72896 South Asia Economic Focus A Review of Economic Developments in South Asian Countries Creating Fiscal Space through Revenue Mobilization June 2012 South Asia Region THE WORLD BANK Washington, D.C. South Asia Economic Focus A Review of Economic Developments in South Asian Countries Creating Fiscal Space through Revenue Mobilization June 2012 South Asia Region THE WORLD BANK Washington, D.C. © 2012 International Bank for Reconstruction and Development / International Development Association or The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpreta- tions, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denom- inations, 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 work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attri- bution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@ worldbank.org. Table of Contents Abbreviations and Acroynyms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii Foreword. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi Chapter 1: Introduction and Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Chapter 2: In Search of Reasons for Large Revenue Gaps in South Asia. . . . . . . . . . . . . . . . . . . . . . .7 Factors That Affect Revenue Gaps. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 A Review of Variables That Affect Revenues in South Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Regression Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Chapter 3: Creating Fiscal Space through Raising Tax Revenue. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Context and Overview of Best Practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Tax Composition and Regimes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Performance Assessment of South Asian Tax Regimes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Tax Administration in South Asia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Tax Expenditures and the Tax Base . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Chapter 4: Nontax Revenue: Structure, Status, and Future . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Structure of NTR in South Asian Countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Analysis of NTR in South Asia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 Options to Increase NTR Collection in South Asia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 Chapter 5: Summary and Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Appendix A: Empirical Estimation of Revenue Gap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Appendix B: Tax-Specific Buoyancies, 2001–10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 Appendix C: Country Studies: Fiscal Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Appendix D: Summary of Debt Sustainability Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 Appendix E: Fiscal Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 Appendix F: Trends of Nontax Revenue. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries iii iv South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Boxes 2.1 Afghanistan: Mobilizing Domestic Revenue through Mining Development. . . . . . . . . . . . . . . . . . . 9 2.2 Inflation and Indexing Exemption Levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2.3 The Informal Economy, Small Firm Size, and the Difficulty of Tax Collection. . . . . . . . . . . . . . . . . 14 3.1 Desirable Features of Tax Policy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 3.2 Tax Reforms in India: A Success Story. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 3.3 Glossary of Key Concepts Used in the Analysis of Taxes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 3.4 Bhutan and Maldives: Revenue Composition and Fiscal Management . . . . . . . . . . . . . . . . . . . . . 29 3.5 Sri Lanka: Tax Exemptions Lead to Tax Base Erosion and Complicate Tax Administration . . . . . . . 35 4.1 User Charges: Efficiency and Equity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Figures 1.1 General Government Revenue, 2000–12. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Public Debt–to–GDP Ratio, 2010. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.3 Total Public Expenditure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 1.4 Public Expenditure on Health Care . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.5 Public Expenditure on Education. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.6 MDG Progress in South Asia by Progress Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 2.1 International Comparison of General Government Revenue. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.2 Revenue Gaps over Time. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.3 International Comparison of Economic Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 2.4 Urbanization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 2.5 Consumer Price Index Inflation: South Asia versus Other Economies . . . . . . . . . . . . . . . . . . . . . 11 2.6 Control of Corruption. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.7 Quality of Bureaucracy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 3.1 Revenue Composition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 3.2 Direct Tax Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 3.3 Domestic Indirect Taxes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 3.4 Trade Taxation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 3.5 Income Tax Yields, CIT Rate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 3.6 VAT Yields Based on GDP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 3.7 Income Tax Yields, FY2000/01–FY2008/09. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 3.8 VAT Yields, FY2001/02–FY2009/10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 4.1 Nontax Revenues in South Asia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 4.2 Composition of NTR, Afghanistan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 4.3 Composition of NTR, Bangladesh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 4.4 Composition of NTR, Bhutan. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 4.5 Composition of NTR, India (Consolidated Central and State Governments) . . . . . . . . . . . . . . . . . . 41 4.6 Composition of NTR, Maldives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 4.7 Composition of NTR, Nepal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 4.8 Main Forms of NTR, Nepal. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 4.9 Composition of NTR, Pakistan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 4.10 Composition of NTR, Sri Lanka . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 4.11 Ratio of User Charges Collected Relative to Government Operation and Maintenance Expenditure in South Asia, 2005–10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 C.1.1 External and Core Budget in Afghanistan, FY2010/11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 C.1.2 Projected Expenditure and Core Budget Allocations in Afghanistan, FY2021/22 . . . . . . . . . . . . . 65 C.2.1 Government Revenue and Expenditure, Bangladesh. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 C.2.2 Financing of Deficit, Bangladesh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 C.3.1 Overall Fiscal Developments, Bhutan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 C.3.2 Revenue Structure, Bhutan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 C.3.3 Expenditure Structure, Bhutan. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 C.4.1 General Government Revenue, India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 Table of Contents v C.4.2 General Government Expenditure, India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 C.4.3 Deficit, India. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 C.5.1 Budgetary Outturns, Maldives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 C.5.2 Revenue Composition, Maldives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 C.5.3 Remunerations, Maldives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 C.6.1 Revenue Composition, Nepal. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 C.6.2 Expenditure, Nepal. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 C.6.3 Capital Formation and Transfers, Nepal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 C.7.1 Creation of Additional Fiscal Space, Pakistan. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 C.7.2 Use of Additional Fiscal Space, Pakistan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 C.7.3 Additional Fiscal Space Needed in Pakistan. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 C.8.1 Financing of Deficit, Sri Lanka . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 C.8.2 Composition of Expenditure, Sri Lanka . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 C.8.3 Revenue Composition, Sri Lanka . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 E.1 Fiscal Trends, Afghanistan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 E.2 Long-term Fiscal Trends, Bangladesh. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 E.3 Fiscal Trends, Bhutan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 E.4 Long-term Fiscal Trends, India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 E.5 Long-term Fiscal Trends, Maldives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 E.6 Fiscal Trends, Nepal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 E.7 Fiscal Trends, Pakistan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 E.8 Long-term Fiscal Trends, Sri Lanka. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 F.1 Trends of Nontax Revenue, Afghanistan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 F.2 Trends of Nontax Revenue, Bangladesh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 F.3 Trends of Nontax Revenue, Bhutan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 F.4 Trends of Nontax Revenue, India. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 F.5 Trends of Nontax Revenue, Maldives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 F.6 Trends of Nontax Revenue, Nepal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 F.7 Trends of Nontax Revenue, Pakistan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 F.8 Trends of Nontax Revenue, Sri Lanka . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 Tables 1.1 How Much Can South Asian Countries Borrow: Baseline Scenarios of IMF−World Bank Debt Sustainability Analyses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.2 Public Expenditure in Social Sectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 2.1 Regression Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 3.1 Personal and Corporate Income Tax Structures of South Asian Countries . . . . . . . . . . . . . . . . . . . 23 3.2 Indirect Taxes of South Asian Countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 3.3 Trade Tax Structure in South Asian Countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 3.4 Comparison of Income Tax Rates across Regions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 3.5 Revenue Buoyancies in South Asia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 3.6 Time Spent on Tax Preparation and Filing, 2008–12 Average . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 3.7 Tax Administration Indicators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 4.1 NTR and Tax Revenue Stabilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 4.2 Cyclicality of NTR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 A.1 Descriptive Statistics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 A.2 Coefficients of Correlation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 A.3 Main Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 A.4 Alternative Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 A.5 Instrumental Variable First-Stage Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 A.6 Instrumental Variable Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 A Note on Fiscal Years and Financial Reporting by South Asian Countries South Asian countries’ national income and product account data refer to fiscal years that cut across calendar years, except for Maldives and Sri Lanka which report according to calendar year. The fiscal years for the respective countries are as follows: • Afghanistan: April 1–March 31 • Bangladesh and Pakistan: July 1–June 30 • Bhutan: July 1–June 30 • India: April 1–March 31 • Maldives and Sri Lanka: January 1–December 31 • Nepal: mid-July–mid-July Reporting practices vary. Bangladesh, Nepal, and Pakistan report their fiscal data in the second year of the split, while India and Afghanistan report at the beginning of the fiscal year split. Abbreviations and Acroynyms ADP Annual Development Program (Bangladesh) BOI Board of Investment (Sri Lanka) BTT business turnover tax CIT corporate income tax CPC Central Processing Center (India) DSA debt sustainability analysis FBR Federal Board of Revenue (Pakistan) FY fiscal year GDP gross domestic product GST goods and services tax IDA International Development Association IMF International Monetary Fund IRD Inland Revenue Department (Nepal) IT information technology IV instrumental variable KTF Kelkar Task Force LIC low-income country MDG Millennium Development Goal MTFF medium-term fiscal framework NBR National Board of Revenue (Bangladesh) NTR nontax revenue OECD Organisation for Economic Co-operation and Development O&M operations and maintenance PIT personal income tax RBI Reserve Bank of India SBP State Bank of Pakistan SOE state-owned enterprise TARP Tax Administration Reform Project THCP Thala Hydro Power Corporation (Bhutan) TRC Tax Reform Committee (India) VAT value added tax South Asia Economic Focus – A Review of Economic Developments in South Asian Countries vii Foreword T his is the second issue of the be close to or approaching the system that is consid- South Asia Economic Focus, a biannual series ered desirable by many researchers and often recom- prepared under the aegis of the office of the chief mended by the World Bank and International Monetary economist for the region, Kalpana Kochhar. This series Fund. Yet, tax productivity—the actual yields of key reviews contemporary developments and policy issues tax instruments—remains low. The conclusions from in South Asia with each volume examining a single these observations are that, in many countries in South subject in depth. The analysis seeks to identify key Asia, tax bases are eroded by exemptions and other tax factors underlying the issue being studied, to outline preferences and tax administration needs to be strength- the policy challenges it poses to different countries, ened. Regarding nontax revenue, it is important that and to offer practical solutions where possible. The user charges for publicly provided goods and services publication is intended for audiences within the World should reflect the cost of their provision. Bank, its partners, and client countries with a view to This volume was prepared by Hisanobu Shishido, generating enriching and fruitful discussions for policy. Kirthisri Rajatha Wijeweera, Bingjie Hu, and Congyan This current volume discusses how fiscal space can Tan, with inputs on tax administration from Daniel be created by mobilizing more revenue in South Asia. Alvarez and Sebastian S. James. Shreehari Dhungana Revenue of many South Asian countries is relatively and Nathan Pamart provided excellent support low, averaging about 15 percent of gross domestic collecting and analyzing data. Prajwal Shahi in the product, while the infrastructure and social develop- World Bank Country Office–Nepal and Daminda ment needs are huge. Increasing revenues would create Fonseka in the World Bank Country Office–Sri Lanka fiscal space for more public investment in physical provided excellent research assistance. Peer reviewers and human capital and lead to faster poverty reduc- were Milan Brambhatt and Daniel Alvarez. David tion and more rapid progress toward the Millennium Gould, Ulrich Bartsch, Roshan Bajryacharya, Debra Development Goals (MDGs). This volume of the South Adams, and Zmarak Shalizi also provided very helpful Asia Economic Focus series examines revenue patterns comments. All World Bank country economists of of South Asian countries, assesses their tax regimes, and South Asian countries provided extensive inputs compares their tax productivity with countries of other for the volume. The authors are responsible for any regions. Tax regimes in South Asia are generally seen to remaining errors. South Asia Economic Focus – A Review of Economic Developments in South Asian Countries ix Executive Summary M ost South Asian countries have However, this book focuses exclusively on revenue low rates of public revenue collection, with mobilization. total revenues averaging about 10–15 percent Income levels and different economic structures of gross domestic product (GDP) compared with an affect revenue levels. The main factors affecting reve- average of 20 percent of GDP in comparable developing nues are identified through quantitative analyses that counties or emerging economies and with even higher relate the revenue-to-GDP ratio to various structural rates in Organisation for Economic Co-operation and other factors using data that cover 168 countries and Development countries. Another feature of the from 1990 to 2008. Findings include the following: (a) fiscal landscape in South Asia is the high deficit-to- higher levels of income per capita are associated with GDP ratio and the resulting high debt-to-GDP ratio. higher revenue-to-GDP ratios; (b) larger GDP shares Debt sustainability analyses show that most countries of agriculture lower the revenue-to-GDP ratio because in South Asia need to reduce debt-to-GDP ratios by of low agricultural income per capita and associated reducing the primary deficit, although absolute debt difficulties in tax collection; (c) larger service shares are levels continue to rise. In addition, the countries also correlated with lower revenue primarily because cannot depend on higher concessionary foreign assis- of the dominance of small and informal firms in the tance because the fiscal situation of donor countries service sector; (d) higher international trade tends to has become increasingly tight. raise revenues, although only marginally because trade The combination of low revenue, limited liberalization often takes place as trade volume rises; borrowing capacity, and stresses on donor aid results (e) countries with significant income from natural in inadequate room for spending. Indeed, most South resources enjoy higher revenues; (f ) urbanization tends Asian governments have been spending significantly to lower revenue collection because it tends to be asso- less than emerging economies or developing coun- ciated with a greater share of informal activities; and tries in other regions at comparable levels of per capita (g) increased transparency and accountability and lower income—including for infrastructure and social sector inflation tend to boost revenue collection. development. This trend is especially stark when seen Revenues of most South Asian countries are against the major gaps in physical infrastructure and significantly below the predicted values even after weak social sector indicators. In addition, South Asian these structural factors are taken into account. South countries are consistent underperformers in achieving Asian economies have a number of features that tend the Millenium Development Goals (MDGs): the region to push down revenue-to-GDP ratios. For example, remains the poorest-performing subregion of Asia and on average, they have lower per capita income, higher one of the poorest-performing regions globally (KÖhler agricultural shares of GDP, and lower international and Stirbu 2008). trade than most regions of the world. GDP shares of South Asian countries’ expenditure needs are the service sector are also higher than those of East Asia enormous. Their high fiscal deficits and outstanding and the Middle East. Although the South Asian coun- government debt provide a strong rationale to create tries have lower urbanization rates, inflation is higher more fiscal space through revenue mobilization.1 Of than in other countries. Efforts by South Asian coun- course, fiscal space can also be created by reducing low tries to lower corruption appear to be about average or priority spending; enhancing the government’s imple- slightly higher than those of their peers. After control- mentation capacity, so that capital investment and social ling for all these factors, the book finds that revenue service provision can be carried out at reduced cost; mobilization of South Asian countries (except for reforming subsidies or transfer programs to make them Bhutan and Maldives) is still below the average for more targeted and efficient; and rationalizing admin- countries at similar income levels by about 4 percent istered prices of publicly provided goods and services. of GDP for India and about 10 percentage points of South Asia Economic Focus – A Review of Economic Developments in South Asian Countries xi xii South Asia Economic Focus – A Review of Economic Developments in South Asian Countries GDP for Bangladesh. Furthermore, the gaps have Asian countries other than Sri Lanka.2 Many remained broadly unchanged for the past 10–15 years. countries have buoyancy well over 1, which Features of tax policy and performance in South means revenue is growing faster than GDP. Asia are assessed to see how these gaps can be Only Sri Lanka has a low figure of 0.6. India narrowed. The best practice tax policy, often included and Pakistan have buoyancy at or close to 1. in World Bank and International Monetary Fund Afghanistan and Bangladesh have high buoy- (IMF) recommendations, is examined, as is the prox- ancy mainly because they began from a very imity of tax regimes of South Asian countries to these low base, and Nepal’s high buoyancy is linked “desirable� tax features. The structure of taxes in South to the recent high growth in remittance, which Asia is benchmarked against other comparable coun- has fueled imports and import-related revenues. tries outside the region. The features of tax collection However, Nepal’s recent extremely high growth in South Asian countries—tax composition, buoyancy, in remittance—reaching 40–50 percent a year— and yields—are then assessed to find the underlying is unlikely to be sustained. causes of low revenue collection. The main findings • Tax yields are very low in South Asia—the crux are as follows: of the problem. Yields of both VAT and income taxes—actual collection as a share of potential • South Asia’s actual tax collection has a pattern collection (the base multiplied by representative closer to that seen typically in lowest-income rates)—are very low in South Asia. Sri Lanka and countries. Income tax collection, especially Bangladesh have the lowest income tax yields, personal income tax collection, is low, reflecting and Pakistan and Bangladesh have the lowest the countries’ low income per capita. This factor VAT yields. Overall, yields of all major South tends to limit the overall tax collection. Trade Asian countries are below the average of low- taxes are still relatively high despite the lowered income countries and below the yields of most import tariff rates. In many other regions, trade East Asian economies. This suggests that tax bases tax collection has declined as a result of trade are narrow mainly because of tax exemptions, liberalization. Low income tax collection and tax preferences, and weak tax administration. resilient trade tariff collection are typically seen • Tax bases in South Asia are narrow because, in the lowest-income countries in part, the sectors that are difficult to tax are • The tax regimes of many countries have been large in the region. Agriculture and informal reformed to conform more closely to the service sectors are large in South Asia relative best practices framework recommended by to other regions. These areas are difficult to the World Bank and IMF. Specifically, many tax, and their large size and substantial infor- countries have adopted or are in the process mality result in difficulty in raising the overall of adopting a value added tax (VAT), with a tax productivity. limited number of nonzero rates. Income tax • The incidence of tax exemptions and holidays rates are not prohibitive and are relatively simple is high in South Asia. Tax exemptions and holi- (except for Pakistan and Sri Lanka). Top corpo- days and other preferences have tended to be rate income tax (CIT) and personal income tax introduced as a response to politically strong (PIT) rates are generally close to each other, pressure groups rather than as a means to address thereby limiting prospects for shifting income proven market failures. Not only do they erode to lower tax brackets. Import tariff rates are low the tax base, but they also raise concerns of effec- on average, and export taxation is limited. In tiveness, leakage, governance, and spillovers— addition, many countries are considering impor- entrenching corruption in tax administration. tant improvements (for example, the goods and • Weak tax administration is another key barrier services tax [GST] in India, comprehensive tax to effective and fair tax collection. Most South reform in Sri Lanka, income tax and VAT reforms Asian countries have not updated their tax laws in Bangladesh, and a possible VAT in Bhutan). or adopted institutional arrangements to incor- • Average buoyancy, or responsiveness of porate modern principles that can accommo- revenue growth to GDP movements, over the date growing complexities of the economy. Tax past decade is at or above 1 for all the South administration is weakened by excessive use of Executive Summary xiii discretion, and the recent episodes of large-scale • Improving administration is essential and can tax evasion argue for the need of better audits happen not only through institutional reform and improved enforcement. but also through simplification of rates, making • Nontax revenues are also an important part of compliance easy and noncompliance costly. South Asian countries’ revenue mobilization, • In Bhutan and Maldives, where revenue depends although they remain unexplored. The countries’ on a single sector (hydropower for Bhutan, dependence on nontax revenues ranges from and tourism for Maldives), it is important that about 15 percent of total revenues in India, Sri revenue sources are diversified. Lanka, and Bangladesh to a high of more than 50 • User fees for government services can be a stable percent in Bhutan and Maldives. Both revenue form of nontax revenue. Efficient user charges fluctuations and the pro-cyclicality of revenues should be set equal to the cost of providing decline when nontax revenues are added to tax additional units of goods and services—such as revenues in most South Asian countries, indi- power, fertilizers, and fuel products. cating that nontax revenues are a stabilizing factor. However, the ratio of user fee collection as a share of the governments’ operation and Notes maintenance spending shows that South Asian 1. “Fiscal space� is defined as the availability of governments are still far from charging appro- budgetary room that allows governments to priate user charges. provide resources for a desired purpose without prejudice to the sustainability of its financial posi- The main policy messages of the volume are as tion Heller (2005). follows: 2. Average buoyancy is derived through regression of log of tax revenue over log of GDP. • With significant requirements for fiscal space, South Asian countries need to seriously consider further measures to strengthen governance, lower inflation, and improve macroeconomic stability, References all of which are, ceteris paribus, conducive to Heller, Peter S. 2005. “Understanding Fiscal Space.� collecting more revenues. Policy Discussion Paper 05/4, International • Existing efforts to improve tax regimes need to Monetary Fund, Washington, DC. be sustained and deepened. Köhler, Gabriele, and Mariana Stirbu. 2008. “Reaching • Considerable attention is necessary for broad- the MDGs in South Asia: An Inventory of Public ening the tax base by reviewing tax exemptions Policies to Overcome Social Exclusion.� UNICEF and preferences, which have tended to erode the Working Paper, United Nations Children’s Fund, tax base over time. Regional Office for South Asia, Kathmandu. I Introduction and Background M ost South Asian countries have Another feature of the fiscal landscape in South low rates of public revenue collection, with total Asia is the high deficit-to-GDP ratio and the resul- revenues averaging about 18 percent of gross tant high debt-to-GDP ratio (figure 1.2). Given these domestic product (GDP)1 compared with an average high levels of outstanding debt, the International of 20 percent of GDP in comparable developing coun- Monetary Fund (IMF)–World Bank debt sustain- tries or emerging economies, and with even higher ability analyses (DSAs) show that countries in South rates in Organisation for Economic Co-operation Asia cannot fund spending through significant addi- and Development (OECD) countries. Over the past tional borrowing because of the resulting increase in three years, the revenue-to-GDP ratio has averaged their debt-to-GDP ratios (table 1.1, see also appendix 8.5 percent in Afghanistan; 11 percent in Bangladesh; D). Most countries need to reduce their debt-to-GDP 14 percent in Nepal, Pakistan, and Sri Lanka; and ratios to maintain sustainability—although the abso- 19 percent in India. Bhutan has recently had a high lute debt levels continue to rise in all countries. During revenue-to-GDP ratio of nearly 24 percent, and the the next six to seven years for which DSAs calibrate revenue-to-GDP ratio in Maldives—which used to be detailed projections, three countries currently with as high as 34 percent in 2007—declined to about 24 large debts (Sri Lanka, India, and Pakistan) will need percent with the onset of the global financial crisis. to reduce their primary deficits significantly. Bangladesh Also, over the past decade, revenue as a share of GDP and Nepal are expected to be able to run slightly larger has largely been stagnant for Bangladesh, Pakistan, and primary deficits to keep their debt-to-GDP ratios at Sri Lanka (figure 1.1). the relatively low levels (44 percent and 33 percent, Figure 1.1  General Government Revenue, 2000–12 (percentage of GDP) 40 35 30 25 20 15 10 5 0 je 12 at 1 3 4 5 6 7 8 9 1 02 0 ) ) im /1 ed ed /0 /1 /0 /0 /0 /0 /0 /0 /0 ro 1/ / st 10 00 09 02 03 04 05 06 07 08 01 ct (p 201 (e 20 20 20 20 20 20 20 20 20 20 20 Afghanistan Bhutan Maldives Pakistan World Bangladesh India Nepal Sri Lanka Emerging and developing economies Sources: World Bank country team data and staff calculations. Note: GDP = gross domestic product. South Asia Economic Focus – A Review of Economic Developments in South Asian Countries 1 2 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries respectively). Bhutan appears to be the only country support from India and other development partners. that has significant future borrowing possibilities, Still, Bhutan will need to diversify its source of reve- which are due to its hydropower potential. High levels nues. In the case of Afghanistan, as discussed later in of external debt are likely to be sustainable because of this book, its revenue from mining and domestic taxes the rising energy demand from India and committed covers only limited recurrent expenses. The scheduled withdrawal of foreign troops in 2014 is expected to lead to a substantial increase in security spending by the Afghan government—a factor that could compli- Figure 1.2  Public Debt–to–GDP Ratio, 2010 cate debt management. Finally, with outstanding debt (debt-to-GDP ratio, %) of over 85 percent of GDP, the situation in Maldives 100 is already deemed unsustainable; hence, Maldives is excluded from table 1.1. 80 The combination of low revenue, highly limited borrowing capacity, and stresses on donor grants 60 results in inadequate room for spending for devel- opment. Indeed, South Asian countries have been 40 spending less than both developing and emerging- market countries in other regions. Figure 1.3 shows 20 government spending as a share of GDP relative to GDP per capita. Most South Asian countries have 0 lower government spending–to–GDP ratios relative to countries at comparable levels of GDP per capita. An a an an s n ka pa ic sh a al d di ve ta si Ja cif (e s ibb n n) ep ut an st lu an ean de In A a is di a exception is Bhutan, whose hydropower development ni Bh N iL k st Ca ica la th ng P al Pa ha ng di d M Sr u Ea he er So fg r Ba strategy has thus far succeeded in providing sufficient t m A xc ia A tin A fiscal revenue to support higher spending. In Maldives, La as mentioned earlier, spending is unsustainably high Sources: Country central banks; World Economic Outlook Database, International Monetary Fund; World Bank staff calculations. relative to available fiscal resources, and the country Note: GDP = gross domestic product. is already at high risk of debt distress. Table 1.1  How Much Can South Asian Countries Borrow: Baseline Scenarios of IMF−World Bank Debt Sustainability Analyses Debt-to-GDP Debt-to-GDP ratio for first ratio for last First-year Last-year Projected year of DSA year of DSA primary primary real GDP Debt spaceb Domestic External debt Country projectiona (%) projection (%) balance (%) balance (%) growth (%) (%) debt (%) (%) Sri Lanka 86.1 64.2 −3.5 −0.1 6.7 55 31.0 23.7 India 76.7 67.8 −5.0 −0.5 8.2 78 73.9 3.8 Bhutan 67.7 111.0 5.3 0.1 8.4 185 44.8 140.3 Pakistan 61.5 43.6 −1.6 −0.3 5.3 40 39.4 0.4 Bangladesh 42.9 43.6 −1.7 −2.2 6.8 29 22.3 7.1 Nepal 35.9 33.2 −1.2 −1.9 3.8 28 24.1 4.0 Afghanistan 8.1 9.2 0.9 −1.2 7.2 9 0.4 8.4 Source: Authors’ calculations. Note: Baseline scenarios are based on International Monetary Fund (IMF)–World Bank debt sustainability analysis (DSA). Maldives is excluded because its debt is judged unsustainable according to the IMF–World Bank DSA of January 2011. GDP = gross domestic product. a. Initial projection years range from fiscal years 2008/09 to 2010/11, depending on available DSAs. b. Space refers to the increase in nominal debt over the seven projection years under the baseline scenario as a share of first-year GDP. Introduction and Background 3 Figure 1.3  Total Public Expenditure Figure 1.5  Public Expenditure on Education (percentage of GDP) (percentage of GDP) Source: Government Financial Statistics Database, International Monetary Source: World Development Indicators, World Bank, 2008. Fund, 2008; World Development Indicators, World Bank, 2008. Note: GDP = gross domestic product; OECD = Organisation for Economic Co- Note: GDP = gross domestic product; OECD = Organisation for Economic Co- operation and Development. operation and Development. Figure 1.4  Public Expenditure on Health Care Table 1.2 also compares total and social sector (percentage of GDP) spending of some South Asian countries with those of countries that have made significant progress in meeting Millennium Development Goals (MDGs). The table confirms that the countries in the region tend to spend less overall and less, in particular, on social sectors. Although South Asian countries’ total spending as a share of GDP is on average less than 2 percentage points below the average of countries that have made significant MDG progress, their social sector spending is less than half of the better performers (4.7 percent of GDP as compared with the leading achievers’ 10.1 percent).2 Despite their governments’ social policy aspi- rations, South Asian countries are consistent under- performers in terms of achieving MDGs: the region Source: World Development Indicators, World Bank, 2008. remains the poorest performing subregion of Asia and Note: GDP = gross domestic product; OECD = Organisation for Economic Co- one of the poorest performing regions globally (Kohler operation and Development. and Stribu 2008) (figure 1.6). Many countries that are trailing in MDG achievement—that is, Afghanistan, Bangladesh, Bhutan, Nepal, and Pakistan—could benefit from higher public spending, particularly Figures 1.4 and 1.5 show the GDP shares of health spending targeted on human development (health, care and education expenditures relative to GDP per education, and social protection). At the same time, capita. Data availability limits the number of countries all countries in the region have the more substantive included in the comparisons, but South Asian coun- challenge of improving the quality and, in particular, tries are again mostly well below values in other coun- the targeting of public spending, which could lead to tries at comparable income levels. better outcomes than raising spending alone. 