenditures little capital spending but mainly nonwage current spending. An (Table 15). Often the clear split between capital and current economic decomposition of public spending in the Uganda ag- spending is not directly available in the ministry budget, and ad- riculture PER (World Bank 2010b) illustrates that development ditional estimates are needed to ensure accurate calculations. TABLE 15: Economic Composition of MAAIF’s Budget (Excluding Grants and Domestic Arrears), 2005/06–2008/09 (Ush billions) 2005/06 2006/07 2007/08 2008/09 REC DEV TOTAL REC DEV TOTAL REC DEV TOTAL REC DEV TOTAL Wage bill 2.5 6.4 8.9 2.5 8.3 10.8 2.5 7.7 10.2 2.6 5.3 7.9 Nonwage current 1.9 41.3 43.2 2.3 39.2 41.5 4.0 45.6 49.6 4.3 35.3 39.6 Capital outlays 0.0 17.2 17.2 0.0 11.0 11.0 0.0 19.4 19.4 0.0 13.2 13.2 MAAIF budget 4.4 64.9 69.3 4.8 58.5 63.3 6.5 72.7 79.2 6.9 53.8 60.7 Rec = recurrent, Dev = development Source: Uganda agriculture PER Task Team estimate, based on MoFPED (various years). PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 38 PA RT V: — EX A M PLES TABLE 16: Economic Composition of Agricultural Sector Budget (Excluding Local Governments) in Tanzania, 2010/11 (Tsh Billions) DEVELOPMENT RECURRENT TOTAL % OF TOTAL CURRENT Wage bill 9.2 22.7 31.9 9.0 Goods & services 49.7 121.5 170.9 48.4 Maintenance 7.9 1.3 9.2 2.6 Current transfers 51.5 58.2 109.7 31.0 Total current expenditure 118.1 203.5 321.7 91.1 CAPITAL Infrastructure 8.9 0.1 9.4 2.6 Equipment 6.8 1.4 8.2 2.3 Other capital 7.0 2.5 9.5 2.7 Studies 4.0 0.5 4.5 1.3 Total capital expenditure 26.7 4.9 31.6 8.9 Total 144.8 208.4 353.2 100.0 Source: Zorya and Francken (2010). TABLE 17: Share of O&M in Recurrent and Total Expenditure in Agriculture O&M O&M (AS % OF RECURRENT AGRICULTURAL (AS % OF TOTAL AGRICULTURAL COUNTRY EXPENDITURE) EXPENDITURE) Ethiopia (2002–06) 34 13 Kenya (2002/03) 25 20 Ghana (2002–04) 11 6 Turkey (1999–2003) 32 2 Vietnam (1997–2002) 25 6 Lao PDR (2002–04) 38 2 Source: Ghana and Kenya from Akroyd and Smith (2007); remaining countries from PERs. A similar situation is found in the Tanzania agriculture PER An important recurrent spending item linked to maintaining (Zorya and Francken 2010), where in 2010/11, the capital capital investment is spending on operations and mainte- spending made up only 18 percent of the development bud- nance. There is a wide variation in the share of O&M spend- get and 9 percent of the total sector budget, compared to ing across countries shown in Table 17, reflecting differ- a 40 percent share of the development budget in the total ences in sector priorities as well as perhaps inconsistencies sector budget (Table 16). in reporting expenditures across countries (as in the Uganda development expenditure above). Example 15: WAGE VS. NONWAGE Further disaggregation of current spending into wage and non- grew from 49 percent of the nonwage budget in 2005/06 to wage spending can provide more insights into spending ef- 80 percent in 2008/09 (Table 18). Similarly, disaggregation of �ciency. Following the Uganda example (World Bank 2010b), the capital outlays provides some interesting insights. The Table 18 provides the wage and nonwage decomposition largest infrastructure elements are livestock and wholesale of the development budget. Goods and services (compris- markets, rural communal roads and bridges, and machinery ing seeds, fertilizers, pesticides, and other agricultural input) and equipment. Investment in irrigation was extremely low. TOOLK IT EX A MP L E 1 6 : P U B LIC VS. P RIVAT E 39 TABLE 18: Decomposition of Wage vs. Nonwage MAAIF Budget, 2005/06–2008/09 (Ush Billions) 2005/06 2006/07 2007/08 2008/09 AVERAGE % SHARE Wage bill 6.41 8.30 7.70 5.30 6.93 11.1% Salaries 0.90 0.99 1.16 0.80 0.96 1.5% Allowances 5.51 7.21 6.44 4.50 5.97 9.3% Nonwage current 41.30 39.16 45.58 35.30 40.33 64.6% Goods and services 20.39 18.47 30.63 28.14 24.41 39.1% Medical and veterinary supplies 3.04 1.72 0.80 0.68 1.56 2.5% Capital outlays 17.17 11.03 19.41 13.23 15.21 24.3% Nonresidential buildings 8.16 4.47 10.03 4.75 6.85 11.0% Other structures (markets) n/a 2.51 5.93 5.30 4.58 7.3% Machinery and equipment 2.21 0.25 0.28 0.02 0.70 1.1% Livestock 0.64 0.30 n/a n/a 0.23 0.4% Irrigation n/a 0.28 0.20 0.01 0.12 0.2% Roads and bridges 4.98 2.70 0.90 0.69 2.31 3.7% Land 1.18 n/a 2.00 1.15 1.15 1.8% Development budget 64.87 58.49 72.69 53.83 62.47 100.0% Source: World Bank 2010b AgPER Task Team estimate, based on MoFPED (various years). Example 16: PUBLIC VS. PRIVATE Many agriculture PERs have not classi�ed expenditures as of public goods (social, productive, and indirectly public or private goods, thereby bypassing an important productive). Included here are programs typically opportunity to engage government planners and policy provided by governments in market economies, makers in an important policy discussion on the appropri- generally because they supply goods or services ate and evolving role of the state and the corresponding whose consumption is neither excludable nor emphases on public and private goods in public expendi- rivalrous. These programs usually include invest- tures. A practical issue in classifying expenditures by type ments like rural infrastructure (excluding on-farm of good is that the PER team must obtain data that are structures) and general social or productive ser- disaggregated to the activity level to make more precise vices such as rural education and health, telecom- classi�cations while recognizing that if disaggregated data munications, information, training and research are not available, expenditures may have to be attributed systems, technology transfer to small producers, to more than one expenditure category. Below are two sanitary systems, natural resource conservation examples drawn from the Mexico (World Bank 2009a) and and environmental programs, emergency pro- Honduras (Anson and Zegarra 2008) agricultural public ex- grams, and operational budgets of institutions. penditure reviews that classi�ed spending categories by 2. Private goods: (see page 16, Table 2–3, of Mexico public and private. PER) presents the breakdown of PEC expenditures according to private goods programs, again broken The Mexico agriculture public expenditure (World Bank down into three subtypes (social, productive, and in- 2009a) presents two tables that classify expenditures along directly productive). Included here are programs that public and private goods: are usually provided (or could be provided) by the 1. Public goods: (see page 15, Table 2–2, of Mexico private sector in market economies (mostly because PER) presents a breakdown of these programs they supply goods or services whose consumption in the Programa Especial Concurrente para el is excludable and rivalrous) and also government Desarrollo Rural Sostenible (PEC) (Mexico’s public programs that provide measurable subsidies to expenditure program for ARD, funded by the individuals or families usually under some targeting federal government) according to three subtypes criteria. Included in this category are, among other PRA C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 40 PA RT V: — EX A M PLES things, subsidies to on-farm productive infrastruc- The Honduras agriculture public expenditure program (Anson ture and equipment, subsidies to rural credit and and Zegarra 2008) classi�ed expenditures using the three- agricultural insurance, input support, marketing category typology of agricultural expenditures employed by support, and compensatory cash transfers. It should FAO in a study covering 16 Latin American countries (Table not be assumed that government funding of private 19). It grouped items of agricultural public expenditure into goods is necessarily undesirable. There are private three categories: private goods, public goods, and equity and goods that are productivity enhancing, and it may poverty-reduction-oriented goods. Some items of expendi- make sense for government to invest in them for ture were assigned to more than one category. For instance, policy purposes, even if they consist of measur- irrigation was classi�ed as a private and public good (with able subsidies to individuals or families. This is, 50 percent of expenditures in each category). Integrated for instance, the case with Oportunidades, which rural development expenditures were classi�ed as equity- enhances human capital, and with many Alianza enhancing (75 percent) and as a private good (25 percent). subsidies, which enhance the productive potential These are some of the judgments that must be made for this of farms. empirical classi�cation. The classi�cation is as follows: TABLE 19: Classi�cation of Public Expenditure by Private, Public, and Equity and Poverty-Reduction Oriented PRIVATE GOODS PUBLIC GOODS EQUITY AND POVERTY-REDUCTION ORIENTED � Commercialization � Training � Targeted rural productive promotion (50%) � Production promotion � Communication and information services � Integrated rural development programs (75%) � Forestry promotion � Soil and natural resources conservation � Public services in rural areas � Fishing and aquaculture promotion � Rural electri�cation � Social infrastructure for rural communities � Targeted rural productive promotion (50%) � Irrigation (50%) � Promotion of community groups � Irrigation (50%) � Land programs (agrarian reform) � Promotion of ethnic groups � Integrated rural development programs (25%) � Associative promotion � Promotion of women � Housing � Promotion of rural families � Water rights regulation � Basic sanitary conditions in rural areas (water and � Agricultural property regularization sanitation) � Education � Water for rural communities � Research � Health and nutrition � Justice � Rural roads � Phyto- and zoo-sanitation � Recreation and sports Source: Anson and Zegarra 2008. Example 17: STATE-OWNED ENTERPRISES It has generally proven successful for the public sector to in Vietnam, where a large number of SOEs are active in the sharply reduce its expenditure allocations to state-owned agricultural sector. The Vietnam Agriculture PER Integrated enterprises (SOEs) unless they perform a public function or Fiduciary Assessment (World Bank 2005b) reports that 319 respond to genuine market failures. Governments are there- SOEs, excluding irrigation-management companies and state fore substantially curtailing public funding for SOEs, espe- forestry enterprises, are active in the agricultural sector. Of cially for revenue-earning entities, which have increasingly this total, just 28 provide public goods. The remainder offer been privatized. services in exchange for fees from users; of these, 171 are considered pro�table, 43 just manage to cover their costs, Expenditure allocations to SOEs tend to have a double- and 105 are unpro�table. The latter are supported by loans displacement effect: they displace expansion of the private from state-owned commercial banks and allocations from sector and markets; and they displace expenditures on vital the state budget. public goods, which the private sector has few incentives to provide. This double displacement has been documented Over the four years prior to 2003, the debts of agricultural SOEs doubled and reached twice the size of the agricultural TOOLK IT EX A M P L E 1 8 : N AT IONAL VS. SUBNAT IONAL 41 budget. The Vietnam Agriculture PER Integrated Fiduciary be provided by the private sector (e.g., agricultural market- Assessment (World Bank 2005b) suggested that a 2.5 per- ing and processing). The crowding out of the private sector cent reduction in the debts of SOEs would result in suf�cient has the effect of discouraging private investment, which is savings to double spending on research and extension. While desperately needed to maintain growth and competitiveness some of these enterprises provide public services, particu- in agriculture. larly in remote areas, many engage in activities that could Example 18: NATIONAL VS. SUBNATIONAL The degree to which disaggregated data are compiled at the of allocations between central and local governments (Tables subnational level varies according to the relative importance 20 and 21). Depending on the funding source, different fund- of subnational expenditures and the availability of data. Local ing channels are used to direct funds to local governments. Of governments receive funding for providing agricultural ser- these, the most signi�cant sources are the conditional grants vices from a number of sources. The Uganda agricultural PER (A2 and A3). (phases 1 and 2) (World Bank 2007a) provides a detailed table TABLE 20: Sources and Channels of Funding for Agriculture and Local Government FUNDING SOURCE FUNDING CHANNEL EXAMPLE A) Government Transfers A1) Unconditional transfers � Unconditional Grant A2) Conditional nondiscretionary transfers � Agricultural Extension Grant (Wage or Nonwage) � National Agriculture Advisory Service A3) Conditional discretionary transfer � Local Government Development Fund � Plan for the modernization of Agriculture Nonsectoral Conditional Grant B) Projects B1) Central projects implemented through LGs � Area Agricultural Modernization Program � District Development Support Program C) Direct Funding to Communities C1) Donor/NGO � Northern Uganda Social Action Fund � Sasakawa–Global 2000 D) Local Revenue D1) Locally generated revenue � Licenses, fees, market dues Source: World Bank (2007a). TABLE 21: District Conditional Grants Transferred to Local Governments for Agriculture (Based on COFOG De�nition) 2001/02–2005/06 (Ush Billions) GRANTS TRANSFERRED TO LOCAL GOVERNMENTS 2001/02 2002/03 2003/04 2004/05 2005/06 Agricultural Extension (wage) 2.16 2.83 3.19 3.89 3.85 Agricultural Extension (nonwage) 2.90 2.73 2.81 2.92 2.78 National Agriculture Advisory Service (District) 2.42 9.32 13.75 15.13 24.87 Plan for the modernization of Agriculture Non Sectoral Conditional Grant 4.44 4.09 4.20 4.01 4.22 Local Government Development Fund 1.62 1.06 1.69 1.42 0.67 Total-District Grants 13.4 20.03 25.64 27.37 36.39 Total COFOG agriculture 131.23 140.64 147.45 108.11 148.11 % transferred to LGs 10.3 14.2 17.4 25.3 24.6 Source: World Bank (2007a). The Lao PDR agriculture PER (Cammack et al. 2008) illus- on understanding the subnational budget processes. Where trates the high proportion of agricultural expenditures allo- agricultural expenditures at subnational levels of government cated through provincial governments relative to the central are comparatively large or growing, it is vital to ensure that government (about 70 percent versus 30 percent on average, these expenditure data are collected. If the data are not be- although the proportion varies by year) (Table 22). Knowledge ing collected systematically, the need for a system to collect, of the relative weights of spending at the national and sub- preserve, and distribute these data should be a key element national levels can help guide the level of attention in PERs of the expenditure analysis. PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 42 PA RT V: — EX A M PLES TABLE 22: Evolution of Agriculture Expenditures, Lao PDR (in Constant 2000/01 KN Billion) 2000/01 2002/03 2004/05 CONSTANT 2000/01 CONSTANT 2000/01 KN CONSTANT 2000/01 KN KN BILLION % BILLION % BILLION % Provinces 509.3 86.5 332.1 83.3 140.3 65.7 Centre 79.7 13.5 66.5 16.7 73.3 34.3 Total Agriculture 589.0 100.0 398.7 100.0 213.6 100.0 Total All Sectors 3547.8 3453.1 3714.0 Source: Cammack et al 2008. Example 19: GEOGRAPHIC ALLOCATIONS The Mexico agriculture PER (World Bank 2009a) shows that to agricultural GDP, while others (Morelos, Baja California the allocation of agriculture public expenditure by states is Sur, Aguascalientes, etc.) receive disproportionately lower correlated to the size of their agricultural sectors. Figure 6 shares. Figure 7 shows a map highlighting the APE received shows that the distribution of Agricultural Public Expenditure by states relative to their agricultural GDP. It shows that the (APE) and agricultural GDP are closely correlated. In fact, states receiving higher proportions of APE are those in the the correlation coef�cient is 0.70, although some states northern and southern parts of the country, whereas the (Tamaulipas, Zacatecas, Tlaxcala, Sinaloa, etc.) receive dis- majority of the central states receive less. proportionately higher shares in relation to their contribution FIGURE 6: State Share of National Agricultural GDP and Agriculture Public Expenditure (2005–06) % (States ordered from left to right by descending level of State Agricultural Public Expenditure received in 2006) 14% 12% 10% 8% 6% 4% 2% 0% Ch ulip a ca uz D acán Fe oo ah o m rit ho o H rnia N Tab ila lif lie a vo co l So co M aju s lif ro N ón Y che Ja ua O go Pu ca is ico n M la Ba Ag C elos ih as Ch cas Tl atán ue la ist n ur ja ue sí Q orn ntes er ra M taro ra n a m alo Ca ca m Co alg ic at Ba G oto eb Q xca V no ua p Ca rre D nta S Ca aya a Za acr ue as lis h rit a R u Le an de ja uas oli Lu éx te G ia ax pe ua o ré or Ta Sin ui ia uc id ur P a a o Sa APE AGDP Source: Scott (2008) TOOLK IT EX A M P L E 2 1 : O F F - B UDGE T E XP E NDIT URE S 43 FIGURE 7: Map of Mexico Indicating the Share of Agriculture Public Expenditure Relative to Agriculture GDP (2006) Dark green states = agriculture public expenditure is more than 16% of agricultural GDP (12 states) Medium green states = agriculture public expenditure is between 15% and 11.1% of agricultural GDP (11 states) Light green states = agriculture public expenditure is less than 11% of agricultural GDP (9 states) Source: World Bank 2009a. Example 20: SAMPLING SUBNATIONAL GOVERNMENTS State and local governments in Nigeria account for about 46 and data, (iii) expressed interest in collaborating with the percent of all public expenditure. The proportion is estimated agriculture PER team, and (iv) location in different geopoliti- to be even higher in the agricultural sector. The Nigeria cal zones. Agriculture PER (World Bank 2008a) team collected expen- diture data at the federal level as well as at the lower tiers of The nonexistence of historical data at the local level con- government. Because subnational data was not readily avail- strained the compilation of time-series data. The original able from a central source, data on state and local govern- study design called for the analysis to cover at least 10 years; ment expenditures were collected at the local level. however, because few data was available prior to 2000, the length of the past chronology was reduced. To compensate, Due to resource constraints, it was impossible to collect all results of other studies by the World Bank and IFPRI, which statistics; therefore, a case study approach was preferred. took a longer view and provided a historical perspective Three states and three local governments were chosen for on the performance of the agricultural sector, were included in-depth analysis, including data collection. The selection in the analysis. The challenge then became that of integrat- was based on the following considerations: (i) high impor- ing the statistics and results of other studies when trying to tance of agriculture in the state’s economy, (ii) capacity reach an overall assessment of sector performance and the within the state’s public institutions to provide information possible role of public expenditures. Example 21: OFF-BUDGET EXPENDITURES Many countries have sizeable off-budget expenditures. and impacts. Many off-budget allocations tend to be donor Failure to consider these expenditures in PERs can yield a �nancing. Assessing the importance of these expenditures distorted view of public expenditure levels, composition, and identifying options for integrating (or at least accounting PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 44 PA RT V: — EX A M PLES TABLE 23: Disbursements to Uganda’s Agricultural Sector by Two Off-Budget Donors, 2000/01–2006/07 (Ush Billions) (agriculture de�ned according to