Report No. 50078-PK Pakistan Tax Policy Report Tapping Tax Bases for Development (In Two Volumes) Volume II: Full Report July 2009 The World Bank The Government of Pakistan Federal Board of Revenue Georgia State University Document of the World Bank TABLE OF CONTENTS Chapter 1: Tapping Tax Bases for Development ................................................................................................... 1 1.1 Introduction ..................................................................................................................................................... 1 1.2 Spending for Results ....................................................................................................................................... 2 1.3 Unsustainable Fiscal Deficits and Unreliable Non-Tax Revenues ................................................................. 3 1.4 Tapping Tax Bases .......................................................................................................................................... 5 1.5 Scope ............................................................................................................................................................... 7 1.6 Process ............................................................................................................................................................ 8 Chapter 2: National Tax Policy Reform ................................................................................................................. 9 2.1 Introduction ..................................................................................................................................................... 9 2.2 General Sales Tax ......................................................................................................................................... 11 2.3 Corporate Income Tax................................................................................................................................... 13 2.4 Individual Income Tax .................................................................................................................................. 15 2.5 Federal Excise Duties ................................................................................................................................... 19 2.6 Customs Duties ............................................................................................................................................. 20 2.7 Provincial Taxation ....................................................................................................................................... 21 2.8 Way Forward................................................................................................................................................. 22 2.9 What Next? ................................................................................................................................................... 29 Chapter 3: Structural Properties ........................................................................................................................... 30 3.1 Introduction ................................................................................................................................................... 30 3.2 Adequacy ...................................................................................................................................................... 32 3.3 Efficiency ...................................................................................................................................................... 40 3.4 Equity ............................................................................................................................................................ 43 3.5 Compliance ................................................................................................................................................... 49 3.6 Policy Reform ............................................................................................................................................... 52 Chapter 4: General Sales Tax ................................................................................................................................ 53 4.1 Introduction ................................................................................................................................................... 53 4.2 Non-Neutrality .............................................................................................................................................. 55 4.3 Weak Legal Vessel ........................................................................................................................................ 61 4.4 Equity ............................................................................................................................................................ 62 4.5 Compliance ................................................................................................................................................... 63 4.6 Policy Options ............................................................................................................................................... 67 Chapter 5: Corporate Income Tax ........................................................................................................................ 68 5.1 Introduction ................................................................................................................................................... 68 5.2 Tax Design and Revenue Trends .................................................................................................................. 70 5.3 Tax Incentives ............................................................................................................................................... 73 5.4 Withholding Taxes ........................................................................................................................................ 76 5.5 Small Companies .......................................................................................................................................... 80 5.6 International Tax Provisions ......................................................................................................................... 81 5.7 Equity ............................................................................................................................................................ 83 5.8 Compliance ................................................................................................................................................... 84 5.9 Policy Options ............................................................................................................................................... 87 Chapter 6: Individual Income Tax ........................................................................................................................ 89 6.1 Introduction ................................................................................................................................................... 89 6.2 Towards a Global Income Tax ...................................................................................................................... 92 6.3 Capital Gains ................................................................................................................................................. 93 i 6.4 Schedules ...................................................................................................................................................... 96 6.5 Tax Credits .................................................................................................................................................... 97 6.6 Equity ............................................................................................................................................................ 98 6.7 Compliance ................................................................................................................................................... 99 6.8 Policy Options ............................................................................................................................................. 101 Chapter 7: Federal Excise Duties ........................................................................................................................ 103 7.1 Introduction ................................................................................................................................................. 103 7.2 Legal Framework ........................................................................................................................................ 105 7.3 Equity and Compliance ............................................................................................................................... 108 7.4 Policy Options ............................................................................................................................................. 109 Chapter 8: Custom Duties .................................................................................................................................... 110 8.1 Introduction ................................................................................................................................................. 110 8.2 Equity and Compliance ............................................................................................................................... 113 8.3 Policy Options ............................................................................................................................................. 114 Chapter 9: Provincial Taxes ................................................................................................................................. 115 9.1 Introduction ................................................................................................................................................. 115 9.2 Tapping Provincial Tax Bases .................................................................................................................... 117 9.3 Sales Tax on Services.................................................................................................................................. 121 9.4 Motor Vehicle Taxes ................................................................................................................................... 126 9.5 Urban Immovable Property Tax.................................................................................................................. 131 9.6 Land and Property Transfer Taxes .............................................................................................................. 139 9.7 Agricultural Income Tax ............................................................................................................................. 144 9.8 Professional Tax .......................................................................................................................................... 149 9.9 Other Taxes ................................................................................................................................................. 151 9.10 Vertical Equity ............................................................................................................................................ 152 9.11 Policy Reform ............................................................................................................................................. 153 FIGURES Figure 1.1: Economic and Fiscal Indicators .............................................................................................................. 1 Figure 1.2: Public Spending and Development Outcomes ........................................................................................ 2 Figure 1.3: Pakistan's Tax and Non-Tax Revenues .................................................................................................. 4 Figure 1.4: Turning Around the Tax-to-GDP Ratio .................................................................................................. 5 Figure 1.5: Tax Collection in Pakistan and other Developing Countries .................................................................. 6 Figure 1.6: Navigational Aid for the Report.............................................................................................................. 7 Figure 2.1: Revenue Impact of GST Broadening .................................................................................................... 12 Figure 2.2: Revenue Impact of Corporate Income Tax Rate Reduction ................................................................. 14 Figure 2.3: Revenue Impact of Individual Income Tax Reform under Current Compliance .................................. 17 Figure 2.4: Revenue Impact of Federal Excise Tax Reform ................................................................................... 19 Figure 2.5: Revenue Impact of Custom Duties Reform .......................................................................................... 20 Figure 2.6: Revenue Impact of Provincial Tax Reform .......................................................................................... 21 Figure2.7: Revenue Impact of National Tax Policy Reform ­ Display One .......................................................... 23 Figure 2.8: Revenue Impact of National Tax Policy Reform ­ Display Two ......................................................... 23 Figure 2.9: Impact of National Tax Reform on Vertical Equity.............................................................................. 23 Figure 2.10: Raising Tax Revenues: Actual Collection, Additional Collection Target, Tax Policy Reform Impact, and Tax Gap .......................................................................................................................................... 24 Figure 2.11: National Tax Reform ­ Size of Tax Policy Impact and Tax Gap for each Tax .................................... 26 Figure 3.1: Inadequate Revenue - Factors behind the Low Tax-to-GDP Ratio ...................................................... 32 Figure 3.2: Contributions of the Three Sectors to GDP and FBR Taxes in 2006/07 .............................................. 33 Figure 3.3: Contributions of Subsectors, Commodities and Heads to FBR Taxes .................................................. 33 Figure 3.4: Long-Term Buoyancy and Stability of FBR Taxes .............................................................................. 34 Figure 3.5: Composition of Federal Tax Collection since 1999/2000..................................................................... 35 ii Figure3.6: Pakistan's Tax Expenditures ................................................................................................................. 37 Figure 3.7: Two Indicators of FBR Tax Collection Performance ........................................................................... 38 Figure3.8: Provincial Tax Collection ..................................................................................................................... 39 Figure 3.9: Pakistan's Doing Business 2008 Rankings ........................................................................................... 42 Figure 3.10: Pakistan's Doing Business 2008 Ranking for Paying Taxes ................................................................ 42 Figure 3.11: Vertical Tax Incidence in 2006/07 ........................................................................................................ 46 Figure 3.12: Horizontal Inequities with Household Consumption Deciles ............................................................... 48 Figure 3.13: Pakistan's Federal Tax Gap (Percent of Actual Tax Collection) .......................................................... 50 Figure 4.1: Pakistan's GST Collection over Time and in International Comparison .............................................. 53 Figure 4.2: Pakistan's GST Productivity in International Comparison ................................................................... 59 Figure 4.3: Contributions of the Three Sectors to GDP and FBR Indirect Taxes in 2006/07 ................................. 60 Figure 4.4: Contributions of Subsectors to FBR Indirect Taxes relative to GDP ................................................... 60 Figure 4.5: Commodity Composition of Gross Collection from GST .................................................................... 60 Figure 4.6: Distribution of Tax Burden and Consumption by Household Expenditure Decile............................... 62 Figure 4.7: Compliance for General Sales Tax ....................................................................................................... 63 Figure 4.8: Distribution of Registered GST Taxpayers by Turnover: Number of Businesses and GST Contributions......................................................................................................................................... 64 Figure 4.9: Return Filing of Registered GST Taxpayers by Sector ........................................................................ 64 Figure 5.1: Pakistan's Corporate Income Tax in International Comparison ........................................................... 69 Figure 5.2: International Comparison of Corporate Income Tax Productivity........................................................ 71 Figure 5.3: Contributions of the Three Sectors to GDP and FBR Direct Taxes in 2006/07 ................................... 71 Figure 5.4: Collection of Corporate Income Tax by Subsectors ............................................................................. 71 Figure 5.5: 2006 Average and Marginal Effective Tax Rates by Industry .............................................................. 74 Figure 5.6: Withholding Tax Collection ................................................................................................................. 77 Figure 5.7: Vertical Incidence of Corporate Income Tax in 2006/07...................................................................... 83 Figure 5.8: Tax Gap of Corporate Income Tax in 2004/05 ..................................................................................... 84 Figure 5.9: Voluntary Payments and Corporate Income Tax Filing ....................................................................... 85 Figure 5.10: Tax Return Filing and Refund Payment for Corporate Income Tax ..................................................... 86 Figure 6.1: Collection of Individual Income Tax .................................................................................................... 90 Figure 6.2: The Notch Problem ............................................................................................................................... 95 Figure 6.3: Incidence of Individual Income Tax in 2006/07 for Salaried and Non-Salaried Individuals ............... 98 Figure 6.4: Overall Incidence of Individual Income Tax in 2006/07 ...................................................................... 98 Figure 6.5: Individual Income Tax Compliance .................................................................................................... 100 Figure 6.6: Trends in Individual Income Tax Filing ............................................................................................. 100 Figure 7.1: Pakistan's Federal Excise Tax Collection Over Time and in International Comparison.................... 103 Figure 7.2: Composition of Federal Excise Tax Collection .................................................................................. 106 Figure 7.3: Equity and Compliance for Federal Excise Taxation ......................................................................... 108 Figure 8.1: Composition of Custom Duty Collection............................................................................................ 110 Figure 8.2: Pakistan's Custom Duty Collection over Time and in International Comparison .............................. 111 Figure 8.3: Equity and Compliance for Custom Duties ........................................................................................ 113 Figure 9.1: Provincial Revenue Composition and Trends ..................................................................................... 115 Figure 9.2: Economic Indicators and Budgetary Position in Punjab and NWFP .................................................. 117 Figure 9.3: Composition of Tax Revenue in Punjab and NWFP .......................................................................... 118 Figure 9.4: Revenue Composition in Punjab and NWFP ...................................................................................... 119 Figure 9.5: Government Spending and Own-Source Revenues in Punjab and NWFP ......................................... 119 Figure 9.6: Sales Tax Collection on Services in Punjab and NWFP in 2005/06................................................... 121 Figure 9.7: Provincial Composition of Sales Tax Collection from Service in 2005/06 ........................................ 121 Figure 9.8: Level and Trends in Motor Vehicle Tax Collection ........................................................................... 126 Figure 9.9: UIPT Collection in Punjab and NWFP in International Comparison ................................................. 131 Figure 9.10: Trends in Land and Property Transfer Collection in Punjab and NWFP ........................................... 139 Figure 9.11: Trends in Agricultural Income Tax Collection ................................................................................... 144 Figure 9.12: Estimated Additional Potential for AIT Collection in 2000/01 based on the 2000 Agricultural Census ............................................................................................................................................................ 145 Figure 9.13: Trends in Professional Tax Collection ............................................................................................... 150 Figure 9.14: Collection of Other Taxes in Punjab and NWFP in 2005/06? ............................................................ 151 iii Figure 9.15: Vertical Equity of Provincial Taxes (2004/05) ................................................................................... 152 Figure 9.16: Revenue Targets for Provincial Tax Collection (% of 2005/06 Provincial Tax Collection) .............. 153 Figure 9.17: Overall Indicative Revenue Impact of Comprehensive Tax Policy Reform in Punjab and NWFP .... 154 Figure 9.18: Indicative Revenue Estimation for Comprehensive Tax Policy Reform in Punjab and NWFP ......... 155 TABLES Table 2.1: National Tax Reform ­ Simulations......................................................................................................... 22 Table 2.2: Sequencing Tax Policy Reform Across Taxes and Issues ....................................................................... 25 Table 2.3: Analytical Preparation of Tax Policy Reform .......................................................................................... 26 Table 2.4: Piecemeal and Comprehensive Tax Policy Matrix .................................................................................. 27 Table 2.5: Sequencing Tax Policy Reform for each Taxes ....................................................................................... 28 Table 3.1: Main Findings of Chapter 3 ..................................................................................................................... 31 Table 3.2: Summary of Tax Incidence Assumptions ................................................................................................ 44 Table 4.1: Main Findings of Chapter 4 ..................................................................................................................... 54 Table 4.2: Policy Reform Options for General Sales Tax ......................................................................................... 67 Table 5.1: Main Findings of Chapter 5 ..................................................................................................................... 69 Table 5.2: Summary of Withholding Tax Regime .................................................................................................... 79 Table 5.3: Pakistan's Double Taxation Treaties........................................................................................................ 81 Table 5.4: Policy Reform Options for Corporate Income Tax .................................................................................. 88 Table 6.1: Main Findings of Chapter 6 ..................................................................................................................... 91 Table 6.2: Taxation of Capital Gains and Dividends in Emerging Economies......................................................... 95 Table 6.3: Policy Reform Options for Individual Income Tax ............................................................................... 102 Table 7.1: Main Findings of Chapter 7 ................................................................................................................... 104 Table 7.2: Policy Reform Options for Federal Excise Tax ..................................................................................... 109 Table 8.1: Main Findings of Chapter VIII .............................................................................................................. 111 Table 8.2: Policy Reform Options for Custom Duties ............................................................................................ 114 Table 9.1: Major Findings of Chapter 9 .................................................................................................................. 116 Table 9.2: Federal, Provincial and Local Revenues ................................................................................................ 119 Table 9.3: Revenue Potential of a Motor Fuel Tax ................................................................................................. 129 Table 9.4: Illustrative Property Tax Reform Program for Punjab ........................................................................... 137 Table 9.5: Illustrative Property Tax Reform Program for NWFP ........................................................................... 137 Table 9.6: Incidence Assumption for Provincial Taxes .......................................................................................... 152 Table 9.7: Piecemeal and Comprehensive Policy Reform Matrix .......................................................................... 159 Table 9.8: Sequencing of Tax Policy Reform ......................................................................................................... 159 BOXES Box 1.1: Basic Services ­ A Public Responsibility .................................................................................................. 1 Box 1.2: Gas and Petroleum Surcharges and Petroleum Development Levy ........................................................... 4 Box 2.1: Egypt's Tax Reforms ............................................................................................................................... 10 Box 2.2: General Sales Tax Reform ­ Broadening the Base .................................................................................. 12 Box 2.3: Corporate Income Tax Reform ­ Bridging Structure ............................................................................... 14 Box 2.4: Corporate Income Tax Reform ­ Two Tier Rate Structure ...................................................................... 14 Box 2.5: Modeling Corporate Income Tax ............................................................................................................. 14 Box 2.6: Individual Income Tax Reform ­ Flat Rate Tax ...................................................................................... 17 Box 2.7: Interpreting the Left Panel of Figure IX.1 ................................................................................................ 17 Box 2.8: Individual Income Tax Reform ­ Two-Tier Rate Structure ..................................................................... 17 Box 2.9: Modeling Individual Income Tax ............................................................................................................. 18 Box 2.10: Modeling General Sales Tax .................................................................................................................... 19 Box 2.11: Federal Excise Tax Reform ...................................................................................................................... 19 Box 2.12: Customs Duties Reform ­ Three-Tier Structure....................................................................................... 20 Box 2.13: Modeling Custom Tariffs ......................................................................................................................... 20 Box 3.1: Brief Overview of Pakistan's Tax System ............................................................................................... 30 iv Box 3.2: Why is Fiscal Policy Procyclical in Developing Countries? .................................................................... 36 Box 3.3: Pakistan's Tax Revenues Compared to OECD and non-OECD Countries .............................................. 36 Box 3.4: The Excess Burden of Taxation ............................................................................................................... 40 Box 3.5: Doing Business 2008................................................................................................................................ 42 Box 3.6: Inequality in Pakistan ............................................................................................................................... 44 Box 3.7: Statutory and Economic Incidence ........................................................................................................... 45 Box 3.8: Tax Incidence Studies on Pakistan ........................................................................................................... 47 Box 3.9: Compliance and Equity ............................................................................................................................ 51 Box 4.1: Taxing Consumption or Income? ............................................................................................................. 55 Box 4.2: External and Internal Neutrality ............................................................................................................... 56 Box 4.3: The "Zero-rating" of the Five Export Sectors .......................................................................................... 57 Box 4.4: Sales Tax Exemptions .............................................................................................................................. 57 Box 4.5: VAT Refund ............................................................................................................................................. 57 Box 4.6: Crediting, Zero-rating, and Exemptions ................................................................................................... 58 Box 4.7: Blacklisting of Businesses ........................................................................................................................ 65 Box 4.8: The Many Types of VAT Evasion and Fraud .......................................................................................... 66 Box 5.1: Should State Bank's Profits be Subject to Income Tax? .......................................................................... 72 Box 5.2: Pakistan's Tax Incentives ......................................................................................................................... 74 Box 5.3: Lessons from International Experience with Corporate Income Taxation ............................................... 75 Box 5.4: Problems of Simplified Tax Systems ....................................................................................................... 78 Box 5.5: Attributing Withholding Tax Collection .................................................................................................. 78 Box 5.6: Three Types of Income Tax Collection.................................................................................................... 85 Box 5.7: Making Sense of the Decline in Voluntary Payments in 2007/08............................................................ 86 Box 6.1: Capital Value Tax .................................................................................................................................... 90 Box 6.2: Social Security Taxes ............................................................................................................................... 90 Box 6.3: Estimating the Revenue Impact of Taxing Stock Market-Related Short-Term Capital Gains ................ 94 Box 6.4: Estimating the Revenue Impact of Individual Income Tax Reforms ..................................................... 102 Box 7.1: Defining Excise Tax Rates ­ In Rem versus Ad Valorem ..................................................................... 106 Box 7.2: Excise Taxation and Non-Essential Consumption ................................................................................. 107 Box 7.3: Reforming Excise Taxation on Tobacco ................................................................................................ 109 Box 8.1: Liberalizing Trade Without Lowering Tax Revenue ............................................................................. 112 Box 8.2: Pakistan's Custom Duties Regime ......................................................................................................... 112 Box 9.1: Provincial GDP Estimates ...................................................................................................................... 117 Box 9.2: Incidence of Sales Tax on Services ........................................................................................................ 125 Box 9.3: Expanding the Provincial Sales Tax Base .............................................................................................. 125 Box 9.4: Place of Supply and the Taxation of Services ........................................................................................ 125 Box 9.5: Overcoming Impediments to the Adoption of a Provincial Motor Fuel Tax ......................................... 130 Box 9.6: Medium Term Reforms of the Urban Immovable Property Tax in Punjab............................................ 133 Box 9.7: Exemptions and Preferential Reductions of UIPT in Punjab and NWFP .............................................. 133 Box 9.8: Estimates of the Revenue Cost of Preferential Treatments in Punjab and NWFP ................................. 133 Box 9.9: Mutations and Registrations ................................................................................................................... 141 Box 9.10: The Patwari............................................................................................................................................. 141 Box 9.11: International Experience with Land and Property Transfer Taxes ......................................................... 143 Box 9.12: Tackling the Problem of Undervaluation ............................................................................................... 143 Box 9.13: Changes in the Net Taxation of Agriculture ........................................................................................... 146 Box 9.14: Revenue Impact of Three Scenarios for the Reform of Agricultural Income Tax ................................. 148 Box 9.15: Setting Provincial Tax Revenue Targets ................................................................................................ 155 Box 9.16: The National Finance Commission Award ............................................................................................ 156 References.............................................................................................................160 ******** v ACKNOWLEDGEMENTS This report is a joint product of a team from the Federal Board of Revenue (FBR), Government of Pakistan, the Andrew Young School of Public Policy (AYSPS) at the Georgia State University, and the World Bank. The views, expressed in the report, are of the authors and not of the Government of Pakistan. The FBR team was headed by Mr. Ahmad Waqar, Secretary Revenue Division and Chairman FBR, and Mr. M. Abdullah Yusuf, former Chairman FBR, and included Mr. Mumtaz Haider Rizvi, Member Fiscal Research and Statistics, Dr. Ather Maqsood Ahmed, former Member Fiscal Research and Statistics, Mrs. Robina Ather Ahmed, Chief Fiscal Research and Statistics, Mr. Umar Wahid, Secretary Fiscal Research and Statistics, Mr. Mir Ahmed Khan, Second Secretary Fiscal Research and Statistics, and Mr. Naeem Ahmed, Second Secretary Fiscal Research and Statistics. The report was prepared by Kaspar Richter (World Bank), and Jorge Martinez-Vazquez (AYSPS). It is based on seven background studies (listed below), drafted by Robina Ather Ahmed, James Alm, Roy Bahl, Musharraf Cyan, Mir Ahmad Khan, Jorge Martinez-Vazquez, Geerten Michelse, Mark Rider, Wayne Thirsk, Umar Wahid and Sally Wallace. The tax revenue simulation results in the report are based on micro-simulation models developed by Mark Rider and Sally Wallace with Harini Kannan. The GST chapter draws extensively on a review of Pakistan's sales tax by Rebecca Millar and Christophe Waerzeggers from June 2008. The report also benefited from the Pakistan tax administration review by Carlos Silvani, Edmund Biber, William Crandall, Wyatt Grant, Orlando Reos and Geoff Seymour from September 2008. Peer reviewer comments from Kai-Alexander Kaiser, Senior Economist, World Bank; Michael Keen, Advisor, International Monetary Fund; Dr. Ahmad Khan, former Member FBR; Russell Krelove, Senior Economist, International Monetary Fund, and Eduardo Ley, Lead Economist, World Bank greatly enhanced the quality of the report. Dr. Ahmad Khan and Dr. A. R. Kemal reviewed the background studies, and Ehtisham Ahmad commented on the concept paper. Anjum Ahmad, Shamsuddin Admad, Mihaly Kopanyi, Hanid Mukhtar, and Saadia Refaqat from the World Bank provided useful feedback. Mirafe Marcos helped greatly by providing the draft chapter of the provincial background study. The team would like to thank Satu Kahkonen, Lead Economist, Miria Pigato, Sector Manager, Ijaz Nabi, former Sector Manager, Yusupha Crookes, Country Director, and Ernesto May, Sector Director, for continued support and guidance throughout all stages of this report. Muhammad Shafiq, Nimanthi Attapattu, and Irum Touqeer handled with great ease all arrangements for the missions and for the processing of the report. The team benefited enormously from the close collaboration of the Government of Pakistan ever since this work was launched in January 2007. The team is very grateful to Mr. M. Abdullah Yusuf, who conceived the idea for this report, and to Dr. Ather Maqsood Ahmed, who developed the framework and coordinated the inputs of the FBR team. The team is greatly indebted for the outstanding support and assistance of FBR's Fiscal Research and Statistics Wing. Its contribution is immense: it co-authored three background studies; provided comments, feedback, and guidance to the team throughout the analysis and report preparation; and facilitated the interaction with other departments, including the Budget Wing, Debt Office, and the Economic Affairs Division of the Ministry of Finance. Background Papers for the Pakistan Tax Policy Report 1. Ahmed, Robina Ather and Rider, Mark. Pakistan's Tax Gap: Estimates by Tax Calculation and Methodology. 2. Alm, James and Khan, Mir Ahmad. Assessing Enterprise Taxation and the Investment Climate in Pakistan. 3. Bahl, Roy, Wallace, Sally and Cyan, Musharraf. Pakistan: Provincial Government Taxation. 4. Martinez-Vazquez, Jorge. Pakistan ­ A Preliminary Assessment of the Federal Tax System. 5. Michelse, Geerten. Pakistan ­ a Globalized Tax World ­ An Analysis of its International Tax Practice. 6. Thirsk, Wayne. Tax Policy in Pakistan: An Assessment of Major Taxes and Options for Reform. 7. Wahid, Umar and Wallace, Sally. Incidence of Taxes in Pakistan: Primer and Estimates CHAPTER 1: TAPPING TAX BASES FOR DEVELOPMENT 1.1 Introduction 1.1. Pakistan's economic development is once again threatened by macroeconomic imbalances. Broadly speaking, high growth in the 1960s was followed by low growth in the 1970s, and high growth in the 1980s by low growth in the 1990s, as macroeconomic vulnerabilities derailed development. Supported by a favorable global environment, Pakistan returned to a strong development record for much of this decade. Growth accelerated and fiscal and social indicators improved (Figure 1.1). But as in the past, the gains proved unsustainable, as economic policies adjusted too little and too late to the deterioration in the external environment. The looming crisis is threatening to undo much of the recent development progress. 1.2. Bolstering tax collection is a central part of the necessary policy adjustment to overcome Pakistan's macroeconomic weaknesses and sustain development. The problems of high fiscal deficit, high current account deficit, and high inflation are linked, in one way or another, to Pakistan's feeble tax revenue effort. Pakistan, with an unevenly developed private sector, is dependent on public investment to close social and infrastructure gaps (Box 1.1). This suggests that there is little space for scaling back the overall level of public spending from around 20 percent of GDP today, even allowing for efficiency gains in the way in which public resources are spent. Funding such public spending through large fiscal deficits or reliance on non-tax revenues runs the risk of extending the volatile economic cycles of the last decades. Instead, the bulk of development financing has to come from mobilizing revenues through tapping tax bases in a sustainable way. Fortunately, as spelled out in the recently completed Poverty Reduction Strategy Paper II, the government is committed to economic stabilization, where rising tax collection goes hand in hand with rising development and social safety net expenditures and declining fiscal deficits. Figure 1.1: Economic and Fiscal Indicators Public Debt, Fiscal Deficit and GDP Growth Consolidated Public Spending (% of GDP) 90 10 15 15 Public Debt Growth (%) (% of GDP) 14 Public Spending other 14 80 9 than Interest & 13 Defense 13 Budget Deficit (% of GDP) 8 70 12 12 7 60 11 11 6 10 10 Public Spending other 50 than Interest & 5 9 Defense & Federal 9 Subsidies 40 8 8 4 30 7 7 3 6 6 20 2 5 5 10 1 4 4 0 0 3 3 1999/2000 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 1999/2000 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 Box 1.1: Basic Services ­ A Public Responsibility Governments around the world contribute to decent living standards by financing, providing, or regulating basic services. Why? First, these services are replete with market failures--with monopoly power, as when students in remote villages have only one nearby school to go to; with externalities, as when an infected child spreads a disease to playmates; and with information problems, as when patients know little about the treatment prescribed by their doctor. Second, stark deprivation of basic needs violates often the basic values of fairness and equity in a society. So the private sector, left to its devices, will not achieve the level of, say, health and education that society desires. The challenge is to see how the government can mobilize adequate resources to meet this fundamental responsibility together with the private sector, communities and other partners. 1 1.2 Spending for Results 1.3. Whatever the achievements in improving living standards since the beginning of this decade, much remains to be done. The rise in public expenditures over this period coincided with a rise in core indicators of household welfare. While many factors contributed to these good outcomes, this pattern does suggest that well- executed government programs can help to further development (Figure 1.2). However, the scale of the outstanding challenges is immense: over 35 million people did not make ends meet in 2004/05; about 10 million children aged 5 to 9 failed to attend primary school in 2006/07; about 2 million infants of 12 to 13 months of age were not fully immunized in 2006/07; and some 100 million people had no access to piped water as source for drinking water in 2006/07. And the sharp increase in food inflation in 2008 is likely to have increased poverty, as the average household in Pakistan spends about half of total household expenditure on food. 1.4. To address these and other needs, the government is planning to ratchet-up spending on human development and infrastructure over the next five years. In the newly drafted Poverty Reduction Strategy Paper II, the Government of Pakistan has announced a number of mega infrastructure projects and programs such as the National Trade Corridor, large-scale dams, upgrading of transmission and distribution network in the energy sector, and rural electrification. In addition, the Fiscal Responsibility and Debt Limitation Act stipulates doubling spending on education and health as percentage of GDP from 2.2 percent of GDP in 2003/04 until 20012/13. 1.5. Ensuring adequate funding for essential social and infrastructure projects is crucial for economic growth. During times of fiscal consolidation, the bulk of the adjustment tends to fall on public expenditures for development. Other expenditures, such as interest payments, defense, and wages are either predetermined or difficult to cut for political reasons, and identifying revenue yielding measures in the short term proves often difficult. Clearly, not all public programs are equally productive. Some projects are poorly selected and procured, or a vehicle of dispensing political favors rather than acquiring productive assets. This is why in case of a fiscal crunch a careful prioritization for cuts rather than an across-the-board reduction is essential to protect growth- promoting programs. However, declining investment is a cause for concern when it results in decreased accumulation of productive public capital, and it is not compensated by increases in private investment. When such growth-promoting spending is cut so much that the present value of future government revenues falls by more than the immediate improvement in the cash deficit, then fiscal adjustment ultimately harms sustainable growth (WBRO 2008). Pakistan's own experience is a case in point. As development spending dropped from 7 percent of GDP in the early 1980s to only 2 percent by the end of the 1990s in response to fiscal problems, trend growth declined from over 6 percent to no more than 3 percent over the same period. Furthermore, provisions for repair and maintenance remained at around 1 percent of development spending, resulting in an erosion of the stock of capital at a fast rate. 1.6. Even allowing for internal savings, Pakistan's medium-term fiscal framework suggests that overall public spending would have to remain at around 20 percent of GDP to fund essential development programs. This would have to be funded from three sources: fiscal deficit; non-tax revenues; and tax revenues. The role of the first two sources should diminish in future, leaving tax revenues to fill the gap. Figure 1.2: Public Spending and Development Outcomes Per Capita Expenditures of the Government, (Rupees, 2007/08 Constant Prices) Selected Indicators of Livling Standards 4,000 4,000 80 80 3,600 3,600 Full Immunization 70 Rate 70 (% Infants 12 to 23 3,200 3,200 Months) 60 60 2,800 2,800 Net Primary 50 Enrolment Rate (% 50 2,400 2,400 Children Aged 5 to 9) 2,000 2,000 40 40 1,600 Development 1,600 Piped Water PRSP 30 (% of Households) 30 1,200 1,200 800 800 20 Poverty Headcount 20 ( %) 400 400 10 10 0 0 87/88 88/89 89/90 90/91 91/92 92/93 93/94 94/95 95/96 96/97 97/98 98/99 99/00 00/01 01/02 02/03 03/04 04/05 05/06 06/07 07/08 0 0 1998/99 2001/02 2004/05 2005/06 2006/07 2 1.3 Unsustainable Fiscal Deficits and Unreliable Non-Tax Revenues 1.7. Large fiscal deficits crowd out private investments and undermine macroeconomic stability. When the government runs a deficit, it meets some of its expenses by issuing bonds. In doing so, it competes with private borrowers for money lent by savers, raising interest rates. This leads to "crowding out" of private investment. The reduction in domestic investment lowers the capital stock and productivity growth and reduces future national income, as the country will invest less in new plant and equipment. This effect is muted when a country has access to foreign capital, because the budget deficit can be financed from abroad, as the rise in interest rates due to government borrowing attracts foreign capital. While foreign inflows may sustain investment and upgrade technology, higher indebtedness to foreigners mean that a higher fraction of a country's output will have to be sent abroad in future rather than being consumed at home and other foreign inflows are less likely to produce such benefits. In addition, as foreigners bid up the price of the Rupee in order to invest them in high-yielding paper, the exchange rate appreciates. This makes imported goods cheaper in Pakistan and exports more expensive abroad, leading to a decline of the trade balance. This effect underlines the "twin deficits" of budget and trade, as seen in Pakistan in 2006/07. The negative consequences of prolonged, large budget deficits are likely to be more severe than just the crowing out of private investment. The inability of the government to restore fiscal balance can cause a fundamental shift in market expectations and a related loss of confidence both at home and abroad, which in turn can generate a self-reinforcing negative cycle between the budgetary situation, financial markets, and the real economy (Rubin, Orszag and Sinai 2004). As seen in Pakistan during 2007/08, the decline in confidence can reduce stock prices and household wealth, raise the costs of financing to business, and reduce private-sector domestic spending. Although it is impossible to know at what point market expectations about the nation's projected fiscal imbalance could trigger negative feedback dynamics between interest rates, exchange rate and economic activity, the harmful impacts on the economy substantially magnify the costs associated with tackling the macroeconomic imbalances. Indeed, the potential costs and fallout from such fiscal disarray provide perhaps the strongest motivation for avoiding large ongoing budget deficits. It is much harder for the political system to reduce deficits than to expand them. As a result of this asymmetry, enacting a public spending increase or tax cut today is costly because they constrain policy-makers' flexibility to respond to unforeseen events in the future. 1.8. While running high fiscal deficits gets in the way of economic growth, perhaps the first place to look for additional resources is non-tax revenues. After all, taxing citizens can be a quick way for governments to become unpopular. Indeed, Pakistan's non-tax revenues have grown faster (15 percent annually in nominal terms) than tax revenues (14 percent) since 1999/2000, and made up in 2007/08 about 30 percent of total revenues of federal and provincial governments (Figure 1.3, left panel) (Box 1.2). In 2007/08, the government received some 4.3 percent of GDP in non-tax revenues, including surcharges and levies, which is more than the collection from direct or sales taxes. Non-tax revenues are likely to become less important in financing public expenditures in future. According to Pakistan's medium-term fiscal framework, they will decline to below 4.0 percent in 2012/13: · Receipts from public property and enterprises ­ traditionally the most important non-tax revenue source - have declined in nominal terms since 2003/04 due to privatization (Figure 1.3, right panel). This trend is set to continue in the future as the selling of public assets to private owners is likely to proceed. · Civil administration revenues grew more than six fold since 2001/02, primarily due to rising defense receipts from logistical support to US operation in the war on terror. Defense receipts are subject to political uncertainty, which led to delays in payments during 2007/08. · The rise in civil administration revenues also reflects higher profits of the State Bank of Pakistan, in part from the privatization of public stakes in banks in the context of the government's privatization agenda. Current plans include the sale of Global Depository Receipts for leading banks, such as Habib Bank and National Bank, so this is likely to remain an important revenue source in the next few years. However, the challenging environment in global credit markets might make an ambitious placement program difficult to implement. The World Bank estimates that gross capital flows to developing countries will decline from US$683 billion in 2007 to US$103 billion in 2008. Furthermore, this source of funding will deplete in the medium run once such bank holdings have been transferred from public to private ownership. · The performance of provincial non-tax receipts has been disappointing in the past, and there is little to suggest this would change in the near future within the current inter-governmental fiscal framework. 3 Figure 1.3: Pakistan's Tax and Non-Tax Revenues Consolidated Fiscal Revenues (Rs Billion, Current Prices) Composition of Consolidated Non-Tax Revenue (% of Overall Non-Tax Revenue) 1,600 100% 13 10 12 15 14 14 12 28 90% 1,400 10 9 Miscellaneous 27 11 11 Non-Tax Revenue Provincial 8 14 80% 11 17 1,200 21 19 28 24 11 70% 16 15 Civil 26 Administration 1,000 Tax Revenue 60% 25 30 29 30 21 Surcharges 27 and Levy 35 800 50% 10 40% 600 16 30% 16 16 68 400 58 54 56 50 20% 52 Property & 200 Enterprises 28 25 10% 26 0 0% 1999/2000 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 1999/2000 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 Box 1.2: Gas and Petroleum Surcharges and Petroleum Development Levy From 1999/2000 to 2004/05, Pakistan's fiscal accounts subsumed gas and petroleum surcharges and petroleum development levy as tax revenues. Since 2005/06, the fiscal accounts categorized such surcharges and levies as non-tax revenues. This report uses a consistent definition over the entire period and classifies these receipts as non-tax revenues. Petroleum and gas surcharges are an unstable source of budgetary revenue because retail prices, as well as margins for the distributors, transporters, and dealers are regulated. In principle, petroleum and gas retail prices are adjusted regularly based on the sum of lagged landed costs of fuel, fixed margins, surcharges fixed in Rupees per liter, and all other taxes, including the 16 percent GST. However, due to political pressures to cushion domestic consumers, the surcharges were compressed as landed costs of fuel and gas increased since 2002. Such surcharges and levies yielded as much as 1.4 percent of GDP in 2002/03, amounted to no more than 0.2 percent of GDP in 2007/08, and are projected to total as much as 1.1 percent of GDP in 2008/09. 4 1.4 Tapping Tax Bases 1.9. A sustainable increase in the government's resource envelop requires higher tax revenue mobilization. Yet, Pakistan's tax collection has failed to improve since the late 1990s. Structural problems, such as a narrow tax base, tax evasion, distrust of the taxpayer vis-à-vis public institutions, and administrative weaknesses, have taken a toll on tax collection. The tax-to-GDP ratio increased from 9.6 percent in 1999/2000 to 10.3 percent in 2007/08 (Figure 1.4). In order to ensure adequate public funding for development priorities while safeguarding macroeconomic stability, the government has endorsed the objective to increase tax collection 13.9 percent of GDP by 2012/13. This commitment is reflected in Pakistan's 2009 Poverty Reduction Strategy Paper. The rise in tax revenues, in addition to a decline in interest payments, will allow the government to reduce the fiscal deficit from 7.4 percent of GDP in 2007/08 to 2.4 percent of GDP in 2012/13. Figure 1.4: Turning Around the Tax-to-GDP Ratio Pakistan's Tax Revenues - Actual and Medium Term Framework (% of GDP) 14.0 14.0 ACTUAL PROJECTION 13.5 13.5 13.0 13.0 12.5 12.5 12.0 12.0 Tax Revenue PRSP-II Tax Revenue with 11.5 11.5 PDL & Surcharges 11.0 11.0 10.5 10.5 10.0 Tax Revenue 10.0 9.5 9.5 9.0 9.0 1999/2000 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 1.10. International experience shows that tax reform can delivery large increases in the tax-to-GDP ratio. While there are other developing countries at Pakistan's income level with similarly low tax-to-GDP ratios, a look across the border suggests that this does not have to be that way (Figure 1.5). The simple average of the tax-to- GDP ratio in Bangladesh, India, Nepal and Sri Lanka ­ countries with similar tax policies and administration ­ is systematically higher than in Pakistan, and the gap increased during this decade. Furthermore, countries like Egypt, India, Thailand, Turkey, and South Africa experienced rapid growth and rising tax ratios, while Pakistan saw tax collection rising just in line with economic growth. And while central government tax collection increased from 13.8 percent in 2000 to 16.5 percent in 2004 in Asian and Pacific countries overall, it remained roughly constant as percent of GDP in Pakistan since the early 2000s. However, increasing tax collection is so much more difficult with anemic rather than strong economic growth. Indeed, on the back of the global economic crisis, Pakistan's tax- to-GDP ratio is set to decline in 2008/09 to 9.2 percent. 