4 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Table 1.2  Public Expenditure in Social Sectors Social sector expenditures Total government Total social sector Country or region expenditure Education Health Social protection expenditure South Asia Bangladesh 11.78 2.04 0.94 0.39 3.37 Bhutan 19.31 2.57 1.59 0.00 4.16 India 28.95 5.07 1.58 0.34 7.00 Nepal 15.46 2.61 0.79 0.66 4.06 Pakistan 16.33 1.05 0.46 0.00 1.50 Sri Lanka 20.59 2.64 1.63 3.88 8.16 South Asian 18.74 2.66 1.17 0.88 4.71 average Leading MDG 20.41 4.11 2.33 3.67 10.11 achieversa Source: Government Finance Statistics (database), International Monetary Fund. Note: GDP = gross domestic product; MDG = Millennium Development Goal. a. Countries that achieved 60 percent or more of Millennium Development Goals include Chile, China, Costa Rica, Mexico, and Thailand. Figure 1.6  MDG Progress in South Asia by Progress South Asian countries’ expenditure needs are thus Assessment enormous. Higher public sector spending is needed not (percentage of countries) only for building physical infrastructure, enhancing human capital, and reaching the MDGs, but also more 100 MDV BGN BGN generally for meeting the countries’ growth promotion BTN BGN LKA BTN NPL LKA and poverty reduction goals. MDV BTN LKA These expenditure needs, in the face of high fiscal 50 LKA IND BTN BGD deficits and outstanding government debt, provide a MDV NPL NPL IND IND PAK IND PAK MDV strong rationale to create more fiscal space to accommo- IND LKA NPL date greater expenditures on infrastructure and social BTN AFG AFG PAK NPL MDV PAK AFG services. Fiscal space is defined as room in a budget that 0 allows a government to provide resources for a desired purpose without harming the sustainability of its finan- cial position (Heller 2005). Education Gender Gender Child Births This book focuses on the creation of fiscal space –50 completion (primary) (secondary) mortality attended through revenue mobilization, although such efforts Seriously off track Off track On track are best made using a comprehensive framework that examines all available sources. Those sources include Achieved Insufficient data (a) reducing lower-priority spending, (b) enhancing Source: Data from World Bank South Asia Human Development Team. the capacity to implement priorities so that capital Note: MDG = Millennium Development Goal. investment and social service provision can be carried out at reduced cost, (c) reforming subsidies or transfer programs to make them more targeted and efficient, and (d) rationalizing administered prices of publicly provided goods and services. Introduction and Background 5 The book is organized as follows: Chapter 2 examines Maldives spent about 50 percent of GDP and the factors that could account for the low revenue collec- has achieved relatively more MDGs than other tion in South Asia. Chapter 3 describes the tax systems regional countries, but this level of spending is in South Asia and assesses them against a number of unsustainable, and the country is making strong benchmarks that are commonly used in the literature, efforts toward fiscal consolidation. such as international comparisons, estimates of buoy- ancy, and tax yields, and then discusses key elements of tax administration in South Asia. Chapter 4 pres- References ents information on nontax revenues, and chapter 5 concludes with some key policy implications. Heller, Peter S. 2005. “Understanding Fiscal Space.� IMF Policy Discussion Paper PDP/05/4, International Monetary Fund, Washington, DC. http://www.imf .org/external/pubs/ft/pdp/2005/pdp04.pdf. Notes Kohler, Gabriele, and Mariana Stribu. 2008. “Reaching 1. The average general government revenue as share the MDGs in South Asia: An Inventory of Public of GDP for South Asia is weighted by GDP in Policies to Overcome Social Exclusion.� Working individual countries. Paper, United Nations Children’s Fund, New York. 2. Simple averages are used rather than weighted aver- World Economic Outlook Database. International ages because MDG achievements, for example, Monetary Fund, Washington, DC. http://www are measured by individual countries. The average .imf.org/external/ns/cs.aspx?id=28. picture looks different when Maldives is included. I II In Search of Reasons for Large Revenue Gaps in South Asia F igure 1.1 in chapter 1 in this volume Figure 2.1 shows that the discrepancy between shows clearly that fiscal revenues in South Asia are actual and predicted revenue value for India and below both world and emerging-country averages. To Nepal is about 5 percent of GDP, and for Bangladesh, further assess where the countries in the region stand with Pakistan, and Sri Lanka, it is 10 percent of GDP or respect to revenue collections, we used regression analysis higher. to look at general government revenue as a share of gross Figure 2.2 describes the evolution of these revenue domestic product (GDP) in relation to GDP per capita. gaps from 1998 to 2008. It is worrisome that many Data used were pooled with cross-section and time-series South Asian countries have not narrowed the gap over information covering 168 countries from 1990 to 2008. this period. The performance of Bangladesh and Sri The results are shown in figure 2.1.1 Except in the case of Lanka has steadily worsened, and Pakistan’s perfor- Bhutan (which exceeds the predicted revenue value) and mance has fluctuated but largely stagnated at low Maldives (which is very close to it), all other South Asian levels. India improved slowly during the second half countries are significantly below the predicted revenue of the 2000s only to deteriorate somewhat with the values given their per capita income.2 2008 global financial crisis. Nepal has seen a steady Figure 2.1  International Comparison of Figure 2.2  Revenue Gaps over Time General Government Revenue (percentage of GDP) (percentage of GDP) -4 -6 -8 -10 -12 00 02 04 6 8 8 0 0 9 20 20 20 20 20 19 Year India Nepal Sri Lanka Source: Government Finance Statistics (database), International Monetary Pakistan Bangladesh Fund, 2008. Note: GDP = gross domestic product; OECD = Organisation for Economic Co- Source: Government Finance Statistics (database), International Monetary Fund, 2008. operation and Development. Note: GDP = gross domestic product. South Asia Economic Focus – A Review of Economic Developments in South Asian Countries 7 8 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries improvement recently, but because those gains are based However, this effect may be weakened as coun- on revenue from remittance-fueled imports, they may tries liberalize trade as income rises. not be sustainable. Demography. Aging societies may suffer from a diminishing productive labor force and, thus, a shrinking tax base. However, they also need revenue to take care of the aged population, and revenue efforts Factors That Affect Revenue Gaps may be more intense. This book examines which rela- This chapter further assesses the possible factors that tionship holds true. could be correlated with the observed revenue gaps— Urbanization. Urbanization is likely to be associated that is, the gaps between the tax effort and the tax with higher productivity and ease of revenue collection potential of South Asian countries. Considerable and, thus, a higher revenue-to-GDP ratio. Urbanization literature correlates revenue performance with a large may also increase revenues associated with royalties and number of developmental, structural, and institu- land rent. If, however, the government cannot reform tax tional indicators.3 The data used and the results of policies to adapt to the rapid changes and therefore capture the regressions are summarized in table 2.1 later in the benefits of urbanization, the effect of urbanization this chapter, and details appear in appendix A. The could be ambiguous, especially because urbanization at factors taken into account in this econometric exer- lower income levels could mean a larger informal sector. cise follow. Natural resource wealth. A country’s efforts to collect Level of economic development (proxied by GDP per revenues may be less costly if the country is rich in capita). Low-income countries may have weaker collec- natural resources. This book considers the effect of tion capacity, so revenue collection is expected to rise natural resource wealth on revenue mobilization by with income per capita. But the relationship may not including a dummy variable for countries where the be linear because trade taxation declines significantly at rents from natural resources make up more than 10 higher income levels (Hattari et al. 2008), and, more percent of GDP. Among South Asian economies, important, preferences for smaller governments may Afghanistan is the only country that is rich in natural be more pronounced at high income levels. resources. (See box 2.1 for the country’s revenue mobi- Economic structure. The structure of an economy lization strategy based on the mining sector.) may affect revenue collection. We examine three struc- Public service provision. The willingness to pay taxes tural factors: the size of the agricultural sector, the size may rise with the government’s provision of quality of the service sector, and the economy’s openness to social services. The analysis used GDP shares of lagged international trade, as follows: public spending on the health and education sectors as proxies for provision of social services. Another • Economies with a large agricultural sector tend proxy is total public expenditure as a share of GDP, to have lower revenue. Not only does farm also lagged by a year. income tend to be small and below the income Human capital (literacy rate). Similarly, the literacy tax thresholds, but also many governments rate—as an indicator of human capital development— exempt all or significant parts of agricultural is included as another proxy for effective delivery of income from taxation. public services. • The share of the service sector could have different Institutions and governance. The quality of gover- effects on revenue mobilization, depending on nance matters with regard to the effectiveness of the the economic development level. That share is public service provision and development outcomes. expected to reduce taxation in developing coun- Public spending and investments in environments with tries because the sector can be based largely on weak institutions and governance often fail to trans- informal activities that are hard to tax. However, late into effective public services (Gupta, Davoodi, and more advanced countries typically have larger Tiongson 2002; Mauro 1998; Rajkumar and Swaroop service sectors and possibly more advanced tax 2008; Tanzi and Davoodi 1997). Thus, the anal- administration systems. ysis includes the index on corruption control. It also • Finally, a more open economy with a higher trade includes the index of quality of bureaucracy because share of GDP is expected to enjoy higher reve- better bureaucracy can be assumed to improve both nues because trade taxes are easier to enforce. public service provision and tax enforcement. In Search of Reasons for Large Revenue Gaps in South Asia 9 A Review of Variables That Affect Another feature of South Asia is its large service Revenues in South Asia sector. Given that the service sector tends to be diffi- cult to tax at lower income levels, the large size of This section reviews the structural, demographic, and the service sector is likely to be associated with lower institutional variables in South Asia and compares them revenue ratios. The value added of services as a share with those of the rest of the world. Figure 2.3 illus- of GDP is approximately 54 percent in South Asia, trates the comparison of economic structures. which is higher than other developing countries in The analysis found that South Asian economies tend East Asia (43 percent).4 to rely more on agriculture than do other countries, A third characteristic of South Asian countries is which could explain the lower tax collection compared their relatively low openness to international trade. to similar countries, all else being equal. Figure 2.3 shows that GDP shares of both exports Box 2.1   Afghanistan: Mobilizing Domestic Revenue through Mining Development Afghanistan’s mineral sector consists of copper, gold, Partial equilibrium analysis shows that the devel- and iron ore, among other metals, as well as construction opment of the Aynak and Hajigak mines have the materials, dimensional stone and gemstones, coal, and potential to generate substantial economic returns. hydrocarbons (mostly natural gas). Damage from war, Afghanistan’s combined revenue stream is estimated chronic neglect, and severe underfunding in the past to average US$322 million per year over 2011–15 and two decades led to plummeting production, which fell far more than US$900 million per year on average until short of Afghanistan’s potential. Existing mineral produc- 2031. The development of 11 additional mines has the tion is currently limited to small coal operations, lime- potential to increase fiscal revenues to over US$1 bil- stone, construction materials, and gemstones and dimen- lion per year on average until 2031. To put these figures sional stone from artisanal and small-scale operations. in perspective, estimates show the mines could con- This situation is expected to change dramatically tribute between 2.0 and 3.0 percent of GDP per year over the next years. The potential of the mining sector to domestic revenues (which stood at 11.3 percent of is being touted in language ranging from “becoming a GDP in 2010/11). very important sector for growth in the country in com- The direct employment effect will be relatively ing years� to the sector that will “lead the economic small, but if one assumes a reasonable development growth process in the future. � Very large numbers are path, jobs created in the public and private service in- being discussed with respect to the value of Afghani- dustries through upstream and downstream links could stan’s minerals, often well over a trillion U.S. dollars— leverage job creation. Given the correct interventions, this for a country with a GDP in 2010 of US$15.9 billion. the effect on domestic revenues, growth, and jobs Although there is some element of truth to all of these could be much larger, depending on the degree of the projections, they refer mostly to the value of minerals mines’ integration into the domestic economy and the in the ground, if one assumes that the minerals are size of the market of domestically produced goods. in large enough concentrations to be profitably mined To prevent the mining sector from becoming an en- and also that the necessary infrastructure either will clave with little spillovers to the rest of the economy, be available or can profitably be built by the operations. the government needs to form partnerships with the In the very long run, those assumptions may well be private sector and reinvest mining revenues strategi- true as mining techniques improve, resources become cally to ensure that communities around the mines scarcer, and the Afghan economy develops. However, and along the major transportation routes benefit from in the very short run, only two of the deposits can be the mining development (for example, through exten- profitably mined with a reasonable degree of certainty: sion of infrastructure). Strengthening public financial the copper deposit at Aynak, just 35.0 kilometers management and transparency structures related to (21.7 miles) south of Kabul, and the iron ore deposit at revenues and expenditures in the mining sector is Hajigak, southwest of Kabul. necessary to avoid the heightened risks of corruption. 10 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Figure 2.3  International Comparison of Economic For other economic variables, figure 2.4 shows that Structures South Asian economies have lower urbanization rates (percentage of GDP) than average, suggesting potentially higher costs of tax 80 administration. As shown earlier (figures 1.3–1.5 in chapter 1 in this volume), most South Asian econo- 70 mies have relatively low levels of total public spending 60 as well as health and education spending as shares of 50 GDP compared with other economies. This difference 40 could be related to fewer incentives for tax compliance. Certain social outcomes, such as literacy rates, can illus- 30 trate this point. The average literacy rate in South Asian 20 countries is about 64 percent, whereas the average for 10 all the nonmember countries of the Organisation for 0 Economic Co-operation and Development (OECD) is rs 82 percent. The weak education outcome may reflect a ci a ª n th in A l A nd ld an M an nd A ah ª si Ce uro ficª Pa si Su Afr t O rica ara b- ica be s ª or A e m sia d tA ra a th a be a a or E em W d idd ª th La nt pe insufficient social service provision. Disappointing an s N le rib ric Ea u S m So Ca e f D E social policy outcomes may further weaken the incen- EC t tive to pay taxes if the public is pessimistic about the Service Agriculture Imports Export use of the tax money.5 Source: World Development Indicators (database), World Bank, 2008. One other potential determinant is inflation, Note: GDP = gross domestic product; OECD = Organisation for Economic Co-operation and Development. a. Developing regions. because it is an alternative to regular revenue collec- tion. Previous studies suggest that high and persistent inflation may negatively affect revenue mobilization by reducing the willingness to pay taxes. Consumer Figure 2.4  Urbanization price inflation in South Asia tends to outpace that in (percentage of GDP) other developing countries and emerging economies (figure 2.5). With respect to governance-related variables, according to the International Country Risk Guide ratings (see figure 2.6), the quality of governance in terms of controlling corruption is above average in South Asia, among both OECD and developing coun- tries. Also, the variable for the bureaucratic quality is relatively good for South Asian countries compared with others (see figure 2.7). The question remains, then, why governments in South Asia are faced with low levels of revenue mobilization.6 Source: World Development Indicators (database), World Bank, 2008. Note: GDP = gross domestic product; OECD = Organisation for Economic Co- Regression Results operation and Development. The dependent variable in the regression analysis is the general government revenue as a share of GDP given the set of country characteristics described above. This and imports are lower in South Asia than in many method can derive revenue gaps after controlling for other developing countries, especially in East Asia and differences in economic structures. The analysis includes Eastern Europe. Because trade taxes are relatively easy some variables under policy makers’ control, such as to enforce, the less open a given economy is to inter- efforts to contain inflation or to fight corruption. If national trade, the more costly it would be for the significant, those efforts should affect fiscal space rela- government to raise taxes. tively quickly. Interpretation includes the size of the In Search of Reasons for Large Revenue Gaps in South Asia 11 Figure 2.5  Consumer Price Index Inflation: South Asia versus Other Economies (%, year-on-year) 20 15 10 5 0 ry ay r ry ay r ry ay r ry ay r ry ay be be be be -5 ua ua ua ua ua M M M M M m m m m n n n n n 07 08 09 10 11 te te te te Ja Ja Ja Ja Ja 20 20 p p p p 20 20 20 Se Se Se Se 07 08 09 10 11 20 20 20 20 20 07 08 09 10 20 20 20 20 Brazil China Emerging markets (average) Indonesia Korea, Rep. South Asia (average) Turkey Source: Global Economic Monitor (database), World Bank, 2011. Figure 2.6  Control of Corruption Figure 2.7  Quality of Bureaucracy (percentage of GDP) (percentage of GDP) Source: International Country Risk Guide (data set), The PRS Group, 2008. Source: International Country Risk Guide (data set), The PRS Group, 2008. Note: GDP = gross domestic product; OECD = Organisation for Economic Co- Note: GDP = gross domestic product; OECD = Organisation for Economic Co- operation and Development. operation and Development. 12 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Table 2.1  Regression Results estimated coefficient of each South Asian country General General dummy as the country’s revenue gap.7 Revenue/GDP Revenue/GDP The main regression results appear in table 2.1; other Variables (1) (2) regression results are given in appendix A. Column (2) Log of GDP per capita 1.254*** 1.039** of table 2.1 includes institutional and governance (0.315) (0.440) indicators such as control of corruption and quality Agriculture share (% of GDP) –0.192*** –0.245*** of bureaucracy. Some country dummies are dropped because they are perfectly correlated with the institu- (0.0351) (0.0347) tional indicators. Services share (% of GDP) –0.115*** –0.222*** The first finding is that higher income level has (0.0292) (0.0389) a significant effect on revenue collection as a share Trade share (% of GDP) 0.0168*** 0.0159*** of GDP. In particular, the results show that a 1-unit (0.00494) (0.00563) increase in logarithm of real GDP per capita (about 170 Ratio between aged and working population (%) 0.858*** 0.839*** percent increase) would lead to an increase in general (0.0421) (0.0478) government revenue of 1 to 2 percentage points of GDP. Log of inflation –0.497** –0.642*** As expected, the economic structure is an important determinant of revenue mobilization. Indeed, a heavy (0.199) (0.224) dependence on the agriculture and service sectors has a Urbanization –0.0308** –0.0151 significant and negative effect on revenue levels. A more (0.0135) (0.0148) open economy tends to enjoy higher revenue collec- Country with resource rents ≥ 10% of GDP 4.757*** 5.920*** tion, but the size of the estimated coefficient is quite (0.681) (0.801) small. These results corroborate the expectation that India –4.206*** –2.435*** higher shares of agriculture and service sectors suggest (0.507) (0.801) a relatively larger informal or hard-to-tax economy and Pakistan -6.696*** –4.002*** are associated with lower revenue. In contrast, trade taxes are easier to collect, so higher trade shares are (0.465) (0.600) linked to higher revenue—but the coefficient is small Bangladesh –10.01*** –5.919*** because trade is increasingly liberalized as trade volume (0.608) (0.681) increases in many countries. Nepal –6.511*** An important finding is that the quality of gover- (0.589) nance matters for revenue mobilization. The anal- Bhutan 16.31*** ysis found that for each unit of improvement in the (1.577) corruption control indicator (which ranges from 0 to 6), the government revenue ratio could increase Sri Lanka –8.076*** –6.970*** 1.8 percentage points. This effect is significant at the (0.937) (0.929) level of 1 percent, which suggests that a possible way Maldives 3.005* to improve fiscal space is to create more transparent (1.667) and accountable institutions.8 Control of corruption 1.756*** Other empirical results also suggest some areas (0.207) for potential improvement in government revenue Quality of bureaucracy –0.212 mobilization. For instance, for each percentage point (0.374) of increase in consumer price inflation, there will be a decrease in the revenue ratio of approximately 0.5 Constant 19.91*** 22.12*** percentage point. This effect is statistically significant (3.378) (3.355) at the level of 1 percent. The analysis suggests that in Number of observations 2,137 1,510 countries with high inflation, revenue collection tends R-squared 0.572 0.693 to be low, which may be related to citizens’ willingness Source: Authors’ calculations. to pay taxes. However, inflation could also be posi- Note: Robust standard errors are in parentheses. GDP = gross domestic product. tively associated with revenue collection if nominal *** p < 0.01, ** p < 0.05, * p < 0.1. In Search of Reasons for Large Revenue Gaps in South Asia 13 Box 2.2   Inflation and Indexing Exemption Levels Inflation could be positively associated with revenue because the exemption level was often adjusted with collection if nominal exemption levels remain constant. inflation. Chinese income tax revenues as a share of Piketty and Qian (2009) evaluate income tax reforms in GDP increased from less than 0.1 percent in 1986 to China and India. The combination of fast income growth about 1.5 percent in 2005 and 2.5 percent in 2008, and an underindexed tax schedule in China suggests while in India, the constant adaptation of exemption that the fraction of the Chinese population subject to levels and income brackets have caused revenues to income tax has increased from less than 0.1 percent in stagnate at about 0.5 percent of GDP . Also, the effective 1986 to about 20.0 percent in 2008, whereas in India, tax rates have risen more in China because the brackets the number has stagnated at about 2.0–3.0 percent stayed the same in nominal terms. exemption levels and income brackets remain constant is significant at 10 percent. Although these are very or are adjusted slowly (see box 2.2). rough measures, they do show the average disparity The estimated effects of the age dependency ratio of South Asian countries relative to their peers, taking on government revenue seem to be dominated by the into account differences in the economic structure, effects of richer countries. Higher dependency ratios demography, institutions, and governance levels. For are seen in more developed and aging societies where discussion purposes, the analysis assumes that these revenues are already high (this variable’s correlation are the amounts of fiscal revenue the countries need coefficient with GDP per capita is above 0.7). A 1.0 to mobilize if they are to catch up with the average percentage point increase in the ratio between the aged of countries that are similar in terms of economic and working populations is associated with a 0.5–0.8 development levels and economic and other struc- percentage point increase in government revenue as a tures. In addition, it is often argued that the size of share of GDP. the informal sector, believed to be large in South Asia, In an alternative model, the analysis controlled for contributes to the lower revenue. However, the nature a proxy for the government provision of public social of the informal sector makes it difficult to measure in services, as it should relate to the willingness to pay each country, and thus it cannot be included in the taxes. Including such proxies turned out to be diffi- regression analysis. Schneider and Buehn (2009) and cult (see details in appendix A). The results show that Schneider, Buehn, and Montenegro (2010) estimated the lagged total government spending has a significant the size of the shadow economy, defined as activities coefficient, but it is also highly correlated with the that avoid tax and other official obligations (see box error terms—indicating an endogeneity problem and 2.3). The size of the shadow economy is not used in possible overestimation of the coefficient. Appendix A the regression analysis also because its estimates are shows an attempt to address this endogeneity problem based on the tax system, thus creating an endoge- using an instrumental variable. Another specification neity problem. used lagged public expenditure on education and health care as proxies for provision of public services. However, the inclusion of these variables created multi- collinearity problems. Notes The regressions included country dummies for 1. Care is needed in interpreting a simple correla- Bangladesh, Bhutan, India, Maldives, Nepal, Pakistan, tion of two variables. This issue is addressed in and Sri Lanka.9 Estimated coefficients of these vari- the next section using multivariate regressions. ables are shown in table 2.1) and all are significant 2. Afghanistan was not included in the sample at 1 percent except for Maldives’ coefficient, which because of inadequate data. 14 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Box 2.3   The Informal Economy, Small Firm Size, and the Difficulty of Tax Collection The informal sector in South Asia is large and ac- Another feature of South Asian economies is the counts for over 80 percent of nonagricultural jobs. high share of GDP coming from the service sector The Planning Commission of the government of India and the fact that the sector is dominated by very small estimates that well over 90 percent of the Indian labor businesses. The service sector accounts for 54 per- force is employed in the informal sector. These num- cent of GDP in South Asia and 55 percent of GDP in bers are indeed high compared to the estimate that India. Moreover, most jobs in the Indian service sector about 50 percent of all jobs in developing countries are in small firms that employ only a few workers (over 900 million workers) are informal. (box figure 2.3.2). In fact, 85 percent of service firms Schneider and Buehn (2009) and Schneider, Buehn, in India have fewer than 10 employees. Even with a and Montenegro (2010) estimate that South Asia’s widening of the service tax base, such small firms shadow economy is, on average, 33 percent of GDP , operating in the informal sector would make collecting not much higher than East Asia’s 32 percent. How- these taxes very challenging. A more promising way to ever, what makes South Asia distinct is the narrow bring the informal sector into the tax net is to reduce dispersion of the size of the shadow economy (box value added tax (VAT) exemptions and strengthen VAT figure 2.3.1). Countries are tightly packed around the administration, because VAT on goods—in particular, average—that is, every country in South Asia has a imports—tends to work as an effective tax on the relatively high share of shadow economy, compared informal sector that does not claim tax credit. with East Asia, which has high dispersion. Box Figure 2.3.2  Domination of Employment by Small Firms in India’s Service Sector, 2001 Box Figure 2.3.1  Shadow Economy and 2006 (percentage of GDP) (percent) 80 60 70 50 60 50 40 40 30 30 20 20 10 10 0 4 10 9 20 9 50 9 10 99 99 0+ 4 10 9 20 9 50 9 10 99 99 0+ –1 –1 1– 5– –4 1– 5– –4 – 0– – 0– 20 20 0 ia Bh ia M tan la s Pa esh an iL l Ea nka ia Vi ina In nam M sia ilip sia Th nes nd Number of workers Sr pa ng ve As d As st Ph ay ne la Ch In Ne u Ba aldi d pi a ki ai et h st 2001 2006 al do ut So Own-account enterprises Establishments Source: Schneider and Buehn (2009) and Schneider, Buehn, and Montenegro Source: World Bank 2011. (2010). Note: GDP = gross domestic product. In Search of Reasons for Large Revenue Gaps in South Asia 15 3. Recent examples include Davoodi and of Armenia’s Stubbornly Low Tax Collection.� Gri­ g orian (2007); Le, Moreno-Dodson, and IMF Working Paper WP/07/106, International Rojchaichaninthorn (2008); Bird, Martinez- Monetary Fund, Washington, DC. Vazquez, and Torgler (2004); and Pessino and Global Economic Monitor (database). World Bank, Fenochietto (2010). Washington, DC. http://econ.worldbank.org/ 4. By contrast, in advanced economies, though WBSITE/EXTERNAL/EXTDEC/EXTDECPR the service sector share of GDP is even higher OSPECTS/0,,contentMDK:22855732~menuPK than in South Asia, tax ratios are significantly :6080253~pagePK:64165401~piPK:64165026~t higher, presumably because of more effective tax heSitePK:476883,00.html. enforcement. Government Finance Statistics (database). Inter­ 5. The analysis used lagged literacy rates to miti- national Monetary Fund, Washington, DC. http:// gate the issue of reverse causality between revenue elibrary-data.imf.org/FindDataReports.aspx?d= collection and education outcomes 33061&e=170809. 6. See appendix A for more details on the construc- Gupta, Sanjeev, Hamid Davoodi, and Erwin Tiongson. tion of these variables. 2002. “Corruption and the Provision of Health 7. Caution in interpretation is needed, because the Care and Education Services.� In Governance, analysis is a comparison with the average perfor- Corruption, and Economic Performance, ed. George mance of a large sample of countries with largely Abed and Sanjeev Gupta, 245–79. Washington, different economic and institutional backgrounds DC: International Monetary Fund. and with diverse demand for higher public expen- Hattari, Rabin, Kalpana Kochhar, Helene Poirson, and ditures. Hence, the estimates of revenue gaps here Peita Topalova. 2008. “Taxes in India—Too Much, should be interpreted as an indication of addi- Too Little, or Just Right?� Mimeo, International tional revenue space. Simple interpretation of the Monetary Fund, Washington, DC. gaps also means that unobserved constraints are IMF (International Monetary Fund). 2011. “Resource ignored (IMF 2011). Mobilization in Developing Countries.� IMF Board 8. However, the estimated effect of bureaucratic Paper, IMF, Washington, DC. quality is significant and negative, which is coun- International Country Risk Guide (data set). The PRS terintuitive, because a better quality of governance Group, East Syracuse, NY. http://www.prsgroup is expected to be positively associated with revenue .com/ICRG.aspx. collection. This may be explained by the fact that Le, Tuan Minh, Blanca Moreno-Dodson, and Jeep in South Asia, the level of education of bureau- Rojchaichaninthorn. 2008. “Expanding Taxable crats, codes of conduct, and civil service rules and Capacity and Reaching Revenue Potential: Cross- regulations may be good on paper, but the actual Country Analysis.� Policy Research Working Paper revenue collection system on the ground may not 4559, World Bank, Washington, DC. function as well in conjunction with wrong incen- Mauro, Paulo. 1998. “Corruption and the Composition tives and poor implementation. of Government Expenditure.� Journal of Public 9. The regressions also included a dummy for Economics 69 (2): 263–79. resource-rich countries, which was highly signif- Pessino, Carola, and Ricardo Fenochietto. 2010. icant and positive, as expected. “Determining Countries’ Tax Effort.� Hacienda Pública Española 195 (4): 65–87. Piketty, Thomas, and Nancy Qian. 2009. “Income Inequality and Progressive Income Taxation in References China and India, 1986–2015.� American Economic Bird, Richard M., Jorge Martinez-Vazquez, and Journal: Applied Economics 1 (2): 53–63. Benno Torgler. 2004. “Societal Institutions and Rajkumar, Andrew Sunil, and Vinaya Swaroop. 2008. Tax Effort in Developing Countries.� ITP Paper “Public Spending and Outcomes: Does Governance 04011, International Tax Program, University of Matter?� Journal of Development Economics 86 Toronto, Toronto. (2008): 96–111. Davoodi, Hamid, and David Grigorian. 2007. “Tax Schneider, Friedrich, and Andreas Buehn. 2009. Potential vs. Tax Effort: A Cross-Country Analysis “Shadow Economies and Corruption All over the 16 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries World: Revised Estimates for 120 Countries.� Tanzi, Vito, and Hamid R. Davoodi. 1997. “Corruption, Economics: The Open-Access, Open-Assessment Public Investment, and Growth.� IMF Working E-Journal 1 (2007-9, Version 2). http://dx.doi Paper 97/139. International Monetary Fund, .org/10.5018/economics-ejournal.ja.2007-9. Washington, DC. Schneider, Friedrich, Andreas Buehn, and Claudio World Bank. 2011. More and Better Jobs in South Asia. Montenegro. 2010. “Shadow Economies All over Washington, DC: World Bank. the World: New Estimates for 162 Countries from World Development Indicators (database). World 1999 to 2007.