COFOG parameters) 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 USAID 15.16 10.32 15.29 23.22 20.86 19.55 13.77 Sida 1.75 1.32 2.24 0.81 0.68 1.33 0.00 Total 16.91 11.64 17.53 24.03 21.53 20.88 13.77 COFOG total: NA 131.23 140.64 147.45 108.11 148.11 NA USAID/Sida as % of COFOG total for the – 8.9 12.5 16.3 19.9 14.1 – agriculture sector Source: USAID and Embassy of Sweden, Kampala, May 2007. Note: NA = not available. Due to the nature of the data provided by the two agencies, it was not possible to review the items of expenditure to make sure that they complied with the COFOG de�nition of agriculture. Following discussions with the program of�cers, it was decided to include 100% of the USAID funding classi�ed as “Agriculture� and none of the “Environment� funds, whereas 75% of Sida’s “rural development/agriculture (natural resources and environment) disbursements� were included. Historical exchange rates were obtained from various UBoS annual reports and from http://ec.europa.eu/budget/inforeuro/ index.cfm?Language=en. for and tracking) them within the overall national budget are plant cuttings, pesticides, veterinary drugs, farm tools, and thus important aspects of an expenditure review, as the crop processing equipment. Based on data collected by FAO examples from Uganda and Russia indicate. (Food Security Group, 2007) during a survey of approximately 25 agencies providing agricultural inputs to approximately The Uganda Agriculture PER (phases 1 and 2) (World Bank 300,000 households in Uganda’s northern districts during 2007a) reports that off-budget funds from donors and NGOs the �rst season of 2007, it is estimated that the total value of contribute signi�cantly to the overall funding of the agricul- inputs supplied amounted to almost Ush 9 billion. Given that tural sector, de�ned according to COFOG criteria. Table 23 agricultural inputs are also supplied in the second season, summarizes development assistance provided by USAID and albeit to a smaller number of farm households, the total an- Sida, two important donors in the sector in Uganda. Neither nual value of agricultural inputs provided by NGOs and other of these agencies channels funds for agriculture through humanitarian agencies is estimated at approximately Ush the national budget. Both operate autonomously, maintain- 14.4 billion. This off-budget expenditure represents almost ing only ad hoc links to the MAAIF and its semiautonomous 10 percent of the total expenditure related to agriculture (de- agencies. The table shows that the volume of funds provided �ned according to COFOG criteria) in 2006/07. Care has to to the sector by just these two off-budget development be taken to avoid double counting in aggregating NGO and partners contributed an additional 10 percent to 20 percent donor off-budget expenditure, however, because a high pro- to the total government budget over several years. In addi- portion of NGO funding is provided by donors. tion to USAID and Sida, several other development partners provide off-budget support to the sector, including GTZ, FAO Once again, this is an important area of the agricultural sector (which also provide on-budget support), UNDP, France, and over which neither the MAAIF nor the District Production and JICA (some of its interventions are also on-budget). Clearly Marketing Directorates have any control. Nor do they have it is dif�cult for the MAAIF to carry out its key functions and any powers or resources to regulate the provision of inputs direct investments in support of sectoral development goals to ensure that quality standards are being met and that all when agencies that provide signi�cant sums of money to the communities bene�t equally from the resources being made sector operate independently, in many instances marginally available. working alongside the MAAIF and its staff. The Russia Agriculture PER (World Bank 2006b) highlights a At the same time, NGOs have become signi�cant players different source of off-budget expenditures driven by govern- in Uganda’s agricultural sector, particularly in recent years ment agencies themselves. These expenditures are signi�- through, among other things, the provision of free agricul- cant in scale, nontransparent, and potentially highly distort- tural inputs to internally displaced people, particularly in the ing. Three types of entities in Russia work with off-budget northern districts. Such inputs have included livestock, seed, expenditures and revenues: (i) government departments that TOOLK IT EX A M P L E 2 2 : G E N DE R DISAGGRE GAT ION 45 collect fees for services provided, (ii) state unitary enterprises, of loose controls on state-owned enterprises and state unitary and (iii) state-owned enterprises. Allowing government agen- enterprises led to a de�ciency in the control and monitoring of cies to collect fees for providing public services and having these agencies. Of most concern is that very little information state-owned enterprises or state unitary enterprises provide on revenues, expenditures, and services provided by these public services on behalf of the government are not in them- organizations is reported back to the government, and almost selves bad practices. However, the quality of services that nothing is available for public review. Because most govern- these sources of revenue help �nance needs to be subject to ment agencies receive government budget funds, they are the same level of regulation and monitoring as directly bud- subject to tighter �nancial control and audit requirements geted expenditures. During the chaos of the 1990s, a pattern than state-owned enterprises or state unitary enterprises. Example 22: GENDER DISAGGREGATION The Bank (World Bank 2009c) provides information on decade, more than 60 countries have undertaken analyses of gender-budget analysis tools that are used to review general public budgets to assess differential incidence and its effect or mainstream budget expenditures (for instance, within the on men and women, as well as to measure their economic annual budget of a nation or of a speci�c sectoral program) contributions (Table 24). or to review expenditures speci�cally targeted to groups of women or men to meet prioritized needs or promote equal Different Approaches to Gender-Informed opportunities. Gender-budget initiatives can be de�ned as Budget Analysis “diverse efforts aimed at breaking down the government’s Approaches have differed in terms of focus, coverage, and budget in order to analyze its impact on women, men, girls, methodology: and boys, as well as on other axes of social differentiation (such as race, ethnicity, class, and caste). Their main purpose ƒ The Women’s Budget Initiative (WBI) in South is to examine whether public expenditures are allocated in an Africa expanded its initial broad focus on the national equitable way, and hence promote gender equality� (Balmori budget to analyses of speci�c budgets for domestic 2003). They can also help to reshape government policy violence prevention, treatment, housing, and child goals and resource allocation. support grant programs, among others. The Uganda Gender Budget Project analysis covered the national The Bank (World Bank 2007b) provides information on budget by sector. In Mexico, the analysis focused on gender-informed public �nance management. In the last TABLE 24: Seven Tools for Gender-Budget Initiatives and Examples of Their Use TOOL APPLICATION Gender-aware policy appraisal Designed to analyze policies and programs from a gender perspective and identify how these policies and the resources allocated to them are likely to reduce or increase gender inequalities. Gender-disaggregated beneficiary assessment Implemented to evaluate the extent to which programs or services meet the needs of actual or potential bene�ciaries, as identi�ed and expressed by the bene�ciaries. Gender-disaggregated public expenditure benefit Used to evaluate the distribution of budget resources among women and men, girls and boys, by estimat- incidence analysis ing the unit costs of a certain service and calculating the extent to which this service is being used by each group. Gender-disaggregated analysis of the impact of the Designed to establish a link between budget allocations, the services provided through them, and the way budget on time use in which different members within a household spend their time. Gender-aware medium-term economic policy Designed to incorporate a gender perspective into the medium-term frameworks of policy development, framework planning, and budgetary allocations, such as by disaggregating variables by gender, combining national income accounts and household income accounts, and highlighting and challenging gender-blind, underlying assumptions about how the economy works. Gender-aware budget statement Generated by government agencies for use in reports on the implications of their expenditures on gender- equity objectives. Disaggregated tax-incidence analysis Used to assess the differential impacts of taxation on women and men, as well as to evaluate the level of revenue raised in relation to the needs and demands for public expenditure. Source: Balmori (2003). PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 46 PA RT V: — EX A M PLES antipoverty programs and public expenditure on health ƒ Tools have included gender disaggregated bene�ciary in several states. Korea and the Philippines analyzed assessments, public-expenditure bene�t incidence women-targeted policies and activities at the local lev- analysis, and tax incidence analysis, among others. el. In Morocco, gender budgeting is being introduced Chile has included gender as a cross-cutting theme also at the local level. In general, the more speci�c in a performance-based national budget and is using or focused the gender-informed budget exercise, the incentives (salary bonuses) for public sector staff as a easier its implementation. tool to achieve measurable results. ƒ Most efforts have covered public expenditures classi- ƒ The Bank has undertaken gender-disaggregated �ed into (i) women-speci�c expenditures, (ii) gender public expenditure reviews in a number of countries, equality expenditures in sectors or line ministries, (iii) including Cambodia, Ghana, Morocco, Paraguay, mainstreamed government expenditures that provide St. Vincent and the Grenadines, Vietnam, and goods or services to the whole community, and (iv) Uganda. Most have combined the use of gender- expenditures to achieve equity in public sector staff disaggregated bene�t incidence analysis with gender rosters. Some also extended coverage to revenues: institutional analyses or gender impact assessments the South African WBI looked at taxation to reduce of public programs. The reviews have shown that bias against women, and a review of the value-added undertaking gender analysis can contribute to better tax (VAT) in Uganda recommended tax relief on items targeted, more ef�cient, and more equitable public used by women in the care economy. expenditure. ƒ The most commonly used method takes the govern- Lesson and challenge. The main lesson from the experience ment’s policy framework and examines it sector by with gender-informed budget analysis is that changing pub- sector, both in terms of utilization of budget expendi- lic policy priorities is a more complex process than point- tures and longer-term impacts on men and women. ing out gender differences and disparities in budgets. The The Uganda analysis compared administration expen- implementation of budget initiatives requires upgrading ditures with public services that citizens received and the technical skills of budget of�cials and gender experts, proposed reallocations within and between sectors. raising public awareness of gender issues to ensure the The Mexico analysis focused on how gender-neutral sustainability of the initiatives, and supporting well-informed programs recognized and addressed the limitations coalitions of NGOs for advocacy. Most importantly, effective women face and whether they covered women’s government agencies are central to their implementation. needs and built their capacities. Ideally, these analy- The key challenge for gender-informed budget analysis and ses should cover the four dimensions of government policy making is moving beyond gender-targeted interven- budgets and their interaction: expenditure, revenue, tions to full and sustained gender mainstreaming in the the macroeconomics of the budget, and participation budget process. in budget decision-making processes. Example 23: COST RECOVERY Cost recovery has an important role in promoting more Despite the water scarcity in many parts of the country, ef�cient use of limited resources and in contributing rev- farmers pay only a portion of the costs involved in deliver- enues to sustain public expenditures. It is always a chal- ing irrigation water. In addition, water charges, although lenge to recover the costs of supplying key goods and theoretically based on the volume of water used, are nor- services, especially if they generate private goods, but mally calculated per hectare because of the lack of water- cost recovery issues are worth pursuing in expenditure measuring devices. The current water tariff structure does analysis. not fully cover the operation, maintenance, and deprecia- tion costs of irrigation infrastructure. Also, farmers pay a From Mexico Water PER (World Bank 2006c): Agriculture small fee when their water use exceeds their concession. uses 76 percent of the available water resources in the This fee does not take into account the opportunity cost country but contributes only 1.9 percent of the water use of water, which would set the price at the marginal cost. fees and bulk water tariffs collected by the National Water Low water pricing and lack of measurement encourage Commission (Comisión Nacional de Agua, CONAGUA). TOOLK IT EX A M P L E 2 3 : C O S T RE COVE RY 47 TABLE 25: Cost Recovery in Irrigation COLLECTION RATE PERCENTAGE OF COST RECOVERED Tunisia 1991 NA 70% of O&M costs Turkey 1998 76% 37% of total costs Colombia 1996 76% 52% of O&M costs Mexico 1998–2002 (IDs) 90% 72% of O&M costs Morocco 2004 (large-scale irrigation) 50–60% 54–190% of O&M costs (depending on the irrigation scheme considered) Italy 1997 NA 60% of total costs Jordan 1999 NA 50% of O&M costs Source: Mexico Water PER (World Bank 2006c). the excessive use and wasting of water, a situation that is be �nanced. Consequently, these important measures are certainly not unique to Mexico. Many countries highly sub- still to be implemented by provincial people’s councils and sidize the cost of irrigation water, with farmers paying only related central agencies. part of the O&M costs. Cost recovery for O&M of irrigation systems is generally better in Mexico than in other devel- Two other problems are the low level of fees actually set and oping countries (Table 25). the low collection rates. The fees are set by the Provincial Peoples’ Committee with the approval of the Provincial Recommendations from the Mexico Water PER (World Bank Peoples’ Council and are usually set at a compromise level 2006c): increase cost-recovery level to at least recover fully below the Irrigation Management Companies’ (IMC) actual O&M costs requirements, taking into account appropriate costs. Collection rates are 80 percent to 90 percent in low- maintenance needs, and curb the over-exploitation of water cost gravity schemes (Mekong Delta) and 50 percent to resources by (i) eliminating or decoupling the energy subsidy 70 percent in high-cost pump schemes (Red River Delta) to the use of electricity, (ii) introducing a small water use fee and in areas where poverty is high. As a result, the avail- for irrigation in the areas that suffer the most from water able funds meet only about half of the estimated necessary resources scarcity, and (iii) generalizing wherever possible annual maintenance, with budgets focused on emergency volumetric water pricing. repairs rather than on a more cost-effective program of rou- tine maintenance. Vietnam Irrigation (World Bank 2005b): While in theory cost recovery from users is expected to fund O&M fully, in prac- Reform is conceptually simple but dif�cult to implement. tice this is not occurring. Subsidies from the government Raising fees to a level that fully covers actual costs and en- budget are still a major source of funds for O&M. These suring the collection of 100 percent of the fees due would subsidies, however, are insuf�cient to cover the de�cit, be impractical, at least in the short to medium term, on and MARD estimates that the state budget allocation to both poverty grounds and on grounds that it would force maintain irrigation infrastructure is only about 60 percent users to pay for major system inef�ciency. Three steps to- of the amount needed. Funding for operating and main- ward a long-term goal of full cost recovery would be (i) cor- taining the water distribution system, from the headworks rect accounting of real costs of O&M to quantify the de�cit, to the village, is a provincial responsibility under Ordinance (ii) fully fund this de�cit through increasing the O&M share 32 (April 2001) and Decree 143-CP (November 2003) on of public expenditure in irrigation on an agreed schedule Exploitation and Protection of Irrigation Schemes. Decree that would phase out the subsidy as the ef�ciency of ser- 143-CP obliges the provinces to set service fees at a level vice delivery improves, and (iii) make payments from the to cover the normal operating expenses of their irrigation government directly to poor communes to pay irrigation management companies. The decree allows for an exemp- fees to the IMCs rather than exempt poor communes from tion or reduction of fees in localities affected by a high fees. The present system of paying the subsidy to the prov- incidence of poverty or by disasters. However, it does not inces does not ensure that the subsidy is passed on to the specify a mechanism by which such exemptions should IMC. This results in service deterioration and impacts IMC PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 48 PA RT V: — EX A M PLES ef�ciency. A fourth step would be to introduce full cost re- irrigation management is resisted by farmers if the section covery packaged with rehabilitation and improved service being turned over to them is in poor condition). delivery. Pilots in Vietnam have shown that participatory Example 24: SIMPLE INDUCTIVE ANALYSIS A simple review of spending allocation data and comparisons 80 percent of expenditures are for salaries, leaving little for with international standards can reveal important insights for operational expenses. About 87 percent (in 2001/02) of the policy consideration. Here are some examples. O&M expenditure for irrigation goes to wages and salaries, thus neglecting actual work (World Bank 2003). Capital vs. Current Spending The economic composition of expenditure matters for Wage vs. Nonwage growth. Using data from 43 developing countries over a Though an efficient capital or recurrent balance is country- 20-year period, Devarajan, Swaroop, and Zou (1996) argue specific, the composition of recurrent spending across that reallocating funds from capital to current expenditures, countries is dominated by salaries and wages. Allocating while keeping the overall level of expenditure constant, can funds to other components of recurrent expenditures, foster long-term economic growth. In Turkey, for example, such as O&M, can bring high returns (Devarajan, Swaroop, 24 percent of rural spending goes to irrigation. The recurrent and Zou 1996). However, there is no predetermined norm share of irrigation investment is 44 percent, while the share for determining the optimal level of spending on O&M. of O&M in recurrent expenditures is only 2 percent (World For example, the effectiveness of agricultural services Bank 2005a Turkey). However, O&M