1.11. Structural weaknesses of Pakistan's tax system heighten its vulnerability to the economic crisis. There are worries that the revenue is raised in an inefficient way by favoring certain sectors and economic activities over others. Such excess burden of taxation can deter people from investing in the most productive sectors and earning more form the resources available, and ultimately get in the way of economic growth. Some sectors are much more heavily taxed compared to their contribution in terms of GDP than other sectors. Agriculture contributes about one fifth of GDP, yet no more than 1 percent in FBR tax revenue. Services make up almost half of economic value added, but only one quarter of central taxes due to the low tax receipts from wholesale, retail, and transport. Given the shortfall in agriculture and services, industry carries the brunt of the tax burden - its tax share is three-times as high as its GDP share. In addition, there are question marks to what extent the tax system, through the way it treats different income classes of people differently, is sufficiently equitable. While some progress has been made, Pakistan's tax code remains complicated and most taxpayers have little knowledge on their obligations. Finally, provincial taxes yield no more than 0.4 percent of GDP, so that district and province governments depend on large fiscal transfers from higher levels to meet their expenditure responsibilities. 5 Figure 1.5: Tax Collection in Pakistan and other Developing Countries Tax Revenue in South Asian Countries (% of GDP), 2000/01 to 2006/07 13.5 13.1 13.0 12.5 12.0 12.0 11.5 11.5 11.0 2006/07 10.5 2003/04 2000/01 10.2 10.0 9.8 9.8 9.5 Pakistan Simple Average of Bangladesh, India, Nepal and Sri Lanka General Government Tax Revenues and Per Capita Income Pakistan and Other Emerging Economies (2003 or 2004) 40 12,000 PPP GNI Per Capita 35 10,000 30 8,000 25 20 6,000 15 4,000 Tax Revenue 10 (% of GDP) 2,000 5 0 0 S. Africa Turkey India China Brazil Russia Mexico Pakistan Thailand Ukraine Egypt Sri Lanka Vietnam Philipp. Banglad. Nepal Indones. Colomb. Growth (%) and Tax Revenue (% of GDP) in Emerging Market Economies, 2002/03 to 2006/07 28 8.0 GDP Growth 26.3 26 (02/03 to 06/07) 23.9 7.0 24 22.9 22.4 22 6.0 20 17.7 18 16.3 5.0 15.8 16 Tax to GDP (%) 14.5 2002/03 2006/07 14.0 14 13.4 4.0 12 10.0 10.2 3.0 10 8 2.0 6 4 1.0 2 0 0.0 Pakistan Egypt Thailand India Turkey South Africa 6 1.5 Scope 1.12. In view of these concerns, it is important to take a thorough look at Pakistan's tax policy reform options. The main message of this report is that Pakistan can take measures to increase the tax to GDP ratio by around 3.5 percentage points over the next five years. In order to ensure a healthy long-run economic development, Pakistan needs to embrace substantial changes in tax policy aimed at increasing the buoyancy of the tax system, broadening the tax bases, reducing distortions and phasing out exemptions. Such tax reforms are also required to deal with the risks stemming from sustained large budget deficits. Failing to act sooner rather than later, only makes the problem more difficult to address without considerable instability, raises the probability of fiscal and financial disarray at some point in the future, and runs the risks of further constraining policy flexibility in future. 1.13. This report provides a comprehensive assessment of Pakistan's tax policies and lays out options its reform. It is based on work commissioned for this report and on contributions of other researchers in Pakistan and elsewhere.1 The report has four parts (Figure 1.6). The first part, this chapter, argues from a macroeconomic perspective that there are no viable alternatives to mobilizing tax revenues for financing government expenditures. The second part, Chapter 2, pulls together the main findings to lay out a roadmap for tax reform. It presents a comprehensive medium-term reform agenda of tapping tax bases for Pakistan's development. The third part, Chapter 3, subjects Pakistan's tax system to a basic health check and assesses the tax system with regard to the four properties of adequacy, efficiency, equity and compliance. The final part, covered in Chapter 4 to Chapter 9, takes each major tax and discusses how it squares up with regard to the properties spelled out in Chapter 3. 1.14. This reports deals only with some of the many aspects of Pakistan's tax system. This coverage reflects the emphasis recommended by the Federal Board of Revenue and peer reviewers at the concept stage of this report, the desire to avoid duplication with existing literature, data constraints, and, last but not least, the need to keep the task manageable. Four limitations are worth highlighting explicitly. First, the report's principal focus is on tax policy issues. Yet, tax policy is intimately related to tax administration, as any implemented tax policy has to pass the test of administrative feasibility. Tax policies should not just generate resources efficiently and equitably, but also ensure broad compliance. This report does not offer a discussion of Pakistan's tax administration, but it provides estimates of tax revenues lost due to tax evasion. Fortunately, the FBR tax administration review (2008) provides an up-to-date assessment of the progress in federal tax administration reform. Second, the report presents an aggregate view. Particular sectors, like oil and gas or banking, face special tax regimes that are not discussed in detail. Being mindful of such sectoral dimensions will be important when translating the reform into tax policy changes. Third, provincial tax polices are discussed for Punjab and NWFP only. As each province faces unique challenges ­ while Sindh has a large tax base through the commercial hub of Karachi, Balochistan is much devoid of any sizable tax base ­, a similar analysis for Sindh and Balochistan would be important, especially in the context of a discussion on the intergovernmental architecture. Finally, politics is a key determinant of the level and structure of taxes in Pakistan, as in other countries, but we do not provide an analysis of the politics of tax reform. We offer no remedies for removing political constraints to good tax structures other than to argue that more transparency and a better appreciation of the gains to be had from good tax policy might lead to better outcomes. Figure 1.6: Navigational Aid for the Report CHAPTER I OBJECTIVE Tapping Tax Bases for Development CHAPTER 2 STRATEGY National Tax Policy Reform CHAPTER 3 PROPERTIES Adequacy Efficiency Equity Compliance CHAPTER 4 CHAPTER 5 CHAPTER 6 CHAPTER 7 CHAPTER 8 CHAPTER 9 TAXES General Sales Corporate Individual Federal Customs Provincial Tax Income Tax Income Tax Excise Duties Duties Taxes 1 Pakistan has a long tradition of tax policy analysis. In their seminal contribution, Ehtisham Ahmad and Nicolas Stern (1991) show how design principles of tax policy can help to assess the impact of tax reform for households, producers and government, and apply the framework to Pakistan. Other contributions include Pasha and Iqbal (1994), Pasha (1995) and various reports of the International Monetary Fund. 7 1.6 Process 1.15. This report is the outcome of a partnership among three parties: the Federal Board of Revenue; the Andrew Young School of Policy Studies (AYSPS) at Georgia State University, and the World Bank. The Government of Pakistan attaches high priority to broadening the tax base in an efficient and equitable manner, and is aware of the need for further improvements in the tax system. After launching a project of the FBR and the World Bank on tax administration reform in 2005, the government sought to complement this work with a parallel initiative on tax policy. In response to the government request, a partnership among the three parties was formed in January 2007. Additional government counterparts include the Ministry of Finance and finance departments in Punjab and Balochistan. The services of AYSPS were generously supported through the UK Department for International Development. 1.16. This collaboration extended beyond the delivery of a report, as process was viewed as a key input for impact. Pakistan's capacity to undertake fiscal policy analysis is limited, despite selective strong points at FBR and elsewhere. Long-term remedies to this situation require initiatives such as training at the Masters level in fiscal policy of FBR and Ministry of Finance personnel, expanding the number of positions dedicated to tax policy analysis in those institutions, and offering more attractive salary and working conditions to attract and retain the right people. Such measures were beyond the scope of this project. Nevertheless, other forms of capacity building took place by the joint work of the AYSPS experts and government tax policy analysis experts, ensuring that the process of preparing the report served as a "teaching tool" for capacity building and a "collaborative tool" to build ownership. Many of these mechanisms also served as dissemination tools to stimulate dialogue. The report's primary audience is the federal and provincial governments. The report is also written for policymakers and development practitioners in Pakistan, as well as for governments and international donors that support tax policy reform in other South Asian countries. The report is expected to generate debate and discussions among the civil society and research communities. 1.17. Timely delivery of analytical work and regular dissemination of intermediary outputs is crucial for stimulating on-going dialogue with counterparts and the broader set of actors. For this reason, the background studies for this report were structured as a series of seven tax policy papers. The AYSPS teams visited Pakistan two times in the course of the preparation of each policy papers. The first mission gathered the basic information, held detailed discussions with government staff and other stakeholders, and came to an agreement with the Pakistan authorities on the coverage of each tax policy study. The second mission presented the draft policy paper and collected comments and feedback for the finalization of the study. The final report will be disseminated through a tax policy workshop. 1.18. The collaboration involved three modes of capacity building. First, three FBR officials spent each three months as visiting scholar at AYSPS to collaborate with the report team on tax policy papers. Second, the AYSPS team delivered in-country training program for civil servant on the methodologies and findings of the tax policy papers. Third, government staff engaged in learning by doing on the policy papers through regular discussions with the report team. In addition, the report team went with a FBR delegation on a study tour to Turkey to learn about its successes and failures in the reforms of their tax systems. 8 CHAPTER 2: NATIONAL TAX POLICY REFORM 2.1 Introduction Embracing Reform 2.1. Pakistan's economic crisis provides the window of opportunity to overcome the inevitable resistance that accompanies comprehensive tax reform. Pakistan's severe fiscal challenges are in good measure the result of structural shortcomings of the tax system. These deficiencies largely imply that the tax system has failed the common interests of taxpayers, tax policy makers, and tax administrators. The tax system's low revenue yield weakens public services; its inefficiencies and unfairness undermines economic growth and voluntary compliance; and its complexities make administration and enforcement difficult: · The tax revenue to GDP ratio is low: The main problem with the tax system is its sustained inability to raise adequate revenues to finance the public sector budget. In fact, over the 2000s tax revenues have barely kept pace with GDP, and the tax revenue to GDP ratio even declined in the last year. Based on international comparisons, Pakistan is a low taxing country. Given the large development needs for social services and public infrastructure, there remains a structural fiscal gap equivalent to about five percent of GDP. · The efficiency costs of the tax system to the economy are high: Pakistan's tax system taxes different economic sectors and types of assets at very different marginal effective tax rates, which create significant distortions. Although some distortions may be intentional, the real consequence of the distortions is that many investment decisions are guided by tax considerations as opposed to economic considerations. The result is an overall less efficient allocation of resources and lower levels of output and rates of economic growth for Pakistan. These distortions are caused, among other issues, by the favorable treatment of certain investments and sectors. · Frequent changes in tax legislation have made the overall tax structure incoherent and tax bases narrow: Tax policy reform in Pakistan has often been piecemeal and disconnected. Many changes in the tax system were catered either to the interest of specific economic sectors or specific firms, or introduced as matter of administrative convenience, as in the case of the zero rating of the five sectors. The government does not know the revenue losses of the incentives given, and has no rigorous system to evaluate the objectives and benefits. These policies have contributed to the narrowing of the tax base and the decrease in tax effort. They have also compromised the fairness of the tax system, reduced voluntary compliance and increased tax. · There are significant horizontal inequities: Currently, the tax laws are such that individuals with the same income or businesses with the same profits can be treated very unequally in terms of the final taxes the law requires them to pay. For example, if your income is in the form of capital gains you entirely escape any tax obligation in contrast to those receiving their income from wages. The horizontal unfairness of the tax system is exaggerated by the uneven application of tax enforcement: while compliant taxpayers bear the full burden of taxes, the failure of tax enforcement mechanisms allows other taxpayers to evade their taxes with a large degree of impunity. · The tax system is overly complex: Even though the basic structure of Pakistan's tax system is broadly in line with international practice, these taxes have become more complex over the years, because of exemptions, other preferential treatments, and ad hoc changes to the structure of those taxes. The extensive use of withholding taxes, often being considered a final tax ­ as opposed to being adjustable against final full liabilities ­ has added to the complexity and arbitrariness of the tax system. · Tax administration remains weak: in spite of partial reforms, there is a long way to go to fully modernize the tax administration. This long-standing bottleneck is characterized by over-reliance on easier "tax handles" through an extensive system of withholding taxes, inefficiencies in administrative structure and operations, and weak enforcement. Over half of the legal tax base still goes untapped. This situation gives rise to the inadequacy of revenue yields and loss of confidence in the fairness of the system due to large horizontal inequities. 2.2. The mild progressivity of the tax system is a positive feature of the current tax system that needs to be preserved. Higher income households bear proportionally a larger share of the tax burden than lower income households. Overall, consumption taxes are distributed in a relatively proportional manner--households across income distributions pay similar percentages of their income in tax. Direct taxes have a more progressive incidence impact on the distribution of after-tax incomes due in part to the high threshold for the individual income tax and the concentration of capital income in the higher deciles of the income distribution. 9 General Directions for Tax Policy Reform 2.3. Reforming Pakistan's tax system on a sustainable basis will require a clear vision for the desired tax structure. This report offers for each of the federal and provincial taxes a description of the current system, a diagnostic of the main problems, and a set of policy options for reform. Most of these reforms could be introduced in the short to medium run on a piecemeal basis. This would go a long way towards addressing many of the current problems with the tax system. Yet, the properties of the tax system hinge not just on the performance of each tax, but on the overall impact of the tax system as a whole. This means that a strategy for the main strands of tax policy reform is crucial for success (Box 2.1). The same approach is required for the reform of tax administration which remains a critical piece of long term strategy for turning around the performance of Pakistan's tax system. 2.4. Fortunately, there is broad consensus on the objectives of and means for fundamental tax reform. Besides increasing tax revenues to finance more and better services and infrastructure, the reform ought to achieve a fairer distribution of tax burden across taxpayers with the same income levels (horizontal equity) and preserve, if not increase, the overall progressivity of the system (vertical equity). The reform should also reduce tax distortions in the economy. And the tax system should be simplified to reduce taxpayer compliance costs and facilitate the task of the tax administration. Even though the tax rates of the major taxes are average by international standards, the current tax bases are too narrow. In combination with a culture of taxpayer non-compliance and an ineffective tax administration, this has led to low and decreasing tax yields. Making these changes sustainable will require refocusing the tax administration from fine tuning tax laws and rules to effective enforcement of the tax laws and rules and the provision of taxpayer services that facilitate voluntary compliance. 2.5. The structural reforms developed in the following paragraphs should be interpreted as general directions for reform that would need to be further refined. They provide an order of magnitude for the potential revenue impact that will have to be elaborated once specific reforms are spelled out. The revenue impacts are calculated by comparing the actual 2007/08 tax collection with the simulated 2007/08 tax collection under specific tax policy reform scenarios. They are expressed either as percentage change in the 2007/08 actual collection, or as the Rupee amount of additional (or reduced) tax collection. The simulations reflect only the direct or static revenue impact from the change in tax rate or tax base, assuming that the behavior of taxpayers and the level of economic activity is the same with and without the tax policy reform. The main reason for making this assumption is that we know not enough about the type and scale of behavioral responses in Pakistan to tax policy changes in order to quantify their likely impact. The calculations also assume that taxpayer compliance remains unchanged relative to current levels. In selected cases, we explore the impact of specific behavioral responses, such as lowering compliance for withholding taxes; and increasing import volumes in response to lower custom tariffs. Box 2.1: Egypt's Tax Reforms Egypt changed the direction of economic policies sharply in July 2004 with the appointment of a pro-reform cabinet. The new cabinet quickly announced broad ranging economic reforms. Taxes were a major pillar of the reform program. Corporate and personal income taxes through 2004 were high -- with statutory rates up to 40 percent, and following tax reforms in Jordan and India, Egypt's top corporate income tax rate had become the highest among the peer group by 1999. The 2005 major tax reform halved the top income and corporate tax rates to 20 percent and put Egypt's rate among the lowest of the peer group. While rate reductions would lower revenues in the short run, other measures introduced by the legislation, notably the elimination of all exemptions and tax holidays, in addition to the increase in prices of domestic energy products and electricity helped to offset this shortfall. The reforms also introduced measures to expand the tax base, including through provisions to encourage the informal economy to legalize its status. In addition, the reforms involved a major overhaul of tax administration, including a move to self-assessment, preparation of a modern VAT to replace a complex sales tax, and the beginning of the integration of the income and sales tax departments. Egypt's tax reform improved revenue collection. Over 2.5 million taxpayers submitted tax returns in 2005/06, a significant increase from 1.7 million in 2005. This increase may have been in part in response to an amnesty scheme for tax evaders. Income tax collection remained constant at 5.9 percent of GDP from 2004/05 to 2006/07, in spite of the reduction in tax rates, and overall tax collection increased from 14.1 percent of GDP in 2004/05 to 14.6 percent of GDP in 2006/07. These tax reforms also contributed to a sharp drop in the importance of tax-related concerns in the 2007 World Economic Forum (WEF) survey. Complaints about the level of tax rates rose from sixth place in 2004 to third in 2005 in the WEF business surveys, before dropping to tenth place in 2007. Similarly, in the 2004 and 2005 WEF surveys, complaints about tax regulations ranked fourth and second place, respectively, and again fourth place in 2006, before dropping to sixth place in the 2007 report. 10 2.2 General Sales Tax 2.6. GST is the most conspicuous underperformer in Pakistan's tax structure, and hence the most promising revenue yielder in structural tax policy reform. Pakistan does not have a broad-based GST on goods and services. There are multiple exemption and zero-rating provisions for the federal GST on goods. Eliminating these provisions would substantially widen the base, and, at the standard GST rate, boost revenues. In addition, the constitution preserves the taxation of services for the provinces. Introducing a federally administered GST on both goods and services would widen the tax base substantially. Such a reform might require a constitutional amendment as well as changes to the revenue sharing formula for the divisible pool between the federal government and the four provinces (Box 9.16). Alternatively, even without changing the constitution, all provinces could agree to delegate the responsibility for the collection of GST on general services to the federal government. They could also agree to increase the share of GST collection on services retained by the federal government from currently 2 percent in order to provide the federal government adequate incentives for tax collections. A third option would be that FBR levies federal excise duties in VAT mode on services, as indeed it does already now for selected services. These revenues would then become part of the divisible pool, where the federal government retains more than half of the receipts. Any of the three reform option would ensure that both federal and provincial governments are far better off in terms of revenue position than under the current system. In addition, the country as a whole would benefit from a more neutral and fairer GST. 2.7. The revenue impact of widening the base of GST for goods and services depends on the number of sectors brought into the tax. We simulate the revenue impact of a general single-rate GST with a zero rate on exports for four exemption scenarios. Since some services are already taxed at the federal level under the form of excise taxes implemented in a GST mode, and other services are already taxed at the provincial level under the provincial service tax, we allow for a set-off of Rs. 17.2 billion in 2007/08. The revenue increases of the broad- based GST with exemptions for government services, healthcare and education are about Rs. 734 billion (Box 2.2, Box 2.10, and left panel of Figure 2.1). If the new GST were also to exempt pharmaceuticals and surgical equipment, the revenue impact is Rs. 714 billion. If the new GST in addition exempts financial services, the revenue increase is Rs. 563 billion. In practice, financial services are often exempt from GST because of the technical difficulties of measuring value added in many financial transactions and intermediation. If the new GST also exempts unprocessed foods in addition to the previous six sectors then the revenue estimate is Rs. 408 billion, about 60 percent of which due to increased taxation of goods (Figure 2.1, right panel). This represents an increase of 108 percent over the actual collection in 2007/08. Unprocessed foods include 11 agricultural sectors, including rice, wheat, sugar cane, pulses, potatoes, vegetables, fruits, seeds, livestock and slaughter products, and fisheries. Processed foods, such as vegetable oils, milled grains, bakery products, and sugar are taxable under this proposal. 11 Figure 2.1: Revenue Impact of GST Broadening Revenue Increase as % of Actual Collection EX: GS+H+E (5%) 2007/08 Revenue Impact of GST Broadening: Goods and Services (Rs. Billion) 800 EX: P+SE+GS+H+E (40%) 700 EX: FS+P+SE +GS+H+E (41%) 600 500 Tax to GDP (%) Services 2002/03 2006/07 E X: UF+FS+P+SE+GS+H+E (108%) 400 300 200 Goods Actual (100%) 100 0 Ex: UF+FS+PSE+GS+H+E Ex: FS+PSE+GS+H+E Ex: PSE+GS+H+E Ex: GS+H+E Box 2.2: General Sales Tax Reform ­ Broadening the Base The standard rate of 16 percent is applied to goods and services. Exports are zero-rated. In addition, certain sectors are exempted. We distinguish four scenarios: (i) Government services, health care and education (EX: GS+H+E); (ii) As under (i), as well as pharmaceuticals and surgical equipment (EX: P+SE+GS+H+E); (iii) As under (ii), as well as financial services (EX: FS+P+SE+GS+H+E); and (iv) As under (iv), as well as unprocessed foods (EX: UF+FS+P+SE+GS+H+E). 12 2.3 Corporate Income Tax 2.8. The general direction for reform should be to broaden the tax base, mainly by rationalizing tax incentives and exemptions and lowering the statutory tax rate. Many countries have been able to increase or maintain tax revenues in recent years through the general reform strategy of reducing corporate income tax rates and broadened tax bases. Although Pakistan has significant potential to broaden the tax base, the data currently available do not allow us to simulate the revenue implications. However, we are able to simulate the revenue impacts of reducing the tax rate. Another reform issue concerns the distortion introduced by the different treatment of debt and equity. Several approaches are available, such as an allowance for shareholder equity or removing the tax deductibility of interest payments, but leave this issue for another study. 2.9. The main reform proposal is to reduce the standard corporate income tax rate from 35 percent to 30 percent for large companies. We consider two options on how to do this. The first simulation keeps the corporate income tax rate for companies with less than Rs. 250 million turnover at 20 percent, as prescribed by the current small company legislation; applies the reduced standard corporate income rate of 30 percent for companies with more than Rs. 750 million turnover; and then uses a bridging tax rate schedule to address the `notch' problem for companies with turnover ranging from Rs. 250 million to Rs. 750 million (Box 2.3). This bridging schedule is designed to ensure that the companies with over Rs. 750 million turnover bracket pay an average and marginal tax rate of 30 percent on their profits, this is similar to the practice followed in other countries, such as the United States. The revenue costs are 17 percent of the actual 2007/08 collections of Rs. 259 billion, equal Rs. 43 billion (Figure 2.2, left panel). The second simulation also keeps the corporate income tax rate for companies with turnover of less than Rs. 250 million at 20 percent, and applies the 30 percent income tax rate to companies with turnover of more than Rs. 250 million (Box 2.4). The revenue loss is 14 percent, or Rs. 38 billion in 2007/08 (Figure 2.2, right panel). 2.10. For both scenarios, we assume that the distinction between old and new small companies is abolished for three reasons. First, it is doubtful that FBR is able to enforce this distinction. In particular, old small companies (i.e. companies with turnover of less than Rs. 250 million that existed already prior to the introduction of the small company concept in 2005) could formally be reconstituted as a new small company. Second, there is no good economic justification for treating differently old and new companies with low turnover. Finally, the revenue loss of reducing the corporate income tax rate from 35 percent to 20 percent for all of companies with low turnover is negligible. Specifically, we estimate these revenue costs to be only Rs. 110 million in 2007/08, equal to no more than -0.04 percent of gross corporate tax receipts. 2.11. The simulations do not account for any behavioral change or compliance response by taxpayers that may result from the reduction in marginal tax rates associated with this proposal. Of course, the reduction in the tax rate from 35 to 30 percent could well lead to an increase in gross receipts through better compliance which in turn would partially offset the estimated revenue loss. 13 Box 2.3: Corporate Income Tax Reform ­ Bridging Structure The corporate income tax rates are as follows: (i) 20 percent for turnover up to Rs. 250 million (ii) 25 percent for turnover of more than Rs. 250 million and up to Rs. 375 million, and (i) (iii) 30 percent for turnover of more than Rs. 375 million and up to Rs. 500 million, and (ii) (iv) 35 percent for turnover of more than Rs. 500 million and up to Rs. 625 million, and (iii) (v) 40 percent for turnover of more than Rs. 625 million and up to Rs. 750 million, and (iv) (vi) 30 percent for turnover of more than Rs. 750 million, and (v) Box 2.4: Corporate Income Tax Reform ­ Two Tier Rate Structure The corporate income tax rate schedule is as follows: (i) 20 percent for turnover up to Rs. 250 million (ii) 30 percent for turnover of more than Rs. 750 million Box 2.5: Modeling Corporate Income Tax The revenue estimates are based on a micro-simulation model that uses 12,928 corporate income tax returns for 2006/07. The model accounts for adjustments to reported tax liabilities (i.e., tax averaging, minimum tax, and so on). There are four main modeling issues. · First, there are a significant number of form use errors and calculation errors in the set of corporate income tax returns used for this model. This may affect the accuracy of the estimates. · Second, the model accounts for adjustments to reported tax liabilities (such as tax averaging, minimum tax) and non-filers, namely corporations that make corporate income tax payments but do not file a corporate income tax return. We assume that the tax attributes of such non-filers are distributed in the same way as the filers. This assumption would be invalid, for example, if small businesses are less likely to file than large businesses. · Third, we assume that corporate income tax compliance remains unchanged compared to 2006/07. Current law compliance is established by simulating actual tax payments using reported taxable profits and current law statutory tax rates with proposed law using reported taxable profits and proposed law tax payments. This is a conservative assumption, as the reduction in the corporate tax rate might improve compliance. · Finally, we need to update the revenue impact from 2006/07 to 2007/08. According to the model, the revenue cost of the reduction in the corporate income tax rate from 35 percent to 30 percent is 17 percent of gross tax receipts in 2006/07. We apply this estimated percentage loss to the actual corporate income tax receipts in 2007/08 over Rs. 259 billion. Implicitly, we are assuming that the distribution of taxable profits, in particular between small and large businesses, is the same in both years. Figure 2.2: Revenue Impact of Corporate Income Tax Rate Reduction Bridging Schedule Two-Tier Structure Rate Reduction (-17%) Rate Reduction (-14%) Actual (100%) Actual (100%) 14 2.4 Individual Income Tax Flat Rate Tax 2.12. Pakistan's individual income tax system requires simplification and rationalization. The individual income tax is collected through a large number of withholding mechanisms, estimated payments, and final tax return reconciliations, while the laws governing the taxation of different forms of individual income are not integrated. In 2006 there were five primary tax returns that individuals could be required to file, each with different tax rate schedules, depending on the type of income and taxpayer. Besides the complexity of returns and rate schedules, individual income taxation is subject to a large array of exemptions and deductions. Even accounting for revenue receipts is complicated by the use of myriad withholding schemes, and by the lack of classification of receipts from current liabilities, past liabilities, and penalties and interest. 2.13. The fundamental direction for reform proposed here is to move toward an integrated individual income tax structure encompassing income of all individuals, including non-incorporated businesses. This entails a radical simplification of the taxation of individual income in Pakistan. The goal behind the integrated individual income tax would be to treat taxpayers, whether they are salaried employees, self-employed businesses, and other non-corporate entities, in a similar way by applying the same tax rate structure to taxable income. Two other reform features should be considered: flat (or flatter) tax rates and dual treatment for capital income. The potential advantages of a flat rate individual income tax have convinced a growing list of countries to adopt this type of reform. Perhaps the most significant attractions for developing countries are the ease of administration and compliance, as well as simplicity in terms of design, for example for inflation adjustments. The empirical evidence on the effects of flat rate income taxes is only emerging but the evidence available for some countries, such as Russia, is that there may be significant positive effects on compliance and smaller effects on real economic activity (Gorodnichenko, Martinez-Vazquez and Sabirianova-Peter 2008). There is also the need to recognize the global economic trends that have made an increasing number of countries, in recent times, treat capital income differently from earned income, in what has become known as a "dual income tax" system. Although there are other reasons, the main argument for treating capital income differently is the high mobility of capital in a globalized world. 2.14. The first reform scenario is the fundamental simplification of the individual income tax through a flat rate of 15 percent (Box 2.6). The income threshold is set to a value approximately equal to twice the per capita income in Pakistan. Preferably, this reform option would provide for a number of specific deductions, but unfortunately we do not have the data to simulate their impact. Instead, we allow for a 10 percent limitation which is our estimate of the level of deductions currently taken for medical expenses and mortgage interest. 2.15. Depending on the assumptions for compliance, the net revenue impact would be an increase of 18 percent to 33 percent of individual income tax actual collections. For 2007/08, the net revenue impact under current compliance would increase individual income tax revenue by 33 percent--or by Rs. 38 billion relative to the actual collection of Rs. 115 billion (Figure 2.3, left panel; Box 2.7 explains how to read this figure). Assuming that compliance and enforcement of the system will remain at current levels might be too optimistic for the withholding tax reforms, as they increase the withholding rates substantially for most withholding categories (Box 2.9). For example, over 90 percent of the revenue impacts among withholding taxes come from withholding on contracts and supply of goods and imports. Under current law, these items are taxed at rates of between 0.75 percent and 6 percent. Under the reform scenario, the tax rate would be a flat 10 percent rate. Assuming that compliance falls by 50 percent compared to the current levels for withholding taxes, the revenue impact of the individual income tax reform would still be Rs. 21 billion. 15 Two-Tier Rate Structure 2.16. The crucial issue with a flat rate tax is that the income tax can become much less progressive. The second reform scenario addresses this issue: It replaces the flat rate income tax rate with a two-tiered rate--12 percent for those with taxable income less than Rs 300,000 and an 18 percent rate for those with taxable income greater than Rs 300,000 (Box 2.8). All other aspects of the individual income tax reform in the first scenario remain unchanged. 2.17. The revenue impact of the two-tier rate system turns out to be fairly similar to the flat rate scenario. The revenue losses from non-withholding and non-capital incomes are reduced, while the revenue gains from withholding and capital incomes are unchanged (Figure 2.3, right panel). The revenue increases for 2007/08 are estimated to be Rs. 39 billion (34 percent of actual individual income tax collection) under the assumption of current compliance, and Rs. 22 billion (19 percent) under the assumption of a drop in compliance of 50 percent compared to the current levels for withholding taxes. The main difference is that individual income tax is now more progressive, although in both cases the tax burden is proportionally larger for the top four population deciles due to the high exempt threshold. 2.18. The reform of the individual income tax could yield higher revenues by adopting higher rates. For example, if we apply a 30 percent rate (instead of 18 percent as before) to taxable income over Rs. 300,000, the revenue increase rises from 34 percent (Rs. 39 billion) to 54 percent (Rs. 62 billion) under current compliance. From the perspective of tax policy design, this modified two-tier tax rate scenario would allow to equalize the highest rate of the individual income tax to the general rate of a reduced corporate profit tax at 30 percent, as discussed next. 16 Box 2.6: Individual Income Tax Reform ­ Flat Rate Tax (i) Flat tax rate of 15 percent for income tax with a minimum exempt income threshold of Rs. 130,500 and a maximum deduction of 10 percent for medical expenses and mortgage interest (ii) Separate tax on capital income, including short term capital gains, at 10 percent also with a flat tax rate (10%) but no minimum income threshold (iii) Rationalization of the withholding system with a uniform 10 percent withholding rate on various withholding taxes for income items but with the elimination of the withholding tax on transactions; and (iv) Elimination of all individual income tax exemptions. Figure 2.3: Revenue Impact of Individual Income Tax Reform under Current Compliance Flat Rate Tax Two Tier Tax Rate Capital Income (4%) Capital Income (4%) Tax Rates Tax Rates Withholding Taxes (30%) Withholding Taxes (30%) (-0%) (-1%) Actual (100%) Actual (100%) Box 2.7: Interpreting the Left Panel of Figure IX.1 Since this chapter will be depicting the revenue impacts of tax policy reform through figures like Figure 2.3, it is worth explaining in some detail how to interpret this figure. The flat rate tax would increase the 2007/08 tax collection from individual income tax as indicated by the arrow pointing upwards. The bulk of the increase is due to higher withholding tax collection (30 percentage points). In addition, capital income contributes 4 percentage points. By contrast, the adjustment in tax rates would lower tax collection by 1 percentage point. This amounts to a net increase of 33 percent relative to the actual individual income tax collection. Box 2.8: Individual Income Tax Reform ­ Two-Tier Rate Structure (i) Tax rate of 12 percent for those with taxable income less than Rs. 300,000; and tax rate of 18 percent for those with taxable income equal or greater than Rs. 300,000 with a minimum exempt income threshold of Rs. 130,500 and a maximum deduction of 10 percent for medical expenses and mortgage interest (ii) Separate tax on capital income, including short term capital gains, at 10 percent also with a flat tax rate (10%) but no minimum income threshold (iii) Rationalization of the withholding system with a uniform 10 percent withholding rate on various withholding taxes for income items but with the elimination of the withholding tax on transactions; and (iv) Elimination of all individual income tax exemptions. 17 Box 2.9: Modeling Individual Income Tax What might be the revenue impact of the tax policy reforms on 2007/08 tax collection? We answer this question with the help of a micro-simulation model that combines the parameters of tax structure, tax base and tax compliance to replicate budgeted tax collection for 2007/08, based on a 2 percent representative sample of 2006 individual income tax returns provided by FBR. Based on the trends for gross wages and salaries and other components of the individual income tax base, we inflate the 2005/06 tax base to proximate 2007/08 tax base. We also incorporate the changes in tax policy parameters between 2005/06 and 2007/08, including the switch from a marginal to a notch rate structure, and changes in thresholds and allowances. There are five main issues in developing this tool. First, the attribution of individual income tax payments to the five types of individual taxpayers is difficult. While the individual taxpayers (self-employed comprising non-salaried and associations of persons; salaried individuals without other sources of income; salaried individuals with other sources of income; self-employed with receipts chargeable under certain section of the income tax ordinance; and retailers with turnover of Rs. 5 million or less) are well identified through separate income tax returns, not all income tax payments come with an income tax return. In particular, many income tax receipts come in form of withholding payments, and there are no clear rules of how to attribute these withholding payments to the different taxpayers (Box 5.5). Since withholding taxes are paid by both individual and corporate taxpayers, these receipts have to be allocated first to individual and corporate income taxes, and then within individual income taxes among the various individual tax returns. We were guided in these attributions by the objective to replicate FBR's reported outcomes for the different taxpayer categories as well as explanations of FBR staff on the attribution rules for withholding tax payments. These considerations lead us to attribute salaried withholding receipts to salaried income tax payers, a small amount of the withholding receipts to retailers, and the bulk to the withholding receipts to the self-employed. Second, a second aspect in modeling withholding receipts concerns the issue of adjustability: do taxpayers incorporate their "adjustable" withholding tax payments in their annual income tax return to lower their overall tax burden? Unfortunately, there are no data available to determine which payments are actually adjusted. The assessment of FBR staff suggests many adjustable withholding tax payments are final payments. Since de jure "final" withholding tax receipts should clearly be treated as final, and many de jure "adjustable" withholding tax receipts appear to be de facto final, we assume that all withholding tax payments attributed to individual taxpayers are effectively final payments. Third, in order to obtain revenue impacts of changes in tax rates, we require an estimate of the tax base. In general, we derive current tax bases through information from income tax returns. For example, in the case of withholding taxes, we divide the reported withheld amount by the statutory tax rate to estimate the current tax base. We use the same procedure for capital income. Tax returns of retailers give us data on annual turnover but not on input costs. For the analysis, we assume a profit rate of 20 percent on turnover. Four, there are modeling issues in terms of the 2005/06 baseline (i.e., gaps between simulated and reported tax liabilities through duplicate returns, input errors, or other factors) and the updating of the 2006 model to a 2008 model (i.e., the changes in tax schedules, thresholds and allowances would have changed the composition of tax returns). Five, the simulations capture only static first-round effects that over time would be dominated by behavioral responses of taxpayers to the changes in tax structure. In particular, we generally assume that compliance and enforcement of the system will remain at the same level as in 2005/06. Such "current law compliance" is established by comparing actual tax payments with full-compliance tax payments under statutory tax rates and the current tax base. Assuming no change in compliance is a reasonable for most tax reforms. However, in the case of the withholding tax reforms, the net effect of the reform increases the withholding rate substantially for most withholding categories. This might make it unlikely that the entire potential revenues would actually be collected. Therefore, we also report revenue estimates assuming that 25 percent and 50 percent of the estimated potential liability for the withholding taxes can actually be collected. 18 2.5 Federal Excise Duties 2.19. There is room to expand the role of excise taxes in Pakistan's revenue collection by enhancing their use in addressing certain negative externalities. Rationalizing the current rate structure of petroleum products, and increasing excise taxation of petroleum products is in line with these objectives. Additional areas of reform include converting ad valorem rates to in rem rates for tobacco products and fully indexing the in rem rates for all excises to the price level, and raising ad valorem rates on selected luxury items (Box 7.2). A related reform objective is the introduction of "green taxes" on effluents to help curb polluting activities. The revenue impact of such reforms depends primarily on the choice of tax rates. For example, the specific reform of tobacco excise taxation presented in Box 7.3 would be to raise excise tax collection from tobacco by about 50 percent, or Rs. 14 billion in 2007/08. Adding additional revenues from selected luxury goods could lift federal excise tax revenues by 24 percent, or Rs. 22 billion in 2007/08. Box 2.10: Modeling General Sales Tax The GST estimation is based on an 81 sector input-output table of the Pakistan economy calibrated to replicate 2004/05 sectoral GDP. The current law tax treatment is compared to the proposed tax treatment on the bases of effective tax rates for each sector. The effective tax rates are derived from four parameter: · the taxable portion: the percent of output that is subject to GST; · the adjustment for threshold: the percent of output that is produced by businesses with gross sales greater than the GST threshold; · compliance rate: the percent of the tax liability that is assumed to be actually paid; and · statutory tax rate. The estimates do not account for any behavioral response that may result from the increase in GST rates. Clearly, increasing the GST rate from 0 to 16 percent in some sectors could decrease the compliance rate and lower the consumption tax base, which in turn would partially offset the estimated revenue increase. Yet, eliminating zero-rated and exempt sectors would also close avenues for tax evasion, and thus increase GST revenues through better compliance. The net effect of these behavioral responses to an increase in the coverage of the GST is ambiguous. Box 2.11: Federal Excise Tax Reform (i) The federal excise rates on tobacco are changed from ad valorem to in rem and increased moderately. (ii) The federal excise rate tax is imposed on a selected list of luxury goods. Figure 2.4: Revenue Impact of Federal Excise Tax Reform Luxury Goods (9%) Tobacco (15%) Actual (100%) 19 2.6 Customs Duties 2.20. The overriding long term goal for customs duties reform is to spur economic growth through trade facilitation. This agenda includes reducing the tariffs dispersion for different commodities in order to bring down the effective rates of protections, and to reduce the overall level of tariffs. Tariff reform will also require reducing the number of rates outside the regular tariff bands and eliminating special exemptions. Since custom duties still contribute a sizable amount to revenue collection, these reforms could have implications for revenue collection. In order to mitigate the risk of revenue shortfalls, custom tariff reform has to be well sequenced with the expansion of the domestic tax base. 2.21. The long term reform scenario creates three tariff rates with duty rates of 0 percent for unprocessed goods, 5 percent for intermediate goods, and 10 percent for final goods. Everything else equal, this reform would reduce revenues by Rs. 128 billion. However, the proposed decrease in tariff rates is likely to increase the value of imports as lower prices would lead to higher volumes. This would not just increase import duties but also GST collections since GST is applied to the import price inclusive of customs duty; as well as federal excise collections since federal excise duties are applied to the import price inclusive of customs duty and GST. Allowing for these behavioral responses, we estimate the net effect of the custom tariffs reform to be a decline in tax revenues of Rs. 65 billion. These estimates assume that the customs services collect GST and federal excise duty at the import stage, even on imported goods with a zero duty rate. Box 2.12: Customs Duties Reform ­ Three-Tier Structure Customs duties are rationalized to establish a three-tier structure: (i) Duty rate of 0 percent for unprocessed goods (ii) Duty rate of 5 percent for intermediate goods (iii) Duty rate of 10 percent for final goods. Box 2.13: Modeling Custom Tariffs We use data on all imports from Germany, India, Japan, and the United States for fiscal 2004/05 at the 8-digit commodity level. The value of imports from these four countries accounts for approximately 30 percent of total imports in fiscal year 2004/05. We adjust the value of imports from these four countries to hit the customs duty forecast of Rs. 151 billion for fiscal year 2007/08. We use an elasticity of -0.5 to further adjust the value of imports for the behavioral response from the decrease in the prices of imports due to the proposed reduction in tariffs. Finally, we account for the potential effect of the change in the point of collection of GST from the customs service to the domestic tax administration for those goods that are not dutiable under the proposal by accounting for the difference in the compliance rates at these two points of collection. Based on the tax gap study, the GST compliance rate is approximately 40 percent when collected by domestic tax administration and approximately 79 percent when collected by customs service. Finally, the proposal accounts for the effect that a change in indirect tax collections would have on national income and if national income changes then direct tax collections should also be expected to change. This effect is due to the following accounting identity: "GDP = national income + indirect business taxes". Assuming that GDP is constant, a change in indirect business taxes implies an offsetting change in national income which is the base of direct taxes. We assume that the effective average marginal direct tax rate is 5 percent. Figure 2.5: Revenue Impact of Custom Duties Reform Tariff Rationalization (-43%) Actual (100%) 20 2.7 Provincial Taxation 2.22. Increasing tax revenue efforts by the subnational governments is crucial to the success of the overall tax reforms. In 2007/08, provincial governments raised only 0.4 percent of GDP in tax revenues. Clearly, it would take a dramatic increase in the subnational tax effort to have a sizable impact on national tax collection. Chapter 9 presents a reform package that could deliver increases of 115 percent and 143 percent in provincial tax collection in Punjab and NWFP, respectively. In view of the similarities of the economies, tax systems and tax policies, it is plausible that increases of the same order would be achievable through similar reforms in Balochistan and Sindh. For the purposes of calibrating the overall impact of provincial tax reform on national tax collection, we assume that provincial tax reforms could boost provincial tax collection by 110 percent, or from Rs. 41 billion to Rs. 86 billion in 2007/08 (Figure 2.6). This estimate excludes any increase from the provincial taxation of services, which is already captured in the GST reform. 2.23. Chapter 9 discussed the formidable challenges involved in encouraging provincial governments to undertake tax policy reforms. One crucial aspect is the need for a change in the incentives for provincial tax collection. Currently, federal transfers through the NFC Award are unrelated to any revenue effort made by provincial governments. The NFC awards could be restructured in order to reward increased provincial revenue mobilization. To be successful, such incentives would have to be large enough to draw out a greater tax effort, but not too large to harm equalization. 2.24. The most important reform area of federal-provincial fiscal relations is reaching an agreement on taxing all goods and services under a general value added tax. Under agreement with the provinces, the federal authorities have already introduced a number of excise taxes on services. What is required now is a more general agreement for the federal authorities to tax goods and services under a broad-based simplified GST ­ with or without constitutional amendment. As we have seen above, such a reform would yield considerable revenues. The two elements for which agreement will be needed are, first, the federal share or aggregate provincial share; and, second, a formula to apportion the provincial share across the provinces. Achieving this will not be easy as such reform would require consensus among provinces and federation. The sizable increase in revenues that a broad based GST on goods and services can generate would be the main incentive for reaching such an agreement. Figure 2.6: Revenue Impact of Provincial Tax Reform Reform Package (110%) Ac tual (100%) 21 2.8 Way Forward Overall Reform Package 2.25. The long term structural tax policy reforms outlined in this chapter could produce substantial revenue increases. In 2007/08, total tax revenue collections stood around Rs. 1.1 trillion or 10.3 percent of GDP. We estimate that the proposed tax policy reforms could raise tax revenues by Rs. 400 billion, equal to 38 percent of 2007/08 national tax collection, or 3.8 percent of GDP (Table 2.1, Figure 2.7 and Figure 2.8). Among the six taxes, we have identified four revenue gainers and two revenue losers: · By far the most important revenue impact would come from GST reform. The adoption of a broad based GST on goods and services in agreement between the federal and provincial governments is therefore the central tax policy reform component. This reform would produce an increase in revenues of Rs. 408 billion, assuming exemptions for seven sectors. This is slightly larger than the overall revenue impact of the national tax policy reforms. · Provincial tax reforms would be the second most important revenue winner, contributing some Rs. 45 billion, or 11 percent of the overall revenue increase. · Introducing a two-tier structure for individual income tax and imposing a 10 percent rate on withholding taxes would raise tax collection by Rs. 34 billion, or 9 percent of the overall revenue increase. · Reforming the federal excise taxation of tobacco would lift tax revenues by Rs. 14 billion, or 4 percent of the overall revenue increase. · The reduction in the corporate income tax rate to 30 percent would reduce tax revenue by Rs. 38 billion, or 10 percent of the overall revenue increase. · The three-tier structure of customs duties would lead to a tax revenue loss of Rs. 65 billion, or 16 percent of the overall revenue increase. 