� Policy Research Working Paper Bank, Washington, DC. http://data.worldbank 5356, World Bank, Washington, DC. .org/data-catalog/world-development-indicators. III Creating Fiscal Space through Raising Tax Revenue Context and Overview of Best Broadening the Tax Base Practices1 Dependence on a narrow base tends to require high tax Tax policies need to be framed against the background rates to generate a given revenue stream (Cetrángolo and of two important factors—efficiency and equity. Taxes Gómez Sabaini 2010; Gordon 2010), creating large costs are distortionary largely because they tend to create to those taxed, incentives to evade taxes, corruption, and, deadweight losses—reducing the population’s welfare ultimately, low revenue levels. Broadening the tax base more than the government can increase it by spending has thus always been a high priority for many developing the revenue. An efficient tax regime minimizes this countries. The tax base can be expanded by bringing distortion, and equity concerns are often addressed untaxed activities and more goods and services into the through a more progressive tax regime. tax net; by simplifying the tax structure, including by In general, taxes that are considered less efficient reducing exemptions; and by strengthening adminis- include (a) excise taxes and turnover taxes that cascade tration. Among indirect taxes, a VAT (with only a few and increase production costs and (b) import tariffs and exemptions and a small number of rates) has become export subsidies and taxes that create a wedge between popular in light of its core merits: it can raise significant domestic and international prices. Likewise, frag- revenue with relatively small economic efficiency losses; mented tax systems between central and local govern- it supports equity objectives; and it is relatively easy to ments and among local governments are also highly administer and to comply with, especially a single-rate inefficient because they increase the costs of produc- VAT. Exemption of basic foods is motivated by equity tion and can distort markets. reasons, and relatively high thresholds limit administra- In the absence of targeted safety nets or adequate tive and compliance costs. social security systems, many governments use tax Direct tax bases can be broadened by (a) expanding regimes to generate more equal patterns of income the types of income covered by taxes and (b) reducing distribution. Setting personal income tax thresholds tax preferences. For corporate income tax (CIT), the high enough to exclude the poor—twice the gross tax base can be expanded by including more firms, but domestic product (GDP) per capita when tax admin- mainly it is expanded by reducing deductions, exemp- istration capacity is limited (World Bank 1991)—or tions, and tax holidays. These incentives are introduced exempting unprocessed agricultural products from a in response to politically strong pressure groups rather value added tax (VAT) obligation are obvious exam- than as a way to address proven market failures. Not ples of such efforts. Other efforts include bringing only do such incentives erode the tax base, but they into the tax net professional income, self-employment also raise concerns of effectiveness, leakage, gover- income, or capital income—quite often those incomes nance, and spillovers—entrenching corruption in tax are earned by the rich. administration. Even some well-intended incentives There are three interrelated aspects regarding raising are quite often ineffective because they cannot override additional revenue through taxes in an efficient way: underlying market forces. Special economic zones, for broadening the tax base, rationalizing the rates, and example, are not sealed areas, because profits can be improving tax administration. transferred from nonholiday to holiday firms. South Asia Economic Focus – A Review of Economic Developments in South Asian Countries 17 18 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Box 3.1   Desirable Features of Tax Policy In lieu of an optimal tax structure, international best • Extend the coverage of the PIT with an effective practice suggests the following benchmark patterns of rate structure that is consistent with the authori- tax policy: ties’ distributional preferences, while keeping the effective maximum PIT rate equal or close to the • Minimize exemptions and tax incentives that uniform CIT effective rate. Keep the rate structure jeopardize revenue and good governance and that simple. generate no clear, offsetting social benefit. • Replace production or sales taxes with a simple • Remove minor taxes and fees that are costly to VAT that has a broad base and a high threshold— administer. minimizing exemptions and the number of rates, • Build CITs that are simple, broad-based, and compet- preferably to one non-zero rate. itive by international standards, and set effective tax • Lower trade tax rates that have fewer bands, rates that are reasonably low and uniform across in- and replace the loss with domestic sources (for vestments. A single statutory rate is recommended. example, VAT). Eliminate export taxes. Adopt an accelerated depreciation schedule. Personal income taxes (PITs) tend to have a narrow reason rates diverge from the single rate is when exemp- base, particularly in developing countries—again, tions are made for basic foods and goods that form a owing to exemptions and deductions but also to low large share of the consumption basket of the poor. income levels—and thus generate little revenue. The Trade taxes are easier to enforce, but they come lower the income per capita, the lower the PIT collec- with high economic efficiency costs. Despite a steady tion (Perry et al. 2006). The low yield can also reflect reduction in trade taxation in South Asia, trade tax poor design; for example, pension contributions or revenue still constitutes an important share of revenues. interest expenditures are deducted, while pension Regarding direct taxes, the main principles involve income and interest income are exempted. In addition, coordinating the top PIT and CIT rates to avoid tax as discussed above, much of professional and entrepre- arbitrage and resulting revenue losses. This CIT rate neurial income and agricultural income are outside the also needs to be internationally competitive. In many tax net, and even if they were included, such incomes countries, nominal rates of between 20 percent and could easily be hidden through, for example, offshore 40 percent are considered consistent with interna- accounts. Addressing these issues could go a long way tional competition. in expanding the tax base. Improving Tax Administration Rationalizing Rate Structures The success of tax administration reforms depends crit- There is no optimal tax structure, but international ically on the design of tax policy and administration best practice suggests benchmark patterns, which are capacity. With simpler structures, administration and summarized in box 3.1. enforcement are much easier, and the scope for corrup- In the case of goods and services tax, best practice tion and leakage is minimized. advice has been to adopt a VAT with a broad base, a In addition to strengthening the administrative appa- single rate, and a fairly high threshold. A single rate ratus, some countries have tried nonstandard approaches of 10–20 percent is often recommended if it has the to improve tax collection and compliance. For example, feature of crediting tax on inputs. The most common the government of India has made efforts to require Creating Fiscal Space through Raising Tax Revenue 19 anyone with a house, a telephone, a credit card, a car, or in national income. The other measure, tax yield, is an expenditure on foreign travel to file personal income the ratio of actual tax collection to the relevant tax tax (Gordon 2010). As noted by Stiglitz (2010), such rate, multiplied by the tax base: R / (TR ∙ TB). In this efforts to design a corruption-resistant tax structure may chapter, the tax bases are proxied by national income, prove essential in future tax reforms in South Asia. R / (TR ∙ Y), as in the following: Another challenge is the large size of the informal sector that is outside the tax net. For instance, according to R / (TR ∙ TB) = (R(TR,TB,TA,Other)) / (TR ∙ TB). the government of India, only 9 percent of workers are employed in the formal sector (Chai and Roy 2006). Tax yield typically reflects the tax system’s produc- For the purposes of exposition, the following simple tivity in collecting revenues, whereas tax buoyancy analytical framework is used to decompose government could be viewed as a measure of the overall respon- revenue (R) into various underlying determinants: siveness of tax revenue to economic growth. Moreover, a number of aspects of tax administra- R ≡ R(TR,TB,TA,Other), tion are analyzed to shed light on better institutional arrangements for tax collection. The next chapter looks where the major contributing factors are tax rate (TR), at opportunities in other sources of revenue (nontax tax base (TB), tax administration (TA), and other revenues) for further improvement in government revenue (Other).2 For exploration of the possible chan- revenue. nels of increasing revenue, the calculation can take the total differential (d) of the revenue formula above: Tax Composition and Regimes dR = ∂R / ∂TR dTR + ∂R / ∂TB dTB + ∂R / ∂TA dTA + ∂R / (∂Other) dOther. Tax Composition The formula lists the possible channels for revenue Figure 3.1 compares the revenue composition of some generation: optimizing the structure of the tax rate, South Asian countries with the average composition broadening the tax base, changing the composition of low-income countries and that of selected East of taxation, improving the quality of tax administra- tion, and increasing nontax revenue. The rest of this chapter explores each channel in turn. Figure 3.1  Revenue Composition South Asia’s tax compositions and rates are discussed (percentage of GDP) first, and a cross-country comparison shows the possibil- 20 ities of more fiscal space for further improvement in the tax structure. In addition, two common measures, tax buoyancy and tax yield, are constructed for each South 15 Asian economy, as well as other comparable economies, to assess the performance of tax systems. Tax buoyancy 10 measures the ratio of the percentage increase in reve- nues to the percentage increase in national income (Y, 5 gross domestic product), (dR / R) � (dY / Y) by substi- tuting the change-in-revenue formula: 0 . an ia ia a h ka p al nd (dR / R) � (dY / Y) = (∂R / ∂TR dTR + ∂R / ∂TB in es Re d es ep an st la In Ch ad on ki , N ai iL ea dTB + ∂R / ∂TA dTA + ∂R / (∂Other) dOther) Pa gl Th d Sr r In n Ko Ba ∙ (Y / R) / dY. Income tax VAT or sales tax Other indirect tax This tax buoyancy is an informative way to measure Customs duties Other the effect of change in the tax rate, tax base, tax admin- Source: World Bank staff calculations based on individual country data. istration, and other revenue, normalized by the change Note: GDP = gross domestic product; VAT = value added tax. 20 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Figure 3.2  Direct Tax Collection on inefficient trade taxes as easy sources of revenue. (percentage of GDP) Indeed, the figures show that, with the exception of 7 India, South Asian countries’ income tax collection tends to be less than other low-income countries. In 6 contrast, trade tax collection in South Asia tends to 5 be larger. In East Asia, the opposite is true: there are more income taxes and less trade taxes—while overall 4 revenue generation is higher. 3 Over time, direct tax collection has been increasing 2 only modestly in South Asia, with the exception of India (figure 3.2). India’s strong performance is 1 due largely to its high economic growth but also to 0 improvements in tax administration (box 3.2). Nepal’s 1 02 6 0 8 improvement is also notable and is based on efforts to 3 9 4 7 5 /0 /0 /1 /0 /0 /0 /0 /0 /0 / 00 01 05 09 07 02 08 03 06 04 improve administration, for example, through estab- 20 20 20 20 20 20 20 20 20 20 FY FY FY FY FY FY FY FY FY FY lishing a large taxpayer unit and improving tax audits. Bangladesh India Nepal In addition, high aggregate demand based on signif- Pakistan Sri Lanka icant remittance inflows is contributing to revenue Source: World Bank staff calculations based on individual country data. growth. Pakistan’s adoption of a low and uniform Note: FY = fiscal year; GDP = gross domestic product. CIT rate (35 percent) had a visible positive effect in fiscal year (FY) 2006/07, but its effect is tapering off. Sri Lanka’s efforts to increase income tax collection— Figure 3.3  Domestic Indirect Taxes for example, by bringing civil servants into the tax net (percentage of GDP) and introducing the economic service charge and an effective minimum CIT payment—had some positive 10 effect, but again, the effort does not appear to have 8 generated a sustained improvement. With regard to indirect taxes, again, the decline in 6 Sri Lanka and the increase in Nepal stand out (figure 3.3). In both India and Pakistan, the decline in the share 4 of indirect taxes from the mid-2000s is attributable to the fall in excise taxes, as part of the stimulus in their 2 FY2008/09 budgets. Indirect tax collection would have fallen more had it not been for special and ad hoc levies. 0 The share of trade taxes has been declining in South Asia, a trend that can be seen in every region in the 2 3 01 9 4 7 0 5 8 6 /0 /0 /0 /1 /0 /0 /0 /0 /0 0/ 01 02 08 09 03 06 04 07 05 0 world, but the speed of decline in the past decade has 20 20 20 20 20 20 20 20 20 20 FY FY FY FY FY FY FY FY FY FY been slow (figure 3.4). In the case of Nepal, the recent Bangladesh India Nepal reversal is due largely to increases in imports financed Pakistan Sri Lanka by large-scale remittances. In Sri Lanka, overall trade Source: World Bank staff calculations based on individual country data. tax has continued to increase because of newly added Note: FY = fiscal year; GDP = gross domestic product. special levies and paratariffs.3 Excluding these paratar- iffs, Sri Lanka’s trade taxation (import tariffs and export taxes) has been on the decline. In effect, these special levies make up for the decline in domestic commodity Asian countries. Perry et al. (2006) and other studies tax collection. Moreover, South Asian countries’ reve- note that the main cause of the gap between poten- nues are more import dependent than these figures indi- tial and actual revenues is a shortfall in direct taxa- cate. In Nepal, the incidence of taxation on trade is even tion—mostly in personal income tax. The countries higher than suggested by tariff revenue—60 percent of with lower income tax collection tend to depend more VAT collection is VAT on imports, and 50 percent of Creating Fiscal Space through Raising Tax Revenue 21 excise revenue is on imports. Similarly, 49 percent of Figure 3.4  Trade Taxation VAT in Bangladesh is trade based, and so are 46 percent (percentage of GDP) of VAT and 19 percent of excises in Sri Lanka. When 4.0 they are levied on final consumption goods, these taxes 3.5 have the same effects as an import tariff. The tax composition of South Asian countries is 3.0 thus that of low collection of income taxes (except for 2.5 India) and of trade taxation that has not declined as 2.0 fast as in other regions, matching the pattern of tax 1.5 collection seen in the lowest-income countries. 1.0 0.5 Tax Regimes 0.0 1 02 6 9 7 0 8 3 4 5 /0 /0 /0 /1 /0 /0 /0 /0 /0 / 00 01 05 08 09 06 07 02 03 04 This section examines how the relatively poor revenue 20 20 20 20 20 20 20 20 20 20 FY FY FY FY FY FY FY FY FY FY performance in South Asia may be related to the tax regimes and the scope for countries in the region to Bangladesh India Nepal raise additional tax revenues efficiently and without Pakistan Sri Lanka increasing distortions. The assessment involves Source: World Bank staff calculations based on individual country data. reviewing the key features of tax regimes of South Note: FY = fiscal year; GDP = gross domestic product. Asian economies and comparing them with the bench- mark “desirable� tax regime discussed in box 3.1. As mentioned before, the tax structure generally considered Box 3.2   Tax Reforms in India: A Success Story India has made significant progress in tax reforms. taxes to a VAT. The number of personal income tax In the past two decades, tax reforms have been rates was reduced to three: 10, 20, and 30 per- guided mainly by two sets of influential commit- cent. The exemption limit was also increased to Rs tee reports—first, the Tax Reform Committee (TRC) 50,000, and further to Rs 100,000. For corporate in- in 1991, and second, the Kelkar Task Force (KTF) come tax, the rates were simplified to three: 20, 30, reports on direct and indirect taxes in 2002—and by and 40 percent in 1992–93; the rates were further the Fiscal Responsibility and Budget Management lowered to 10, 20, and 30 percent in 1997–98. To Act of 2004. broaden the base further, the government imposed a Following the balance-of-payments crisis of 1991, minimum alternative tax in 1997–98, equivalent to 30 the TRC laid out a road map for tax reform that percent of book profits on the tax-exempt compa- would be consistent with the other significant struc- nies. The structure of import duties was simplified tural reforms being undertaken. The principles of the gradually after the TRC recommended a reduction in reforms were to broaden the tax base, lower rates the number and level of tariffs to 5, 10, 15, 20, 25, and reduce their dispersion, and make administration 30, and 50 percent by 1997–98. The top rate was simpler and more effective. The main reforms were reduced from 400 percent to 40 percent by 1997–98 (a) to reduce the share of trade taxes in total rev- and further to 12.5 percent in 2006–07. The KTF enue, (b) to increase the share of direct taxation, and continued the thrust of these reforms in the early (c) to reform indirect taxation by transforming excise 2000s. (continued) 22 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Box 3.2   Tax Reforms in India: A Success Story (continued) Tax Administration Reforms Through a tax credit mechanism, this tax is collected on value added in goods and services at each stage India has made significant strides in improving tax of sale or purchase in the supply chain. The imple- administration, and three initiatives stand out in this mentation of GST could lead to the abolition of other respect. The first initiative is the introduction of a tax taxes—such as octroi, central sales tax, state-level deduction at source as a way to reach the groups that sales tax, and so on—thus avoiding the multiple layers are hard to tax. The second initiative is to strengthen of taxation that currently exist in India. The combined information and accountability systems through the GST rate is proposed at about 16 percent. Experts say issuance of the permanent account number and the that GST is likely to improve tax collection rates and to introduction of the Tax Information Network. The third boost India’s economic development by breaking tax initiative is to use nonstandard measures to improve barriers between states and integrating India through compliance. Specifically, all residents of large cities a uniform tax rate. are required to file a tax return, as long as one of the India’s biggest challenge now is to expand the tax following six conditions is met: ownership of a house, net from the present number of approximately 50 ownership of a car, ownership of a club membership, million taxpayers (out of the 1.2 billion population and ownership of a credit card, participation in foreign the estimated 300 million in the middle class). This travel, or ownership of a telephone. India has suc- goal would require greater effort to expand the reach ceeded in using information technology and outsourc- of the tax administration to cover areas outside the big ing to strengthen tax administration. The Central cities as well as those transactions that are conducted Processing Center (CPC) in Bangalore has been very predominantly in cash (for example, real estate trans- successful in processing a majority of the income tax actions). Figure 3.2.1 suggests the potential gain from returns, with more than 80 percent of the taxes col- such an expansion of the tax net. lected being electronically filed. The CPC is managed by a team of the Income Tax Department, with the data entry and validation carried out by an external Box Figure 3.2.1  Share of India’s Population vendor. The wealth of data from the Tax Information That Pays Income Taxes Network and the CPC is now enabling the Income Tax (percent) Department to implement a credible risk-based audit system. 0.12 0.10 Reforms in the Pipeline 0.08 The Direct Tax Code is expected to be updated soon 0.06 (for the first time since 1961). With a goal of fur- 0.04 ther broadening the base and keeping the structure 0.02 simple, the revised code would phase out cesses and 0.00 surcharges and limit deductions, exemptions, and 01 98 92 02 89 94 04 91 00 99 93 03 90 95 86 97 20 7 88 96 06 08 05 87 0 20 19 19 19 19 19 20 19 19 20 19 19 20 19 19 20 19 19 20 20 20 19 concessions. On the indirect tax, India is preparing a Share of population with Share of population goods and services tax (GST) system. The GST is a top 1 percent income subject to income tax comprehensive tax on the manufacture, sale, and con- Source: Piketty and Qian 2009. sumption of goods and services at the national level. Note: GDP = gross domestic product. Creating Fiscal Space through Raising Tax Revenue 23 Table 3.1  Personal and Corporate Income Tax Structures of South Asian Countries Minimum income of PIT exemption maximum Maximum Number of PIT level (multiple PIT bracket CIT rates or Maximum PIT rates (including Maximum PIT of GDP per (multiple of Number of CIT representative rate minus CIT Country zero rate) rate (%) capita) GDP per capita) rates CIT rate (%) rate (%) Afghanistan — — — — — — — Bangladesh 5 25 180,000 1,180,000 6 27.5 -2.5 (4.2) (28.1) Bhutan 4 25 100,000 1,000,000 1 30 -5 (1.1) (11.2) India 4 30 160,000 500,000 2 33.99 -3.99 (3.3) (10.3) Maldives — — — — 1 25 — Nepal 4 25 160,000 2,500,000 1 25 0 (3.2) (49.5) Pakistana 17 25 350,000 (3.7) 4,550,000 6 35 -10 (48.6) Sri Lanka 7 24 0 3,000,000 6 28 -4 (12.7) Sources: World Bank country data; individual government websites; http://www.taxrate.com. Note: CIT = corporate income tax; GDP = gross domestic product; PIT = personal income tax; — = not available. a. There are two PIT rate schedules: one for wage earners and the other for the rest of the public. by researchers as desirable, and often recommended and Sri Lanka, in particular, have relatively by the World Bank and the International Monetary complex CIT rate structures with different Fund (IMF), has the following features: (a) a VAT sectors assigned different rates. The complexity with minimum exemptions and number of rates or of the Sri Lankan system is derived from various (b) a uniform CIT rate that is equal to or close to the concessionary rates granted to many industries maximum PIT rate (see box 3.1). (see the last section of this chapter on exemp- Tables 3.1–3.4 summarize the general features of tions and incentives), and they erode the tax tax regimes in each country in South Asia. From these base significantly tables, we can see that the tax policy in major South • Finally, almost all countries except for Pakistan Asian economies has moved toward the benchmark seem to have representative CIT rates close to the framework discussed in box 3.1. maximum PIT rates (25–35 percent), which are also within the range considered as internation- • The PIT structure is not very complex (except ally competitive. If we assume that the similar for Pakistan, which has two different rate sched- nominal rates mean similar effective rates, the ules, one of which has 17 tax brackets); the taxa- proximity between these two rates reduces the tion thresholds are relatively high at about three likelihood of taxpayers shifting declared income times the GDP per capita on average; and the to reduce tax liabilities. maximum rate is at reasonable levels of 20–30 Table 3.2 describes the representative indirect tax in percent—and none of the tax regimes has prohib- each country. Again, most countries appear to follow itively high rates. the benchmarks discussed in box 3.1. They tend to use • The CIT structure is also simple. Most coun- single or a minimum number of nonzero rates, and tries have one or two rates as representative, these rates—at 12–16 percent—are reasonable within although most have some exceptions. Pakistan the norm established through the experience of other 24 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Table 3.2  Indirect Taxes of South Asian Countries Range of Representative Number of nonzero rates Registration Zero-rated Country VAT or sales tax rate (%) nonzero rates (%) threshold Exemptions activities Afghanistan — — — — — — — Bangladesh VAT 15 1 15 2,000,000 Exports, cottage industry Bhutan Sales — — — — — — India VAT a 12.5 3 1.0–12.5 500,000 a — — Maldives None — — — — — — Nepal VAT 13 1 13 2,000,000 Agricultural Exports goods and others Pakistan Sales 16 1 16 7,500,000 — — Sri Lanka VAT 12b 3 5–18 1,800,000 — — Sources: World Bank country data; individual government websites; http://www.taxrate.com. Note: VAT = value added tax; — = not available. a. Reforms are planned (see box 3.2). b. Rate was lowered from 15 percent in 2011. Table 3.3  Trade Tax Structure in South Asian Countries Trade Average import Number of tariff Maximum import restrictiveness Export taxation or Country tariff (%) slabs tariff (%) indexa restriction Comments Afghanistan — — — — — — Bangladesh 23.9 4 25 9 Garment, 0.6% Average paratariff, Others, 0.7% 8% There are specific rates Bhutan 21.9b 100 — — — India 18.5 4 100 12.5 — Paratariff rates, about 4% Maldives 16.9 5 200 — — — Nepal 12.4 6 80 Wood products, — 70–200% Export bans on sand and gravel Pakistan 14.7 5 100 10 — — Sri Lanka 11.4 3 30 8 — Numerous paratariff rates Sources: World Bank country data; Country Policy and Institutional Assessment (CPIA), World Bank; individual government websites; http://www.taxrate.com. Note: — = not available. a. See Ivanic and Pitigala 2011. b. Bhutan’s effective rate is lower because 90 percent of trade is covered by the free trade agreement with India. Creating Fiscal Space through Raising Tax Revenue 25 Table 3.4  Comparison of Income Tax Rates across Regions (percent) Region Average CIT general rate Average minimum marginal PIT rate Average maximum marginal PIT rate East Asia and Pacific 24.0 9.0 29.0 Central Europe and Central Asia 15.8 13.3 19.4 Latin America and the Caribbean 28.7 13.5 29.2 Middle East and North Africa 24.9 10.4 26.0 South Asia 32.1 8.4 26.7 Sub-Saharan Africa 30.3 10.1 35.2 Western Europe 23.6 15.6 36.9 United States and Canada 26.5 12.5 32.0 Sources: World Bank staff calculations; individual government websites; http://www.taxrate.com. Note: CIT = corporate income tax; PIT = personal income tax. countries. Only India’s VATs are fragmented; rates Performance Assessment of South and administration differ by state. The government is Asian Tax Regimes well aware of the differences and has designed a signif- icant tax reform to rectify the situation—including the In addition to international comparisons, we conduct reform of the direct tax system and introduction of two additional analytical exercises to assess tax regimes vastly improved goods and services tax (GST) imple- in South Asia. 4 Buoyancy measures the ratio of mented as a VAT. India’s tax reform experience and percentage increase in revenues to percentage increase future plans are discussed in box 3.2. in GDP.5 Tax efficiency measures the ratio of actual Table 3.3 describes trade taxation in South Asia. tax collection to potential revenue, with the base In general, the average (unweighted) tariff rates are multiplied by a representative tax rate. A glossary declining and are between 10 percent and 25 percent, of tax analysis terms used in this report is shown in with Nepal and Sri Lanka at the lower end and box 3.3. Bangladesh on the high end. The average announced tariff of Bhutan is also high, but the majority of its trade is with India and is under the free trade agreement. Buoyancy So its effective rate is much lower. The maximum rate for some countries is high—200 percent in Maldives, Buoyancy is a useful metric to examine the performance 100 percent in Bhutan and India, and 80 percent in of a tax regime (table 3.5).6 The concept of buoyancy Nepal. These high rates are usually reserved for a few of revenue is an important aspect of the tax structure, luxury items, including vehicles. The number of tariff and it may be useful as a descriptive tool, which could slabs is not excessive either. Still, many countries in the lead to further questions and a more detailed examina- region are considered to have highly restrictive trade tion of particular taxes in any country. In our concep- regimes (Ivanic and Pitigala 2011) through nontariff tual framework of this analysis, buoyancy is the sum barriers or as a result of trade facilitation problems, of (a) percentage change in the average tax rate; (b) but a fuller examination of these issues is beyond the change in the size and composition of tax bases relative scope of this book. to GDP; and (c) change in tax administration efforts. Table 3.4 shows a comparison of general CIT It also measures the level of progressivity as income rates and lowest and highest marginal PIT rates across increases. Tax-specific buoyancies for each country are various regions. The average CIT rate is high in South shown in appendix B.7 Asia, but the PIT rates are lower in South Asia than Abstracting from Afghanistan, where revenues are in most other regions. rising from very low levels, the highest buoyancy is that 26 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Box 3.3   Glossary of Key Concepts Used in the Analysis of Taxes Tax buoyancy. A country’s percentage increase in tax Tax yield. A measure of how productive a tax is in revenue divided by percentage increase in GDP (or collecting revenues. Yield is the actual revenue as a some measure of national income). share of potential revenue; the latter is proxied by the product of representative tax rates and the target tax Tax elasticity. A country’s percentage increase in tax base, such as GDP or imports. revenue (if we assume that no change in the tax re- gime—tax rates, coverage, and so forth—takes place) Tax regime. Description of a country’s tax system as a as a share of the percentage increase in GDP (or some whole, including tax instruments employed, rates used, measure of national income, as a potential tax base). bases intended, and exemptions offered, among others. Table 3.5  Revenue Buoyancies in South Asia recent past when GDP is used as the base for calcula- Country Overall revenue buoyancy tion. The same is true for trade tax. But when private consumption is used as the base for VAT and total Afghanistan 1.87a trade is used as the base for trade tax, respective buoy- Bangladesh 1.31 ancies show either no improvement (VAT) or very slow Bhutan 1.44 improvement (trade tax), indicating consumption and India 1.00 imports have been growing much faster than GDP. In Maldives 1.31 fact, appendix B indicates that consumption seems to outpace GDP in many countries in South Asia. Nepal 1.50 For Bangladesh, buoyancies of both tax revenue and Pakistan 1.15 nontax revenue are consistently above 1.2, and both Sri Lanka 0.61 income tax and tax on goods and services have relatively Source: World Bank staff calculations based on individual country data. high buoyancy rates—indicating that Bangladesh’s a. Simple average of annual buoyancies. revenue structure is stable and that major components are growing at above GDP growth rates. Only trade taxes have less-than-unitary buoyancy. The country’s low overall revenue performance may in fact be the of Nepal, with Bhutan, Maldives, and Bangladesh not result of a low base and low GDP growth rates (until far behind. Pakistan and India have close to unitary the recent past) rather than low buoyancy. buoyancy. Sri Lanka has the weakest buoyancy at 0.6, India’s average revenue buoyancy masks differences consistent with its declining revenue trend. across direct and indirect taxes and also changes over Nepal’s high buoyancy is based on the robust perfor- time. Overall revenue buoyancy is 1.0, but income mance of domestic indirect taxes throughout the period taxes are more buoyant, at 1.6 on average for the past as well as on recent improvements in income tax collec- decade. Over time, however, income tax revenue buoy- tion. Import duty collection has also increased recently, ancy peaked at 2.4 (FY2006/07) before declining to much more rapidly than GDP, because of remittance- about 1.2 over the past three years and to below 1.0 fueled imports. In fact, because 60 percent of VAT over the past two years. Indirect tax buoyancy has also revenue and 50 percent of excises are on imported declined recently. Buoyancy for trade tax has been weak goods, Nepal’s improvement in performance has been throughout the period. import dependent. These points are illustrated clearly in Pakistan has a pattern of buoyancy similar to India’s. appendix B, where tax-specific buoyancies are discussed. Over the past decade, the average buoyancy for indi- Nepal’s VAT buoyancy improves significantly for the rect tax was slightly above 1.0, and for income tax it Creating Fiscal Space through Raising Tax Revenue 27 was 1.3. Over the past three years, the buoyancy for Figure 3.5  Income Tax Yields, CIT Rate income taxes weakened to 0.9, but the indirect tax buoyancy remained close to unitary. Trade tax buoy- 0.35 ancy has also been weak in Pakistan. 0.30 Sri Lanka’s low buoyancy seems to come from poor 0.25 performance of all tax instruments. No individual tax category has buoyancy above 1.0, whether averages are 0.20 taken over the past 10 years or 3 years. Buoyancy is partic- 0.15 ularly low for VAT. The highest tax-related buoyancy, 0.9 for income tax, is also weaker than other countries 0.10 in South Asia, and even that has been declining recently. 0.05 0.00 . sh ep ka a a e am Tax Efficiency n es a nd al in si ag di ta de an ep in ,R ne la Ch tn In is er pp la ai iL N k ea e do av Pa ng Vi Th ili Sr r In Ko Ph C Ba In this section, each of the major tax categories is LI analyzed by calibrating tax yields in an effort to iden- LIC Average South Asia East Asia tify which group of taxes has lower productivity and Sources: World Bank staff calculations based on South Asian country data; IMF 2011 for East Asian where additional fiscal space could be created (see data. Purohit 2006). These yields are ratios of actual collec- Note: CIT = corporate income tax; LIC = low-income country. tion to the relevant tax rate multiplied by GDP (proxy for tax base).8 Discussion of primarily yields of income tax and VAT (figures 3.5 and 3.6) includes a compar- ison with several East Asian countries (IMF 2011). Figure 3.6  VAT Yields Based on GDP These figures confirm that most South Asian coun- 1.2 tries have yields lower than the average of low-income countries (LICs) or the yields in East Asian countries. 1.0 In the case of income tax, Sri Lanka and Bangladesh are particularly weak, with tax yields in the range of 0.8 less than 10 percent. Pakistan and Nepal are about 10 0.6 percent, and India stands out at 17 percent. But even India’s efficiency is below the average of LICs and well 0.4 below most East Asian comparators, whose yields range from 15 percent to 33 percent. 