expenditure in irriga- is believed to be reduced if expenditures on wages and tion has been found to yield higher returns than new capital salaries exceed 60 percent of the total recurrent budget, investments. In India, for example, rates of return for O&M leaving O&M with a small share of total agricultural ex- expenditures on irrigation have been found to be between penditure (World Bank 2005a). While international bench- 29 percent and 40 percent (Pradhan 1996). Diminishing marks suggest a ratio of salary-operational expenditure of marginal returns on excessive capital investments render 60 to 40 (or 1.5:1), the ratios for salary-nonsalary costs in them unproductive. This argument could also be made for Pakistan’s North West Frontier Province for 2002/03 were excessive recurrent spending, especially when salaries and at least 3 to 1 and as much as 32 to 1 for the Agriculture wages represent a large share of recurrent expenditures. and Livestock Departments (World Bank 2004 Pakistan). For example, in India, the intensity of current spending (the In Bangladesh, 90 percent of the revenue budget of the share of expenditure in agricultural state domestic product) Department of Agriculture Extension goes to staff sala- outweighs the intensity of capital expenditures by more ries. To improve the effectiveness of extension services than double (an average for all states of 5.6 percent for cur- in the country, it is recommended that at least 30 percent rent and 2.2 percent for capital expenditures from 1998/99 of the total budget of the Department is allocated to cover to 2000/01). In the state of Punjab in India, more than operational costs of the service (World Bank 2005c). Example 25: SIMPLE CONGRUENCE ANALYSIS The Honduras agricultural PER (Anson and Zegarra 2008) of alignment of budgetary allocation with the program (in includes a simple assessment of whether the stated ag- percentage terms). Expenditures on rural development ricultural strategy—the Special Food Security Program were excluded because their disaggregated levels were not (Programa Especial de Seguridad Alimentaría, PESA)—is known. The results are reflected in the table below, show- consistent with actual budgetary allocations to the agri- ing that only about 30 percent of all sectoral expenditures cultural sector. A simple methodology was used: the ex- are in line with PESA strategic priorities (Table 26). This pat- penditures of the various sectoral agencies in 2004 were tern is common to many countries. This example illustrates aggregated and allocated according to strategic themes a simple congruence analysis that has been used in several outlined in PESA, which enabled an estimate of the extent analyses of public spending. TOOLK IT EX A M P L E 2 6 : E F F I CIE NCY GAINS F ROM SWIT CHING FROM PR IVATE TO PUBLIC GOOD S 49 TABLE 26: Comparison of Budgetary Allocation to Strategy Elements in Honduras, 2004 STRATEGIC THEME ACCORDING TO THE SPECIAL FOOD SECURITY PROGRAM AMOUNT (US$ MILLION) Market Development and Trade Agreements – Agricultural Phytosanitary 0.15 Technology Innovation 3.00 Agricultural Education 1.17 Agricultural/Rural Finance 6.70 Development of Rural and Irrigation Infrastructure 5.30 Sustainability of Natural Resources 20.0 Access to Land, Judicial Security, and Social Equity 6.75 Subtotal 43.07 Total Public Expenditures (Five Years) 144 Proportion of Public Expenditure to Support Strategic Themes 30% Source: Anson and Zegarra (2008). Example 26: EFFICIENCY GAINS FROM SWITCHING FROM PRIVATE TO PUBLIC GOODS The below extract from the Mexico PER (World Bank 2009a) of increasing the share of expenditures going to subsidies highlights the ef�ciency gains from switching from private to were large and negative (and highly statistically signi�cant). public goods (Valdes 2008b). A reallocation of 10 percentage points of public expenditures from subsidies to public goods would increase per capita Recent econometric work �nds that for many countries agricultural income by about 2.3 percent without increasing in Latin America and the Caribbean it is crucial to shift total expenditures. On the poverty side, the analysis found rural public expenditures from large subsidies going to that the mix of expenditures had no statistically signi�cant speci�c groups of producers and towards the increased direct effect on poverty, in part because of the poor state provision of public goods. To illustrate just how impor- of rural poverty data. The main impact of the mix of public tant the mix of private and public good spending can be, expenditures on poverty is indirect, through the effect on per Ramon López14 used the database from the Food and capita agricultural income. Overall, the �ndings from Lopez’s Agriculture Organization’s Latin American and Caribbean cross-country analysis for Latin America are consistent with Regional Of�ce to classify expenditures by several coun- other analyses for India and China, where spending on rural tries over 1985–2001. Public goods included technology roads and agricultural research was found to have large pov- generation and transfer, soil conservation, sanitary and erty alleviation effects. phytosanitary protection, communications and informa- tion services, rural infrastructure, and social services The allocation criteria for expenditures are not complex in (education and health). Private goods included commod- theory—public goods should be a priority, and the coverage, ity-speci�c items, marketing assistance and promotion, targeting, and cost-effectiveness of transfers should be part subsidized credit, and irrigation. of the evaluation of any project. But in practice it is compli- cated to design public spending because the policy maker The key message from the analysis is that while government requires empirical measures of the relevant “elasticities,� expenditures do have a positive effect on agricultural per reflecting how agricultural growth and poverty reduction capita income, the composition of those expenditures is im- respond to expenditures on various program categories. portant. The long-run effects on per capita agricultural GDP From a national development perspective, policy makers require information on how national welfare responds to ag- ricultural development. Policy design is further complicated 14 The work by Ramon Lopez of the University of Maryland was pioneering in this area and �rst developed for the WB regional re- by the heterogeneity of farming conditions with respect to port “Beyond the City, The Rural Contribution to Development� initial product mixes and levels of productivity, access to in 2005, and later published as Lopez and Galinato (2007). PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 50 PA RT V: — EX A M PLES infrastructure, soil fertility, and access to markets. This het- �nance, research, and extension, will determine the ability erogeneity stems in part from the natural resource base and of small-scale farmers to take advantage of opportunities in geography, but also in part from the approach governments both domestic and international markets. Finally, although have taken in the past to expenditures on public goods. Past the role of governments as providers of public goods is investments in infrastructure and institutions, such as tele- well-established, the performance of governments is often communications and transport, and the support network of disheartening. Example 27: MARGINAL RETURNS TO PUBLIC SPENDING ACROSS SECTORS Estimating marginal returns to public spending across sec- equations, they estimate the marginal impact of different tors is a fairly data-intensive exercise. IFPRI has used this types of government expenditures. For example, for the approach in many countries with results summarized in Fan study on India, marginal returns for an additional Rs 100 bil- et al. (2007). The country studies typically develop a simulta- lion of government expenditure were estimated for various neous equation model to estimate the various effects of gov- types of public spending. These estimates can be used to ernment expenditure on production, inequality, and poverty compare the relative bene�ts of equal incremental increases through different channels. Through subsequent differential in expenditure on different items (Table 27). TABLE 27: Marginal Returns to Public Spending in India AGRICULTURE TOTAL FACTOR POVERTY (% POINT) PRODUCTIVITY (% POINT) NUMBER OF POOR REDUCED (PER 100 BILLION RUPEES AT 1993 PRICES) (PER MILLION RUPEES) TYPES OF SPENDING RANK RANK RANK R&D −0.45 2 6.01 1 84.5 2 Irrigation −0.05 7 0.61 4 9.7 7 Roads −0.65 1 2.37 2 123.8 1 Education −0.22 3 0.62 3 41.0 3 Power −0.003 8 0.12 8 3.8 8 Soil and Water −0.12 5 0.43 6 22.6 5 Rural Development −0.13 4 0.49 5 25.5 4 Health −0.09 6 0.38 7 17.8 6 Source: Fan, Hazell and Thorat (2000). Example 28: ELASTICITY APPROACH An IFPRI analysis used an elasticity approach to estimate the based on these growth rates and the expenditure elasticities agricultural public spending required to achieve the agricultur- of growth. The model simulations account for the impact of al growth needed to meet the �rst Millennium Development the nonagricultural sector by using nonagricultural-growth Goal. Estimates were provided by individual African countries elasticity of poverty. The share of the nonagricultural sec- (Fan et al 2008). The IFPRI study used both growth-poverty tor is expected to increase over time, and thus its impact and expenditure-growth elasticities to estimate the required on poverty reduction may also increase. Therefore, the agricultural growth rates and corresponding public expendi- study also considers the nonagricultural sector in order to tures needed to achieve this goal. The required agricultural avoid overestimating the agricultural growth and spending growth rates are estimated using elasticities of poverty with required to achieve the �rst Millennium Development Goal. respect to both agricultural and non-agricultural growth, and The simulations were conducted for the 30 countries in Sub- the additional spending needed in agriculture is calculated Saharan Africa in which the agricultural sector contributes at TOOLK IT EX A M P L E 2 9 : U N D ERSTANDING T HE P OL IT ICAL E CONOMY OF PUB LIC SPEND ING 51 least 10 percent of the gross domestic product (GDP) and availability of expenditure data. (More details on the method- where the majority of the poor depends upon agriculture for ology can be found at: http://www.ifpri.org/sites/default/�les/ their livelihood. The choice of countries was governed by the publications/ifpridp00751.pdf.) Example 29: UNDERSTANDING THE POLITICAL ECONOMY OF PUBLIC SPENDING The determination of a country’s public budget and its vote in disparity between the agricultural and nonagricultural sec- parliament (for example) is a political process. Understanding tors, worsened by India’s relatively low agricultural growth the political economy of public spending can help guide rec- rate. Making electricity free is a politically convenient ommendations for improvement. The World Development instrument to transfer income to the agricultural sector. Report 2008: Agriculture for Development highlights this as- Unlike other policy instruments, it does not require imple- pect for India and Zambia as presented here. The recent book mentation by the (often ineffective) public administration. Political Economy of Agricultural Price Distortions, edited by Farmers who buy water from pump owners—a consider- Kym Anderson (2010), provides the rich set of conceptual able portion of farmers in most states—potentially bene�t and econometric studies using both the new developments from the subsidy, too, which increases the attractiveness in general political economy theory and the new set of panel of this policy instrument for politicians who want to win data on agricultural distortions to improve understanding of state elections. why governments intervene in agricultural markets in the way they do. Addressing jointly the quality of electricity supply and its cost is a key element of reforming the subsidy policy. However, Birner, Sharma, and Palaniswami (2006) reports on issues because of widespread power theft and losses, states lack related to electricity subsidies to agriculture in India. With the credibility to deliver better service in exchange for higher 55 percent to 60 percent of India’s irrigated land supplied by prices. One option would be to decentralize the energy sup- groundwater, electricity for tube-well pumps is an important ply to local governments or community groups, relying on lo- input. Most state governments provide electricity to farmers cal accountability to improve electricity quality. Elite capture at a subsidized flat rate, and often for free. But the quality must still be prevented, but this community-oriented option of service is poor because of erratic and limited supply and has the potential to break the political impasse. This exempli- voltage fluctuations, which can result in crop losses from �es a trade-off between potential ef�ciency cost from the forgone irrigation and damaged pumping equipment. loss of economies of scale in decentralized generation and not making any progress at all. The electricity subsidies to agriculture are also �scally drain- ing and environmentally damaging. In Punjab, electricity In Zambia, no strong opposing coalition exists for the elimina- subsidies to agriculture in 2002/03 were 7 percent of state tion of high spending on fertilizer subsidies (Beintema and expenditures. Together with other policies that promote others 2004; Govereh and others 2006; Pletcher 2000). About water-intensive crops such as rice, the electricity subsidies 5 percent of Zambia’s national budget goes to agriculture. In contribute to the overexploitation of groundwater. About �scal year 2005, more than half the agriculture budget was 60 percent of the state’s groundwater resources are already spent on the Fertilizer Support Program (37 percent) and overexploited, with extraction rates exceeding recharge crop marketing (for maize) under the Food Reserve Agency rates, which is clearly not sustainable. (15 percent). Only 3 percent of the budget went to irrigation development and other rural infrastructure, and 11 percent Increasing electricity prices and introducing metering are went to operating costs, which included agricultural research technically and economically sound, but they are not politi- and extension. Spending on agricultural research and devel- cally feasible, so far. Larger farmers obviously bene�t more opment fell from about 1.2 percent of agricultural GDP in from the subsidy, and they have political influence, but there 1985 to about 0.5 percent in 2000. is more to these subsidies. Why is spending on fertilizer subsidies so high? There are The introduction of subsidies followed massive farmer no powerful groups that would bene�t from its elimination, protests against electricity price increases in the 1980s. despite its being an economically unproductive use of public Now, their continuation responds to the increasing income resources. This contrasts with early reforms in maize milling, PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 52 PA RT V: — EX A M PLES where the private sector gained signi�cantly from privatiza- fertilizer through the government program were predomi- tion and strongly supported the reform. Under the fertilizer nantly civil service employees in a program intended to be program, traders often bene�t. targeted at the poor. Parliamentarians also bene�ted, some- times informing groups of farmers that there was no need to A 2002/03 household survey showed that only 29 percent repay loans on fertilizer received. of farmers acquired fertilizer. Of those farmers, 59 percent acquired it through private dealers and 36 percent obtained The economic costs of the program are high, both from it through the government Fertilizer Support Program. Both lower spending in higher-productivity areas such as agricul- groups had higher income and wealth and were close to tural research, extension, and infrastructure and from slower tarmac roads and district centers. However, those receiving diversi�cation away from maize production. Example 30: AGRICULTURE SECTOR-WIDE APPROACHES Donors account for a large share of public spending in the ag- supported agriculture in 2005) largely through “basket fund- riculture sectors of many low-income countries. In an effort ing� (pooling donor resources) guided by an agreed agricul- to improve the alignment of donor support with government tural development program. Ghana has also progressed in strategies and to reduce the transaction cost of aid (reducing harmonizing development assistance and aligning it to its multiple planning and reporting requirements, different pro- agricultural development strategy. curement and �nancial management systems), agriculture sector-wide approaches (SWAp) are being increasingly ad- Nicaragua’s sector-wide PRORURAL15 program was launched opted by governments and donors. The Paris Declaration and in 2005. The government, the private sector, and 15 donors Accra According on aid effectiveness are supportive of this (which supply more than 90 percent of donor assistance for approach across all sectors. In agriculture, about 18 countries agriculture) signed a Code of Conduct to promote country lead- have adopted a SWAp. ership, harmonization, and alignment in supporting national priorities in the agricultural sector. A common fund, guided by In Uganda, a coherent country-led poverty reduction strategy a Memorandum of Understanding (MOU), was established was supported by a sound agricultural strategy and institu- in 2006 to pull together contributions of donor agencies into tional reforms. The management of aid flows for a coherent a single account, used according to the priorities de�ned by pro-poor expenditure strategy, including a strategy for rural Nicaraguan institutions. The setup of a sectoral Medium-Term areas, has resulted in relatively stable long-term commit- Expenditure Framework (MTEF) and an annual budget plan ments by donors. In Tanzania, government leadership has provided the basis for approving disbursements, using the overcome fragmentation (17 bilateral and multilateral donors government’ �nancial and procurement procedures. Example 31: INSTITUTIONAL COORDINATION MECHANISMS An important aspect of some agriculture PERs has been to and Rural Affairs; and (iv) the Ministry of Environment and provide a description of the role of agricultural public institu- Forestry and the General Directorate of Forestry. tions (at various tiers of government), an assessment of their effectiveness, and the arrangements for coordination among Ministries have �eld of�ces at provincial and district levels, institutions. which are under the authority of the provincial governors (vali) and district subgovernors (kaymakam). The provincial The Turkey agriculture PER (World Bank 2005a) shows that administration operates on the principle of deconcentration the public spending in the rural sector is channeled through (i.e., a delegation of power from the central administration both central and local government agencies. Central gov- to the provincial level). Provincial governors and subgover- ernment in Turkey is administered through a series of min- nors are appointed by the central government and have au- istries and af�liated agencies which include (i) the General thority over the �eld of�ces of the central administration. Directorate of Rural Services, accounting for the largest share of the 2003 rural sector budget; (ii) the State Hydraulics 15 Experiences and results from the Global Donor Platform’s and Waterworks Organization; (iii) the Ministry of Agriculture in-country facilitation sevices: http://www.donorplatform.org/ component/option,com_docman/task,doc_view/gid,396. TOOLK IT EX A M P L E 3 2 : S P E CIF IC RE COMME