2.26. These numbers are only indicative and rely on crucial assumptions. For example, the analysis is based on the economic situation of the last few years, where economic growth was robust, inflation low, businesses did well, and government spending on the rise. Pakistan's economic situation today is different: growth is sluggish, inflation high, enterprise profits dropping, and government spending on the decline. In addition, the actual outcomes of tax policy reform would differ because of the changes in the behavior of taxpayers and the economy in response to the changes in tax policy. Finally, some components of a tax policy reform could not be calibrated. In particular, the losses in corporate income tax revenue from the reduction in the rate would be at least in part compensated by a broadening of the tax base. 2.27. While the revenue impact of these reforms remains to some degree uncertain, these structural reforms are in line with principles of good tax policy and provide other benefits than just higher tax collection. First, while the reforms would increase tax burdens for all income groups, the proportional increase is largest for the seventh to the ninth household decile (Figure 2.9). The overall distribution of tax burdens would remain progressive. Second, the new tax system would yield a more horizontally equitable tax system due to the removal of many exemptions and other preferential tax treatments. This would also improve the overall investment climate in the country, as the equalization and general lowering of the marginal effective tax rates across sectors reduces the excess burden of taxation on the economy. The reformed tax structure would also be easier to administer, which would lead to improved enforcement and compliance, and therefore more horizontal equity, and more certain revenue flows. Table 2.1: National Tax Reform ­ Simulations Taxes Simulations General Sales Tax (GST) Seven-Sector Exemptions Corporate Income Tax (CIT) Two-Tier Structure Individual Income Tax (IIT) Two-Tier Structure Federal Excise Duties (FED) Tobacco Excise Customs Duties (CD) Three-Tier Structure Provincial Taxes (PT) Reform Package 22 Figure2.7: Revenue Impact of National Tax Policy Reform ­ Display One CIT (-4%) CD (-6%) FED (1%) PT (4%) ITT (3%) GST (39%) Actual (100%) Figure 2.8: Revenue Impact of National Tax Policy Reform ­ Display Two 2007/08 Revenue Impact of National Tax Reform 40 400 35 350 30 300 25 250 20 200 Rs. Billion 15 150 10 100 % of 2007/08 National Tax Collection 5 50 0 0 -50 -5 GST PT IIT FED CIT CD -100 -10 Figure 2.9: Impact of National Tax Reform on Vertical Equity Incidence of National Taxation Pre- and Post-Reform by Household Consumption Deciles 45 20 Effective Tax Rate (% of GDP) 18 40 16 35 14 30 12 25 10 20 8 15 PRE-REFORM POST-REFORM 6 10 4 Share in Total Taxes (%) 5 2 0 0 Bottom 2 3 4 5 6 7 8 9 Highest 23 Meeting the Revenue Target 2.28. Pursuing the twin track reforms of tax policy and tax administration would put the government in good stead to meet its medium-term revenue collection target. Our simulations show that a fairly simple and comprehensive tax policy reform package could boost tax-to-GDP ratio by 3.5 percentage points of GDP to meet the 2012/13 tax collection target of 13.9 percent of GDP (Figure 2.10). However, two caveats are important. The simulations relate to the economic situation prior to the current economic crisis. These simulations also assume no changes in enforcement and compliance relative to current levels. Tax compliance could increase with a more simplified and uniform tax system, but it could also fall with the overall increase in the tax burden. This makes it so important to complement the reforms of tax policy with the reform of tax administration. Enhancing tax enforcement would contribute to revenue increases through a reduction in the tax gap. After all, our estimates suggest that the revenue potential from eliminating the federal tax gap alone (7.2 percent of GDP) is even larger than from implementing the national tax policy reform package (3.8 percent of GDP) ­ even allowing for the fact that a full elimination of the tax gap is not desirable, since tax administration and tax compliance are costly. Figure 2.10: Raising Tax Revenues: Actual Collection, Additional Collection Target, Tax Policy Reform Impact, and Tax Gap Actual (100%) Additional Collection Target (35%) Federal Tax Gap (72%) Tax Policy Reform (38%) 24 Sequencing Tax Policy Reform 2.29. Any successful reform has to take along three core stakeholders: the tax administration, the taxpayer and the policy maker. First, meeting this goal will not only be about changes in tax policy but also about changes in tax administration. In the past, reform initiatives failed because there was no adequate focus on ensuring that tax policy reforms can be well administered. Federal tax administrators have already made good progress in a number of areas, but more needs to be done to complete FBR's modernization program over the next few years (Government of Pakistan 2008). At the provincial level, such a reform effort still has to be launched. Second, although overall tax burdens in Pakistan are low, the distribution of those tax burdens among households and among economic sectors is uneven. Those taxpayers already paying taxes and having to compete in an increasingly globalized world are likely to oppose reforms if these reforms are perceived as further increasing their tax burdens while doing little to tax other individuals and businesses that now are paying no or little taxes. Third, the fiscal position of the government is weak and could deteriorate further in view of the economic slowdown. This will make policy makers reluctant to provide leadership on tax reforms that may be considered uncertain in their revenue impact, especially if they face opposition from interests groups. 2.30. Because some of the ills afflicting Pakistan's tax system have been the results of past policy measures, it will be important to spend time and effort to generate broad consensus for the reform package. Many tax incentives and preferential treatments that have been around for some time have become entrenched and will be hard to remove. The political economy of tax reforms in other countries suggests the importance of having a broad package that asks for the sacrifices of many but also offers general advantages including simpler taxes and reduced tax rates. In the case of a gradual implementation of the reforms (Table 2.4), there is a risk that the politically easier reforms will be picked first and the more controversial reforms will never be delayed and ultimately derailed. One way to guard against this risk is to separate the approval of the tax reform package, which could be done upfront for the whole agenda, and the timing of its implementation, which could be done sequentially. This approach could also help in overcoming resistance from particular groups, such as sectors affected by the removal of tax preferences, by highlighting the overall gains from the whole reform effort. 2.31. The right sequencing of tax policy reform will be important to sustain reforms. If some of the expected outcomes of the reforms do not fully materialize, a backlash against change could stall the reform implementation. This highlights the importance of a solid analytical preparation (Table 2.3) and roadmap for the successful reform of each major tax (Table 2.5). 2.32. Ensuring early gains should be part of the implementation strategy. The broadening of tax bases should have priority compared to rationalization of taxes, and especially compared to tax rate reduction. Across taxes, GST reform should be the first priority, followed by reforms of provincial taxes, individual income tax and federal excise tax, while the reforms of custom duties and corporate income tax should have the lowest priority (Table 2.2). For example, trade reform with a reduction and unification of the customs tariffs is desirable, but it should be put on hold until it is certain that the revenue losses from trade liberalization would be offset by larger collections from domestic taxes, especially the GST. When we combine tax by tax the revenue potential from tax policy reform with the potential from tax gap elimination, corporate income tax and GST are the most important tax reform areas (Figure 2.11). Key milestones for implementing the reform agenda are the 2009/10 and 2010/11 budget laws, as Pakistan typically introduces changes to tax legislation jointly with the budget law. Table 2.2: Sequencing Tax Policy Reform Across Taxes and Issues Broadening of Tax Bases Rationalization of Taxes Reduction in Tax Rates FIRST PRIORITY THIRD PRIORITY GST PT, IIT, FED CD, CIT 25 Figure 2.11: National Tax Reform ­ Size of Tax Policy Impact and Tax Gap for each Tax 2007/08 Revenue Potential from Tax Policy Reform and Tax Gap (Rs. Billion) 600 600 500 500 Tax Policy & Tax Gap 400 400 300 300 Tax Gap 200 200 100 100 0 0 Tax Policy -100 -100 CIT GST FED IIT CD -200 -200 Table 2.3: Analytical Preparation of Tax Policy Reform All Taxes · Harmonize tax definitions and tax procedures for all domestic taxes and modernize the delegation framework for the revenue laws · Conduct economic analysis of specific proposals for tax policy changes, including expansion of the tax bases and adjustment in rates. This work includes the following: Individual Income Tax · Develop reformed system of tax credits · Develop transition plan from current system of multiple tax brackets to two-tier tax rates Corporate Income Tax · Establish clear attribution of withholding tax collection to salaried taxpayers, non-salaried taxpayers and corporations · Review each withholding tax with regard to its revenue impact, base, rate and adjustability · Evaluate treatment of debt and equity · Assess potential to widen corporate income tax base through the elimination of exemptions, investment incentives and special treatments General Sales Tax · Develop plan for widening successively the tax base · Develop revenue sharing formula between federal government and provincial governments Federal Excise Duties · Assess scope for increasing taxation of luxury goods and services · Develop reformed system of petroleum products Custom Duties · Assess scope for expanding the base of dutiable imports · Develop transition plan from current multiple duties rates structure to three-tier duty rate structure Provincial Taxes · Conduct provincial tax policy analyses for Balochistan and Sindh · Develop reform proposals for revised National Finance Commission Award that include incentives for provincial revenue effort 26 Table 2.4: Piecemeal and Comprehensive Tax Policy Matrix Reform Type Proposal Individual Income Tax · Introduce taxation of short-term stock market related capital gains · Tax stockbrokers' income according to the non-salaried income tax schedule · Broaden the tax base by reducing the zero-rate income thresholds in real terms · Replace the two basic schedules with single schedule accompanied by tax credits for earning labor income or for female labor force participation · Remove notch problem by reverting to conventional method of taxing personal incomes Piecemeal progressively · Reduce significantly the number of tax brackets · Streamline system of personal tax credits (charity; pensions; purchases of new shares; purchases of housing) · Ensure that tax credits deliver the same amount of tax relief regardless of the taxpayer's taxable income Comprehensive · Introduce broad-based two-tier individual income tax Corporate Income Tax · Limit the use of tax exemptions and tax incentives for industrial policy · Make withholding taxes adjustable for formal economy · Limit favorable withholding tax rates for informal economy Piecemeal · Eliminate withholding taxes that generate only small revenues · Narrow definition for small businesses by reducing turnover threshold in real terms · Make small businesses withholding agents · Strengthen rules for international tax provisions Comprehensive · Introduce corporate income tax with lower rate and wider tax base General Sales Tax · Redraft the sales tax act to convert general sales tax into modern VAT, designed as shared federal- Piecemeal and provincial tax covering both goods and services and under federal administration Comprehensive · Expand successively the base of the sales tax Federal Excise Duties · Bring federal excise legislation in line with the sales tax law · Convert the existing ad valorem rates into in rem rates for tobacco products Piecemeal and · Index in rem rates for inflation Comprehensive · Employ excise duties as green tax and simplify the rates structure for petroleum products · Bring additional luxury goods and services under excise taxation Customs Duties · Combine tariff reductions with a sequence of separate measures to avoid overall revenue losses, including elimination of existing tariff exemptions (especially at the 6-8 digit level), increases in Piecemeal excisable imports; and expansion of the yield from general sales tax · Reduce tariff dispersion, especially at disaggregated level Comprehensive · Introduce customs duties with three tariff rates for unprocessed, intermediate and processed goods. Provincial Taxes Piecemeal · Strengthen major provincial taxes selectively Comprehensive · Strengthen consolidated system of provincial taxes and provide incentives for revenue effort 27 Table 2.5: Sequencing Tax Policy Reform for each Taxes Time-Period Proposal Individual Income Tax · Tax stockbrokers' income according to the non-salaried income tax schedule · Broaden the tax base by reducing the zero-rate income thresholds in real terms Short-Term · Remove notch problem by reverting to conventional method of taxing personal incomes progressively · Introduce taxation of short-term stock market related capital gains · Replace the two basic schedules with single schedule accompanied by tax credits for earning labor income or for female labor force participation Medium-Term · Streamline system of personal tax credits · Ensure that tax credits deliver the same amount of tax relief regardless of the taxpayer's taxable income · Introduce broad-based two-tier individual income tax Corporate Income Tax · Limit the use of tax exemptions and tax incentives for industrial policy · Narrow definition for small businesses by reducing turnover threshold in real terms Short-Term · Make small businesses withholding agents · Eliminate withholding taxes that generate only small revenues · Make withholding taxes adjustable, especially for formal economy · Limit favorable withholding tax rates for informal economy Medium-Term · Strengthen rules for international tax provisions · Lower tax rate when tax base wide enough to ensure adequate revenues General Sales Tax · Redraft the sales tax act to convert general sales tax into modern VAT, designed as shared Short-Term federal-provincial tax covering both goods and services and under federal administration Medium-Term · Expand successively the base of the sales tax Federal Excise Duties · Convert the existing ad valorem rates into in rem rates for tobacco products · Index in rem rates for inflation Short-Term · Employ excise duties as green tax · Bring additional luxury goods and services under excise taxation · Simplify the rates structure for petroleum products Medium-Term · Bring federal excise legislation in line with the sales tax law Customs Duties · Eliminate tariff exemptions, especially at the 6-8 digit level Short-Term · Reduce tariff dispersion, especially at disaggregated level · Introduce customs duties with three tariff rates for unprocessed, intermediate and processed Medium-Term goods once tax base wide enough to ensure adequate revenues Provincial Taxes Short-Term · Strengthen major provincial taxes selectively Medium-Term · Strengthen consolidated system of provincial taxes and provide incentives for revenue effort 28 2.9 What Next? 2.33. In view of Pakistan's federal structure, overcoming bottlenecks of tax policy reform will require national consensus. One option for generating such consensus would be through blue ribbon commissions: one at the federal level and four at the provincial level (the provincial commissions are discussed in Chapter 9). These commissions should have broad representation from federal and subnational levels, and public and private sectors. Many countries have successfully implemented tax policy reforms through broad-based tax reform commissions with representation of stakeholders from inside and outside of government, with the mandate to develop time- bound action plans to the government for comprehensive tax reform. This model often includes the appointment a technical committee to provide the commission with high standards of background research, options, and simulation of results. Pakistan already has a strong tradition of appointing commissions and task forces for tax policy and tax administration reform (Task Force on Tax Administration Reform 2001; the Commission on Tax Reform 1998; the Pasha Committee 1994; and the Tax Reforms Committee 1991). Alternatively, rather than setting up blue ribbon commissions, Pakistan could generate a consensus around tax reform by drawing on existing institutions and commissions, including the Ministry of Finance with its cross-ministerial convening power, the National Finance Commission, with its federal and provincial representation, the Economic Advisor Council constituted by the Prime Minister in May 2008, and the Planning Commission's Panel of Economists constituted by the Deputy Chairman of the Planning Commission in September 2008. 2.34. The goal of the federal commission would be to develop a comprehensive tax policy reform program, based on the government's objectives and national and international experience in tax reform. The terms of reference would include three major tasks: developing a roadmap for reform to put in place an equitable and efficient tax structure that generates revenues adequate for Pakistan's development needs; addressing the problems with tax assignment across layers of government; and providing incentives for subnational revenue effort. Key milestones for implementing the reform agenda are the 2009/10 and 2010/11 budget laws, as Pakistan typically introduces changes to tax legislation jointly with the budget law; as well as the completion of the new, eight, National Finance Commission award which is expected during 2009. 2.35. Beyond promoting the work of the federal and provincial commissions, the government needs to strengthen its capacity for tax policy analysis and planning. The scope and quality of tax policy formulations depends critically on the availability of well trained and experienced staff to conduct different forms of fiscal analyses ranging from revenue forecasts to micro-simulations based on taxpayer return data. A main weakness of government's tax policy practice has been its inability to withstand the pressures from vested interest groups, which has introduced a number of concessions, and eventually distortions, into the tax system. These concessions have often found their way into tax legislation because of the government's inability to evaluate the impact of these measures on the tax system and the economy has a whole. 2.36. There is also a need for strengthening the fiscal analysis capacity both at FBR and the currently dormant revenue division at the Ministry of Finance. The work at FBR would be more micro-oriented based on tax return data, and the work of the revenue division would be more macro-oriented to ensure consistency with the fiscal and macroeconomic frameworks. The revenue division should take lead responsibility in the formulation of fiscal policy. This would allow FBR to focus on its task of administrating and enforcing tax legislation. 29 CHAPTER 3: STRUCTURAL PROPERTIES 3.1 Introduction 3.1. Tax systems are not just about raising revenues. This primary objective of financing government expenditures is sometimes termed "adequacy". Corresponding, dynamic concepts relate to the ability to generate adequate ("buoyancy" and "elasticity") and predictable ("stability") resources over time in line with growing expenditure requirements. But while raising revenue is important, taxes should in general not influence the behavior of those who pay them unduly; in choosing taxes, a common goal is to minimize the interference of taxes in the economic decisions of individuals and firms ("efficiency"). Another objective is to distribute the burden of taxation in a way that meets with a society's notions of fairness; such "equity" is typically defined in terms of "ability to pay", such that those with greater ability should pay greater taxes ("vertical equity"), and those with equal ability should pay equal taxes ("horizontal equity"). Finally, taxpayers should adhere with the tax system ("compliance"). This means the tax system should make it easy to comply with tax laws and be simple to administer. 3.2. This chapter subjects Pakistan's tax system to a basic health check and assesses the tax system with regard to the four dimensions of adequacy, efficiency, equity and compliance. While many features of Pakistan's tax system are in line with international practice (Table 3.1), the examination also reveals important structural problems. For starters, the tax system has failed to deliver sustained increases in the tax-to-GDP ratio over the last few decades. While the revenue performance improved over the last few years, the pace of progress is slow and indeed halted altogether in 2007/08. Today's tax-to-GDP ratio remains inadequate to meet Pakistan's development agenda. These shortcomings are, in part, the direct result of exemptions and special regimes, which, in combination with administrative problems, also weaken efficiency, equity and compliance and undermine the overall coherence of the tax system. Box 3.1: Brief Overview of Pakistan's Tax System In 2007/08, Pakistan collected approximately Rs. 1,050 bn in tax receipts, of which the provinces collected some Rs. 40 bn (4 percent of national tax revenues). Not unlike other developing countries, about two-fifth (Rs. 389 bn or 39 percent) of the federal taxes come from direct taxes, and three-fifth (Rs. 617 bn) from indirect taxes. A direct tax is one paid directly to the government by the persons (juristic or natural) on whom it is imposed. An indirect tax is one which is collected by intermediaries who turn over the proceeds to the government. Federal direct taxes consist of an income tax, payroll taxes, a capital value tax, and a wealth tax. About 95 percent of the direct tax revenues come from the income tax. Taxable income under the individual income tax consists of income derived from wages and salaries, rental income, business income, capital gains, except for stock market capital gains, and other sources of income. There is a progressive tax rate schedule that increases gradually from zero to 20 percent. There is a long list of deductions, exemptions, credits, and preferential tax rates. The base of the corporate income tax is conventionally defined as income earned from the sales of goods and services, and any income received in the form of rents, interest, management fees or royalties. A fairly usual list of business expenses is deductible from chargeable income. Pakistan, with the exception of small companies, applies a uniform corporate tax rate of 35 percent. In addition, there is an extensive network of withholding measures that include many forms of presumptive taxation. To a large extent, Pakistan's withholding tax regime consists of familiar withholding taxes on income at source which has proven to be an effective means of increasing tax compliance in many countries. But in addition to these familiar source based withholding taxes, there is an important, in revenue terms, regime of what are essentially excise and transactions-type taxes, which are characterized in the reports and financial statements of the FBR as withholding taxes. This part of the withholding regime may be unique to Pakistan. Some of these withholding taxes are considered to be final, and others are considered to be advanced or adjustable taxes. The indirect tax system consists of a sales tax, a federal excise tax, and import duties. Sales tax, in the form of a invoice-credit value-added tax, is imposed on all traders, other than those involved in providing services, with a turnover above the registration threshold of Rs. 5 mn ($62,500). A zero rate regime applies to exports and a number of other sectors, and some sectors are exempt. The sales tax rate was increased from 15 percent to 16 percent in 2008/09. The federal excise tax base consists of 48 excisable goods. These can be grouped into six categories, vegetable oils, carbonated beverages, tobacco products, cement, and 17 different types of oils and fuels, and certain luxury items. Finally, customs duties are applied to imports. The tariff schedule generally provides for duty rates, ranging from 0 to 25 percent in 5 percentage point increments, i.e., 0 percent, 5 percent, 10 percent, and so on, with tariff peaking beyond these rates for imports of automobiles, automobile parts, aerated beverages, and tobacco products. While the constitution assigns the federal government the more buoyant and easy-to-reach taxes, provincial taxes include the property and property transfer taxes, motor vehicle taxes, the agriculture income tax, and the sales tax on services. 30 Table 3.1: Main Findings of Chapter 3 Topic Findings Tax-to-GDP Ratio Pakistan's development needs require that the tax-to-GDP ratio should increase by at least 3.2 percent of GDP over the next five years. This would raise Pakistan's tax effort to 13.2 percent of GDP, roughly equal with the simple average of the tax collection in Bangladesh, India, Nepal and Sri Lanka, as well as the figure for Indonesia. Narrow Tax Base In 2006/07, relative to its share in GDP, industry paid 60 times more taxes than agriculture, and 5 times more taxes than services. Among the 24 subsectors, there are 10 subsectors whose tax contribution is below its GDP contribution: textile, ownership of dwellings, agriculture, construction, transport and communication, wholesale and retail trade, food and beverages, public administration, other services and telecom. All four major federal taxes rely heavily on the performance of a few sectors. Long-Term Buoyancy The overall buoyancy of the FBR taxes declined from above unity in the 1960s and 1970s to below unity in the 1980s and 1990s, and increased above unity in the 2000s, although it remains below the levels of the 1960s and 1970s. At 1.05, it is rather low: a one percent increase in GDP implies an increase in the tax-to-GDP ratio of only 0.05 percent (that is 5 percent of 1 percent). Short-Term Buoyancy At 9.6 percent of GDP, federal tax collection is about the same in 2007/08 as in 2002/03, and only some 0.5 percentage points of GDP higher than in 1999/2000. Federal Tax According to government estimates, tax expenditures of the federal government, calculated Expenditures as the estimated revenue loss in percent of GDP, declined from around 0.8 percent in 2000/01 to 0.33 percent in 2005/06, and then increased to 2.1 percent in 2006/07 and 0.9 percent in 2007/08. Federal Revenue Since 1999/2000, fiscal year outcomes for FBR collection fell short of what should have Targets been expected under the realized growth factors and budgeted revenue measures in four out of the last eight years. Provincial Tax Effort The share of provincial taxes in provincial revenues is very low. In 2006/07, it was no more than 2 percent in Balochistan, 4 percent for NWFP, 9 percent for Sindh, and 8 percent for Punjab. Sindh collected no more than Rs. 390 in own-source tax revenues from each of its citizens, Punjab Rs. 230, NWFP Rs. 110 and Balochistan Rs. 100. Excess Burden Pakistan's tax system provides economic agents with an uneven playing field. The special treatment of particular sectors and sources of incomes through differences in tax bases and tax rates is widespread. This distorts the allocation of resources as too many resources are allocated in the lightly taxed sectors and too few in the more heavily taxed sectors. Investment Climate In the area of paying taxes, the 2008 Doing Business survey ranks Pakistan only 146th out of 178 countries. Out of the ten areas covered, the area of "paying taxes" is Pakistan's 2nd worst ranking. Vertical Equity Pakistan's overall tax burden is somewhat progressive. The lowest decile of households pays 2 percent of all taxes, while the highest decile pays 41 percent of all taxes. This compares to a share in total income of less than 3 percent for the lowest deciles, and more than 32 percent for the highest decile. Horizontal Equity Pakistan's tax system has a large degree of horizontal inequity. For example, among households in the same income group, some households consume a larger share of low-tax goods and services than others, and some households earn a larger share of low-tax income sources, such as income derived from capital, than others. Tax Gap Pakistan's federal tax gap in 2007/08 is estimated to be 76 percent of actual tax receipts. This sum amounts to Rs. 4,600 worth of cheating by every man, woman and child in Pakistan. Tax Compliance Since taxpayers become more compliant when they are more likely to be caught, the large size of the tax gap suggests that Pakistan's enforcement measures are vastly inadequate. Taxpayers are also likely to be more compliance when they perceive that the money is put to good use in the delivery of public services. 31 3.2 Adequacy Tax-to-GDP Ratio 3.3. The government intends to increase Pakistan's tax-to-GDP ratio by at least 3.2 percent of GDP over the next five years. Chapter I argued that Pakistan's development needs require an additional tax effort of this magnitude. This would raise Pakistan's tax effort to 13.2 percent of GDP, roughly equal with tax collection in Sri Lanka. This target is clearly ambitious. Only few countries managed to make such large and rapid progress. India and South Africa did increase tax collection by about 4 percentage points of GDP between 2002 and 2007, but they are rare exceptions. Even then, they were helped by rapidly growing tax bases of sales, incomes and profits on the back of a strong economic expansion. Figure 3.1: Inadequate Revenue - Factors behind the Low Tax-to-GDP Ratio Narrow Tax Base Low Buoyancy Large T ax Exemptions Low Tax Compliance Tax Policy Tax Administration 3.4. The reasons for the low tax effort in Pakistan relate to large tax exemptions and low tax compliance. As Pakistan's tax rates are not unusual by international standards, the low tax-to-GDP ratio is linked to the narrow effective tax base. The effective tax base is narrow as many economic activities are lightly taxed or fully exempt by law (large tax exemptions); in addition, many economic activities contribute only a part or even nothing of their legal tax obligation (low tax compliance reduces the effective tax base relative to the statutory tax base). Large tax exemptions and low tax compliance often imply that the growing activities in Pakistan's economy contribute too little to the overall tax effort. For example, the recent reduction of tax rates, such as the introduction of the 15 percent single rate for GST and the move to a uniform 35 percent tax rate on companies and individuals, was not compensated through a corresponding enlargement of tax bases. Similarly, the reduction in duty rates on imports and excisable goods did not go hand in hand with an expansion in tax collection from other domestic goods and services. The narrowness of the tax bases adds to the low buoyancy of the tax system. 3.5. Tax exemptions are the result of large federal tax expenditures and the poor mobilization of the provincial tax base. There are various ways to categorize the overlapping and mutually reinforcing factors behind large tax exemptions, but it might be useful to highlight two. First, tax policy itself has contributed to lackluster collections by defining very narrow tax bases for the major federal taxes. This is reflected in a large number of for tax expenditure, as well as, in many years, fairly low tax policy measures introduced at the time of the budget. Second, provincial own-source revenue mobilization is weak, as discussed in detail in Chapter 9. According to the 1973 Constitution, agricultural income, capital gains on physical assets, the consumption of services, motor vehicles, and urban property are exclusively provincial tax bases. Yet, almost no tax is imposed on the first two bases, little is mobilized from wholesale, retail, and transport, and Pakistan's housing market boom over the last few years appears to have bypassed completely the subnational exchequers. The poor provincial performance is linked to weaknesses in the overall intergovernmental fiscal architecture. 3.6. Low compliance with tax laws reflects a weak tax administration and a large shadow economy. Some taxpayers evade taxes simply because of greed. Others evade taxes because they see little point in funding a government that they view as poorly governed and at times corrupt. For both groups, FBR's limited capacity to audit tax returns and enforce tax payments, including in the hard-to-tax sectors, further reduces the incentives to comply. The weaknesses in tax administration are so damaging for tax collection because of the structure of the economy. There is a large underground economy and informal sector, including agriculture and retail, where the bulk of transactions is in-cash or in-kind leaving behind little paper trail. The issue of compliance is taken up again at the end of this chapter, but for now we are turning the focus on the narrow tax base. 32 Narrow Tax Base 3.7. Differences in compliance and exemption levels as well as tax rates imply that some economic activities are taxed more heavily than others. In 2006/07, agriculture contributed about one fifth of GDP yet no more than one percent in FBR's tax revenue (Figure 3.2). Services made up almost three-fifth of economic value added but only one third of central taxes. Given the shortfall in agriculture and services, industry carries the brunt of the tax burden. Relative to its share in GDP, industry pays 60 times more taxes than agriculture, and 5 times more taxes than services. Breaking down the sectors further shows even more accentuated differences (Figure 3.3, left panel). Among the 24 subsectors, there are 10 subsectors whose tax contribution is below their GDP contribution: textile (whose tax contribution is negative), ownership of dwelling, agriculture, construction, transport and communication, wholesale and retail trade, food and beverages, public administration, other services and telecom. Clearly, this does not in itself demonstrate that these activities are under-taxed. After all, GDP is not synonymous with tax base, and some activities, like petroleum, are heavily taxed for good reasons. Also, the calculations concern FBR taxes only (although accounting for provincial taxes would not make much of a difference as they make up only 5 percent of overall taxes). Nevertheless, these differences in tax and GDP contributions are indicative of the potential for widening the tax base across activities and the direction of required tax reforms. Figure 3.2: Contributions of the Three Sectors to GDP and FBR Taxes in 2006/07 GDP (%) FBR Taxes (%) Services (33%) Services (58%) Industry (66%) Industry (23%) Agriculture (20%) Agriculture (1%) 3.8. The narrowness of the tax base is also apparent from the dependence on a few revenue spinners. All four major federal taxes rely heavily on a few commodities (Figure 3.3, right panel). For example, the top six withholding taxes collected 43 percent of all direct taxes in 2007/08. Striking is also the growing concentration of import taxes on a small group of commodities. The top 10 commodities contributed 81 percent of import sales tax collection in the first nine months of 2007/08, compared to only 50 percent in 2002/03; likewise, the share of the top 5 commodities in customs duties increased from 40 percent to 57 percent over this period. Overall, tax collection is heavily dependent on imports. The joint contribution from the sale tax on imports, custom duties, and withholding taxes on imports is around 55 percent of overall taxes. Add to this, as we have seen above, manufacturing, and you get the two principal bases for federal taxes. This is why any reductions in growth of manufacturing or dutiable import can put FBR's annual collection targets at risk. Figure 3.3: Contributions of Subsectors, Commodities and Heads to FBR Taxes Ratios of Tax Share to GDP Share of 24 Sectors in 2006/07 (%) 29.5 7 Contribution of Top Commodities and Heads to FBR Taxes (%) 90 90 6 80 Top 10 Commodities in 80 5 Import Sales Tax Top 10 Commodities in 4 Domestic Sales Tax 70 70 3 2 60 Top 5 Commodities in 60 Customs Duties Tax Share Above GDP Share 1 Tax Share Below GDP Share 50 50 0 Top 6 Withholding Other Manufacturing Automobiles Finance & Insurance Machinery Fertilizer Ownership of Dwelling Construction Telecom Textiles Chemicals Edible Oils Agriculture Sugar Food & Beverages Other Services Petroleum Products Electricity & Gas Distrib. Wholesale & Retail Trade Iron & Steel Public Admin. & Defense Mining & Quarrying Transp., Stor. & Commun. Cigarrettes Cement Heads in Direct Taxes 40 40 30 30 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 33 Long-Term Buoyancy 3.9. Today's narrow tax base is the result of the low buoyancy of Pakistan's tax system over the decades. As the demand for public services tends to increase as per capita income rises, tax systems should be able to deliver over time automatic growth in fiscal revenues relative to the size of the economy. If the proportional change in tax revenues exceeds the proportional change in economic activity, then, everything else equal, the tax-to-GDP ratio would rise with per capita income. Unfortunately, measuring such tax elasticity is far from straightforward, as the observed changes in tax revenues are not just the result of changes in the tax base arising from changes in economic activity. Instead, they could also reflect changes in the structure of taxes, such as tax rates or the definition of the tax base, as well as changes in the tax administration, such as a stricter enforcement of the tax laws. Since it is difficult to disentangle these effects, analysts typically look at tax buoyancy, which measures the ratio of the proportional change in tax revenues to the proportional change in GDP. Figure 3.4 (left panel) looks at the average buoyancy measures over the last five decades, using a World Bank GDP series that adjusts for the periodic rebasing in national value added. The buoyancy measures are derived by regressing the natural logarithm of the tax revenue series on the natural logarithm of the GDP series. Buoyancy above unity indicates tax revenue growth in excess of GDP growth. The overall buoyancy of the FBR taxes declined from above unity in the 1960s and 1970s to below unity in the 1980s and 1990s, and increased above unity in the 2000s, although it remains below the levels of the 1960s and 1970s, and at 1.05 it remains rather low: a one percent increase in GDP implies an increase in the tax-to- GDP ratio of only 0.05 percent (that is 5 percent of 1 percent). There are distinct differences in the buoyancy of the four main taxes. Direct taxes outgrew GDP in the last two decades, and sales taxes in the last three decades, but not previously. Custom duties grew far slower than GDP during the trade liberalization of the 1990s, but not otherwise. Excises were by far the most buoyant revenue source in the 1960s, grew slower than GDP since the 1970s, and were the least buoyant revenue source in the 2000s. 3.10. Low revenue stability compounds low buoyancy. Tax revenues display a high degree of volatility. The coefficient of variation ­ measured as standard deviation divided by the mean ­ equals to 0.33 in the 2000s, just somewhat below the 0.39 in the 1960s (Figure 3.4, right panel). The tax-to-GDP ratio changed on average by 0.4 percentage points from one year to the next since 1959/60, and by 0.3 percentage points since 1999/2000. Such volatility makes it difficult to forecast tax revenue collection accurately and undermines budget planning. Recently, direct taxes, with their dependence on enterprise profitability, and custom duties, with their dependence on import volumes, were the most volatile revenue sources. Figure 3.4: Long-Term Buoyancy and Stability of FBR Taxes Buoyancy of FBR Taxes (%) Coefficient of Variation of FBR Taxes (%) 2.4 0.8 FBR Direct 2.2 Sales 0.7 Excise Customs 2.0 FBR 1.8 0.6 Direct Sales 0.56 1.6 Excise Customs 0.5 1.4 1.33 0.4 0.39 0.38 0.38 1.2 1.09 1.05 0.35 1.0 0.93 0.95 0.3 0.8 0.6 0.2 0.4 0.1 0.2 0.0 0.0 59/60-69/70 69/70-78/79 79/80-89/90 89/90-99/00 99/00-07/08 59/60-68/69 69/70-78/79 79/80-88/89 89/90-98/99 99/00-07/08 34 Short-Term Buoyancy 3.11. In a growing economy, the tax collection machinery has to run fast just to stand still. From 1999/2000 to 2002/03, nominal GDP increased by 8.4 percent annually. Hence, just in order to keep the tax-to- GDP ratio constant, nominal tax collection had to increase by the same rate. In fact, federal tax collection grew by 10 percent annually, so that the tax-to-GDP ratio increased from 9.1 percent in 1999/2000 to 9.6 percent in 2002/03. From 2002/03 to 2007/08, federal tax collection grew faster (17 percent annually), yet nominal GDP expanded also faster (17 percent annually), so the tax-to-GDP ratio remained constant at 9.6 percent. This is surprising. After all, international experience suggests that tax collection is pro-cyclical in developing economies (Box 3.2). 3.12. Breaking down the federal taxes into the main components suggests little structural change since 1999/2000. The most important adjustment is the decline in excise duties due to a reduction in rates and the number of excisable goods (Figure 3.5, left panel). With the steady switch from import substitution to trade integration, Pakistan's simple average tariff rates fell from over 20 in 1999 to 15 percent in 2007. As a result, custom duties remained roughly constant around 1.5 percent of GDP in spite of rising import volumes. At the same time, the performance of the sales tax was disappointing. After increasing from 3.1 percent of GDP in 1999/2000 to 4.1 percent of GDP in 2002/03, it declined subsequently to 3.6 percent of GDP in 2007/08. It took until 2006/07 before the good profitability of the corporate sector translated into higher income tax collection, and only some of the gains could be maintained in 2007/08. Overall, at 9.6 percent of GDP, federal tax collection is the same in 2007/08 as it was in 2002/03, and only some 0.5 percentage points of GDP higher than in 1999/2000. Direct taxes and GST are the most important revenue sources (Figure 3.5, right panel), contributing close to two-fifths of federal collection each. The balance is picked up by excise duties and custom duties. This structure is broadly similar to the average of non-OECD countries (Box 3.3). Figure 3.5: Composition of Federal Tax Collection since 1999/2000 Composition of Federal Taxes (% of GDP) Composition of Federal Tax Revenue (% of Federal Tax Revenue) 0 10 4.5 0 0 Overall Federal 100% 1 1 Non-FBR 12 9 Sales 18 17 17 16 15 4.0 90% 15 19 19 Customs 8 12 3.5 80% 9 8 9 10 12 8 Direct 9 16 Excise 7 70% 3.0 6 60% 2.5 41 36 37 5 39 42 42 41 50% 34 40 Customs 2.0 Sales 4 40% 1.5 3 30% Excises 1.0 Direct 2 20% 39 39 33 35 32 31 31 31 31 1 0.5 Non-FBR 10% 0 0.0 0% 1999/2000 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 1999/2000 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 35 Box 3.2: Why is Fiscal Policy Procyclical in Developing Countries? A standard normative prescription of fiscal policy is that it should be countercyclical. During bad times, the government should increase spending (due to automatic stabilizers such as social safety net spending) and lower taxes (even with constant tax rates due to progressive taxation) to spur the economy out of the downturn; during good times, the government should reign in spending and increase taxes to smoothen economic activity and save for the rainy day. Empirical evidence for the last decades shows that OECD countries, but not developing countries, conform to this pattern, although recent global trends in tax policy, such as the move towards reductions in tax rates in view of a robust world economy and globalization may have weakened fiscal stabilizers also for developed countries. The literature has proposed economic and political economy reasons for the puzzling behavior of fiscal policy in developing countries. First, as credit market restrictions prevent developing countries from borrowing during bad times, they are required to cut back expenditures and raise taxes during a downturn. Second, the fluctuations in the tax base are much larger in developing countries than in developed countries. During good times, pressures for spending increases become hard to resist in view of plentiful tax revenues. Alternatively, voters may want to avoid wasteful spending through lower tax rates, which again results in procyclical fiscal policy. How does this discussion on the benefits of counter-cyclical fiscal policy square with the discussion in Chapter 1 on the need for fiscal consolidation in Pakistan? Chapter 1 focuses on the effects of ongoing, sustained budget deficits. Whatever decisions are made about short-run fiscal policy when the economy is weak, the objective should be budget balance over the business cycle. This requires tax collections to grow faster than GDP when the economy is strong. Sources: Alesina, Campante and Tabellini (2007); Norregaard and Khan (2007), Perotti (2007), and Talvi and Vegh (2005). Box 3.3: Pakistan's Tax Revenues Compared to OECD and non-OECD Countries What stands out in Pakistan's tax structure in international perspective? Drawing on the database of the IMF Fiscal Affairs Department, we compare tax collection in Pakistan with the unweighted average of 39 non-OECD countries and 30 OECD countries across the period of 2002 to 2006 (2002/03 to 2006/07 for Pakistan). Two features stand out. First, looking at the breakdown of tax collection as percent of GDP, with one exception, Pakistan uniformly collects less than both non-OECD and OECD countries for each of the six categories. To a large extent, this is explained by the higher income levels in both OECD and non-OECD countries than in Pakistan. Nevertheless, the gap in the tax-to-GDP ratio is sizable: non-OECD countries collect relative to GDP on average about twice as much in taxes as Pakistan, which suggest significant headroom for raising the tax-to-GDP ratio in Pakistan. Second, comparing the contribution of each tax to overall tax collection, Pakistan's low shares of other income taxes, including individual income tax, and other taxes, including property taxes, are noteworthy. Tax Revenue in OECD and non-OECD Countries (% of GDP), 2002 to 2006 Tax Revenue in OECD and Non-OECD Countries (% of Tax Revenues), 2002 to 2006 10 40 Pakistan Non- 9 OECD 35 OECD 8 Non- OECD 30 OECD 7 25 6 Pakistan 5 20 4 15 3 10 2 5 1 0 0 Corporate Other Income General Sales Excises Custom Duties Other Taxes Corporate Other Income General Sales Excises Custom Duties Other Taxes Income Tax Tax Tax/VAT Income Tax Tax Tax/VAT 36 Federal Tax Expenditures 3.13. Tax exemptions are one of the principal reasons for a narrow tax base with low buoyancy. They come in different shapes and forms, including concessions, special regimes, and incentives, but they all amount to an exemption from certain taxes that the tax authorities would otherwise collect from an individual or an organization. While they are part of government fiscal activity, they may go unnoticed because the revenue foregone does not explicitly show as spending. If used properly, tax expenditures can play an important role in helping implement policy priorities. However, the assessment of tax exemptions is often complicated since reporting and accounting practices fall far short of what is used for government expenditures. This makes it difficult, if not impossible, to evaluate the cost, efficiency, and equity impact of tax exemptions. This lack of scrutiny may invite fiscal opportunism through exemptions for special interest groups. The political economy of granting tax exemption is simple but powerful. While the benefits of exemptions are highly concentrated among the members of the interest group, the costs are dispersed among all taxpayers. 3.14. Many countries report regularly on one type of tax exemptions: tax expenditures. They are exemptions or concessions that fall outside a tax norm or benchmark. The tax norm includes the rate structure, accounting conventions, deductibility of compulsory payments, and provisions to facilitate tax administration. Identifying certain provisions in the tax law as tax expenditures is less straightforward as it may seem, since distinguishing integral tax rules and deviations from such rules involves invariably many judgment calls. In addition, tax norms are defined differently across countries, making it difficult to make comparisons. 3.15. Official figures suggest that the reduction in Pakistan's tax expenditures has recently been reversed. Pakistan reports regularly on tax expenditures in the annual Economic Surveys (Figure3.6). According to these reports, tax expenditures, calculated as the estimated revenue loss in percent of GDP, declined from around 0.8 percent in 2000/01 to 0.33 percent in 2005/06, and then increased to 2.1 percent in 2006/07 and 0.9 percent in 2007/08. The sharp variation in these estimates over the last three years is surprising. Looking at the components of tax expenditures, the variation is primarily due to different estimates for revenue losses from the exemption of capital gains. This underscores the conceptual difficulties in estimating tax expenditures consistently over time, as well as the need for a continued reduction of tax expenditures. Figure3.6: Pakistan's Tax Expenditures Government Estimates of Revenue Loss from Tax Expenditures (% of GDP) 2.5 2.0 Customs Duties 1.5 Sales Tax 1.0 Income Tax 0.5 0.0 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/8 37 Federal Revenue Targets 3.16. Over-estimated revenue forecasts have contributed to reducing the pressure on eliminating tax exemptions. Unrealistic forecasts contribute to poor budget management, such as expenditure arrears and a stop- and-go expenditure policy, and make meaningful medium-term budget planning difficult. Inaccurate budget plans also weaken transparency and diminish public accountability of fiscal operations. Yet, more often than not, developing countries tend to overestimate their tax revenues. This might be because of political pressures to boost public expenditures without adjusting the overall fiscal balance, ambitious revenue forecasting as incentive device for a hard-to-control tax administration, or unanticipated economic difficulties. Administrative shortcomings play also a role, such as weak planning and meager resources for the implementation of budget measures, or inadequate information analysis of the impact of measures. Indeed, Pakistan's federal budget revenue estimates systematically deviated from actual revenue receipts during the 1990s. Then, from 2002/03 to 2006/07, FBR broke with the past and outperformed its collection targets (Figure 3.7, left panel). In 2007/08, tax collection fell short of the budgeted target once again, although by a small margin only. The over-projection of tax collection runs the risk that inadequate tax policy measures are taken at the time of the budget. Pakistan's federal tax policy measures amounted to at most Rs. 11 billion between 1999/2000 to 2005/06, and were even negative in three years. The 2007/08 and 2008/09 budgets, with the greater focus on raising the tax-to-GDP ratio, incorporated more substantive policy measures, amounting to Rs. 35 billion and Rs. 70 billion, respectively. 3.17. Tax performance should be seen in the light of evolving economic circumstances which sometimes deviate substantially from the initial estimates. A rough ex post examination of tax by tax outcomes since 1999/2000 is revealing (Figure 3.7, right panel).2 Actual FBR collection fell short of what could have been expected under the realized growth factors and budgeted revenue measures in five out of the last nine years. In particular, FBR met the revenue collection targets from 2003/04 to 2005/06, but the better-than-expected economic performance suggests that revenue collection might have been even larger. Likewise, higher-than-expected inflation implied higher-than-expected nominal growth factors in 2007/08, notwithstanding the otherwise difficult economic conditions. 3.18. Excise duties and sales taxes have underperformed in recent years. There are large differences in performance across taxes. Excise collections were consistently lower than what could have been expected, perhaps due to underestimation of the impact of relief measures at the budget stage. Sales taxes have been underperforming since 2003/04, which might point to an erosion of the tax base due to special regimes and zero-rating. Custom duties have fallen short of expectation up to 2001/02 and in 2004/05, and direct taxes display overall little systemic pattern. The good performance of direct taxes in 2006/07 was driven by high banking and energy sector profits. Figure 3.7: Two Indicators of FBR Tax Collection Performance FBR Tax Collection Relative to Original Target Ex-Post Expected FBR Tax Collection (Shortfall as % of Actual Collection) 4 40 40 Excise 2 Actual growth in proxy tax base and budgeted estimates 30 of revenue measures should have resulted in a higher 30 FBR tax collection than actual outcome. 0 20 Customs 20 Percent Surplus (+) or Shortfall (-) -2 10 Direct 10 -4 FBR -6 0 0 -8 -10 -10 -10 -20 -20 -12 Actual growth in proxy tax base and budgeted estimates Sales of revenue measures should have resulted in a lower -30 FBR tax collection than actual outcome. -30 -14 1990/91 1991/92 1992/93 1993/94 1994/95 1995/96 1996/97 1997/98 1998/99 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 1999/2000 2000/01 2001/02 20002/03 20003/04 2004/05 2005/06 2006/07 2007/08 -40 -40 2 The ex post analysis builds on these ingredients. FBR taxes are divided into direct tax, sales tax, custom duties and excise duties. For each of these taxes, the baseline is defined as last fiscal year's collection. The baseline is then multiplied with the actual nominal growth factors. These are as follows. Direct taxes: nominal GDP; import sales tax: nominal import growth; domestic sales tax: large-scale manufacturing growth and inflation; customs: dutiable imports growth and inflation; excise duties: large-scale manufacturing growth of selected items and inflation. Adding net budgetary measures to the baseline multiplied with nominal growth factors gives the ex-post expected revenues, which are then compared to actual revenue. 38 Provincial Tax Effort 3.19. Weak mobilization of provincial taxes also contributes to the overall low tax effort. At a time when federal government taxes are less than 10 percent of GDP, there is a natural question whether the revenue collection of provincial governments is adequate. Two considerations might suggest they are not. First, while the federal government carries out the bulk of the revenue collection, provincial governments deliver a large chunk of the public services. Provincial expenditures represent close to 30 percent of national government expenditures, yet provincial tax revenues contribute just 4 percent of national tax revenues (Figure 3.8, left panel). The share of provincial taxes in provincial revenues is very low. In 2006/07, it was no more than 2 percent in Balochistan, 4 percent for NWFP, 9 percent for Sindh, and 8 percent for Punjab. The gap is filled primarily by large federal transfers to the provinces based on the National Finance Commission Award. The large vertical imbalances between federal and provincial government weaken financial management and fiscal accountability to the public. This divide could increase even further in case of a deletion of the concurrent list which would expand provincial expenditure responsibilities. The provincial expenditure-revenue imbalance means that the tax system violates the benefit principle, one of two fundamental tax principles (the other one being ability to pay). According to this principle, people should be taxed according to the benefit they receive from the government services financed by the tax revenue raised. Second, all provinces carry structural fiscal deficits which they finance, among others, by drawing from balances created by unfilled positions and by slow disbursements of project funds. These deficits are not sustainable in the long run and call for greater own-source revenue efforts. In 2006/07, Sindh collected no more than Rs. 390 in own-source tax revenues from each of its citizens, Punjab Rs. 230, NWFP Rs. 110 and Balochistan Rs. 100 (Figure 3.8, right panel). 3.20. The barriers to increasing revenue mobilization by provincial governments are formidable. They range from structural problems with the present tax system, to administrative shortcomings, to the absence of incentives for provincial governments to increase their tax effort. There are also important constitutional limitations for revenue assignments between levels of government. Revenues are assigned between the federal government and the provinces by specific constitutional provisions. The federal-provincial assignment is in general determined through the federal and concurrent lists. Revenue sources that are mentioned in the federal list belong to the federation only; those in the concurrent list are a shared base for which both the federation and the provinces can develop legal instruments to tax the base; and those neither mentioned in the federal or concurrent lists belong to the province only. While the constitution assigns the federal government the more buoyant and easy-to-reach taxes, provincial taxes are sufficiently broad based to form a substantially more revenue productive tax system than currently the case. The provincial taxes include the property and property transfer taxes, motor vehicle taxes, the sales tax on services, and the agriculture income tax. However, tax exemptions and preferences have narrowed existing tax bases, and many taxes are subject to specific rates. As a result, as discussed in detail in Chapter 9, all of these taxes yield little revenues. Figure3.8: Provincial Tax Collection Provincial Tax Revenues (% of GDP and % of National Tax Revenues) Provincial Per Capita Tax Revenue in 2007/08 0.60 6 500 Percent of GDP Percent of National Sindh 0.55 Tax Revenues 450 0.50 5 400 0.45 350 0.40 4 300 0.35 0.30 3 250 Punjab 0.25 200 0.20 2 150 NWFP 0.15 Balochistan 100 0.10 1 50 0.05 0.00 0 0 1999/2000 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 39 3.3 Efficiency Excess Burden 3.21. Taxation inevitably distorts the allocation of resources in the economy. The excess burden of taxation, one of the most important concepts in public economics, is the economic loss that society suffers as the result of a tax, over and above the revenue it collects. All taxes lead to such inefficiency ­ apart from the hypothetical lump sum tax which equals a fixed amount that cannot be varied by any action of the individual. Taxation introduces distortions as it changes relative prices. This leads producers and consumers to change their behavior in order to reduce the amount of tax they must pay. As a result, society incurs a welfare loss and a reduction in income, since the allocation of resources is inefficient (Box 3.4). Such excess burden might not only affect the scale of production. Other forms of excess burden arise, for example, if the tax system induces a company to invest in buildings when, in the absence of tax preferences, it would have invested in machinery and equipment; or when companies are encouraged by the tax system to finance their capital through debt rather than equity, and so on. Individual taxpayers are also exposed to excess burdens when the tax system alters their savings and labor supply decisions. Box 3.4: The Excess Burden of Taxation In a market economy, consumers and producers react to market prices. A consumer would often be willing to pay even more for a good than the price charged in a store. Suppose, for example, the going price in the market for a CD is Rs. 1000. Suppose also that one consumer is willing to pay Rs. 1500 for the CD, and another consumer is willing to pay Rs. 1200 for the CD. By paying the going market price of Rs. 1000, both individuals benefit, the first by Rs. 500 and the second by Rs. 200, for a total benefit of Rs. 700. This "benefit" is called the consumer surplus. Now, consider what happens if an excise tax of Rs. 300 per CD is imposed, making the going market price now Rs. 1300. The person who was willing to pay Rs. 1500 will still purchase the CD and will pay a tax of Rs. 300, but he will now enjoy a benefit of only Rs. 200. The person who was willing to pay Rs. 1200 will no longer purchase the CD because its cost (Rs. 1300) is now greater than the amount she was willing to pay (Rs. 1200). The tax on the CD has generated revenues of Rs. 300, but the tax has also made both individuals worse off: the total benefit has fallen by Rs. 500, from Rs. 700 to only Rs. 200. Because the decline in the total benefit of Rs. 500 is greater than the tax revenues of Rs. 300, there is an "excess burden" that equals the difference, or Rs. 200. This outcome is inefficient because waiving the tax on the consumer willing to pay Rs. 1200 would leave the government or CD seller no worse off, but the consumer better off. Most all taxes create this excess burden because most taxes cause agents to change their behavior. If we add up the losses to everyone in society, we arrive at an estimate of the excess burden, or the efficiency cost, of taxes. 