0.2 In the case of VAT (or sales tax), all South Asian countries are below East Asian comparators and well 0.0 . ep ka sh n es a a e m al nd ia below the LIC average. Pakistan and Bangladesh have di in ta ag ep an na es de ,R in la In Ch is er pp n N iL et ai la k ea do av Pa yields of about 20 percent, India and Sri Lanka about Th ng Vi ili Sr r Ko In Ph C Ba LI 22 percent, and Nepal 36 percent. The LIC average is LIC Average South Asia East Asia about 40 percent, and East Asian country yields are sometimes near or exceed 100 percent. Sources: World Bank staff calculations based on South Asian country data; IMF 2011 for East Asian Over time, tax yields in South Asian countries have data. Note: GDP = gross domestic product; LIC = low-income country; VAT = value added tax. been improving. For income tax, India has shown its largest improvement over the past 10 years, and Nepal has experienced rapid improvement since FY2005/06 (figure 3.7). In the case of VAT, only Nepal experienced later in this chapter for a more detailed discussion of a rapid increase, again since FY2005/06 (figure 3.8). Sri Lanka). Income tax yields in Pakistan and Sri Lanka Bangladesh’s yields are improving steadily but slowly have stagnated since about FY2006/07, and India and for both taxes. In contrast, VAT yields in Sri Lanka Pakistan have shown no significant improvement in have plummeted over the past seven years (see box 3.5 VAT yields throughout the decade. It is of particular 28 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Figure 3.7   Income Tax Yields, FY2000/01–FY2008/09 issues of the two countries are somewhat different from those of the rest of South Asia. 0.20 0.16 Tax Administration in South Asia 0.12 Weak tax administration is a key barrier to effective and fair tax collection in South Asia. Together with the 0.08 narrow tax base, it is a major cause of low tax yields in South Asia. Most countries have not updated their tax 0.04 laws to incorporate modern principles and the growing 0.00 complexities of the economy. According to experiences in a range of developing countries over the past two 1 8 9 02 7 3 4 5 6 /0 /0 /0 /0 /0 /0 /0 /0 / 00 decades, accepted best practices generally include the 07 08 01 06 02 03 04 05 20 20 20 20 20 20 20 20 20 FY FY FY FY following elements. FY FY FY FY FY Bangladesh India Nepal Institutional arrangements. A critical mass of tax Pakistan Sri Lanka administration around the world has to some degree moved toward establishing semiautonomous legal enti- Source: World Bank staff calculations based on individual country data. Note: CIT = corporate income tax; FY = fiscal year. ties (revenue authorities) to respond to poor civil service institutions, political interference, or rigid budget and financial regulations that prevail in line ministries. The experience—although not uniform across coun- tries—has generally been positive. In contrast, in South Asia, tax administrators are attached to the ministry of Figure 3.8  VAT Yields, FY2001/02–FY2009/10 finance with varying degrees of independence. Even 0.5 when the administrators have some degree of inde- pendence, the critical power of human and finan- 0.4 cial resource management remains with the ministry of finance. The tax officials are also part of the civil 0.3 service, and their pay and promotion are determined through general human resource policies governing 0.2 the civil service. Organization. Another important recent reform has 0.1 been the shift from organizations focused on the type of tax they collect to a functional (for example, audit, 0.0 legal, or collection) or client-based organization (for 8 9 4 5 0 3 6 7 2 /0 /0 /0 /0 /1 /0 /0 example, large taxpayers). In South Asia, a tax admin- /0 /0 07 08 03 04 09 02 05 06 01 20 20 20 20 20 20 20 20 20 istrator’s responsibilities are still based on the type of FY FY FY FY FY FY FY FY FY Bangladesh India Nepal taxes, although many countries have introduced large taxpayer units. Pakistan Sri Lanka Use of information technology (IT) to deliver services Source: World Bank staff calculations based on individual country data. to taxpayers. The use of IT applications can enhance Note: FY = fiscal year; GDP = gross domestic product; VAT = value added tax. tax administration. Examples include electronic filing, e-payment, e-refunds, and virtual tax offices. Financial transactions in South Asia still take place largely in cash, concern that Bangladesh, Pakistan, and Sri Lanka suffer making it difficult to control tax evasion and gover- from low yields of both taxes. nance problems such as bribery. The tax administra- In addition, box 3.4 discusses separately revenue tion and the ministry of finance mandate invoicing and composition and fiscal management issues of Bhutan reporting for large transactions conducted in cash, but and Maldives because the revenue patterns and fiscal the success of those requirements remains highly elusive. Creating Fiscal Space through Raising Tax Revenue 29 Strategic management. To respond to higher demand required, collection enforcement, and independent by stakeholders for more efficient and accountable appeal mechanisms to balance the powers of inves- tax collection, many tax administrators in developing tigation and collection with the rights of taxpayers. countries have begun to produce multiperiod business In most South Asian countries, such procedures are plans and transparent reports of performance outcomes. onerous, and the legal system is ill-equipped to adju- Best international practices include the formation of dicate tax cases expeditiously. tax analysis units that produce reports on key perfor- Audit and compliance enforcement. Key to effec- mance products, such as tax gap analysis, revenue fore- tive administration are effective risk management and casting, and tax expenditure budgeting. In general, these enforcement strategies to deter, detect, and address practices have not been widely adopted in South Asia. noncompliance. In South Asia, compliance costs Simplification of legal procedures to resolve tax remain very high. disputes. Key elements include the existence of a single For a sense of how South Asia compares with other tax procedure code, a reasonable penalty structure, a countries in tax administration, table 3.6 presents reduced number of tax returns and payments that are information on the time spent on filing and paying Box 3.4   Bhutan and Maldives: Revenue Composition and Fiscal Management This box discusses the composition of revenues in In Maldives, tourism-related revenues are domi- Bhutan and Maldives separately from the other South nant. The tourism tax was nominally fixed until recent- Asian countries because of their very high depen- ly. Revenues were based on the number of tourism dence on nontax revenues and the very narrow bases person-nights, import duties were in large part levied of their tax regimes. on imports by resorts, and a large portion of nontax Revenue performance in Bhutan has been good. receipts were lease payments made by resort islands. The average revenue of the past four years was 23.5 Income tax receipts are almost exclusively bank profit percent of GDP , well above the regional average (figure tax collection and form a tiny portion of total revenue. 3.4.1). Nontax revenue is high (above 10 percent of Both tax and nontax revenue suffered considerably GDP), with profit transfers and dividends related to with the onset of the global financial crisis that started hydropower projects. Indirect taxes (in particular, in 2008—with revenue dropping to 26 percent of GDP excises) and income taxes have performed well— in 2009 from near 50 percent the year before (figure again, the major driver of this performance has been 3.4.2). The crisis underscored the two weaknesses of the hydropower sector. In fact, high fiscal revenue has the tax system in Maldives—that is, high dependency been generated mostly by the hydropower sector and on tourism income, which could be highly elastic and has been a key contributor to higher public invest- thus volatile; and the lack of a strong domestic tax ment. But revenues from the hydropower sector are base (such as sales or income taxes). The government volatile annually and seasonally within a year. The chal- undertook a major tax reform drive in 2010 aimed at lenge for the Bhutanese authorities in the short run moving to a more broad-based tax system. According- is to manage fiscal outlays well in the environment of ly, a goods and services tax (GST) on tourism became revenue volatility within a fiscal year. The medium-term effective in January 2011 that covers a broad array of challenges revolve around the need to create fiscal tourist consumables (and not just bed-nights), and a space to finance investment as aid inflows decline. domestic GST also came into effect in October 2011. Expenditure now stands at 32 percent of GDP and is A broader CIT regime spanning beyond the banks was set to continue rising with the 20 percent civil service also implemented in July 2011. Together these mea- wage increases in 2009 and 2011. Developing non- sures are expected to add approximately 12.5 percent hydropower-related revenues is important, and these of GDP in revenue over the medium term—effectively are most likely to come from the introduction of a VAT bringing the revenue-to-GDP ratio close to about 34 and improvements in tax administration. percent. (continued) 30 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Box 3.4   Bhutan and Maldives: Revenue Composition and Fiscal Management (continued) Box Figure 3.4.1  Bhutan: Revenue Box Figure 3.4.2  Maldives: Revenue Composition, FY1998/99–FY2009/10 Composition, FY1998/99–FY2009/10 (percentage of GDP) (percentage of GDP) 25 50 20 40 15 30 10 20 5 10 0 0 0 20 9 20 1 20 3 20 4 20 2 20 5 20 7 20 8 99 9 FY 000 20 6 /1 FY 0/0 20 2 FY 8/0 20 9 FY 2/0 20 1 /0 FY 1/0 20 3 FY 4/0 0 20 4 20 5 FY 6/0 20 6 /0 20 8 19 8/9 99 9 FY 5/0 FY 000 20 7 FY 1/0 FY 0/0 /1 FY 8/0 FY 2/0 FY 3/0 FY 4/0 FY 5/0 FY 7/0 09 19 8/9 /0 03 07 /2 0 0 0 0 09 0 0 9 0 06 0 /2 20 0 0 0 0 0 0 19 0 9 20 19 FY FY FY FY FY FY FY Income taxes Tourism taxes Income taxes Taxes on goods and services Import duties Nontax revenues Trade taxes Other Nontax revenue Source: World Bank staff calculations. Source: World Bank staff calculations. Note: GDP = gross domestic product. Note: GDP = gross domestic product. taxes. On average, taxpayers in South Asia spend over than in developed countries where the higher degree of 330 hours per year filing and paying taxes, consider- computerization and electronic processing could result ably more than in East Asia and Pacific and countries in fewer staff involved in tax administration; and (c) of the Organisation for Economic Co-operation and the number of active taxpayers per tax administration Development, and broadly in line with Latin America employee is lower in South Asia than in most regions. and the Caribbean, Eastern Europe and Central Asia, However, the interpretation of this last number is not and Sub-Saharan Africa. Within South Asia, Bangladesh, straightforward because it measures both staffing of tax Nepal, and Pakistan have the highest numbers. administration and extent of tax compliance. We also look at other measures of the quality of We take a somewhat closer look at Bangladesh, tax administration systems, such as the cost of tax Nepal, and Pakistan as follows. We examine how closely collection as a share of tax revenue, the number of they conform to the best practices outlined above. tax administration personnel per 1,000 people living in the country, and the number of taxpayers per tax administration employee (table 3.7). The numbers in the table show that (a) tax administration is more Bangladesh costly in South Asia than in East Asia and Pacific, Bangladesh’s tax administration operates in a largely Central Europe and Central Asia, Latin America and manual environment, and the compliance burden is the Caribbean, and developed regions; (b) the number onerous. Tax disputes are held in the court system for of tax administration personnel per 1,000 residents in several years, locking up tax revenue and creating cash South Asia is lower than in other regions, lower even flow problems for businesses. Because of a long-standing Creating Fiscal Space through Raising Tax Revenue 31 legal dispute, the National Board of Revenue (NBR) Table 3.6   Time Spent on Tax Preparation and Filing, has been unable to recruit enough newly trained offi- 2008–12 Average cers for many years, further impeding overall opera- Country or region Time spent tions and service delivery. South Asia (excluding Maldives)a 331 A recent report on tax administration highlighted Afghanistan 275 the following reasons for poor income tax collection: Bangladesh 322 • Weakness in the income tax law, such as the exces- Bhutan 274 sive use of discretion and withholding India 265 • The proliferation of tax incentives and concessions Nepal 364 • Poor compliance arising from weak enforcement Pakistan 560 • Limited use of taxpayer information from other taxes paid, as well as third-party information Sri Lanka 256 • Excessive use of presumptive taxes as a final tax, East Asia and Pacific 237 irrespective of taxpayer size China 535 Indonesia 266 Similar problems of weak enforcement also plague Korea, Rep. 253 the VAT collected on domestic goods because the system is designed primarily like the old excise system: Malaysia 147 • More physical monitoring of businesses, rather Thailand 264 than the use of books, accounts, and secondary Latin America and the Caribbean 390 information Eastern Europe and Central Asia 302 • Treatment of each unit of a multiunit business Middle East and North Africa 188 as a separate business entity • Use of deemed value added based on a percentage Sub-Saharan Africa 318 of sales (truncated VAT) OECD high income 186 • Tax liability based on approved prices rather Sources: World Bank 2007, 2008, 2009, 2010, 2011. than actual prices Note: Data are for the time it takes to prepare, file, and pay (or withhold) the corporate income tax, • Overreliance on presumptive VAT the value added tax, and social security contributions (in hours per year). a. Maldives imposes no direct taxes. Separately, exchange of information between the VAT and income tax offices is limited, thus reducing Table 3.7  Tax Administration Indicators Average number of active Average administration cost Average number of tax staff per taxpayers per tax administration Region (% of revenue) 1,000 residents employee East Asia and Pacific 1.2 0.4 568.1 Central Europe and Central Asia 1.1 1.0 355.4 Latin America and the Caribbean 1.3 0.4 388.6 Middle East and North Africa 3.2 0.5 391.8 South Asiaa 1.8 0.3 258.2 Sub-Saharan Africa 2.9 0.3 257.1 Western Europe 0.9 1.3 764.9 United States and Canada 1.5 0.7 1,593.9 Source: USAID 2010. a. Data are available for Bangladesh, Bhutan, India, and Sri Lanka. 32 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries information for the NBR to use in assessing the compli- true e-filing system, which is being implemented in the ance risk of its taxpayer base. These problems exist next reform phase. These IT reforms improve service because NBR’s organizational structure is not condu- delivery, reduce face-to-face contact between tax officials cive to efficient tax administration. Essentially, the and taxpayers, and thus reduce corruption possibilities. two wings—the indirect tax wing (VAT-customs) and The Inland Revenue Department (IRD) has intro- the income tax wing—operate independently and are duced a new performance management system to staffed by officers of different cadres of the civil services. reward good performance. The primary aims of these This division hampers organizational reform efforts, as reforms have been to reduce the contact between the evidenced in the issues related to the creation of the taxpayers and the tax administration and, as a result, to large taxpayer unit. Although this division has resulted reduce possibilities of corruption. The IRD has imple- in better monitoring of the primary revenue contribu- mented a risk-based audit system, but the system does tors, efficiency is limited, because separate large taxpayer not use third-party information and needs upgrading. units—one for income tax and the other for VAT— The next phase of IT reforms involves better reporting had to be created. of third-party information as well as that of the different Past efforts to automate the tax processes have been taxes to the IRD that would then feed into an effec- largely unsuccessful. There has been no comprehensive tive risk-based audit system. effort to automate the entire tax administration. Such These reforms have had the effect of raising tax buoy- an effort would require a change in business processes, ancy (in 2009, revenues rose by 25 percent, in part helped including reorganization of the entire NBR. Because by a high remittance rate and a system of one-time reorganization could upset the balance of power between voluntary declaration of income). However, the overall the separate tax wings, as well as the powers of the tax tax collection as a percentage of revenue continues to be officials, such an effort has been resisted. One example low, and there is scope for increases in the tax collection. of the difficulties in starting coordinated reforms is the There are too few registrants (450,000 for CIT, and less demand for the income tax wing and the VAT wing to than 100,000 for VAT), and fewer—about 50 percent have separate systems of tax identification numbers while of income tax registrants and 80 percent of VAT regis- each wing currently has a separate plan for automation. trants—actually file returns. Also, a significant weak- The NBR is developing a modernization plan that ness is the collection of VAT and customs duty at the covers reforms over the next five years. Together with border. Because the border—especially that with India— the organizational reforms, the key reforms include new is very porous, there is substantial leakage of revenue VAT and income tax laws and automation of processing. owing to smuggling. Those imports that come through The NBR has now embraced the idea of outsourcing the the proper channel are in most cases heavily underin- routine tasks that could be automated, such as the issu- voiced, which has been the subject of a recent national ance of tax identification numbers, filing of tax returns, VAT fraud investigation.9 and electronic payment of tax. The hope is that these As a result, the efforts of the IRD, while being reforms would address some of the problems that the tax somewhat successful with the internal tax administra- processes face. The NBR is aiming to achieve a tax-to- tion, are not matched by the tax collection efforts of GDP ratio of 13 percent by 2016 through these reforms. the Department of Customs at the borders. There is a move to bring customs and the IRD under a central revenue board, with the hope of improving the effec- Nepal tiveness of tax collection. A means to exchange infor- mation between the Indian and Nepalese customs Nepal has undertaken a series of tax administration authorities is necessary to have any credible effect on reforms over the past decade, starting with the estab- tax evasion in transactions at the border. This is one lishment of a functional organization structure that of the next big challenges for the Nepal government. includes a large taxpayer office. Several reforms through the use of IT have recently been implemented, including an e-PAN (permanent account number) system that Pakistan allows taxpayers to register online and an e-filing system that allows them to file tax returns online. However, Weak revenue mobilization has been Pakistan’s core the lack of an e-payment system means this is not a development challenge for many years because of an Creating Fiscal Space through Raising Tax Revenue 33 inefficient tax administration, a narrow tax base, a translated into demonstrably higher revenues. TARP’s complex and nontransparent tax system, and corruption original revenue target has not been met owing to a and tax evasion. To address the administration issues, the mix of unfavorable economic circumstances, including government has invested significantly in physical and a challenging environment with frequent changes IT-related infrastructure to support a functionally based in political authorities and senior FBR staff, weak organization operating with revised business processes in implementation capacity, and problems with basic key areas: taxpayer registration, filing, payment, enforce- infrastructure such as electricity. Other features that ment, legal matters, and refunds. These are assisted by have adversely affected the outcomes so far include the World Bank’s Tax Administration Reform Project (a) an overly ambitious agenda; (b) optimistic tax (TARP), which began in 2004. policy assumptions that earlier government commit- TARP has helped the government make significant ments—for example, introduction of a modern VAT progress in enforcement, IT, organization, and human and effective simplification of the tax regime—would resources, thus establishing the foundations for future be implemented as envisaged; (c) underestimated and deeper reforms. The project has adequately funded legal constraints (for enforcement and for the FBR’s improvements in physical and IT infrastructure and organizational restructuring); and (d) underestimated properly focused on taxpayer services. Specifically, the staff resistance to the FBR’s integration initiatives. following reforms have been achieved through the Managerial problems have also been significant: insuf- implementation of TARP: ficient follow-up of reform actions, limited coordina- tion between headquarters and field formations, and • Service to taxpayers by the Federal Board of insufficient ownership and accountability—leading to Revenue (FBR) has significantly improved and incomplete implementation of the committed reforms is changing the image of the FBR in society. with no one held accountable. At a more technical level, • Implementation of a new functionally integrated underutilization of newly deployed IT systems and weak organization, as opposed to the previous orga- management and implementation of newly designed nization based on tax type, will allow improved core business processes have also been problematic. operations at the FBR. Nonetheless, the government of Pakistan’s efforts, • The amount of refunds processed has increased aided by TARP, have laid the foundation for further dramatically, reaching PRs 106.4 billion in progress. Once the primary management and legal 2010–11, an increase of 233 percent above the issues are resolved, the FBR can be in a better posi- 2009–10 level and equivalent to 0.59 percent tion to jump ahead to significantly improve its opera- of GDP (the highest in recent years). tions and, as a result, increase tax revenues. The earlier • Enforcement has improved significantly through envisaged tax policy changes and legal improvements (a) introducing the new taxpayer registration would further add to the tax revenue collection and system, thereby increasing registered taxpayers by efficiency of the system. 25 percent since 2008; (b) introducing e-filing, which led to a substantial reduction in nonfilers; (c) reducing tax arrears (by 22 percent in 2011); and (d) restarting audit activities, which had Tax Expenditures and the Tax Base been suspended for about four years. In the analytical framework introduced at the start of • The FBR infrastructure was drastically improved this chapter, the third factor in having an efficient tax by modernizing buildings and by improving the system (after tax rates and tax administration) is the IT infrastructure to expand communications and tax base. In developing countries, tax bases are most computerization, which allowed the introduc- often affected by the practice of exempting the whole tion of the Integrated Tax Management System or part of tax liabilities for certain individuals or firms, at all field offices. or for certain sectors. This practice is also widespread • The FBR approved a human resources manage- in South Asia. These incentives are also known as tax ment policy framework in 2010. expenditures because they are equivalent to expendi- tures made through the tax system—as if collecting Although the FBR has achieved these notable proper taxes and then returning to selected taxpayers results, concerns remain, because results have not been the amount of prescribed tax incentives as spending. 34 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Tax expenditures are typically introduced with the The perception is that in South Asia, tax exemptions intention to achieve equity objectives or to enhance and tax expenditures are higher than international aver- investment and promote growth by correcting real or ages, but most developing countries, including countries perceived market failures. Quite often, the objectives are in South Asia, do not maintain systematic data on tax to attract foreign direct investment, promote exports, expenditures. Even among developed countries, where and encourage investment, as well as to promote equity. the data are compiled, methodologies vary so much that Offsetting these potential benefits are a number of comparing tax expenditures is difficult. For example, negative factors: tax expenditures reduce the tax base tax expenditures range from 2 percent of GDP in the and lower revenue, they escape the spending prioriti- Netherlands to 4 percent in Australia and Canada to zation process, and they complicate tax administration. approximately 7 percent in the United States, and data In turn, this complication provides scope for abuse of problems are even more severe in developing countries. the tax system by taxpayers, bureaucrats, and govern- Mortaza and Begum (2006) report that tax expendi- ment officials, reducing revenue further (Swift 2006). tures in 2005 were 2.5 percent of GDP in Bangladesh, An examination of individual tax instruments clarifies 4.5 percent in India, and 0.4 percent in Pakistan. Swift the distortions associated with tax incentives as follows: (2006) reports that tax expenditures were 5 percent in Turkey (2004) and well over 10 percent in China • CIT-related incentives (tax holidays or lower (2002). The IMF reports Nepal’s tax expenditures as 2.2 CIT rates) are typically used to attract invest- percent in 2011. These figures are the result of totally ment. But these incentives tend to benefit the different calculation methods and, for developing coun- most profitable firms, which arguably would tries, they are based on many assumptions. One obser- have invested even without the incentives. In vation we can make is that South Asian countries may addition, providing CIT incentives to foreign not have clearly higher shares of tax expenditures than firms could simply subsidize their home country countries in other regions. treasuries, because most South Asian countries That being said, in one clear case—Sri Lanka— have treaties to eliminate double taxation with generous tax incentives have continuously eroded the most foreign countries. Even without such trea- tax base and reduced revenue over the course of the ties, CIT rates tend to converge to some interna- past 10–15 years (see a detailed discussion in box 3.5). tional norms, and having lower rates than those norms merely benefits foreign governments. CIT incentives are also costly because transfer pricing Notes is easy for both domestic and foreign investors. • Incentives related to PIT are used mostly for equity 1. The best practices are based on IMF (2011), Tanzi reasons. Because most of the poor in developing and Zee (2000), Barbone et al (1999), and World countries do not pay PIT in any case, most coun- Bank (1991). tries tend to achieve equity objectives by exempting 2. Tax rate describes the ratio at which a business agriculture from PIT. Although these exemptions or person is taxed. More generally, it includes may cause distortions and some income diver- the composition of different types of taxes in the sion, agricultural income tax exemptions tend national fiscal system, such as income tax, corpo- to be politically charged and difficult to reform. rate tax, and property tax. Tax base refers to the • Incentives related to indirect taxes (including assessed value of a set of assets, investments, or import taxes) often involve exemptions of income that is subject to taxation, such as the certain sectors, and these exemptions tend to value of real estate. Tax administration denotes the increase the scope for abuse and circumvention. overall fiscal architecture that manages, supervises, Tax credits and investment allowances are more and enforces the collection process of the taxes, transparent incentives, and accelerated depre- including tax legislation, statutes, and conven- ciation tends to be neutral over time, both as tions. Other revenue is the nontax revenue. an investor incentive and for revenue collection 3. Paratariffs include import and export license fees, (Tanzi and Zee 2000). However, South Asian port and airport development levies, cess (or countries do not appear to use these better alter- commodity) levies, motor vehicle concessionary natives very much. levies, and others. Creating Fiscal Space through Raising Tax Revenue 35 Box 3.5   Sri Lanka: Tax Exemptions Lead to Tax Base Erosion and Complicate Tax Administration As noted earlier, Sri Lanka stands out among South The Growing Scope of Exemptions Asian countries as having experienced a steady decline in tax revenues. Among other reasons, this In the late 1970s, as part of its export promotion decline is because of a long history of frequent and ad strategy, Sri Lanka introduced an elaborate tax incen- hoc changes to the tax policy. Those changes, together tives regime, which was administered through the with the growing number of exemptions and incen- Board of Investment (BOI). BOI-registered compa- tives—largely administered by agencies other than nies were typically located within export processing the tax authorities—have complicated the system and zones and were exempt from certain provisions of undermined revenue administration. the Inland Revenue Act, Customs Ordinance, and exchange control regulations. In addition, these A History of Frequent Changes companies had the right to employ expatriates, to borrow abroad, and to repatriate capital from their As shown in figure 3.5.1, tax collection worsened investments. significantly after the early 1990s. This erosion in rev- In 1982, the BOI also took over the customs enues coincided with an important change in the tax functions within the zones, allowing BOI-registered structure; that is, the country made a transition to a companies to circumvent the perceived inefficiencies GST system from the erstwhile turnover tax system.a of the Customs Department. At the beginning of the The Inland Revenue Department was ill-equipped to 1990s, tax incentives offered under the BOI Act were manage the transition because it lacked the required IT extended well beyond the export-oriented sector to facilities and the necessary systems and controls. The offer a wide range of incentives aimed at general GST and the subsequent VAT administration continued investment promotion. These included tax holidays, to play catch-up in subsequent years but were faced concessionary corporate income tax rates, generous with increased challenges owing to frequent changes depreciation allowances, and import duty and excise in tax rates and the growing number of exemptions.b exemptions (figure 3.5.2). Box Figure 3.5.1  Sri Lanka’s Tax Collections Box Figure 3.5.2  Number of Projects with over Time BOI Tax Holidays (percentage of GDP) (number of projects) 20 800 700 15 600 500 10 400 300 200 5 100 0 0 90 80 10 5 00 5 5 0 9 8 19 20 19 19 1 20 01 19 0 00 20 9 20 9 19 8 20 8 19 20 6 20 5 19 4 20 4 19 3 2003 20 19 2 20 2 10 20 7 20 19 6 95 19 7 19 9 9 9 0 9 0 0 0 9 0 9 9 0 0 9 9 20 19 19 Sources: Ministry of Finance and Treasury; Department of Census and Statis- Source: Government of Sri Lanka. tics, Central Bank of Sri Lanka. Note: BOI = Board of Investment. Note: GDP = gross domestic product. (continued) 36 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Box 3.5   Sri Lanka: Tax Exemptions Lead to Tax Base Erosion and Complicate Tax Administration (continued) In addition to incentives provided under the BOI and noncompliance is that the number of individu- Act, Sri Lanka also introduced tax incentives under als registered for income tax has declined to below the Inland Revenue Act. Although originally these 200,000, one-fifth of potential (table 3.5.1).c incentives were focused on areas not covered by the After the end of the war in 2010, the government of BOI Act, there has been significant overlap in recent Sri Lanka began devising a new policy for investment years. Tax incentives under the Inland Revenue Act promotion and broad-based reform of the BOI. It was covered the following areas: the manufacture and deemed that the country no longer requires tax incen- export of goods deemed nontraditional exports; agri- tives as a primary enticement for foreign investment, culture, fisheries, and agroprocessing; industrial tools thus shifting the emphasis away from tax concessions and machinery; small- and large-scale infrastructure toward greater predictability for investors and increas- projects; rehabilitation of underperforming industries; ing the ease of doing business. As the first step toward and companies engaged in research and development. The incentives usually took the form of tax holidays of 2 to 12 years’ duration, depending on the sector, and a period of lower corporate tax rates once the tax Box Figure 3.5.3  Efficiency of Tax Collection holidays expired. Sri Lanka also offers significant accel- in Sri Lanka, 2004–10 erated depreciation allowances, as well as import duty (cost per Re 100 of revenue collected) exemptions on raw materials and capital goods. 0.5 The Result: Weakened Efficiency of Tax Administration 0.4 The frequent changes to tax rates and increased 0.3 exemptions have complicated the tax administration, undermining its efficiency over time—calculated as the ratio of the total recurrent cost of revenue collections 0.2 to taxes collected (figure 3.5.3). Efficiency appears to have improved somewhat in 2010, but it remains to be 0.1 seen whether this improvement will be sustained. The increased complexity of the tax system—an 0.0 onerous number of forms, continuous follow-up by tax 2004 2005 2006 2007 2008 2009 2010 officials, and so forth—has been a principal cause of poor compliance. One indication of possible evasion Source: Sri Lanka Inland Revenue Department. Box Table 3.5.1  Number of Individuals and Resident Companies Registered to Pay Income Tax in Sri Lanka, 2004–10 Number Number of registered taxpayers of tax filers 2004 2005 2006 2007 2008 2009 2010 Resident 26,344 21,644 23,405 24,132 25,573 24,043 29,833 companies Individuals 143,704 120,570 136,783 163,438 200,418 210,399 192,451 Source: Inland Revenue Authority Annual Performance Reports, 2004–10. Creating Fiscal Space through Raising Tax Revenue 37 this goal, the government in mid-2010 instructed the BOI of the tax base would add as much as 1.0–1.5 percentage to refrain from granting further concessions to invest- points of GDP into revenue in the coming years. ment projects and to work with relevant ministries to In addition, following the recommendations of the develop a new investment promotion strategy for the Presidential Tax Commission, the budget for 2011 intro- country. This step has marginally reduced the number of duced major changes to the Sri Lankan tax structure, projects with tax holidays. A sizable portion of existing with the aim of simplifying the tax structure, improving concessions are set to expire over the next two years, compliance, and generating additional revenue. These with nearly all of them expiring by 2015. According to the are positive steps, and their effect on revenues should IMF , the proposed reforms and the ensuing broadening also be measurable in the coming years. a. See also appendix C for Sri Lanka. The system of GST combined with VAT represents the single largest revenue source for Sri Lanka’s government. b. Revenue buoyancy of the VAT system in the country has been poor—continuously below unity—from the system’s inception. c. This number excludes PAYE (pay as you earn) taxpayers, for whom tax is deducted with the payroll. 4. See Gallagher (2005) for a discussion of different References measures used to benchmark tax systems. 5. Tax elasticities, which measure the same ratio while Barbone, Luca, Arindam Das-Gupta, Luc De Wulf, maintaining tax policy and rates unchanged, are a and Anna Hansson. 1999. “Reforming Tax Systems: superior concept for discussing the performance The World Bank Record in the 1990s.� Policy of the tax regime. However, data for deriving tax Research Working Paper 2237, November 2009, elasticities are not readily available for many coun- Washington D.C. tries in South Asia. Cetrángolo, Oscar, and Juan Carlos Gómez Sabaini. 6. Buoyancies are derived in two ways. One is 2010. “Tax Policy in Argentina: Between Solvency regressing log revenue on log GDP over 10 years and Emergency.� In Taxation in Developing using both nominal and real values. The other is Countries: Six Case Studies and Policy Implications, the average of annual buoyancies calculated as ed. Roger H. Gordon, 62–108. New York: percentage revenue increases over GDP growth Columbia University Press. rates. We report mostly regression analyses in real Chai, Joseph C. H., and Kartik C. Roy. 2006. Economic terms, because these two methods yield qualita- Reform in China and India: Development Experience tively similar results. But simple annual averages in a Comparative Perspective. Northampton, MA: are also used in discussion, as needed. Edward Elgar Publishing. 7. Buoyancies shown in appendix B are the average Gallagher, Mark. 2005. “Benchmarking Tax Systems.� of annual figures because some countries had Public Administration and Development 25 (2): limited data. 125–44, doi:10.1002/pad.353. 8. Calibrations have been carried out with other Gordon, Roger H. 2010. “Taxes and Development: bases. These alternative bases include private Experiences of India vs. China, and Lessons for consumption for VAT, nonagricultural GDP for Other Developing Countries.� In Taxation in income tax, and imports for import tariffs. The Developing Countries: Six Case Studies and Policy overall conclusions do not change regardless of Implications, ed. Roger H. Gordon, 37–61. New the choice of bases. York: Columbia University Press. 9. In Nepal, many cases of tax evasion have been International Monetary Fund (IMF). 2011. “Nepal: identified in the recent past. Most notably, evaders 2011 Article IV Consultation.� IMF Country often used fraudulent receipts to claim huge credits. Report 11/318, IMF, Washington, DC. A concern for enforcing compliance is that those Ivanic, Marcos, and Nihal Pitigala. 2011. “South Asia found guilty of tax evasion have not been prosecuted, Economic Integration.� PowerPoint presentation, making the cost of evasion and noncompliance low. World Bank, Washington, DC. 38 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Mortaza, M. Golam, and Lutfunnahar Begum. 2006. Tanzi, Vito, and Howell H. Zee. 2000. “Tax Policy for “Tax Expenditures in Bangladesh: An Introductory Emerging Markets: Developing Countries.� IMF Analysis.