NDAT IONS T O GOVER NMENT BUD GET GUID ELINES 53 Local administrations also have a certain responsibility established a coordinating framework for ARD public ex- for expenditures in Turkey. They are coordinated by the penditures through the Interministerial Commission for Ministry of the Interior through its General Directorate Sustainable Rural Development (CIDRS), headed by the of Local Authorities. Estimates from the State Planning Ministry of Agriculture, Livestock, Rural Development, Organization indicate that total consolidated expenditures Fisheries, and Food. The expectation was that CIDRS of local administrations (all sectors) represent 4 percent to 5 would develop a rural development strategy and serve as percent of GNP.16 Moreover, they account for 30 percent of a forum to coordinate the government’s ARD activities. A the investment budget; however, their investment spending strategy, however, has not been developed, and CIDRS has is focused almost exclusively on urban areas, mainly drink- not yet actually managed to orient and coordinate the allo- ing water supply and urban transportation. Furthermore, cation of federal resources to ARD programs. Thus the PEC only Special Provincial Administrations, which account for (Mexico’s public expenditure program for ARD, funded by 16 percent of local administrations’ total budget, spend for the federal government) annually prepared by CIDRS has, road and water supply investments in rural areas. to date, not functioned as a planning tool. It operates as an annual budget exercise where the budget lines of all The Ghana Agriculture PEIR (Kolavalli et al. 2009) summarizes rural development programs are tabulated together. PEC in a flow chart how ministries, departments, and agencies in appears as an annex of the annual federal budget. the agricultural sector are organized. The Ministry of Food and Agriculture (MoFA) is the lead agency for the agricultural sector. An Agriculture Sectoral Working Group (SWG) for promot- MoFA has seven technical directorates and three subvented ing interinstitutional coordination of the budgetary process agencies: the Ghana Irrigation Development Authority (GIDA), was established in Uganda (World Bank 2010b). The SWG the Grain and Legume Development Board, and the Veterinary has the task of reviewing sector strategies and invest- Council. GIDA is represented in the Ghana Water Resources ment programs, recommending projects for submission Commission, which coordinates water resources development to the Plan for Modernization of Agriculture Secretariat across sectors. At the regional level, MoFA is represented by and Development Committee that are in line with sector the Regional Agricultural Development Units and at the district plans, preparing agriculture Sector Budget Framework level by the District Agricultural Development Units. Paper (SBFP) as a basis for compiling the annual budget for the sector and enabling the MTEF for the sector to evolve In summary, the Ministry of Food and Agriculture is part over time, providing the main forum for the sector-wide ap- of a complex institutional landscape in which the re- proach to planning and budgeting for the agricultural sector, sponsibility for important agricultural functions, such as identifying policy issues for consideration and action by the agricultural research and cocoa development, is shared Ministry Top Policy Management, providing information for with agencies that are located under other ministries. the Joint Government of Uganda Donor Reviews, and moni- Likewise, MoFA has to function in a complex decentral- toring budget implementation vis-à-vis the aims and objec- ized system. This implies considerable coordination chal- tives set out in the Budget Framework Paper (BFP). The lenges for MoFA because the Ministry has the overall SWG is chaired by the Permanent Secretary of the Ministry mandate to lead the sector. of Agriculture, Animal Industries, and Fisheries. Typically, The Mexico Agriculture PER (World Bank 2009a) describes the SWG is conveyed after the distribution of the Budget and assesses the interinstitutional coordination mecha- Call Circular and may meet also during preparation of the nism for the Agriculture and Rural Development (ARD) Budget Framework Paper (BFP). Funding for the SWG ac- sector. The Sustainable Rural Development Law of 2001 tivities is budgeted from within the sector MTEF ceilings. Example 32: SPECIFIC RECOMMENDATIONS TO GOVERNMENT BUDGET GUIDELINES An expenditure analysis should provide speci�c recommen- of institutional and staff incentives and capacities. Below are dation to government budget guidelines, through a careful examples from the macro (Turkey) and sectoral level (Uganda review of actual practices and underlying constraints in terms and Ghana). 16 This compares to central government consolidated budget expenditures, which represent around 40 percent of GNP. PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 54 PA RT V: — EX A M PLES The Turkey Agriculture PER (World Bank 2005a) discusses assessment of the budgetary processes for agriculture key results of the macro-level PEIR (World Bank 2001) along that uses the sectoral Mid-Term Expenditure Framework with constraints to further progress. It was noted that, (MTEF) at the Ministry of Agriculture and the local level. after the review, the Government of Turkey took important The MTEF is a major component of the Public Financial steps toward improving budget coverage and classi�ca- Management Reform Program (1999) and aims at (i) link- tion, enhancing the budget preparation process, restoring ing budget expenditures more closely to national priorities its credibility with line ministries, strengthening the capac- through a national strategic planning process, (ii) integrat- ity to formulate policies and budgets within a medium-term ing aid-�nance programs and policies into mainstream perspective, and improving budget execution and �nancial sector-resource allocation processes, (iii) elaborating three- accountability. Recommendation centered on the Public year sector expenditure proposals based on anticipated Financial Management and Control Law, enacted in late 2003, resources availability, and (iv) developing an activity- or which provided a legal framework for structural and institu- performance-based budget system that enhances the tional reforms to upgrade public expenditure management. government’s ability to monitor expenditure ef�ciency and Speci�c recommendations related to the need to clarify roles effectiveness by linking expenditure directly with outputs. and responsibilities of public administrations and to transfer The PEIR highlights that at both the central and local lev- some responsibilities to the private sector. els, the rigidity of the organizational process, the lack of planning capacity, and the heavy administrative costs are The Uganda Agriculture PER (phases 1 and 2) (World Bank unlikely to yield substantial improvements to the budget- 2007a) comprehensively assesses the budgetary pro- ary process and that serious weaknesses remain even 10 cesses followed in the agricultural sector. The Study Team years after the introduction of MTEF. examined the most recent Sector Budget Framework Paper (SBFP) guidelines and suggested modi�cations to improve The review found that persistent divergences between bud- the SBFP’s analytical content and internal consistency. get allocations and disbursements were a government-wide Recommendations followed the structure of the headings problem and that disbursement rates varied across spending of the 2007/08 SBFP and related to the sector priorities, categories, with budget allocations for personal emoluments performance, budget allocations and objectives, mid-term more likely to be ful�lled than the two complementary ex- budget priorities, expected outputs, performance indicators, penditure categories (services and investments). The team and planned activities. In particular, the Team suggested underlined that MoFA tried to face these budgetary chal- improving the methodology of the SBFP by (i) providing a lenges through enhancing �nancial control via means such more detailed explanation of expenditures and shortfalls, (ii) as periodic internal audits. However, it was recommended clearly linking indicators to speci�c outputs and activities, that to be effective, (i) the audit should have been more and (iii) prioritizing activities requiring additional resources. frequent and (ii) the monitoring teams should be composed by adequate staff and well-endowed to overcome logistical The Ghana Agriculture Public Expenditure and Institutional challenges. Review (PEIR) (Kolavalli et al. 2009) contains a qualitative Example 33: INCENTIVE MECHANISMS TO IMPROVE BUDGET PERFORMANCE An example of an incentive mechanism to improve bud- minimum conditions. The minimum conditions and perfor- get performance is the system of local government grant mance measures relate to several functional areas: �nancial transfers introduced by the Government of Tanzania (United management, �scal capacity, planning and budgeting, trans- Republic of Tanzania 2006). The Local Government Capital parency and accountability, human resource development, Development Grant (LGCDG) system provides grant �nanc- procurement, project implementation, and council functional ing for local capital improvements conditioned on local gov- processes. LGAs are assessed against these measures an- ernment authorities (LGAs) meeting minimum requirements nually, the outcome of which determines the levels of grants that ensure that funds transferred to them are properly used. received in the subsequent year. LGAs who do not meet the minimum conditions do not get capital grant transfers. LGAs In addition, the LGCDG system includes performance mea- who do meet the minimum conditions get the grants, with sures to encourage continued improvements beyond the TOOLK IT EX A M P L E 3 4 : B U D GE T E D VS. ACT UAL 55 those that show improved performance measures during the The basic premise of the LGCDG system is to avoid wast- annual assessment getting higher levels of grant transfers. age or misallocation of scarce public resources by giving All LGAs have access to a capacity-building grant that can resources to those LGAs that have demonstrated capacity be used for training (e.g., on �nancial management, planning to manage well, and by building the capacity of all LGAs. and budgeting, procurement, etc.) to improve performance This system was developed and is being implemented in scores in subsequent assessments. The performance as- close collaboration with development partners, who are also sessment and associated capacity building funds are aligned supporting the LGCDG system. The performance assess- to improving capacity across the entire budget cycle—plan- ments have subsequently been extended to include mea- ning, budgeting, procurement, implementation, accounting, sures related to agriculture through the Agriculture Sector and auditing. Development Programme. Example 34: BUDGETED VS. ACTUAL The agriculture PERs reviewed in World Bank (2010d) actually disbursed to the ministry and (2) low disbursement showed large discrepancies between planned and actual by the ministry to intended activities. There is a myriad of budget execution far exceeding accepted international stan- underlying causes for these low rates. For the �rst factor of dards developed under the Public Expenditure and Financial discrepancy between the approved amount and the disburse- Accountability partnership, which says that actual expen- ment to the ministry, these include late release of funds, ditures should deviate no more than 10 percent from the cutbacks in approved budgets due to revenue shortfalls or budget to qualify as ef�cient execution (see Table 28). A unforeseen demands on available funds, and so on. For the review of Agriculture PERs has found that typically execution second factor of low disbursement within the ministry, these rates are higher for recurrent costs, of which a large portion include procurement delays, low accountability and nontrans- consists of salaries, compared to capital costs. In terms of parent �scal relations between state and local government, funding sources, execution is higher for nationally funded weak internal institutions and implementation plans involving programs as opposed to foreign-�nanced, which tend to the various executing agencies, weak monitoring systems to have stricter �duciary requirements that delay disbursement. track the delayed disbursements of approved funds, poor ex- The gap can arise from two factors: (1) discrepancy between penditure recording, unauthorized expenditures, poor internal the amount of budget approved and the amount of budget controls, no uniform formats, late reporting, and so on. TABLE 28: Disbursements as a Share of Budget Allocations (%) AVERAGE EXECUTION RATE 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 (%) HONDURAS Agriculture & – 67.7 64.9 56.6 73.4 56.8 44.9 60.1 57.3 62.4 59.4 Forestry NIGERIA Agriculture – – – – 91 58 60 85 103 – 79.0 ETHIOPIA Agriculture Rural 83 86 72 71 80 79 85 82 – – 79.8 UGANDA Ag., Animal – – – – 85.2 82.1 103.5 160.5 118.7 90.4 106.7 Industries, & Fisheries Source: Brzeska and Fan (2009). Note: Recurrent cost only. PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 56 PA RT V: — EX A M PLES Example 35: TIMING OF RELEASES The Uganda Agriculture PER (World Bank 2010b) shows releases made on a regular quarterly basis are usually 98 that in addition to the shortfalls in allocations, the effec- percent to 99 percent of budget allocations. The seasonal tiveness of the public expenditures was weakened by un- nature of many agricultural operations, however, means certainty in seasonal cash flows. The development budget that NAADS District expenditures peak in the �rst and third typically is released in four equal installments in the �rst quarters of the �nancial year, when crops are planted (see month of each quarter. However, the cash budget system Figure 8). One effect of cash-budget ceilings and limited restricts quarterly fund releases to cash flow availability. cash releases to the Ministry of Agriculture, Animal Industry, While this practice helps to ensure macroeconomic stability, and Fishery (MAAIF) is that some anticipated donor project it may destabilize activities at the central and local govern- expenditures cannot be undertaken. At the national level, it ment levels. Because wage payments and Poverty Action used to be the case that some donor projects required the Fund (PAF)–related expenditures have �rst call on available government to provide counterpart funds to pay for various funds, nonwage recurrent expenditures and government items before donor funds were disbursed. This restriction development expenditures on non-PAF activities tend to be appears to have been relaxed by most, if not all, development affected the most. partners. It is necessary that, within the overall cash flow available, the Ministry of Finance, Planning, and Economic Cash flow within the National Agricultural Advisory Service Development (MoFPED) should cater to the relatively small (NAADS) provides a good example of how seasonal require- but particular cash flow requirements of agencies such as ments can affect expenditures. NAADS District expenditure NAADS to improve its operational effectiveness. is classi�ed as a Poverty Action Fund activity, and as such, FIGURE 8: NAADS (Districts) Budgeted and Actual Cash Flows, 2005/06 (Ush Millions) 1,000,000 Budgeted cash flow 900,000 Actually released 800,000 700,000 600,000 Ush millions 500,000 400,000 300,000 200,000 100,000 0 il y r er r r t ne ly ay ry ch us be be be ar pr ob Ju ua Ju M ar ug nu em em em A ct M br A Ja ec O pt ov Fe Se D N Source: World Bank (2010b). TOOLK IT EX A M P L E 3 7 : P U B LIC E XP E NDIT URE T RACKING SURVEY 57 Example 36: APPROACHES TO GATHERING DISTRICT- LEVEL DATA The Ghana PEIR (Kolavalli et al. 2009) illustrates various ap- included questions designed to elicit their perceptions of proaches and sources for collecting the required information. various issues relevant to their work, the importance of vari- The expenditure review primarily involved collecting and ous factors in their work environment, and the accountability analyzing data, whereas the institutional review was more structure. The survey was completed by 67 staff members. participatory. The participatory methods included two con- The study also used data collected through a survey by sultations with senior management and consultations with the Institute of Statistical, Social, and Economic Research a group of retired staff. The study team chose to examine the workings of district of�ces, as organizational strengths (ISSER) and IFPRI as part of the project “Making Rural and weakness are likely to be most noticeable at the points Service Provision Work for the Poor,� focusing on rural of service delivery. After a preliminary document review, the water supply and agricultural extension. The survey cov- team did a scoping study at the East Akim District Agricultural ered households, elected and appointed District Assembly Development Unit (DADU) to understand how the district of- members, District Assembly staff, farmer-based organi- �ces function and to identify suitable approaches for captur- zations, agricultural extension agents, and organizations ing critical aspects of how the DADUs functioned. involved in rural water supply. The data presented in this report were collected in four districts (Wassa West, Following the scoping study, the team presented the research Wassa Amen� East, Tolon Kumbungu, and West Gonja). plan to regional and national directors in a meeting organized in Data on public agricultural expenditures were obtained Accra. They recommended four districts for case studies, one in primarily from the Controller and Accountant General’s each agro-ecological zone, to capture any differences in the chal- Department, MoFA, the Council for Scienti�c and Industrial lenges faced in delivering services: Dangme East in the coastal Research, the Ghana Cocoa Board, the Ministry of Road zone, Wassa Amen� West in the forest zone, Wenchi in the Transport, the Ghana Statistical Service, and many other Transition Zone, and West Mamprusi in the Savannah Zone. institutions, in addition to published sources. Information from the agencies and case study interviews mentioned The DADUs in these districts were requested to compile the earlier was used to analyze factors relating to institutional required information before the team visited. During the �rst performance. three weeks of February, a team of consultants visited the districts to interact with staff and collect the information. In In examining expenditures, several analyses were done: the districts, senior of�cers were interviewed individually, and the group discussions were held with the �eld staff. ƒ Trend analysis, to assess the size and composition of Data collected from the district of�ces included (1) details of public (mainly government) expenditure in the agricul- performance assessment for one year; (2) details of dates tural sector of recruitment, promotions, and current positions of all staff; ƒ Unit cost analysis, to assess the ef�ciency of provid- and (3) annual work plans, progress, and �nancial reports. ing public goods and services in the sector The staff was asked to individually provide the following ƒ Bene�t incidence analysis, to assess the effective- information: (1) a list of activities carried out in the previous ness of delivery and utilization of public goods and �ve days, including distances traveled, time spent interact- services in the sector ing with farmers, number of farmers contacted, and other ƒ Case studies, to identify the binding constraints for information; and (2) the training and promotions received improving public