3.22. Pakistan's tax system imposes a particularly high excess burden. While taxes lead to an excess burden, they are also inevitable. Without them, there would be no funds to finance public investment projects or to pay for transfer payments. But, as the burden imposed on taxpayers exceeds the tax they pay, it is important to ensure that the benefits arising from public spending outweigh the efficiency losses. More importantly, for the same amount of revenue collected, some taxation systems are more distortionary than others. Pakistan's tax system provides economic agents with an uneven playing field. The special treatment of particular sectors and sources of incomes through differences in tax bases and tax rates is quite common. This distorts the allocation of resources as too many resources will be allocated in the lightly taxed sectors and too few in the more heavily taxed sectors. The revenues lost due to these special treatments need to be made up with higher taxes in other activities or sectors in the economy. This is more distorting than a more even treatment of activities and sectors for the following reason: the excess burden increases disproportionately with the tax rate; indeed, excess burden increases approximately proportional to the square of the tax rate. In other words, doubling the tax rate leads to a quadrupling of the excess burden of taxation, tripling the tax rate results in a nine fold increase in excess burden, and so on. Hence, policies which erode the tax base and lead to increases the tax rate need special justification. Likewise, tax-financed expenditure and redistribution policies must provide benefits that outweigh the efficiency costs of raising such revenue. 3.23. The system of corporate taxes in Pakistan introduces a wide range of distortions in firm behavior. As we will discuss in detail in Chapter 5, these distortions are illustrated most clearly by the large variation in average and marginal effective tax rates by sector and by asset type. Together with the extensive system of tax incentives and exemptions, the preferential tax rate for small businesses, and the use of the presumptive tax regime for some companies, the corporate tax gives preferential treatment to different types of investment and to different sectors, thereby leading firms to base their investment decisions mainly on tax considerations rather than on market forces. 40 Investment Climate 3.24. Paying taxes is among the worst dimensions of Pakistan's investment climate. While no business likes to pay taxes, businesses also have a hard time to be productive without schools, hospitals, courts, roads, airports or other public infrastructure and services that are paid for by taxes. The 2008 Doing Business survey (Box 3.5) allows us to look at ten areas of formal business regulations and processes across 178 countries. Business regulations affect directly the productivity of investment and economic activity of micro to large enterprises. They cover a range of issues, from the costs to starting a business to the costs of closing a business. How do business regulations in Pakistan, as laid out on paper, compare with those in 177 other countries? Two broad findings stand out (Figure 3.9). First, overall, Pakistan's ranking at 76 is respectable. For example, with the exception of Maldives, all other South Asian countries do worse. Second, Pakistan does remarkably poorly on paying taxes. Ranked 146th, Pakistan is in the bottom quintile of the countries. Out of the ten topics, this is Pakistan's 2nd worst ranking, only the category "enforcing contracts" puts Pakistan even lower. 3.25. Given Pakistan's low ranking on paying taxes, it is worth unpacking the indicator. It has two main aspects ­ the compliance costs, as measured by the number of business taxes to pay and the time spent in paying taxes; and the financial costs, as measured by the total tax rate. For both dimensions, the focus should go beyond corporate income taxes. All taxes borne and collected by businesses should be recognized along with the related tax compliance costs. In particular, the total tax rate includes all taxes that are paid by companies, from corporate income taxes, property taxes and labor taxes to all applicable deductions and exemptions. Sales and consumption taxes are included in terms of compliance costs, but not as part of the analysis of the total tax rates as they are not considered to be borne by the business. 3.26. The large compliance costs with the tax system are responsible for Pakistan's poor rating on paying taxes. Pakistan's tax system subjects the stylized modest-sized business TaxpayerCo to 47 payments across 10 different taxes per year, and to some 560 hours per year to ensure tax compliance (Figure 3.10). The General Sales Tax imposes by far the largest compliance costs ­ 480 hours are spent in dealing with GST regulations alone. By contrast, the same company would be subject to only 35 payments and would spend only 264 hours complying with tax regulations in Thailand. While tax compliance is cumbersome, Pakistan's total tax rate is more in line with international benchmarks. The total tax rate is a standardized indicator that is used to gauge the overall burden of corporate taxes after adjusting for necessary exemptions and the like. Pakistan's total tax rate of 40.7 percent is significantly lower than that of India, Egypt and Turkey. This is remarkable as Pakistan's corporate rate is higher than these comparator countries and underlines the importance of considering all taxes paid by businesses. However, it is worth highlighting that the total tax rate is based on a stylized business, and therefore does not capture differences in the structure of the enterprise population. International comparisons based on actual tax return data are provided in Chapter 5. 3.27. While Pakistan's ranking on paying taxes is disappointing, the link between taxation and economic performance is complex. Taking growth as one measure of the investment climate, the extensive literature on the linkages between taxation and economic growth is still somewhat inconclusive. At the macroeconomic level, the connection between policy variables and economic growth tends to be unstable across all time periods, estimation methods, or specifications. There is some suggestion that irrespective of the overall level of taxation, the composition does matter, but even here could be a trade-off: while greater reliance on indirect taxes might reduce growth, greater reliance on direct taxes might reduce inequality (Alm and Rogers 2008). One reason for the lack of robust results is that these relationships may also vary with the level of economic development of the countries, and, particularly, with the quality of governance. Another reason is that the impact may also depend on how the additional taxes are spent. For example, long-term growth might benefit from investments in infrastructure but not from outlays for transfer payments. 3.28. Whereas the macroeconomic impacts of taxation might be disputed, there is greater consensus on the microeconomic impacts. In particular, simple, broad-based tax systems tend to be better at encouraging employment, saving and investment. In addition, imposing cumbersome procedures or high tax rates on forms of internationally mobile capital, whether human, physical, or financial, is likely to be damaging to the growth 41 prospects to an economy. This suggests that Pakistan should be cautious in manipulating tax policies to increase economic growth. Box 3.5: Doing Business 2008 The World Bank and PricewaterhouseCoopers have recently conducted a survey on the ease or difficulty in conducting business in 178 countries around the world. Doing Business 2008 is based on a "case study company" approach, in which a standardized, common company is constructed, and then the specific institutional features of each country are applied to this identical company in order to determine how easy (or difficult) it is to conduct business in the country. There are various dimensions along which doing business is examined: starting a business, dealing with licenses, employing workers, registering property, getting credit, protecting investors, trading across borders, enforcing contracts, closing a business, and paying taxes. Specifically, the company TaxpayerCo is assumed to: · be a limited liability, taxable company in its second year of operation · be domestically owned, with five resident owners · be engaged in general industrial and commercial activities (e.g., producing and selling ceramic flower pots) · have 60 resident employees (4 managers, 8 assistants, and 48 workers) · have a turnover of 1050 times income per capita · distribute 50 percent of its profits as dividends. Figure 3.9: Pakistan's Doing Business 2008 Rankings Pakistan's Rankings in Doing Business 2008 out of 178 Countries Rankings in Doing Business and Paying Taxes 2008 - Pakistan and Comparitors 170 180 160 154 170 165 Ease of Paying Taxes 150 146 158 160 Ease of Doing Business 150 140 150 146 132 130 140 130 126 120 120 120 110 107 111 110 106 100 101 93 94 100 88 92 90 89 Pakistan's Average Ranking = 76 90 81 80 80 76 70 68 70 61 60 59 60 56 57 54 60 51 50 50 40 35 40 30 30 24 30 15 20 19 20 10 1 10 0 Thailand Bangladesh Turkey India Sri Lanka South Korea South Africa Maldives Nepal Pakistan Malaysia Egypt 0 Protecting Closing a Starting a Getting Registering Dealing Trading Employing Paying Enforcing investors business business credit property with across workers taxes contracts licenses borders Figure 3.10: Pakistan's Doing Business 2008 Ranking for Paying Taxes Rankings in Paying Taxes 2008 by Component - Pakistan and Comparitors 180 170 Time to Comply Total Tax Rate 160 156 Tax Payments in Year 150 138 140 130 120 110 100 90 80 79 70 60 50 40 30 20 10 0 Korea Maldives Nepal Turkey India Thailand Pakistan Malaysia South Africa Egypt Bangladesh Sri Lanka 42 3.4 Equity Tax Incidence 3.29. Inequalities in incomes have long been a stark fact of life in many developing countries, including Pakistan. Wide inequalities in living standards are a concern because severe deprivations violate a society's sense of fairness. In addition, since they often arise from unequal opportunities to gain skills, jobs, or market access, they have a tendency to persist. And because inequity leads to an inefficient use of resources and to less effective institutions, it hinders long-term development. It follows that there is a legitimate role for public action in the pursuit of equity, notwithstanding the primacy of individual freedoms and of the role of markets in allocating resources (WDR 2006). 3.30. Equity matters for taxation. After all, "ability-to-pay" is one fundamental tax principle. With the top household decile earning about ten times the income of the bottom household decile, and inequality rising noticeably from the early to the mid-2000s (Box 3.6), there are concerns about the distributional impact of Pakistan's tax system. Even though there is no consensus regarding the "right" amount of equity in a tax system, most people seem to believe that fairness dictates that at the very least, low income individuals should not pay more of their income in tax than high income individuals. This notion is closely tied to the concept of progressivity. A "progressive" tax is a tax where the average burden faced by the household or individual increases with income. Thus, a high income taxpayer would not only have a larger tax liability but also pay a larger share of her income in tax. This results in a redistribution of the income, as, relative to the pre-tax income shares, the share in total post- tax income increases for low income taxpayers and decreases for high income taxpayers. 3.31. Without a thorough empirical analysis, it is difficult to sort out the net effect of different factors on the tax incidence. Tax incidence is the study of who bears the burden of taxation. Who pays Pakistan's taxes? Do they fall inordinately on low-income families, or are they protected as they are primarily working in the hard-to-tax informal sector and agriculture? Or is the tax burden borne disproportionately by the higher income classes, who also own most of the capital in the country, or do they benefit from legal exemptions and evade taxes? A priori, one might expect that the tax system in Pakistan is regressive. After all, the tax structure is weighted heavily toward consumption taxes, there is no tax on capital gains, and the individual income tax weighs lightly in revenue- collection mix. But the level of tax collection is low in Pakistan by international standards, and the tax administration focuses on the "easy-to-tax." This includes large companies, visibly higher income families, international trade, payrolls and consumption taxes in general. 3.32. Taxation is only one of the fiscal instruments that can make the income distribution more equitable. The most important tool is government expenditures, such as health programs, education programs, and direct subsidies for housing and food. In general, public programs might be the better way to reach equity goals than taxation. While efficiency considerations suggest tax systems should ensure a level playing field across sectors and activities, targeting particular beneficiaries is more natural when designing public programs. Furthermore, the effectiveness of public expenditure programs, although not free of problems of fraud and waste, is not undermined by issues of evasion and avoidance that are widespread in tax systems. Many potentially taxable transactions take place in the informal sector, and imposing heavy taxes on the formal sector can push these activities outside the reach of the tax administration, either to the informal sector or outside the country. Finally, government spending in Pakistan is almost double the amount of tax collection, and has a greater potential to impact the income distribution. 3.33. While public programs might be the better tool to achieve equity objectives, it is important to remember that they often rely on funding from taxation. Hence, the incidence of taxes should ideally be considered jointly with the incidence of public spending. For example, if a broad-based tax is used to fund a public spending program that benefits mostly poor people, the overall impact of these policies could well make fiscal policy more equitable. An analysis of the combined effect of spending and taxation on equity is beyond the scope of this report. Nevertheless, assessing the equity of the tax system remains essential in its own right. It is one part of the comprehensive assessment of spending and taxation. More importantly, from a dynamic perspective, it also provides the benchmark for evaluating the impact of changes in tax policy on equity while keeping public spending unchanged. This scenario may well apply to Pakistan. Given the dire fiscal situation, any additional tax effort should be used to reduce the fiscal deficit without expanding public spending. 43 Box 3.6: Inequality in Pakistan Equity has much in common with poverty reduction and development. Access to schools, health centers, roads, credits and markets provides pathways out of deprivation, and as poor people make use of such opportunities, the growth prospects of the country improve. Given the linkages between equity, growth and poverty reduction, it is worthwhile asking how Pakistan compares with other countries in terms of equity. Unfortunately, there are numerous conceptual, methodological, and definitional issues when comparing equity across countries. For starters, from an equity perspective, the distribution of opportunities matters more than the distribution of outcomes. But opportunities, which are potentials rather than actuals, are harder to observe and measure than outcomes. Even the measurement of outcomes is fraught with difficulties. For example, the definition of the welfare indicator differs from one country to the next, and national surveys in different countries might vary widely in their ability to capture adequately the top- end of the welfare distribution. In view of these difficulties, meticulous attention to concepts, definitions and the details of survey methodology is required for cross-country comparisons. Perhaps the best available source for such analysis is the database generated for the 2006 World Development Report on Equity and Development. It was created using primary data from nationally representative surveys with sufficiently comprehensive definitions of income or consumption, the most commonly used summary measures of welfare. This data suggests that Pakistan is among the most equal countries in the world. Among the 126 countries with survey-based income or consumption data, Pakistan has the 11th lowest level of inequality -- as measured by the Gini coefficient (ranging from zero to unity, with zero indicating full equality), the most commonly used index of inequality -- on par with Denmark, Norway and Uzbekistan. Similarly, among the 80 countries with consumption as welfare measure, Pakistan ranks 8th. The same holds in regional comparisons. The Gini coefficient of Pakistan (0.27) is lower than the ones of Bangladesh (0.31), India (0.33), Nepal (0.36) and Sri Lanka (0.38). Clearly, Pakistan's government has every right to be proud of this achievement. Nevertheless, it is important to put these comparisons into context. First, Pakistan's inequality index has remained roughly unchanged from the early 1990s to the early 2000s, a period of economic stagnation with little progress in poverty reduction. By contrast, it increased by about 10 percent from 0.27 in 2001/02 to 0.30 in 2005/06, a period of economic expansion and noticeable progress in poverty reduction. In other words, the slow pace of changes in economic structure and productivity in the past might have favored a fairly benign level of consumption inequality. Second, consumption is only one dimension of well-being, and low inequality of consumption is far from synonymous with high equity of opportunities. Indeed, there is evidence that Pakistan might do rather poorly in international comparisons along the dimensions of education and health. The WDR 2006 suggests that Pakistan ranks only 98th out of 111 countries in terms of inequality in the years of schooling, and that the gap in immunization rates between the bottom and the top wealth quintile in Pakistan is among the largest in the world. Table 3.2: Summary of Tax Incidence Assumptions Tax Labor Capital Consumption Individual income 100% for tax on 100% for taxes on dividends, salaried individuals rents; distributed by allocation to distributed by income from capital, income allocation to wages from property from the PSLM in PSLM Corporate Income 50% distributed by 50% distributed by allocation to Alternative: 50%; distributed by allocation among 34 allocation to wages capital income from the PSLM consumption groups. (formal and informal) GST 100%: Distributed by 34 categories of household consumption Customs 100%; Distributed by 34 categories of household consumption, merged with customs duty classification Excise 100%; Distributed by consumption based on disaggregated excise tax receipts. Taxes on capital 100% on capital allocated by and property PSLM defined capital income; (dividends, some alternative assumption: 50 and interest income, 100% of the tax on property to renters for taxes on rental rental income) income, allocated based on rental expenses from PSLM 44 Modeling Tax Incidence 3.34. Since taxes cause individuals and firms to change their behaviors, answering the question of "who bears the final burden of a tax?" is less straightforward than it may seem. Fortunately, there is wide agreement among economists on basic principles that guide the modeling of tax incidence. First, the economic incidence (the change in real income resulting from a tax) rather than the statutory incidence (the legal requirement to remit a tax) matters (Box 3.7). This also implies that individuals ­ shareholders, workers, landlords, interest income recipients, and consumers ­ rather than corporations (which bear the statutory incidence for corporate income tax) bear the tax burden. Second, the incidence of a tax depends on the possibility of shifting the tax burden through changes in behavior of individuals. The burden of taxation is channeled mainly through changes in prices, and the incidence depends upon the responsiveness of consumers and producers to changes in these prices. When a tax is imposed, individuals will adjust their behavior to reduce their tax liabilities ("tax shifting"). Those who are better able to adjust their behavior are better able to shift the burden to others and will bear less of the burden of the tax. Enterprises will typically attempt to shift increases in indirect taxes to consumers by increasing prices by the same amount but might not always be able to do this without losing customers. If consumers respond little to a change in gasoline prices, then consumers will bear more of the incidence of a tax on gasoline. Similarly, if workers are able to reduce their work effort or to shift their labor to untaxed sectors in response to a individual income tax, then workers will bear less of the burden of that tax. Box 3.7: Statutory and Economic Incidence Assume that the price of a bottle of a soft drink is Rs. 10 and the government imposes a tax of Rs. 1 per bottle. Every time a bottle is purchased Rs. 1 goes to the tax collector. Suppose that after the imposition of the tax, the tax induces a rise in the price of soft drinks to Rs. 11 per bottle. Clearly, the seller receives the same amount per bottle as he did before the tax and, in this case, consumers pay the entire tax in the form of high prices. Suppose now that after-tax price of the soft drink rises to Rs. 10.50. Since the seller has to turn over Rs. 1 to government, the seller keeps a net of Rs 9.50 - he is worse off by 50 paisa per bottle. Consumers are also worse off, because they have to pay 50 paisa more per bottle than before the imposition of the tax. In this case the producer and consumer share the tax burden. For both cases, the statutory incidence is on the seller. But the situation differs with response to who really bears the burden, because the price may change in response to the tax. The economic incidence of a tax is the change in the distribution of real income induced by a tax. The process by which the statutory incidence of a tax is moved from those legally responsible to those who bear the economic burden is referred to a "tax shifting". If a tax is shifted to consumers via higher product prices, then the tax is said to be "shifted forward"; if the tax is borne instead by workers or other input suppliers, then the tax is said to be "shifted backward". 3.35. Building on these principles, our approach uses conventional methods to reflect particularities of Pakistan's tax structure. First, in line with much of the household welfare literature within and outside of Pakistan, the household is taken as unit of analysis, as decisions regarding work, consumption, and savings are often made within the context of family finances. Second, current consumption is the measure of income. It is a proxy for permanent income since expenditures tend to be closely aligned with average income over longer periods. It also tends to be more accurately measured in household surveys than current income. Income needs to be "grossed up" for taxes to ensure that effective tax rates are expressed in terms of gross income. Third, as we will discuss later in this chapter, the compliance with tax laws is poor, and the effective tax rate can deviate widely from the statutory tax rate through the cascading effects of input taxes, zero-rating, exemptions and evasion (this is explained in Chapter 4). To account for these two issues, we adopt the attribution approach to calculate taxes paid by a household. We take the household record data and distribute actual tax receipts to the appropriate income and consumption items. This approach cannot estimate as easily the effect of changes in rules and regulations of the tax system as the simulation model which calibrates the potential tax liability for each household based on statutory tax rules and regulation. However, the tax policy changes of interest for this report can still be captured due to an appropriate disaggregation of income and consumption. Fourth, we take account of different compliance among income groups for the individual income tax based on a sample of actual FBR tax returns. Finally, the analysis makes standard assumptions for tax shifting, and tests the sensitivity of findings to variations in these assumptions (Table 3.2). 3.36. The incidence analysis covers all federal and provincial taxes at the level of the receipts as of 2006/07. The principal data source is the 2004/05 Pakistan Living Standard Measurement Survey (PSLM). Additional data includes the 2005/06 Labor Force Survey and FBR tax return samples from 2003/04 to 2005/06. We "age" nominal income and consumption variables from 2004/05 to 2006/07 levels based on the 2005/06 PSLM and nominal GDP growth from 2005/06 to 2006/07. 45 Vertical Equity 3.37. Pakistan's overall tax burden is somewhat progressive. The distribution of the total taxes paid across deciles shows where taxes come from. Low income households pay a small percentage of total tax receipts. The lowest decile pays 2 percent of all taxes, while the highest decile pays 41 percent of all taxes (Figure 3.11, left panel). This compares to a share in total income (scaled up to align with GDP) of less than 3 percent for the lowest deciles, and of more than 32 percent for the highest decile. Indeed, as we move from low to high income deciles, the share of taxes tends to increase more than their share of income. Hence, the effective tax rates tend to increase from low to high income deciles. The percent of gross income paid in tax increases for higher income deciles. The tax system is progressive, although the effective tax rates increase noticeably only at the highest decile. For the interpretation of Figure 3.11, and similar figures in the report, it is important to remember that the effective tax rates are derived by dividing tax payments by gross income, where gross income is scaled up to equal GDP (as opposed to the usual practice of using household income or consumption within deciles). Using GDP. This ensures that the average effective tax rates equals to the national tax to GDP ratio in 2006/07. 3.38. The result remains broadly unchanged when households are sorted by per capita consumption rather than household consumption. In some families, household consumption might be larger just because there are a large number of household members. Hence, dividing household consumption by the number of household member might be a better way of ranking households with respect to their income levels. Sorting households by a per capita consumption moves larger families at a given expenditure level into lower deciles. Since larger families consume more on average than smaller families, and higher consumption is linked to higher taxes, this increases the overall share of taxes paid in the lower deciles (Figure 3.11, right panel). Nevertheless, the effective tax rates are quite similar to before, indicating that consumption and taxes across deciles have adjusted in roughly equal proportions. Additional sensitivity analyses with regard to the distribution of capital income, home-produced consumption, and withholding taxes suggest that it would take large changes in the assumptions for economic incidence to change the level of progressivity of a tax system significantly. 3.39. The social desirability of the observed tax incidence is a normative question. Does the modest degree of progressivity reflect a reasonable distribution of tax burdens? Since it involves value judgments, the people of Pakistan ultimately have to answer this question. Making international comparisons is also not straightforward, as country studies often analyze different taxes, make different assumptions about the incidence of taxes, or use different measures of income. Nevertheless, a number of country studies report overall incidence results that suggest that many systems are slightly progressive. From this perspective, the distribution of taxes in Pakistan is not out of line with the international experience. For tax policy reform, it is useful to take the progressivity of the tax system as a reference point. Furthermore, since the pursuit of equity through taxation implies typically efficiency costs, the desirability of changes in the tax system require an evaluation with regard to both equity and efficiency. Figure 3.11: Vertical Tax Incidence in 2006/07 Tax Incidence for Household Consumption Deciles (Percent) Tax Incidence for Per Capita Consumption Deciles (Percent) 45 14 45 14 Effective Tax Rate Effective Tax Rate (% of GDP) (% of GDP) 40 40 13 13 35 35 30 12 30 12 25 25 11 11 20 20 15 10 15 10 10 10 Share in Total Taxes Share in Total Taxes (%) 9 (%) 9 5 5 0 8 0 8 Bottom 2 3 4 5 6 7 8 9 Highest Bottom 2 3 4 5 6 7 8 9 Highest 46 Box 3.8: Tax Incidence Studies on Pakistan There is a significant literature on the distribution of tax burden in Pakistan, as summarized in the table below. These studies can be classified according to various characteristics, including micro versus sectoral level, the coverage of taxes, simulated versus distributed revenues, adjustment for compliance, and incidence assumptions. The results presented in this report are roughly in line with previous micro- level studies, but three differences are worth highlighting. First, most previous studies have focused on central taxes and largely ignored provincial-level revenues. In addition, the focus of at least two important studies is on consumption taxes only. This report expands this analysis to include income taxes (corporate, individual, and provincial taxes) which are a very important component of Pakistan's tax system. The inclusion of the income taxes leads to a distribution of a more progressive burden than found in previous studies. Second, this burden study adjusts for non-compliance where possible. A few existing studies attempt to control for tax evasion of consumption taxes, but the previous literature has not attempted to control for non-compliance in the case of other taxes. Adjusting for non-compliance with income tax lowers the progressivity of the income tax somewhat. Third, part from to these differences in the vertical equity analysis, the report also provides analysis of horizontal inequities in the system, which may be the cause of revenue loss through tax expenditures and may also magnify taxpayers' indifference toward the tax system. Key Results of Literature Study Author Method, Tax Coverage, Results Other information Incidence of Taxes in Wahid and Micro-level, central plus The overall tax burden is progressive. Most Pakistan (2008) Wallace subnational revenue, distributed progressivity comes from the individual actual revenue, compliance income tax and taxes on corporate income. adjusted, total expenditures The distribution of indirect taxes is roughly used as a proxy for total proportional to total expenditures (income). income, detailed results by tax Provincial taxes are slightly progressive. type The Distributional Ahmad and Micro-level, uses input-output Finds that reform to a broad based VAT Consequences of a Tax Ludlow model to estimate shifting with (replacing other central taxes) provides the Reform on a VAT for focus on VAT reforms (as largest welfare increases to low income Pakistan (1989) replacement for some or all households other taxes), estimates changes in effective tax rates by population deciles Tax Incidence Analysis of Shah and Micro-level, trade taxes, Alternative incidence assumptions are used to Developing Countries: An Whalley individual income taxes and highlight the sensitivity of results; trade taxes Alternative View (1991) corporate income taxes (central are regressive under traditional assumptions level) but progressive if the incidence is on importers; the individual income tax may be shifted to rural labor, making it regressive; the corporate income tax is somewhat progressive under alternative assumptions. Social Incidence of the Refaqat Micro-level, VAT, simulates VAT (GST) is slightly progressive when General Sales Tax in taxes paid households are ranked by consumption; When Pakistan (2003) households are ranked according to their reported current income annual incidence is regressive although lifetime incidence remains progressive Tariff Reduction and Siddiqui Use a CGE model to analyze Reductions of tariffs on industrial imports Functional Income and Iqbal impact of tariff reductions, use would reduce regressivity (improve income Distributio in Pakistan: A of representative households distribution) CGE Model (2001) Distribution of Agricultural B.A. Azhar Sectoral and by farm size, The distribution of agricultural taxes is Tax Burden within the central and subnational taxes on regressive Sector (1992) agriculture, value added in sector is the base Pakistan: A Preliminary Martinez- Central taxes, uses household Overall burden is progressive, largely due to Assessment of the Federal Vazquez data by quintiles the individual income tax Tax System (2006) 47 Horizontal Equity 3.40. Horizontal equity refers to the tax treatment of similarly situated individuals. It contrasts with vertical equity, which is about differences in people's ability to pay are appropriately reflected in taxation. One might expect that families that earn about the same amount of income would pay a similar amount of tax to government. However, this is not always the case. Consider a simple example of two similar families. A tax system may give tax breaks to capital income (exempting dividend income or applying a lower tax rate to interest income). If the two families each have Rs. 200,000 income but family "A" has all wage income and family "B" has 50 percent of their income in wages and 50 percent in capital income, then family "B" will pay less tax than family "A". This is referred to as horizontal inequity. 3.41. Horizontal inequities reduce tax compliance and the efficiency of the tax system. There are numerous areas in the tax system where "like" households are treated "unlike" for tax purposes. This issue can lead to myriad problems. Taxpayers do not like to feel that they have been taken advantage of. If their neighbors pay less tax, they may strive to pay less tax as well. It also contributes to distortionary legal behavior. If capital is afforded preferential treatment compared to labor income, then people will try to shift their income to capital holdings. Exemptions and tax evasion are the two most important sources of horizontal inequities. Pakistan's tax system includes a long list of exemptions, such as exemption of capital gains from the sale of securities, different treatment of some forms of consumption, such as is the case for the zero rating of some goods under the GST, and different treatment of industrial and agriculture sectors. Tax evasion, the topic of the next section, is reflected in the large size of the informal sector and underreporting of registered taxpayers. 3.42. The evidence suggests that Pakistan's tax system is marred by a large substantial degree of horizontal inequity. While it is difficult to develop an estimate of horizontal equity of the entire tax system, we can explore evidence of these inequities by examining the distribution of various items within a household decile. For example, there are a number of goods that are exempt from GST. If a household consumes relatively more untaxed goods, that household will pay less in GST than a similar household that consumes more taxable goods. The vertical bars in Figure 3.12 represent minimum (equal to zero), average and maximum values of the variables for each decile. The consumption shares of low-taxed goods are highly uneven across deciles. According to the 2004/05 PSLM, the average share of consumption of low-taxed goods is about 25 percent of total consumption. Some households report virtually zero consumption of these goods, while others report 90 percent or more. The same holds, to a lesser extent, in case of services which are largely untaxed. The consumption shares of services are low but there are households that purchase a multiple of the services than other households within the same decile. Capital is another item that receives preferential treatment from taxation in terms of capital gains, shares in initial public offerings, and a general difficulty inherent in taxing capital. Families with little capital income, all else equal, are not able to take advantage of these tax preferences and will therefore have a higher tax burden. There is quite a wide distribution associated with the holding of capital throughout the income distribution, pointing again to large horizontal inequity in the tax system. Figure 3.12: Horizontal Inequities with Household Consumption Deciles Share of Low-Taxed Items in Household Consumption by Decile (Percent) 100 90 80 Maximum 70 60 50 40 30 Mean 20 10 Minimum 0 Bottom 2 3 4 5 6 7 8 9 Highest 48 3.5 Compliance Tax Gap 3.43. The tax gap provides a useful measure of the extent of tax evasion in a country. People and firms fail to comply with tax laws for any number of reasons, such as illness, distractions, ignorance, sloth, and greed. There is a difference between tax avoidance and tax evasion. Tax avoidance refers to the use of the tax law to minimize tax liabilities and is perfectly legal. Tax evasion, or non-compliance with the tax law, refers to the non-payment of lawful tax liabilities and is illegal. A tax gap does not arise from tax avoidance, it arises from non-compliance, whatever the underlying motivation. Given this operational definition, it should be clear that in the unlikely event that everyone fully complies with a country's tax system, actual revenue would be exactly equal to the tax system's potential revenue, and there would be no tax gap. As tax evasion increases, the country's tax gap increases. Thus, the size of a country's tax gap is directly related to the extent of tax evasion in the country. 3.44. A large tax gap suggests that a tax system is likely to underperform in terms of revenues and efficiency. Lower revenue is the most immediate result of tax evasion, but it also leads to lower efficiency. Since tax evasion requires government to use a higher tax rate to raise a given level of revenue, tax evasion also increases the excess burden since, as discussed previously in this chapter, this burden is approximately proportional to the square of the tax rate. Tax evasion also provides an incentive to engage in those activities for which it is relatively easy to evade taxes. For example, because the income from self-employment can be done on a cash basis and is therefore harder to detect, this occupation is more attractive than otherwise. The opportunity for noncompliance can distort resource allocation in a variety of other ways, such as causing companies that otherwise would not find it attractive to set up a financial subsidiary, or set up operations in a tax haven, to facilitate evasion. Another source of excess burden are the resources taxpayers expend to implement and camouflage noncompliance, and the resources the tax authority expends to address this. 3.45. Tax evasion harms also equity. According to the principle of horizontal equity, a tax system should impose similar tax burdens on similarly situated individuals. When some taxpayers comply with a tax system and other similarly situated individuals do not, there are adverse consequences for the horizontal equity of a tax system. Tax evasion also compromises the vertical equity of a tax system. In developed countries, high income individuals are often less tax compliant than low income individuals because they can deploy greater resources to shield their income from the tax authorities. At the same time, Pakistan's large informal sector allows also low income earners and many middle income professionals to evade tax and escape detection. This makes the overall impact of tax evasion on vertical equity ambiguous. Variations in compliance rates by income class can to some extent be offset by adjustments in the rate schedule, but it is practically impossible to offset variations within an income class, so that evasion creates horizontal inequity because equally well-off people end up with different tax burdens. 3.46. A large tax gap also suggests weak tax administration. The best tax administration is not simply one that collects the most revenue. How the revenue is raised may be equally important. A poor-quality tax administration may collect large amounts from the easy-to-reach sectors and little amounts from the hard-to-reach sectors. The level of collection is therefore a somewhat inadequate measure of the effectiveness of tax administration. A more accurate measure is the size of the tax gap, and how this tax gap varies among different groups of tax payers. 49 Sizing the Tax Gap 3.47. How to estimate the amount of tax that goes uncollected due to tax evasion? As discussed, the tax gap is the difference between potential and actual tax revenues, where potential revenues are the amount of tax that the government should expect to collect if everyone fully complies with the tax law. It is a simple matter to get actual tax collections by type of tax, so the trick is to obtain a reasonably accurate measure of potential tax revenues. Our basic strategy is to use micro-simulation models to estimate the potential revenues from the major federal taxes of which there are only a hand full. Such modeling requires micro-economic data with information about the relevant tax bases, particularly sources of household income and expenditures, and a tax calculator to simulate the arithmetic calculations of computing tax liabilities by type of tax. The payoff for going through this time-consuming exercise with large data sets is detailed information on the rate of compliance by type of tax. Such knowledge is helpful in targeting the scarce resources of tax enforcement and tax reform. 3.48. The projected Pakistan's federal tax gap in 2007/08 is estimated to be approximately 79 percent of actual tax receipts. This sum amounts to over Rs. 796 billion, or some Rs. 4,800 worth of cheating by every man, woman and child in Pakistan (Figure 3.13). This is likely to be a rather conservative estimate of Pakistan's `true' federal tax gap, as subsequent chapters, which provide tax-by-tax explanations of how these estimates were derived, will make clear. The number does not include any tax gap estimates for the workers welfare fund, workers profit participation fund, the capital value tax, or the wealth tax, which represent less than 2 percent of total federal tax receipts. Data limitations also rule out the analysis to provincial taxes, which account for about 4 percent of Pakistan's total tax collection. It is difficult to compare Pakistan's tax gap estimate with other countries, as there are not many reliable tax gap studies that use similar methodologies. Figure 3.13: Pakistan's Federal Tax Gap (Percent of Actual Tax Collection) Gap (79%) Actual (100%) 3.49. The large size of Pakistan's tax gap suggests that increasing the country's tax effort in an equitable and efficient way requires reforms of both tax policy and tax administration. It could suggest collusion between tax payers and tax collectors that lowers official tax receipts for private side payments, or more generally for a low tax morale that points to the need of improving the transparency and accountability of the public sector overall. The finding of more evidence for non-compliance with direct taxes than indirect taxes is fairly common, as income taxes tend to be more difficult to administer. The disparity would have been even larger without the withholding tax regime that is subsumed as part of the direct tax system, even though some of these taxes are arguably de facto indirect taxes. The withholding tax regime masks serious compliance problems, as many salaried individuals are paying withholding tax on electricity, telephone bills and bank withdrawals without claiming the adjustment for these payments on the income tax return. 50 Closing the Tax Gap 3.50. International evidence suggests that tax evasion depends on enforcement strategies. It is a fair guess that FBR is disliked by many Pakistani businessmen, professionals and workers. But many people who dislike FBR probably do so for the wrong reasons. They think it is a harsh and cruel agency, but in fact it is not nearly as harsh and cruel as it should be. At least, this is an important lesson from the US Internal Revenue Service (I.R.S.)'s National Research Program (Dubner and Levitt 2006). In this three-year study, some 46,000 randomly selected tax returns from 2001 were intensively reviewed. The study found a tax gap of $345 billion, or nearly one-fifth of all taxes collected by the I.R.S. While this tax gap is sizable, most people are compliant, and some people cheat far more than others. This becomes clear when studying the net misreporting percentage which measures the amount that was misreported on every major line item on those 46,000 returns. In the "wages, salaries, tips" category, taxpayers underreport no more than only 1 percent of the actual income. Yet, in the "nonfarm proprietor income" category, 57 percent of the income goes unreported. Why such a large difference between the wage earner and a self-employed? The only person reporting the self-employed's income to I.R.S. is the self-employed himself; for the wage earner, his employer is filling in a form to let the I.R.S. know exactly how much he has been paid. And the wage earner's taxes are automatically withheld from his every check, while the self-employed has all year to decide if, and how much, he will pay. It is not that the average self-employed worker is less honest than the average wage earner. Instead, the self-employed knows that the only chance the I.R.S. has of learning his true income and expenditures is to audit him. But since the I.R.S audit rate is so small -- the agency conducts face-to- face audits with no more than 0.20 percent of all individual taxpayers -- he can be pretty confident to go ahead and cheat. The stark differences in compliance rates across taxable items that line up closely with detection rates suggest strongly that people pay their taxes not so much because it is the right thing to do, but because they fear getting caught if they do not. A combination of good technology (employer reporting and withholding) and poor logic (most non-cheaters overestimate their chances of being audited) makes the system work. So the compliant taxpayer should dislike FBR not because it is too vigilant, but because it is not nearly vigilant enough. 3.51. Apart from better enforcement, boosting tax morale might be another way to increase tax compliance. In some countries, people appear to be far more compliant than the low audit rates would suggest. One explanation is that tax compliance is voluntary due the intrinsic motivation of citizens to pay taxes. Such tax morale depends on their attitudes toward the state (Box 3.9). If the general population feels that the tax system treats them fairly and that they are getting good value for their taxes, their willingness to pay tax increases, and vice versa. In this view, increasing tax morale, all other things held constant, should increase voluntary tax compliance. However, increasing the transparency and accountability of the government is likely to work only over the long- term. In the meantime, there is little alternative to an efficient, even-handed and enforcing tax administration. 3.52. One important issue is how much resources should be devoted to enforcing tax laws. What should be the extent of audit coverage, the strategy to choose audit targets and the penalty imposed on detected evasion? Just as it is not optimal to station a police officer at each street corner to eliminate robbery and jaywalking, it is not optimal to completely eliminate tax evasion. The tax gap estimates are not measures of the potential for additional enforcement yields because some would not be cost-effective to collect. In addition, the closing of the tax gap would also have to be evaluated from the perspective of the overall fiscal stance, so as to ensure that it is aligned with tax policy measures and does not lead to an inappropriate fiscal tightening. Nevertheless, the size of the tax gap suggests that Pakistan's enforcement measures to date are vastly inadequate. Box 3.9: Compliance and Equity A key ingredient for a well functioning public sector is a societal understanding that the quality of public services depends on everyone paying their share of the tax burden. Where this perception fails, the social compact breaks down, and tax avoidance and evasion become widespread. This leads to a vicious circle of free-riding and tax rate increases, with adverse consequences for the public finances, the quality of service delivery, and social cohesion. It follows that the same institutions that influence the quality and breadth of service delivery also affect the overall tax effort. Voice and accountability can strengthen the tax effort, as the services provided become a reflection of the desires of the broader electorate rather than of a privileged few. For similar reasons, high inequality in the distribution of political power and wealth may be prejudicial to the tax effort. Low tax revenues in some countries may reflect the low solidarity of the elite with middle- and lower- income groups: the small, wealthiest segment of the population is unwilling to pay more in taxes to provide public services, because the elite can procure many private substitutes for publicly-provided services. Source: WDR 2006. 51 3.6 Policy Reform 3.53. The Pakistan tax system underperforms with regard to adequacy, efficiency, equity and compliance. Faced with frequent revenue shortfalls, the tax system has evolved over time in a piecemeal manner without adequate consideration of how the pieces fit together. Inadequate tax collection is linked to high exemptions and low compliance, as they drive a wedge between the potential and actual tax yield. Widespread exemptions and special treatment regimes reduce tax revenues and distort the allocation of investment across sectors and asset types. They also contribute to tax evasion which distorts resource allocation and undermines vertical and horizontal equity. In addition, weak compliance leads to a tax gap between what taxpayers owe, as determined by the tax law through the statutory tax rates, and what the taxpayer actually remit to the fiscal authority. Such tax evasion involves intentionally underreporting income, revenues or wealth, overstating those items subject to tax benefits, or by failing to file appropriate tax returns. With exemptions and evasion, the tax base remains narrow and tax rates high. 3.54. While institutional transparency and accountability are important for success in revenue collection, technical aspects of public finance are crucial for a well-functioning tax system. From the perspective of efficiency, the primary contribution of taxation is in ensuring a level playing field for economic activities. From the perspective of equity, it is in providing the resources to fund equitable public spending. From the perspective of compliance, it is in ensuring low costs of compliance (and large costs of non-compliance). For all three, administrative feasibility is vital. It might appear difficult to reform a tax system in a way that it meets these multiple objectives. Fortunately, the properties of efficiency, equity, and ease of compliance lead, in large measure, to the same basic prescriptions for a reform of the tax system. Let us take as given that a particular amount of tax revenues is required to finance government expenditures. Tax reform should then be guided by three basic principles:3 · Make the tax base as broad as possible. A broad and comprehensive tax base with a minimum number of special exemptions and deductions in which every unit is taxable generally provides a good design for the taxation of income, consumption, payroll, or asset wealth. With a broad tax base, similar forms of income and expenditure face the same tax burden resulting in a level playing field. This minimizes the impact of the tax system on choices of tax payers. It also leads to horizontal tax equity, or similar tax burdens for people in similar circumstances. To gain political support for a broad and comprehensive tax base, such reforms can be sold as a package of a removing the special exemptions and deductions to fund lower tax rates. · Keep tax rates as low as possible. The broader the base, the lower the rate needed to generate a given revenue level. And lower rates lower efficiency costs. The general rule is that the excess burden of taxes increases proportionally to the square of the tax rate, so halving the tax rate implies a decline in the excess burden to a quarter. From an efficiency perspective, it is better to impose a single rate on a broad base of taxpayers, rather than dividing that base into segments and imposing different rates on each one. A broad-based consumption tax, for example, will still discourage labor supply on the margin, but choices between tradable and non- tradable goods and services will not be distorted, if all are taxed at the same rate. Income tax bases should also be broad, treating all incomes, from every source, as uniformly as possible. This need not imply redistribution of income to the rich if the broadening of the tax base focuses on eliminating loopholes that are exploited by those who can afford to employ tax lawyers. · Make compliance simple and non-compliance expensive. A system with a broad tax base and low rates is also simple. This helps to improve voluntary compliance, as such a system is viewed as more fair and reduces the incentives for clever tax avoidance and evasion schemes. It also helps tax administration, as, with a given amount of resources, it is easier to manage a system with a few rules and regulations than a system with many rates and regimes. But the carrot of simplicity is not enough and needs to go along with the stick of enforcement. This requires, among others, reliable data bases, field audits, and penalties against non-filing or under-stating of tax returns. 3 Reviewing fifty years of experience in developing countries with the BBLR approach of broader bases and lower rates to tax reform, Richard Bird (2008) concludes that this concept holds up "fairly well". 52 CHAPTER 4: GENERAL SALES TAX 4.1 Introduction 4.1. The most important global change in countries' taxation system over the last few decades has been the spread of the value added tax (VAT). The number of countries with value-added taxes increased from 48 in 1989 to 143 in 2007. The United States is the only OECD country without a VAT, although some 45 out of the 50 US states impose a retail sales tax and several states use a form of value-added taxation. The growing reliance on value added taxes in developing countries reflects, among others, a substitution away from turnover and sales taxes as well as trade taxes. The rising importance of VAT also marks a shift from income- to consumption-based taxation in the light of increasing competition among countries for mobile tax bases, especially capital, and the greater administrative ease of indirect taxes compared to direct taxes (Box 4.1). 4.2. The performance of Pakistan's VAT, the general sales tax (GST), has been disappointing. GST revenue increased from 3.1 percent of GDP in 1999/2000 to 4.1 percent of GDP in 2002/03, but then declined to 3.5 percent of GDP in 2007/08 (Figure 4.1). Since 2006/07, the GST share in FBR taxes has dropped below 40 percent and fallen behind the share of direct taxes. International comparisons confirm that GST collections in Pakistan are weak. For example, India and Sri Lanka collect around 1.5 to 2.5 percent of GDP more through GST than Pakistan. The lower collection is not because of lower rates. Until 2007/08, Pakistan charged a standard GST rate of 15 percent on sales price inclusive of any federal excise duty and/or customs duty, the same rate as in Sri Lanka and higher than the 12.5 percent rate in India. Pakistan introduced general sales tax rates of 17.5 and 20 percent for some items in 2007/08, and increased the standard rate even to 16 percent in 2008/09. 