� Policy Note PN 0706, Policy Analysis Working Paper WP/00/35, International Monetary Unit, Bangladesh Bank, Dhaka. Fund, Washington, DC. Perry, Guillermo E., Omar S. Arias, J. Humberto López, USAID (U.S. Agency for International Development). William F. Maloney, and Luis Servén. 2006. Poverty 2010. Fiscal Reform and Economic Governance Reduction and Growth: Virtuous and Vicious Circles. Project, Collecting Taxes: 2009/10 data set. http:// Washington, DC: World Bank www.fiscalreform.net/index.php?option=com_cont Piketty, Thomas, and Nancy Qian. 2009. “Income ent&task=view&id=759&Itemid=134. Inequality and Progressive Income Taxation in World Bank. 1991. Lessons of Tax Reform. Washington, China and India, 1986–2015.� American Economic DC: World Bank. Journal: Applied Economics 1 (2): 53–63. ———. 2007. Doing Business 2008. Washington, DC: Purohit, Mahesh. 2006. “Tax Efforts and Taxable World Bank. Capacity of Central and State Governments.� ———. 2008. Doing Business 2009. Washington, DC: Economic and Political Weekly 41(8): 747–51, World Bank. 753–55. ———. 2009. Doing Business 2010: Reforming through Stiglitz, Joseph E. 2010. “Development-Oriented Tax Difficult Times. Washington, DC: World Bank. Policy.� In Taxation in Developing Countries: Six ———. 2010. Doing Business 2011: Making a Case Studies and Policy Implications, ed. Roger H. Difference for Entrepreneurs. Washington, DC: Gordon, 11–36. New York: Columbia University World Bank. Press. ———. 2011. Doing Business 2012: Doing Business Swift, Zhicheng Li. 2006. “Managing the Effects of in a More Transparent World. Washington, DC: Tax Expenditures on National Budgets.� Policy World Bank. Research Working Paper WPS 3927, World Bank, Washington, DC. IV Nontax Revenue: Structure, Status, and Future I ssues related to nontax revenue revenue collections (total revenue was just 11 percent of (NTR) remain a relatively unexplored area in the GDP in 2010), NTR nevertheless reflects an important South Asian context. In marked contrast to the source of revenue, accounting for 20 percent of total wealth of literature and analytical work available on revenue. User fee collection in Afghanistan is partic- tax revenues, analysis and discussions on NTR are ularly high—higher than in any other South Asian sparse. However, in recent times there has been some country—and accounts for nearly two-thirds of NTR renewed interest on the topic, particularly in the wake collections (figure 4.2). of increased recognition that NTR could be a vital tool in expanding fiscal space and could help in the process of macroeconomic stabilization.1 Governments derive Bangladesh NTR from a diverse collection of sources ranging from NTR has increased steadily since the 1990s, growing an interest earnings, property income, user fees or charges, average of 12 percent between fiscal year (FY) 1991 and and sale of property (assets). In many South Asian FY2010. But the share of NTR to GDP has increased countries, profit transfers and dividends from state- only marginally, from 1.8 percent during the 1990s to owned enterprises (SOEs)2 also contribute substantial about 2 percent in the 2000s. Much of this increase revenue, and in Pakistan and Bangladesh, defense- is due to contributions from defense receipts, interest related receipts are also significant. South Asian countries depend to varying degrees on NTR, ranging from a low of about 15 percent of Figure 4.1  Nontax Revenues in South Asia total revenues in countries such as Bangladesh, India, (nontax revenue buoyancy) and Sri Lanka to a high of around 50 percent in coun- tries such as Bhutan and Maldives.3 Similar variation is 2.0 Pakistan evident in NTR as a share of gross domestic product Bhutan (GDP). NTR buoyancy in all countries, barring India 1.6 Afghanistan and Sri Lanka, is greater than unity, with Sri Lanka again Nepal recording the lowest buoyancy (figure 4.1). The discus- Maldives 1.2 Bangladesh sion that follows looks first at the individual structures of NTR in South Asia together with their evolution India 0.8 over time and then probes some analytical aspects— Sri Lanka such as revenue stability, cyclicality, and optimality of various user charges—with a view to suggesting avenues 0.4 for further enhancing NTR. 0.0 0 2 4 6 8 10 12 14 Structure of NTR in South Asian Nontax revenue (% of nominal GDP) Countries Sources: World Bank, CEIC. Afghanistan Note: GDP = gross domestic product. Bouyancy was calculated using log regression from data span- ning a number of years (number of years varied for countries and was based on data availability). In Afghanistan, NTR has accounted for about Nontax revenue as percentage of GDP was based on the latest year of GDP (or estimates) available, 2 percent of GDP. In a country of extremely low overall which, in most cases, was 2010. South Asia Economic Focus – A Review of Economic Developments in South Asian Countries 39 40 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Figure 4.2  Composition of NTR, Afghanistan administration side, agencies responsible for collecting (Afghanis) NTR suffer from significant capacity shortfalls that 20,000 impede revenue collection. Strengthening these capac- ities, together with revising user fees, could further increase the NTR contribution to total revenue. 15,000 10,000 Bhutan The share of NTR in total revenues of Bhutan averaged 46 percent over the past decade. The NTR is mainly 5,000 generated through dividends from state-owned Druk Green Power Corporation and through profit trans- 0 fers from the Tala Hydro Power Corporation (THPC), the Royal Monetary Authority of Bhutan, and the 09 08 10 07 06 04 03 05 20 20 20 20 20 20 20 20 Department of Lottery (figure 4.4). The significant Property income User fees or charges drop in NTR in FY2009/10 was due to a decline in Fines, penalties, and forfeits Other profit transfers. The drop largely reflected the corpo- Source: World Bank. ratization and privatization of THPC, with the entity Note: NTR = nontax revenue. becoming a taxpaying (and dividend-paying) unit and therefore no longer making direct profit contributions. In recent times, interest receipts of the government have Figure 4.3  Composition of NTR, Bangladesh also increased. These are mainly generated through (Crore taka) loans on-lent by the government of India to various 20,000 state enterprises. Interest receipts were a significant 2.5 percent of GDP in FY2009/10 and are projected to remain at similar levels in FY2010/11. 15,000 10,000 India India records substantial NTR collection, averaging 5,000 2.8 percent of GDP and contributing 17 percent of total revenue. Over two decades (1990–2010), NTR revenue 0 has seen an average annual growth of 13.5 percent, faster 0 20 9 1 98 20 0 20 08 20 7 20 4 20 3 20 01 20 2 99 9 20 6 20 5 than most other countries other than Maldives and Sri /1 FY 8/0 FY 6/0 FY 03/0 FY 2/0 FY 00 FY 1/0 19 8/9 FY 5/0 FY 4/0 FY 00/ 7/ FY 07/ 09 /2 0 9 0 0 0 FY 99 0 0 19 Lanka. User charges for fiscal, economic, and social FY FY services continue to remain the primary form of NTR Dividends and profits Interest collection, although the relative contribution of user Administrative charges and user fees Defense receipts charges has fluctuated over time (figure 4.5). Interest Rent income Other receipts were a significant form of NTR collection in Sources: World Bank; CEIC. the mid-1980s, but that contribution has since declined. Note: FY = fiscal year; NTR = nontax revenue. One crore = 10 million. Dividends and profits have also increased following periods of high economic growth and increased corpo- rate profitability, including for public enterprises. earnings, and administrative and user fees (figure 4.3). There is still considerable leeway to increase NTR Maldives collection, particularly through administrative and user charges, which have remained unrevised for many In the case of Maldives, NTR accounted for over years (see Centre for Policy Dialogue 2011). On the 50 percent of total revenue during the past decade Nontax Revenue: Structure, Status, and Future 41 (2001–10), the highest in South Asia. The composition Figure 4.4  Composition of NTR, Bhutan of NTR in Maldives is highly weighted toward island (Ngultrums, million) resort lease receipts and income from state enterprises, 8,000 which together accounted for over two-thirds of total 7,000 NTR and one-third of total revenue (figure 4.6). NTR is likely to remain a significant source of revenue in the 6,000 coming years if one considers the relatively long duration 5,000 of island resort leases (35–50 years, with many granted 4,000 after the mid-1990s), as well as the financial position of state-owned enterprises. However, the government’s 3,000 plans of privatizing some of the SOEs in the coming 2,000 years could see a moderation of such revenues and some 1,000 transformation of NTR into tax revenue.4 0 /11 0 5 3 9 4 7 8 02 6 /1 /0 /0 /0 /0 /0 /0 /0 10 / 09 04 02 08 03 06 07 01 05 20 20 20 20 20 20 20 20 20 20 Nepal FY FY FY FY FY FY FY FY FY FY Since economic reforms commenced in early 1990s, User charges Dividends Profit transfers Nepal has seen progressive improvement in the ratio of Interest Other overall revenue to GDP (from about 10 percent of GDP in the early 1990s to 15 percent of GDP by 2010), with Source: World Bank. Note: FY = fiscal year; NTR = nontax revenue. collections of both tax and NTR rising. NTR collection a. Data for FY2010/11 are provisional. grew by an average of 13 percent from the mid-1990s, slightly lower than the 15 percent growth in tax reve- nues. NTR currently accounts for about 13 percent of total revenue or about 2 percent of GDP (down from about 22 percent of total revenue and 2.5 percent of Figure 4.5  Composition of NTR, India (Consolidated GDP contributions in 2003/04). In Nepal, the two Central and State Governments) principal forms of NTR—user fees and dividends from (Indian rupees, billion) state stakes in commercial enterprises—account for 2,000 over 70 percent of total NTR revenues (figures 4.7 and 4.8). In turn, user fees rely on two fundamental drivers of the economy: tourism and migrant labor. Prior to 1,500 2006, tourism-related charges (for example, visa and mountain trekking fees) were on the decline owing to falling tourist arrivals in view of the Maoist insurgency.5 1,000 However, since 2006, Nepal tourism has been on the rise and with it the tourism-related revenues, which reached 500 0.2 percent of GDP in 2010. Passport fees currently account for 10 percent of all NTR. Considerable labor migration also occurred during the period of the conflict 0 2 5 5 09 9 FY 9/2 /99 FY 200 /04 FY 99 /94 FY 00 /06 FY 00 /08 FY 99 /91 FY200 /01 FY 99 /93 FY 99 /95 FY200 /07 FY Y1 97 7 FY 99 /92 FY 99 /96 19 99 /98 FY 00 00 FY 98 /87 FY 98 /89 FY 00 /03 0 FY 198 /86 FY 198 /88 FY 199 /90 FY 00 /02 FY 200 4/0 and, in particular, after 2001. The associated increase 20 8/0 19 6/9 /1 2 0 9 8 0 3 1 3 2 7 0 1 2 1 4 6 1 1 1 5 1 6 1 8 2 2 FY 985 7 9 2 1 in the issuance of passports thus led to a steep rise in 1 1 FY F passport fees. Labor migration continues to increase User fees Interest income Dividends and profit rapidly in Nepal, leading to a dearth of labor in the country, among other issues. Dividend contributions Profits of RBI Other stemming from state ownership in commercial enter- Sources: World Bank; CEIC. prises contribute about one-third of NTR in Nepal. Note: FY = fiscal year; NTR = nontax revenue; RBI = Reserve Bank of India. Many of these contributions are from Nepal Telecom and the Nepal Rastra Bank, the central bank for Nepal.6 42 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Figure 4.6  Composition of NTR, Maldives Pakistan (Rufiyaa, million) In Pakistan, NTR is a significant form of revenue, 4,000 accounting for 25 percent of total revenue. This represents 3,500 the highest NTR ratio outside of Bhutan and Maldives. In recent years, Pakistan’s central bank—the State Bank 3,000 of Pakistan—has emerged as the largest source of NTR, 2,500 followed by dividend receipts and defense ministry 2,000 income (figure 4.9). The Oil and Gas Development Corporation of Pakistan, an oil mining, refining, and 1,500 retailing company with 75 percent ownership by the state, 1,000 makes the largest contribution to dividends, followed by 500 Government Holdings (pvt) Limited, a fully state-owned holding company for upstream petroleum projects in 0 Pakistan. The defense ministry contribution to NTR 01 10 08 05 02 09 06 03 04 07 largely stems from reimbursements and other payments 20 20 20 20 20 20 20 20 20 20 by the United States, from the Coalition Support Fund, Resort lease rents Income from SOEs Other for expenditure for and services rendered by the Pakistan Sources: Ministry of Finance and Treasury, Maldives; World Bank. military. These receipts were sizable: PR 116 billion in Note: NTR = nontax revenue, SOE = state-owned enterprise. FY2010 (or 0.6 percent of GDP). However, the defense contribution has fallen in FY2011 and is expected to decline further in the coming years. Figure 4.7  Composition of NTR, Nepal (Nepalese rupees, thousands) Sri Lanka 25,000,000 Sri Lanka records the lowest ratios of NTR to total revenue and NTR to GDP in South Asia. NTR buoy- 20,000,000 ancy is also the lowest. Over the past two decades, NTR grew about 14 percent annually, slightly more than the 15,000,000 growth in tax revenues (13 percent). Much of the NTR collection in Sri Lanka has traditionally been related 10,000,000 to property income, which comprises rents earned on government buildings, interest, profits and dividends 5,000,000 from SOEs (and other private enterprises with govern- ment stakes), lottery proceeds (net), and profit trans- 0 fers from the Central Bank of Sri Lanka (figure 4.10). Until recently, user charges have not been a key contrib- 20 3 20 08 20 9 0 20 01 FY 199 8 99 99 20 04 20 0 20 02 20 05 20 06 19 95 19 6 20 07 19 97 FY 2/0 /1 FY 8/0 9 FY 200 FY 5/9 FY 00/ FY 07/ FY 97/ 19 8/ FY 03/ FY 01/ FY 04/ FY 05/ / FY 06/ 09 FY 96/ 94 0 0 utor to NTR, but their contribution has been on the 9 / 19 FY FY increase since 2005. Given the country’s poor tax collec- Duties and fees Penalties, fines, and forfeitures tion record, with tax buoyancy considerably less than Receipts from sales and rent of government property, services, and commodities unity, the government has become increasingly aware Dividends Interest Royalties and sales of government property of the role NTR can play in ramping up overall revenue Donations, gifts, and miscellaneous income in the country. Indeed, in the 2012 budget proposals, Source: World Bank. almost two-thirds of new revenue proposals were NTR. Note: FY = fiscal year; NTR = nontax revenue. Analysis of NTR in South Asia The analysis of NTR includes assessment of NTR and revenue stability, cyclicality, and user charges. Nontax Revenue: Structure, Status, and Future 43 NTR and Revenue Stability Figure 4.8  Main Forms of NTR, Nepal (Nepalese rupees, thousands) Although tax and NTR are both important to govern- ments, an important consideration is the element of 8,000,000 stability that NTR brings to overall revenue structure. 7,000,000 A comparative analysis of tax and nontax revenues in 6,000,000 different countries in South Asia shows that the two 5,000,000 are mostly uncorrelated across different countries, 4,000,000 with the degree of negative linear association being the greatest in countries such as Bhutan and Pakistan 3,000,000 (table 4.1). Moreover, in all countries except Maldives, 2,000,000 the revenue stabilization coefficient shows that NTR 1,000,000 helps to stabilize revenue position.7 0 Table 4.1 shows that the stabilization role of NTR 20 /01 19 /96 0 20 /03 20 /04 99 99 20 /02 2 /05 20 00 20 /08 20 /09 19 /95 1 97 FY 19 /98 FY 006 6 20 /07 /1 2 /0 19 98/ FY 96/ FY 00 FY /20 is strongest in the case of Afghanistan and Nepal. In 09 FY 95 FY 02 FY 03 FY 01 FY 04 FY 07 FY 08 FY 94 FY 997 FY 005 19 the case of Maldives, the high dependence of all forms of revenue (tax and nontax) on tourism is reflected in FY the highest correlation between the tax and nontax Dividends Tourism Fees Passport Fees revenues.8 This dependence has been the principal Source: World Bank. cause for concern in the country, which until recently Note: FY = fiscal year; NTR = nontax revenue. did not have a diversified tax base, and no sales tax or corporate tax regime (barring the commercial banks). However, the tax reforms undertaken by the Maldives government in 2010 saw the introduction of both a goods and services tax and a corporate income tax regime. These initiatives would most likely improve Figure 4.9  Composition of NTR, Pakistan (Pakistani rupee, million) revenue stability indicators. 600,000 500,000 Cyclicality of NTR Relative to Tax Revenues 400,000 A simple correlation-based analysis was used to measure 300,000 the respective cyclicality of tax revenues and NTR. Results 200,000 show that both tax revenues and NTR are highly and positively correlated with GDP for all countries in South 100,000 Asia (table 4.2).9 However, in most countries, barring Pakistan and India, the correlation is less for NTR than 0 FY 8 11 FY 3 10 06 02 09 1 04 FY 7 FY 5 for tax revenues, suggesting that although both forms 0 00 0 0 0 20 20 20 20 20 20 20 20 20 20 2 FY FY FY FY FY FY FY of revenue are procyclical, NTR tends to be less procy- clical than tax revenues. The latter phenomenon is stron- Dividends SBP profits Defense income Interest income gest in the case of Sri Lanka, Afghanistan, and Maldives, Sales proceeds and royalties User fees Other suggesting that at times of economic downturn, there Sources: World Bank; CEIC. could be some element of insulation from NTR. Indeed, Note: FY = fiscal year; NTR = nontax revenue; SBP = State Bank of Pakistan. in the case of Sri Lanka in 2009, when tax revenue growth was a sluggish 6 percent in the face of the slowdown in growth, NTR grew by 16 percent. The growth was largely owing to profit transfers and dividend payments from state-owned banks, which realized windfall gains from sales of government bonds (and from realization 44 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Figure 4.10  Composition of NTR, Sri Lanka social, general, and fiscal) perhaps represent one of the (Lankan rupee, million) more reliable and stable forms of NTR.10 However, 100,000 many governments in South Asia have struggled to establish appropriate user charges commensurate with 80,000 cost recovery and users’ ability to pay (see box 4.1). Attempts to introduce appropriate user charges are 60,000 often dogged by lack of information, political inter- ference, and, in situations in which user fees already exist, the difficulty of raising such fees and charges to 40,000 reflect higher costs.11 A broad evaluation of the sufficiency of the user 20,000 fees can be done by computing the ratio of user fees collected relative to government expenditure on oper- 0 ation and maintenance. Figure 4.11 shows that most 20 2 1 3 2 1 20 8 FY 992 10 1 1 1 8 20 9 FY 00 FY 994 2 3 FY 09 FY 007 FY 995 FY 990 FY 997 FY 004 2 6 FY 05 FY 96 FY 00 FY 00 FY 99 FY 00 FY 99 FY 99 FY 99 FY 0 FY 00 20 countries, barring Afghanistan, have low collection 20 19 1 1 1 2 2 1 1 1 2 FY ratios—in most cases, below 30 percent. Bhutan and Property income Social security contributions Nepal record particularly low rates of collection. Fees and administration charges Other Several studies conducted at a subnational level in India also show user charges remaining well below cost Source: World Bank. Note: FY = fiscal year; NTR = nontax revenue. recovery rates (Ghimire 2005; Purohit and Purohit 2010). The Das-Gupta (2005) analysis of cost recovery patterns in different states in India finds (inconclusively) that cost recovery across various services (economic and social) is positively correlated with income of the Table 4.1  NTR and Tax Revenue Stabilization states, with richer states having better cost recovery Nontax revenue Correlation of change in tax ratios. He also shows that, in general, cost recovery stabilization effect Stabilization coefficient and change in NTR rates in the provision of economic services are much Afghanistan –0.0503 0.2644 higher than cost recovery rates in the provision of Bangladesh –0.0040 –0.0278 social services. A more recent assessment conducted by Bhutan –0.0106 –0.5139 PricewaterhouseCoopers on behalf of the state govern- ment of Madhya Pradesh also comes to the same conclu- India –0.0029 0.5594 sion. In addition, the PricewaterhouseCoopers study Maldives 0.0022 0.7105 also notes different cost recovery rates among states Nepal –0.0220 –0.0705 for economic and social services: states such as Tamil Pakistan –0.0086 –0.2952 Nadu and Andhra Pradesh are leading in cost recovery Sri Lanka –0.0061 –0.0842 in economic sectors, such as forestry and mining, and Uttar Pradesh leads in cost recovery in social sectors Source: World Bank staff calculations. Note: NTR = nontax revenue. such as education (PricewaterhouseCoopers 2010). Options to Increase NTR Collection of capital gains on the bond portfolios) in an environ- in South Asia ment of falling interest rates. As discussed earlier, in many South Asian countries (other than Bhutan and Maldives) considerable scope exists to increase NTR, which could significantly User Charges improve fiscal space. The following are some of the Compared with other forms of NTR, user fees for measures that various governments and subnational various services provided by the government (economic, entities could consider in terms of raising NTR. Nontax Revenue: Structure, Status, and Future 45 Greater use of state-owned land. In many countries in Table 4.2  Cyclicality of NTR South Asia, the government continues to own substan- Coefficient of correlation GDP (excluding tial amounts of land. In Sri Lanka, for example, state Country with GDP GDP agriculture) ownership of land is estimated at over 80 percent. These Afghanistan Total revenue 0.97 0.97 lands could be put into productive use in collabora- NTR 0.74 0.72 tion with the private sector. Improved operational performance of SOEs. Many Tax revenue 0.99 0.99 South Asian countries still have a heavy presence of Bangladesh Total revenue 0.99 1.00 SOEs, whose revenues do not cover costs and are thus NTR 0.95 0.96 reliant on government. Improving the operational Tax revenue 0.99 1.00 performance of these enterprises will aid public finances Bhutan Total revenue 0.97 0.96 in three ways: (a) reduction in the need for budgetary transfers; (b) regular servicing of government loans NTR 0.85 0.93 that realize interest income for the government; and Tax revenue 0.88 0.93 (c) increased profit transfers and dividends. India Total revenue 0.99 0.99 Optimization of user charges. As noted earlier, NTR 0.98 0.98 cost recovery rates of various services provided by Tax revenue 0.98 0.98 the government in different countries across South Asia remain low. One of the fundamental reasons Maldives Total revenue 0.90 0.88 for this low rate is the lack of proper pricing policies NTR 0.81 0.79 for various user charges, commensurate with both Tax revenue 0.95 0.94 the cost of providing such services and the users’ Nepal Total revenue 0.96 0.97 ability to pay. In the technical literature, three types NTR 0.91 0.95 of pricing policies are often advocated for govern- ment services: (a) total absorption costing,12 which Tax revenue 0.94 0.95 attempts to pitch user charges with total costs associ- Pakistan Total revenue 0.99 0.99 ated with the provision of respective services (that is, NTR 0.96 0.97 fixed plus variable); (b) marginal cost pricing, which Tax revenue 0.96 0.95 essentially attempts to pitch prices closer to the incre- Sri Lanka Total revenue 0.98 0.97 mental costs; and (c) activity-based costing, which is somewhat of a modern innovation that attempts to NTR 0.69 0.68 apportion costs to various production processes (or Tax revenue 0.97 0.96 activities) that underlie the output in order to arrive at Source: World Bank staff calculations. the costing. Of these various policies, total absorption Note: GDP = gross domestic product; NTR = nontax revenue. The cyclicality was calculated over the costing is not often used (except as an academic exer- years for which data were available. cise to ascertain the total cost associated with providing a service), whereas marginal cost pricing is often the method of choice for pricing policies. However, in South Asia, most user charges are set arbitrarily and an assessment of measures for expanding nontax with only infrequent revisions to keep up with cost revenue in 2005. In addition, several state govern- (Ghimire 2005). As such, proper costing policies for ments in India have undertaken several assess- user charges, particularly for economic services, could ments of NTR (PricewaterhouseCoopers 2010; considerably improve NTR receipts. Government of Karnataka 2004) 2. Central banks of respective countries also contribute profit transfers. 3. Nontax revenues are taken as a period average Notes 2001–10. 1. For example, the Advisory Committee for Macro­ 4. Transformation could occur when the entities economic Management in Nepal, in collaboration go off the government’s books and start paying with the Asian Development Bank, undertook corporate income taxes. 46 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Box 4.1   User Charges: Efficiency and Equity Properly designed user charges can boost revenue differences, and the resulting cost to the budget has without causing efficiency losses or creating distor- exceeded 2 percent of GDP . In Bangladesh, SOE tions. Efficient user charges are set equal to the losses that arose from adjusting administered prices cost of the public sector providing an additional unit too slowly, thus not covering input cost, reached of goods or services. Developing countries (includ- 3 percent of GDP , and the budget paid 60 percent ing those in South Asia) often do not charge these of those losses. These SOEs are in the areas of cost-recovery prices in services such as power. They petroleum, fertilizer, power, and airlines. In particular, also often subsidize fertilizers and fuels. Demand for slow-moving fuel and fertilizer price adjustment in the such goods would be reduced if cost recovery prices face of rapidly rising input costs are a challenge to were charged. The resulting lower cost creates excess fiscal sustainability. demand, and the benefits of the subsidy fall dispropor- Nepal faces a very similar situation in terms of slow tionately on the rich. If user charges are appropriately adjustment of oil prices and power tariffs. The loss designed, both efficiency and equity objectives can be to the budget for each could be as high as 2 percent served and more revenue obtained. of GDP . (These SOEs’ losses are first funded through For example, in Pakistan, electricity is provided by loans, but ultimately, they are shouldered by the eight distribution companies, and the uniform power budget.) The chronicled and large losses of Sri Lanka’s tariff is determined on the base of the most efficient Ceylon Electricity Board are similarly funded by public of these companies. The government picks up the sector banks and pose a threat to the budget. Figure 4.11  Ratio of User Charges Collected Relative to 5. The Maoist insurrection began in 1996, when Government Operation and Maintenance Expenditure in the Communist Party of Nepal, Maoist, began South Asia, 2005–10a an armed struggle. The insurrection ended with (percent) the Comprehensive Peace Accord signed in November 2006. Afghanistan 6. Nepal Telecom is the leading and profitable Bangladesh telecommunication operator with a 90 percent Bhutan government stake. 7. Based on methodology proposed by Haughton India (1998), the revenue stabilization coefficient shows Pakistan how the overall coefficient of variation for total Nepal revenue is affected by the inclusion of NTR. A Sri Lanka negative value indicates that NTR contributes to Maldives stabilizing the revenue position. 8. See also the Maldives country analysis in appendix C. 0 10 20 30 40 50 60 9. Running the correlation estimates with GDP Sources: World Bank, http://data.worldbank.org; CEIC. (excluding the agriculture sector in each country) a. In some countries, a shorter time span was used owing to lack of data. yielded very similar results. 10. This is in comparison to sources such as interest income, profits, and dividends, which are highly variable and subject to economic cycles. 11. Recourse to cost recovery could well lead to a distortionary impact on the tax system, whereas Nontax Revenue: Structure, Status, and Future 47 recourse to users’ ability to pay could affect debt Haughton, Jonathan. 1998. “Estimating Tax Buoyancy, sustainability. Elasticity and Stability, EAGER/PSGE–Excise 12. Total absorption costing is also referred to as full Project, Methodological Note 1, USAID, April absorption costing in some literature. 1998. Government of Karnataka. 2004. “Revenue Reforms Commission, Final Report.� Revenue Reforms References Commission, Government of Karnataka, Bangalore. http://www.kar.nic.in/finance/others/final-rrc.pdf. Centre for Policy Dialogue. 2011. “State of the Purohit, Mahesh, and Vishnu Kanta Purohit. 2010. Bangladesh Economy in FY2010–11.� CPD, Non-tax Sources in India: Issues in Pricing and Dhaka. http://www.cpd.org.bd/downloads/ Delivery of Services. New Delhi: Gayatri Publications. IRBD%20FY11_First%20Reading.pdf. PricewaterhouseCoopers. 2010. “Financing Das-Gupta, Arindam. 2005. “Non-Tax Revenues in Growth: Study on Avenues for Enhancing Indian States: Principles and Case Studies.� Internal Non Tax Revenue—Strengthening Performance report submitted to Asian Development Bank, Management in Government, Government of Manila. http://www.pdfport.com/view/744148- Madhya Pradesh.� PricewaterhouseCoopers, non-tax-revenues-in-indian-states-principles- New York. http://www.dif.mp.gov.in/spmg/ and-case-studies.html. AugmentingNonTaxRevenueReport.pdf. Ghimire, Parasad Madhab. 2005. “Measures for Expanding Non-tax Revenue.� Policy Paper 5, Economic Policy Network, Kathmandu, Nepal. V Summary and Conclusions M ost countries in South Asia important efforts are described in box 3.2 in chapter need to create additional fiscal space in a fiscally 3 in this volume. In Sri Lanka, reform plans include sustainable way. The region urgently needs simplifying the import tariff through the elimination to build physical infrastructure and enhance human of surcharges and extra tariff duties, rationalizing VAT capital because it suffers from a high incidence of rates, and broadening the VAT and income tax base. poverty and slow progress in achieving the Millennium State enterprise reform will also be promoted to intro- Development Goals (MDGs). Yet government spending duce more commercial management. The government and, in particular, social sector and infrastructure of Bangladesh is contemplating fundamental changes spending, are much lower than in other countries at in income tax and VAT policy, with technical assistance comparable income levels, owing in large part to the from the World Bank and IMF. Pakistan has sched- countries’ limited revenue mobilization. Given that uled the introduction of a broad-based VAT, which outstanding public debt is already high in many coun- envisages a single rate with minimum zero rating. tries in the region, debt sustainability studies indicate Simplifying and rationalizing PIT and CIT struc- that there is limited scope for future borrowing in the tures are also a high priority that should be pursued as region. At the same time, donor grants are unlikely to soon as possible. Developing nonhydropower-related expand. The only options for the regional countries revenue is important in Bhutan, and the government would be increasing revenue collection and prioritizing is considering introduction of a VAT. and enhancing efficiency of spending. Increased transparency and accountability and lower The tax regimes of many regional countries have inflation would boost revenue collection. The regression been reformed to conform more closely with the best analysis conducted for this book shows that control- practices framework recommended by the World ling corruption by increasing transparency and account- Bank and the International Monetary Fund (IMF). ability is associated with significant revenue increases. Specifically, many countries have adopted or are in the The same is true when the country has macroeconomic process of adopting a value added tax (VAT), with a discipline and inflation is reasonably low. These are limited number of nonzero rates. Income tax rates are actions under the control of the government, and South not prohibitive and are relatively simple (except for Asian countries that need added fiscal space could seri- those of Pakistan and Sri Lanka). Top rates for corpo- ously consider further measures to address governance rate income tax (CIT) and personal income tax (PIT) issues and improve macroeconomic stability. are generally close to each other, limiting prospects for Reducing and rationalizing tax exemptions and shifting income to lower tax brackets. Tariff rates are other tax preferences would likewise contribute to low on average, and export taxation is limited. higher revenues, with the same or even lower tax rates. Yet total tax collection is low, income tax collection All countries in the region should review their tax is limited, and trade tax still occupies a relatively high expenditures. In particular, in Sri Lanka, where revenue share of the total revenue. This pattern is typical of buoyancy has been the weakest, a critical task will be to lower-income countries where administration capacity review and significantly reduce its Board of Investment is limited. The low collection rate is due to poor yields and others’ tax concession scheme (including ad hoc of key instruments such as VAT and income taxes. tax holidays on VAT and CITs, and exemptions on Yields are low largely because of limited tax adminis- trade and excise taxes). Similarly, in Bangladesh and tration capacity and eroded tax bases caused by signif- Nepal, a necessary condition for increasing revenues icant exemptions and tax preferences. and fiscal sustainability is the reduction of exemptions Many countries already have plans to improve and incentives. Tax expenditures are estimated by the their tax systems, and it is important that these plans Bangladeshi authorities as 5 percent of gross domestic continue to be implemented and deepened. India’s product (GDP), and in Nepal, the IMF estimates tax South Asia Economic Focus – A Review of Economic Developments in South Asian Countries 49 50 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries expenditures as 2.2 percent of GDP. Replacing tax Regarding nontax revenue, key issues in many coun- holidays and exemptions with tax credits for invest- tries are poor performance of state-owned enterprises ment (if needs are confirmed) and allowing acceler- (SOEs) and user charges that do not cover the cost of ated depreciation are viable choices for the countries. provision. Many regional countries have numerous Tax administration should be strengthened in many SOEs that suffer from financial losses and must depend countries. Improving administration is essential: the on government support, instead of providing resources authorities need to make the rates simple, compliance to the government in the form of profit or dividends. easy, and noncompliance costly. Many actions can be Also, many of these SOEs provide goods and services to considered. For example, the internal organization of the public without charging the full cost of provision. Bangladesh’s tax authority should be rationalized and Most notably, cost recovery by enterprises providing a comprehensive, institutionwide reform launched. fertilizers, power, and petroleum products is often In Nepal, both registrants and tax filers need to be weak and results in significant losses to the national increased, tax audits need to be made more efficient, and budget. These problems need to be tackled. Making tax evaders who have been recently identified need to be SOE management commercial is one step in the proper prosecuted. In Pakistan, recent reforms in information direction. In any event, the countries should allow full technology and other systems should be translated into cost recovery for provision of these goods and services— higher revenue by strengthening enforcement—some- subsidizing these prices almost always benefits the rich thing that requires a change in the culture of tax offi- much more than the poor. cials. Finally, the government of India’s current efforts, In some countries, revenue sources must be diver- which require any individual who owns a house, a tele- sified. For example, in Bhutan and Maldives, revenues phone, a credit card, a car, or a club membership or has depend on a single sector (hydropower and tourism, expenditures on foreign travel to file personal income respectively). In