agricultural expenditure management during the staff member’s employment with the ministry. and the delivery and utilization of public goods and In addition, staff responded to a survey instrument that services in the sector Example 37: PUBLIC EXPENDITURE TRACKING SURVEY The Uganda Agriculture PER (World Bank 2010b) includes of the major development projects in the agricultural sector an excellent analysis in which a Public Expenditure Tracking between 2005/06 and 2007/08. The methodology involves Survey (PETS) was used to help assess the technical ef�ciency four key steps, outlined below in an extract from the Review. PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 58 PA RT V: — EX A M PLES FIGURE 9: Relationship between Inputs, Outputs, and Outcomes in Agriculture Financial Government Donor Household Private traders, Inputs funds funds contributions NGOs and other funds Tangible Inputs Vaccine Capital/finance Services (advisory, etc.) Outputs Increased Healthier Improved access to output cattle inputs and finance Outcomes Increased agricultural productivity Increased farm income Reduced poverty Source: Uganda Agriculture PER Task Team. Step 1: Define the result chain suitable for PETS in evaluation must cover a suf�cient timeframe to allow agriculture. To improve ef�ciency in agriculture—in outcomes to be realized and properly documented. other words, to get more outputs and outcomes without spending more—one needs to de�ne inputs, outputs, This kind of rigorous impact evaluation is presented for and outcomes. Figure 9 shows how these terms can be the NAADS program in the report. Similar detailed evalu- used for agriculture and how they are related. ations of additional programs were not feasible given the time and resources available. Instead, the PETS method- Outcomes are what stakeholders should be inter- ology was used to identify sources of inef�ciency in the ested in achieving, inputs are the means to getting delivery of public services. The PETS methodology was there, and outputs are milestones along the way. complemented by focus group discussions with project The evaluation of outcomes of most agricultural public bene�ciaries, service providers, agricultural staff in the expenditures is critically important but also dif�cult, and local government and MAAIF, project management unit generally such information is not available in Uganda. staff, and political leaders to obtain qualitative feedback There is a time lag in agricultural production, and it is on service quality and project performance. This combi- challenging to link speci�c inputs directly to marginal nation of quantitative (PETS) and qualitative (focus group) values of products. For example, when a vendor is con- information provides an understanding of the factors tracted and undertakes civil work (inputs) to construct a affecting the quality of service provision and makes it wholesale market (output), it is important to have infor- possible to evaluate the operational ef�ciency of project mation about how traders and farmers use the market implementation. For evaluating NAADS, the quantitative and the market’s effect on transaction costs (outcomes). results of the impact evaluation were combined with When farmers use cattle vaccine (input), they expect qualitative assessments in selected districts. that cattle will be healthier (output) and that healthier cattle will be more productive (outcome). Each of these Step 2: Set up the operational framework for analyz- result chains requires a rigorous impact evaluation com- ing technical efficiency. For operational purposes, it is paring targeted and nontargeted households, and the important to distinguish between waste, leakage, and TOOLK IT EX A M P L E 3 8 : C O S T E F F E CT IVE NE SS 59 inef�ciency. Waste and leakage can be dealt with in a dif- PETS, which tracks the flow of funds between different ferent way from inef�ciency—for example, by imposing layers of government. For instance, the procured cattle stronger control systems, developing better incentives do not end up with the designated bene�ciary but with against malpractice, and improving accountability. Public someone else. Leakages in public service increase the service inef�ciency, which stems from poor choices re- cost (input) of achieving a given level of service (output) lated to policy, technology, or management, is harder to and result in a deteriorating measure of ef�ciency. But detect, with assessment demanding more knowledge since leakage is not a consequence of choices in policy, of the service in question. technology, or management, for operational purposes, leakage should be detected separately. Inef�ciency can arise because changing circumstances can cause the optimal choice or combination of inputs Step 3: Select the indicators of efficiency. The most or the appropriate type of output to change over time. simple but highly useful indicator of ef�ciency is unit For example, a government can be doing the right thing cost (the cost of producing one unit of output). Other but using the wrong mechanism (delivery instrument), things being equal, reducing unit cost will improve the resulting in inef�ciency. In contrast, waste occurs when ef�ciency of production. While knowing the level of unit a completely unnecessary and avoidable cost is borne costs is important, it is also necessary to know whether by the public sector. It can arise from many sources, the current level of unit costs is appropriate and to know such as weaknesses in the system, weak capacity, low which factors most affect variations in unit cost across levels of management accountability, improper planning, different projects. There are several ways to assess and corruption. Examples of waste include (1) duplica- whether the level of current unit costs is appropriate. tion of administrative functions, either between MAAIF For instance, unit costs for procured goods can be com- and local governments or between MAAIF and other pared with market prices. Unit costs of centrally and ministries and agencies; (2) unnecessary delays and locally prepared items may also be compared. For civil contract disputes in project implementation, which lead work such as building roads, unit costs can be compared to cost overruns; (3) improper appraisal and feasibility with other projects and guidelines from the Ministry of work, which lead to delays and cost overruns; and (4) Infrastructure. poor asset maintenance, which leads to early replace- ment of physical capital such as irrigation, market, and Step 4: Analyze the findings and develop hypotheses road infrastructure. of what drives inefficiency. It is one thing to identify the factors that mechanically drive unit costs up or down. It Leakages occur when released funds are not spent is quite another thing to identify the underlying causes on the inputs for which they were intended. One way of these factors and prescribe remedial actions at the to look for leakage is to track the flow of �nancial and local, MAAIF, and national government levels. physical resources. This is the methodology used in Example 38: COST EFFECTIVENESS The Ghana Agriculture PEIR (Kolavalli et al. 2009) shows and headquarter levels). Table 29 gives details of the out- how, using the case studies as a representation of their puts, total costs, and unit costs in 2006, as well as the respective regions, the study team estimated the unit assumptions used in estimating the costs. The total costs cost of providing agricultural extension services, including were not signi�cantly different across the three regions farm and home visits, demonstrations, and farmer train- considered here. ing. The cost associated with these services includes the direct operational costs (fuel, materials, and so forth), cost Unit costs can be estimated for other investments in the of the time spent by the agricultural extension agents sector, and comparisons with other countries or regions can (which can be prorated based on their annual salaries and be useful. For example, WDR 2008 shows the unit costs bene�ts), and indirect or overhead costs associated with over time and between Sub-Saharan Africa, which is much supervision and management of the agriculture extension higher than non-Sub-Saharan African countries (see Table agents (accounting for time and related operational costs 30). Understanding the source of these difference can help and investments at the regional, extension directorate, guide improvements to future cost effectiveness. PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 60 PA RT V: — EX A M PLES TABLE 29: Cost of Extension Services Provided by MoFA in Selected Regions, 2006 BRONG-AHAFO REGION NORTHERN REGION WESTERN REGION COSTS (GH¢) Salaries and bene�ts 2,061,752 2,264,505 1,841,860 Operations 203,744 223,780 182,014 Overheada 124,959 137,247 111,631 Total 2,390,455 2,625,532 2,135,506 ACHIEVEMENTS Number of home and farm visits 20,871 71,934 29,076 Number of demonstrations 6,012 628 1,600 Number of farmers trained 58,175 45,526 23,750 UNIT COST (GH¢ PER FARMER REACHED)B Equally weightedc 52 64 93 Unequally weightedd 66 51 85 Source: Authors’ calculation, based on Government Financial Statistics (Of�ce of the Accountant General); 2006 MoFA releases to cost centers; and 2006 annual reports of MoFA cost centers. Note: (a) Based on 11% total headquarters (including directorates) expenditure allocation to the Extension Services Directorate, and 9.5%, 10.5%, and 8.5% of total RADU and DADU expenditure allocation to the Brong-Ahafo, Northern, and Western RADUs and DADUs, respectively; (b) assumes 10 farmers reached in each demonstration; (c) assumes equal weights to home visits, demonstrations, and training in terms of time spent by AEA in those activities; and (d) assumes more weight to home visits (60%) than to demonstrations (25%) and training (15%). TABLE 30: Comparison of Cost Per Hectare of Irrigation 1970-74 1975-79 1980-84 1985-89 1990-94 1995-99 Sub-Saharan Africa Number of projects 3 9 11 15 4 3 Cost per hectare (2000 US$) 4,684 24,496 11,319 7,669 8,287 8,347 Non-Sub-Saharan Africa Number of projects 21 66 75 41 49 6 Cost per hectare (2000 US$) 3,433 4,152 5,174 2,252 3,222 3,506 Sources: African Development Bank and others (2007); Carter and Danert (2006); IFAD (2005); International Water Management Institute (2005); World Bank (2006t). Note: Rates of return on externally �nanced irrigation projects in Sub-Saharan Africa and the rest of the world (two-thirds of which were in Asia) during 1970–99. Example 39: EXPLAINING INEFFICIENCIES The Uganda agriculture PER (World Bank 2010b) presents an to be hampered from the very beginning. MAAIF needs to be analysis of the broad country-level policy environment and much more proactive in bringing these issues to national atten- the sector-level and project-speci�c factors that affect the tion and looking for concerted remedies at the national and local technical ef�ciency of projects. levels. In Uganda, the environment is not entirely conducive to implementing agricultural projects. It takes about one-and-a-half Analysis on country-level policy environment: The prereq- years for Parliament to ratify a loan. Delayed rati�cation re- uisites for launching projects quickly and smoothly are often duces project bene�ts. For the National Livestock Productivity beyond the immediate influence of the Ministry of Agriculture, Improvement Project, for example, a year’s delay in project Animal Industry, and Fishery (MAAIF), including rati�cation by implementation is estimated to reduce the economic rate of Parliament, the provision of counterpart funds, and overall pro- return by 5 percent (from 23 percent to 18 percent). The un- curement and �duciary capacity in the public sector. If foreign timely release of counterpart funds by the Ministry of Finance, loans are rati�ed without delay, and if counterpart funds are Planning, and Economic Development reduces the quality of provided in full and on time, MAAIF and local governments are project implementation. As indicated previously, outturns of enabled to deliver goods and services ef�ciently to bene�cia- government funds ranged from 10 percent to 50 percent for ries. If the opposite is true, a project’s effectiveness is likely TOOLK IT EX A M P L E 4 0 : MO NIT ORING AND E VAL UAT ION 61 the period under review. After Parliament rati�es a project, one for such a large share of the development projects reviewed year is typically needed to establish a procurement and man- (81 percent of MAAIF’s capital budget), MAAIF needs to take agement unit that meets domestic and local requirements and immediate action to devise remedies to deal with these prob- to open special project accounts, especially if a project includes lems. These lessons from experience should be taken into more than one ministry. Most of these issues are relevant to all account when new investment projects are prepared. projects, not just agricultural projects. Yet even if these prob- lems are beyond MAAIF’s immediate control, MAAIF can still Other important lessons on technical inef�ciencies emerge raise them with MFEDP and propose concrete remedies. from studies of how goods and services are delivered to farm- ers and frontline service providers. First, the unit costs of goods Analysis on sector-level factors: The impact of individual proj- procured at the central level is usually 20 percent to 50 percent ects is considerably circumscribed by the lack of a strategic higher than comparable market prices or unit costs at the local approach for using public expenditures to support agriculture. level. Second, the delivery of centrally procured goods is prone One example is the current strategy for improving farmers’ ac- to wastage. The mortality rate of livestock distributed through cess to inputs through subsidized input distribution by develop- several projects is unacceptably high (ranging from 7 percent ment projects, complemented by farmer training. This strategy to 38 percent across interventions). In several instances, a fails to resolve the real causes of farmers’ poor access to in- signi�cant number of cattle and goats died in the �rst months puts, including high transaction costs (caused by poor rural in- after delivery. The distribution of vaccines entailed low value frastructure), slow progress in micro�nance development, and for money and substantial wastage because vaccines were weak technical and business capacity of private agro-dealers. overpriced and storage facilities were inadequate. Records of Once the projects are phased out, inputs are still not accessible. �nancial and physical resource transfers from the central man- Project expenditures are likely to have been wasted, given that agement units to the districts often do not match, suggesting they could have been used for infrastructure or other public wastage, diversion, or leakages, although the source of the investments with more sustainable outcomes. problem remains unclear (possibilities include poor record keeping and thus low accountability, diversion from intended Analysis on project specific factors: Most MAAIF devel- bene�ciaries, corruption, resource misallocation, improper opment projects with infrastructure components are char- accounting, or simply insuf�cient information from project acterized by long delays (three to �ve years) in building and implementation units). Leakages were more pronounced rehabilitating rural infrastructure. Aside from the issues just for in-kind transfers than �nancial transfers, suggesting that mentioned, important problem areas include improper ap- inputs and goods were more easily directed away from their praisal and feasibility work, poor coordination of preparation intended bene�ciaries or uses. A key lesson from these expe- and implementation between MAAIF and local governments, riences is to decentralize procurement to reduce costs and to inadequate operating budgets for local technical staff, ineffec- place resources directly at the command of the bene�ciaries tive project procurement, and problems related to land tenure.17 to improve accountability. For example, such costs are lower These result in high cost overruns, low-quality work, and for NAADS because most procurement is decentralized and other kinds of waste. Given that capital expenditures account involves the local communities and bene�ciaries. Example 40: MONITORING AND EVALUATION There is general agreement that weak M&E systems are one sector is crucial to (i) generate the disaggregated information of the most fragile links in the entire chain of expenditure required to ensure more ef�cient and equitable sectoral alloca- management. While this weakness can be observed in all sec- tions, (ii) improve the implementation of expenditure programs, tors of an economy, it is especially serious in the agricultural and (iii) enhance the governance and accountability of expendi- sector, given its great diversity of actors and stakeholders. A tures (using various mechanisms). Expenditure data disaggre- Monitoring and Evaluation system (M&E) for the agricultural gated by program, project, activity, or region are essential for more rigorous expenditure analyses. A phased approach can 17 Land tenure problems are often related to improper appraisal and be taken because performance-based budgeting systems take weak coordination of activities with local governments. time to develop. Emphasis tends to be placed on improving PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 62 PA RT V: — EX A M PLES monitoring in high-priority programs and projects, which will but the actual implementation is facing practical challenges, certainly contribute to more ef�cient public expenditures. especially in ARD programs. These are related to (i) the limited availability of experienced talent in M&E, which hinders the The Ghana agriculture PEIR (Kolavalli et al. 2009) contains good implementation and oversight of good M&E; (ii) the insuf�cient insights into underlying problems with M&E and describes feedback of the evaluation results into managerial and budget- strategies for enhancing the role of M&E systems as instru- ary decision making; (iii) the fact that evaluations are limited to ments for promoting greater ef�ciencies and impacts on agri- single programs and not at the level of the entire PEC to track cultural expenditures. The review identi�es a need to sharpen progress in rural development; (iv) the absence of a uni�ed the focus on a few strategic outcome indicators. The Food and information system and database for different ARD programs, Agriculture Sector Development Policy II (FASDEP II) includes or even different agricultural support programs, to coordinate a comprehensive list of almost 60 indicators. Collecting and the distribution of resources at the bene�ciary level; and (v) an- managing information on such a broad range of indicators, even nual evaluations that are costly to implement and unnecessary though each indicator might be useful in some sense, is a rather in most cases. The review also contains recommendations challenging task. MoFA may wish to focus M&E reporting by to address some of these issues: (i) evaluation of single pro- strategically selecting a few key outcome-related indicators in grams and of the entire PEC, (ii) development and integration addition to selected key output-oriented indicators. There is a of databases of the bene�ciaries of ARD programs, (iii) deeper need to “mainstream� M&E indicators with other management program evaluations at longer intervals instead of the current processes by making them relevant for managers as source annual cycle, and (iv) systematic use of evaluation results for information to perform their tasks more effectively. It would managerial and budgetary decisions based on action agendas be useful to have indicators that are relevant for managers in agreed upon by the evaluators, the implementing agency, and