4.3. The GST on goods is collected by the federal government. After deducting a collection fee (5 percent), one-sixth of the revenues are used as pass-through grant to local governments. The inter-provincial distribution of this pool is on the basis of the 1998/99 octroi collections. The remaining five-sixths of the revenue become part of the general divisible pool and is distributed between the federal government and the provinces (vertical sharing) and among the provinces horizontally according to the NFC formula. Figure 4.1: Pakistan's GST Collection over Time and in International Comparison Pakistan's Tax Collection of General Sales Tax General Sales Tax in Pakistan and Other Emerging Economies (2006/07 or latest available) 4.5 70 8 12,000 % of Indirect Taxes PPP GNI Per Capita % of GDP % of FBR Taxes 4.0 7 60 10,000 3.5 6 50 8,000 3.0 5 40 2.5 4 6,000 2.0 % of GDP 30 3 4,000 1.5 2 20 1.0 2,000 1 10 0.5 0 0 Bangladesh Nepal Pakistan Thailand Egypt India Sri Lanka South 0.0 0 Africa 1999/2000 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 For India: IMF estimates of state-level VAT and state-level sales taxes in 2004/05. 4.4. There is a strong case for overhauling GST to strengthen revenues, improve efficiency and equity, and tighten compliance. This chapter first highlights different ways in which Pakistan's GST deviates from a modern VAT (Table 4.1). These departures lead to non-neutralities with negative impacts on efficiency, horizontal equity, and revenue performance. One of the reasons for the large deviations between Pakistan's GST and a modern VAT are the excessive discretionary powers granted to the tax administration through the sales tax law; this law suffers also from other technical and drafting shortcomings. The GST base is fragmented through generous exemptions and zero-rating provisions and the constitutional assignment of services as provincial tax base. The main conclusion of this analysis is that there is a strong case for a wholesale reform of Pakistan's GST. 53 Table 4.1: Main Findings of Chapter 4 Topic Findings Revenue Trends · As percent of GDP, the GST revenue increased from 3.1 percent of GDP in 1999/2000 to 4.1 percent of GDP in 2002/03, but then declined to 3.6 percent of GDP in 2007/08. Tax Design · The standard rate for GST was increased in 2008/09 from 15 percent to 16 percent. · Pakistan's sales tax is not neutral vis-à-vis businesses because different taxpayers are subject to different rules. · Pakistan's constitution allows the federal government to impose the sales tax only on goods, while sales of services, a provincial tax base, go largely untaxed. · Many sectors and commodities are zero-rated, including the whole chain of activities of the five export-oriented sectors, from the import stage to the export stage. · Exemptions apply to edible foodstuffs, unprocessed and unpackaged goods, machinery, equipment, and vehicles. · The GST threshold of Rs. 5 million turnover for businesses imposes registration requirements on few businesses only. Low Efficiency and · Pakistan's GST productivity of 24 percent implies that it collects less than one Narrow Base quarter compared to what it would collect in case of a uniform standard GST on all value added. · Industry contributes in indirect taxes more than thrice its GDP share, while services and agriculture contribute far less than their GDP shares. · Petroleum products, cigarettes, edible oils, automobiles, other manufacturing, iron and steel and cement contributed in 2006/07 three-fifths of all indirect tax revenues but accounted jointly only six percent of GDP. · There is an increasing reliance on a few commodities in GST collection, which makes the revenue base more volatile. Weak Legal Vessel · The sales tax act grants enormous discretion to the FBR in administering the provisions of the tax. · Certain valuation rules in the sales tax act are unusual. · The language and certain definitions are unnecessarily complicated and at variant with the Income Tax Ordinance 2001. Equity · The tax incidence of GST is moderately progressive. The effective tax as percent of GDP increases from 3.3 percent for the bottom decile to 3.8 percent for the top decile. Compliance · The tax gap for GST is estimated to be 30 percent of actual collection. This suggests weak GST enforcement of registration, return filing, and accuracy of filed taxes, and fraudulent use of tax refunds. · Close to 90 percent of the sales tax collection comes from businesses with more than 300 million rupees in turnover. These large businesses represent fewer than 2,000 units, or less than 3 percent of all filing businesses. · According to the 2008 Doing Business report, the by far highest compliance costs among Pakistan's taxes are imposed by GST regulations: the model business spends some 560 hours per year to ensure tax compliance, out of which 480 hours are spent in dealing with GST regulations. · While Pakistan's income tax administration has embraced self-assessment, the GST administration is still based on complete verification of invoices. 54 4.2 Non-Neutrality Taxpayer Categories and Services 4.5. Pakistan's GST deviates from a comprehensive VAT in important ways. In principle, a VAT is a tax on the value created in goods or services by a business at any stage of production, distribution or sales (Keen 2007). Value-added is simply the difference between the value of sales and the value of intermediate inputs in the form of goods, services, or capital. Such a consumption VAT has desirable efficiency properties, which explains in good part the global expansion of VAT in recent decades. Since the whole value added chain is covered by the tax, there is no distortion of production, distribution or sales choices. Since the purchases of capital goods are included in the inputs and hence excluded in the VAT tax base, there is no distortion in the rate of returns on capital. And since all consumption purchases are treated equally, there is neutrality also across all forms of consumption. In summary, a general VAT achieves in theory internal and external neutrality (Box 4.2). However, in practice, the VAT systems of almost all countries are not fully aligned with such a comprehensive VAT. Pakistan is no exception, but the deviations are particularly stark. Four issues are worth highlighting: taxpayer categories, services, zero-rating and exemptions, and thresholds. 4.6. First, Pakistan's sales tax is not neutral vis-à-vis business activities because different taxpayers are subject to different rules. The sales tax requires categorization of taxpayers into classes, such as importers, commercial importers, manufacturers, wholesalers, retailers, and combinations such as wholesaler-cum-retailer, importer-cum-manufacturer, and manufacturer-cum-exporter. In addition, special schemes apply to particular industry sectors, including vehicle dealers, cotton ginners, steel-smelters re-rollers and ship breakers, and telecommunication companies (in relation to mobile phones), to name but a few. The different regimes include different registration requirements for different types of taxpayers (for example, the GST threshold discussed below applies only to manufacturers and retailers, but not to other sectors or businesses liable to excise taxes in sales tax mode), different tax rates, different compliance requirements, different input tax credit entitlements, different tax bases, and different methods of collection, including withholding. In some cases, the operation of the sales tax has been so modified that, for particular types of taxpayer, that the legal provisions appear more similar to those in a law imposing excise taxes or customs duties. 4.7. Second, Pakistan's constitution allows the federal government to impose the sales tax only on goods. The sales tax on services is provincial prerogative. Following an agreement of the federal government with all four provinces in 1999/2000, the federal government administers and collects the provincial sales tax in combination with GST, and transfers the receipts to the provinces. Since the federal government retains a collection fee of only 2 percent, it has little incentive to collect these taxes effectively or expand the base to other services. As a result, the sales tax still needs to be firmly established in the service sector. Retail trade, construction, legal consultancy, accounting and many transportation services are at best taxed very lightly, even though the federal government levies excises in sales tax mode on some services. This issue is discussed in greater detail in Chapter 9. Box 4.1: Taxing Consumption or Income? Whether to tax households based on their income or on their consumption is one of the central questions of tax design. The debate over alternative tax bases involves both philosophical arguments about what constitutes a fair measure of ability to pay and economic arguments about the relative efficiency of different tax bases. Consumption tax supporters argue that the amount that an individual draws from the economy's resource pool should determine his or her tax burden. They also point out that an income tax levies a `double tax' on saving, since saved income is taxed both when it is earned and when the savings yield a return to capital. Proponents of income taxation argue that the change in an individual's command over resources between one period and the next is an appropriate measure of `ability to pay', even if those resources are not immediately consumed. Moreover, they argue that changes in resources should be taxed regardless of whether they arise from labor income or from the returns to past saving. Income taxes create two distortions: one between the before-tax and the after-tax real product wage, which distorts the labor­leisure margin, and one between the before-tax and the after-tax real rate of return to saving. The latter distorts the lifetime allocation of consumption relative to the pattern that would be chosen if the return to delaying consumption equaled the economy's pre-tax marginal product of capital. Shifting from an income tax to a consumption tax eliminates the second distortion. The key analytical issue in evaluating the welfare consequences of replacing an income tax with a consumption tax is therefore measuring the efficiency costs associated with the taxation of saving and investment. Source: Auerbach (2006). 55 Zero-rating and Exemptions 4.8. A number of sectors and commodities are zero-rated. Since 2005/06, FBR applies "zero-rating" (or more accurately described, exempting, as discussed in Box 4.3) from GST to the entire value chain, covering both domestic and export activities, of the five main export sectors, which are textile products, leather articles, carpets, sports goods, and surgical goods. Exports, which by definition are consumed abroad, are usually not subject to GST in order to avoid double taxation on final consumption, so this provision might seem innocuous. After all, the bulk of the products in these five sectors are exported, and the GST collection had resulted in a cumbersome and potentially fraudulent refund process. This step improved cash flow in these sectors and restricted falsified claims (Box 4.5). However, the measure has reinforced the narrowness of the tax base and undermines tax administration in four ways. First, while FBR imposed a three percent single stage retail sales tax on domestic sales of products from these five sectors, the enforcement record of taxes at the retail level remains very weak. In most cases, the domestic sales of textiles, carpets or leather articles are likely to escape taxation even at this low rate. Second, as the zero-rating applies throughout the whole production chain, the tax authority has to decide whether a particular input is used downstream for these five export sectors or any other sectors. This provision opens the door for misuse. Third, the zero-rating of the entire value chain has also done away with the invoice paper trail that helps in documenting economic activity and bringing taxpayers in to the tax net. Finally, the zero-rating policy has gone beyond these five export-oriented sectors. In 2005/06, zero-rating was applied to tractors and all inputs used to produce tractors; as well as to all raw materials, components and parts used in manufacturing plant and equipment. In 2006/07, buses and taxis were zero-rated for both imports and local supply. In 2007/08, electricity was zero- rated, as net GST collection was negative. In addition to fraudulent tax refunds, the negative cash flow from electricity could be explained by the delays in notifying increases in electricity tariffs as production costs rose on the back of higher fuel costs. 4.9. While the extent of zero-rating has generally increased over the last years, the sales tax includes at the same time significant restrictions on input tax adjustments. This is understandably a major area of concern to the compliant business community. 4.10. The list of exemptions remains long in spite of some reductions in recent years. Exemptions to the sales tax are primarily aimed at reducing the sales tax burden on edible foodstuffs, unprocessed and unpackaged goods, but also machinery and equipment (Box 4.4). The exemption of imports of many capital items and raw materials results in the effective conversion of many exempt supplies into zero-rated supplies. Such supplies could be partly taxed if the inputs of goods used by the suppliers were taxed. This is usually considered to be one desirable effect of exemption: that it taxes the inputs of the supplier where it is difficult or undesirable to tax the outputs. Furthermore, GST is not applied to any goods consumed in the Provincial and Federal Administered Tribal Areas and the Northern Areas. Box 4.2: External and Internal Neutrality The design of a modern VAT is guided by the principles of external and internal neutrality. External neutrality ensures that the tax burden on domestic consumption is the same irrespective of the origin of the goods or services consumed. It also ensures that the government receives tax revenue from domestic consumption. External neutrality is achieved by applying the destination principle, which is currently the preferred jurisdictional principle for value added taxes. This involves taxing imports at the same rate as domestically produced goods or services; and zero-rating exports. Internal neutrality ensures that consumers bear the same tax burden on all consumption and that taxpayers are taxed in the same way irrespective of the business structure adopted or the way in which supplies are made. It is best achieved by a broad based tax applied to all goods and services, all types of taxpayer, and all types of transaction at a single rate. In particular, internal neutrality in the legal design of a VAT requires: · minimal use of multiple rates; · minimal use of exemptions and zero-rating; · broad coverage of taxpayers by using broad concepts of `taxable activity' and `taxable person'; · application to each and every stage in production and distribution chain, including the retail stage; · a broad input tax credit mechanism designed to ensure that each taxpayer pays taxes only on its value added; and · a broad and inclusive concept of `taxable value' so that the tax burden as a percentage of consumption closely approximates the rate of tax. 56 Box 4.3: The "Zero-rating" of the Five Export Sectors FBR suspected the fraudulent processing of tax refunds from the five export sectors because net GST collections from these sectors were not only negative--as it may be expected for exporters under zero-rating--but far more negative than could be reasonably expected. FBR solved this problem by suspending all tax refunds to these sectors while at the same time exempting these sectors from filing GST returns. Therefore, even though even the FBR refers to this policy as "zero-rating", properly speaking, the five sectors were exempted and not zero rated. That is, the five sectors gained the exemption from filing GST but also lost the right to get refunds on GST taxes paid on their inputs incorporated in the exported products. Box 4.4: Sales Tax Exemptions According to the Sales Tax Act the following commodities are exempt from sales tax: · live animals including poultry, bovine meat, sheep and goats; fish, excluding live fish; eggs; · live plants; pulses, excluding those that are bottled, canned or processed; red chiles, tumeric and ginger unless sold in retailing packing bearing brand names; seeds, fruits and spores used for sowing; cinchone bark; sugar beet and sugar cane; · edible oils; fruit juices unless canned, bottled or packaged; dairy products; ice and water unless bearing brand names; · poultry and cattle feed; · table sale unless retailed under a brand name; ultrasound gel; adult diapers; dextrose and saline infusion; dialysis and angioplasty equipment and supplies; · machinery, plant, and equipment not manufactured locally, including agricultural machinery; · trailers or semi-trailers; vehicles like compressed natural gas buses, ambulances, and firefighting vehicles; · national defense equipment and supplies; spare parts and equipment for aircraft, air navigation and marine pilotage; · supplies consumed in-house for manufacture of goods; iron and steel scrap; incinerators; bricks; · newspapers, journals, magazines etc. excluding directories; · currency notes, stocks and bonds; gold or silver in unworked condition; · computer software; · imported goods of international agencies; imports of household effects by rulers of Gulf Sheikdoms; goods imported by public hospitals and non-profit organizations; and · goods produced in Pakistan, exported and then re-imported within on year. Box 4.5: VAT Refund Taxpayers whose input tax exceeds their output tax should receive refunds. This should be generally the case for exporters, because their output tax is nero and their input tax should be positive. Refund levels relative to GST collection vary around the world. A country's trade balance profile will affect refund levels, with net exporting countries likely to have higher refund levels than net importing countries. In the UK, Netherlands, Sweden, Russia, Hungary, South Africa, Canada, refunds exceed 40 percent of gross VAT collections; in Africa, Asia, and Latin America, refund rates of less than 20 percent of gross collections are more typical. In Pakistan, the GST refund rate has dropped sharply after the introduction of zero- rating/exemption of the five export-oriented sectors in 2005/06. General Sales Tax Refunds as Percentage of Gross Sales Tax Collection 50 50 45 45 Domestic Refund as 40 % of Gross Domestic 40 GST 35 35 30 30 25 25 Overall Refund as % of Total GST 20 20 15 15 10 10 5 5 0 0 1999/2000 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 57 Thresholds 4.11. An inherent problem with any sales tax is the appropriate tax treatment of small firms. The use of thresholds is a means for exempting small enterprises from sales tax registration because the revenue to be gained from such taxpayers is minimal, while the costs of compliance and administration are high. In Pakistan, GST is imposed on businesses in manufacturing and retail with a turnover above the registration threshold of Rs. 5 million (around $62,500). This threshold is higher than in many industrialized countries, but not uncommon to thresholds in other developing countries. According to the 2001/03 Economic Census, only some 20,000 businesses out of 1,567,000 businesses in the wholesale, retail, restaurant and hotels sectors had sales revenue in excess of Rs. 5 million. 4.12. Choosing the appropriate threshold requires a judicious balancing of administrative and compliance costs and additional tax revenue. Too high a threshold compromises the basic objective of raising revenue. Too low a threshold may have the authorities overwhelmed by the difficulties of implementation. Keen and Mintz (2004) show that the optimal threshold depends on the distribution of firms in the economy: the larger the concentration of sales in large companies, the higher the threshold should be set. The optimal threshold balances the increase in revenues from a reduction in threshold with the costs of administration for the tax authorities and compliance costs for the tax payer. In addition, a low threshold has economic costs as it might inhibit firms to grow to their optimal size. However, if the threshold is too high there would be far too many unregistered businesses that would capture a large market share at the expense of the registered dealers. After all, unregistered businesses have a price advantage, as they can offer the same product cheaper as they do not have to remit VAT to the tax authorities. Faced with competition from a large number of unregistered dealers, even registered businesses might find it difficult to pass on the burden of the tax to the next buyers. Under such circumstances, they will be tempted to evade VAT. Many countries, including Pakistan recently, effectively lower the GST threshold by keeping the threshold unchanged in nominal terms irrespective of inflation. Box 4.6: Crediting, Zero-rating, and Exemptions The mechanism of input tax credit (or input tax adjustment) is a key element in the design of a modern VAT. As most countries, Pakistan collects VAT (or GST) through the credit method. VAT is applied to the value of gross sales less the credit for taxes paid by registered suppliers of intermediate inputs to the business. This approach is self-enforcing as businesses need to obtain an invoice from suppliers in order to be able to claim credit for VAT payments on inputs. In a consumption VAT, full input tax credits are allowed at the time of import or purchase for all inputs, including capital and fixed assets. VAT exemptions mean that businesses do not pay any VAT on their sales and do not receive any credits for VAT paid by their suppliers. Exemptions give rise to revenue, efficiency, and administrative problems: · Revenue loss: the government misses out tax receipts as the exempt sector or product is excluded from the VAT base. · Distortion of input choices through cascading or rate escalation: If a supplier is exempt from VAT, then the buyer of the exempt supply will not receive any credit for the inputs to the exempt supply. In other words, the un-reclaimable input tax is thus included in the price of the exempt supply, making it partially taxed. A business to business transaction leads then to cascading or rate escalation: the effective VAT rate is higher than the nominal rate for the buyer of inputs from an exempt sector. In other words, exemptions operating at an intermediate stage of production increase the effective tax rate. For example, small businesses that are below the VAT registration threshold experience rate escalation for supplies that they obtain from registered non-exempt businesses. Exemptions operating at the retail stage, the final stage of the value-adding chain before consumption, reduce the effective tax rate. · Incentive to self-supply: an exempt sector can reduce its tax liability further through vertical integration as this increases the share of the value added of the supply that is not liable to VAT. · Administrative costs: exemptions increase the complexity of the tax system and create opportunities for tax evasion. The standard recommendation is to keep VAT exemptions as simple as possible; be recorded in terms of who and what is exempt and how much revenue is foregone; and be evaluated regularly to ensure they achieve their stated objectives (Bird and Gendron 2007). Typically exempted supplies of goods in OECD include medical and dental supplies; certain supplies by charities and non-profit making bodies; and certain supplies relating to sports, education, culture and religion. Zero-rating is an exemption from paying tax on sales with credit for the tax paid on inputs. It means that businesses apply a zero rate to their value added and receive a credit for the VAT paid by their suppliers. Zero-rating has no impact at all but on the final stage of the value chain, since taxes lost at one stage will be recovered at the subsequent stage. By contrast, zero-rating at the retail stage means the good remains entirely untaxed by VAT. The standard recommendation is to limit zero-rating to exports and perhaps key items consumed by poor people. 58 Low Efficiency 4.13. Beyond the revenue losses to the tax administration, the numerous holes in Pakistan's GST system imply that its efficiency properties are compromised. A standard measurement of a country's sales tax efficiency is the so-called productivity, which is defined as the share of GST (or VAT) revenues as percent of GDP relative to the standard GST rate. Pakistan's GST productivity of 24 percent implies that it collects less than one quarter compared to what it would collect in case of a uniform standard GST on all value added. The rating is low by international comparison (Figure 4.2). The weak performance is in large measure due to the small size of the domestic tax base. In particular, while private domestic consumption is more than five times as large as consumption-related imports, the revenue collections from imports and domestic activity are roughly the same. Once exemptions, evasion, and zero rating are taken into account, only a part of domestic consumption, primarily from manufacturing and imports, is captured in the tax base. 4.14. The indirect tax burden is tilted towards industry, as most services escape adequate taxation. This is borne out by a sectoral analysis of tax and GDP shares. The sectoral breakdown of tax collection is available only for direct taxes overall. Since GST accounts for close to three-fifths of overall indirect tax collections, the pattern should look similar for GST alone. Industry contributes more than thrice its GDP share, while services and agriculture contribute far less than their GDP shares (Figure 4.3). The imbalances between industry and services are far larger for indirect taxes than for direct taxes, as indirect taxes perform poorly in generating revenues from services: Furthermore, the tax base is concentrated in a few subsectors, leaving revenue collection vulnerable to fluctuations in the profitability of these subsectors. Petroleum products, cigarettes, edible oils, automobiles, other manufacturing, iron and steel and cement contributed in 2006/07 three-fifths of all indirect tax revenues but accounted jointly only six percent of GDP (Figure 4.4). 4.15. There is an increasing reliance on a few commodities in GST collection, which makes the revenue base more volatile. The top five import items (petroleum products, plastic and articles, edible oil/ghee, vehicles and auto parts, cigarettes) contributed in 2007/08 about 70 percent of all collection, compared to just under one third in 2002/03 (Figure 4.5). The contribution from petroleum products alone increased over this period from 9 percent to 38 percent. A parallel trend is evident from domestic sales tax, driven by the rising reliance on telecom and petroleum products. Figure 4.2: Pakistan's GST Productivity in International Comparison Value Added Tax in Pakistan and Other Emerging Economies (2006/07 or latest available) 60 12,000 PPP GNI Per Capita 55 50 VAT Productivity 10,000 45 40 8,000 35 30 6,000 25 20 4,000 15 10 2,000 5 0 0 Bangladesh Pakistan Nepal Sri Lanka India Egypt Turkey South Africa Thailand 59 Figure 4.3: Contributions of the Three Sectors to GDP and FBR Indirect Taxes in 2006/07 GDP (%) FBR (%) Services (22%) Services (58%) Industry (77%) Industry (23%) Agriculture (20%) Agriculture (1.3%) Figure 4.4: Contributions of Subsectors to FBR Indirect Taxes relative to GDP Ratios of Indirect Tax Share to GDP Share of 24 Sectors in 2006/07 (%) 21.3 10 9 8 7 6 5 4 3 2 Tax Share Above GDP 1 Tax Share Below GDP Share 0 s g n e e . s e e s r . s g r y t l g s s s e n o r c n e s d e m e b i l n a r n e n e l i s t l i i l i u t n u c n a g o z i r a i g e e e t i l i e t c t l t l a i e r a c l i t c i y u n i S r b O t u x e c u r m v r f T r e l t s i r r S m u t o e e d e w u r c u m e e l e e r D m a h e & c l r o T t i r s o D i v T e e u c C a m b i a r D s n S a t e F s h a n f o d g i P f n g I C r & e b a C Q M o u t o o A e . r I n u E C m & & h n R & G & a A p i C e . r t i g u c O & d & n M e l h n o t m e o y i r o s r a S d l o t i n e r e n A a F c i h t n i , . c s i M t e F p i e l r t O P w s l c O n b o e a u h l r P W E T Figure 4.5: Commodity Composition of Gross Collection from GST Pakistan's Gross Collection from Domestic Sales Tax (%) Pakistan's Gross Collection from Import Sales Tax (%) 100% 100% 90% 90% Other Import Products 80% 80% Other Dom. Products 70% 70% 60% 60% Cigarettes 50% 50% Vehicles & Auto Parts Cigarettes Edible Oil/Ghee 40% Sugar 40% Plastic and Articles Natural Gas 30% 30% 20% POL Products 20% POL Products 10% 10% Telecom Services 0% 0% 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 60 4.3 Weak Legal Vessel 4.16. The sales tax act grants enormous discretion to the FBR in administering the provisions of the tax. The current law grants broad powers to the tax administration to vary almost all aspects of the tax without parliamentary approval. While tax laws in many other countries contain provisions for delegated powers to the tax administration, they are typically only in relation to administrative matters and rules and regulations dealing with detail rather than issues related to the tax base, tax rate, or taxpayer. In Pakistan, the tax administration has exercised far greater powers under the sales tax with the result that the law can be, and frequently is, changed in significant ways through the issue of administrative orders. 4.17. Many of the articles in the sales tax act begin with a "notwithstanding" clause proclaiming that, whatever may be the content of the article, may be overturned at any time by a decision of the FBR. In order for the taxpayer to understand how the sales tax works, he or she has to read and digest carefully the Sales Tax Notifications and Statutory Regulatory Ordinances (SROs) published regularly in the Gazette that reflect FBR decisions on how the sales tax should take effect any point in time. Many of the notifications and SROs are relatively innocuous involving clarifications of payment procedures, previous SROs and changes in the tax forms. Others, however, engage in fine tuning of the sales tax act and initiate sales tax policy changes. In the first nine months of 2007, about 75 changes in the sales tax law occurred through issuing new SROs. By comparison, there were 20 new SROs for the income tax and 15 for the excise tax. Many of these SROs reflect a tendency towards, as already mentioned, granting more extensive zero-rating and exemption status and relying more on presumptive methods. For example, a SRO from 2005 stated that pesticides were to be taxed with a 30 percent fixed value addition. Also, in the same year, the FBR fixed the price of potassium fertilizers, both domestic and imported, at Rs. 4,610 per ton. FBR also began to impose sales tax on the "printed" retail price set by a manufacturer, thus pre- collecting tax that should have been paid at later stages. In 2005/06, this method of tax collection was applied to toiletries, tea, confectionery and footwear. 4.18. Certain valuation rules in the sales tax act are unusual. A modern VAT should operate as a tax on consumption expenditure. This means that the use of market value rules should be only a last resort. At present, Pakistan values supplies at market value if there is even a small element of in-kind consideration. A more appropriate approach would be to value the consideration given for a supply, so that where consideration is in-kind, the value of the supply is the market value of the in-kind consideration rather than of the supply. In any case, there are significant deviations even from the valuation rules in the sales tax act itself because of the use of multiplicity of special regimes in effect. 4.19. Other weaknesses of the sales tax act include: · The sales tax is applied to "applications to own use" but the provisions implementing this are not clearly set out and the underlying policy rational is not made apparent. · Clear tax accounting rules with regard to applicable time of supply rules are missing. · Some definitions are used to implement material provisions. For example, the definition of `manufacturer or producer' has provisions regarding liability and collection, while the definition of `retailer' includes rules on price notification and turnover. · There are inconsistencies in terminology, including references to `deduction, credit, or adjustment' of input tax in a single sentence, when all three terms are apparently intended to refer to the same thing. In some cases there is repetition of effect in the legal provisions. · Non-adherence to the original structure of the sales tax act has led to new provisions being inserted in unexpected places. · The language and certain definitions could be simplified and brought into line with the plain English style of the Income Tax Ordinance 2001. This includes the definitions of taxable persons, firm, business, and undertaking, and the concepts of taxable person, taxable activity, and taxable supply. 61 4.4 Equity 4.20. The main criticism leveled at GST is that it is regressive. Since poor people spend a higher fraction of their income on consumption than non-poor people, a consumption tax is likely to impose a higher proportional burden on poor people (Refaqat 2005). However, the regressive nature of GST is mitigated in a multi-rate system, where reduced tax rates apply to essential consumption goods, such as food and other basic necessities. While poor and non-poor people benefit equally on their purchases of such commodities, poor people benefit more as they spend a larger fraction of their income on essential consumption goods. As we will see, this feature makes Pakistan's GST fairly proportional in its incidence across household deciles. 4.21. Calculating the tax incidence of GST is fairly straightforward. Since the same procedure is used for the two other consumption taxes, federal excises and custom duties, it is worth explaining the procedure briefly. The calculations rely on two assumptions. GST is passed forward to consumer who carries the final tax burden; and tax evasion is distributed among the population in the same proportion as consumption expenditures. The steps are as follows. The 2004/05 PSLM provides detailed information for each household on their consumption baskets. FBR also reports GST receipts in great detail. We aggregate consumption items into 32 categories so that we can match consumption expenditures with tax receipts. For example, the PSLM contains detailed information on the consumption of milk products: fresh milk, lassi, powdered milk, cheese, etc. We summarize these codes into one "milk product" group that we can pair with GST receipts from "dairy produce." Next, we estimate the consumption of a particular category, sum these expenditures for each category by household decile, and divide the total amount of consumption of the good in that decile by the total consumption of that good in all deciles. This gives us an estimate of the relative share of a consumption category for each expenditure decile. Then, we distribute GST receipts for each of the 32 categories based on the relative amount of consumption in each decile. For example, returning to milk products, according to PSLM data, the lowest expenditure decile accounts for 3.5 percent of all milk consumption in Pakistan. Hence, we allocate 3.5 percent of consumption taxes related to dairy produce to that expenditure group. 4.22. The tax incidence of GST is moderately progressive. Figure 4.6 (left panel) displays the results when households are lined up according to expenditure deciles. The effective tax as percent of GDP increases from 3.3 percent for the bottom decile to 3.8 percent for the top decile. The bottom decile contributes 3.0 percent of total consumption but only 2.8 percent of total GST, the top decile contributes 32 percent of total consumption and 34 percent of total GST. The moderate progressivity does not suggest that Pakistan's GST is a broad-based consumption tax. Instead, it reflects the fact that goods that are predominantly consumed by households in the lower deciles attract lower GST rates, or are even exempt altogether, than other goods. For example, major food categories are largely exempt under GST, and the consumption of food is far more concentrated in the lower deciles than overall consumption: the bottom 50 percent households account for 29 percent of food consumption, but only for 19 percent of overall consumption (Figure 4.6, right panel). Figure 4.6: Distribution of Tax Burden and Consumption by Household Expenditure Decile Incidence of General Sales Tax by Household Consumption Deciles Distribution of Food and Overall Consumption in 2004/05 (% of Total) 35 3.8 35 35 Effective Tax Rate (% of GDP) 3.7 30 30 30 3.6 25 25 25 3.5 20 20 20 3.4 15 15 15 3.3 10 10 Food 10 3.2 Share in Total Taxes 5 (%) 5 5 3.1 Consumption 0 3 0 0 Bottom 2 3 4 5 6 7 8 9 Highest Bottom 2 3 4 5 6 7 8 9 Highest 62 4.5 Compliance Tax Gap 4.23. The estimation of the tax gap for GST draws on a detailed input-output table. This table provides information on total use or gross sales for each of the 81 sectors, in addition to the values of primary and intermediate use, imports and exports, and investment expenditures. This gives us all the necessary data to model Pakistan's potential sales tax base from domestic sales, including taxable supplies, input credits, and refunds on exports. Our methodology provides disaggregated estimates of the sales tax gaps from domestic sales for a large number of economic sectors. This makes it a useful tool for identifying non-compliant sectors and allocating enforcement resources to these sectors. However, there are a number of limitations. In particular, the most recent input-output table of Pakistan's economy, as available with the Federal Bureau of Statistics, is for 1989/90. We update this model to reflect the level of economic activity by sector to values for 2004/05, using ratios and relationships from the 1989/90 model and 2004/05 national accounts data for the 81 sectors. This is in some ways an ad-hoc adjustment, and it will be important to redo the sales tax gap estimates whenever an input-output table becomes available that captures more accurately the today's structure of Pakistan's economy, including the current input-use coefficients of each sector. Furthermore, the model does not provide a separate estimate for the import part of the sales tax gap. In order to obtain an estimate of the total sales tax gap, we assume that imports are subject to the same rate of non-compliance as that for domestic sales. 4.24. The tax gap for GST is estimated to be 30 percent of actual collection, or Rs. 113 billion in 2007/08. At first sight, this might seem like a fairly moderate tax gap in view of the low yield of GST. However, it is important to remember that that tax gap is calibrated with reference to the existing legislation. Any tax exemptions through constitutional provisions, special input tax credits, concessional rates, or tax expenditures are not captured. In fact, from the 81 sectors in the input-output model, no more than 45 have a positive taxable portion. They account for one third of all total output in the model, which reflects that the bulk of the agricultural and service sectors are legally outside the GST framework. Figure 4.7: Compliance for General Sales Tax Actual Tax Collection and Tax Gap of General Sales Tax Gap (30%) Actual (100%) 63 Enforcement 4.25. The size of the GST tax gap suggests weak GST enforcement of registration, return filing, and accuracy of return filing. Just as GST evasion strategies are manifold (Box 4.8), the GST enforcement strategy has to be comprehensive. The first step is to ensure that all businesses with turnover above the GST threshold are registered. The next concern is to ensure that all registered businesses actually file GST returns. Some 130,710 businesses were registered for the sales tax in 2006/07. Among those registered, 88,675 filed regular non-zero monthly GST returns. In other words, some 36,365 business, or about one in three of all registered businesses, filed either zero returns or no returns at all (Figure 4.8, left panel). Clearly, there is a need for a special examination of firms in this category to ascertain their tax potential. Figure 4.9 gives us some insights about the sectoral distribution of filing compliance. This indicator gives the number of registered taxpayers relative to the number of returns files for GST. Manufacturers, importers and exporters have the highest level of filing compliance. This is expected for exporters as they are typically seek refunds for GST paid on inputs. Manufacturers and importers may represent a better "tax handle" since they are likely to be more visible to the tax authorities. A third issue is to ensure that businesses file accurate VAT returns based on valid invoices (Box 4.7). Close to 90 percent of the sales tax collection comes from businesses with more than Rs. 300 million in turnover (Figure 4.8, right panel). These large businesses represent fewer than 2,000 units, or less than 3 percent of all filing businesses. By contrast, some 32,000 businesses contributed less than 2 percent of all sales tax revenue. From this perspective, an immediate priority is to make sure that businesses with large potential GST liabilities file accurately. Figure 4.8: Distribution of Registered GST Taxpayers by Turnover: Number of Businesses and GST Contributions Distribution of Number of GST Taxpayers by Turnover in Million Rupees for 2006/07 Distribution of GST Taxpayers by GST Contributions in Million Rupees for 2006/07 10-20 5-10 1-5 Less than 1 > 300 100-200 200-300 20-50 Non/Null Filers 50-100 50-100 20-50 100-200 10-20 Non/Null Filers 200-300 5-10 1-5 Less than 1 > 300 Figure 4.9: Return Filing of Registered GST Taxpayers by Sector Registration and Return Filing for General Sales Tax in 2006/07 by Sector 40,000 Filing (% of 80 Registered) 35,000 70 30,000 60 25,000 50 20,000 40 15,000 30 Filing (#) 10,000 Registered (#) 20 5,000 10 0 0 Service Distributors Exporters Manufacturers Wholesalers & Importers Providers Retailers 64 Self-Assessment 4.26. A joint effort from governments, businesses, and consumers is important to ensure good GST compliance. For businesses, the compliance cost include personnel and material for producing GST documentation, preparing GST accounts, filing of GST returns, and interacting with the tax administration. The evidence from the 2008 Doing Business survey suggests that these steps involve a large effort. As discussed in Chapter 3, the by far highest compliance costs among Pakistan's taxes are imposed by GST regulations ­ the model business spends some 560 hours per year to comply with the tax rules, out of which 480 hours are spent in dealing with GST regulations (Figure 3.10). These compliance costs arise from the complexity of the legislation, lack of guidance from tax administration, and ­ for large enterprises ­ the lack of standard principles on the treatment of international trade of goods. 4.27. While Pakistan's income tax is based on self-assessment, GST enforcement is still based on complete verification of invoices. The international experience with invoice verification has been disappointing. Relying on cross-matching of invoice numbers of business transactions places a too heavy reliance on documentation, and too little reliance on verifying the underlying transactions. Instead, modern VAT administrations endorse self- assessment. This means that the taxpayer assesses its requirement to register; the tax treatment of its outputs; its entitlement to input tax deductions; and its entitlement to refunds. The tax administration is centred on a risk assessment approach to target structured audits. This implies that tax administrators check returns for accuracy based on the reported information but there is no systematic cross-checking of invoices outside of the audit situation. Supporting documentation are not lodged with returns but only requested in case of an audit. Box 4.7: Blacklisting of Businesses Pakistan's GST rules provide for `blacklisting' and suspension of registration in case a business is found to have issued a `fake invoice'. The consequences of blacklisting a supplier lie primarily with the supplier's customers. They will be denied an input tax credit for purchases from the blacklisted firm while the suspension is in force. Although not provided by law, the denial of input tax credits may apply also retrospectively to prior purchases. Most countries do not provide for blacklisting in the enforcement of the VAT. If its use can be justified, the primary consequences should lie with the supplier rather than customers, unless of course those customers have been complicit in fraud or indifferent to the possibility of fraud. However, there is already adequate provision in the sales tax act to cover the latter situation. As a matter of equity, innocent customers should not be punished for the default of their suppliers. Thus, customers should not be denied input tax credits retrospectively if they purchased in good faith and paid the price of the goods or services purchased. While a blacklisting is in force, customers should only be denied input tax credits for valid purchases if the blacklisting is known to them. To this end, FBR should be required to publish an up-to-date list of blacklisted suppliers and no consequences should befall genuine customers prior to the date of publication. 65 Box 4.8: The Many Types of VAT Evasion and Fraud · Under-reported sales. A trader may report only a proportion of sales, falsifying records and accounts to match, or may make some sales `off the books' entirely. Under a VAT, they may or may not issue an invoice. If they do, the customer, if registered for VAT, may seek the corresponding credit. The ultimate success of the evasion then depends on the inability of the revenue authorities to discover that more VAT invoices have been issued but not declared by their issuer. · Failure to register. The commonest such cases under the VAT are relatively small businesses operating close to the level of turnover at which registration becomes compulsory, that fail to register, saving both the VAT for which they would be liable and VAT compliance costs. "Ghosts"--traders wholly unknown to the revenue authorities--may be able to evade income taxes as well as VAT. · Misclassification of commodities. When traders have sales that are liable to tax at different rates, or some of which are exempt items under the VAT, they may reduce their liability by exaggerating the proportion of sales in the lower-taxed categories. · Omission of self-deliveries. Goods or services produced by the business and consumed by the proprietor or employees, in principle taxable, may not be declared. · Tax collected but not remitted. This may be possible either through false accounting (under-reported sales, as above), by engineering bankruptcy before tax is paid, or in other ways. More particularly, the `missing trader' frauds typically involve registered businesses charging their customers VAT but disappearing before paying tax to the authorities. · Imported goods not brought into tax. If tax is not levied at the border, then there is a potential gain from purchasing imported goods bearing no tax and then reselling them in the home market. · False claims for credit or refund. This is the most obvious way to exploit the credit mechanism: "A VAT invoice..." as Bird (1993) puts it, is "a check written on the government." The universal zero-rating of exports means that fraudulent claims to have exported commodities are an intrinsic difficulty for the VAT. · Credit claimed for VAT on purchases that are not creditable. This arises in two main forms. First, when traders supply a variety of outputs, some subject to VAT and others exempt, they have an incentive to allocate inputs to production of the taxed items (in respect of which input tax credit is available) rather than the exempt (for which it is not). Second, items bought for private consumption may be misrepresented as business inputs, allowing the VAT to be recovered (and income tax liability reduced). · Bogus traders. Companies may be set up solely to generate invoices that allow recovery of VAT. Such "invoice mills" exploit the practical impossibility of crosschecking every invoice against evidence that earlier tax has been paid. Source: Keen and Smith (2007). 66 4.6 Policy Options 4.28. GST is Pakistan's tax with the largest revenue potential from tax policy reform. Our analysis makes the case for the urgent need to reform GST in order to strengthen the government's revenue position. The reforms will also help economic growth by supporting businesses through a level playing field for GST and simplified compliance. The reform entails the broadening of the GST base via the inclusion of services and the elimination of exemptions and special treatments of sectors. The induction of services into GST will require revenue sharing with provinces. Other GST reform components include the simplification of the tax, the modernization of legislation, the improvement of the use of administrative resources, and an open communication between authorities and businesses. Ultimately, GST reform is impossible without strong government commitment to change (Table 4.2). 4.29. Pakistan needs to overhaul its existing sales tax legislation along the lines of a modern VAT. The sales tax act should be substantially redrafted in the form of a new general sales tax act. The new GST law should repeal and replace the current act while incorporating relevant existing provisions, terminology, and concepts where possible, and should aim to bring the existing GST in line with international best practice. The draft law could be released for public consultation well in advance of its introduction to parliament, so that stakeholders can provide the benefits of their insights and recommendations. This would require agreement on a suitable time frame for implementation. 4.30. Prior to the implementation of any legislative change, a detailed economic analysis should be conducted to determine the revenue implications of changing from the current sales tax to a modern VAT- style tax. This study should provide recommendations on the following issues: · extension of the tax base to services; · the use of exemption and zero-rating, and the treatment of refunds; · the threshold for registration and whether voluntary registration should be allowed; and · the extent to which simplified schemes may be used to tax certain sectors or taxpayers more efficiently. 4.31. The administration of GST on a self-assessment basis should be integral part of the new act. In line with FBR's experience on reforming income tax, this would require: · A review of all aspects of the administration of sales tax; · Technical assistance to equip FBR personnel to effectively implement administration of sales tax on a self-assessment basis with all the necessary support for taxpayer education; · Establishment of an effective audit and enforcement function; and · Establishment of effective internal and external appeal and review mechanisms. 4.32. Strengthening the GST legislation would also require curtailing FBR's administrative powers. This would include deleting the broad powers to effectively legislate by administrative order from the legislation; and limiting rule making powers to deal with subsidiary matters only. Neither the government nor FBR should be empowered to effect substantive changes to the tax without parliamentary approval. Table 4.2: Policy Reform Options for General Sales Tax Topic Findings Legislation · Take policy decision to convert general sales tax into shared federal-provincial tax covering both goods and services and under federal administration · Redraft the sales tax act in the form of a modern VAT. Tax Base · Expand successively the base of the sales tax Administration · Broad powers to effectively legislate by administrative order should be deleted from the legislation · Ensure that the new sales tax will be administered effectively on a self-assessment basis 67 CHAPTER 5: CORPORATE INCOME TAX 5.1 Introduction 5.1. Globalization restricts the ability of governments to set tax policies independently of those in other countries. Pakistan's economy takes part in the increasingly dense network of regional and global activity. Over the last two years, Pakistan received around $10 billion annual net capital inflows, over half of which was in the form of foreign direct investment; this compares to net capital outflows of around $0.5 billion at the beginning of the decade. Over the same period, trade flows as percent of GDP increased from around 15 percent to 30 percent. Saving and investment, and goods and services, move with increasing ease across borders. This has important implications for the system of enterprise taxation. Tax bases are now significantly more mobile, especially in the case of capital income. In addition, the measurement and assignment of tax bases have become more difficult. For example, the product a multinational company may be designed, produced, assembled and sold in multiple countries. This allows the firm to manipulate prices across production stages and countries to minimize its tax liabilities. These factors give rise to the possibility of tax competition, with governments setting their corporate taxation policies strategically in response to the tax policies of other countries. This competition can, in the extreme, lead to a race to the bottom. Similarly, as companies decide where to invest capital, governments are increasingly judged relative to their neighbors, as investors often take a country's tax system as a signal for its approach towards business. Whether through tax competition or tax mimicking, globalization has accentuated the need for continuous adaptation and deeper international cooperation, and raised the cost of standing still. 5.2. An almost global tax reform movement over the last two decades is the fall in the statutory corporate tax rates. This drop is evident in all regions, although in different degrees (Norregaard and Khan 2007). In the countries of the OECD and Asian Pacific, the average top statutory corporate tax rate declined from close to 40 percent in the early 1990s to around 30 percent in 2006. The trend reflects the increasing tax competition and capital mobility among national jurisdictions, backed by the easing of budget constraints due to strong economic fundamentals. While the reductions in corporate income tax rates were global, the adjustments in the tax base diverged between developed and developing countries. In OECD countries, the decline in statutory rates has come together with a broadening of the tax base through a scaling back of general deductions and exemptions and special incentives. By contrast, in many developing countries, the tax bases have narrowed due to increasing reliance on tax holidays, free trade zones, and tax breaks. As a result, corporate tax revenues have increased in the last decade in OECD countries, but declined in many developing countries, with Asia-Pacific countries lying somewhere in- between. 5.3. In spite of the recent improvement in Pakistan's revenue performance, there is room for further gains. The chapter looks at the recent performance of Pakistan's corporate income taxes in the light of these international trends (Table 5.1). Tax incentives lead to revenue losses, yet their impact on investment is quite uncertain. They also contribute to large variations in the effective tax burden of companies across sectors, and, together with withholding taxes and small company taxes, result in a complex system of enterprise taxation. These distortions lower the efficiency and equity of the corporate income tax, and, helped by weak administrative enforcement, also reduce compliance. Changes in tax policy and administration, including measures in the area of international tax provisions, have the potential to deliver large increases in revenue collection of corporate income taxes while making the corporate income tax less discriminatory and distortionary. 