such cases, efforts should be intensi- tax, would be a viable option for all regional countries fied to develop broad-based domestic commodity taxa- to consider for future tax reforms. tion and income tax bases outside the particular sector. Appendix A Empirical Estimation of Revenue Gap T his appendix documents our the lagged total public expenditure as a share of GDP econometric analysis of the determinants of and the lagged literacy rates as proxies for the delivery revenue collection using cross-country panel of public services. In addition, we also attempt to data. We estimate the effect of each determinant on identify the effect of public social services on revenue general government revenue as a share of gross domestic collection using the instrumental variable approach product (GDP) and compare the revenue mobiliza- (tables A.5 and A.6). A description of how we define tion performance in South Asian countries to that in and construct the explanatory variables and our iden- other countries. tification strategy follows. Methodology Variables Our objective is to assess the country characteristics that may affect revenue collection and obtain an esti- Income Per Capita mate of the difference between actual and theoret- There is likely an ambiguous relationship between the ical revenue share of GDP as predicted by the model. level of economic development and the revenue ratio Following the previous work by Hattari et al. (2008) of GDP. On one hand, low-income countries may on empirical determinants of government revenues, we have weaker governance quality and therefore lower tax estimate the following base model: enforcement. Thus, government revenue ratios may be lower for developing countries as opposed to advanced Ri,t = α + β0yi,t + β1agrii,t + β2servi,t + β3openi,t + β4depi,t economies. On the other hand, the size of the public + β5infi,t + β6urbi,t + β7pub_servicei,t–1 + β8resi + β9c_ sector may gradually decrease as an economy becomes corrupi,t + β10b_qualityi,t + β11India + β12Pakistan + more advanced. In addition, previous studies found β13Bangladesh + β14Nepal + β15Bhutan + β16 SriLanka that domestic and consumption tax collection may + β17Maldives + εi,t go up as income increases, whereas trade tax revenue will likely decrease (Hattari et al. 2008). The idea is where R represents the general government revenue that low-income countries tend to be more depen- share of GDP, i represents country i, and t represents dent on trade tax, which is easier to enforce; however, year t. We consider the following factors that affect together with economic development, the tax system revenue mobilization: income; economic structure becomes more mature, and the government will even- variables such as share of agriculture, services, and tually become less dependent on trade tax as an effec- trade; age dependency ratio; inflation; urbanization tive income tax system develops. rate; natural resource wealth; provision of public social services; control of corruption; and quality of bureau- cracy (tables A.1 and A.2). Economic Structure Variables: Because the provision of public services may be Agriculture, Service, and Trade affected by current revenue levels, we exclude this variable in our main specification (table A.3) to avoid Shares of GDP reverse causality issues. However, in an alternative The relative importance of various structural compo- model (table A.4), we include lagged public expen- nents of a given economy is expected to affect revenue diture on education and health care as well as lagged collection. For example, economies with a large agri- literacy rates as a proxy for the provision of public cultural sector tend to have lower levels of government social services. In another specification, we use both revenues. The extent to which the economy relies on the South Asia Economic Focus – A Review of Economic Developments in South Asian Countries 51 52 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Table A.1  Descriptive Statistics Standard Variable Aspect Mean Deviation Minimum Maximum Observations General govern- Overall 30.40778 13.29829 0 104.985 N = 2,576 ment revenue (% Between 11.83621 6.169 78.51609 n = 152 of GDP) Within 5.050031 –2.410313 73.61365 T-bar = 16.9474 Log of GDP per Overall 7.766549 1.68261 4.627772 10.9442 N = 2,511 capita Between 1.607975 4.75731 10.74395 n = 150 Within 0.1932531 6.687083 9.499189 T-bar = 16.74 Log of inflation Overall 1.637022 1.149965 –4.017384 9.371577 N = 2,430 rate Between 0.7950967 0.1489474 3.846098 n = 150 Within 0.8359273 –2.639228 8.200928 T-bar = 16.2 Agriculture share Overall 16.07911 14.609 0.0686389 75.52297 N = 2,418 (% of GDP) Between 14.42469 0.1503392 66.63165 n = 147 Within 4.334289 –14.74362 52.1281 T-bar = 16.449 Services share (% Overall 55.13941 13.86794 2.317229 85.07104 N = 2,414 of GDP) Between 13.6697 14.97186 81.01225 n = 147 Within 4.36611 35.83666 88.96269 T-bar = 16.4218 Trade share (% of Overall 85.06636 47.21732 0.3088029 438.0917 N = 2,497 GDP) Between 48.10277 0.8397979 406.4641 n = 151 Within 15.5512 –15.38271 221.7431 T-bar = 16.5364 Age dependency Overall 11.47004 7.183365 1.261699 32.85413 N = 2,443 ratio (%) Between 6.82474 1.519147 26.99179 n = 145 Within 1.1276 3.665023 23.08289 T-bar = 16.8483 Urbanization rate Overall 52.40846 24.31855 6.56 100.0 N = 2,569 (%) Between 23.58173 8.377647 100.0 n = 151 Within 2.522055 40.52639 65.84703 T-bar = 17.0132 Literacy rate (% of Overall 78.51824 21.50302 9.391331 99.8 N = 285 population age 15 Between 20.55526 19.03188 99.7836 n = 118 and older) Within 3.611468 63.62789 91.76023 T-bar = 2.41525 Sources: World Development Indicators, 2010 (database), World Bank; Government Finance Statistics, 2008 (database), International Monetary Fund. Note: GDP = gross domestic product; N = total number of observations; n = number of countries. service sector also matters. In developed countries where international trade is another important determinant of both service sectors and tax administration systems are the revenue ratio. A more open economy with a higher sophisticated, the revenue collection levels may be high. trade share of GDP is expected to enjoy the less costly By contrast, in many developing countries where a large enforcement of tariffs and therefore higher revenue share of the economy is based on small or informal collection. In our analysis, we control for the impor- service sector businesses, tax collections are difficult to tance of the agriculture and service sectors and for the enforce. The degree to which the economy is open to openness of the economy measured by share of GDP. Table A.2  Coefficients of Correlation Lag of General literacy Lag of govern­ rate (% of Lag of educa­ Lag of ment Agri­ Age popula­ public tion health revenue Log of culture Services Trade depend­ Urbani­ tion age expend­ expend­ expend­ Control of Quality of (% of GDP per share (% share (% share (% ency Log of zation Resource 15 and iture (% iture (% iture (% corrup­ bureau­ GDP) capita of GDP) of GDP) of GDP) ratio (%) inflation rate (%) dummy older) of GDP) of GDP) of GDP) tion cracy General 1 govern­ment revenue (% of GDP) Log of GDP per 0.5702 1 capita Agri­culture –0.5084 –0.8382 1 share (% of GDP) Services share 0.3304 0.5965 –0.6621 1 (% of GDP) Trade share (% 0.1938 0.1941 –0.2709 0.1240 1 of GDP) Age depend­ 0.5825 0.7219 –0.5800 0.5991 0.0340 1 ency ratio (%) Log of inflation –0.2840 –0.4070 0.3639 –0.3423 –0.1475 –0.2645 1 Urbani­zation 0.4821 0.7853 –0.7066 0.4576 0.0950 0.5598 –0.2831 1 rate (%) Resource 0.1061 0.0475 –0.0570 –0.4154 0.0484 –0.2687 0.0575 0.1513 1 dummy Lag of literacy 0.2688 0.5190 –0.4683 0.3301 0.2293 0.3923 0.0931 0.4313 0.0928 1 rate (% of popula­tion age 15 and older) Lag of public 0.7974 0.4776 –0.4043 0.3622 0.1241 0.5744 –0.2015 0.384 –0.0095 0.3732 1 expend­iture (% of GDP) Lag of educa- 0.5062 0.1826 –0.1604 0.3081 0.1977 0.1761 –0.1806 0.1057 –0.1613 0.1959 0.5498 1 tion expend­ iture (% of GDP) Lag of health 0.6292 0.5382 –0.4966 0.5428 0.0429 0.6955 –0.2631 0.4817 –0.2038 0.3818 0.6698 0.5459 1 expend­iture (% of GDP) Control of 0.5327 0.6309 –0.4414 0.4921 0.0446 0.5934 –0.2381 0.4303 –0.2224 0.2299 0.5203 0.3383 0.6023 1 corrup­tion Quality of 0.5850 0.8056 –0.6522 0.6140 0.1627 0.6921 –0.3819 0.5482 –0.1645 0.4269 0.5521 0.3835 0.6569 0.7177 1 Appendix A. Empirical Estimation of Revenue Gap bureau­cracy 53 Source: See table A.1. Note: GDP = gross domestic product. 54 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Table A.3  Main Results Variables (1) (2) (3) (4) Log of GDP per capita 1.254*** 1.039** 1.756* 3.071** (0.315) (0.440) (0.962) (1.404) Agriculture share (% of GDP) −0.192*** −0.245*** −0.139 −0.197* (0.0351) (0.0347) (0.104) (0.101) Service share (% of GDP) −0.115*** −0.222*** −0.0293 −0.233* (0.0292) (0.0389) (0.0925) (0.123) Trade share (% of GDP) 0.0168*** 0.0159*** 0.0139 0.0080 (0.00494) (0.00563) (0.0130) (0.0153) Age dependency ratio (%) 0.858*** 0.839*** 0.543*** 0.473*** (0.0421) (0.0478) (0.122) (0.143) Log of inflation −0.497** −0.642*** −0.185 −1.047 (0.199) (0.224) (0.616) (0.703) Urban population (% of total) −0.0308** −0.0151 −0.0246 −0.0455 (0.0135) (0.0148) (0.0365) (0.0527) Resource dummy 4.757*** 5.920*** 4.808** 2.991 (1 if rent > 10% of GDP) (0.681) (0.801) (1.871) (2.297) India −4.206*** −2.435*** −2.703 −1.718 (0.507) (0.801) (1.663) (2.476) Pakistan −6.696*** −4.002*** −5.974*** −5.495*** (0.465) (0.600) (1.618) (1.939) Bangladesh −10.010*** −5.919*** −9.491*** −8.310*** (0.608) (0.681) (2.316) (2.069) Nepal −6.511*** −6.380*** (0.589) (1.181) Bhutan 16.31*** 20.46*** (1.577) (7.444) Sri Lanka −8.076*** −6.970*** −5.690* −4.853 (0.937) (0.929) (3.209) (3.286) Maldives 3.005* 3.689 (1.667) (4.033) Control of corruption 1.756*** 0.733 (0.207) (0.687) Quality of bureaucracy −0.212 −0.697 (0.374) (1.192) Lag of literacy rate (% of population −0.0240 −0.0373 age 15 and older) (0.0346) (0.0396) Constant 19.91*** 22.12*** 13.33 20.06** (3.378) (3.355) (10.140) (9.490) Number of observations 2,137 1,510 233 160 R−squared 0.572 0.693 0.418 0.463 Source: Authors. Note: Robust standard errors are in parentheses. GDP = gross domestic product. *** p < 0.01, ** p < 0.05, * p < 0.1 Appendix A. Empirical Estimation of Revenue Gap 55 Table A.4  Alternative Model Variables (1) (2) (3) (4) Log of GDP per capita −0.503 0.855 1.915 3.773** (0.430) (0.734) (1.524) (1.834) Agriculture share (% of GDP) −0.292*** −0.229*** −0.105 0.0155 (0.0579) (0.0738) (0.157) (0.147) Service share (% of GDP) −0.125*** −0.154** 0.00790 −0.134 (0.0430) (0.0608) (0.119) (0.165) Trade share (% of GDP) −0.00258 0.00532 0.0136 0.0115 (0.00698) (0.00856) (0.0236) (0.0322) Age dependency ratio (%) 0.668*** 0.704*** 0.561*** 0.463** (0.0520) (0.0613) (0.175) (0.210) Log of inflation −0.0666 −0.301 −0.371 −1.924* (0.359) (0.423) (1.129) (1.017) Urban population (% of total) −0.0388** −0.0436* −0.0218 −0.0479 (0.0192) (0.0226) (0.0615) (0.0733) Resource dummy (1 if rent > 10% 6.992*** 6.837*** 3.952 3.897 of GDP) (1.180) (1.359) (2.477) (2.740) Lag of education spending 1.678*** 1.646*** 0.884 −0.0733 (0.220) (0.280) (0.592) (0.659) Lag of health spending 1.788*** 1.496*** 0.842 1.855** (0.253) (0.279) (0.772) (0.764) Control of corruption 1.260*** −0.0989 (0.373) (1.039) Quality of bureaucracy −1.574** −0.148 (0.771) (1.601) Lag of literacy rate (% of population −0.130* −0.137* age 15 and older) (0.0747) (0.0698) India −1.731* 1.521 −0.738 1.006 (1.023) (1.623) (2.699) (3.346) Pakistan −0.977 1.327 −6.386** −5.688* (0.720) (1.047) (2.870) (3.248) Bangladesh −6.047*** −2.786** −9.399*** −9.616** (0.815) (1.256) (3.320) (3.763) Nepal −5.489** −7.632*** (2.143) (2.064) Bhutan 6.195*** 5.080 (1.526) (3.544) Sri Lanka −3.850*** −4.869*** (0.746) (1.166) Maldives 1.445 7.291 (2.162) (4.943) Constant 23.76*** 14.05** 11.61 10.58 (4.850) (5.937) (16.86) (14.69) Number of observations 826 622 121 88 R-squared 0.659 0.713 0.420 0.501 Source: Authors. Note: Robust standard errors are in parentheses. GDP = gross domestic product. *** p < 0.01, ** p < 0.05, * p < 0.1 56 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Table A.5  Instrumental Variable First-Stage Results Variables (1) (2) (3) (4) Legal_United Kingdom 1.590* −2.206** −3.048 (0.942) (1.038) (2.911) Legal_France 1.086 −1.874** −3.602** −6.510** (0.860) (0.939) (1.536) (2.869) Legal_Germany 3.240** −3.139** (1.364) (1.446) Legal_Scandinavia 10.67*** 3.997*** (1.227) (1.288) Legal_Socialist −0.806 (2.530) Log of GDP per capita −1.327*** −0.133 −2.018 1.571 (0.505) (0.501) (1.450) (1.513) Agriculture value added share (% −0.176*** −0.180*** −0.394*** −0.227* of GDP) (0.0471) (0.0457) (0.119) (0.122) Service value added share (% of 0.0817*** −0.0457 0.0797 0.0295 GDP) (0.0265) (0.0313) (0.0648) (0.0745) Trade share (% of GDP) 0.00865 −0.000484 −0.0155 −0.0192 (0.00588) (0.00619) (0.0177) (0.0212) Age dependency ratio (%) 0.888*** 0.773*** 0.639*** 0.502*** (0.0600) (0.0613) (0.158) (0.170) Log of inflation 0.535** 0.0501 0.444 0.260 (0.223) (0.231) (0.898) (1.007) Urban population (% of total) 0.0194 −0.00303 −0.0153 −0.0258 (0.0159) (0.0177) (0.0422) (0.0503) Resource indicator (1 if resource 2.159** 1.976** 3.061 2.555 rents ≥ 10% of GDP) (0.843) (0.850) (2.028) (1.984) India −0.270 1.261 −3.473 1.064 (1.205) (1.256) (2.261) (2.933) Pakistan −6.895*** −4.006*** −11.27*** −8.825*** (0.919) (0.907) (2.086) (2.084) Bangladesh −11.29*** −7.607*** −15.29*** −13.33*** (0.859) (0.804) (2.283) (2.216) Nepal −8.242*** −10.12*** (1.753) (1.943) Bhutan 17.32*** 6.531*** (1.672) (1.547) Sri Lanka −2.680** −1.147 −7.980*** −5.214* (1.362) (1.332) (2.814) (2.670) Maldives 3.459* 8.983 (2.072) (6.167) Control of corruption 1.421*** 0.321 (0.230) (0.769) Quality of bureaucracy 0.390 −1.479 (0.382) (1.082) Literacy rate (% of population age −0.0341 −0.0490 15 or older) (0.0481) (0.0513) Constant 25.21*** 24.35*** 44.60*** 25.95** (3.984) (3.715) (12.24) (11.04) Number of observations 1,953 1,471 222 178 R-squared 0.509 0.651 0.408 0.479 Source: Authors. Note: Robust standard errors are in parentheses. GDP = gross domestic product. *** p < 0.01, ** p < 0.05, * p < 0.1 Appendix A. Empirical Estimation of Revenue Gap 57 Table A.6  Instrumental Variable Results Variables (1) (2) (3) (4) Lag of government expenditure 1.396*** 1.549*** 1.369*** 1.092*** share of GDP (0.105) (0.151) (0.440) (0.286) Log of GDP per capita 0.459 1.450*** 2.361 2.836*** (0.377) (0.533) (1.561) (0.998) Agriculture value added share (% −0.103** 0.0103 0.133 −0.0483 of GDP) (0.0418) (0.0589) (0.205) (0.106) Service value added share (% of −0.216*** −0.132*** −0.214*** −0.185*** GDP) (0.0272) (0.0392) (0.0804) (0.0654) Trade share (% of GDP) 0.0107** 0.0123** 0.0238** 0.0283*** (0.00442) (0.00545) (0.0120) (0.00995) Age dependency ratio (%) −0.342*** −0.503*** −0.353 −0.129 (0.0994) (0.139) (0.311) (0.230) Log of inflation −1.087*** −0.835*** −0.971 −0.680 (0.253) (0.295) (0.772) (0.595) Urban population (% of total) −0.00634 −0.0121 0.0506 0.0147 (0.0152) (0.0201) (0.0412) (0.0378) Resource dummy (resource rents ≥ 1.467* 1.170 0.786 0.981 10% of GDP) (0.884) (1.050) (2.108) (1.443) Control of corruption −1.003*** −0.548 (0.354) (0.575) Quality of bureaucracy −1.007** −2.721*** (0.445) (1.022) India −4.691*** −4.119** −3.726* −0.551 (1.429) (1.729) (2.127) (1.954) Pakistan 1.978* 1.607 2.572 0.221 (1.126) (1.190) (4.316) (2.053) Bangladesh 4.590*** 4.944*** 5.534 1.815 (1.414) (1.365) (6.371) (3.392) Nepal 5.105* 0 5.690 0 (2.640) (0) (4.431) (0) Bhutan −6.426** 0 −6.078 0 (2.898) (0) (4.073) (0) Sri Lanka −4.292*** −4.463*** −0.642 −2.316 (1.133) (1.333) (4.026) (2.481) Maldives −1.773 0 −7.659 0 (2.168) (0) (5.965) (0) Literacy rate (% of population age 15 −0.0149 −0.0632 and older) (0.0445) (0.0414) Constant 0.482 −10.69** −17.64 −2.846 (3.695) (5.428) (21.75) (11.17) Number of observations 1,953 1,471 222 178 R-squared 0.529 0.601 0.599 0.788 Overidentification test statistics Chi-sq(3) = 5.57 Chi-sq(3) = 5.22 Chi-sq(2) = 0.0002 Chi-sq(2) = 0.002 p = 0.13 p = 0.16 p = 0.99 p = 0.99 Source: Authors. Note: Robust standard errors are in parentheses. GDP = gross domestic product. *** p < 0.01, ** p < 0.05, * p < 0.1 58 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Age Dependency Ratio public services. If we assume that taxpayers guide their choice using information on the outcomes of public We also account for the age dependency ratio, defined services, a better human capital development indicator as the ratio between the size of the working population would suggest a higher willingness to pay taxes. In our in the economy and the number of dependents. An revenue regression, we include the lagged literacy rate aging society may suffer from a diminishing produc- as an explanatory variable. tive labor force and thus a shrinking income tax base, although the effect of aging on other types of taxes, such as consumption tax, is not clearly predicted by theory. Holding other determinants constant, we Natural Resource Wealth expect a higher age dependency ratio to have a nega- Following previous studies (IMF 2011), we construct tive effect on income tax revenue. an indicator for those countries that are rich in natural resources. It may be less costly for the government to collect revenues from the natural resource sectors if Urbanization Rate the country enjoys advantageous natural endowments. Defined as the urban population share of the total population, urbanization is likely to have a positive effect on revenue mobilization because it is positively Control of Corruption and the associated with development of the formal sector of Quality of Bureaucracy an economy. Therefore, urbanization helps reduce the The quality of governance matters for the effectiveness cost of tax enforcement. Another reason that urban- of the public sector and thus development outcomes. ization may work in favor of revenue collection relates Recent studies have shown that public spending and to the mobilization of nontax revenues associated with investments in weak institutional and governance envi- royalty and land rent revenues. However, in the mean- ronments often fail to translate into effective public time, reforms in tax policies may not be able to adapt services. Strong evidence of better human develop- to the rapid development in order to capture the bene- ment outcomes as a result of public sector spending is fits of urbanization. We expect urbanization to have found in better-governed countries (for example, Gupta, an ambiguous effect on revenue collection. Davoodi, and Tiongson 2002; Rajkumar and Swaroop 2008). Previous studies also reveal that corruption could distort the composition of government expenditures, Public Social Spending biasing composition toward public projects that are easier to extract rents from and, ultimately, undermining The willingness to pay taxes may depend on the govern- the effectiveness of public social spending (for example, ment’s investment in the social sector. Taxpayers could Mauro 1998; Tanzi and Davoodi 1997). Therefore, a observe the provision of public services and make an country with lower quality of governance is expected to economic decision as to whether to comply with the suffer from lower willingness to pay taxes owing to its tax laws. In our main analysis, we use the lagged public ineffective provision of public services. Another reason expenditure on education and health care as a share of that governance matters is that the transparency and GDP as a proxy for public social spending. We also accountability of the institutional environment are crit- estimate alternative models using lagged total public ical for the enforcement of tax collection. expenditure as an indicator of the public investment in the social sector. Identification Strategy Literacy Rate Although we used the lags of public expenditure vari- Along similar lines of reasoning, the level of human ables as proxies for the provision of public social services, capital development such as the literacy rate could be our estimates could still be biased if the lagged expendi- positively associated with the level of revenue collec- ture variables are correlated with the error term. Indeed, tion because it could be a proxy for effective delivery of the previous public spending may be affected by the Appendix A. Empirical Estimation of Revenue Gap 59 government capacity to raise revenues, and therefore 2009). Comprehensive economic, social, and demo- influenced by the current revenue collection if the ability graphic information is available for 152 countries, with to mobilize revenues is persistent over time. To address annual observations for the period from 1980 to 2008. the possible endogeneity of the expenditure variables, Descriptive statistics appear in table A.1. we propose using the historical legal origins as instru- Income per capita and inflation rates are measured ments for public expenditure in the social sector. Our using natural logs. The economic structure variables are assumption is that historical origin of the legal system measured as a share of GDP in percentage terms. The may be a good proxy for the government tendency to age dependency and urbanization ratios are measured play a significant role in the economy and thus invest in a standard way, the former being the ratio between more in public resources. We expected that these legal old (ages 64 and older) and working population (ages origin variables satisfy the exclusion constraint because 15–64) and the latter being the ratio between the urban legal origin may not directly affect revenue collection population and the total. The indicator for natural through channels other than the provision of public resources is defined to equal 1 if the rents collected services. Overidentification test statistics also suggest from the natural resource sectors are greater than that the proposed instruments are plausible.1 A two- 10 percent of GDP, and 0 otherwise. stage-least-squares (2SLS) method is used to estimate It is worth noting that our governance variables from our model with lagged expenditure variables as deter- the International Country Risk Guide (ICRG) are measured minants of revenue. First-stage results appear in table in a special way (PRS Group 2009). The ICRG ratings are A.5. The 2SLS estimates are presented in table A.6. based on three categories of risk: political, financial, and Some of the instrumental variable (IV) results are economic. A separate index is constructed for each of the indeed consistent with our main findings in table A.3. three categories. For purposes of our analysis, we focus on In particular, the GDP shares of agriculture, services, the first risk group, political risks. In the construction of and international trades are significant determinants the political risk rating, 12 variables covering both polit- of revenue with the same signs. We find that countries ical and social characteristics are measured. We use two with higher agriculture or service value added as share of of these variables—control of corruption and bureau- GDP are associated with lower revenue-to-GDP ratios cracy quality—in our empirical analysis of the effect of and that countries with a higher trade share of GDP tend governance on revenue mobilization. to enjoy higher revenue. Another consistent finding is that higher inflation rates are associated with lower reve- nues. The estimated effect of inflation is now higher and Control of Corruption more significant than in the main results. The effect of The corruption variable could take a value that ranges natural resource wealth is also in line with our baseline from 0 to 6, with 6 representing the best control of results. However, other results differ from our baseline corruption and 0 standing for the worst rating. This in table A.3. For instance, the estimated effect of GDP variable is a measure of corruption within the polit- per capita is insignificant in columns 1 and 3, without ical system. The ICRG measure of corruption not only controlling for the governance indicators. IV results takes into account financial corruption in the form of on the country dummies (table A.6) are also inconsis- special payments and bribes related to trade licenses, tent with our baseline results in table A.3. Bhutan and exchange controls, tax assessment, police protection, or Maldives show up with negative estimated coefficients, loans, but also is more focused on the type of corrup- which is counterintuitive. India now has a larger nega- tion, in the form of excessive patronage, nepotism, job tive estimated coefficient. Pakistan, Bangladesh, and reservations, secret party funding, and suspiciously close Nepal now have positive estimated coefficients. ties between politics and business. A higher score for this variable stands for lower risks and better control of corruption in the political system. Data and Measurement of Variables Our main data sources are the World Development Bureaucracy Quality Indicators (World Bank 2010) and Global Financial Statistics (IMF 2010), which are supplemented by The range of this variable is from 0 to 4. A higher the International Country Risk Guide (PRS Group value of this variable is assigned to countries where the 60 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries bureaucracy has better capabilities to govern without Hattari, Rabin, Kalpana Kochhar, Helene Poirson, and abrupt changes in policy or interruption in govern- Peita Topalova. 2008. “Taxes in India—Too Much, ment services. The institutional strength and quality of Too Little, or Just Right?� Report, International the bureaucracy is an important determinant of risks Monetary Fund, Washington, DC. and may minimize revisions of policy when govern- IMF (International Monetary Fund). 2011. ments change. Higher bureaucracy quality is likely to Government Finance Statistics (database). http:// be associated with more efficient management and elibrary-data.imf.org/FindDataReports.aspx?d= allocation of resources and more effective delivery of 33061&e=170809. public services. Mauro, Paulo. 1998. “Corruption and the Composition of Government Expenditure.� Journal of Public Economics 69 (2): 263–79. Note PRS Group. 2009. International Country Risk Guide. Syracuse, NY: PRS Group. 1. For instance, in the main instrumental variable Rajkumar, Andrew Sunil, and Vinaya Swaroop. 2008. specification, the overidentification test statistics “Public Spending and Outcomes: Does Governance Chi-sq(2) = 0.02 P-value = 0.9. Matter?� Journal of Development Economics 86 (2008): 96–111. Tanzi, Vito, and Hamid R. Davoodi. 1997. “Corruption, References Public Investment, and Growth.� IMF Working Gupta, Sanjeev, Hamid Davoodi, and Erwin Tiongson. Paper 97/139. International Monetary Fund, 2002. “Corruption and the Provision of Health Washington, DC. Care and Education Services.� In Governance, World Bank. 2010. World Development Indicators Corruption and Economic Performance, ed. George 2010 (database). World Bank, Washington, DC. Abed and Sanjeev Gupta, 245–72. Washington, http://data.worldbank.org/data-catalog/world DC: International Monetary Fund. -development-indicators/wdi-2010. Appendix B Tax-Specific Buoyancies, 2001–10 GDP as base Relevant basea Country Tax All years Last 3 years All years Last 3 years Bangladesh Income tax 1.57 1.70 1.45 1.67 Domestic indirect tax 1.18 1.39 1.23 1.32 VAT and sales tax 1.36 1.36 1.42 1.29 Trade tax 0.85 0.58 −2.20 −6.04 (omitting extremesb) 0.59 0.58 Bhutan Income tax 2.17 5.09 0.77 0.79 (omitting extremes )b 1.08 2.19 Domestic indirect tax 2.19 2.56 2.77 1.38 VAT and sales tax 1.36 1.88 1.33 0.61 Trade tax 1.28 1.16 −0.30 0.89 (omitting extremesb) 0.24 India Income tax 1.59 1.24 1.46 1.24 Domestic indirect tax 0.95 0.50 1.09 0.47 VAT and sales tax 0.96 0.65 1.09 0.60 Trade tax 0.49 0.02 2.36 6.90 (omitting extremes )b 0.18 0.45 Maldives Income tax −0.40 0.37 0.59 0.27 (omitting extremes )b 0.15 0.03 Domestic indirect tax 1.51 0.42 0.22 −2.79 Trade tax 0.30 −1.65 0.83 0.46 (omitting extremes )b 1.16 0.80 Nepal Income tax 1.17 1.70 1.16 1.93 Domestic indirect tax 1.55 1.76 1.71 1.70 VAT and sales tax 1.50 1.60 1.57 1.55 Trade tax 0.85 1.81 0.92 1.33 Pakistan Income tax 1.32 0.89 1.23 0.88 Domestic indirect tax 1.07 1.30 1.22 1.13 VAT and sales tax 1.03 0.98 1.13 0.85 Trade tax 0.66 0.41 −1.64 0.50 (omitting extremes )b 0.87 (continued) South Asia Economic Focus – A Review of Economic Developments in South Asian Countries 61 62 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries GDP as base Relevant basea Country Tax All years Last 3 years All years Last 3 years Sri Lanka Income tax 0.97 0.55 0.94 0.55 Domestic indirect tax 0.80 0.75 0.55 −0.16 VAT and sales tax 0.55 0.21 −1.77 −4.57 (omitting extremesb) 1.00 0.98 Trade tax 0.84 0.70 0.55 −0.34 Source: Authors’ calculations. Note: GDP = gross domestic product; VAT = value added tax. a. Base for income tax is nonagricultural income; for indirect tax, base is private consumption; and for trade tax, base is total trade. b. Omitting figures whose absolute value exceeds 5. Appendix C Country Studies: Fiscal Issues Country Study C.1  Afghanistan aid dependency, and use of parallel systems to circum- vent limited government absorptive capacity have Afghanistan is experiencing its second major transition hindered aid delivery and the building of a more effec- in a decade. After the collapse of the Taliban regime tive Afghan state. in 2001, the country is now facing the drawdown of most international military forces over the coming three years, along with a likely decline in civilian aid. These events will shape the country’s fiscal and economic Public Expenditure landscape as the government assumes full responsi- Public expenditures in Afghanistan include the oper- bility for security by 2014. ating budget, the development budget (together consti- tuting the core budget) and the external budget (see figure C.1.1). The operating budget consists largely Fiscal Space of the government wage bill plus nonwage O&M Afghanistan has limited instruments at its disposal costs, funded by domestic revenue, multidonor trust to create fiscal space. With a budget that is heavily funds, and donors’ budget support. The core develop- dependent on donor financing, fiscal management will ment budget comprises donor-financed projects and remain a challenge. The country will continue to rely programs executed by the government, plus a small on external funding for its operational budget over at portion funded by budget support and domestic reve- least the next 10 years. In spite of the constraints, the nues. Starting from negligible levels 10 years ago, the government could explore ways in which to create some core development budget currently finances about fiscal space for its development goals through further US$1 billion of development activities yearly, and prioritization and reduction of expenditures (mostly the external budget finances close to US$14 billion the operating budget). (budget for fiscal year [FY]2010/11), of which the bulk is spent on security. Afghanistan’s budgetary performance has been Aid impressive in many respects. Sound public finan- cial management systems have been introduced and Currently, annual aid (estimated at US$15.7 billion in progressively strengthened, and Afghanistan’s ratings 2010) is equal to about 100 percent of Afghanistan’s according to the Public Expenditure and Financial gross domestic product (GDP). Aid has funded the Accountability framework are above those of coun- delivery of essential services, including education and tries at its income level. In this context, fiscal disci- health, infrastructure investments, and government pline has been maintained, including control over the administration. Substantial improvements in the lives civilian wage bill and staffing. of Afghans have occurred over the past 10 years as a result of this effort. But these inflows have financed significant asset accumulation, which in turn suggests significant liabilities in terms of operations and main- Revenue tenance (O&M). These expenditures are currently In the future, Afghanistan will need to continue mobi- financed by donors off-budget, but they would need lizing more domestic revenue and further increase to come on-budget when donors exit. In addition, aid the revenue-to-GDP ratio. Revenue has grown by 20 has been so high that inevitable waste and corruption, percent of GDP per year during the past four years, South Asia Economic Focus – A Review of Economic Developments in South Asian Countries 63 64 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Figure C.1.1  External and Core Budget in Afghanistan, FY2010/11 a. External budget Nonsecurity Security US$5.2 billion US$8.6 billion b. Core budget Development 23% grants Security Operating US$1.3 billion Grants Security grants: US$2.35 billion 9% CSTC-A 12% LOTFA 8% ARTF RCW Domestic 48% revenues Development FY2010/11 Core budget financing US$945 million Source: Ministry of Finance and World Bank staff calculations. Note: Security budget of US$1.3 billion is part of the entire Operating budget of US$2.35 billion. ARTF = Afghanistan Reconstruction Trust Fund; CSTC-A = Combined Security Transition Command–Afghanistan; FY = fiscal year; LOTFA = Law and Order Trust Fund for Afghanistan; RCW = Recurrent Cost Window. compared to only 3 percent of GDP in FY2002/03, and spending liabilities that will come on-budget during is now equivalent to 10 percent of GDP. In FY2010/11, the transition process. collection reached an all-time high of US$1.7 billion, exceeding the International Monetary Fund (IMF) target. Behind these successful collection efforts is the Security Expenditure rapid growth of tax and customs duties (Ministry of Finance [[year?]]). However, additional revenue mobi- If the size of the security sector (Afghan National Army lization efforts and the government’s ambitious revenue and Afghan National Police) remains at currently agreed- targets will not be sufficient to finance the recurrent upon levels, Afghanistan will need to rely on continuing Country Studies: Fiscal Issues 65 international funding to pay for most security costs, Figure C.1.2  Projected Expenditure and Core Budget which will pressure operating expenditures (salary and Allocations in Afghanistan, FY2021/22 nonsalary). Security spending, mostly salaries, domi- (percentage of GDP) nated the total core budget in FY2010/11, and nearly 40 percent of the on-budget spending was allocated a. Projected expenditure to security. This share will grow over the coming years 60 as the Afghan national security forces aim to reach a target level of 352,000 troops by November 2012. With 50 Development such high security spending and with a likely decline Others in external financing over the upcoming years, there is 40 O&M civilian a real risk that priority development and O&M and 30 Civil service wages service delivery expenditures will be crowded out. 20 O&M security Domestic revenues, Financing Gap 10 17.5 Security wages In FY2021/22, Afghanistan’s financing gap is projected 0 at over 25 percent of GDP. After peaking at over 40 21 1 20 /10 2 20 /06 20 /13 20 /09 20 /11 20 /20 20 /07 20 /08 20 /14 16 6 20 /18 20 /12 20 /19 20 /15 20 /17 20 0/2 /2 20 5/1 09 10 12 05 08 19 13 06 07 18 percent of GDP in FY2014/15, the financing gap is 17 11 14 2 1 20 projected to gradually decline over time when expected mining revenues come onstream, reaching about a quarter of GDP in FY2021/22 (figure C.1.2, panel a). Even if one assumes that ambitious targets for robust b. Projected core budget allocations as a percentage of GDP growth in domestic revenue are met (with a projected Total Afghan core budget expenditures rise from 10 percent of GDP to more than 17 percent (43.0) of GDP a decade from now), there will be an unman- ageable fiscal gap. This gap arises primarily as a result Operating Development of rising security sector spending, for O&M and wages (33.0) (10.0) (for both the Afghan National Army and the Afghan National Police). In addition, nonsecurity spending Wages and Salaries O&M Others will contribute, to a lesser extent, to the O&M liabil- (16.0) (14.5) (2.5) ities associated with the handover of donor-built assets and with an increasing government payroll (as the pay Security Security (7.0) (10.5) and grading reform is completed and some additional expenditures are taken on to develop a senior civil Nonsecurity Nonsecurity service cadre at higher wages). Security spending is (9.0) (4.0) projected at over 17.5 percent of GDP in FY2021/22 (about equal to the total projected domestic revenue in that year); the civilian wage bill is projected at 9.0 percent; civilian nonwage O&M expenditure at 4.0 Sources: Ministry of Finance, Afghanistan; World Bank staff calculations. percent; core development budget at 10.0 percent; Note: GDP = gross domestic product; O&M = operations and maintenance. and other spending at 2.5 percent of GDP (figure C.1.2, panel b).1 Civilian aid will need to be spent more selectively for development and O&M. Various combinations percent of GDP). Donors would finance the remaining of aid and domestic resources for different spending bulk of security costs plus a more highly prioritized categories are possible. But a reasonable approach to core development budget. In this regard, low-income manage this financing gap could be for domestic reve- countries receive, on average, approximately 9 percent nues to cover the full civilian operating budget and to of gross national income in nonsecurity development pay for part of the security costs (at the current level of 3 assistance. Afghanistan would require close to three 66 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries times this level of total aid. If foreign assistance of this external financing are projected to be 3 percent and 2 magnitude for security and civilian expenditures is not percent of GDP in FY2012, respectively. forthcoming, then the government of Afghanistan will need to make some extremely difficult choices that could adversely affect service delivery. Revenue Strong revenue performance helped contain the deficit within the original budget target in FY2011. Overall reve- Country Study C.2  Bangladesh nues grew by 20.2 percent. Indirect tax collections from Sustainable budget deficits and their financing have both domestic activities and imports exceeded expecta- been among the strongest elements in Bangladesh’s tions, with growth of more than 29.4 and 22.1 percent, macroeconomic management in the past. The budget respectively. Income tax collections grew by 33 percent.2 deficit and public debt as a percentage of GDP remain The revenue-to-GDP ratio increased from 10.9 percent among the lowest in the region. in FY2010 to 11.6 percent in FY2011, and the tax-to- GDP ratio increased from 9.0 percent to 10.1 percent. Achieving a tax collection in excess of 10 percent of Fiscal Deficit and Debt GDP is a significant accomplishment for Bangladesh. The gains in tax revenue mainly reflect tax buoyancy Bangladesh has maintained a good record of fiscal and better implementation of administrative reforms in discipline over the past decade, despite severe pressures the areas of automation, registration, and enforcement. on public resources. In the context of a revenue-to- At present, indirect taxes constitute 71.3 percent GDP ratio of only 11.6 percent in FY2011, however, of tax revenue collection. Reliance on import-based Bangladesh needs to raise more revenue to meet its taxes has declined in recent years, from 50.0 percent social spending needs, to provide relief for its recur- in FY2005 to 35.4 percent in FY2011. Recent reforms rent natural disasters, and to balance the effect of the in tax administration have tended to boost the collec- global financial crisis. Total revenues increased by about tion of direct taxes. Despite these efforts, direct taxes 1 percent of GDP between FY2006 and FY2011 (with constitute only 28.7 percent of total taxes. most of the increase occurring in FY2011), and total spending increased by nearly 2 percent of GDP over the same period (figure C.2.1). Public debt fell from 47 percent of GDP in FY2006 to 43 percent of GDP in Expenditure FY2011 as a result of continued solid economic growth. Total public expenditure in FY2011 remained below Recently, there has been a shift toward a higher fiscal the revised budget target because of the underuse of deficit and a sharp increase in domestic financing. The Bangladesh’s Annual Development Program (ADP) overall fiscal deficit in FY2011 reached 4.3 percent of expenditures. A rise in subsidies to agriculture, power, GDP—0.6 percentage point higher than the previous fuel, food, and exports contributed to accelerating year. The increase in the overall deficit in FY2011 came non-ADP expenditures. Including grants and pensions, entirely from the primary deficit, which increased total subsidies and current transfers amounted to 4.0 from 1.6 percent of GDP in FY2010 to 2.4 percent percent of GDP in FY2011, compared with 3.8 percent in FY2011. The increase in the primary deficit reflects in FY2010. The increase in subsidies was more than the 1.6 percentage point rise in the expenditure-to- offset by a 0.3 percentage point decline of GDP in GDP ratio owing to significant increases in subsidies, non-ADP capital spending. The policy response to the transfers, and capital expenditures. growing subsidy and quasi-fiscal deficit has been in the The primary deficit is projected to rise to 3 percent proper direction, in view of the growing fiscal burden on of GDP in the FY2012 budget. At the current nominal account of fuel-related subsidies, though the response is effective interest rate of about 4.4 percent and GDP not quite adequate. Administered prices of diesel, furnace growth of 6.7 percent, this is sustainable in the sense oil, and fertilizer were increased by 38.6 percent, 130.8 that the debt-to-GDP ratio will remain stable at about percent, and 66.7 percent, respectively, in 2011. The 40 percent. The overall budget deficit is projected to average bulk power tariff was increased by 7 percent in rise to 5 percent of GDP, and domestic borrowing and August 2011 and by 14.4 percent in February 2012. Country Studies: Fiscal Issues 67 Nondiscretionary expenditures also are increasing. Figure C.2.1  Government Revenue and Expenditure, The expenditure on interest payment is projected to Bangladesh increase from Tk 153 billion (1.9 percent of GDP) in (percentage of GDP) the FY2011 budget to Tk 180 billion (2.0 percent of GDP, 11.0 percent of total expenditure) in the FY2012 20 budget. Salary and allowances for public servants is 18 projected to rise by 9.8 percent in FY2012. Total expenditure 16 Safety net programs continue to occupy a large 14 share of the budget, but there is an intention to reduce 12 Deficit duplication in coverage. The expenditure on social 10 safety nets has steadily increased, from 2.1 percent in Total revenue 8 FY2008 to 2.7 percent in FY2011. It has increased by 6 about 8.0 percent in nominal terms in the FY2012 budget, but as a share of GDP, the government plans 4 to reduce this expenditure to 2.5 percent in FY2012. 2 The biggest component of development expendi- 0 FY 8 FY 0 FY 0 FY 4 FY 2 FY 9 tures, the ADP, declined from 6.5 percent of GDP in FY 1 12 FY 1 FY 7 FY 06 FY 5 FY 03 1 1 0 0 0 0 0 0 0 0 20 20 20 20 20 20 20 20 20 20 20 20 20 FY2001 to 3.2 percent in FY 2009 before increasing FY to 4.2 percent of GDP in FY 2011. Average ADP use rate (percentage of total allocation spent) in the past Source: Ministry of Finance, Bangladesh. Note: FY = fiscal year (July to June); GDP = gross domestic product. Values for 2012 are from the 10 years has been 78 percent. Since FY2010, it has budget. improved to more than 80 percent. The ADP has been overprogrammed for FY2012, with a nominal increase of more than 40 percent over FY2011. Allocation for operations and maintenance remain Figure C.2.2  Financing of Deficit, Bangladesh inadequate. Historically, the lack of a proper balance (percentage of GDP) between maintenance and capital expenditures has contributed to the deteriorating quality of public assets 3.0 and service provision. The allocation for repairs, main- 2.5 tenance, and rehabilitation has been increased by only 8.9 percent in the FY2012 budget. 2.0 1.5 1.0 Deficit Financing The deficit was largely financed from domestic sources 0.5 (figure C.2.2). Foreign financing in FY2011 is estimated 0.0 at 0.9 percent of GDP, compared with 1.3 percent the previous year. Domestic financing of the deficit increased –0.5 FY2006 FY2007 FY2008 FY2009 FY2010 FY2011 FY2012 from 2.4 percent of GDP in FY2010 to 3.4 percent of Domestic bank borrowing External financing GDP in FY2011. With nonbank financing declining from 2.7 percent of GDP in FY2010 to 0.7 percent in Domestic nonbank borrowing FY2011 owing to weak demand for National Savings Source: Ministry of Finance, Bangladesh. Directorate certificates, the share of bank financing in Note: FY = fiscal year (July to June); GDP = gross domestic product. Values for 2012 are from the total domestic financing of the deficit increased from budget. −0.3 percent of GDP in FY2010 to 2.6 percent in FY2011. A major concern is the rise in monetary financing of the deficit to 2.1 percent of GDP, compared with a repayment of central government debt to the 68 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Figure C.3.1  Overall Fiscal Developments, Bhutan Country Study C.3  Bhutan (percentage of GDP) a. Fiscal deficit Fiscal management in Bhutan has largely been sound until recently. The overall budget framework is under- 4 pinned by the sound constitutional principle of covering current expenditures with domestic reve- 2 nues. In addition to fiscal discipline, sizable inflows of donor assistance have resulted in small fiscal surpluses 0 from FY2006/07 to FY2009/10. However, the fiscal deficit widened significantly in FY2010/11 to reach –2 4.8 percent of GDP, due in large part to rising capital –4 expenditures. (figure C.3.1). –6 Revenue –8 Total revenue (domestic revenue and grants) averaged 11 8 6 9 0 7 5 /0 /0 /0 /1 /0 /0 / 36 percent of GDP for the past decade. Receipts from 10 07 05 08 09 06 04 20 20 20 20 20 20 20 grants and nontax revenue (NTR) are much larger FY FY FY FY FY FY FY than tax revenue collections (figure C.3.2). Grants b. Revenue and expenditure and NTR collectively account for about two-thirds 30 of total revenue, with Bhutan recording the highest grant receipts. Of those receipts, domestic revenues 25 (tax and nontax revenues) as a share of GDP aver- 20 aged about 21 percent, with grants accounting for the rest. During the past decade, grants and electricity 15 revenues accounted for as much as two-thirds of the total revenues. Annually, the total revenues are vola- 10 tile because of the fluctuating grant inflows and the weather.3 Even within a year, annual revenues fluctuate 5 because of seasonality in hydropower revenues (in a 0 normal year, more than half of revenue is generated in the third quarter because of the rainfall pattern) 6 0 09 11 8 7 5 /1 /0 /0 /0 /0 0/ 8/ creating cash management problems. Development of 09 05 07 06 04 01 00 20 20 20 20 20 2 2 FY non-hydropower-related taxes such as a value added FY FY FY FY FY FY tax (VAT) and a rationalized income tax regime are Current expenditure Domestic revenue important priorities in this environment, as discussed Source: Budget Policy and Fiscal Framework Statement. in the main text. Note: FY = fiscal year; GDP = gross domestic product. Expenditure Total expenditures (excluding net lending) accounted for approximately 37 percent of GDP in the past Bangladesh Bank last year that amounted to nearly 1.0 decade. During this period, current expenditures percent of GDP. The rise in domestic financing at a time have been roughly half the total expenditures (figure when the liquidity in the banking system came under C.3.3). For many years current expenditures have pressure has raised concerns about crowding out bank largely been managed within the domestic revenue credit to the private sector. Also, monetary financing as required by the constitution. However, in recent raises the risk of fueling inflation in an already high years, current spending has been on the rise. Wages inflation environment. and salaries, representing 40 percent of total recurrent Country Studies: Fiscal Issues 69 expenditures (8 percent of GDP), have grown by an Figure C.3.2  Revenue Structure, Bhutan average 21 percent from FY2007/08 to FY2010/11. 35,000 Growth in expenditures on general goods and services has also been sizable, averaging 17 percent from 30,000 FY2007/08 to FY2010/11. As a result, the excess of 25,000 domestic revenue over the current spending has become narrower. Capital expenditure accounts for the rest, 20,000 high at about 20 percent of GDP. 15,000 10,000 Country Study C.4  India 5,000 India has good revenue mobilization, but has room 0 11 8 0 9 to improve composition and efficiency of spending. 6 7 5 /1 /0 /0 /0 /0 /0 / 10 09 07 08 05 06 04 20 In the second half of the 2000s, India made 20 20 20 20 20 20 FY FY FY FY FY FY FY remarkable progress in fiscal consolidation supported by fiscal responsibility legislation, after tradition- Tax revenue Nontax revenue Grants ally facing high deficits and the resulting high debt Source: Budget Policy and Fiscal Framework Statement. burden. Fiscal consolidation was achieved through Note: FY = fiscal year; NTR = normal tax rate. revenue mobilization efforts and expenditure restraint, which together improved the fiscal balance by nearly 6 percent of GDP (figure C.4.1). Revenue increased mainly because of the introduction of a general VAT on goods. Expenditures contracted partly because of Figure C.3.3  Expenditure Structure, Bhutan stagnating civil service salaries. 40,000 High public sector salary increases and fiscal stim- ulus measures eroded some of the gains in the wake of 33,000 the global financial crisis in 2008, but India’s govern- 26,000 ment has since reasserted its commitment to consolida- tion and prudent macroeconomic management. In the 19,000 wake of the crisis, some excises and import tariffs were lowered in late 2008 and early 2009. On the expendi- 12,000 ture side, salary adjustments were recommended by the 5,000 6th Pay Commission in 2008—independently from the global financial crisis—to be applied retroactively –2,000 10 9 11 from FY2006/07.4 Arrears payments and the higher 08 6 5 7 0 /0 0 /0 0/ 9/ 8/ 7/ 4/ 05 06 01 0 0 00 0 salaries themselves, as well as some additional (small) 20 20 20 20 20 2 2 FY FY FY FY FY FY FY fiscal stimulus measures, increased overall spending by Current expenditure Capital expenditure about 2.5 percent of GDP in FY2008/09 compared Net lending Current domestic revenue with FY2007/08 (figure C.4.2). Despite the widened fiscal stance since the global Note: FY = fiscal year; GDP = gross domestic product. financial crisis, debt indicators have continued to improve (see figure C.4.3). The ratio of public debt to GDP has declined significantly, to 68 percent of GDP in FY2010/11, from a high of 87 percent of GDP in the declining interest bill coming from favorable debt FY2003/04. External debt accounts for less than 4 dynamics and the erosion of civil service salaries, the percent of GDP. ratio of capital spending to GDP declined throughout Although the overall efforts at fiscal consolidation the decade, because of increasing transfers and subsi- were commendable, the 2000s saw a significant dete- dies. Transfers and subsidies result from inadequate rioration in the composition of spending. Despite pass-through of rising prices for food and energy items, 70 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Figure C.4.1  General Government Revenue, India because India’s social safety net relies to a large extent (percentage of GDP) on the distribution of subsidized food and nonfood items (for example, cereals, cooking oil, kerosene, liqui- 25 fied petroleum gas [LPG], and diesel). Fiscal space is also eroded by under-recovery of the 20 costs of providing goods and services by public sector entities. Most prominently, public sector oil compa- 15 nies are footing part of the bill of LPG and kerosene subsidies. Meanwhile, a crisis is brewing in the power 10 sector, where regulators have been unable to raise elec- tricity prices in line with rising input costs or to entice 5 power distribution companies to increase metered and paid-for electricity consumption as a share of overall 0 production. Accumulated losses have put the finan- 20 3 /11 20 0 20 5 1 00 20 2 20 04 20 8 20 9 20 6 20 7 /1 /0 /0 /0 /0 /0 cial viability of state electricity boards in jeopardy, /0 /0 /0 / 10 20 09 00 02 04 01 03 07 08 05 06 9/ 20 20 with lenders becoming reluctant to provide working 9 19 capital, which has resulted in increasing cross-arrears. Direct taxes Indirect taxes Nontax revenue In a similar situation, the central government paid the Source: Ministry of Finance, Government of India. equivalent of 1.6 percent of GDP in 2002 to settle bills Note: GDP = gross domestic product. of the state electricity board. Looking forward, the 13th Finance Commission presented a bold strategy of “expansionary fiscal consol- idation.� Its aim is to promote growth, while giving assurances that the state “will continue to mobilize and Figure C.4.2  General Government Expenditure, India deploy a significant proportion of resources to promote (percentage of GDP) public welfare� (India MOF 2009). The commission 30 proposes to adopt fiscal rules for an overall adjustment in the general government deficit. Capital spending 25 and net lending would rise sharply to 6.4 percent of GDP by FY2014/15, higher than at any time since the 20 late 1980s. For the first time, the government adopted 15 an explicit debt ceiling, with the consolidated govern- ment debt-to-GDP ratio scheduled to fall to 68 percent 10 by FY2014/15, from 75 percent in FY2008/09.5 The transfer of revenue from the central pool of taxes to 5 the state governments has been increased to 32.0 0 percent from 30.5 percent, and as in previous years, the commission proposed the use of a multitude of addi- 20 3 00 /11 20 0 20 8 20 9 20 01 20 6 20 7 20 02 20 4 20 05 /1 /0 /0 /0 /0 /0 /0 / / 20 / 10 09 00 02 07 08 05 06 01 03 04 tional grants to promote changes in financial manage- 9/ 20 9 19 Interest Transfers Subsidies ment in the states. Fiscal consolidation would be built in part on Consolidations Capital an increase in revenue, which would be achieved by Source: Ministry of Finance, Government of India. the introduction of a goods and services tax (GST) Note: GDP = gross domestic product. to replace and expand the current VAT. The GST is envisaged as a game-changing factor in revenue mobilization not only by expanding the tax base and simplifying the tax code, but also, more important, by integrating markets across India—through removal of barriers to interstate movement of goods and services and leveling of the playing field across jurisdictions. Country Studies: Fiscal Issues 71 The GST is therefore seen as promoting higher GDP Figure C.4.3  Deficit, India growth and mobilizing a higher percentage of GDP (percentage of GDP) in taxes. Although the central and state governments 12 agree on the desirability of the GST, discussions are ongoing regarding the modalities. These include the 10 framework, rate structure, taxes to be subsumed under the new GST, list of items to be exempted, treatment 8 of interstate services, threshold limits, and tax admin- 6 istration. Constitutional changes will be required to enable the central government to levy taxes beyond 4 the manufacturing stage and to empower the states to levy taxes on services. An institutional mechanism 2 needs to be built to ensure compliance. On the expenditure side, strategic resource alloca- 0 00 /11 20 0 1 20 9 20 6 3 02 4 20 08 20 05 20 7 /1 /0 /0 /0 /0 /0 /0 tion in India is affected by the lack of well-developed / 20 / / 10 09 00 08 05 02 01 03 07 04 06 9/ 20 20 20 20 20 sector strategies based on government objectives, the 9 19 development and costing of programs to achieve those objectives, and the linking of resource allocation to the priorities specified in sector strategies. Source: Ministry of Finance, Government of India. Note: GDP = gross domestic product. In the federal arrangement, the subnational govern- ments have wide-ranging responsibilities with regard to service delivery. The central government, however, intervenes in the state subjects through specially designed central schemes to improve the frontline service delivery. The increasing funding of centrally sponsored schemes reduces scope for local innovation, Country Study C.5  Maldives improvements in targeting, monitoring, and evalua- Fiscal Policy tion. The role of the central government in contrib- uting to efficient service delivery through effective Expansionary fiscal policy, particularly in the years monitoring of transfers to implementing agencies, following the 2004 tsunami, has put a major strain in providing guidance through policy measures, and on macroeconomic stability and has remained a major in evaluating the performance in these services has destabilizing factor for the Maldives economy. The become important. spending increase initially was on reconstruction efforts, Important weaknesses of the public financial but increasingly thereafter, it was on matters unrelated management system, which limit fiscal discipline, to reconstruction (or the tsunami). In particular, the include (a) the absence of a multiyear perspective in rapid growth under the public sector remunerations the expenditure planning that indicates future year bill6 (reflecting an increase in both the number of commitments, (b) a lack of effective fiscal risk assess- personnel and the salaries and wages) contributed to ment at an aggregate level, (c) the unevenness and high expenditures that were not sustainable in a fragile the late spike in the annual spending pattern, (d) the revenue structure. Recurrent expenditure, which was surrender of money at the end of the fiscal year in an 20.6 percent of GDP in 2004, shot up to 39.0 percent annual lapsable budget cycle owing to a lack of effec- of GDP in 2009, and the budget deficit exploded from tive program management in budget implementation, a benign 1.2 percent of GDP to an unprecedented (e) an absence of a hard budget constraint, and (f ) 23.4 percent of GDP (figure C.5.1). The onset of the the weak internal control and internal audit system. global financial crisis, and the consequent drying up Although an external audit system is well established of tourist arrivals and fall in external financial flows, in the country and facilitates the legislative branch meant that the economy had to adjust—the hard way in exercising control over the executive branch, the (figure C.5.2). process of scrutinizing the audit reports has deterio- Indeed, the country underwent a painful adjust- rated, adversely affecting its effectiveness. ment process, underpinned by wage cuts across the 72 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Figure C.5.1  Budgetary Outturns, Maldives public sector, hiring freezes, and increase in utility (percentage of GDP) prices. The process was supported by an IMF program (December 2009) and was complemented by budget 60 support operations from both the World Bank and the 50 Asian Development Bank. In 2010, the budget deficit 40 was reduced to 12.7 percent of GDP. However, with the 30 full restoration of public sector wages and other expen- 20 diture measures that came onstream, further consolida- tion over the medium term is expected to be slow and 10 gradual. As a result of these concerns, the IMF program 0 has been in a state of hiatus since April 2010, and the –10 latest debt sustainability analysis has rated Maldives as –20 at high risk of debt distress, essentially making it inel- –30 11 igible for nonconcessional borrowing. 10 02 03 01 09 04 06 07 08 05 20 20 20 20 20 20 20 20 20 20 20 Expenditure Revenue Budget deficit High Expenditures Source: Ministry of Finance and Treasury, Maldives; Monetary Authority, Maldives. In the immediate aftermath of the tsunami, social Note: GDP = gross domestic product welfare expenditures of the government (quite under- standably) rose, but they had largely disappeared by 2007. More worrying, however, was the substan- tial growth in public sector employment and remu- neration, an outcome unrelated to the tsunami. The public sector workforce in Maldives had been steadily Figure C.5.2  Revenue Composition, Maldives rising, from about 27,000 in 2000 to about 32,000 (Rf, millions) by end-2010, suggesting that 1 person in 10 in the 10,000 country was employed in the public sector (one of the highest ratios in the world). On top of this figure, 8,000 the public sector, particularly the civil service (the largest segment of Maldives’ public sector), period- 6,000 ically received generous increments—far in excess of any reasonable magnitude of merit increase or cost-of- 4,000 living adjustment. General pay increases and promo- tions granted from 2004 to February 2009 doubled 2,000 the government’s remuneration bill, rising from 9.6 percent of GDP in 2004 to 22.6 percent of GDP by 0 2009 (figure C.5.3). Currently, more than 40 percent 10 03 04 01 06 08 11 09 02 05 07 of total expenditure is accounted for by remunerations, 20 20 20 20 20 20 20 20 20 20 20 Import tax Tourism tax Other tax Resort lease rents greatly limiting fiscal space. Dividends from SOE Other nontax Capital revenue Grants Source: Ministry of Finance and Treasury, Maldives. Revenue Note: SOE = state-owned enterprise. Maldives’ revenue structure is almost entirely depen- dent on the tourism sector, with both major tax and major nontax revenues having high dependence on the sector. Dividends from state-owned enterprises are perhaps the only exception, accounting for 15 percent of total government revenue, but even these have a Country Studies: Fiscal Issues 73 vicarious dependence on the tourism sector. Import Figure C.5.3  Remunerations, Maldives duties and tourism tax account for nearly 85 percent of (percent) (percentage of GDP) tax revenue (40 percent of total revenue), while resort 200 25 lease payments account for nearly 17 percent of total revenue. Maldives’ total revenue (inclusive of grants) 20 averaged 26 percent of GDP prior to the tsunami 150 (2001–04), but the significant inflow of grant assis- 15 tance in the wake of the tsunami, coupled with robust 100 tourist arrivals after 2006,7 pushed revenues closer to 10 40 percent of GDP by 2007. However, both tax and nontax revenue suffered 50 5 considerably with the onset of the global financial crisis, with revenue dropping to 26 percent of GDP in 0 0 2009. The crisis underscored the vulnerability of the 01 04 10 11 08 06 02 09 03 07 05 20 20 20 20 20 20 20 20 20 20 20 tax system in Maldives, in view of both a high depen- dency on volatile sources and the lack of a domestic Remuneration bill and tax revenues (left-hand axis) tax backbone (such as sales tax and income tax regimes, Total remunerations bill (right-hand axis) barring a bank profits tax). For many years, institu- tional capacity limitations had inhibited the introduc- Source: Ministry of Finance and Treasury, Maldives. Note: GDP = gross domestic product. tion of a modern, broad-based tax structure and an efficient tax administration. The new government of President Mohammed Nasheed, which came to power in the midst of the financial crisis and in the face of a growing fiscal and balance of payments (BOP) crisis at home, wanted expeditiously to undertake tax reforms with the support of the IMF.8 However, in the face of considerable political opposition,9 the reform measures were considerably delayed, exposing the country to much vulnerability. Nevertheless, in 2010, the Majlis Maldives is faced with an enormous challenge of endorsed the long-anticipated ad valorem tax on regaining the macroeconomic stability that is vital to tourism (tourism GST) and the business profits tax, alleviating the vulnerabilities the economy is currently and in 2011, crucial income tax legislation was also facing: (a) the growing crisis in Europe, the largest presented to parliament. source of the tourist market (more than 60 percent of arrivals); and (b) the unsustainable size of the civil service. Therefore, the single remedial measure the Fiscal Space country will need to focus on in the coming years will be the civil service reforms—pruning the size of the Creating fiscal space in Maldives in the coming years civil service and reducing the current remunerations would be considerably challenging. Even though bill, which are clearly at unsustainable levels. revenue-generating measures have finally been imple- mented, the shift is not expected to result in a signifi- cant reduction in fiscal deficits over the coming years Country Study C.6  Nepal because of the parallel increase in expenditure on account of new measures passed by the parliament10 In Nepal, fiscal management has been commend- and the government. Faced with large fiscal deficits able. Fiscal prudence has been maintained by keeping in 2008–09, the government undertook significant net domestic borrowing to about 2 percent of GDP monetary financing, which became a significant desta- (considered by the IMF and the World Bank to be the bilizing factor—it broke the Maldivian rufiyaa peg to sustainable level). In addition, revenue rose significantly, the U.S. dollar in April 2011 and stoked strong infla- from 11 percent of GDP in FY2006 to 15 percent in tionary pressures. FY2011 (figure C.6.1). 74 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Figure C.6.1  Revenue Composition, Nepal Revenue (percentage of GDP) Regarding revenue, there are weaknesses. The struc- 18 ture is highly dependent on imports. Import-related 16 revenues (tariff revenue plus VAT and excise taxes on 14 imports) are nearly half of the total revenue. Good 12 performance in the recent past can be explained in 10 large part by high growth in remittances that fueled imports. Another weakness is the large number of tax 8 incentives and exemptions. These tax expenditures 6 rose recently as the government announced additional 4 incentives for investors and exporters in the FY2011 2 budget. Such expenditures erode the tax base, weaken public finances, and do not effectively address pro- 0 growth or pro-poor objectives. 08 10 im 11 09 07 ge 2 06 e) ud 01 t) st 0 Improving the current system can be achieved 20 20 20 at 20 20 (e FY2 (b 2 FY FY FY FY FY FY mainly through reducing tax exemptions and incen- tives and increasing the prevalence of domestic-source Income tax VAT International trade tax taxes. The IMF estimates that eliminating some VAT Excise tax Other Nontax revenue exemptions and corporate tax incentives would add 2.2 percent of GDP to revenues. With remittances Note: FY = fiscal year; GDP = gross domestic product;, VAT = value added tax. slowing down (and IMF’s long-lasting tax administra- tion coming to an end), other prospects for increasing revenue are limited. In addition, new developments are being identi- Figure C.6.2  Expenditure, Nepal (percentage of GDP) fied. In 2011, authorities unearthed large-scale VAT evasion by major business houses, mostly through the 25 use of fake receipts for credit, but more sophisticated evasion methods are also being employed. Identifying 20 these methods and stopping or reducing tax evasion have become a high priority. 15 10 Expenditure Regarding expenditure, total spending has been kept 5 consistent with revenues, foreign aid, and sustainable deficits. Figure C.6.2, indicates fairly good practice— 0 the overall capital budget is rising while wages are kept constant as a share of GDP. 11 10 09 08 06 ge 2 07 ud 01 20 t) 20 20 20 20 20 However, some major issues also confront Nepal— (b Y2 FY FY FY FY FY FY F inadequate transparency and limited implementa- Staff salary Other current salaries tion capacity, both of which jeopardize the quality Capital formation Other capital expenditure of spending. Regarding transparency, spending known as transfer Note: FY = fiscal year; GDP = gross domestic product. payments (both current and capital) is rising fast and is now the largest expenditure item, amounting to 9 percent of GDP (or 38 percent of total expenditure). Because monitoring of transfer payments is difficult and recipients consider them as their entitlements, Country Studies: Fiscal Issues 75 budgetary transparency has declined as rigidities have Figure C.6.3  Capital Formation and Transfers, Nepal risen. Transfers also include a large (but unknown) sum (percentage of GDP) of teacher salaries. Because the government does not 10 budget contingencies for possible outlays, such as the 8.9 8.9 cost of postconflict army integration or potential cost 8.0 8.1 of financial sector restructuring, the elevated rigidity 8 raises the risk that long-maintained fiscal prudence will be jeopardized. 6 5.3 5.6 The second problem is the limited implementa- 4.4 4.6 4.3 tion capacity of the public sector. Spending in the 4 3.4 3.1 3.2 line item classified as fixed capital formation has been 2.3 2.3 limited to about 2 percent of GDP in the past and is 2.0 1.9 2 now 3.0–3.5 percent (figure C.6.3). Even though aid financing increased significantly, implementation is lagging. Efforts to spend increased foreign financing have 0 ge 1 09 10 08 ge 2 07 06 05 ud 01 in fact contributed to the fast rise of transfer payments. ud 01 F ) t) t 20 20 20 20 20 20 (b 2 (b Y2 FY FY FY FY FY FY Public financial management has not been improving, FY but instead is exacerbating transparency and governance problems and increasing implementation difficulties. Note: FY = fiscal year; GDP = gross domestic product. Capital formation Transfers Nepal’s National Planning Commission (NPC) and the United Nations Development Programme (UNDP) recently issued the Millennium Development Goals Needs Assessment for Nepal 2010 (NPC and UNDP 2011). The assessment states that Nepal needs addi- tional resources amounting to 2.8 percent of GDP this year and that the need rises to 5.8 percent of GDP by again facing sizable challenges. Some of these problems 2015. The report concludes that such resources are are caused by external developments, for example, the unlikely to be forthcoming. adverse security situation, natural disasters, and the Two observations about the needs assessment should commodity and financial market shocks emanating be noted. First, the report’s pessimistic conclusion from international markets, but some of them are notwithstanding, streamlining incentives and exemp- internal and policy related. One important avenue tions alone would give additional resources amounting to addressing these problems is by putting the coun- to 2.2 percent of GDP, and if the authorities were to stop try’s fiscal house in order, which is critical both for oil-product subsidies, another 1.0–1.5 percent would reestablishing economic stability and for instigating be added. With further efforts to prevent evasion and economic recovery. rationalize transfer payments, filling the resource gap to meet the Millennium Development Goals (MDGs) would be within the authorities’ reach (with domestic Expenditure resources). Second, if the existing problem of limited implementation capacity is not addressed, having access Within an international context, Pakistan has a rela- to this large amount of additional resources would not tively low level of budgetary expenditure because, in enable Nepal to reach the goals anyway. Therefore, the part, the bulk of the public corporate sector is outside conclusion remains that the most important issue in the budget and the size of general government is not Nepal is to address implementation capacity. excessive. In addition, this low level is partly an outcome of the difficulties that successive governments have faced in mobilizing revenues. However, this relatively low level of public spending has its own implications Country Study C.7 Pakistan for the quality and quantity of services delivered by the After a period of respectable economic and fiscal perfor- government departments and institutions—the poor mance in the early 2000s, the Pakistani economy is social indicators and infrastructural shortages that the 76 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Figure C.7.1  Creation of Additional Fiscal Space, Pakistan country has experienced over the past several years. The (percentage of GDP) government has created initiatives and formulated plans to improve the outreach and quality of social services 4 1.8 and public infrastructure and to address the poverty and 3 0.5 vulnerability concerns that have been heightened by a poorly performing economy and persistent double-digit 2 0.4 inflation. In addition, perhaps the foremost priority is 0.2 2.7 to reduce the fiscal deficit and reestablish fiscal stability. 