districts and regions to monitor the activities and outcomes. a third party. Adherence to these agendas should be monitored The Mexico Agriculture PER (World Bank 2009a) devotes an closely. As a logical follow-up to the law on sustainable rural entire chapter to assessing M&E and institutional aspects of development (Ley de Desarrollo Rural Sustentable), it was ARD programs. The rapid development of the ambitious insti- suggested to start a “rural proo�ng� system to anticipate and tutional M&E systems, especially since 2006, is remarkable, monitor national policies’ impact on the rural sector. Example 41: INCIDENCE ANALYSIS The following extract from the Mexico PER (World Bank ordering corresponding to the different data sources. In the 2009a) provides an example of an incidence analysis with case of administrative data, producers are ordered by land- a primary focus on the equity of public spending across pro- holding, which is the only measure of wealth and welfare ducers and households. reported in these data. In the case of the national house- hold surveys, bene�ts received are analyzed by population Data Requirements: The availability of household and deciles ordered by (current) income per capita. Household producer databases reporting both agricultural support pro- and producer data allow coverage of the principal ARD pro- grams and a relevant measure of household and producer grams, including the principal agricultural support programs well-being or wealth to allow distributive analysis is limited. and rural social programs like Oportunidades and the Three kinds of data sources are used, which are comple- mentary but not strictly comparable: (i) general household surveys including ARD transfers (the National Household 18 General household surveys have the important advantage of be- ing nationally representative and including high-quality data on Income and Expenditure Survey, or ENIGH, 2004 and 2006 income and other measures of household well-being, but their surveys and Encuesta Nacional sobre los Niveles de Vida sample size is not designed to capture speci�c transfer programs de los Hogares, or ENNVIH, 2002 survey), (ii) evaluation accurately, especially when these have limited coverage or con- centrate a large share of their bene�ts among a relatively small surveys for certain programs (Alianza, Oportunidades); proportion of producers. The other two sources are of course and (iii) administrative data of the programs (Programa designed to capture the program bene�ciaries and transfers ac- de Apoyos Directos al Campo, or PROCAMPO, Ingreso curately but are not nationally representative and generally con- tain limited or no household income data. The analysis obtained Objetivo, Alianza Rural Development Program, or PDR).18 from the three sources must therefore be interpreted carefully The distribution of bene�ts is analyzed using two types of and in a complementary way. TOOLK IT EX A M P L E 4 1 : I N C I D E NCE ANALYSIS 63 FIGURE 10: Comparison of Distribution of PROCAMPO and Oportunidades Transfer for Rural Population Decile (% 2006) (Ordered from Left to Right by Pretransfer Income Per Capita) Rural distribution 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% 80.0% 90.0% 100.0% Oportunidades Procampo Income Source: Scott (2008); calculations based on data by ENIGH 2006. Seasonal Employment Program (PET), as well as a number progressive than the income distribution, Oportunidades of other smaller rural development programs. Using the is highly progressive (pro-poor) while PROCAMPO is distribution of irrigated land as a proxy for the distribution regressive. of hydrological, hydro-agricultural, and tarifa 9 electricity subsidies, the agricultural support programs covered in Analysis using household surveys and administrative this incidence analysis represent 75 percent of Agricultural data (9 social and rural development programs): ARD Public Expenditure. expenditures are an inef�cient redistributive instrument, whereas agriculture public expenditure is not only inef�- First analysis using household surveys (PROCAMPO cient but actually contributes to increasing income inequal- and Oportunidades): Figure 10 compares the distribu- ity. To assess the global impact of ARD expenditures on tion of PROCAMPO and Oportunidades resources using rural income inequality, the agriculture public expenditure ENIGH data.19 The contrast between Oportunidades and programs are contrasted with nine social and rural devel- PROCAMPO is evident from their concentration curves opment programs reported in ENIGH 2006 and in a special in the income space. Although both programs are more “Social Module� commissioned by the Ministry of Social Development (SEDESOL) with ENIGH 2004. Concentration coef�cients (CCs) are used to measure equity.20 As shown 19 The distribution of PROCAMPO bene�ts differs substantially according to whether ENIGH data or administrative data of the Agricultural Commercialization Supports and Services program 20 The concentration coef�cient (CC) is a generalized form of the (Apoyos y Servicios a la Comercialización Agropecuaria, ASER- Gini coef�cient that shows the concentration of a particular CA) are used. The latter data result is a substantially more re- income source x (in this case payment from ARD programs) gressive distribution than the former. This is because, like most when recipients are ranked by an index y (in this case pretransfer household surveys, ENIGH does not capture household incomes income). The CC ranges from −1 when all transfers are received at the top of the distribution very well. In view of this difference, by the poorest households, through 0 when all households we report in some sections two sets of results for PROCAMPO, receive the same amount of transfer income, to +1 when all one using ENIGH data and the other using ASERCA data. transfers are received by the richest households. PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 64 PA RT V: — EX A M PLES FIGURE 11: Concentration Coef�cients of ARD Expenditures, Income and Land (2006) Oportunidades -0.315 Piso Firme -0.312 IMSS-Oportunidades -0.315 PET -0.309 Total RD -0.265 Desayunos DIF -0.232 Despensa DIF -0.142 Seguro Popular -0.137 Vivienda Rural -0.073 Procampo (ENIGH) 0.117 Opciones Productivas 0.121 Total ARD 0.238 Pre-transfer income 0.349 Credito a la Palabra 0.454 Procampo (ASERCA) 0.502 Rainfed land 0.503 Total land 0.524 Total APE 0.584 Irrigated land 0.657 Ingreso Objetivo 0.809 -45% -25% -5% 15% 35% 55% 75% Source: Scott (2008), using ENIGH 2006. Note: Data for 2004 are used for the following programs: Piso Firme (Firm Foundation), PET, the Pantries and Breakfasts (Desayunos y Despensas) Program of the National Service for Integral Family Development (Desarrollo Integral de la Familia, DIF), Vivienda Rural (Rural Housing), Opciones Productivas, and Crédito a la Palabra. ENIGH 2004 (Social Module), and ASERCA Bene�ciary databases. FIGURE 12: Concentration Curves of Major Agricultural Programs and Land (%) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Procampo Ingreso Objetivo Alianza-Desarrollo Rural Irrigated Land Rainfed Land Source: Scott (2008). TOOLK IT EX A M P L E 4 2 : I M PA CT E VAL UAT ION 65 in Figure 11, the CC of ARD expenditures is regressive derived from the administrative data for PROCAMPO, in absolute terms (CC = 0.24) but still progressive rela- Ingreso Objetivo, and Alianza PDR. For comparative tive to the distribution of pretransfer income (CC = 0.35). purposes, the distribution of rain-fed and irrigated Agriculture public expenditure is more regressive (CC = land is also shown. There is an extreme concentra- 0.58) than the distribution of pretransfer income (CC = tion of bene�ts for all programs. The poorest pro- 0.35), thus actually contributing to increasing income ducer decile (in terms of land) receives a tenth of a inequality. By contrast, rural development expenditure percentage point of Ingreso Objetivo and similarly (RDE) is progressive (pro-poor) in absolute terms (CC = insigni�cant fractions of the other programs. At the −0.26), with the notable exception of the two productive other end of the distribution, the richest producer SEDESOL programs, Opciones Productivas (Productive decile receives the following shares of transfers: Alternatives) (CC = 0.12) and Crédito a la Palabra (Credit (i) 45 percent of PROCAMPO, (ii) 55 percent of the on Demand) (CC = 0.45). Alianza PDR, (iii) 60 percent of energy and hydro- logical subsidies (proxied by irrigated land), and Analysis using program evaluation surveys and (iv) 80 percent of Ingreso Objetivo. administrative data: To compare the principal APE programs, Figure 12 presents concentration curves Example 42: IMPACT EVALUATION Most agriculture PERs lack resources to undertake special- analyze the impact of agricultural research and extension in ized data-intensive, quantitative impact evaluation. When lim- the country. The agriculture PER also undertakes qualitative ited data are available, the evaluation of results often relies assessment of the core agricultural subsectors in relation to on existing impact studies or on a comprehensive literature their institutional and public �nance characteristics. These review of key �ndings of the impacts of public expenditures subsectors include agricultural input supply and subsidies, of similar projects undertaken elsewhere. agricultural �nancial services, and national strategic food reserves. The Vietnam Agriculture PER (Volume II) (World Bank 2005b) relied on existing impact studies showing that The Ghana Agriculture PER (Benin et al. 2008) highlights government expenditures have been crucial to growth that lack of information limits the possibility to conduct im- and poverty reduction in Vietnam. Two studies have mea- pact analysis on key subsectors. Required data include (i) sured the impact of this spending. The first, by Fan et time-series data, disaggregated by region and district, on al. (2004), found that government spending on irrigation, government and other public expenditures on the various roads, and agricultural research has contributed to both agricultural subsectors (crops, livestock, �shery, forestry, and agriculture growth and poverty reduction. The second natural resource management) and (ii) data on functions or study, by Barker et al. (2002), estimated determinants activities (research, extension and training, marketing, inputs of agriculture growth for the same period and found that such as seed, fertilizer, and chemicals, and infrastructure like public investment in irrigation was the most important irrigation, feeder roads, marketing information system, and source of agriculture growth (accounting for 28 percent post-harvest handling). However, the Ghana agriculture PER of the growth), followed by agriculture research (27 per- uses various quantitative methods to assess the effective- cent). An important finding from these two studies is that ness of public spending, including bene�t-incidence analysis while irrigation has been the largest source of growth, the of impacts at the local level. Data from the �fth Ghana Living aggregate irrigation investment program has become un- Standards Survey were used to understand the geographi- economic. The conclusion is that the sector expenditure cal distribution of households’ income and showed that the program should select only those irrigation investments effects of ongoing poverty reduction strategy were mainly with acceptable rates of return and that savings should bene�ting the south of the country, deepening the divide be reallocated to high-return activities, particularly agricul- between north and south. tural research. Econometric analysis from Benin et al. (2008) was used to The Nigeria Agriculture PER (World Bank 2008a) summarizes guide new efforts to target agricultural and rural development key �ndings from a comprehensive literature review used to analysis (Figure 13). A cost-bene�t analysis was undertaken. PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 66 PA RT V: — EX A M PLES FIGURE 13: Bene�t-Cost Analysis of Public different types of agriculture expenditures (e.g., extension, Expenditures in Ghana research, input support, or irrigation) in different regions. 47.4 Empirical Results: The Mexico Agriculture PER (World Bank 2009a) presents empirical results showing that the impact of 33.1 Mexico’s agricultural public expenditure (APE) on agricultural 22.5 20.7 growth seems comparatively small, that the impact of APE 16.8 on total factor productivity (TFP) is also low, and that a nega- tive correlation exists between public spending on private goods and state agricultural growth (Figure 14). Coastal Forest South North Ghana The APE on TFP was also found to be low for Mexico. This Savanna Savanna was calculated by comparing growth rates in agricultural Source: Benin et al. (2008). Note: Bene�ts are measured as the total value of household agricultural production and income (from crops. livestock, forestry, GDP and TFP from 1981–2001 with average APE and AGDP �shing, hunting, and so forth) per capita in 2005–06. Costs are Government expenditure rates for 1985–2001, ordering countries by the of Ghana and AgSSIP expenditures at the regional and district levels between 2000 and 2006. latter (Figure 15). Results showed that (i) different agro-ecological zones have A negative correlation between spending in private ag- different comparative advantages in production and that (ii) ricultural goods and state-level AGDP growth was found. there is a trade-off between allocating resources to areas Controlling for the levels of mechanization, fertilization, and where the growth effects are highest (Southern Savannah) expenditures in public goods, a 10 percent increase in APE and areas where the prevalence of poverty is highest in private goods as a percentage of the value of agricultural (Northern Savannah). Results suggested that an effective production is associated with a 2.6 percent reduction in strategy should take into account returns associated with AGDP growth. FIGURE 14: Agricultural Growth Rate in Latin America and Caribbean in Relation to Agricultural Public Expenditure (as a Percentage of GDP), Average 1997–2001 7.0 6.0 Dominican Rep. Panama 5.0 AGDP Growth Rates (%) Peru Nicaragua 4.0 Brazil Ecuador Costa Rica 3.0 Chile Guatemala Argentina LAC 2.0 Paraguay Bolivia Mexico Venezuela 1.0 Colombia Honduras El Salvador 0.0 0.0 5.0 10.0 15.0 20.0 25.0 % of AGDP Source: Valdes (2008a) using FAO (2005). Agricultural GDP growth rates are based on World Development Indicators Database TOOLK IT EX A M P L E 4 3 : MO RE P UBL IC- P RIVAT E PART NE RSHIPS 67 FIGURE 15: Average Annual Agricultural GDP and TFP Growth Rates and Distribution of APE/ AGDP (as a Percentage, 1981/5–2001) 5.0% 40% 35% 4.0% 30% 3.0% 25% GDP, TFP growth APE/GDP (%) 2.0% 20% 15% 1.0% 10% 0% 5% o ua Re y Co blic ca a il en e la ca a s Ec u r ia Pa na ay a do ra m il al bi ic an gua r az ue liv Pe Ri ai ag ti gu Ch m du ex na m ua Br pu en m ez Bo ar in Uru te ra lo sta M Pa on Ja rg ic ua Co H A N V G ic -10% 0% om D GDP 1981-2001 TFP 1981-2001 APE/GDP 1985-2001 Source: APE and Agricultural GDP from FAO (2005); Agricultural TFP growth rates from Avila and Evenson (2004). Example 43: PUBLIC-PRIVATE PARTNERSHIPS Different types of public-private partnerships (PPPs) are and infrastructure, as with Banrural, which provides important for delivering better and more ef�cient agricul- �nancial services to smallholders in Guatemala. Not all tural services, for sharing the cost of public expenditures, such programs are suitable for targeting the poor, but and for stimulating private investment. PPPs are part of a they can still free up public resources, which can then broader strategy for the state to shift toward more effec- be targeted toward the poor under other institutional tive roles and arrangements to improve the ef�ciency and arrangements. equity of public expenditures on agriculture. The partner- ƒ PPPs involve third sector organizations, such as ship approaches listed below are useful for enhancing the producer organizations, along with public agencies composition and impact of public expenditures and broad- and private businesses, as with Ghana’s Sustainable ening the participation of stakeholders. These approaches Uptake of Cassava as an Industrial Commodity can be applied in different ways, accompanied by appro- Project. priate funding mechanisms and “rules of the game,� de- ƒ Devolving management authority to user groups is pending on the particular country and sector (WDR 2008, widely applied in natural resource and irrigation man- para. 11.39): agement, as with the irrigation scheme of the Of�ce du Niger, in Mali. The opportunities and challenges of ƒ Contracting out is suitable for functions that require devolution to user groups are comparable to those of public �nance but not necessarily public provision. It is community-driven development. increasingly used for agricultural advisory services, as ƒ Privatization works best where state involvement in Uganda. is no longer needed. Veterinary services are a good ƒ PPPs go beyond outsourcing, creating joint responsi- example. Across 10 Sub-Saharan countries, the bilities for �nancing and providing agricultural services PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 68 PA RT V: — EX A M PLES number of private veterinarians increased from Service cooperatives, formed and owned by producers, can pro- 70 in the mid-1980s to 1,780 in 2004. Community- vide pro-poor agricultural services. In India, dairy cooperatives based paraveterinarians have improved service provide services to more than 12 million households, bene�ting provision, too. women particularly because of their role in dairy farming. Example 44: MATCHING GRANTS TO LEVERAGE RESOURCES Morris et al. (2007) provides an overview of several prin- �rms. Business development grants are gener- ciples that should guide effective matching grant schemes. ally designed to improve business performance by Matching grants are another form of market-smart subsidies encouraging investment in business processes or and can be used to fund start-up activities. Under the typical building staff capacity. In the fertilizer sector, they matching grant program, the recipient of the matching grant can be used to encourage risk-taking, for example, by (which can be an individual, a group of individuals, a �rm, stipulating that recipients must expand fertilizer distri- or a not-for-pro�t organization) is required to mobilize local bution into new market areas or carry new product resources, which are then combined with external funding lines. (the matching grant) to pursue a de�ned set of objectives. 