68 Table 5.1: Main Findings of Chapter 5 Topic Findings Revenue Trends · Statutory corporate income rates have fallen across the world in the last two decades, whereas tax exemptions have decreased in developed countries and increased in some developing countries. · Pakistan's corporate income tax rates have come down markedly. In spite of the rate reductions, revenue collections from the corporate income tax collection improved from 1.2 percent of GDP in 2000/01 to 2.5 percent of GDP in 2007/08. Tax Design · Pakistan's applies a sensible mix of resident-based and source-based taxation. The rules governing the tax base of the corporate income tax are broadly in line with international practice. · The contribution of the industrial sector to direct tax collection in 2006/07 was more than twice its GDP share. Effective Tax Rates · The average and marginal effective tax rates vary greatly by sector and are relatively high. Withholding Taxes · Withholding taxes come in three varieties (income tax is withheld at source; presumptive withholding taxes in excise mode, and adjustable withholding taxes in excise mode). These withholding taxes account jointly for more than half of all income tax revenue. Small Companies · There is a favorable tax treatment for small companies. To qualify for this regime, companies must have annual turnover of less than Rs. 250 million (around $3.1 million), no more than 250 employees, and be set up in recent years. International Tax · The system is largely in line with international practice. Provisions · Pakistan has entered into 54 double tax agreements with other countries. Equity · Under standard assumptions for the final incidence of corporate taxes, the effective tax rate as percent of GDP increases moderately from 1.1 percent for the bottom decile to 1.4 percent for the 9th decile. It jumps to 3.9 percent for the top decile. Compliance · The tax gap is estimated to be about 218 percent of the actual corporate income tax payments. For 2007/08, this would imply a tax gap of over Rs. 450 billion. · In 2006, FBR had 23,100 registered companies, compared to around 47,500 registered at the Securities and Exchange Commission of Pakistan. Among FBR's cases, slightly less than half filed a tax return. And among those who filed, two-third declared either a negative or zero taxable income, and one-third positive taxable income. Figure 5.1: Pakistan's Corporate Income Tax in International Comparison Pakistan's Gross Tax Collection of Corporate Income Tax Corporate Income Tax in Pakistan and Other Emerging Economies 3.0 80 (2006/07 or latest available) % of FBR 9 12,000 PPP GNI Per Capita Gross Direct Taxes 2.7 % of FBR Gross Taxes 70 8 2.4 10,000 % of GDP 60 7 2.1 6 8,000 50 1.8 5 1.5 40 6,000 % of GDP 4 1.2 30 3 4,000 0.9 20 0.6 2 2,000 10 0.3 1 0.0 0 0 0 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 Bangladesh Sri Lanka Nepal Egypt Turkey Pakistan India Thailand South Africa 69 5.2 Tax Design and Revenue Trends 5.4. Pakistan's applies a sensible mix of resident-based and source-based taxation. In a global economy, there is a distinction between where income is earned (source) and where its owner resides (residence). Accordingly, countries may seek to tax corporate income on source-basis (taxing income arising from within its own borders), or residence-basis (taxing the worldwide income of resident company). Most countries impose combination of the two, and Pakistan is no exception. Pakistan's corporate income tax is imposed on the world- wide profits of resident companies, which have control and management located in Pakistan, and the Pakistan source profits of non-resident companies. From a revenue perspective, taxing the world-wide income of resident enterprises is important as they increasingly earn profits abroad. From a theoretical perspective, a residence-based system is consistent with the principle of capital export neutrality which requires that the investment tax rates be the same regardless of where the investments take place. This system, if applied across the world and combined with global capital mobility, ensures global production efficiency by equalizing pre-tax marginal returns to investment across countries. At the same time, taxing non-resident companies for income earned within Pakistan broadens the tax base without harming Pakistan's attractiveness for foreign investment. This is because multinational companies often receive tax credits in their home country for any taxes incurred abroad. In such a situation, failing to impose corporate income tax on profits earned within Pakistan would simply transfer taxes to the capital-exporting country. 5.5. The rules governing the tax base of the corporate income tax are broadly in line with international practice. The corporate income tax collects revenue from a firm's profits. They are measured as net accounting profits, which equals gross revenues less operating costs plus capital adjustments including interest expenses, depreciation allowances, and inventory adjustments. The base is defined as active business income earned from the sale of goods and services, and any passive income received in the form of rents, interest, management fees or royalties. Deductions from chargeable income include, among others, interest expenses on company debt; expenses for research and development and training; irrecoverable debts; cost of sales, rents, repairs and royalties; depreciation allowances for tangible fixed assets; amortization of the cost of intangible assets; and social welfare contributions. State Bank of Pakistan's profits are exempt following international practice (Box 5.1). 5.6. Following international trends, Pakistan's corporate income tax rates have come down markedly. In the early 1990s, banking, public, and private companies were taxed at the rate of 66 percent, 44 percent, and 55 percent, respectively. By 2007, Pakistan adopted a uniform corporate tax rate of 35 percent on taxable profits for both public and private sector companies. The main exception to this uniform rate approach is small companies which are taxed at a lower rate, as discussed below. 5.7. In spite of the rate reductions, collections from the corporate income tax improved during this decade on the back of the economic upturn. It increased from 1.2 percent of GDP in 2000/01 to 2.5 percent of GDP in 2007/08 (Figure 5.1). The share of the corporate income tax jumped from one-half to two-thirds in gross direct taxes, and from one-seventh to one-quarter in gross FBR taxes. By international standards, this is a fairly impressive revenue effort in view of Pakistan's per capita income level, although the revenue productivity, calculated as corporate income tax revenue as percent of GDP divided by the statutory tax rate, is below countries like Turkey, India, Thailand and South Africa (Figure 5.2). Revenue collection improved sharply in 2006/07, primarily due to the doubling of nominal tax collection from the financial sector, in addition to a solid performance from public companies in the oil and gas sector and foreign companies. 5.8. The direct tax burden falls more on some economic activities than on others, which harms efficiency and equity. This is borne out by an analysis of tax and GDP shares across the three sectors, repeating the tax base analysis of Chapter 3 but restricted to direct taxes. The sectoral breakdown of tax collection is available only for direct taxes overall. Since the corporate income tax accounts for two-thirds of overall direct tax collections, the pattern should look similar for the corporate income tax. The contribution of agriculture to federal direct taxes is negligible, as agricultural income is a provincial tax base. However, even accounting for provincial taxes would not change the picture substantially, as provinces barely tax agricultural income. By contrast, industry's tax contribution is more than twice its GDP share, and services' tax contribution is just below its GDP share (Figure 5.3). As a result, the tax base remains tilted towards a few subsectors, so that revenue collection remains vulnerable to fluctuations in the profitability of these subsectors. Finance, petroleum products, telecommunication and edible oil in 2006/07 contributed three-fifths of all corporate income tax revenues but accounted only for one-seventh of GDP (Figure 5.4). 70 Figure 5.2: International Comparison of Corporate Income Tax Productivity Corporate Income Tax in Pakistan and Other Emerging Economies (2006/07 or latest available) 30 12,000 PPP GNI Per Capita 25 10,000 20 8,000 15 CIT Productivity 6,000 10 4,000 5 2,000 0 0 Bangladesh Sri Lanka Nepal Egypt Pakistan Turkey India Thailand South Africa Figure 5.3: Contributions of the Three Sectors to GDP and FBR Direct Taxes in 2006/07 GDP (%) FBR (%) Services (58%) Services (51%) Industry (23%) Industry (49%) Agriculture (20%) Agriculture (0.2%) Figure 5.4: Collection of Corporate Income Tax by Subsectors Ratios of Corporate Income Tax Share to GDP Share in 2005/06 and 2006/07 (%) 16 27 28 Pakistan's Gross Tax Collection of Corporate Income Tax 11 100% Foreign 10 Companies 90% 9 8 Public Sector 80% companies 7 6 70% 5 60% 4 2005/06 3 2006/07 50% Private Tax Share Above GDP Share 2 Companies Tax Share Below GDP Share 40% 1 0 30% Sugar Fertilizer Others Edible Oils Automobiles Telecom Transport, Stor. & Commun. Textiles Petroleum Prod. Wholesale & Retail Electricity & Gas Distribution Finance & Insur. Iron & Steel Construction Cement Cigarrettes 20% Banking 10% Companies 0% 2005/06 2006/07 2007/08 (Jul-Mar) 71 Box 5.1: Should State Bank's Profits be Subject to Income Tax? In the first half of 2007/08, the State Bank of Pakistan, Pakistan's central bank, remitted profits in the amount of Rs. 47.3 billion to the federal treasury. These profits were exempted from income tax, so the whole amount was booked as non-tax revenue. Without this exemption and assuming a the standard corporate tax rate of 35 percent, corporate income tax receipts would have increased by Rs. 16.6 billion. This would have boosted the annual projected tax-to-GDP ratio by around 0.15 percent. Thus this question has been often asked: should the State Bank profits be subject to ordinary taxation, mainly the corporate income tax, or should they be exempt, given the special status of central banks? The taxation or exemption of central bank profits matters for the revenue position of the federal government in a number of cases. First, in some countries not including Pakistan, central banks are not fully owned by the government, and the taxing of profits distributed to "other" shareholders would make a difference to the final inflow of government revenues. Second, in Pakistan, income tax receipts are part of the divisible pool, where over two-fifths of the resources are redistributed on a population basis to the provinces in line with the January 2006 Amendment of the National Finance Commission Award. By contrast, State Bank profits are currently counted as civil administration receipts and accrue fully to the federal government. Finally, even in those cases where the revenues received in the end by the federal governments are the same whether or not central bank profits are taxed, the "credit" for collecting some of those revenues would go to different institutions within the central government. If the State Bank profits are taxed, the recognition would go to the Federal Board of Revenue. This may be seen as a superficial issue, but it is unlikely to be so for tax administrations under pressure to increase actual tax collections. The general consensus of the economic literature and international finance institutions is to exempt central bank profits from corporate income taxation. The transfer of realized profits to the government should take place after ensuring an appropriate level of reserves. However, there are two cases in which the taxation of central bank profits is recommendable: when the central bank is owned partially by private shareholders, and when the central bank engages in commercial activities competing with private financial institutions. There are two main reasons for exempting central bank profits from income tax. First, profits and losses from the operations of central banks are generally quite different from those from financial institutions, or other types of businesses. Central Banks do not, and should not, maximize profits, and their level of profits is not an indication of economic performance or profitability. Central bank profits are more often than not related to the implementation of monetary policy. Profits are also significantly related to the monopoly enjoyed by central banks in issuing national currency. On the other hand, economic losses, at least in purely accounting terms, do not impose restrictions to the functioning of central banks because these institutions do not even need capital to perform their main functions although at the limit, a deteriorated financial position can endanger the ability of the monetary authorities to effectively carry out a sound monetary policy. Second, central banks generally require a minimum level of liquidity in order to fulfill their mandate. Because the results of a central bank can be volatile, the central bank may need to recapitalize profits or augment reserves, and these needs are likely to vary from year to year. It is in this context that extracting a fixed proportion of the central bank profits can be seen as limiting its ability to keep an adequate balance of assets and liabilities. Furthermore, regardless of the application of a corporate income tax to central bank profits, only realized profits should be transferred to the treasury. If the tax on net income is applied on non-realized profits, the central bank might lack the funds to disburse its tax liabilities timely and impose an undesirable pressure on its financial accounts. What is the general practice in the international experience on this issue? The table below shows a summary of the practices regarding taxation of central bank profits in 101 countries, based on financial statements of the central bank in each country. Some 90 countries exempt their central banks from the corporate income tax, while 11 countries do tax central bank profits. Most countries also arrange for the transfer of the central bank profits to the government after appropriate allowances for reserve accumulation have been made. Among the group of 11 countries with corporate income taxation, most of them are developed countries where the economy is considered stable and the monetary authority expected to display a great deal of autonomy and financial strength. Is the Central Bank Subject to Corporate Income Taxation? YES Austria, Belgium, France, Greece, Italy, Japan, Philippines, Portugal, South Africa, Turkey, and UK. NO Afghanistan, Albania, Argentina, Australia, Barbados, Belarus, Belize, Benin, Burkina, Cote D'Ivoire, Guinea- Bissau, Mali, Niger, Senegal, Togo, Bermuda, Bhutan, Bolivia, Botswana, Brazil, Bulgaria, Cameroon, Central African Republic, Chad, Congo, Equatorial Guinea and Gabon, Canada, Chile, Colombia, Croatia, Cyprus, Czech Republic, Denmark, Dominican Republic, East Caribbean Area, Estonia, Finland, Georgia, Germany, Ghana, Guatemala, Guyana, Honduras, Hong Kong, Hungary, Iceland, India, Iraq, Ireland, Jamaica, Jordan, Kazakhstan, Kenya, Korea, Kuwait, Latvia, Lithuania, Luxembourg, Macedonia, Malaysia, Mongolia, Namibia, Nepal, Netherlands, New Zealand, Norway, Pakistan, Papua New Guinea, Paraguay, Peru, Poland, Russia, Rwanda, Serbia, Singapore, Slovakia, Slovenia, Spain, Sri Lanka, Sudan, Sweden, Switzerland, Trinidad and Tobago, Uganda, United States, Uruguay, Venezuela, and Zimbabwe. 72 5.3 Tax Incentives Effective Tax Rates 5.9. Investment decisions depend not only on the statutory corporate income tax rate, but on the combined effect of enterprise tax rates and tax bases. Low statutory rates may not necessarily imply that the tax regimes are friendly towards investment. According to 2008 Doing Business survey data for 175 countries, corporate income taxes account for no more than 36 percent of the total enterprise tax burden. In particular, tax incentives can lower the burden of the corporate income tax substantially. The last two decades have seen an increasing expansion of tax holidays, free trade zones and tax breaks in Pakistan. The 2001 Income Tax Ordinance contains an extensive list of exemptions and concessions for the corporate sector (Box 5.2). 5.10. There are two core measures to take account both the tax rate and the tax base in the calculation of the effective tax burden on investment. The first key measure is the marginal effective tax rate (METR) which refers to the intensive margin of capital formation: it is the additional tax paid by a firm when it decides to invest one more unit of capital that is just breaking even. A positive METR indicates that investment is discouraged; a negative METR indicates that investment is subsidized. The second key measure is the average effective tax rate (AETR) which refers to the extensive margin of capital: it is total taxes paid as a fraction of gross corporate income. AETR is relevant for understanding the capital location decisions of firms. 5.11. The AETRs vary by sector and are relatively high. The left panel of Figure 5.5 shows the AETR paid by corporations that filed tax returns in 2006-07 for the main economic sectors. Taxes paid include income and profit taxes, of which the main components are advance tax payments and withholding taxes. The calculations are based only on those returns that showed gross income of Rs. 1 million and above and positive tax payments. The statutory corporate tax rates in this tax year were 39 percent for banking and financial institutions, 37 percent for private companies, 35 percent for public companies, and 20 percent for small companies. The overall average tax rate including withholding taxes is 34.3 percent, or very close to the statutory tax rate of 35 percent. However, there is substantial variation in AETRs across sectors, ranging from a high of nearly 46 percent for beverages to a low of 14 percent for cement. Some AETRs exceed the relevant top statutory tax rates due to the inclusion of withholding taxes. The variations in AETRs reflect the differential treatment of sectors in the 2001 Income Tax Ordinance. For example, the relatively high AETR in banking and financial institutions stems from the application of the 10 percent capital gains tax to these firms. The relatively low AETRs on cement firms and on transport firms are attributable to the high depreciation allowances (e.g., the initial allowance) that these firms typically claim; correspondingly, insurance companies face a relatively high AETR because the magnitude of depreciation allowances for these firms is relatively low. Sectors that are machinery-intensive also have relatively high AETRs because of the differential application of high customs duties on machinery. Comparisons of Pakistan's AETR with European countries suggest that the calculated AETR are relatively high. 5.12. Similarly, there is substantial variation in the METRs. The right panel of Figure 5.5 shows calculations of METRs for the manufacturing sector under various assumptions about the type of investor. The main tax factors contributing to intersectoral tax distortions include the level of various tax rates and the variance in tax rates and tax allowances across sectors. METRs tend to increase in a fairly linear way with the statutory corporate tax rate. Many of the same factors that generated large differences in AETRs across sectors are also present here. For example, the high METR in banking and financial institutions is due largely to the 10 percent capital gains tax that is assumed to apply to these firms. Also, the relatively low METR on petroleum (underground) firms arises because of the use of 100 percent expensing of capital purchases in this sector, and the relatively low METR on petroleum (offshore) firms is due to the partial expensing of machinery purchases. Similarly, firms in the power generation sector face a lower METR because these firms are assumed to be exempt from several taxes (e.g., the tax on profits) that apply to most other sectors. Sectors that are especially machinery-intensive tend to have higher METRs because of the differential application of high customs duties on machinery; this tendency may be partially offset for sectors with large purchases of capital (e.g., fertilizer, iron and steel, transport, power generation), thereby generating lower METRs, in part because of the high initial allowance that these firms are allowed. Similarly, sectors that are especially inventory-intensive tend also to have higher METRs because these sectors are differentially and negatively affected by the general sales tax. Similar calculation of METRs by asset type clearly shows that assets are treated quite differently by the various features of the tax system. International comparisons suggest that the METRs in Pakistan are often higher than in several other countries. 73 Box 5.2: Pakistan's Tax Incentives Firms operating in export processing zone receive a five year tax holiday and a 75 percent tax exemption after the holiday expires; firms operating in special industrial zone enjoy an unlimited income tax rebate of 75 percent. In addition, any income that is chargeable under the head "capital gains" derived by a person from an industrial undertaking set up in an export processing zone is exempt from payment of income tax. Firms in export processing zones are also exempt from customs duties and sales taxes under certain conditions. Certain types of income receive favorable tax treatment via exemption from the income tax. Several types of preferential tax rates are given. The notion of "group taxation" was recently introduced and provides direct tax incentives. There is also a provision of the law to provide an exemption for the permanent establishment of non-resident petroleum exploration and production companies. The law provides an exemption from the withholding tax on any payment received by an oil distribution company or an oil refinery for supply of its petroleum products. In order to bring at par the non-resident petroleum exploration and production companies, the permanent establishments of the non-resident companies have also been exempted from withholding tax on the supply of crude oil and gas. There are some special provisions that work through indirect taxes. There is zero rating of the sales tax for a number of goods and industries. There are also many exemptions in the general sales tax and incentives in custom duties. Figure 5.5: 2006 Average and Marginal Effective Tax Rates by Industry 2006 Average Effective Tax Rates by Industry 2006 Marginal Effective Tax Rates by Sector and Country of Origin of Investor 50 50 60 45 45 55 50 USA 40 40 With WHT 45 Pakistan 35 35 Canada Without WHT 40 30 30 35 25 25 30 25 20 20 20 15 15 15 10 10 10 5 5 5 0 0 0 Transport Cigarettes Gas Power Generation Fertilizer Hotels & Restaurants Textile Beverages Petroleum (Undergrd) Sugar Cement Automobiles Telecomm Insurance Iron and Steel Petroleum (Offshore) Chemicals Banking & Finance Textile Tobacco OVERALL Automobile Transport Telecom Chemicals Insurance Petroleum Fertilizer Other Banking & Finance Hotels & Restaurants Beverages Cement Sugar Pharmaceuticals Iron and Steel Gas 2006 Marginal Effective Tax Rates by Asset Type for Investors from Pakistan 2001 Statutory, Average and Marginal Effective Tax Rates in Europe 80 100 156 70 90 Statutory Corporate AETR Tax Rate METR Building Inventory 80 60 Machinery Land 70 50 60 40 50 30 40 20 30 10 20 0 10 Textile Automobiles Insurance Chemicals Transport Petroleum (Offshore) Gas Power Generation Cigarettes Banking & Finance Telecommunications Beverages Fertilizer Hotels & Restaurants Cement Sugar Iron and Steel 0 Denmark Australia Netherlands Italy Luxembourg Portugal Greece Belgium Germany Sweden Ireland France Spain UK Finland 74 Low Rates or High Incentives 5.13. The high level and large sectoral variations in AETRs and METRs carry economic and fiscal costs. A principle of good tax policy holds that the tax system should raise revenues with minimal interference in the decisions of firms (Zee, Stotsky and Ley 2002). When a business changes its investments solely because of the existence of a tax, this tax imposes an excess burden. Together with the extensive system of tax incentives and exemptions, the preferential tax rate for small businesses, and the use of the presumptive tax regime for some companies, the corporate tax gives preferential treatment to different types of investment and to different sectors, thereby leading firms to base their investment decisions more heavily on tax considerations rather than on market forces. They translate into large variations in AETRs and in METRs by sector and asset, which bias investment decisions. The variations would be even larger if differences in compliance by firms would be taken into account. Widespread tax evasion means that many firms face an effective tax rate of zero. Tax incentives also carry a fiscal price tag in that they reduce tax collection and add to administrative costs. They open the door for political favoritism of foreign and large corporations, as smaller firms lack the resources and influence to get special treatment. Yet, proponents of tax incentive argue many countries feel they have little choice but to grant incentives in order to compete with rival countries, especially when the country has little in the way of market size or resource endowment to attract foreign investors. They also claimed that tax incentive also would ultimately pay for themselves through higher capital accumulation and growth. The potential benefits would include increases in investment, gains from industrialization, the creation of well-paying jobs, the transfer of technology, and increases in revenues from the growth in the tax base. 5.14. While tax incentives can in particular circumstances stimulate investment, a country's overall investment climate is more important for the success or failure of industries. Since Pakistan has not undertaken a rigorous benefit-cost analysis of tax incentives, it is useful to refer to international experience. The empirical literature has thrown up a variety of findings depending on country sample, time period, data and methodology, and this work remains controversial and unresolved. Nevertheless, there is growing evidence that the main effect of tax incentives is on the transfer of income across jurisdictions (via such mechanisms as transfer pricing and financial policies) rather than the transfer of real economic activity. The emerging consensus is that tax incentives have a fairly limited effect in attracting investment because other aspects of the business environment matter more. They include macroeconomic stability, the quality of infrastructure, the skill level of the labor force, location, the size of the domestic market, regulatory environment and the rule of law. 5.15. The best way to encourage investment for a small capital-importing country like Pakistan is to lower the statutory tax rate on corporate income, rather than to offer selective tax incentives (Box 5.3). A small open economy like Pakistan must compete with investment opportunities in other countries, so it must offer the foreign investor the going after-tax rate of return. The narrowing of the corporate tax bases, whether through the influence of lobbying groups, tax competition or tax benchmarking, suggests there is sizable revenue potential from broadening the tax base which could provide room for further reductions in the statutory tax rate in revenue-neutral fashion. Box 5.3: Lessons from International Experience with Corporate Income Taxation Some general lessons from the recent tax literature related to the corporate income tax and tax incentives are as follows: · Avoid the use of tax incentives for investment. Incentives have little impact on real investment, and seem mainly to reduce tax revenues. Even where they do attract investment, the investment is not often socially productive. · Resist the temptation to promote industrial policies through the tax system. Such attempts are likely to result in a proliferation of tax incentives to enterprises and in the preferential treatment of particular groups of taxpayers, which will lead to other groups requesting them. Also, it is difficult to get rid of tax incentives once they are in place. · Keep enterprise tax rates in line with those of neighboring countries, and with those of the capital-exporting countries: neither higher nor lower. Higher tax rates reduce the capital stock, and lower tax rates lead to strategic behavior by other countries, which simply may reduce tax revenues in all countries. Recent experience shows that it is difficult, if not impossible, to prevent the movement of capital, either into or out of a country. 75 5.4 Withholding Taxes Three Types 5.16. Tax administrations around the world resort to withholding taxes to collect revenues from informal and often non-compliant taxpayers. Since the early 1980s, Pakistan has developed an extensive array of withholding taxes for specific taxpayers in which gross receipts (and not income) are taxed at a specified rate. Such taxes are used for a variety of reasons. They include reducing the compliance costs on taxpayers, especially for small-scale entrepreneurs; simplifying tax administration by removing taxpayers from the tax rolls and by providing direct measures of tax liabilities; and reducing corruption by eliminating official discretion in assessing tax liabilities. 5.17. A withholding tax provides for the payment of tax before, or even without, the filing of any tax returns. Withholding agents are required to deduct such tax payments at source. In Pakistan, they include government, companies other than small companies, individuals with turnover of Rs. 25 million or more, association of persons with turnover of Rs. 50 million or more, foreign contractors or consultants, importers and exporters. The number of withholding taxes increased from 6 in 1979/80 to 12 in 1989/90 and to 24 in 2000/01, and declined since then to 18 in 2007/08 (Table 5.2). Given the extensive array of withholding taxes, it is not surprising that this method of tax collection contributes more than half of all income tax revenue (Figure 5.6, left panel). The withholding tax on construction contracts alone accounts for about 37 percent of all withholding tax collections, or almost 20 percent of gross federal income tax, even though construction contributes only 2.5 percent to domestic value added. 5.18. Pakistan's withholding tax regime can be divided into three regimes: income tax is withheld at source; final presumptive withholding taxes in excise mode, and adjustable withholding taxes in excise mode. While the first type of withholding tax is based on income, the second and third types are based on something else than income, such as gross receipts. First, a number of conventional taxes withhold income tax at source. They include withholding on wages and salaries (at variable rates), bank interest (10 percent), dividends (10 percent), gross rents (variable rates), purchases and sales of listed shares (0.01 percent), royalties and management fees paid abroad (15 percent), and prizes and lottery winnings (10 percent). The withholding on interest and rents, but not dividends, are final taxes. Such withholding simplifies tax administration and reduces the number of taxpayers required to file tax returns. Second, other withholding taxes are in the nature of excise taxes and presumptive in that they constitute final tax liabilities. Examples include withholding taxes on construction contracts (6 percent rate), imports (2 percent), exports (1 percent), income of petrol station operators (10 percent); brokerage fees; and the purchase of goods and services by a withholding agent other than a company (1.5 percent for the sale of rice, cotton seed, or edible oils, 3.5 percent for other goods, 2 percent on transport services and 6 percent on other services). Presumptive taxation can be a fairly efficient and equitable tool to increase compliance among the `hard-to-tax', such as the self-employed, professionals, and small businesses. Third, another category of withholding taxes are advanced income tax payments in the excise tax mode. Such adjustable payments include the purchase of goods and services of a company, withholding against cash withdrawals from commercial banks (0.3 percent of daily withdrawals in excess of Rs. 25,000), car purchases (variable rates), stock market transactions (0.01% on the purchase, sale, and trading), electricity bills (variable rates) and telephone bills (10 percent). 76 Figure 5.6: Withholding Tax Collection Pakistan's Gross Tax Collection of Income Tax Pakistan's Gross Tax Collection from Withholding Taxes Cash Withdrawal 100% 100% Electricity Collection on Demand Export 90% 90% Interest/Securities 80% 80% Telephone 70% Voluntary Payments 70% Salary 60% 60% Other WHT 50% 50% 40% 40% Imports 30% 30% Withholding Taxes 20% 20% Contracts 10% 10% 0% 0% 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 Weaknesses 5.19. While withholding taxes have long become a regular fixture in Pakistan's tax system, they have a number of drawbacks. Many of them mirror difficulties with such simplified tax systems in other countries (Box 5.4). Even though withholding taxes are formally part of the direct taxes (Box 5.5), they have become in many instances effectively indirect taxes. The crucial issue is to what extent withholding taxes are advance tax payments against which credit can be taken in the calculation and submission of a final tax return (Table 5.2). The final withholding on rent, interest and prizes gives Pakistan's global income tax elements of a scheduler income tax in the spirit of the dual tax system, where capital income is taxed at a lower rate than labor income. In addition, some payments are adjustable in principle, but it is far from clear that taxpayers would typically be claiming the resulting tax credits. This, for example, includes salaried individuals who pay withholding taxes on their telephone and electricity bills. An important advantage of creditable withholding taxes is that they provide an incentive for businesses to become registered taxpayers. They impose no additional charge on taxpayers in the formal sector, so long as the amount withhold is fully creditable and, if it exceeds final liability, refundable. At the same time, they do impose an additional charge on taxpayers in the informal sector. 5.20. Second, because the methods for determining tax liability differ from one regime to another, it is possible for nimble taxpayers to take advantage of these differences in the tax rules. One widely used trick is to resort to over-invoicing in order to shift taxable income towards more lightly taxed regimes. For example, taxes on transportation services are a fixed amount per passenger seat or per unit of vehicle weight. If a company supplying these services to a corporation over-invoices the value of these services it will not raise its own taxes but will save taxes for its customer. Similarly, a low-rate presumptive tax on a firm on gross sales also encourages over-invoicing by that firm in its sales to corporations under the regular tax regime. Finally, as withholding on interest is final even for taxpayers with business income, it creates another strong incentive for tax-avoiding arbitrage transactions (Zee 2005). Two taxpayers can provide each other with identical loans in order to eliminate their income tax liabilities through the deductibility of interest payments. Such tax evasion is very difficult to detect, especially if the back-to-back loan used to wipe out any income tax obligation is arranged through a third party. 5.21. Third, extensive exceptions make the withholding tax regime complex. To give just one example, importers have to withhold at a rate of two percent. Yet, there is a reduced rate of one percent for importers purchasing business inputs for their own use, fertilizer for the manufacture of fertilizer, motor vehicles for the production of motor vehicles; as well as for import houses under certain conditions. 77 Box 5.4: Problems of Simplified Tax Systems Incentives. First, if the tax liability under the simplified regime is less than under the regular tax regime, then businesses have an incentive to enter the simplified regime to reduce the tax burden rather than just the compliance cost. Second, firms at the threshold have an incentive to fragment their operations in order to meet the requirements, as a marginal increase of turnover across the threshold generates an enormous increase in tax liability as taxpayers enter into the regular tax system. This, if taxes to be paid under the simplified regime are higher than under the regular regime, then the incentives are reversed in the right direction. Thresholds. Most simplified regimes have thresholds based upon turnover and/or upon employee size, and only entities that do not exceed these thresholds are eligible for the simplified regime. The establishment of appropriate levels for these thresholds is crucial. A threshold that is set too high will undermine the regular tax system; a threshold that is set too low will fail to achieve the goal of simplification for many small taxpayers. In Pakistan, the threshold has in fact been set high, and so that the simplified regime is likely to include viable and ongoing enterprises that are fully capable of being taxed under the regular tax regime. A related issue is that there is no apparent mechanism by which taxpayers "graduate" to the regular tax system, except when they exceed the thresholds. In many countries, there is a limit on the number of years that a taxpayer can participate. However, there are no such graduation provisions in Pakistan. Finally, any threshold must be revised periodically in light of changed economic circumstances (e.g., inflation, economic growth). Legal Eligibility and its Enforcement. There are significant financial incentives for taxpayers to elect the simplified regime rather than the regular tax system, even if they do not meet the legal qualifications. However, in Pakistan, there appears to be little verification that those taxpayers who elect to be taxed under the simplified regime are in fact legally eligible for such participation. This suggests that many participants have illegally moved to, or remained in, the lower cost tax system. Fairness. The simplified regime contributes to horizontal and vertical inequities in the tax system. Two retail firms with equal turnover are taxed very differently if one is subject to the regular tax system and the other is in the simplified regime. Reforms. Ensure that the tax burdens under the simplified regular tax regimes are comparable; reevaluate turnover thresholds, so as to ensure that only "small" taxpayers are eligible; and adjust the thresholds for changing economic circumstance; and improve enforcement of the simplified regime to ensure that only eligible firms participate. Box 5.5: Attributing Withholding Tax Collection How does FBR attribute withholding tax collections to the different income taxes, namely to salaried individuals, non-salaried individuals, and corporations? Based on discussions with FBR staff at several district offices, it appears that the withholding tax on wages and salaries is attributed to the individual income tax on salaried individuals, while collections from other withholding taxes are attributed to the individual income tax on non-salaried individuals. However, the sum of collections from the remaining withholding taxes is greater than the total income tax on non-salaried individuals, which suggest the actual practice is different or the explanation incomplete. 78 Table 5.2: Summary of Withholding Tax Regime Section Title of Section Rate Mode 148 Imports · Collected by Collector of Customs for Value of Goods imported 2% Adjustable · Collected by Collector of Customs for value of goods specified 1% Adjustable · On import of Urea fertilizer a nd Pulses 1% Adjustable 149 Salary Slab Rates Adjustable 150 Dividends paid by a Resident Company 10 % Adjustable Dividends paid by a Non-Resident Company in Mining operations 7.5 % Adjustable Dividends distributed by purchase of power project privatized by WAPDA 7.5 % Adjustable Dividends distributed on shares of power genera tion company 7.5 % Adjustable 151 Profit on Debt u/s 150(1)(a)-on deposit/certificate under National Savings or Post Office 10 % Final u/s 150(1)(b)- Paid by baking company/financial institution on account/deposit 10 % Final u/s 150(1)(c)-Profit paid on a security issued by Federal/Prov./Local Gov. 10 % Adjustable u/s 150(1)(d)-Profit paid on bond etc. issued by bank or financia l institution 10 % Final 152 Payments to Non-Residents 15% Adjustable 153 Payment of Goods and Ser vice s (a) Sales of Goods - Sale of rice, cotton seed or edible oil 1.5 % Final Sale of other goods 3.5 % Final (b) Rendering of Services -Transport Services 2% Final Other Services 6% Final Sub-section (1A)-Payment by an exporter to a non-resident 0.5 % Final For Companies 2%&6% Adjustable (c) Execution of Contrac ts 6% Final 154 Exports Sub-section (1)-On realization of foreign exchange proceeds on Export of goods 1% Final Sub-section (3)- On proceeds on inland back-to-back LC/other arrangement 1% Final Sub-section (3A)-B y EPZ at the time of export of goods 1% Final Sub-section (3B)]-Deducted by a direct exporter on payment to indirec t exporter 1% Final [Sub-section (2)]-On realization of foreign exchange proceeds on Sale of Goods 5% Final 155 Income f rom Property Slab Rates Final 156 Prizes and winnings (Prize Bonds) 10 % Final Prize and winnings (raffle, lottery, prize, cross-word puzzle) 20 % Final 156A Petroleum Products 10 % Final 156B Withdrawal of balance under Pension Fund Specified Adjustable 231A Cash Withdrawal from a bank 0.3 % Adjustable 231B Purchase of Motor Cars Slab Rates Adjustable 233 Brokerage and Commission 10 % Final Brokerage and Commission ­ Advertising Agents 5% Final 233A Collection of Tax by a Stock Exchange Registered in Pakistan Under Sub-section 1(a)-On purchase of shares 0.01 % Minimum Under Sub-section 1(b)-On Sale of shares 0.01 % Minimum Under Sub-section 1(c)-On Trading of shares 0.01 % Minimum Under Sub-section 1(d)-On financing of carryover trades (Badla) in 10 % Adjustable shares 234 Transport Business Slab Rates Final 234A CNG Stations 4% Final 235 Electricity Consumption Slab Rates Minimum/ Adjustable 236 Telephone users (Bill Subscriber) 10 % Adjustable Telephone users (Prepaid Cards) 10 % Adjustable Source: Adapted and updated from FBR Quarterly Review, September to December 2007. 79 5.5 Small Companies 5.22. Small companies benefit from a reduced corporate income tax rates. They are defined as companies with annual turnover of less than Rs. 250 million (around $3.5 million) and no more than 250 employees at any time during the tax year. In addition, the paid-up capital plus undistributed reserves must not exceed Rs. 25 million and it must not be formed as a result of fragmentation of an existing business. The impact of these recently introduced provisions on the corporate tax system appears to be minor so far, as only few companies are registered as such small companies. For example, small companies contributed less than 1 percent of gross corporate income tax revenues in 2006/07. Nevertheless, this sector is likely to become more important over time. The tax treatment of small businesses brings up three policies issues. 5.23. First, the definition of small businesses is extensive. A small business, as defined by Pakistan's tax law, is not "small" when compared with the legal definition of smallness in other countries. Elsewhere, "small" almost never describes firms with more than $1 million in turnover and more than fifty employees. 5.24. Second, small businesses are exempt from any withholding requirements. The argument in favor of this exemption is that it encourages the corporatization of small businesses operated by individuals and partnerships that are currently outside the grasp of all regulatory agencies. Since individuals and associations of persons with less than Rs. 25 million and Rs. 50 million in turnover, respectively, are not required to withhold any tax on their payments of goods and services or execution of contracts, this exemption privilege should be extended to small business as well in order to provide a level playing field. However, the logical conclusion of this argument is a race to the bottom and no withholding anywhere. This exemption potentially threatens the integrity of the withholding system and the current regime of presumptive taxation. Large and small businesses that would be subject to withholding on their sale of output to other large businesses can escape the withholding web by structuring their sales to small businesses. It also implies that small businesses are not required to withhold income taxes from the wage and salary incomes paid to their employees, who are hence likely to escape individual income taxation. If the policy goal is to entice small businesses out of the informal economy, there are better instruments for achieving this objective than a withholding exemption. For example, the Small Business Administration in the US caters to the borrowing needs of small businesses, as long as these companies provide the same types of documentation that a corporation would be expected to provide. 5.25. Third, until 2007/08, small companies benefited from a lower statutory rate of 20 percent. Providing a concessionary tax rate for small companies appears to represent an obvious tax distortion biasing investment decisions toward small business. However, this preferential treatment has been justified as a means of offsetting another distortion, namely the failure of credit markets to channel adequate financial resources to small business. Potential lenders to small business may overlook them because their business prospects are more difficult for lenders to evaluate. It is on these grounds that many countries, including Canada and the USA, offer concessionary tax rates to small business. Nevertheless, Pakistan stood out until 2007/08 by taxing the entire business income at the 35 percent rate once the threshold is exceeded. The nature of this problem can be depicted by a simple numerical example. Consider a small company with a turnover just below the threshold of Rs. 240 million and a profit margin of 10 percent on that turnover. Under 2006/07 legislation, its profits of Rs. 24 million are taxed at 20 percent yielding Rs. 4.8 million in revenue. Assuming this company increases its turnover by Rs. 20 million, and now exceeds the threshold by Rs. 10 million. It would be taxed on profits of Rs. 26 million at a rate of 35 percent and face a new tax bill of 9.1 million. Some Rs. 2 million of additional profit would encounter an additional tax of Rs. 4.3 million! Clearly, it is far better to remain a small business than to grow into a larger one. The decision to remain small could impose avoidable economic costs on the economy, as it may prevent firms from exploiting economies of scale and scope. To avoid this "notch" problem, Pakistan changed in 2007/08 the tax system for small companies by introducing rates of 25 percent for taxable income from Rs. 250 million to Rs. 350 million, 30 percent from Rs. 350 million to Rs. 500 million, and 35 percent from Rs. 500 million onwards. 80 5.6 International Tax Provisions Protecting Domestic Tax Base 5.26. Pakistan's system of international tax provisions is to a large extent in line with international practice. The Income Tax Ordinance 2001 (Ordinance) contains a number of provisions dealing with international issues of the domestic tax system. They lay the basis for the protection of the domestic tax base. In addition, Pakistan has concluded over 50 double tax agreements which complement these domestic rules. 5.27. The Ordinance contains necessary provisions to safeguard the domestic tax system in the world of international business, although some of these concepts have become more difficult to apply. In particular, the Ordinance includes provisions that protect the Pakistan tax base from erosion through practices such as transfer pricing and thin capitalization. It also contains domestic withholding taxes on payments abroad, such as management fees and interest payments. Furthermore, key-provisions, like residency, geographical source, and international profit allocation, are in place and well structured. Such principles were initially developed internationally in response to the first industrialization wave of the late 1900s, and are still embedded in today's OECD or UN model conventions. While these concepts have been adopted over the years, they have become increasingly flawed with economic globalization, deregulation and technological advances. With the help of new technologies, multinational operating taxpayers are increasingly able to arbitrary decide whether they want to establish residence or source income in a certain jurisdiction. Furthermore, international tax rules have relied heavily on the arm's length principle for profit allocation. Yet, this method establishes a heavy burden on both tax administration and taxpayers as the arm's length or uncontrolled price is often difficult to establish. This is especially the case when there are economic reasons for performing certain activities internally within a multinational company rather than through outsourcing. Some of these points are discussed in greater detail below. 5.28. Pakistan has entered into 54 double tax agreements with other countries. Such agreements provide certainty to international investors, specify clear income allocation rules, spell out the method of double tax relief, and contain clear rules on the taxation of investment income (Table 5.3). Following common practice among developing countries, Pakistan uses the UN model convention. In addition, Pakistan has entered into a multilateral agreement with Bangladesh, Bhutan, India, the Maldives, Nepal, and Sri Lanka ­ South Asian Association for Regional Cooperation (SAARC) ­ which contains provisions on the exchange of information and assistance in the area of tax administration and enforcement. In accordance with international practice, the double tax agreement provisions prevail over domestic tax provisions in so far as they provide for double tax relief, the determination of Pakistan source income of nonresidents, the allocation of income to Pakistan sources, the determination of income, and the exchange of information. Table 5.3: Pakistan's Double Taxation Treaties Austria Iran Norway Syria Azerbaijan Ireland Oman Tajikistan Bangladesh Italy Philippines Thailand Belgium Japan Poland Tunisia Bosnia Herzegovina Kazakhstan Portugal Turkey Canada Korea Qatar Turkmenistan China Kuwait Romania United Arab Emirates Denmark Libya Singapore United Kingdom Egypt Malaysia South Africa United States Finland Malta Sri Lanka Uzbekistan France Mauritius Sweden Yemen Germany Nepal Switzerland Vietnam Hungary The Netherlands Saudi Arabia Indonesia Nigeria SAARC 81 Weaknesses 5.29. The Ordinance lacks implementation rules regarding the establishment of residency. The Ordinance defines under what circumstances a person is treated as a resident in Pakistan. These circumstances are related to the number of days a person is present in Pakistan during the tax year. However, the 2002 Income Tax Rules­ a set of regulatory implementation rules attached to the 2001 Income Tax Ordinance ­ do not make sufficiently clear how to count the days individuals are present in Pakistan. A similar issue is present in the definition of resident companies. The Ordinance defines that in general all companies established under Pakistani laws are resident companies. Any other company will be treated as a resident company if control and management of the affairs is situated in Pakistan. However, there is no guidance regarding the term "central management and control". 5.30. Geographical source rules are important as they form the link between a geographical area and income, but they are difficult to implement for international finance, e-commerce and other businesses. The Ordinance lays out source rules that link income to Pakistan, such as the availability of a permanent establishment (business income), the place where the activities which generated the income take place (employment income), and the place where the assets are located out of which the income is subtracted (dividend, royalty, technical service fee). Some income can be sourced relatively easy as the instrument through which it is generated has a "tangible" character, such as business assets like machinery. Other income may be more difficult to allocate due to their "intangible" character, including financial activities and services. For example, the liberalization of capital markets makes it easier to shift the geographical source of income generated through financial instruments, and global trading activities typically take place in more than one of the three main time zones and involves customers in more than one jurisdiction. Likewise, concepts such permanent establishment do not adequately link profit to a geographical area in the case of e-commerce, such as on-line trading of goods, services or financial instruments, electronic fund transfers, and electronic data exchanges between companies and within a company. Since geographical considerations are irrelevant for e-commerce, source rules become ineffective for income allocation. Servers can easily be located anywhere and their location is generally unknown and unimportant in a business transaction. Moreover, tax administrations have difficulties in tracing where transactions take place. 5.31. The Income Tax Rules do not provide any provisions to guide the tax authorities on turnkey projects. In such projects, separate entities are responsible for setting up and operating a plant or equipment. While the taxpayer often contends that the contract is divisible and that there can be no tax liability in Pakistan for the offshore supply of equipment and services, the tax authorities typically argue that the contract is composite and integrated and that the nonresident taxpayer is liable to pay tax in Pakistan for the offshore supplies as well. 5.32. Implementing the "transfer pricing" principle is also fraught with practical difficulties. The Ordinance allows the tax commissioner to make profit corrections in situations where associates have not dealt with each other at "arm's length". The arm's length method refers to an independent market price: the comparable uncontrolled price. This principle is also laid down in the OECD- and UN model conventions and in Pakistan's double tax agreements. However, the comparable uncontrolled price is rarely available and certainly arbitrary in cases of intangible assets and service fees. But alternative methods also have drawbacks, as information on branch- specific gross margins or mark-ups are typically not available to the tax administration, and intercompany transactions deserve special attention to prevent tax avoidance. Hence, the transfer pricing concept is important but should be applied only as last resort to prevent misuse and uncertainty for foreign investors. 5.33. Pakistan's double taxation agreements have similar weaknesses as the domestic provisions. For example, they contain rules that allocate taxing rights to the country of residence and/or the country of source. In general, the UN model convention follows the domestic definition as long as this definition is based on domicile, residence, place of incorporation, place of management or any other criterion of a similar nature. In case an individual is resident of both contracting countries, Pakistan has adopted the concept of "effective place of management" as the tie-breaker provision in most double taxation agreements, which is slightly different from its domestic concept of "central place of management and control". However, the agreements are not accompanied by detailed implementation rules to guide the interpretation of terms like "permanent home", "center of vital interests", "habitual abode", and "effective place of management". Similarly, they do not specify whether reduced withholding tax rates should be implemented at source through a statement of residence ex ante or through a refund procedure ex post. 82 5.7 Equity 5.34. The corporate income tax is progressive. In order to see how the 2006/07 corporate income tax payments are distributed across household deciles, we have to make assumptions about its incidence. Initially, we assume that the tax burden of corporate income tax is shared equally by wage earners and capital holders; and the capital component is distributed in line with income from property and asset sales, as captured in the 2004/05 Pakistan Living Standard Measurement Survey. As we use actual tax receipts, the tax incidence reflects non- compliance. Since there is no information on whether non-compliance increases or decreases as we move up the income distribution, we assume the level of non-compliance remains constant across household deciles. The effective tax rate as percent of GDP increases moderately from 1.1 percent for the bottom decile to 1.4 percent for the 9th decile. It jumps to 3.9 percent for the top decile (Figure 5.7, left panel). This spike is linked to the capital portion, where 93 percent of the tax is borne in the highest decile, while the labor portion is distributed fairly proportional to income. Almost four-fifth of the corporate tax collection comes from the top decile. 5.35. A sensitivity analysis confirms the progressive nature of corporate income tax. Given the impact of the lopsided distribution of capital on the derive tax incidence, it is worthwhile to verify the findings under an alternative assumption. Pakistan's 2001/02 Household Income and Expenditure Survey contains more detailed information on the distribution of assets and capital than the 2004/05 Pakistan Living Standard Measurement Survey. Using the 2001/02 distribution of capital holdings, the corporate income tax becomes somewhat less progressive. For the top decile, the effective tax rate decreases to 3.2 percent of GDP and to 48 percent as a share of tax collection (Figure5.7, right panel). Nevertheless, the overall progressive pattern remains unchanged. Figure 5.7: Vertical Incidence of Corporate Income Tax in 2006/07 Incidence of Corporate Income Tax by Household Consumption Deciles Incidence of Corporate Income Tax under Alternative Capital Income 60 4 Distribution 60 4 55 3.6 55 Effective Tax Rate 3.6 50 (% of GDP) 3.2 50 3.2 45 45 2.8 40 2.8 40 2.4 35 2.4 35 30 2 30 2 Effective Tax Rate 25 (% of GDP) 25 1.6 1.6 20 20 1.2 1.2 15 15 0.8 0.8 10 10 Share in Total Taxes 0.4 Share in Total Taxes 0.4 5 5 (%) (%) 0 0 0 0 Bottom 2 3 4 5 6 7 8 9 Highest Bottom 2 3 4 5 6 7 8 9 Highest 5.36. Beyond the incidence of the corporate income tax across income groups, the tax structure introduces significant vertical and horizontal inequities. As we have seen, the corporate income tax discriminates among firms, largely because of the existence of tax preferences that are available to some firms and sectors and not to others. In addition to the formal provisions for tax relief, there is discretionary relief on a case-by-case basis. Similarly, the opportunities for tax evasion differ considerably between firms in the formal and the informal sectors, between "large" and "small" firms, and between domestic and multinational firms (due for example to differences in transfer pricing capabilities). This weakens confidence in the system, and may encourage some companies to evade and avoid taxes. It also leads to economic inefficiencies, as companies make different economic choices in order to capture tax advantages. 