1 1.2 Creating additional fiscal space, therefore, is critical for 0 0.2 Pakistan’s plans of restabilizing the economy, acceler- –0.4 –0.6 ating economic growth, and reducing poverty. –0.8 –1 –0.9 –2 FY2003/04–FY2006/07 FY2007/08–FY2010/11 Fiscal Space Nontax revenue Interest payments Although there is no perfect definition of fiscal space, for the purpose of elaboration, one possible measure Tax revenue Fiscal deficit Defense of fiscal space could be the level of fiscal resources Source: Ministry of Finance, Pakistan. currently used, or required, to finance noncommitted Note: FY = fiscal year; GDP = gross domestic product. public expenditure. In the Pakistani context, this type of expenditure would mean noninterest, nondefense expenditure.11 After a period of poor economic and fiscal perfor- mance in the 1990s, Pakistan, through economic reforms in the early 2000s, took some important Figure C.7.2  Use of Additional Fiscal Space, Pakistan (percentage of GDP) strides toward improving the overall economic situa- tion. The main factor behind this positive change was 4 the government’s attempt to enhance the fiscal space by containing low-priority expenditures and public 3 1.8 debt. During this period (FY2003/04–FY2006/07), the establishment, defense, interest, and subsidy expendi- 2 0.4 1.1 tures all declined. As such, a fiscal space of 0.4 percent of GDP per year was created to enhance the noncom- 1 1.5 1.4 mitted expenditure (see figure C.7.1). However, this followed a decline of 0.8 percent of GDP per year in 0 –0.3 –0.7 –0.6 fiscal deficit. Regarding use, there was an annual increase –1 –0.3 of 1.5 percent of GDP in development spending. With overall fiscal space amounting to only about one-quarter –2 of this increase, it was the decline in subsidies (0.3 FY2003/04–FY2006/07 FY2007/08–FY2010/11 percent of GDP per year) and other recurrent expen- ditures (by 0.7 percent of GDP per year) that made Subsidies Development Other current this possible (figure C.7.2). Source: Ministry of Finance, Pakistan. Most of the gains from economic and fiscal reforms Note: FY = fiscal year; GDP = gross domestic product. of FY2003/04–FY2006/07 were squandered during the next four years as the worsening security and polit- ical environment prohibited the government from providing an adequate response to the commodity price shock of FY2007/08, which badly destabilized the economy. Because the government remained reluctant to pass the sharply higher commodity and energy prices Country Studies: Fiscal Issues 77 on to the consumer, the subsidy expenditure increased Figure C.7.3  Additional Fiscal Space Needed in Pakistan numerous times. Although, according to our definition, (percentage of GDP) the large fiscal space of 1.8 percent of GDP per year 10 Total spending, 8.6 was created, this was solely through a sharp increase in fiscal deficit (by 2.7 percent of GDP per year).12 8 Required additional 0.4 As per Similarly, the use of fiscal space was in sharp contrast fiscal space, 4.0 2.5 stabilization program and to the earlier period, the subsidies increased by 1.4 6 Public spending, 4.6 GoP As per MTDF , percent of GDP per year, and other recurrent expen- 0.7 Vision 2030, 0.3 1.0 ditures rose by 1.1 percent of GDP per year (owing 4 0.7 and other initiatives to rising establishment costs), whereas development 2.8 spending declined by 0.6 percent of GDP per year. 2 3.3 As per FRDL Act 1.5 0 FY2007/08–FY2010/11, FY2011/12–FY2014/15, Fiscal Challenges average average Pakistan today faces a dual challenge of stabilizing the Physical infrastructure Human development Transport economy and accelerating and sustaining economic Electricity Irrigation Fiscal stability Social safety nets growth. The former requires reducing the fiscal deficit from about 6.5 percent of GDP per year to about 4.0 Source: Ministry of Finance, Pakistan. percent of GDP per year over the next four years. Note: FRDL Act = Fiscal Responsibility and Debt Limitation Act; FY = fiscal year; GDP = gross domes- The latter rather calls for large investments in human tic product; GoP = Government of Pakistan; MTDF = medium-term development framework. resources and infrastructure. For instance, to meet MDGs, Pakistan would have to increase its social spending by 0.5 percent of GDP per year (figure C.7.3). To accelerate economic growth, significantly higher levels of public investment are needed, espe- two-thirds of which has been financed out of domestic cially in the power, water, and transport sectors. On borrowings (figure C.8.1). Debt-to-GDP ratios peaked the aggregate, Pakistan would have to generate an at more than 103 percent in 2001 but then gradually additional fiscal space of about 4 percent of GDP per declined as a result of high inflation and strengthening year to achieve its stabilization and economic growth of the rupee.13 Yet by 2010, the debt-to-GDP ratio was objectives. This fiscal space could be partially achieved 82 percent, the highest in South Asia. Before 2007, the by increasing revenues by 0.6 percent GDP, reducing predominant form of foreign borrowing consisted of untargeted subsidies by 1.5 percent of GDP, and concessional borrowings from multilateral and bilateral lowering the interest payments by 0.4 percent of GDP. partners, but since 2007, increased foreign commer- The remaining 1.5 percent of GDP could come from cial borrowings were seen. The government securities enhancing the efficiency and effectiveness of govern- market, consisting of rupee-denominated treasury bills ment expenditure through budgetary, fiscal manage- and treasury bonds, was opened up for foreigners in ment, and procurement reforms and strengthening 2007, with subscriptions reaching US$2.2 billion by of the public-private partnership framework in areas end-2010. The government also successfully placed traditionally reserved for the public sector. The ques- several eurobond issues from 2007, bringing in US$2 tion remains, however, whether in the present polit- billion from foreign investors. The total foreign-held ical environment the government will have the will and debt reached 36 percent of GDP in 2010, of which motivation to undertake the desired reforms. nonconcessionary commercial debt accounted for about one-quarter, or 9 percent of GDP. Country Study C.8  Sri Lanka Sri Lanka has run persistently high budget deficits. The Fiscal Structure average fiscal deficit (for the central government) has The broad fiscal structure has not undergone much been about 8 percent of GDP from 1990 to 2010, about transformation over the past few decades. Just three 78 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Figure C.8.1  Financing of Deficit, Sri Lanka items of recurrent expenditure—public sector remu- (SL Re, billions) neration; interest bills; and subsidies and transfers, including pension payments—account for nearly 500,000 two-thirds of total expenditure, and from 2001, total 400,000 expenditure has consistently exhausted total revenue generated (figure C.8.2). Despite the ongoing conflict, 300,000 from 1990 the country has maintained an average capital expenditure of 6.5 percent of GDP, much of 200,000 it devoted to public investments in social and phys- 100,000 ical infrastructure. The medium-term fiscal framework (MTFF) does not envisage any significant reduction in 0 the recurrent expenditure items (although faster GDP growth proportionate to these expenditures is expected –100,000 to reduce relative levels of these expenditures over the 20 1 20 9 00 19 2 1994 20 9 10 19 8 20 04 19 1 20 06 93 20 8 19 5 19 7 20 3 19 0 96 2002 20 5 2007 0 9 9 0 9 medium term, albeit not significantly). However, the 9 0 9 9 0 9 0 20 19 19 19 MTFF anticipates that capital spending of 6.0–6.5 Foreign financing Domestic financing percent of GDP can be sustained, thus meeting infra- structure development needs. Sri Lanka’s tax structure remains unbalanced and laden with administrative complexity. Since 1990, the number of taxes oper- ating in the country has grown steadily—to about 25 taxes by 2010. The contribution of indirect taxes to total taxes has remained persistently high, at about 81 percent in 2010. Despite the significant growth in the Figure C.8.2  Composition of Expenditure, Sri Lanka economy—averaging more than 6 percent since 2000— (percent) the tax base has not kept pace (see discussion below) and tax elasticity remains below unity. 35 30 25 Revenue 20 A progressive reduction in revenue generation largely reflects the effects of falling domestic taxes and trade 15 taxes (figure C.8.3). A GST replaced the business turn- 10 over tax (BTT) system in 1998, with tax rates set at 5 below break-even levels.14 In 2001, as part of the IMF program, the GST was replaced by a VAT, but the new 0 system still was weaker than the earlier BTT in terms 20 01 20 0 20 3 10 2099 20 4 20 05 20 6 20 8 19 6 19 0 09 97 2002 19 1 19 8 19 2 19 3 19 4 19 5 07 0 0 0 0 0 9 9 9 9 9 9 9 9 of revenue generation. In years following the intro- 20 19 19 20 Wages and salaries Interest payments Transfer payments duction of the VAT, the system became inherently Other current expenditure Capital and net lending Revenue complex, with constant changes to both the structure and the rates.15 At the same time, the list of exemp- Sources: CEIC Data database, http://www.ceicdata.com/; CBSL. tions under VAT also grew as a result of industry lobbying. These factors considerably undermined the revenue-generating ability of the tax. Trade-related taxes, composed of both customs duties (direct tax) and other taxes (the most significant of which is the Port and Airport Development Levy), have also seen a marked decline over the years. Under World Trade Organization commitments and the South Asian Country Studies: Fiscal Issues 79 Preferential Trade Agreement, Sri Lanka consider- Figure C.8.3  Revenue Composition, Sri Lanka ably lowered its base tariff rates in the mid-1990s. As (percent) a result, customs duty collections were halved, from 25 about 5.0 percent of GDP in 1990 to 2.5 percent of GDP by end-1999. Collections continued to fall 20 thereafter to about 1 percent of GDP by 2010, partly as a result of free trade agreements with India and 15 Pakistan and partly owing to the rise in duty exemp- tions granted to Board of Investment (BOI)–approved 10 enterprises. In recent times, efforts have been made to further simplify the duty structure (on a three- 5 band tariff structure) and to prune customs duties on a wide array of consumer durable goods, including 0 motor vehicles and electronic items. Tax statistics also 20 1 19 8 20 3 99 19 3 20 0 94 20 6 19 5 96 10 19 0 20 8 20 2 91 09 19 2 04 20 5 07 19 7 0 9 0 9 0 9 9 0 0 0 9 0 9 20 19 19 19 20 20 19 20 show a gradual lowering of income taxes in relation to total taxes despite the introduction of new direct tax Taxes on trade Taxes on domestic goods and services measures. Tax avoidance is considerable in Sri Lanka, Taxes on incomes and dividends Other taxes Nontax revenue with a substantial informal segment and with ad hoc tax exemptions and tax waivers granted over the years under the BOI having significantly curtailed the coun- try’s revenue potential.16 Fiscal Space The end of the 30-year armed conflict in May 2009 opened up a new and bright vista in the country’s tax administration18 was implemented in part with the postindependence history—and with it the need for 2011 budget, though much more needs to be done. greater fiscal space. The need for greater fiscal space In line with the commission’s recommendations, the in Sri Lanka is largely twofold: (a) undertake public government made its first strides toward a simpler and investments in infrastructure—particularly in roads, more broad-based tax system. The VAT was converged power generation, and ports and airports—that is into a single rate, and several “nuisance� taxes (such essential to sustain the greater than 8 percent growth as the bank debits tax) were removed. However, going momentum anticipated by the government, and (b) forward, the government may want to think of broad- have sufficient ability to engage in countercyclical poli- ening the VAT base further (to include the wholesale cies in an increasingly uncertain external environment. and the retail sectors) and, as a key consideration, In the wake of the global financial crisis, the country pruning the current (excessive) list of exemptions entered into a program with the IMF in 2009 and under the VAT. has made significant progress.17 The program envis- Beyond additional revenue raising, other means of ages considerable fiscal consolidation that is reflected creating fiscal space are likely to have limited effect. in the MTFF. However, staying on course with the Under the IMF program, the government undertook program targets is likely to be a substantial challenge, to improve the efficiency of public enterprises19 in a bid particularly after 2012. to reduce fiscal transfers to those enterprises. However, For methods to generate fiscal space, revenue-raising since the onset of the program, this effort has made measures (both tax and nontax) clearly appear to be only limited progress. The country’s debt dynamics the dominant practice, as reflected in the MTFF. The remain favorable, allowing for greater leverage in the medium-term objective is to raise tax revenue to more future (see appendix D), but this projection is subject than 15 percent of GDP from the current level of less to much uncertainty, such as the possibility of both than 13 percent. The Presidential Tax Commission’s an emerging-market credit downgrade as a result final recommendation for reforming the tax system and of current global conditions and a rollover risk in 80 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries consideration of both low (Macaulay) duration of the Party emerged as a minority to former President debt portfolio20 and likely high future interest rates.21 Maumoon Gayoom’s party. The relationship of the government and private sector 10. The “scorched earth� tactics adopted by highly has also witnessed many ups and downs in recent partisan Majlis—whose majority is with the oppo- times, with the government giving no concrete signals sition—have seen a number of bills passed that (or undertaking any strong reforms) to underline that were highly inimical to the government’s fiscal the private sector is the engine of growth. Indeed, the consolidation plans. government action on this front has been exactly the 11. Here, the average fiscal space is calculated using opposite: venturing into areas hitherto undertaken the average values for the main fiscal variables (in by the private sector, increasing encroachment into terms of GDP ratios) for three successive four-year the private sector commercial banks setting up new periods—FY1999/00–FY2002/03; FY2003/04– state-owned enterprises to compete with the private FY2006/07; and FY2007/08–FY2010/11;taking sector; and, more recently, passing laws to expro- their differences across two adjoining periods; and priate lands and businesses deemed underperforming then summing up these differences according to or underused. the fiscal space definition given in this section on Pakistan. 12. In net terms, the space available for noncommitted expenditure declined by 0.9 percent of GDP per Notes year (that is, 2.7 less 1.8 percent of GDP per year). 1. Other operating expenditures include pension, 13. High inflation reflects faster denominator growth, transfer, interest payments, and liabilities associ- and strengthening of currency results from real ated with Kabul Bank. exchange rate appreciation. 2. This has been a result of intensifying tax collec- 14. In 1998, a GST was introduced with one rate at tion from the existing base rather than bringing 12.5 percent to replace the existing three-band new taxpayers into the net. turnover tax rates of 6, 8, and 18 percent. The 3. Nontax revenue dropped considerably because of revenue-neutral level was closer to 14 percent. poor rainfall in FY2010/11. 15. A simple two-band structure at inception became a 4. Pay commissions recommend adjustment three-band structure in 2004, a five-band structure of civil service salaries once every 10 years. by 2006, and then a four-band structure in 2010. Recommendations from the pay commissions have 16. In addition to duty concessions, BOI-approved generally been accepted by the central government enterprises were also entitled to corporate (income as well as the state governments. In between the tax) holidays ranging from 5 to 10 years, suggesting adjustments by the pay commissions, civil service that these entities largely remained outside of pay is only partially adjusted for inflation through the tax net. the dearness allowance. 17. The original 30-month standby arrangement of 5. The target of 68 percent of GDP was achieved US$2.9 billion has seen eight tranche disburse- already in FY2010/11, partly because of changes ments up to June 2011, the furthest the country in definitions and accounting rules. has ever progressed under an IMF program. 6. The bill encompassed salaries and wages, allow- 18. Tax administration reforms have been drawn from ances, and benefits of the public sector work force. the ongoing Asian Development Bank–funded 7. Tourism revenue yielded both higher tax revenue Fiscal Management Efficiency Project, which is (because of higher import duty collections for working on aspects of simplification of proce- resort imports and higher tourism tax collections) dures and processes, improvement in the manage- and higher nontax revenue (because of the rapid ment information system, tax audits, and human surge of new resort lease contracts granted to cater resource development. to the growing demand). 19. Particular targets were the loss-making Ceylon 8. Maldives entered into an IMF program in Electricity Board and the Ceylon Petroleum December 2009. Corporation. 9. In the parliamentary (Majlis) elections held in 20. Macaulay duration is a technical term used to May 2009, the government’s Maldives Democratic describe weighted average time to maturity. This is Country Studies: Fiscal Issues 81 distinct from modified duration, which is used by References investment analysts to calculate interest rate sensi- tivity of bond prices. The duration of outstanding Afghanistan MOF (Ministry of Finance). [[AQ: Add government domestic debt portfolio was 1.74 in year]]. “Fiscal Bulletin Q1.� Afghanistan MOF, Kabul. 2010, down from 1.77 in 2009. India MOF (Ministry of Finance). 2009. “Thirteenth 21. In response to the global financial crisis, the Finance Commission Report.� Ministry of Finance, Central Bank of Sri Lanka has undertaken consid- Government of India, New Delhi. erable monetary easing from 2008 to 2011. This NPC (National Planning Commission) and UNDP approach has driven demand pressures and is (United Nations Development Programme). 2011. likely to be reflected in higher inflation from Millennium Development Goals Needs Assessment for 2012 onward. Nepal 2010. NPC and UNDP, Kathmandu, Nepal Appendix D Summary of Debt Sustainability Analyses Domestic public External public Countries Report date debt debt Remarks Debt projections Suggestions Afghanistan August 16, 2010 Almost all debts External debt is In the medium Public debt is projected to Risk of distress is high. are foreign; debt 8.1% of GDP term, external increase by about 108% Revenue from mining and has fallen signifi- at end-FY2010. debt is projected (Af 62 billion, US$1.3 taxes are able to cover cantly following Current account to be 9% by billion) from 2011 to 2016; only operation expense. debt relief pro- deficit of 1.7% of 2016 under almost all the increase Debt-to-GDP ratio is 8%. grams. Primary GDP is projected the baseline contributes to a rise in However, the withdrawal deficit is 1.1% of to turn into scenario. external debt (about of NATO troops scheduled GDP . Real GDP surplus and de- 103%, equivalent to Af 60 in 2014 would lead to a growth has been teriorate over the trillion, US$1.2 billion); rise potentially significant in- strong (22.5%). medium term. in domestic public debts crease in security spend- The economy is contributes to a relatively ing, which would make heavily depen- small part of the increase the deficit unsustainably dent on donor (Af 2 trillion, US$0.05 high. There is projected to support. billion). be minimal fiscal space, and debt is not likely to be sustainable unless the current trend is reversed. Bangladesh October 17, 2011 Domestic debt External debt is In the medium Public debt is projected Risk of debt distress is is 21.3% of GDP 21.6% of GDP term under the to increase by about low, but analysis reveals at end-FY2011. at end-FY2011. baseline sce- 70% (Tk 2 trillion, US$32 less favorable indicators Public and public- Current account nario, external billion) from 2012 to 2017; on public debt because of guaranteed surplus of 1.1% debt is projected rise in external public a continued low domestic external debt GDP is projected to be 18.1% by debt is projected to be revenue base. Strong had a face value to slip into a 2017, whereas about 32%, (Tk 0.5 tril- increase in remittances in equivalent to small current public debt will lion, US$7.8 billion); rise recent years and moder- 21.6% of GDP account deficit increase slightly in domestic public debt ate strong projection at end-FY2011. of about 0.4% of to 43.6%. Pri- is projected to be 105% in exports would allow Strong real GDP GDP . mary deficit will (Tk 1.5 trillion, US$24.6 more room for external growth reflects be 2.2% of GDP , billion). debt. The current level of stepped-up on par with the public debt is sensitive investment and debt-stabilizing to alternative scenarios, reforms. primary deficit. with progress on fiscal reforms, higher liabilities from infrastructure finance, and recapitaliza- tion of state-owned enter- prises. There is limited public debt space unless measures to mobilize do- mestic revenues become effective. (continued) South Asia Economic Focus – A Review of Economic Developments in South Asian Countries 83 84 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Domestic public External public Countries Report date debt debt Remarks Debt projections Suggestions Bhutan May 13, 2011 Domestic debt External debt In 2015, public Public debt is projected Bhutan is in moderate is 2.1% of GDP is 65.6% of debt is projected to increase by 270% (Nu risk of distress (with at end-2008/09. GDP . Mostly to be 111% 100 billion, US$1.8 billion) high external debt for A low-income foreign-currency- (baseline sce- from 2010 to 2015; rise the development of the country, Bhutan denominated nario) and 90.4% in external public debt hydropower sector). has strong GDP debts are for for external debt, is projected to be over Public debt is projected growth and a low the hydropower under the base- 200% (Nu 80 billion, to increase over time; by interest rate. sector. line scenario. US$1.3 billion); increase 2015, the debt level is in domestic debt is pro- projected to reach 111%. jected to be about Nu 26 Owing to strong growth billion (US$0.45 billion). (real GDP growth of 8% in medium term), rising energy demand from India, committed donor support, and high inter- national reserves, the high level of public debt is likely to be sustained. However, domestic financing has been small in comparison with external financing, which allows for possibilities of domestic debts. India December 6, Domestic debt is External debt is In 2015/16, Public debt is projected Current level of public 2010 72.5% of GDP at 4.2% of GDP . public debt is to increase by about debt is relatively high end-2010/11. Real Current account projected to be 100% (Rs 50 trillion, compared to other GDP growth is deficit is 4%. 67.8% according US$886 billion) from 2010 emerging economies. 7.7%; real inter- to the debt-sta- to 2015; rise in external With strong projected est rate is 4%, bilizing baseline public debt is projected growth (real GDP growth projected to be scenario. to be about 90% (Rs 2.5 over 8%) and low inter- low till 2012; and trillion, US$43 billion); rise est rates, the space for primary deficit in domestic public debt domestic debt exists but is 5%. amounts to 102% (Rs 48 is quite limited. However, trillion, US$844 billion). the external public debt remains contained (19% of GDP). Given the high average growth and low external debt, even a combination of adverse shocks would still leave external debt lower than 30% of GDP , which may allow for possibilities of external debt. Appendix D. Summary of Debt Sustainability Analyses 85 Domestic public External public Countries Report date debt debt Remarks Debt projections Suggestions Maldives January 24, 2011 Domestic debt External debt is In 2015, public Public debt is projected Maldives is in high risk of is 43.8% of GDP 41.8% of GDP at debt is projected to increase by 170% distress. Under current at end 2010. The end-2009, similar to be 137% (Rf 35 billion, US$1.9 policies, public debt is economy is pro- to the amount (baseline billion) from 2010 to projected to rise and jected to recover. of domestic scenario) and 2015; rise in external reach 306% of GDP by Fiscal deficit is public debt, 125.8% for public debt is projected 2030. The negative effect 18% of GDP , which makes the external debt, to be about 118% (Rf 13 of the economic crisis and there are economy vulner- under the base- billion, US$0.7 billion), on exports and tourism high real interest able to capital line scenario. with an even higher rise receipts, combined with rates. flow sudden in domestic public debt excessive government stops. Current of 205% (Rf 22 billion, spending, would lead account deficit is US$1.2 million). to large domestic and 30% of GDP . external financing require- ments (primary deficit projected to be 15.4%). The current debt path is unsustainable, and in the medium term, there will be minimal room for fur- ther debt, unless further fiscal consolidation could reverse this trend. Nepal September 26, Domestic debt External debt is In 2015, public Public debt is projected Nepal is in moderate risk 2011 is 13.5% of GDP 22.4% of GDP , debt is projected to increase by about 18% of distress (with risks of at end-2010. mostly from to be 33.2% (Rs 333 billion, US$4.8 large fiscal spending to A low-income World Bank and (baseline sce- billion) from 2011 to 2016; address financial sector country, Nepal Asian Develop- nario). The pri- moderate rise in external vulnerabilities). The net has moderate ment Bank, with mary deficit will public debt is projected domestic borrowing in growth and low small appre- be 1.9% of GDP , to be 18% (Rs 48 billion, the immediate future (negative) real ciation projected higher than the US$0.8 billion); domestic rises only slightly from interest rates. and strong remit- debt-stabilizing public debt is projected to 2.4% to 2.7% of GDP , tance flows. primary balance rise about 180% (Rs 285 and in the long term, (baseline). billion, US$4 billion). net domestic borrow- ing remains at 2.5% of GDP . Primary expendi- ture seems suppressed throughout the projection period. Overall, external debt is projected to decline over time, and only a slight increase in domestic debt is al- lowed (this to offset the low revenue increase projected in this version of the DSA). (continued) 86 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Domestic public External public Countries Report date debt debt Remarks Debt projections Suggestions Pakistan July 10, 2011 Domestic debt External debt is In 2015/16, total Public debt is projected With improving projec- is 31.7% of GDP 29.8% of GDP , public debt is to increase by 65% by tions of growth, from at end-2009/10, which makes expected to be 2015/16, nearly Rs 6 tril- 2.8% for 2011 to 5.0% by and it has been up most of the 43.6% (baseline lion (US$23.4 billion) debt 2015, the current public stable over the external debt. scenario) and level. Rise in external debt level is likely to be past few years. Currency has 47.5% (no policy public debt is projected sustained. The current Real interest rate depreciated, and change scenar- to be small at 1.4% Rs account has improved; low is negative. current account io). The primary 61 billion, US$13 billion) however, with depreciat- Real GDP growth balance has balance will be while domestic public ing currency, the current is weak, and been declining 2% of GDP debt will contribute to level of external debt there is a small higher than the most of the increase of tends to be vulnerable. surplus. debt-stabilizing public debt at 124% (Rs Domestic debt is project- primary balance 5.8 trillion, US$37 billion). ed to decline over time. under the base- However, DSA projec- line scenario. tions permit only a small increase for external debt because more external exposure would increase vulnerability to exchange rate shocks. Sri Lanka September 13, Domestic debt is External debt is In 2015, public Public debt is projected Although strong econom- 2010 36.4% of GDP at 49.6% of GDP . debt is expected to increase by 62% from ic growth (close to 8%) end-2009. Real The exchange to be 63.7% 2010 to 2015 (SL Rs 2.6 allows the country to is- interest rate is rate has depreci- (baseline sce- trillion, US$17.6 billion). sue more domestic debts low (negative), ated sharply, nario) and 71.1% Rise in external public than in the past, the and GDP growth and the current (no policy change debt is projected to be overall public debt level is is very strong. account balance scenario). The 65% (more than SL Rs already very high (more The deficit is has widened to primary balance 1.1 trillion, US$7.6 billion), than 80% in the past few about 3%. 7.8% of GDP will be 2.6% of with a similar increase in years). Fiscal deficit was owing to a GDP , higher than domestic public debt of nearly 10% in 2009 but huge increase the debt-stabi- 63% (SL Rs 1.5 trillion, was down to 8% in 2010. in trade deficit, lizing primary US$10 billion). Moreover, the foreign which makes the balance (baseline component of debt is economy vulner- scenario). rising at a considerable able to sudden pace. More than half of stops. the government debt is foreign debt; much of Sri Lanka’s US$5.5 billion gross official reserves are borrowed; and the government is borrow- ing from offshore, which could lead to much vulnerability in terms of exchange rate risk and bunching of repayments. Overall, there is domestic debt space, but it is limited, and raising more external debt would make the economy more vulnerable. Source: World Bank staff calculations based on International Development Association–International Monetary Fund debt sustainability analyses. Note: DSA = debt sustainability analysis; FY = fiscal year; GDP = gross domestic product; NATO = North Atlantic Treaty Organization. Appendix E Fiscal Trends Figure E.1  Fiscal Trends, Afghanistan (percentage of GDP) 20 15 10 5 0 –5 –10 –15 FY2003/04 FY2004/05 FY2005/06 FY2006/07 FY2007/08 FY2008/09 FY2009/10 FY2010/11 Capital expenditure Current expenditure Deficit before grants Total revenue Sources: Ministry of Finance and World Bank database. Note: FY = fiscal year; GDP = gross domestic product. Figure E.2  Long-term Fiscal Trends, Bangladesh (percentage of GDP) 15 12 9 6 3 0 91 FY 1 FY 4 FY 0 92 FY 2 FY 9 FY 9 FY 7 FY 3 im 11 98 FY 8 FY 6 06 FY 4 0 97 95 FY 5 10 93 e) 0 9 0 9 0 0 0 0 0 9 0 9 0 st 0 19 20 19 19 19 20 20 20 20 19 20 19 20 19 20 20 19 20 19 19 at 20 (e FY2 FY FY FY FY FY FY FY FY FY Capital expenditure Current expenditure Deficit before grants Total revenue Sources: Ministry of Finance and World Bank database. Note: FY = fiscal year (July to June); GDP = gross domestic product. South Asia Economic Focus – A Review of Economic Developments in South Asian Countries 87 88 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Figure E.3  Fiscal Trends, Bhutan (percentage of GDP) 25 20 15 10 5 0 –5 –10 –15 –20 FY2002/03 FY2003/04 FY2004/05 FY2005/06 FY2006/07 FY2007/08 FY2008/09 FY2009/10 Capital expenditure Current expenditure Deficit after grants Defciit before grants Total revenue Sources: Ministry of Finance and World Bank database. Note: FY = fiscal year; GDP = gross domestic product. Figure E.4  Long-term Fiscal Trends, India (percentage of GDP) 30 25 20 15 10 5 0 –5 –10 –15 20 2 20 0 0 19 86 99 9 20 03 19 2 20 4 20 1 19 0 20 9 19 3 FY 199 8 20 07 19 91 19 6 20 05 19 9 19 94 20 8 0 19 7 19 87 20 6 19 5 19 8 FY 1/0 FY 0/0 19 8/9 FY 1/9 FY 3/0 FY 9/9 FY 8/0 /1 FY 2/9 FY 7/9 FY 5/9 FY 8/8 FY 7/0 FY /9 FY 5/0 FY 4/9 FY 7/8 5/ FY 2/ FY /20 FY 06/ FY 0/ FY 4/ / FY 86/ 09 93 96 0 8 0 9 0 0 8 0 9 9 9 9 0 8 0 0 9 8 19 FY FY FY Capital expenditure Current expenditure Deficit before grants Total revenue Sources: Ministry of Finance and World Bank database. Note: FY = fiscal year; GDP = gross domestic product. Fiscal Trends 89 Figure E.5  Long-term Fiscal Trends, Maldives (percentage of GDP) 40 35 30 25 20 15 10 5 0 –5 –10 –15 –20 –25 FY1994 FY1995 FY1996 FY1997 FY1998 FY1999 FY2000 FY2001 FY2002 FY2003 FY2004 FY2005 FY2006 FY2007 FY2008 FY2009 FY2010 Capital expenditure Current expenditure Deficit before grants Total revenue Sources: Ministry of Finance and World Bank database. Note: GDP = gross domestic product. Figure E.6  Fiscal Trends, Nepal (percentage of GDP) 20 15 10 5 0 –5 –10 FY1999 FY2000 FY2001 FY2002 FY2003 FY2004 FY2005 FY2006 FY2007 FY2008 FY2009 FY2010 (estimate) Capital formation Current expenditure Deficit before grants Total revenue Sources: Ministry of Finance and World Bank database. Note: FY = fiscal year; GDP = gross domestic product. 90 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Figure E.7  Fiscal Trends, Pakistan (percentage of GDP) 20 15 10 5 0 –5 –10 5 3 11 1 00 2 0 8 4 9 6 02 7 /0 /0 /1 /0 /1 /0 /0 /0 /0 /0 / 20 10 / 00 04 09 02 11 07 03 08 05 01 06 20 9/ 20 20 20 20 20 20 20 20 20 20 20 9 FY FY FY FY FY FY 19 FY FY FY FY FY FY FY Capital expenditure Current expenditure Deficit before grants Total revenue Sources: Ministry of Finance and World Bank database. Note: FY = fiscal year; GDP = gross domestic product. Figure E.8  Long-term Fiscal Trends, Sri Lanka (percentage of GDP) 25 20 15 10 5 0 –5 –10 –15 1 9 10 3 01 02 0 07 5 04 2 08 6 05 09 7 06 4 00 03 8 9 9 9 9 9 9 9 9 9 9 20 19 19 20 19 19 20 19 19 19 20 20 19 20 19 20 19 20 20 20 20 FY FY FY FY FY FY FY FY FY FY FY FY FY FY FY FY FY FY FY FY FY Capital expenditure Current expenditure Deficit before grants Total revenue Sources: Ministry of Finance and World Bank database. Note: FY = fiscal year; GDP = gross domestic product. Appendix F Trends of Nontax Revenue Figure F.1  Trends of Nontax Revenue, Afghanistan (NTR in total revenue) 40 35 30 25 20 15 10 y = –0.071ln(x) + 0.3627 R² = 0.59133 5 0 FY2003/04 FY2004/05 FY2005/06 FY2006/07 FY2007/08 FY2008/09 FY2009/10 FY2010/11 Source: World Bank databases. Note: FY = fiscal year; GDP = gross domestic product. Figure F.2  Trends of Nontax Revenue, Bangladesh (NTR in total revenue) 21.0 20.5 20.0 19.5 19.0 y = 2E–05x² – 0.0003x + 0.1991 18.5 R² = 0.00286 18.0 FY1997 FY1998 FY1999 FY2000 FY2001 FY2002 FY2003 FY2004 FY2005 FY2006 FY2007 FY2008 FY2009 Source: World Bank databases. Note: FY = fiscal year; GDP = gross domestic product. South Asia Economic Focus – A Review of Economic Developments in South Asian Countries 91 92 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Figure F.3  Trends of Nontax Revenue, Bhutan (NTR in total revenue) 60 50 40 30 20 y = –2E–05x² – 0.0003x + 0.1828 10 R² = 0.35094 0 FY2001/02 FY2002/03 FY2003/04 FY2004/05 FY2005/06 FY2006/07 FY2007/08 FY2008/09 FY2009/10 FY2010/11 Source: World Bank databases. Note: FY = fiscal year; GDP = gross domestic product. Figure F.4  Trends of Nontax Revenue, India (NTR in total revenue) 22 20 18 16 14 y = –2E–05x² – 0.0003x + 0.1828 R² = 0.35094 12 10 1 20 1 20 0 19 88 20 02 99 9 19 9 20 08 20 3 0 19 97 19 5 19 0 20 9 20 04 19 3 20 07 19 6 19 6 20 05 19 4 19 2 19 7 20 6 FY 199 8 FY 0/9 FY 0/0 /1 FY 200 19 8/9 FY 8/8 FY 2/0 FY 4/9 FY 9/9 FY 8/0 FY 2/9 FY 5/9 FY 5/8 FY 3/9 FY 1/9 FY 6/8 FY 5/0 FY 7/9 FY 87/ FY 01/ FY 07/ FY 96/ / FY 06/ FY 04/ 09 03 9 0 8 0 9 8 0 9 9 8 9 9 8 0 9 / 19 19 FY FY Source: World Bank databases. Note: FY = fiscal year; GDP = gross domestic product. Trends of Nontax Revenue 93 Figure F.5  Trends of Nontax Revenue, Maldives (NTR in total revenue) 60 55 50 45 40 y = –0.0018x² + 0.0113x + 0.5348 R² = 0.71626 35 30 FY2001 FY2002 FY2003 FY2004 FY2005 FY2006 FY2007 FY2008 FY2009 FY2010 FY2011 Source: World Bank databases. Note: FY = fiscal year; GDP = gross domestic product. Figure F.6  Trends of Nontax Revenue, Nepal (NTR in total revenue) 24 22 20 18 16 14 y = –0.0009x² + 0.014x + 0.1432 R² = 0.63691 12 10 8 at 1 3 9 1 0 2 9 8 95 0 4 6 6 7 5 7 /9 im /1 /0 /0 /0 /1 /0 /9 /0 00 /0 /9 /0 /0 /0 /9 / e) 97 st 0 02 08 00 09 01 98 07 94 03 95 05 06 04 96 /2 (e 01 19 20 20 20 20 20 19 20 19 99 20 19 20 20 20 19 2 FY FY FY FY FY FY FY FY FY FY 19 FY FY FY FY FY FY FY Source: World Bank databases. Note: FY = fiscal year; GDP = gross domestic product. 94 South Asia Economic Focus – A Review of Economic Developments in South Asian Countries Figure F.7  Trends of Nontax Revenue, Pakistan (NTR in total revenue) 31 29 27 25 23 21 19 y = –0.0016x² + 0.027x + 0.1662 17 R² = 0.68499 15 5 3 11 1 00 8 6 0 02 7 12 4 9 /0 /0 /1 /0 /0 /0 /0 /0 /0 / / 20 / 10 00 04 09 02 07 05 01 06 11 03 08 20 9/ 20 20 20 20 20 20 20 20 20 20 20 9 FY FY FY FY FY FY FY FY FY 19 FY FY FY FY Source: World Bank databases. Note: FY = fiscal year; GDP = gross domestic product. Figure F.8  Trends of Nontax Revenue, Sri Lanka (NTR in total revenue) 18 16 14 12 10 8 y = –0.0004x² + 0.0085x + 0.0881 R² = 0.36292 6 01 04 9 3 09 06 4 00 8 1 07 2 08 6 05 10 02 7 0 03 5 9 9 9 9 9 9 9 9 9 9 20 19 19 20 20 19 19 20 19 20 19 19 20 20 19 20 19 20 20 19 20 FY FY FY FY FY FY FY FY FY FY FY FY FY FY FY FY FY FY FY FY FY Source: World Bank databases. Note: FY = fiscal year; GDP = gross domestic product. THE WORLD BANK 1818 H Street, N.W. Washington, DC 20433