3. Community-driven development (CDD) grants can Because matching grants are usually provided only once to a be used to support agricultural income generation at given recipient, they do not imply a recurring commitment of the community level. Direct transfers for productive resources on the part of the granting agency, meaning that activities, including investment in soil fertility, may it is usually easy to devise an exit strategy. Most matching be justi�ed if they reduce poverty by targeting the grant programs simply terminate as soon as their funds are very poor and if they stimulate some private sector fully disbursed. Matching grants can be particularly useful development. By encouraging recipients to engage in situations where credit markets are weak, �nancial risk in collective action to obtain technical support or is high, or both. In such cases, the public or donor funds are arrange the purchase of inputs or market outputs, used to leverage private investments until �nancial and risk CDD grants can provide options that are not open to markets emerge. individuals acting alone. Matching grants have some drawbacks, however. Matching grants can be used for promoting fertilizer market Administration of matching grant programs can be costly, development at three levels in the supply chain. They can be because it is often dif�cult to distinguish between needy used to foster (1) agricultural innovation, (2) business devel- applicants who really are unable to mobilize investment re- opment, and (3) community-driven investment. sources on their own and more well-heeled applicants who 1. Agricultural innovation grants can be awarded—usu- are perfectly capable of accessing commercial credit. If this ally through competitive procedures—directly to pro- distinction is not maintained, matching grants can end up go- ducer organizations or to partnerships of farmers and ing to creditworthy recipients, thereby crowding out private research and extension services to develop and test sector lending. Finally, matching grant programs are ripe for innovations related to soil fertility and other types political manipulation, and administrators of matching grant of innovations, depending on farmers’ priorities. programs often come under pressure to steer grants to Farmers’ involvement ensures that local demand those with political and economic influence. increasingly guides the development and dissemina- Because matching grants are generally market smart, or at tion of technology. The approach shows promise, but worst market neutral, they have been growing in popular- it requires substantial local public and private �nancial ity. However, considerable care is needed in designing a support and a highly professional unit to administer matching grant program (McKean and Ostrom 1995; Van der the program. Grants may be managed completely Meer and Noordam 2004). Experience suggests that grant by farmer groups to allow them to test and adapt programs are much more likely to be successful when some new technologies, as is being implemented for soil simple guidelines are followed: fertility management and complementary practices in Tanzania. ƒ The type of grant selected should be tailored to local 2. Business development grants can be awarded circumstances, including the quality of local technical competitively to individual entrepreneurs or to entire expertise. TOOLK IT EX A M P L E 4 5 : E N V I RONME NTAL SE RVICE S AND CL IMATE CH A NGE 69 ƒ The economic rationale for public co�nancing should and staf�ng costs, administering grants, M&E, be articulated clearly. institutional development, and training to prepare ƒ Rigorous and transparent eligibility criteria and assess- proposals. ment procedures are important, as is competent fund ƒ Often there are trade-offs between ensuring that management with clear objectives and procedures. operations are cost-effective and ensuring that they ƒ To ensure that proposals address private sector are conducted with accountability and transparency, priorities, an initial investment may be desirable to and some balance must be sought to ensure that both build the capacity of potential recipients so that they objectives are met. can develop and defend proposals that identify key ƒ Grant funding is typically most effective when it is problems, critically evaluate alternative solutions, and complemented by other funding mechanisms. In justify grant funding based on clear net bene�ts in many cases, block funding from the public sector will economic and social terms. still be needed to address core public goods. ƒ A signi�cant learning period should be allowed to enable ƒ From the beginning, grant schemes should have stakeholders to gain experience in working with the a clear disengagement strategy. Proposals should grant scheme and to make adjustments as necessary. include action steps that can be monitored, with ƒ Administrative costs must be controlled rigorously milestones and targets indicating when objectives to create a sustainable private market for support have been achieved. Generally grants should be services within a limited period. Fifteen percent for a �xed period, such as three years, and not appears to be an international norm for start-up renewable. Example 45: ENVIRONMENTAL SERVICES AND CLIMATE CHANGE The agriculture sector is expected to be heavily impacted on emissions, mandated best practices, property rights al- by climate change requiring for adaptation measures to be location, and support for institutions for common property taken. At the same time, the sector accounts for over 30 management. percent of greenhouse gas emissions and uses over 70 percent In agriculture and forestry, environmental taxes are of available water resources. Even though countries must ad- rarely used due to the high cost of monitoring emissions. dress these environmental issues, mainstreaming has been Emissions permits outside agriculture tend to be pollutant slow, given the complex scienti�c nature of climate change speci�c. Better known in the natural resource sector are and the uncertainty associated with it and due to the fact that best management practices to reduce environmental dam- an international carbon trading system is not in place nor are age. Such practices are now more common—for example, natural resources properly priced. when managing native forests—and are often pushed by The Mexico PER (World Bank 2009a) has the following tech- tourism-linked considerations and by consumer preferences nical note on payment for Environmental Services in Natural in wealthier countries valuing not only the product but the Resource Sectors: Managing natural assets well is dif�cult environmental friendliness of production. In fact, there is a growing emphasis on private, third-party certi�cation and because the value of many—if not most—environmental voluntary implementation of standards, as with organic services cannot be internalized by private actors; they are produce in fresh fruits and vegetables for exports. This is externalities and public goods. Individuals in society general- due in part to the normal slowness of governments as many ly value natural assets, but the incentives guiding private ac- developing countries adjust to the design and enforcement tion rarely reflect their value to society as a whole. The key to of best practices. better managing natural assets is to design institutions that can measure the social bene�ts and costs of environmen- Recent policies have emphasized both market-oriented and tal services and translate that information into appropriate government-supported subsidies of environmental goods incentives for individual decision makers, such as farmers. and services. These policies, such as incentives for better The classic policy proposals for better management focused land management, focus on Payments for Environment on reducing negative externalities are environmental taxes Services (PES) provided by those who bene�t from such PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 70 PA RT V: — EX A M PLES services, including local, regional, and global bene�ciaries. more than 280,000 hectares of private forests, with more PES programs can be found increasingly in both developed than 800,000 hectares pending, spending US$57 million and developing countries, ranging from contracts between between 1997 and 2002. Forest conservation contracts paid farmers and local industries whose pro�ts are sensitive to over US$210 per hectare, and reforestation contracts paid water availability, to contracts between one country’s indus- US$538 per hectare. Funding came primarily from fuel taxes, try and groups in distant countries seeking carbon sequestra- complemented by international donors. tion. According to the FAO, there are four main PES markets: climate change mitigation, watershed services, biodiversity One special consideration is the potential role of the PES in conservation, and landscape aesthetics. In developing coun- alleviating poverty. Associated with poverty are low levels tries’ markets for reducing carbon emission and conserving of education, ill-de�ned property rights over natural assets, biodiversity are potential sources of new revenue for agri- and a lack of capital that could support households during culture. The ultimate bene�ciaries of environmental services long-term investments in natural assets (such as planting are usually dispersed, so intermediaries such as govern- trees). Poverty often leads to environmental degradation. ments and international public and private concerned enti- Being able to target poor farmers for PES would bring both ties are required. To date, PES programs have been pushed environmental bene�ts and poverty reduction. Yet while by governments and are geographically focused. They are targeting the poor on paper might be logical, implement- usually direct payments in exchange for better land manage- ing a PES plan would be constrained by the very structural ment, such as soil and water conservation measures, and characteristics that link poverty to environmental degrada- tree plantations for carbon sequestration. Bene�ciaries can tion in the �rst place: ill-de�ned property rights that make pay via a surcharge to water bills or with fees for park visi- contracting and long-term projects uncertain for both buyer tors. There are two key steps for designing ef�cient PES pro- and seller of the service; small farm holdings that increase grams. First is targeting farmers or other managers of natural the transaction costs per unit of compensation and reduce assets whose adoption of better management practices the viability of better management (such as crop rotation) can yield the highest environmental bene�ts for the lowest and thus increase the required compensation to entice costs. Second is the structure of compensation, which not farmers to better management practices; and lack of credit only induces better management but also is self-enforcing by as a short-term backup makes farmers more risk averse to reducing monitoring costs and assuring compliance. adopting a new system of management. Nevertheless, the Bank, while recognizing that PES is not in itself an antipov- Perhaps the most studied case of PES in Latin America is the erty strategy, has suggested some pro-poor considerations Costa Rican National Forestry Financing Fund, which began in PES design. In particular, it suggests that one could de- in 1997 and has led to similar efforts in other countries. The vise speci�c mechanisms to counter high transaction costs, fund of�cially recognizes that forest owners provide a bundle understand the social context in which a PES would be of environmental services, including watershed protection, applied to avoid adverse impacts on the poor, and design biodiversity conservation, scenic beauty, and carbon �xation appropriate remedial measures. Making PES programs pro- and sequestration. The PES system arose when the govern- poor, however, might increase their costs, so funding must ment was searching for sustainable funding for both forest be found beyond the bene�ciaries of environment services. conservation and aid to the forestry sector. A 2002 evaluation Donors who are more directly concerned about poverty of the program found that the Costa Rica PES had contracted could �nance these additional costs. Example 46: DECOUPLING AGRICULTURAL SUBSIDIES Decoupled payments are less distorting than output-linked follow market signals and produce goods for which there is forms of support such as tariff protection. The process on demand but also earn a baseline living income in the form of decoupling has varied signi�cantly by commodity. The most a government transfer. common strategy that has been used by countries to phase out price supports is to provide farmers with direct cash trans- Key lessons learned from countries that have success- fers, shifting from production-based support to direct income fully eliminated crop price supports include (i) adequately support. Under a direct transfer program, payment reception compensate farmers for potential income loss, (ii) educate is not at all tied to production. This means that farmers can them about new support programs, and (iii) provide technical TOOLK IT EX A M P L E 4 7 : T H E CASE F OR AND AGAINST F E RT ILIZER SUBS ID IES 71 assistance to help them adjust. Turkey and Mexico have both write-offs, and substantial reduction in state-�nanced crop used direct payments to replace production subsidies: Turkey purchases and �scal transfers to cover the losses of state for several crops, one of the most important being tobacco, agricultural enterprises. At the same time, the DIS Program and Mexico for staple food crops. The transition process was expanded to become the main instrument of rural income made easier in Mexico, which offered technical assistance support. to help farmers invest their direct payments in productive Simulations for Lebanon and Greece show that decoupling activities with high rates of return. payments from production would likely lead to widespread The Turkey Agriculture PER (World Bank 2006e) noted that abandonment of tobacco cultivation (IDRC 2008; Psaltopoulos the government started in 2000 to change the system of ag- and Eudokia 2005) and that the household level impacts of ricultural support to promote �scal stabilization and allocative decoupling support programs from production will depend on ef�ciency. There was great concern that farmers would not what alternative economic activities are available and what be adequately compensated when subsidies were replaced requirements are attached to receiving direct payments. The with a direct income support (DIS) program, creating a major Lebanon Agriculture PER (World Bank 2010c) noted that, challenge for reform. By the end of 2002, the implementa- although tobacco subsidy removal would undoubtedly cause tion of the reform of agricultural transfers had signi�cantly re- friction; negative effects could be mitigated by tying receipt duced arti�cial incentives for inputs and particular crops and of uncoupled support to farmers remaining on their land by switched the main focus of agricultural transfers to the DIS accompanying a DIS program with a wide-spread information Program. By 2004, the annual �scal cost had been reduced campaign to ensure that eligible farmers understand how to from US$6.1 billion (3.1 percent of GDP) in 1999 to US$2.4 receive payment and by providing technical assistance to billion (0.8 percent of GDP). This has been achieved largely help farmers invest in alternatives to tobacco production that through the elimination of credit subsidies, reduction of debt stimulate investment and labor demand. Example 47: THE CASE FOR AND AGAINST FERTILIZER SUBSIDIES Fertilizer subsidies often comprise a signi�cant share of pub- that fertilizer subsidies often: (i) crowd out the private lic budgets, increasingly so in many African countries. In this sector by undermining incentives for �rms to invest in context, it may be useful to understand the arguments for production and marketing; (ii) encourage rent seeking, and against fertilizer subsidies to help dialogues with govern- making them magnets for corruption and abuse; (iii) tend ment counterparts. This section is from Morris et al. 2007. to be captured by wealthier farmers, leading to a regres- sive distribution of bene�ts; (iv) encourage inef�cient Arguments in favor of fertilizer subsidies: Three main use of fertilizer at the farm level; and (v) are dif�cult to economic arguments are typically used to justify the use target, resulting in a high level of leakage to neighboring of fertilizer subsidies. First, they can kick-start innova- countries. tion and stimulate greater demand for fertilizer. Second, they can help overcome missing or imperfect markets High administrative and fiscal costs: In addition to for fertilizer. Third, they can compensate farmers for these more practical problems, the larger issue—and positive externalities. A fourth argument is more politi- perhaps the most relevant from a public expenditure cal: Fertilizer subsidies are needed to level the playing point of view—is that fertilizer subsidy programs have �eld for farmers who must compete in global markets extremely high administrative and �scal costs. Most against competitors who bene�t from subsidies and empirical studies that have assessed fertilizer subsidy other forms of support in their own countries. programs have concluded that the costs outweigh the bene�ts. This is likely to be the case in Nigeria as well, Common problems with fertilizer subsidy programs: where considerable portions of the federal and state Nigeria’s fertilizer subsidy programs have not been rigor- agriculture budgets have been used to support fertil- ously evaluated, but anecdotal evidence suggests that izer subsidy programs. Much of the fertilizer distributed these programs have experienced the same problems through these programs seems to have gone to large- that have plagued fertilizer subsidy programs every- scale farmers who were capable of purchasing fertilizer where. Studies carried out in many countries have shown PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 72 PA RT V: — EX A M PLES on commercial terms, rather than to the resource-con- measures to improve the supply of fertilizer, but also strained smallholders whom the schemes were in- measures to strengthen demand for fertilizer. For this tended to bene�t. Meanwhile, there was public under- reason, building fertilizer markets must go hand-in- investment in core public goods and services, including hand with building output markets and linking farmers agricultural research and extension, market information to those markets. systems, and rural infrastructure. Guiding principles for design of “market-smart� When subsidies may be justified: This is not to say subsidies: When used as part of a broader strategy to that fertilizer subsidies should never be used. Although address the binding constraints on supply and demand, the long-term objective of policy makers must be to well-designed fertilizer subsidies can play a useful role. support the emergence of viable fertilizer markets led But fertilizer subsidies should always be “market smart,� by the private sector, subsidies may be justi�able on so that they contribute to the development of viable and a temporary basis to help overcome market failures. competitive private-sector-led fertilizer markets. They Yet, the �rst policy option should be to address the should be targeted at resource-poor farmers to encour- underlying cause of market failure/problem rather than age incremental use of fertilizer by those who would trying to compensate for it with subsidies. If they are otherwise not use it; otherwise commercial sales will be used, however, it should be in ways that encourage displaced. They should not distort relative prices of dif- the ef�cient uptake of fertilizer as part of an integrated ferent inputs so as to encourage economically inef�cient package of improved crop-production technologies. use. And they should always be temporary, introduced Fertilizer promotion programs therefore must be com- for a limited period, with a clear schedule for phasing out prehensive and multifaceted, encompassing not only when they have achieved their purpose. Example 48: ACCOUNTABILITY MECHANISMS Experience shows that appropriate accountability mecha- the top political leadership, especially the President’s Of�ce, nisms (applied at various levels and for diverse stakeholders) was high. The budget processes also introduced account- can play an important role in enhancing the ef�ciency and ability to parliamentarians, particularly for the ful�llment of impact of expenditures. Up to now, relatively few countries targets in the budget statement. However, accountability to consistently and effectively used sound accountability mech- the District Assembly has not yet been established in any anisms. The Ghana Agriculture PEIR (Kolavalli et al. 2009) formal sense, even though �eld staff interact with elected generated useful information and