83 5.8 Compliance Tax Gap 5.37. The potential revenue from corporate income tax is calculated on the basis of an input-output model with 81 sectors of Pakistan's economy. This model provides information on the total use or gross sales for each sector; on business expenses that are deductible from corporate taxable income, specifically the costs of inputs, compensation of employees, and indirect taxes; and on depreciation allowances based on fixed capital expenditures and statutory depreciation rules. Not all of the 81 sectors are subject to the corporate income tax because of a variety of reasons. For example, some industries might be dominated by small businesses, which could include handicrafts, textiles, and light manufacturing. Others, such as electricity, might effectively be exempted from corporate income tax, even though the enterprises in this sector typically assume the corporate form. Yet other sectors are dominated by public institutions, namely the central monetary authority, public administration and defense, and education and healthcare. Agriculture is not subject to the federal income tax. In order to err on the side of caution, we have excluded the estimated profits from all these sectors. Once we eliminate these sectors, we are left with 24 sectors, which in our view are dominated by large enterprises subject to corporate income tax. They cover eight major economic sectors: mining, chemicals, cement, metal products, equipment manufacturers, construction, transportation and communications, and financial services. We then apply a tax calculator for the 2004/05 corporate income tax system on the estimates of taxable income for each of these sectors, and compare the obtained potential tax liabilities with actual tax payments. 5.38. Even though the methodology is likely to underestimate actual tax liabilities, the estimated tax gap is very large. We estimate the tax gap to be about 218 percent of the actual corporate income tax payments in 2004/05. Translated into 2007/08 numbers, this would imply a tax gap of over Rs. 450 billion, or about 45 percent of actual federal tax collection. Clearly, the size of the tax gap is worrisome. It represents a conservative estimate because we have eliminated sectors, such as the textile sector, that includes corporatized firms, and the estimate of depreciation allowances as a percent of gross receipts appears to be rather high. There are also other signs of low compliance with the corporate income tax. Nearly 30 percent of actual corporate income tax receipts come from withholding on construction contracts, and withholding tax on imports and exports accounts for an additional 60 percent of actual corporate tax receipts. As with the individual income on non-salaried individuals, it appears that the withholding tax regime is masking serious problems with non-compliance in the corporate sector. Figure 5.8: Tax Gap of Corporate Income Tax in 2004/05 Actual Tax Collection and Tax Gap of Corporate Income Tax Gap (218%) Actual (100%) 84 Voluntary Payments 5.39. The trends in voluntary payments suggest there is more room to improve tax compliance. With the introduction of the universal self-assessment system as part of the 2001 Income Tax Ordinance, it was expected that voluntary compliance would play an increasing role in tax collection (Box 5.6). Yet, the share of voluntary payments in gross tax collections, after rising from 33 percent in 2002/03 to 36 percent in 2003/04, remained constant until 2005/06, while the share of withholding tax collection increased at the expense of collection on demand (Figure 5.9). At the same time, the share of companies with national tax numbers filing income tax returns declined from 69 percent in 2004 to 54 percent in 2006, as the growth in national tax numbers exceeds the growth in income return filers. Voluntary payments increased sharply in 2006/07, but these gains could only partly be sustained in 2007/08 as the Rupee amount of payments with returns declined (Box 5.7), even though the number of return filers increased by another 21 percent. 5.40. With the number of return filers rising and tax payments declining, there are concerns about the accuracy of tax returns. In 2006, FBR had 23,100 registered companies, compared to around 47,500 registered at the Securities and Exchange Commission of Pakistan. While some of the discrepancy could reflect outdated data, it does suggest that a large number of active companies do not have a national taxpayer number. Among FBR's cases, slightly less than half filed a tax return. And among those who filed, two-third declared either a negative or zero taxable income, and one-third positive taxable income; only one-in-eight companies declared a taxable income in excess of Rs. 1 million (Figure 5.10). The record for foreign companies is even more dismal: only 17 of 494 foreign companies filed a return in 2006. Judging by these tax returns, one would imagine that 2006 was a difficult year for the economy. Yet, it was the fourth year of a solid recovery, with strong growth, modest inflation, rising household incomes and falling poverty. The 2006 picture does not seem to be an exception. A FBR analysis of 2005 tax returns found a similar pattern. This study also revealed that almost three-fifth of all tax returns lacked basic information such as the national identification card number, name, address and telephone number of company. Fortunately, the recent drive to pay off refunds improves incentives for tax return filing (Figure 5.10). Figure 5.9: Voluntary Payments and Corporate Income Tax Filing Pakistan's Gross Tax Collection of Income Tax Corporate Income Tax Filing 100% 30,000 70 Compliance (%) Collection on Demand 90% 60 25,000 80% Income Tax Filers (#) 50 70% Voluntary Payments National Tax Number 20,000 (#) 60% 40 50% 15,000 30 40% 10,000 30% 20 Withholding Taxes 20% 5,000 10 10% 0 0 0% 2004/05 2005/06 2006/07 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 Box 5.6: Three Types of Income Tax Collection Income tax collection comes in three forms: voluntary payments, either as advance tax or as tax with returns; collection on demand; or withholding taxes. Under the universal self-assessment scheme, all taxpayers automatically qualify for self- assessment. With the pay-as-you-earn concept, the taxpayer determines the expected income for the current year, and derives the amount of tax due on the income earned. This is the basis for a monthly advance payment. Then, once a year and for most companies in December, companies are required to file their annual return for the preceding financial year to discharge their total tax liability. Advance taxes and some taxes withheld at source during the course of the year are adjustable against the final tax liability at the time of filing. This is the part of the voluntary payment related to return filing. If a company has paid more than its final tax liability, it can apply the excess against future tax liability. The returns received are not examined, or assessed, in detail at the time of receipt. Instead, the tax returns are considered as final unless their cases are selected for audit through a risk assessment based on pre-announced parameters. If such an audit leads to an additional tax demand, the department proceeds with "collection on demand". Taxpayers are required to maintain accounts and related documents for ensuring an effective tax audit. Finally, voluntary payments and collection on demand are complemented through withholding taxes, the most important pillar of income tax collection. 85 Figure 5.10: Tax Return Filing and Refund Payment for Corporate Income Tax 2006 Tax Return Filing of 47,500 SECP-Registered Companies Refunds of Income Tax Payments 35 15 Positive Tax Return (10%) 30 Zero Tax Return (13%) 12 % of Gross Income Tax 25 Collection Negative Tax Return (7%) 9 20 Rs. Billion No Tax Return (31%) 15 6 10 3 5 No FBR Registration (39%) 0 0 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 (Jul-Mar) Box 5.7: Making Sense of the Decline in Voluntary Payments in 2007/08 The number of income tax returns increased by 21 percent from January 2007 to January 2008. Yet, the collection of tax payments collapsed by 84 percent from Rs. 43 billion to Rs. 7 billion. What caused the sharp decline in payments with returns? The disappointing collection may be in part related to the extended business closure for the 2007 Eid holidays and the national mourning in the last week of December. But the principal reason is likely to be the change in the advance tax regime. In the past, taxpayers submitted advance tax on the basis of their last year's turnover for 80 percent of their tax liabilities, with the balance remitted along with the filing of the return. In 2007, FBR introduced two changes in the advance tax regime. The tax basis shifted from last year's turnover to the current year's expected profits, the minimum payment increased from 80 percent to the full 100 percent of the estimated tax liability. This left little payment still to be done at the time of the tax filing. In particular, the advance payments from January to June 2007 captured in part tax payments that under the old regime would have come due only in 2007/08. The net effect of the regime change was to inflate advance payments in 2006/07 at the expense of payments with returns in 2007/08. In addition, the strong performance of banking, telecommunication and oil and gas during 2006 fuelled the collections with returns in 2006, while the first signs of an economic slowdown might have reduced these in 2007. Two Components of Voluntary Payments of Corporate Income Tax (Rs. Million) Average Payment Per Corporate Income Tax Return (Rs. Million) 180,000 4.0 160,000 3.5 Advance 140,000 3.0 120,000 With Return 2.5 100,000 2.0 80,000 1.5 60,000 1.0 40,000 0.5 20,000 0 0.0 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 July 2006 to January 2007 July 2007 to January 2008 86 5.9 Policy Options 5.41. The corporate income tax system requires change in policy and administration, as well as the strengthening of international tax provisions (Table 5.4). The main priority should be to broaden tax bases and lower tax rates. Specifically, the statutory tax rate of the corporate income tax could be reduced. This would bring Pakistan more in line with peer countries and help to reduce significantly the distorting effects of the corporate tax. At the same time, expanding the base of the corporate income tax, mainly by reducing or rationalizing the use of tax incentives and exemptions, could ensure that the reduction in the statutory rates does not lead to revenue loss, and bolster efficiency and equity. In this regard, it would be important to undertake a complete examination of the costs and benefits of tax incentives and exemptions. In addition, the thresholds for small businesses could be reexamined in order to ensure that only truly small businesses qualify. 5.42. The second priority of tax policy reform is to rationalize the withholding tax system. The extensive use of withholding taxes increases the effective tax rate on companies; creates opportunities for tax evasion schemes; mask serious compliance issues; and, through the many special stipulations, complicate the tax system. In many cases, they also have become final liabilities. As a result, many taxpayers are caught in the withholding tax net, even though their incomes puts them in the zero-rate bracket under the regular tax schedule. Some withholding taxes resemble often more excise taxes than income taxes, as in the case of withholding on imports, exports and sales of goods and services. As a general rule, whenever they are imposed on the formal economy, withholding taxes should be considered an advance payment of the final tax liability. Withholding taxes that generate trivial amounts of revenues could well be eliminated. Finally, clarifying FBR's rules for attributing withholding taxes to the different income tax heads will improve the understanding of their impact of the system of direct taxation, and hence improve FBR's ability to identify and implement good performance policies. 5.43. The main priority for tax administration is to improve compliance. In July 2002, the universal self assessment scheme for income tax was introduced with the laudable goals of restoring the confidence of taxpayers, minimizing the contact between taxpayers and tax collectors, and augmenting income tax revenues. While good progress was made on the first two objectives, there is still a long way to go to meet the third objective. The tax gap analysis, along with much anecdotal evidence, suggests that companies are able to evade tax liabilities. FBR needs to launch a comprehensive effort to ensure that taxpayers meet their legal obligations ranging from registration, keeping accurate records, and filing timely returns. Stringent penalties need to be imposed in the event of non-filing. Non-filers should be assessed under best judgment presumptions with the proviso that the assessment must be paid in advance of any attempt of appeal. Deception should be deterred through the risk of detection. This is where the new program of field audits for selected corporate cases comes in. With the planning stage completed, the full implementation of audit program can now being rolled out speedily. Audit should initially focus on large corporate entities. Corporate income taxes account for about 70 percent of direct tax collection, even though corporate income tax filers count for less than one percent of all income tax filers. The bulk of the corporate income tax comes from a handful of companies in oil and gas, banking, and telecom ­ some of them with a close link to the public sector. 5.44. The discussion on international tax provisions suggests three areas of reforms. First, there is a need for policy level guidance on new sectors, covering residency, e-commerce, banking and financial activities, turnkey projects, and the establishment of a reliable transfer price, in the 2002 Income Tax Rules. Second, a number of steps could help to limit the use of transfer pricing provisions to exclusively anti-avoidance situations. Staff in large taxpayer units lacks up-to-date information about the developments in doing business in a globalized environment, as well as practical training and sharing of experiences on the operational issues they imply. The number of tax officers dealing with transfer pricing issues should be limited. For example, the large taxpayer units in Karachi and Lahore could set up specialized transfer pricing teams of two to three persons. Other tax officers outside this specialist teams could refer cases that include transfer pricing issues to these teams. At the policy level, a special team within the Federal Board of Revenue could be established to monitor the international developments in the area of transfer pricing. In addition, the option of entering in so-called advance pricing agreements, which provide certainty to multinational enterprises regarding the pricing of their internal transactions, could be explored. Third, double taxation agreements should be provided with more detailed implementation rules. For example, the current agreements provide no guidance on the interpretation of terms like "center of vital interests" and "effective place of management"; or on the application of reduced withholding tax rates. 87 Table 5.4: Policy Reform Options for Corporate Income Tax Topic Proposals · Limit the use of tax incentives for investment and industrial policy Tax Incentives and Tax Rates · Lower the statutory tax rate and expand the tax base to prevent revenue losses · Make withholding taxes adjustable, especially for formal economy · Limit favorable withholding tax rates for informal economy · Eliminate withholding taxes that generate only small revenues Withholding Taxes · Establish clear attribution for withholding tax collection to salaried taxpayers, non- salaried taxpayers, and corporations · Narrow definition for small businesses by reducing turnover threshold in real terms Small Companies · Make small businesses withholding agents · Provide policy level guidance on new sectors, covering residency, e-commerce, banking and financial activities, turnkey projects, and the establishment of a reliable transfer price, International Tax in the 2002 Income Tax Rules. Provisions · Limit the use of transfer pricing provisions to exclusively anti-avoidance situations · Provide more detailed implementation rules for double taxation agreements · Launch comprehensive risk-based compliance strategy. · Roll-out field audits speedily, with an initial focus on large corporations. Compliance · Ensure a viable appeals mechanism 88 CHAPTER 6: INDIVIDUAL INCOME TAX 6.1 Introduction 6.1. The individual income tax is typically not popular among taxpayers, yet it needs to be part and parcel of the tax system. As a direct tax on individuals, rather than an indirect tax on transactions, it relies on the tax administration taking a slice of a hard-earned paycheck of workers and the business income of the self- employed. At the same time, the individual income tax is central to a society's notion of raising revenue in a fair and equitable way. By allowing the assessment of individual economic circumstances, the income tax is naturally better suited for making a tax structure progressive than indirect taxes. The economic literature underlines the crucial role of the individual income tax. But just how progressive should it be? Early research established a surprising result under the assumption that all individuals agree about how much income is worth in terms of well- being (or in economic jargon utility), and the well-being of one individual can be traded off against the utility of another. Then post-tax incomes would have to be equalized across individuals to ensure that the utility loss through income tax of an additional unit of income is the same across individuals (Edgeworth 1897). However, this result does not take into account the distortionary impact of taxation on the generation of income. If disincentive effects increase with the level of tax rates then tax rates on income, at the margin, should remain fairly low (Mirrlees 1971). Nevertheless, even from the point of efficiency, Atkinson and Stiglitz (1976) show that, under fairly standard assumptions, government cannot improve on a progressive tax on labor income using differential consumption taxes. While this result might not have a direct translation into tax policy, as there are important considerations ­ including tax administration ­ that the Atkinson-Stiglitz framework does not consider, it confirms that the individual income tax should be an integral part of the tax system. 6.2. Pakistan's tax yield from individual income tax is inadequately low and declined since the early 2000s. Partly due to a reduction in income tax rates, gross revenues from the individual income tax declined from 1.5 percent of GDP in 2000/01 to 1.1 percent of GDP in 2007/08 (Figure 6.1). The individual income tax accounts for around 11 percent of overall FBR tax revenues and 29 percent of direct tax revenues, which include also the corporate income tax (discussed in Chapter 5), capital value tax (Box 6.1) and social security taxes (Box 6.2). This level of revenue mobilization would seem to put Pakistan somewhere in the middle rank among emerging market economies. For example, apart from India, the other major South Asian economies collect less as percentage of GDP. However, Pakistan's standing might actually be worse than what the numbers suggests for three reasons. First, Pakistan's individual income tax is measured in gross receipts, which include refunds. Second, there are two types of taxpayers liable to individual income tax in Pakistan: workers and salaried individuals; and small unincorporated businesses and associations of persons, including firms with a profit sharing agreement. This may be in part due to a more expansive definition of individual income tax than the ones used in other developing countries. Indeed, FBR refers to the individual income tax as non-corporate income tax. Finally and most importantly, the individual income tax collection includes a share of the funds mobilized through withholding taxes. Some of the withholding taxes are presumptive, imposing a fixed charge on certain transactions. This arrangement turns such withholding taxes effectively into indirect taxes. Since the heavy reliance on withholding taxes ­ which in many instances are indirect taxes ­ is unusual, this moves up Pakistan's rank in country comparisons on individual income tax collections. This chapter will identify reasons for the low revenue yield of the individual income tax. They are linked to weaknesses in tax design and tax compliance (Table 6.1). The main policy conclusion is that there is scope for raising more tax revenue, while at the same time broadening its base, simplifying its design, and preserving its progressivity. Since companies provide the largest share of withholding taxes, they are discussed in Chapter 5. 89 Figure 6.1: Collection of Individual Income Tax Pakistan's Gross Tax Collection of Individual Income Tax Individual Income Tax in Pakistan and Other Emerging Economies 1.6 55 (2006/07 or latest available) % of GDP As % of FBR Gross 9 12,000 Direct Taxes PPP GNI Per Capita 50 1.4 As % of FBR Gross 8 Taxes 45 10,000 1.2 7 40 35 6 8,000 1.0 30 5 0.8 6,000 25 4 % of GDP 0.6 20 3 4,000 15 0.4 2 10 2,000 0.2 5 1 0.0 0 0 0 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 Nepal Egypt Bangladesh Sri Lanka Pakistan India Thailand Turkey South Africa Box 6.1: Capital Value Tax Pakistan used to have a general wealth tax that generated very small amounts of revenue and was abolished in July of 2001. Puny amounts of tax continue to trickle in from this tax as a result of unsuccessful appeals. However, selective wealth taxes continue to exist in the form of capital value taxes applied to the ownership and purchase of cars, stocks and real estate. These taxes are imposed at the time these particular assets are acquired. In 2006/07, these taxes contribute only about Rs. 8 billion, or about one percent of federal tax revenues: · Three capital value taxes of 0.01 percent each are attached to the share transactions at the stock exchanges. · Car purchases are taxed at five percent of their sales value. · A two percent tax is charged on the sales value of immovable urban property larger than 500 square meters. A fixed value of Rs. 50 per square meter applies if no value can be attached to the property or the purchaser has no national tax number. One issue with these taxes is their design. For example, the tax on stock transactions is rather complicated and may not be the best approach to taxing capital income. The alternative of the capital gains tax is discussed later in this chapter. But overall, the crucial policy question is whether this group of assets is taxed too lightly. A number of other developing countries impose heavier taxes on these assets, including Turkey, and generate significant revenue from these sources. Box 6.2: Social Security Taxes Pakistan does not have a national social security system financed primarily by payroll taxes. Instead, it has a patchwork of private and public pension schemes. Public sector employees make compulsory contributions to a pay-as-you-go public pension fund, which are matched by the government. In addition to enterprise pension schemes, private sector employees have two sources of social security support: · The Workers Welfare Fund collects a two percent tax on the after-tax income of companies for the Ministry of Labor. The Ministry uses these funds to provide benefits for low income workers ranging from housing to dowry payments and maternity benefits. · The Workers Profit Participation Fund is funded by a five percent of the after-tax income of large companies and multi- national enterprises and distributes the proceeds primarily for the benefit of low wage workers within the company. Any residual not distributed is transferred to the Workers Welfare Fund. · The cost effectiveness of these two programs for low wage workers in the formal sector have not been systematically evaluated. Anecdotal evidence suggests low utilization of funds, weak project management and overlap with other public agencies. 90 Table 6.1: Main Findings of Chapter 6 Topic Findings Collection Pakistan's tax yield from individual income tax declined from 1.5 percent of GDP in 2000/01 to 1.1 percent of GDP in 2007/08. It accounts for around 11 percent of overall FBR tax revenues and 29 percent of direct tax revenues. Design Taxable income consists of world-wide income of residents derived from five different sources. There is an extensive list of exemptions, although the overall costs of exemptions estimated to amount to less than Rs. 1 billion. The tax applies differential rates for different tax payers and different income sources. Capital Gains Capital gains on the sale of shares listed on any of Pakistan's three stock exchanges are exempt until June 30, 2010. Schedules There are two basic tax schedules with large and different thresholds for men and women, spikes in marginal tax rates for non-salaried taxpayers, and a large number of tax brackets. Separate rates apply for rental income, dividends, interest income. Credits Tax credits are granted up to certain limits, come in four varieties, and work like a form of deduction, since the value of the credit increases with the taxpayer's average tax rate Equity One reason for the progressivity of individual income tax is that income tax rates increase with income levels. But there are other reasons. The taxation of salaried income is progressive primarily due to large zero-rate thresholds. The taxation of non-salaried income is progressive as higher consumption deciles are more likely to earn non-wage income. Compliance The tax gap is 97 percent of actual receipts for salaried taxpayers, and 15 percent of actual receipts for non-salaried taxpayers, including all other sources of individual income. This would translate into a tax gap of Rs. 31 billion, 27 percent of individual tax revenues, and 3 percent of federal tax revenues in 2007/08. 91 6.2 Towards a Global Income Tax 6.3. In principle, the individual income tax is imposed on a comprehensive and global income base, with noticeable exceptions. A basic choice in the design of a individual income tax is whether the aggregate of different source of income should be taxed through a single rate schedule (global income approach) or different sources of income should be taxed through different rate schedule (schedular approach). Invariably, countries' individual income tax systems fall somewhere in-between, and Pakistan is no exception. Taxable income in Pakistan consists of the world-wide income of residents derived from five different sources: wages and salaries, property (rental) income, business income, capital gains, and other sources of income, such as dividends and interest payments. There are three notable deviations from the global income approach. 6.4. First, there is an extensive list of exemptions from the income tax base. The second schedule of the 2001 Income Tax Ordinance includes 70 pages of exemptions and concessions. The most important ones include the following: · Capital gains on the sale of shares listed on any of Pakistan's three stock exchanges are exempt until June 30, 2010. · Capital gains arising from the sale of real estate unless they derive from the main business activity of the seller; · Any income earned by mutual funds, investment companies, or institutes; and income from holding federal securities; · All agricultural income (unless the taxpayer also earns more than Rs. 80,000 of non-agricultural income); · Any non-military pension, and all armed forces pensions; · Any annuity payments and any payments from the workers participation fund; and · Any fringe benefits for employees of transportation companies, schools, hospitals, hotels and restaurants. In addition, there are also tax credits for four certain expenditure items. 6.5. Second, there are different tax rates for different taxpayers. For example, the rate schedules differ for taxpayers depending on whether they earn more or less than 50 percent of their income from salaries. In addition, male salary employees have a lower income tax threshold than female salary employees. There are also preferential tax rates for particular groups of taxpayers. Examples include senior citizens (50 percent of ordinary tax) and teachers and researchers (75 percent of ordinary tax). 6.6. Third, there are different tax rates for different sources of income. In particular, there are progressive rate schedules for withholding taxes on income from property and differential rates for withholding taxes on company dividends. Withholding taxes on rental income are final (meaning the tax liability can not be adjusted as part of the income tax return), whereas withholding taxes on dividend income are adjustable. There is also a long list of economic transactions subject to withholding taxes, some of which are final. 6.7. Only some of the departures from the global income approach appear well justified. The concept of global income taxation guided for a long time tax reform discussions, as, after all, it followed directly from the principle of horizontal equity ­ an individual's overall taxable capacity, irrespective of the sources of her income, should determine her tax liability. However, as globalization has underscored the high mobility of capital and the low mobility of labor, the efficiency costs of taxing capital heavily have become more apparent. In the late 1980s, the Scandinavian countries began introducing dual tax systems, which combine a low and uniform taxation of capital income with a higher and progressive taxation of labor income. Such systems have since become popular also in other regions of the world. Pakistan's special treatment for dividend, interest, rental income and capital gains can be justified in this light. Nevertheless, as we will discuss in the following paragraphs, there are deviations from the global income approach that are not justified by efficiency considerations. In particular, the system of tax credits and the capital gains exemptions appear unusually generous. In addition, some schedular components could be simplified. 92 6.3 Capital Gains 6.8. One important deviation from the global income concept in Pakistan is the exemption of capital gains from the sales of stocks listed in stock exchanges. The debate over exempting stock market related capital gains inevitably brings out the thorny issue of the differential treatment of this form of corporate earnings and the usual (full or less than full) double taxation of capital income earned by corporations and distributed as dividends. A corporation's decision to retain its after-tax income will ordinarily produce a higher stock market valuation, or capital gain, in the amount of the retention. If capital gains are taxable at the individual investor level, corporate earnings would be subject first to the corporate income tax and then to capital gains tax upon the distribution of after-tax income to corporate shareholders. This would suggest that a capital gain arising from retained earnings should not be subject to individual tax to avoid double taxation. However, as stock ownership is concentrated with high income earners, most countries impose a tax on dividend pay outs, although at a fairly moderate rate (Table 6.2). In that case, some tax on capital gains is warranted in order to avoid distorting corporate distribution polices in the direction of a greater amount of retentions. Since Pakistan imposes a modest rate of tax on dividends, it might also want to consider imposing a similar tax on stock market related capital gains. 6.9. Since developing capital markets is crucial for developing the economy, Pakistan should tread cautiously in taxing long term capital gains. Introducing a capital gains tax would create an incentive for listed companies in Pakistan to register their shares on neighboring stock exchanges and take their capital gains on these alternative stock exchanges where they would not be taxed. In addition, a new capital gains tax would increase the cost of capital in Pakistan since capital gains form an important part of the investment return earned by investors. However, such concerns should not be overstated. During 2006 and 2007, the new capital raised at Karachi Stock Exchange was only less than $200 million. In addition, the bulk of the trading is highly leveraged based on financing provided through the so-called Continuous Funding System. This would suggest that the threat of large amounts of capital moving off-shore is limited. What would a reasonable compromise look like between the goals of promoting the development of a viable stock market and bringing capital gains into the tax net? None of the main emerging economies in Asia and Europe impose significant taxes on long-term stock market related capital gains (Table 6.2), partly because a significant fraction of long-term gains could in fact represent only the effects of inflation rather than real economic gains. In a world of mobile capital resources, Pakistan would be well advised not to break this mould. As for short term capital gains, which are derived over the course of a single year, Pakistan could follow the examples for India and Turkey and impose a moderate tax on this source of income. Taxing short- term capital gains moderately while exempting long-term capital gains would help to encourage long term investment and discourage short term volatility. Estimating the short-term tax revenue impact of such a measure is as much science as speculation, as it involves predicting the stock market. Over the longer-term, the revenue impact should be positive, even after allowing for the elimination of transaction taxes that are currently in place (Box 6.3). 6.10. The taxation of capital gains raises also administrative issues. First, capital gains, especially when derived from sophisticated financial instruments, can be difficult to measure correctly. Tax administrators have to be able to determine precisely when a share was purchased and when it was sold. A reporting system has to be in place that verifies the declarations of short-term capital gains on individual tax returns with information submitted by stockbrokers. Second, real capital gains should ideally be taxable under an income tax as they accrue. Unfortunately, no country in the world has figured out how to do this in an administratively elegant and completely sensible fashion. Instead, capital gains are taxed typically on the basis of when they are realized, and often at preferential rates to compensate for the bunching that occurs when gains accrued over several years are realized in a single year. Third, as capital gains become taxable, capital losses should become deductible from tax liabilities. Since the timing of the realization of capital losses for tax purposes are at the discretion of the taxpayer, tax rules are required for the treatment of such losses to prevent abuse. 6.11. Whatever tax regime is used for capital gains, stockbrokers' income should be tax as regular business income. Capital gains received by those in the regular course of the business operations, such as stockbrokers and other dealers of financial and nonfinancial assets, are truly business income and should be taxed as such. Pakistan is one of the few countries where the financial intermediaries trading income falls under the definition of capital gains, and hence their business income remains untaxed. 93 Box 6.3: Estimating the Revenue Impact of Taxing Stock Market-Related Short-Term Capital Gains What might be the revenue impact of taxing short-term capital gains from stock market transactions? Since tax collection depends on the performance of the underlying tax base, we would have to predict the performance of Pakistan's stock exchanges to estimate the potential yield for 2007/08. Since the future is uncertain, we turn to the recent past and calculate ex- post, admittedly crudely, the potential tax collection from a short-term capital gains tax of 10 percent. To keep the exercise simple, we will make a number of initial assumptions. · We consider only the Karachi Stock Exchange, Pakistan's main trading center. · The value of traded stock can fluctuate widely on a daily basis, from as low as Rs. 1 billion to as high as Rs. 100 billion. We will assume a constant daily trading value of Rs. 25 billion, roughly in line with the value traded in 2006/07. · All traded stocks have been held for a period of less than one year, so that all capital gains are short-term capital gains. · The imposition of the short-term capital gains does not alter the behavior of brokers and investors at the Karachi Stock Exchange. Under these assumptions, the tax collection in 2006/07 would have been Rs. 95 billion, as shown in the figure below. While this might sound like a large number, it is not out of line with previous years. In fact, the average nominal tax collection during 2003/04 to 2006/07 would have been some Rs. 105 billion. These numbers would be at least another 10 percent higher if we include the other two stock markets in Lahore and Islamabad as well as any trading in other financial assets than equity. What explains this sizable tax yield? The main factor is the large increases in stock prices over the period under consideration. The index of the Karachi Stock Exchange rose annually by between 34 percent and 55 percent in these four years. Illustrative Calculation of Potential Tax Collection from Short-Term Capital Gains Tax 70 140 60 KSE 100-Index 120 (% Annual Growth) Potential Tax Collection 100% Capital Gains Short-Term (Rs bn) 50 100 Potential Tax Collection 40 12% Capital Gains 80 Short-Term (Rs bn) 30 60 20 40 10 20 0 0 -10 -20 2003/04 2004/05 2005/06 2006/07 2007/08 -20 -40 But there are also good reasons why these numbers almost certainly overestimate the revenue impact. First of all, stock markets can also go down. Indeed, in 2007/08, the KSE 100-index declined by 11 percent, which would have implied a net capital loss of around Rs. 27 billion. Second, the introduction of the short-term capital gains might go at the expense of the current taxation of the stock market, although some countries have in place taxes on both capital gains and capital value. In Pakistan, all capital gains, and indeed even capital losses, are taxed indirectly through the imposition of a presumptive withholding tax on the purchase of shares (0.01 percent of purchase value); the sale of shares (0.01 percent of sales value); and the trading of shares (0.01 percent of the traded value). In 2006/07, FBR collected Rs. 2.2 billion through these taxes. Finally, the imposition of a short term capital gains tax will change stock market investment behavior. The return to short term investments in the stock market would be lower, which might lead to an overall decline in trading values, as well as lengthen holding durations. This is likely to lower volatility but also reduce the tax yield. Taking the US as a benchmark, no more than 12 percent of all capital gains in the 1999 stock market were short-term. This share is low in part because of the vast depth and breadth of the US financial markets, but also in part because of the tax incentive to realize capital gains only after at least one year: short-term capital gains in the United States are taxed like ordinary income and long-term capital gains are taxed at the preferential rate of 15 percent. Applying the 12 percent share to Pakistan would reduce the tax yield in 2006/07 from Rs. 95 billion to Rs. 12 billion (and on the upside, the tax losses in 2007/08 would be reduced from Rs. 27 billion to Rs. 3 billion). 94 Table 6.2: Taxation of Capital Gains and Dividends in Emerging Economies Country Tax Rate on Dividends Short Term Gains Long Term Gains Brazil 15% 15% 15% India a/ 10% 15% zero Indonesia a/ 10% zero zero Korea Ordinary income 11% for small and medium 11% for small and medium Malaysia Exempt zero zero Philippines b/ 11% zero zero Russia 10% 13% residents, 30% non-residents 13% residents, 30% non-residents Sri Lanka 15% zero zero Thailand 10% zero zero Turkey 10% Ordinary income zero Vietnam 10% zero zero a/ Indonesia imposes a final withholding tax of 0.1 percent on the sale of stock market shares and 0.5 percent on the sale of new shares, whether sold or not. India also has a similar securities transation tax. b/ Philippines taxes only the capital gains of non-resident ali ens at rates of 0.5 percent for listed shares and 5 to 10 percent on unlisted shares. Figure 6.2: The Notch Problem Income Tax Bracket and Maximum and Minimum Marginal Tax Rate For Male Non-Wage Taxpayers 60 6,000,000 Maximum Marginal Tax Rate (At Low Bracket) 50 5,000,000 Marginal Tax Rate at Low Bracket Marginal Tax Rate at High Bracket Minimum Marginal Tax Income Bracket Tax Rate and Rate (At High Bracket) 40 4,000,000 30 3,000,000 20 2,000,000 Tax Rate (%) 10 1,000,000 0 0 1,000 to 1,300 800 to 1,000 100 to 110 110 to 125 125 to 150 150 to 175 175 to 200 200 to 300 300 to 400 400 to 500 500 to 600 600 to 800 Above 1,300 Below 100 Income Tax Brackets (in Rs '000) 95 6.4 Schedules 6.12. The individual income tax contains many schedules. · Taxpayers, who earn less than half of their income from wages and salaries, face a fourteen bracket rate schedule with a rate of zero for incomes of less than Rs. 100,000, and a maximum rate of 25 percent for income over Rs. 1,300,000. By contrast, if a taxpayer earns more than half of his or her income in the form of wages and salary, he or she is subject to a separate twenty bracket rate schedule, with a higher zero rate bracket (up to Rs. 180,000) and a lower top rate (20 percent for incomes in excess of Rs. 8,650,000). · For both salaried and non-salaried taxpayers, the rate schedule applies to the total, rather than the incremental, amount of income. The marginal relief concept introduced in the 2008/09 mitigates this problem, although in a cumbersome way, for salaried taxpayers, but not for non-salaried taxpayers. · Female taxpayers face the same rate schedules as male taxpayers with the exception of thresholds which are increased to Rs. 125,000 for non-salaried, and Rs. 240,000 for salaried taxpayers. · Furthermore, separate rates apply also to rental income, dividends, interest income and lottery earnings. For instance, rental income is taxed at 5 to 15 percent rates with no allowance given for any costs of earning income or for any other taxes paid on that income. Dividends are taxed at a flat ten percent rate. 6.13. The multiple separate tax schedules complicate the income tax system. The steeper rate schedule for non-salaried taxpayers than salaried taxpayers invites evasion through income misclassification. For example, with a taxable income of just over Rs. 1,000,000, the tax liability is Rs. 210,000 for non-salaried taxpayers but only Rs. 90,000 for salaried taxpayers. A simpler system, which would retain a lighter taxation of salaried than non-salaried taxpayers, would be to have a single schedule, accompanied by tax credits (or deductions) for wage income. Similarly, the differential thresholds for male and female taxpayers could be unified in favor of a tax credit for female labor force participation; however, in this case there would be no substantive difference between using differential thresholds or tax credits. 6.14. The zero rate bracket of Rs. 180,000 for male wage earners is unusually large. The high threshold excludes not just low-income taxpayers who would impose an undue burden on the tax administration but also the majority of wage earners from paying income taxes. The average monthly wage in Pakistan is about Rs. 6,500, which translates into an annual income of Rs. 78,000. Workers earning twice the average wage are liable for no income tax. Consequently, Pakistan reaps little tax revenue from labor income and has to compensate for this through higher taxation of business incomes. Furthermore, as the low and middle income groups, which are the primary users of basic public services, pay no income taxes, one important channel for accountability of public service delivery is undermined. Determining the "right" level of income tax exemption is primarily about trading off administrative costs with collection benefits. A useful rule of thumb encountered in the international tax literature is that exemption levels should be set somewhere in the vicinity of twice a country's level of per capita income. This rule would indicate an upper bound for the zero rate bracket of about Rs. 130,000 for wages earners. 6.15. The application of progressive marginal tax rate to the entire taxable income, rather than on increments, creates "notch" problems. Figure 6.2 illustrates the issue. Assume a male non-salaried taxpayer's taxable income increases by one rupee from Rs. 109,999 to Rs. 110,000, which moves him from the 0.5 percent tax rate to the 1 percent tax rate bracket. His previous tax bill was about Rs. 550 (0.5 percent of Rs. 109,999), and his new tax bill is Rs. 1,100 (1 percent of Rs. 110,000). Hence, his marginal tax for the additional one Rupee earned equals Rs. 550, which is a marginal tax rate of about 55,000 percent! By any standard, this is a highly punitive and unfair marginal tax burden. By contrast, if the taxpayer's income increases to Rs. 125,000, the highest income in this bracket, then his marginal tax rate is only 4.7 percent. Similarly, if taxable income rises from Rs. 1,299,999 to Rs. 1,300,000, the marginal tax rate becomes a whopping 5,200,020 percent. If taxable income rises from Rs. 1,299,999 to Rs. 1,500,000, then the marginal tax rate is 51 percent. Clearly, under current arrangements, taxpayers have a strong incentive to conceal any increases in income from the tax collector in order to avoid moving into a higher tax bracket. Being compliant can mean paying more in taxes than what you earn more in income. 6.16. The income tax schedules have also a large number of tax brackets. For the past several decades, many countries have moved towards the adoption of flatter tax rate structures, employing only a handful of tax brackets. Canada, for example, has operated with only three tax brackets since the 1980s. Some developing economies, from Jamaica to Latvia and Ukraine, have gone farther and have adopted single rate individual income tax systems. While they tend to make the distribution of tax burdens less progressive, these reforms can help to increase taxpayer compliance in some countries without major distributional costs. 96 6.5 Tax Credits 6.17. Tax credits are granted up to certain limits and come in four varieties: · Charitable contributions to sports, religious, cultural, welfare, medical and technology organizations; · Contributions to an approved pension fund; · Investment in new shares (IPOs) held for more than one year; and · Any interest in, share in rent, or a share in appreciation of a house paid for by a loan from a recognized lender. These tax credits are targeted allowances in that they encourage certain activities. 6.18. A generous array of tax credits whittles away the income tax base. A high income taxpayer who takes advantage of all of these available credits in a single tax year would have little difficulty in reducing his or her income tax liabilities to zero. In addition, tax credits are designed in a way that they resemble tax deductions whose value increases with the size of the taxpayer's marginal (and average) tax rate, and hence they confer larger tax savings for those with higher incomes. Overall, Pakistan's personal tax credits are too numerous and generous, too complex and regressive in design, and exact heavy administrative costs on the tax authorities and compliance costs on the taxpayers. 6.19. Most counties provide either a tax credit or a tax deduction for charitable donations, but two issues might be considered. First, the current definition of a charitable institution is wide by including, for example, sports and cultural activities. Second, income of non-governmental organization, to the extent it is derived from commercial activities, should be subject to taxation. 6.20. In view of the tax exemption for pension income, the tax credit for pension contributions is unusual. Pakistan grants favorable tax treatment to a patchwork of public and private pension schemes, as discussed in Box 6.2. Up to certain limits, contributions are creditable against tax liability and benefits are non-taxable when paid out. The tax credit provides a tax saving with no offsetting flow of future tax payments. The exemption of non- contributory pension benefits is unjustified from the point of view of horizontal equity and biases compensation packages toward retirement compensation. Even in the case of contributory pension benefits, these should be taxed provided that contributions to the retirement fund were made tax free. Standard praxis in most countries is to tax either pension contributions or, more frequently, pension benefits. 6.21. The tax credits for buying new shares and housing provide only a tax saving for the taxpayer and no corresponding tax obligations at some future point. This happens because capital gains reaped from holding shares and the implicit rental income of housing are both exempt from taxation. 6.22. Independently of any changes Pakistan might make in the number of tax credits it offers, the system could be simplified. An allowance could be given either as deductions to income, or as credits against the tax liabilities. There is an important difference between the two concepts. In principle, under a progressive tax rate structure, the value of a deduction increases with a taxpayer's applicable rate band, while tax credits deliver the same amount of tax relief regardless of the taxpayer's taxable income. A credit for charitable donations, for example, might allow a credit equal to ten percent of the donation. Taxpayers making the same donation would be able to claim the same amount of tax saving no matter what their taxable income status. In Pakistan, however, tax credits work like a form of deduction, since the value of the credit increases with the taxpayer's average tax rate. This practice goes against the common view that high income earners should not be granted more valuable targeted allowances than low income earners. 97 6.6 Equity 6.23. The individual income tax is progressive. Figure 6.3 shows the effective tax rates by household consumption quintiles separately for salaried and non-salaried taxpayers, based on the tax incidence model introduced in Chapter 3. The salaried income tax, which accounts for 15 percent of the overall individual income tax collection, falls predominantly on the top income deciles. The large threshold level implies that the effective tax rates on the first six deciles are close to zero. Most of the tax burden falls on the top consumption decile which accounts for over three-quarters of the overall tax collections from salaried taxpayers. However, with salaried income accounting only one tenth of non-salaried income, the effective tax rates of salaried taxpayers pale in importance compared to those of the non-salaried taxpayers. The non-salaried income tax is also progressive, as higher consumption deciles are more likely to have non-wage sources of income. Adding the two taxes together, the individual income tax becomes progressive throughout the full decile range (Figure 6.4). The effective tax rate increases from 0.6 percent of GDP for the bottom decile to 1.4 percent of GDP for the top decile. 6.24. Differential tax evasion across the income distribution has a bearing on the progressivity of the tax system. In light of the discussion on tax compliance in the next section, is there a need to adjust our results? No, these findings already take account of compliance differences among income groups for the individual income tax, based on a comparison of simulated tax payments using household survey data and actual tax returns. Figure 6.3: Incidence of Individual Income Tax in 2006/07 for Salaried and Non-Salaried Individuals Incidence of Salaried Income Tax by Household Consumption Deciles Incidence of Non-Salaried Income Tax by Household Consumption Deciles 80 1.2 80 1.2 Effective Tax Rate (% of GDP) 1.1 1.1 Share in Total Taxes 70 (%) 70 1 1 60 0.9 60 0.9 0.8 0.8 50 Effective Tax Rate 50 (% of GDP) 0.7 0.7 40 0.6 40 0.6 0.5 0.5 30 30 0.4 0.4 20 0.3 20 0.3 0.2 Share in Total Taxes 0.2 10 10 (%) 0.1 0.1 0 0 0 0 Bottom 2 3 4 5 6 7 8 9 Highest Bottom 2 3 4 5 6 7 8 9 Highest Figure 6.4: Overall Incidence of Individual Income Tax in 2006/07 Incidence of Salaried and Non-Salaried Income Tax by Household Consumption Deciles 50 1.4 Effective Tax Rate (% of GDP) 1.2 40 1 30 0.8 0.6 20 0.4 10 0.2 Share in Total Taxes (%) 0 0 Bottom 2 3 4 5 6 7 8 9 Highest 98 6.7 Compliance 6.25. Since the adoption of the 2001 Income Tax Ordinance, the federal income tax system operates on a self-assessment basis. This means that the government expects taxpayers to determine their own tax obligations and to pay voluntarily whatever is due. The Universal Self-Assessment Scheme places the onus on taxpayers, and the government avoids the costly alternative of determining each individual's tax liability and doing whatever it must to collect it. The tax returns are considered as final unless their cases are selected for audit through pre- announced audit parameters. The frequency and effectiveness of taxpayers' audits are important determinants of voluntary compliance, yet tax audits remain weak in Pakistan. One cost of relying so heavily on the voluntary compliance of taxpayers without well-structured audit programs is that a large share of the tax is not paid. At least, this is suggested by our estimate for the tax gap of individual income tax. 6.26. We calculate the tax gap through a three-step procedure. First, the Labor Force Survey (for salaried individuals) and the Pakistan Living Standard Measurement Survey (for non-salaried individuals, associations of persons) provide us with representative samples of the potential tax base that is not subject to intentional reporting errors for tax purposes. Second, we apply a tax calculator for the 2004/05 individual income tax system on the estimates of taxable income for each observation in the sample. The tax calculator makes, whenever possible given available information, allowances for tax exemptions and credits. Concessionary rates for senior citizens and teachers make little difference as their incomes are typically below the threshold for income tax. Third, individual income tax collection also includes tax receipts from rental income, capital gains, interest payments, dividends and other income sources. Since we lack information on these sources in the Pakistan Living Standard Measurement Survey, we assume somewhat arbitrarily the same rate of non-compliance as for self-employed income. 6.27. The estimated tax gap is large for salaried and small for non-salaried taxpayers, perhaps due to the attribution of withholding taxes to individual income tax from non-salaried taxpayers. We estimate the tax gap to be about 97 percent of actual receipts for salaried taxpayers, and 15 percent of actual receipts for non- salaried taxpayers, including all other sources of individual income (Figure 6.5). This would translate into a tax gap of Rs. 31 billion in 2007/08, equal to 27 percent of individual tax revenues, and 3 percent of federal tax revenues. The estimated level of non-compliance among the non-salaried taxpayers appears to be too low by international standards. For example, in the United States, the tax gap for the self-employed is estimated to be about 40 percent. As the informal economy is larger and tax enforcement weaker in Pakistan than in the United States, one would expect a tax gap at least as high as in the United States. One explanation for this counter-intuitive result is FBR's accounting of receipts from Pakistan's extensive withholding tax regime. Some three-fifth of actual income tax revenues from non-salaried individuals comes from withholding taxes. Yet, the withholding taxes on telephone, electricity bills and bank withdrawals undoubtedly collect significant revenue also from the salaried individuals. Therefore, attributing all of the revenue from these collection heads exclusively to non-salaried individuals may be masking a larger tax gap for the self-employed. At the same time, this attribution may be leading us to mistakenly overstate the tax gap for salaried individuals. Similarly, receipts for dividends and interest come largely from bank withholding, which covers both the corporate and non-corporate sector. Attributing all these withholding taxes to non-salaried taxpayers underestimates the tax gap for non-salaried taxpayers and overestimates it for other taxpayers. 6.28. Concerns about low taxpayer compliance are borne out by the low, though increasing, number of tax return filers. While the number of national tax number holders for non-corporate income tax has risen since the beginning of this decade, the number of income tax filers has increased only since 2003/04, leaving the 2006/07 compliance rate of 66 percent some 20 percent below the 1999/2000 level (Figure 6.6). However, the effective compliance rate might be higher, at around 80 percent, given that up to 500,000 of those national tax numbers could be invalid. Taxes are also paid via withholding schemes, such as for electricity, telephone and cash withdrawals, which are not captured in the number of tax return filers. Nevertheless, a number of 2.1 million income tax filers is low in view of employed workforce of close to 50 million, some 10 million of which are clerks, technicians, professionals, senior officials and managers. 99 Figure 6.5: Individual Income Tax Compliance Individual Income Tax Gap (Percent of Actual Collection) Gap (29%) Actua l (100%) Salaried Income Tax Gap (Percent of Actual Collection) Non-Salaried Income Tax Gap (Percent of Actual Collection) Gap (15%) Gap (97%) Actual (100%) Actual (100%) Figure 6.6: Trends in Individual Income Tax Filing Income Tax Filing 2800 90 National Tax Number ('000) 80 2400 Active Compliance (%) Active National Tax 70 Number ('000) 2000 60 Income Tax Filers ('000) 1600 50 40 1200 30 800 20 400 10 0 0 1999/2000 2000/2001 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 100 6.8 Policy Options 6.29. There are three priorities in the reform of the individual income tax. The changes could be implemented separately and in piecemeal fashion (Box 6.4). At the same time, as argued in Chapter 2, there are good merits for overhauling the individual income tax system more fundamentally in the direction of a single tax rate system. 