constructive suggestions District Assembly members. Food and Agriculture Budget related to accountability mechanisms. It was noted that support triggered accountability to donors as well. It was MoFA was making some efforts to increase the accountabil- noted this process should be strengthened and strongly ity of various units in the organization. These included (i) the linked to outcomes. The review also showed that routine establishment of annual “performance review� workshops, processes that could enhance responsibility (such as regular to which donors and other stakeholders were also invited; individual assessments and evaluation of the quality of the and (ii) the development of plans, setting up of targets, and work performed) could have been improved and that ac- assessment of performance in relation to previously speci�ed knowledgement of good performances would increase staff outcomes. The review found that MoFA’s accountability to accountability. Example 49: INSTITUTIONALIZING PERS An emerging lesson on the sustainable impacts of PERs is reviews to the Government of Rwanda is to institutional- the importance of encouraging governments to institutional- ize PERs in Rwanda (World Bank 2009d). It is important ize PERs as part of their annual budgetary process. In addi- to recognize that these efforts need to be coordinated at tion, deeper analytical reviews are required periodically even the macro level, in close collaboration with the Ministry of in the countries where annual reviews are the norm because Finance, and that they must ensure a strong macro-sectoral in-depth PERs cannot be carried every year. A core compo- linkage and complementarity. In cases where there is strong nent of the World Bank’s technical assistance in expenditure commitment and leadership at the level of the ministry of TOOLK IT EX A M P L E 5 0 : D I S S E MINAT ION ST RAT E GY 73 agriculture, some aspects of institutionalization could be initi- These sector BFPs would then feed into development of the ated (as part of the annual budgetary process) and perhaps macro-level BFP, as would the macro-level budget review and serve as a catalyst for the Ministry of Finance to take broader any related PERs conducted at the macro level. Figure 16 leadership. summarizes the associated proposed calendar. Experience from countries that have institutionalized PERs Based on experience in institutionalizing PERs in various shows that aligning these reviews to the budget calendar countries, the process appears to require that eight compo- and ensuring results feed into budget preparation and policy nents and supporting processes are in place: (i) an overall discussions are critical for capacity development. The PER framework and procedure with clearly articulated roles and process in Tanzania, Ghana, and Uganda has, in general, in- responsibilities for the key stakeholders, (ii) a mechanism for formed policy discussions and prioritization for preparation of coordinating quality assurance and dissemination, (iii) plan- annual budgets and also for the medium-term expenditure ning and budgeting for PERs, (iv) capacity development to plan. Rwanda is also planning to adopt a similar approach with ensure ongoing sustainability of the process, (v) M&E of the sector budget reviews that would feed into sector working results to feed into ongoing strengthening of the process, (vi) group meetings and in turn would feed into preparation of the a plan for the development of needed data and reference ma- sector Budget Framework Papers (BFPs) that rationalize areas terials, (vii) a process of independent reviews of the reports, of focus in the sector budget and review past performance. and (viii) partnership with donors. FIGURE 16: Government of Rwanda Proposed Budget Calendar Review & Consultation Cabinet / Parliament Discussion PER completed Workshop on -Sector PERs and Consolidated PER -Macro and current year and sector reports consolidated priorities discussed in Cabinet report PER Areas of focus drawn from previous year JBSR/ Sector reviews BFP Parliament discussion, Sector/ joint EDPRS reviews (include Draft reflects PER Review/M&E APR results and sector basic analysis of Final APR reflecting reviews spending or PER Sector Budget Review-SBRs) and strategy updates Inter- BFP/ draft budget 2nd BCC with Ministerial to Parliament Budget Budget Revised MTEF meeting on Draft budget to Final Macro-review ceilings consultation and ceilings budget Parliament 3rd budget to Execution Report defined Macro-framework Revised priorities BCC / final Parliament MTEF Macro-framework BCC updated based on budget to National ceilings; PER final updated Sector SBRs/ sector Cabinet and BFP/budget donor MTEFs reviews Parliament to Cabinet commitments July Aug Sept Oct Nov Dec Jan Feb Mar Apr May Jun Sept. Joint Budget Support Review April Joint Budget Support Review (JBSR) (JBSR) Source: World Bank 2009d. Example 50: DISSEMINATION STRATEGY The dissemination strategy for a PER will depend on the the objectives of the work being undertaken, (ii) determine target audience to be reached. A dissemination strategy who could use this information (e.g., policy makers, donors, used by IFPRI is to (i) determine the target audience against researchers, or other stakeholders), and (iii) determine who PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 74 PA RT V: — EX A M PLES FIGURE 17: Match Publication to Audience Anson, R., and E. Zegarra. 2008. Honduras: Public Expenditure Assessment and Strategy for an Enhanced Agricultural and Publication Forestry Sector. Prepared for the Joint DFID/World Bank Partnership on Agriculture Public Expenditure Review in collabo- Research Report ration with RUTA. The Expert Avila, A. F., and R. E. Evenson. 2004. Total Factor Productivity Growth in Agriculture: The Role of Technological Capital. Chap. Food Policy Report The Insider 31 in R. E. Evenson, P. Pingali, and T. P. Schultz, eds., Handbook of Agricultural Economics. Amsterdam: Elsevier. Balmori, H. H.. 2003. “Gender and Budgets: Overview Report.� Issue Brief The Interested BRIDGE (Development-Gender), Institute of Development Studies. Brighton: University of Sussex. Also available at www.bridge.ids. Press ac.uk. Release The Layperson Barker, R., C. Ringler, Nguyen Minh Tien, and M. Rosegrant, 2002. “VN-4: Macro Policies and Investment Priorities for Audience Irrigated Agriculture in Vietnam,� National Component Paper Source: von Grebmer et al. 2005. for the Project on “lrrigation; Investment, Fiscal Policy, and Water Resource Allocation in Indonesia and Vietnam�, IFPRI Project No. 2635-000, Country Report, Vol. 1. Vietnam: Asian wants to receive the information, determining what type of Development Bank. policy dissemination activities are needed such as press in- Beintema, N., E. Castelo-Magalhaes, H. Elliot, and M. Mwala. 2004. terviews, research reports, and policy briefs (von Grebmer et “Zambia�. Washington, DC: International Food Policy Research al. 2005). Institute. As an example, when IFPRI released its projections of the Benin, S. A. Nin Pratt, S. Fan, C. Breisinger, T. Mogues, J. Thurlow, and X. Diao. 2008. “Measuring the Growth and Poverty global food outlook for 2020, it created different communica- Reduction Impacts of Public Investments in Agriculture and tion products for different audiences. For the experts, a full Rural Areas.� Regional Strategic Analysis and Knowledge report, 206 pages, containing complex, detailed information Support Systems (RESAKSS) Working Paper 7. Washington, was published. This report is intended for an important but DC: International Food Policy Research Institute. (http://www. resakss.org/publications/ DiscussionP7.pdf). very limited scienti�c audience. For an interested, informed audience, IFPRI published a 16-page report, 2020 Global Birner, R., N. Sharma, and N. Palaniswamy. 2006. “The Political Economy of Electricity Supply to Agriculture in Andhra Pradesh Food Outlook: Trends, Alternatives, and Choices. It contained and Punjab.� Washington, DC: International Policy Research the essential �ndings, enhanced with color graphics aimed Institute. at policy makers and journalists. For the popular press and Brzeska, J., and S. Fan. 2009, “Agricultural Public Expenditure Reviews: general public, it published a 2-page press release containing A Synthesis of Six Country Cases and Two Cross-Country Studies,� only the main messages through sound bites or bullet point- Mimeo.com. type information. This approach appeals to the mass media Cammack, T., M. Fowler, and C. D. Phomdouangsy. 2008. “Lao PDR and high-level policy makers who don’t have time to read Public Expenditure Study.� Prepared for the Joint DFID/World longer reports (see Figure 17). Bank Partnership on Agriculture Public Expenditure Review. Carter, R. , and K. Danert. 2006. “Planning for Small-Scale Irrigation Intervention.� FARM-Africa, Working Paper Series 4. London. REFERENCES Devarajan, S., V. Swaroop, and H. Zou. 1996. “The Composition of African Development Bank, Food and Agriculture Organization, Public Expenditure and Economic Growth.� Journal of Monetary International Fund for Agricultural Development, International Economics, vol. 37, 313–44. Water Management Institute, and W`orld Bank. 2007. “Investment in Agricultural Water for Poverty Reduction and Economic Growth Fan, S., B. Yu, and A. Saurkar. 2008. “Public Spending in in Sub-Saharan Africa.� Washington, DC: World Bank. Developing Countries: Trends, Composition, and Changes.� In Public Expenditure, Growth, and Poverty in Developing Akroyd, S., and L. Smith. 2007. “Review of Public Spending to Countries: Issues, Methods, and Findings. Baltimore: Johns Agriculture.� A joint DFID/World Bank study. Hopkins University Press. Anderson, K., ed. 2010. The Political Economy of the Agricultural Fan, S., M. Johnson, A. Saurkar, T. and Makombe. 2008. “Investing Price Distortions. New York: Cambridge University Press. in African Agriculture to Halve Poverty by 2015.� IFPRI Discussion Paper 000751. Washington, DC: International Food Anderson, K., ed. 2009. Distortions to Agricultural Incentives: A Global Policy Research Institute. Perspective, 1955 to 2007. Washington, DC: World Bank. TOOLK IT REFERENCES 75 Fan, S., and J. Brzeska. 2007. “How Can Government Promote Pro- McKean, M., and E. Ostrom. 1995. “Common Property Regimes in Poor Agricultural Growth?� Synthesis of IFPRI Case Studies. the Forest: Just a Relic from the Past?� Unasylva 180. http:// Module 1 of joint DFID/World Bank study. www.fao.org/docrep/ v3960e/v3960e03.htm#TopOfPage. Fan, S., P. Lan Huong, and T. Quang Long. 2004. Government Spending Morris, M., V. Kelly, R. Kopicki, D. Byerlee. 2007. “Fertilizer Use in and Poverty Reduction in Vietnam. Draft Report Prepared for the African Agriculture.� Directions in Development. Washington, World Bank-Funded Project “Pro-Poor Spending in Vietnam,� by DC: The World Bank. International Food Policy Research Institute, Washington, DC OECD. 2004. “Agricultural Policies in OECD Countries: At a and Central Institute for Economic Management, Hanoi. Glance.� Paris: Organization for Economic Cooperation and Fan, S., P. Hazell, and S. Thorat. 2000. “Government Spending, Growth, Development. and Poverty in Rural India.� American Journal of Agricultural OECD. 2010. “Measuring Aid to Agriculture.� http://www.oecd.org/ Economics 82 (4):1038–51. dataoecd/54/38/44116307.pdf. Food and Agriculture Organization (FAO). 2005. Base de Datos de Pletcher, J.. 2000. “The Politics of Liberalizing Zambia’s Maize Estadisticas e Indicadores del Gasto Publico Agricola y Rural Markets.� World Development 28(1):129–42. (GPAGRURAL), O�cina regional para American Latina y el Caribe, Santiago. Pradhan, S. 1996. “Evaluating Public Spending: A Framework for Public Expenditure Reviews.� World Bank Discussion Papers Food Security Group. 2007. “Food Security Activities Implemented by 323. Washington, DC: The World Bank. Humanitarian Agencies (Excluding Food Aid) in Uganda During the First Season of 2007: Mid-Season Data Update.� Kampala: FAO. Psaltopoulos, D., and B. Eudokia. 2005. “Modeling the Effect of Trade Policy Scenarios on Multi-functionality in Greek Agriculture: Golait, R., and S. M. Lokare. 2008. “Capital Adequacy in Indian A Social Accounting Matrix Approach.� September 1. CEPS Agriculture.� Reserve Bank of India Occasional Papers, Vol. 29, ENARPRI Working Papers No. 14. Available online at http://aei. No. 1. pitt.edu/6724/01/1262_14.pdf. Govereh, J., J. J. Shawa, E. Malawo, and T. S. Jayne. 2006. “Raising Scott, J. 2008. “Assessment of ARD Public Expenditures.� the Productivity of Public Investments in Zambia’s Agricultural Background paper for the Mexico Agricultural and Rural Sector.� International Development Collaborative Working Paper Development Public Expenditure Review, Report No. 51902-MX. Series ZM-FSRP-WP-20. Lansing, MI: Michigan State University Washington, DC: The World Bank. Press. Tangermann, S. 2006. “OECD Work on Agricultural Policies in Brazil, International Development Research Center (IDRC). 2008. “Raw China, India and South Africa.� February 27. Presentation at the Tobacco Price Subsidy: Implications for Farm Income and Natural Rural Week of the World Bank. Washington, DC: The World Productive Resources.� Technical Report Number 103454-002. Bank. Ottawa. United Republic of Tanzania (2006). “Local Government Capital International Fund for Agricultural Development (IFAD). 2005. Development Grant System: Manual for the Assessment of “Agricultural Water Development for Poverty Reduction in Councils against Minimum Access Conditions and Performance Eastern and Southern Africa.� Measurement Criteria.� Prime Minister’s Of�ce—Regional International Water Management Institute (IWMI). 2005. “Lessons from Administration and Local Government. Irrigation Investment Experiences: Cost-Reducing and Performance- Valdes, A. 2008a. “Agricultural Public Spending: Description and Enhancing Options for Sub-Saharan Africa.� Pretoria: IWMI. Assessment Relevant to Latin America.� Background paper Kolavalli, S., R. Birner, S. Benin, L. Horowitz, S. Babu, K. Asenso- for the Mexico Agricultural and Rural Development Public Okyere, N. M. Thompson, and J. Poku. 2009. “Public Expenditure Expenditure Review, Report No. 51902-MX. Washington, DC: and Institutional Review.� Ghana’s Ministry of Food and The World Bank. Agriculture. Ghana Strategy Support Program (GSSP) Background Valdes, A. 2008b. “Payments for Environmental Services in Natural Paper 17. Washington, DC: International Food Policy Research Resource Sectors.� Background paper for the Mexico Agricultural Institute. and Rural Development Public Expenditure Review, Report No. López, R. 2004. “Why Governments Should Stop Non-social 51902-MX. Washington, DC: The World Bank. Subsidies: Measuring the Consequences for Rural Latin Van der Meer, K., and M. Noordam. 2004. “The Use of Grants America.� World Bank Policy Research Paper 3609. Washington, to Address Market Failures: A Review of World Bank Rural DC: The World Bank. Development Projects.� Agriculture and Rural Development López, R., and G.I. Galinato. 2007. “Should Governments Stop Discussion Paper 27. Washington, DC: The World Bank. Subsidies to Private Goods? Evidence from Rural Latin America.� von Grebmer, K., K., S. Babu, V. Rhoe, and M. Rubenstein. 2005. Journal of Public Economics, 91(5-6): 1071–94. Communicating Food Policy Research: A Guidebook. Washington, Matthews, A., P. Claquin, and J. Opolot. 2006. “Distortions to DC: International Food Policy Research Institute. Agricultural Incentives in Uganda.� Washington, DC: World Bank World Bank. 1998. “The Impact of Public Expenditure Reviews: Agricultural Distortions Research Project. An Evaluation.� Report No. 18573. Operations Evaluation Department. Washington, DC: The World Bank. PR A C T I T I O N E R S ’ TOOL KIT F OR AGRICULT URE P UBLIC EX PEND ITUR E A NA LYS IS 76 R EFER ENC ES World Bank. 2001. “Turkey Public Expenditure and Institutional World Bank. 2009a. “Mexico Agricultural and Rural Development Review.� Reforming Budgetary Institutions for Effective Public Expenditure Review.� Report No. 51902-MX. Washington, Government, Report No. 22530-TU. Washington, DC: The World DC: The World Bank. Bank. World Bank. 2009b. “Mozambique Agriculture Public Expenditure World Bank. 2003. “India Revitalizing Punjab’s Agriculture.� Report Review.� Washington, DC: The World Bank. No. 37069. Washington, DC: The World Bank. World Bank. 2009c. Gender in Agriculture Sourcebook. The World World Bank. 2005a. “Turkey Policy and Investment Priorities for Bank, Food and Agriculture Organization, and International Fund Agriculture and Rural Development.� ECSSD/ECA. Washington, for Agricultural Development. DC: The World Bank. World Bank. 2009d. “Institutionalizing Public Expenditure Reviews: World Bank 2005b. “Vietnam: Managing Public Expenditure for A Guidance Note with Applications to Rwanda.� Technical Poverty Reduction and Growth.� Report No. 30035-VN, Volume Assistance for Public Expenditure Management. Washington, 2: Sectoral Issues. Washington, DC: The World Bank. DC: The World Bank. World Bank 2005c. “Revitalizing the Agricultural Technology System World Bank, 2009e. “United Republic of Tanzania Public Expenditure in Bangladesh.� Report No. 34543-BD. Washington, DC: The and Financial Accountability Review 2008.� Report No. 48042- World Bank. TZ. Washington, DC: The World Bank. World Bank. 2006a. “Ethiopia Agricultural and Rural Development: World Bank. 2010a. “World Development Indicators.� Washington, Public Expenditure Review for 1997–98 and 2005–06.� DC: The World Bank. Washington, DC: The World Bank. World Bank. 2010b. “Uganda: Agriculture Public Expenditure World Bank. 2006b. “Russian Federation Enhancing the Impact of Review.� Report No. 53702-UG. Washington, DC: The World Public Support to Agriculture and Rural Sectors.� Report No. Bank. 39213. Washington, DC: The World Bank. World Bank. 2010c. “Lebanon Agriculture Sector Note: Aligning World Bank. 2006c. “Mexico Water Public Expenditure Review.� Public Expenditures with Comparative Advantage.� Washington, Report No. 36942-MX. Washington, DC: The World Bank. DC: The World Bank. World Bank. 2006d. “Reengaging in Agricultural Water Management: World Bank. 2010d. “How Do We Improve Public Expenditure Challenges and Options.� Washington, DC: The World Bank. in Agriculture?� A Synthesis Report of the DFID-World Bank Partnership on Agriculture Public Expenditure Review. World Bank. 2006e. “Turkey Public Expenditure Review.� Report Washington, DC: The World Bank. No. 36764-TR. Washington, DC: The World Bank. World Bank. 2010e. “Zambia: Impact Assessment of the Fertilizer World Bank. 2007a. “Uganda: Agriculture Sector Public Expenditure Support Program: Analysis of Effectiveness and Ef�ciency.� Review Phases 1 and 2.� Joint DFID/World Bank Partnership on Washington, DC: The World Bank. Agriculture Public Expenditure Review. London: Oxford Policy Management. World Development Report (WDR). 2008. “Agriculture for Development.� Washington, DC: The World Bank. World Bank. 2007b. “Promoting Gender Equality and Women’s Empowerment.� In Global Monitoring Report 2007: Confronting Zorya, S. 2006. “Improving Agricultural Fiscal Policy in Ukraine.� the Challenges of Gender Equality and Fragile States, Chapter 3. ECSSD Working Paper 36970. Washington, DC: The World Bank. Washington, DC: The World Bank. Zorya, S. and N. Francken. 2010. “Tanzania: Rapid Budget Analysis World Bank. 2008. “Nigeria Agriculture Public Expenditure Review.� for Annual Review 2010/11—Agriculture Sector.� World Bank Report No. 44000-NG. Washington, DC: The World Bank. and Irish Aid Report, Dar es Salaam, Tanzania. TOOLK IT A G R I C U L T U R A L A N D R U R A L D E V E L O P M E N T This is a joint publication with: Agriculture and Rural Development (ARD) 1818 H Street, NW Washington, D.C. 20433 USA Telephone: 202-477-1000 Internet: www.worldbank.org/ard