6.30. First, streamlining credits and phasing out many exemptions would move it closer to a fully-fledged broad-based income tax. The tax credits could be reshaped so that the amount of targeted allowances remains unchanged irrespective of the income level of the beneficiary taxpayer. While most exemptions and credits are relatively minor, there is one noticeable exception. Imposing a modest tax on short-term stock market related capital gains would seem attractive from multiple perspectives: it benefits the national exchequer by generating revenues over the medium term; it improves efficiency by making companies indifferent between retained earnings and paying out dividends; and it improves equity by making the tax system more progressive. Arguably, it would also be good politics, as it would be a symbol for a government ready to tackle powerful lobbies in the national interest. Whatever is done with the capital gains tax, there is every reason to tax the incomes of stockbrokers according to the regular schedule for non-salaried taxpayers 6.31. Second, the tax design can be simplified. The individual income tax operates with two basic schedules for recipients of wage and non-wage or business income. There are anywhere from fourteen to twenty different tax brackets each having large but different zero rate brackets for men and women. The high threshold level could be addressed over time by keeping the zero-band constant in nominal terms in spite of inflation. The individual income tax applies higher marginal tax rates to the total amount of a non-salaried taxpayer's income. This notch problem can result in punitively high effective rates of marginal taxation and create perverse incentives to accurately report additional amounts of taxable income. The newly introduced marginal tax relief mitigates this problem to some extent, but it further complicates the tax system and applies only for salaried taxpayers. 6.32. Third, tax compliance could be improved for both salaried and non-salaried taxpayers. Revenues from individual income tax could rise from 1.1 percent of GDP to 1.4 percent of GDP simply by closing the tax gap. Improving voluntary compliance through simplification in the tax system will go some way. However, there is a clear need for greater enforcement efforts given the large tax gap for salaried taxpayers (97 percent of actual collections); and the large potential tax gap for non-salaried taxpayers (non-salaried tax contributions account for 85 percent of all individual income tax collection; and withholding taxes camouflage a large compliance problem). A comprehensive effort, ensuring that taxpayers meet their legal obligations ranging from registration, keeping accurate records, and filing timely returns, is required. In particular, effective audits would generate revenues directly through collection on demand, and, more importantly, provide a positive indirect deterrent effect for taxpayers. 101 Box 6.4: Estimating the Revenue Impact of Individual Income Tax Reforms The revenue impact of the three policy measures would be as follow: · Removing the notch structure by making rates applicable only to taxable income within the respective tax brackets would reduce revenues by Rs 9.4 billion (6.5 percent of target revenues for 2007/08). Thus, restoring a conventional approach to the rate schedule is a revenue loser. This result is intuitive, given that higher tax rates apply then only for a slice rather than the entirety of the taxable income. This effect would be mitigated to the extent that a marginal tax structure would reduce the incentive for taxpayers to underreport their taxable income. · Eliminating all reductions and credits would increase revenue by Rs. 0.9 billion (0.6 percent). Pakistan's tax credits and tax reductions, while complicating the tax system considerably, are not particularly costly. · Eliminating the tax credit for the purchase of new stock would increase revenue by Rs. 0.1 billion (0.1 percent). Table 6.3: Policy Reform Options for Individual Income Tax Topic Proposals Capital Gains · Introduce taxation of short-term stock market related capital gains · Tax stockbrokers' income according to the non-salaried income tax schedule Schedules · Broaden the tax base by reducing the zero-rate income thresholds in real terms · Replace the two basic schedules with single schedule accompanied by tax credits for earning labor income or for female labor force participation · Remove notch problem by reverting to conventional method of taxing personal incomes progressively · Reduce significantly the number of tax brackets Tax Credits · Streamline system of personal tax credits (charity; pensions; purchases of new shares; purchases of housing) · Ensure that tax credits deliver the same amount of tax relief regardless of the taxpayer's taxable income Administration · Improve accounting system for withholding taxes to salaried and non-salaried taxpayers · Implement comprehensive risk-based compliance strategy · Expedite audit programs, especially for withholding of taxes from salaried workers 102 CHAPTER 7: FEDERAL EXCISE DUTIES 7.1 Introduction 7.1. Excise duties are taxes on the sale or use of specific goods and services, such as tobacco and petrol. Their design does not follow the standard prescription of a broad base and low rates. Instead, a good excise system is invariably one that generates revenue from a narrow base with relatively low administrative costs. This lesson from practice is also grounded in economic theory which suggests that excise taxes should be highly selective, narrowly targeting a few goods. Ramsey (1927) founded the modern theory of optimal taxation with an excise taxation model. With identical consumer preferences, optimal excise tax rates vary inversely with elasticities of demand for taxed goods: the larger the demand reduction in response to price increases, the larger the efficiency costs, and hence the lower the optimal excise tax rate. However, Atkinson and Stiglitz (1976) showed that, under simplified assumptions, nothing is gained from supplementing an income tax with differentiated excise taxes. But the Atkinson-Stiglitz framework does not consider three important roles in which excise taxes can strengthen a country's taxation system. First, they levy revenues according to the benefit principle of taxation. For example, petrol fuel taxes ­ used in many countries but not imposed in Pakistan ­ are similar to user fees, as the tax burden increases with the use of public roads. Second, their consumption entails negative externalities on society. For example, car drivers add to road congestion and environmental pollution, justifying taxation of petroleum products. Third, excise taxes discourage the consumption of potentially harmful substances, such as tobacco or alcohol. 7.2. In line with international trends, Pakistan has scaled back its reliance on excise taxes. With the rise in income and value-added taxes, the reliance on excise taxes has declined around the world over the last decades, as countries have more and more restricted excise taxation to goods like petroleum products, tobacco and alcohol. In 2004, the US federal government collected only 4 percent of its total tax revenues from excise taxes, close to half of which came from petroleum products. But there are exceptions to this general pattern. Perhaps the one of the starkest example is Turkey. Turkey's high excise taxes on traditional excisable goods as well as luxury goods collect close to 8 percent of GDP, and contribute nearly one-third of all tax receipts. And for petroleum products, the steep rise in international fuel prices has put pressure on governments to moderate domestic price increases through a reduction in petroleum taxes. In Pakistan, federal excise taxes declined from 1.5 percent of GDP in 1999/2000 to 0.9 percent in 2007/08, and now contribute 8 percent of FBR taxes (Figure 7.1). 7.3. This chapter highlights four weaknesses in Pakistan's excise taxation. They related to the legal framework, the scope of excise taxation, the composition of the rates structure, and compliance (Table 7.1). Figure 7.1: Pakistan's Federal Excise Tax Collection Over Time and in International Comparison Pakistan's Tax Collection of Federal Excise Duties Federal Excise Duties in Pakistan and Other Emerging Economies 1.5 30 % of GDP (2006/07 or latest available) 8 12,000 1.4 PPP GNI Per Capita 1.3 25 7 1.2 10,000 % of Indirect Taxes 1.1 6 % of FBR Taxes 1.0 20 8,000 0.9 5 0.8 15 4 6,000 0.7 % of GDP 0.6 3 0.5 10 4,000 0.4 2 0.3 2,000 5 1 0.2 0.1 0 0 0.0 0 Pakistan Bangladesh Nepal South India Sri Lanka Thailand Turkey 1999/2000 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 Africa 103 Table 7.1: Main Findings of Chapter 7 Topic Findings Revenue Trends · In Pakistan, federal excise taxes declined from 1.5 percent of GDP in 1999/2000 to 0.9 percent in 2007/08, and now contribute 9 percent of FBR taxes. Tax Design · The excise tax base consists of seven categories: cigarettes, cement, vegetable oils, carbonated beverages, seventeen different types of oils and fuels, toiletries and cosmetics, and services. · The law is a weak legal vessel in that it grants extensive powers to the FBR to change the nature, rate and manner of collection of the tax at any time. In Rem versus Ad · Pakistan's treatment of excisable products is in line with standard Valorem Rates prescriptions with the exception of tobacco products. Luxury Consumption · Excise taxation of luxury consumption could bring in additional revenues and improve the progressivity of the distribution of tax burdens. Equity · Excises are regressive. The effective tax rate declines from 1.1 percent of GDP for the bottom decile to 0.6 percent of GDP to the top decile. Compliance · The federal excise tax gap is approximately 69 percent of actual federal excise tax revenues. 104 7.2 Legal Framework 7.4. The 2005 federal excise tax law defines all of the important excise tax concepts and the procedures for complying with the tax. The law replaced the earlier system of physical excise control with a new system in which all clearances are now self-assessed. Taxpayers are domestic producers of excisable goods and services and purchasers of excisable imports of goods. The excise tax base consists of 48 goods and, until recently, 8 services that are described in the law. Although the number of goods seems large, they can be grouped into seven categories: cigarettes, cement, vegetable oils, carbonated beverages, seventeen different types of oils and fuels, toiletries and cosmetics, and services (Figure 7.2). With the exception of cement and services, this list of excisable products is fairly standard from international perspective. The law outlines a moderate list of exemptions from excise tax. Most of them are sensible, such as the exemptions given to international agencies and organizations and to purchases made by outbound ships and planes; others, such as the exemption for hydraulic cement imported or purchased by petroleum or energy sector companies, appear ad hoc. 7.5. An interesting feature of excise taxation in Pakistan is that it mimics in a number of instances the behavior of a value-added tax. Excisable services are generally taxed at the standard sales tax rate of 16 percent. Excisable services include TV advertising, inland air travel, inland air freight, first class or air conditioned train travel, shipping agents, a broad array of telecom services, and all types of insurance services except life. In 2006/07, additional services were added to the excise list, including cable TV; money changer, non-fund or fee related banking services; and international air travel. The excise duty applied to most services is treated as if it were a sales tax, turning it into a creditable charge to any sales tax registrant. 7.6. The resemblance of excise taxes to the general sales tax holds also for some goods. For example, vegetable oils is taxed at the standard sales tax rate of 16 percent, to be collected under sales tax mode and any tax paid by a sales tax registrant is entitled to an input tax credit. Moreover, the law provides for zero rating of all excisable goods that are exported. In addition, it allows a producer of excisable goods to deduct from tax liability any amount of excise that may have been applied to the inputs used by the producer. 7.7. However, the federal excise legislation can be improved in a number of regards, many mirroring weaknesses in the sales tax legislation. The law is a weak legal vessel in that it grants extensive powers to the FBR to change the nature, rate and manner of collection of the tax at any time. Some provisions allow the tax authorities to interfere in matters that should be decided exclusively by supplier and customer. Some definitions, including the definition of the sales tax mode, could be simplified and others require cross-referencing to the sales tax law. In addition, there is a good case to transform the tobacco excise tax from ad valorem to in rem (Box 7.1 and Box 7.3) and to simplify the rate structure for petroleum products (Box 7.2). Finally, the use of federal excises in sales tax mode is unlikely to be an appropriate substitute for expanding the scope of coverage from selected services to services overall through a broad-based VAT. 7.8. Over the long term, Pakistan may consider introducing green taxes as part of the excise roster. Green taxes are taxes on "bads" and are one of the few "good" taxes that have ever been invented by mankind. The "bads" are environmental externalities representing unwanted by-products and wastes from production processes that reflect differences between the social and private costs of producing goods and services. These differences arise because private producers have no cost incentives to take into account the environmental degradation, for example lower air and water quality, that may be associated with their production activities. Green taxes are designed to make producers aware of the social costs they impose on the economy and to impart incentives to generate lower levels of pollution. For example, taxes on water and air effluents are intended to make reduction in pollution privately profitable, thereby improve air, water and ground quality and, at the same time, generate revenue that can be used to reduce the revenues collected from other taxes that typically impair economic efficiency. Among others, this would require imposing higher excise duties on leaded than unleaded petroleum products. Despite their advantages, green taxes may prove to be difficult to implement in practice. Because information about the cost and ease of abating pollution is not easy to obtain, it is difficult to predict how much pollution will be reduced at any given green tax rate. For this reason, among others, environmental regulation often goes hand-in-hand with, and sometimes replaces, green taxation. 105 Figure 7.2: Composition of Federal Excise Tax Collection Pakistan's Net Tax Collection from Federal Excise Taxes (%) 100% Other Categories 90% Petroleum Products 80% Beverages 70% Natural Gas 60% Services 50% Cement 40% 30% 20% Cigarettes 10% 0% 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 Box 7.1: Defining Excise Tax Rates ­ In Rem versus Ad Valorem The in rem (or specific) rates and ad valorem (price related) rates have both their strengths and weaknesses. In rem taxation has administrative advantages. The tax inspector has only to count the volume of production and can avoid the vexatious issue of having to place a value on the product for excise tax purposes. In addition, an in rem approach is more suitable for products with negative externalities. For petroleum products, for example, the environmental harm attributable to fuel use is more closely related to the liters of petrol consumed than to the value of the petrol itself. Whenever items are taxed on in rem basis, it is important to index these rates to the annual inflation rate and avoid any erosion of excise tax collections through general price increases. By contrast, ad valorem taxation has the advantage that taxpayers pay according to the value of what they consume; and hence taxes paid also go up automatically with the price of a product. This is particularly appealing if the intention is to tax heterogeneous goods where the consumer has the choice between cheap and expensive varieties, such as consumer durables and toiletries. Pakistan's treatment of excisable products is in line with these assessments with the exceptions of tobacco products and one category of petroleum products. In rem rates are used for most petroleum products, and ad valorem rates for edible oils, beverages, tobacco products and beauty products. The highest ad valorem rate is on high priced cigarettes, at 63 percent of the retail price, followed by carbonated beverages at 50 percent. · For edible oils, and aerated beverages and their ingredients, retaining the current ad valorem approach is justified in light of the considerable heterogeneity of products found in the product group. · For tobacco products, converting the existing ad valorem rates into equivalent in rem rates appears advisable from the perspective of both administration (given the difficulty of attaching values accurately for similar products) and efficiency (given the of petroleum use). · For petroleum products and their derivatives, most excise tax categories are already taxed on an in rem basis. However, there is a large number of different rates on seventeen different kinds of oil and fuel. Too many fine distinctions pry open the door for tax evasion through misclassification. Most countries operate with a much simpler system of tax rates for petroleum products and their derivatives. Another anomaly remains with "other mineral oils excluding sewing machine oil" which is currently taxed at a fifteen per cent ad valorem rate. Unless there is large heterogeneity within this category, this product could well be taxed on in rem basis. · The only other excisable products subject to ad valorem treatment are beauty products and toiletries as well as the category "other" under the heading organic composite solvents and thinners. These categories display large product heterogeneity which suggests that the current ad valorem treatment is appropriate. 106 Box 7.2: Excise Taxation and Non-Essential Consumption While standard tax policy advice is to limit excise taxation to a few items, bringing non-essential goods and services under excise taxation is attractive in view of Pakistan's revenue needs. Apart from reaping higher revenues for the national exchequer, such excise taxation might also introduce a greater progressivity into the distribution of tax burdens, which, as we will see later in the chapter, is regressive for excise taxes. Finally, such taxation could also help to curtail imports, and thereby reduce the demand for foreign exchange, without undue protectionism for the domestic economy. Since services are a provincial tax base, leveling excises on services has implication for the inter-governmental revenue allocation. This issue is taken up in Chapter 9 in the discussion of provincial taxes. A number of considerations should guide the selection of household consumption items for inclusion into a new excise tax base of non-essential goods and services. The first is the definition of non-essential consumption itself. Non-essential consumption should be defined as one having a high income elasticity of demand, preferably much higher than one. This makes sure that it is indeed consumed foremost by rich people. Second, the item chosen should have a relatively low price elasticity of demand. This would restrict losses in efficiencies arising from large changes in consumer demand. This criterion implies that the excise tax base chosen should encompass broad expenditure categories. One would not, for example, want to tax cigarette lighters or only certain kinds of jewelry such as diamond necklaces. What kinds of goods and services are most likely to satisfy the criteria described above? In the area of medical and dental services one could impose an excise tax on the value of non-essential services such as cosmetic surgery or tooth implants. Personal care services probably offer a wider range of taxation opportunities such as taxing beauty parlors, sports clubs and recreational facilities. Legal services are another possibility but here one would want to design the tax in a way that would avoid it becoming a tax on doing business. Non-business or non-routine legal services such as legal fees for divorce proceedings or civil demands would be better candidates for excise taxation. Recreational equipment, such as golf clubs, would also probably fit neatly into the criteria described above. Finally, following Turkey's example, it would make sense to extend the scope of excise taxation on modern communication services, such as the use of cell phones, and transportation services to not only include cars but also motorcycles, pleasure boats, airplanes and helicopters. As for the structure and level of the tax rate, an ad valorem rate of 16 per cent, in line with the standard GST rate, would seem appropriate. This rate is commonly applied to other excisable goods and is not high enough to encourage rampant evasion and smuggling. The table below shows some of the potential categories of services and commodities to which a non-essential excise tax could be applied. These are likely to have high income demand elasticities, although most of them will have medium to high price demand elasticities. The revenue impact is derived based on a 16 percent ad valorem rate, and 50 percent effective compliance rate, based on the 2004/05 PLSM expenditure data inflated up to 2007/08. Potential 2007/08 Tax Revenue from Excise Duties on Selected Non-Essential Goods and Services (16% Rate and 50% Compliance) 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 Transportation Medical and Personal Care Car, Bicycle, Mobile Phones Legal Expenses IT Purchases Dental Motorcycle & Recretional Equipment 107 7.3 Equity and Compliance 7.9. One concern about excise taxation is their potential regressivity. For one, excise duty rates do not rise with consumption in the way that income tax rates rise with income. And since poor people spend higher fractions of their incomes than non-poor people, taxes based on expenditures rather than income may put greater relative burdens on poor people. Ultimately, the distributional impact of excise taxes depends on the income elasticities of demand for goods and services subject to excise taxation. Of course, from an efficiency point of view, it is desirable that the buyers of products do not adjust their purchases much in response to the change in prices due to the imposition of the excise taxes. Such "inelasticity" of demand limits the excess burden, or the distortive impact of imposing excise taxes on these products. At the same time, it also means that consumers carry the tax burden in terms of higher after-tax prices. For our analysis, we assume that the tax burden falls fully on consumption, based on disaggregated excise tax receipts. Excises are regressive. The effective tax rate declines from 1.1 percent of GDP for the bottom decile to 0.6 percent of GDP to the top decile. The regressivity comes mainly through the tax on tobacco, which takes up a larger share of expenditures in lower deciles than in higher deciles (Figure 7.3, left panel). Figure 7.3: Equity and Compliance for Federal Excise Taxation Incidence of Federal Excise Tax by Household Consumption Deciles Actual Tax Collection and Tax Gap of Federal Excise Duties 25 1.2 Effective Tax Rate (% of GDP) 1 20 0.8 Gap (69%) 15 0.6 10 0.4 Share in Total Taxes (%) 5 0.2 Actual (100%) 0 0 Bottom 2 3 4 5 6 7 8 9 Highest 7.10. The federal tax gap is estimated to be over two-third of actual tax collection. The estimate for the tax gap from federal excise taxes is derived as follows. From the 2004/05 PSLM, we derive the value of total consumption for 4 out of 10 broad categories of excisable goods, specifically aerated water and beverages, tobacco products, edible oils, and cosmetics and perfumes. Since the values of consumption reported in the PSLM reflect final retail sales, and the federal excise taxes are collected at the manufacturing stage, we take 80 percent of the PSLM values to account for markups by wholesalers and retailers. The value of these commodities, adjusted for the retail and wholesale markups, account for approximately 52 percent of total federal excise tax revenues in 2004/05. We multiply these values by the statutory excise tax rate to estimate potential revenues, and apply the household weight to the sum of the sample values to obtain estimates of the corresponding population values. For those commodities for which there is no PSLM category, we assume that their tax gap is equal to the average tax gap for the excisable goods with data. The federal excise tax gap is approximately 69 percent of actual federal excise tax revenues (Figure 7.3, right panel), or Rs. 63 billion in 2007/08. This tax gap is relatively large, since, for example, all the federal excise revenue from edible oils comes from imports. These estimates may even understate the `true' federal excise tax gap because some of the excisable goods, such as cooking oils and aerated water and beverages, are often used for intermediate use, and their consumption is not captured in the PSLM. 108 7.4 Policy Options 7.11. Federal excise taxes play an integral role in Pakistan's tax system. It would seem easy to dismiss their importance. Among the major taxes, they contribute least in terms of revenue collection, and their importance might well decline further in view of the political pressures to reduce excise rates of petroleum products in the light of high fuel prices. Yet, it is crucial to bolster the role of excise taxes (Table 7.2). They could serve well the traditional excise tax goals of raising revenue, discouraging undesirable types of consumption, pricing some public services such as roads, and improving the overall progressiveness of taxation. Maintaining sizable excise taxes on petroleum products is justified on efficiency grounds, as the use of petroleum is harmful for the environment and for the balance of payments. Expanding the coverage of excises by including services ­ and possibly selected non- essential goods in view of Pakistan's dire revenue needs ­ that so far have escaped effective taxation could help to fix a major whole in the tax system and boosts revenues along the way. As discussed in Chapter 2, this should be an interim arrangement only until a fully fledged value added tax, covering both goods and services, is put in place. Additional areas of reform include modernizing the legislative framework and bringing it in line with any changes in the sales tax law; converting ad valorem rates to in rem rates for tobacco products; and simplifying the rates structure for petroleum products. Like for the other taxes, ensuring good compliance with the federal excise taxes is crucial. Table 7.2: Policy Reform Options for Federal Excise Tax Topic Proposals Legal Framework · Strengthen the federal excise legislation by restricting the extensive powers of FBR, simplifying definitions and bringing the law in line with the sales tax law. Rates · Converting ad valorem rates into in rem rates for tobacco products · Index in rem rates for inflation · Simplify the rates structure for petroleum products Non-Essential Consumption · Bring additional non-essential goods and services under excise taxation Compliance · Strengthen enforcement Box 7.3: Reforming Excise Taxation on Tobacco Cigarettes are inexpensive in Pakistan. The most popular brands cost PKR 18.40 per pack of 20 (USD 0.25 per pack at the current exchange rate). In part because cigarettes are inexpensive, annual consumption of cigarettes has increased from 292 cigarettes per capita in 1994 to 406 cigarettes per capita in 2007. This raises serious health concerns. A proven intervention to reduce smoking is to raise the price of cigarettes relative to other products by increasing excises on cigarettes. Although Pakistan adjusts its cigarette excises almost every year, the rates have not kept pace with inflation and the growth of per capita incomes. Domestic cigarettes are classified into three tiers based on the retail price before GST. Each tier is subject to a different excise regime. Cigarettes in Tier I (the lowest-priced cigarettes) are subject only to a specific excise (based on quantity). Cigarettes in Tier II (cigarettes in the mid-price range) are subject to mixed regime comprising a specific excise and an incremental ad valorem excise (based on value). Cigarettes in Tier III (the highest-priced cigarettes) are subject only to an ad valorem excise. Pakistan could return to a two-tier regime similar to that abandoned in 2001, and thereby reduce cigarette consumption, increase government revenue, and simplify the excise regime. In particular, the specific excise could be increased from Rs. 6.34 to Rs. 15.00 per pack of 20 cigarettes, and the bracket for the first tier would be increased from Rs. 14.86 to R.s 28.00. For cigarettes in the second tier, the excise tax would be 63 percent of the retail price before GST. Going forward, the specific rate and the bracket between the two tiers would be automatically indexed for inflation. Under the proposal, 80 percent of all cigarettes would be in the first tier and subject to the specific excise. Assuming the excise tax is fully passed through to consumers, adoption of this proposal will lead to a 50 percent increase in the price of the most popular brands. Consumption of cigarettes would decline by 18 percent, providing significant health benefits, and the government's revenue from cigarette excises would increase by 47 percent from the forecasted 2008/09 revenue of Rs. 35 billion to Rs. 52 billion. Source: Petit and Sunley (2008). 109 CHAPTER 8: CUSTOM DUTIES 8.1 Introduction 8.1. Customs duties are levied on the dutiable value of goods imported into Pakistan. They raise revenue and provide protection to domestic producers against foreign producers. The tariff schedule provides for duty rates, ranging from 0 to 25 percent in 5 percentage point increments. There are exceptions to this general rule. The rates for automobiles, automobile parts, beverage alcohol, and tobacco products are between 25 and 100 percent. In addition, exemptions from import duties have eroded the custom duties base. Only five commodity groups contribute about 55 percent of gross duty collection (Figure 8.1), and the share of dutiable imports in overall imports declined from over 70 percent in 2004/05 to just over 50 percent in 2007/08. 8.2. Over the last two decades, the government's reliance on custom duties' collection has reduced substantially. As part of the trade liberalization reforms launched in the late 1980s, the government made substantial cuts in import tariff rates across all sectors of the economy and simplified the tariff structure by reducing the number of tariff bands. The trade-weighted average tariff rate fell from over 50 percent in 1995/96 to around 11 percent in 2007/08. Pakistan now has the lowest average tariff rates of the three large South Asian economies (Bangladesh, India, and Pakistan). Where once most products faced customs duties of 50 percent or more, now the median tariff rate is 10 percent, 40 percent of all tariff lines lie in the zero and 5 percent bands, and 25 percent is the maximum standard tariff rate (Box 8.2). These reforms reduced the customs duties' share in federal tax revenue from 47 percent in 1990/91 to 18 percent in 1999/2000 and 15 percent in 2007/08 (Figure 8.2, left panel). Custom duties contributed about 1.4 percent of GDP to federal tax collection in 2007/08, compared to 1.6 percent in 1999/2000. Pakistan's experience mirrors those of other countries that have liberalized their trade regimes (Box 8.1). Its revenue collection from custom duties as percent of GDP is lower than in other major South Asian economies, reflecting the stronger trade liberalization (Figure 8.2, right panel). Because sales and excise tax apply to the tariff laden value of an import, lower tariffs also contribute indirectly to lower yields (in absolute terms) from both sales and excise taxes. This chapter looks at the equity and compliance dimensions of custom duties, and then proposes a broad strategy for reform (Table 8.1). Figure 8.1: Composition of Custom Duty Collection Pakistan's Gross Tax Collection from Custom Duties (%) 100% 90% 80% Other Items 70% 60% Mechanical Machinery 50% Electric Machinery 40% Edible Oil and Waxes 30% POL Products 20% 10% Vehicles 0% 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 110 Figure 8.2: Pakistan's Custom Duty Collection over Time and in International Comparison Pakistan's Tax Collection of Custom Duties Custom Duties in Pakistan and Other Emerging Economies (2006/07 or latest 2.0 30 available) % of GDP 12,000 % of Indirect Taxes 3 PPP GNI Per Capita 1.8 25 1.6 10,000 % of FBR Taxes 1.4 20 8,000 1.2 2 1.0 15 % of GDP 6,000 0.8 10 0.6 1 4,000 0.4 5 2,000 0.2 0.0 0 0 0 1999/2000 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 Turkey South Africa Pakistan Egypt Thailand Sri Lanka India Bangladesh Nepal Table 8.1: Main Findings of Chapter 8 Topic Findings Revenue Trends · The customs duties' share in federal tax revenue declined from 47 percent in 1990/91 to 18 percent in 1999/2000 and 15 percent in 2007/08. · Pakistan's experience mirrors those of other countries that have liberalized their trade regimes. Equity · Custom duties are fairly proportional. The bottom decile accounts for 3 percent of all custom duties paid and faces an effective tax rate of 1.46 percent; the top decile accounts for 33 percent of all custom duties paid and faces an effective tax rate of 1.49 percent. Compliance · The estimated import duty gap is 21 percent of actual duty revenues but there are large difference from one country to another. 111 Box 8.1: Liberalizing Trade Without Lowering Tax Revenue While trade liberalization ultimately leads to declining trade tax revenues ­ after all, free trade refers to a situation of no trade taxes, it does not have to lead to declining tax revenues. Three results from international experience are worth highlighting. First, recent econometric work suggests that high-income countries have been able to recuperate the revenue losses from trade taxes through domestic taxes; middle-income countries have made up most of the revenue losses; and low-income countries ­ which depend especially on trade tax revenues ­ have recovered at best only 30 percent. This striking difference in country performance across income groups is most likely related to the greater administrative capacity in high-income and middle- income countries to strengthen nonborder tax collections relative to low-income countries. Furthermore, potential domestic tax bases, income taxes in particular, are more easily tapped in high-income and middle-income countries, where formal, non- subsistent economic activity is more widespread. Second, the presence of a VAT does not in itself appear to enhance the ability to recover revenue. In contrast, what does seem to matter is the quality of the design and implementation of the VAT. In particular, having a VAT with a single rate and few exemptions increases the likelihood of revenue recovery. Third, countries that have successfully recovered the revenue loss from trade liberalization have often relied on both consumption and direct taxes to make up the shortfall. Overall, countries with a successful record in this area have: (1) carefully sequenced trade liberalization with domestic tax reform; (2) typically supported trade reform initiatives with improved customs and tax administration; and (3) been able to mobilize the necessary and continued political commitment to reform. Source: Based on Norregaard and Khan (2007) and Baunsgaard and Keen (2005). Box 8.2: Pakistan's Custom Duties Regime Key features of Pakistan's custom duties regime include the following: · Maximum duty rate of 25 percent · Continuous reduction within slabs (e.g., from higher to lower) · Creation of a 0 percent slab for specified primary raw materials · 5 percent rate applied mostly to primary raw materials · 10 percent rate applied to mostly secondary raw materials · Customs duty of 5 percent on plant, machinery, and equipment not manufactured locally for information technology, construction, tourism, and others · Customs duty rate of 0 percent on agriculture machinery · Exemption of crude petroleum oils, urea, and agricultural tractors · Various bilateral free trade agreements and membership in South Asia Free Trade Agreement · Complete harmonization of Pakistani customs with the World Trade Organization and the World Customs Organization. Firms in export processing zones are also exempt from customs duties for plant, machinery, and equipment and raw materials under certain conditions. There are exemptions at the 6-8 digit heading level of the Pakistan Customs Tariff. This is unusual and cumbersome from the perspective of administration and compliance. 112 8.2 Equity and Compliance 8.3. The incidence of customs duties, like the general sales tax and unlike the federal excise duties, is fairly proportional under the assumption that the consumer carries the tax burden. The calculation of the tax burden of custom duties follows the same steps as for sales and federal excise taxes. The bottom decile accounts for 3 percent of all custom duties paid and faces an effective tax rate of 1.46 percent; the top decile accounts for 33 percent of all custom duties paid and faces an effective tax rate of 1.49 percent (Figure 8.3, left panel). Figure 8.3: Equity and Compliance for Custom Duties Actual Tax Collection and Tax Gap of Custom Duties Incidence of Custom Duties by Household Consumption Deciles 35 1.8 Effective Tax Rate (% of GDP) Gap (21%) 1.6 30 1.4 25 1.2 20 1 15 0.8 Actua l (100%) 0.6 10 0.4 Share in Total Taxes (%) 5 0.2 0 0 Bottom 2 3 4 5 6 7 8 9 Highest 8.4. The estimation of the tax gap for custom duties relies on comparing foreign export values to Pakistan with Pakistani import values. Imports to Pakistan from, say, China are not just recorded by Pakistani custom officials at the import stage, but also by Chinese custom officials at the export stage. This allows us to verify the import values from China reported in Pakistan with the export values to Pakistan reported in China. But why would Chinese trade statistics be more accurate than Pakistani trade statistics? Because of an asymmetry in custom duties ­ they are typically imposed on imports but not on exports. Hence, traders face an incentive to underreport the dutiable value of imports to avoid paying duties, but there is no corresponding incentive to misreport export values with regard to custom duties. The import duty gap is derived on a three-step procedure based on trade information between Pakistan and its leading trade partners. First, we take 6-digit commodity level export data to Pakistan in 2004/05 from the United Nations Comtrade database for China, Germany, India, Japan, Malaysia, and the United States. They account collectively for about 50 percent of Pakistan's import values. Then, we calculate effective duty rates at the 6-digit commodity level for each of these country based on 2004/05 FBR data. Finally, we estimate potential tax revenues by multiplying reported export values in our sample by the corresponding effective duty rates, and compare this number to actual tax revenues from custom duties. To obtain an overall import duty gap across all of Pakistan's trading partners, we weigh each country's estimated duty gap by the corresponding country's share in total trade with Pakistan. 8.5. The estimated import duty gap is 21 percent of actual duty revenues, or Rs. 32 billion in 2007/08. This gap provides useful information by commodity and country for allocating enforcement resources (Figure 8.3, right panel). This import duty gap is likely to be an underestimate of the `true' import duty gap, as it accounts only for customs duty evasion through underreporting of import values. Two alternative methods are not captured: the smuggling of goods (which would not be reflected in the official data also of the exporting country) or the misclassification of goods (which puts imports of a given value into a lower custom duty rate category). 8.6. The import duty gap varies widely from country to country. There are two countries with a positive duty gap. It equals 83 percent of actual customs revenue for China and 13 percent for Germany. In contrast, the estimated potential import duties for the United States, India, Japan and Malaysia are less than the estimate of actual import duties, although the gaps are relatively small in percentage terms. One explanation could be that firms operating in countries with high corporate tax rates, like the United States and Japan, may have an incentive to overstate the value of exports in order to shift income abroad for tax purposes. Similarly, firms operating in countries with a value added tax, like Malaysia, may have an incentive to overstate the value of exports to exaggerate value added export refunds. 113 8.3 Policy Options 8.7. In the short term, proceeding prudently with any further reduction of customs tariffs is desirable. Tariff reduction leads to revenue losses as long as the government is no able to compensate these revenues from domestic tax sources. While continuing with carefully paced tariff reductions, Pakistan could consider a sequence of separate measures to avoid overall revenue losses (Table 8.2). It could initially eliminate existing tariff exemptions in the system; then increase excise rates on excisable imports to balance out tariff reductions; and finally adjust the scope and rate of the general sales tax to meet remaining revenue needs. 8.8. In the long term, FBR is set to focus on trade facilitation rather than revenue generation, with the expectation that lower tariffs will spur economic growth and revenue gains elsewhere in the economy. This policy will help reduce the differences in the effective rates of protections of different sectors of the economy, reduce distortions in the allocation of economic resources, and lead to increases in economic welfare. A progressive substitution of customs tariff collection with revenues from domestic taxes could also make the tax system more progressive. One reform priorities lies in further reducing tariff dispersion, by reducing the number of rates outside the regular tariff bands and eliminating special exemptions. For example, in 2007/08, some 471 tariff lines out of the overall 6,774 tariff lines were above 25 percent. 8.9. In addition, Pakistan might want to consider altering its tariff schedule in the direction of more uniform rates. Right now, lower rates apply in general to raw materials and higher rates on finished goods. Such escalated tariff structure leads to effective rates of tariff protection that are much higher for final goods than the nominal tariff rates indicate. A uniform tariff applied to all imports would remove differences in effective protection rates within the import substituting sector and provide broad based protection to the entire manufacturing sector. Infant industry protection would be replaced by infant economy protection providing equal treatment for all activities engaged in import substitution. More uniform tariffs also reduce the pressure of special interests lobbying for preferential tariff treatment. This reduction in rent seeking opportunities would not only save resources directly due to the lower reward for lobbying activity but also result in less opportunities for corruption in the customs offices and elsewhere. Misclassification of imports at the border entry would no longer be an issue and customs clearance procedures would be simplified and less costly due to the need for fewer inspections. What might a uniform tariff for Pakistan look like? Presumably, it should be revenue neutral as it would be difficult to make up any lost tariff revenue from other revenue sources. The current tariff regime generates revenues of 1.4 percent of GDP. Dutiable imports as a share of GDP are around 10 percent, so a revenue neutral uniform tariff would be around 14 percent (1.4/0.10). Switching from the current to a uniform tariff schedule is obviously a large policy change. An interim goal would be to reduce fine distinctions between commodities at the 8-digit level, as duty rate differentials of this magnitude create incentives for firms to misclassify imports to avoid paying duty. Table 8.2: Policy Reform Options for Custom Duties Topic Proposals Tariff Levels · Combine tariff reductions with a sequence of separate measures to avoid overall revenue losses, including elimination of existing tariff exemptions (especially at the 6-8 digit level), increases in excisable imports; and expansion of the yield from general sales tax Uniformity of Rates · Reduce tariff dispersion, especially at disaggregated level Compliance · Cooperate with trading partners to ensure accurate trade flow records 114 CHAPTER 9: PROVINCIAL TAXES 9.1 Introduction 9.1. Pakistan's intergovernmental fiscal system is out of balance. Provincial governments spend about sixteen times as much as they collect in terms of taxes. If we add in other provincial own-source revenues, provincial spending is still six times higher than provincial revenues (Figure 9.1, left panel). It is doubtful that local residents see much connection between the level of taxes they pay their provincial government and the expenditure benefits they receive. This means that the government misses out on one of the most important advantages of fiscal decentralization ­ taxpayers holding their elected provincial officials accountable for the quality of services delivered. 9.2. A second dimension of fiscal imbalance at the provincial level is that the tax revenue effort compared to the size of the tax bases. While provincial tax bases include the consumption of services, property, motor vehicles, agriculture and professions, the provinces obtained in 2007/08 no more than 0.4 percent of GDP in tax revenue, equivalent to only 4 percent of national tax collection (Figure 9.1, right panel). This mismatch arises primarily because the intergovernmental fiscal architecture provides little incentive for the provinces to collect their own taxes, and because the skills of the provincial tax administration are inadequate for their hard-to-collect tax bases. 9.3. This chapter reviews the status of revenue mobilization by provincial governments in Pakistan and assesses structural reform options that might yield higher revenues. It is based on case studies of Punjab and North-West Frontier, two provinces that reflect a good deal, but certainly not all, of the rich provincial diversity in Pakistan. The main conclusion is that tax reforms in both provinces have the potential to increase provincial tax collection from 0.4 percent of GDP to 1 percent of GDP, in line with government objectives. Figure 9.1: Provincial Revenue Composition and Trends Resources for 2007/08 Provincial Expenditures (%) Provincial Tax Revenues (% of GDP and % of National Tax Revenues) 0.60 6 Percent of GDP Percent of National 0.55 Tax Revenues 0.50 5 0.45 0.40 4 0.35 Federal Transfers, Grants and Loans (82%) 0.30 3 0.25 0.20 2 0.15 0.10 1 0.05 Other Prov. Rev. (12%) 0.00 0 Provincial Taxes (6%) 1999/2000 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 115 Table 9.1: Major Findings of Chapter 9 Topic Findings Tapping · Punjab and NWFP run large budget deficits in 2006/07 (about 1.5 and 3.5 percent of provincial GDP, Provi ncial respectively). Tax Bases · In 2007/08, the share of federal transfers, loans and grants was over 80 percent of all receipts for Punjab and just under 75 percent for NWFP. · Own-source revenue mobilization is weak due to a hard-to-reach and inelastic tax base, weak tax administration, large federal transfers, federal encroachment on provincial tax bases, and low political will to tax provincial constituencies. Sales Tax on · Provincial governments levy the sales tax on services, and collected in 2005/06 24 Rupees per capita Services in Punjab and 20 Rupees per capita in NWFP. · The collection is far below potential because of political opposition to tax increases, the headquarter problem, weak tax administration, and some federal encroachment on provincial tax bases.. Motor · Punjab and NWFP collected 46 and 32 Rupees per capita in 2005/06 in motor vehicle taxes. Vehicle Tax · Major issues include clarifying the underlying goals of motor vehicle taxation; the low level of collection; leakages in the tax base; and weak administration. UIPT · Punjab and NWFP collected 12 and 14 Rupees per capita in 2005/06 in Urban Immovable Property Tax. · Major issues include low revenue yield and growth due to the narrowness of the tax base, tax preferences and undervaluation of properties. In addition, the distribution of the tax burden is inefficient in Punjab as it falls more heavily on improvements than on land, there is ambivalence about whether UIPT is a local or provincial tax, and the state of administration is poor. Land and · Punjab and NWFP collected 125 and 32 Rupees per capita in 2005/06 in land and property transfer Property taxes. Transfer · Major issues include low and inelastic tax collection, the undervaluation of property transfers in both Taxes urban and rural areas, large exemptions, and poor record keeping. Agricultural · Punjab and NWFP collected 7 and 3 Rupees per capita in 2005/06 in agricultural income tax. Income Tax · Major issues include low revenue yield due to large exemptions in Punjab, an inequitable tax structure that makes no distinction by crop type, and lack of enforcement due to weak administration.. Professional · Punjab and NWFP collected 3 and 4 Rupees per capita in 2005/06 in professional tax. Tax · Major issues include the unclear goal of the tax, low revenue yield, and inadequate staffing and automation in tax offices.. Other Taxes · Punjab and NWFP collected 17 and 8 Rupees per capita in 2005/06 in other taxes. · Major issues include unclear rational for other taxes, low revenue yield and weak enforcement. Vertical · The overall impact of provincial taxes on the income distribution is mildly progressive. The effective Equity tax rate of provincial taxes in 2004/05 increased moderately from 0.32 percent of GDP for the bottom decile to 0.36 percent for the sixth decile, and then rose more sharply to reach 0.54 percent for the top decile. 116 9.2 Tapping Provincial Tax Bases Structural Fiscal Deficits 9.4. Economic activity, tax bases and expenditure needs in Punjab and NWFP differ starkly. Per capita GDP in Punjab is close to one fifth higher than in NWFP (Figure 9.2, left panel). Punjab also has higher household income and lower poverty than NWFP. The tax base in Punjab is larger, more urban, and easier to reach than in NWFP. The potential for raising tax revenues would thus appear to be greater in Punjab than in NWFP, while expenditures needs ­ at least from a social and poverty perspective ­ would appear to be greater in NWFP than in Punjab. Nevertheless, tax collection is a steep challenge in either province: both provinces are marked a large informal sector, comprising small businesses and agriculture and accounting for close to three-quarter of total employment, and by stark inter-district disparities. 9.5. Punjab and NWFP face recurrent structural fiscal deficits and rely heavily on borrowing and postponing planned spending to close the gap between actual 2revenues and expenditures. A recasting of the 2006/07 provincial budgets reveals that both Punjab and NWFP carry large current deficits that are financed by borrowing and drawing down reserves. Punjab's overall deficit, including capital expenditures, stands at Rs. 78 billion, and NWFP's at Rs. 27 billion. According to provincial GDP estimates (Box 9.1), the budget deficit amounts to 1.5 percent of GDP in Punjab, and 3.5 percent of GDP in NWFP. Provincial borrowing finances 44 percent of Punjab's deficit and 66 percent of NWFP's deficit (Figure 9.2, right panel). The rest, the uncovered deficit, is covered by a drawdown of existing balances which have accumulated in part because of unfilled positions and low utilization rates of development projects. Thus far, these balances have been adequate to cover the fiscal deficit, but there is no guarantee that they will continue to do so in future. In addition, this practice violates basic principles of sound public expenditure management and harms the development impact of public spending. Clearly, bringing provincial finances onto a sustainable footing calls for significant restructuring of the tax system. Figure 9.2: Economic Indicators and Budgetary Position in Punjab and NWFP Selected Economic Indicators in Punjab and NWFP (2007 or Latest Available) Budgetary Position of Punjab and NWFP in 2006/07 (%) 110 100% Wheat Trading Punjab Uncovered Operations 100 Deficit 90% Capital Financed 90 Expenditure Deficit NWFP 80% 80 70% 70 60% 60 50% 50 Current General Revenue 40 40% Revenue Expenditure Receipts 30 30% 20 20% 10 10% 0 Population (Million) Per Capita GDP Per Capita Poverty (% of Urbanization (% of 0% (Pakistan=100) Consumption Population) Population) Punjab NWFP Punjab NWFP (Pakistan=100) RESOURCES EXPENDITURE Box 9.1: Provincial GDP Estimates There are no official provincial GDP estimates in Pakistan. However, provinces have calculated their own provincial GDP, using sectoral data and assumptions about regional shares and growth rates relative to national GDP. As these calculations give only rough approximations of provincial GDP, this chapter uses provincial GDP solely for level comparisons in a given year, such as, what is the size of tax collection relative to economic activity. For comparisons over time, the chapter looks primarily at real per capita tax collection, defined as provincial tax collection divided by the provincial population and adjusted for inflation. 117 Revenue Assignment and Structure 9.6. Revenue assignment in Pakistan is derived from the constitution and certain statues. A revenue source can be assigned to the federation or a province by an express provision in an article of the constitution or through classification in the federal or the concurrent lists of subjects (Table 9.2). The tax structure and administration in the provinces is similar in many respects due to central influence and the adoptions of Punjab tax laws by other provinces. Provinces have assigned some of their revenue sources to local governments through statues like the 2001 Local Government Ordinances. 9.7. Provincial tax revenues are derived from only a few taxable sources. While there are more than a dozen tax sources available to provincial governments, most tax revenues in Punjab and NWFP come from property transfers, motor vehicles, and sales tax on services (Figure 9.3, left panel). As we will argue in this chapter, a tax- by-tax analysis reveals that the current tax structure does not reflect the full revenue potential of all taxable sources. The provinces have access to some taxes that have broad enough bases and potentially enough built-in growth to form a more revenue productive tax system. 9.8. The provinces rely heavily on intergovernmental fiscal transfers from the National Financial Commission Award to fund expenditures. While neither Punjab nor NWFP excel in mobilizing resources from their own tax base, the per capita tax collection in Punjab was about twice as large as in NWFP in 2007/08. At the same time, NWFP fares better in own-source non-tax revenues due to profits from hydro electricity. But the by far most important revenue source is federal transfers, grants and loans, the bulk of which is distributed on equal per capita basis. Transfers accounted in 2007/08 for 81 percent of all provincial revenues in Punjab, and 74 percent in NWFP (Figure 9.4). As a result, the level of per capita provincial government expenditures is almost the same across the provinces (Figure 9.3, right panel). As federal payments increased over the last years, provincial tax collection fell behind (Figure 9.5): in Punjab, the share of provincial taxes in overall provincial revenues declined from 12 percent in 2003/04 to 6 percent in 2007/08; and in NWFP, from 4 percent in 2003/04 to 3 percent in 2007/08. Per capita tax collection in Punjab fell in real terms from Rs. 300 in 2004/05 to only Rs. 230 in 2007/08, while it stagnated around Rs. 120 in NWFP. 9.9. The limited role of provincial governments in national tax collection is unusual by international standards, although intergovernmental fiscal systems vary greatly from country to country. The low reliance of Pakistan's provincial governments on their own sources of revenue is also the model of choice in countries such as Russia, Indonesia and China. Countries such as Argentina, Brazil and India follow the more conventional advice and decentralize significantly more revenue raising power to their state governments. Similarly, the assignment of expenditure responsibilities to the provincial level differs across countries. Pakistan's expenditure decentralization of around 30 percent is not so unusual by comparison with some other large federations. Nevertheless, the combination of the large provincial expenditure share and the low provincial own-source revenue share in Pakistan stands out from international perspective. Figure 9.3: Composition of Tax Revenue in Punjab and NWFP Provincial Tax Composition for 2005/06 (%) Per Capita Provincial Revenues and Expenditures in Punjab and NWFP in 100% 2007/08 (Rs. Per Capita) Other Indirect Taxes 4,000 90% 3,600 80% Stamp Duties 3,200 70% INDIRECT TAXES 2,800 60% GST on Services 2,400 50% 2,000 Motor Vehicle Tax 40% 1,600 30% Other Direct Taxes 1,200 NWFP Land Revenue 20% DIRECT 800 Registration Fee TAXES 10% (Transfer of Property) 400 Punjab Urban Property Tax 0% 0 Punjab NWFP Expenditures Other Provincial Revenues Provincial Taxes 118 Table 9.2: Federal, Provincial and Local Revenues Federal Provincial Local 1. Personal income tax (except agr. income) Excise duty on alcohol, liquor, narcotics Tax on the transfer of 2. Corporate income tax Sales tax on services immovable property 3. Customs Tax on professions (Tehsil and Town 4. Sales tax goods Motor vehicle tax Councils) 5. Excise duty (except on alcohol, narcotics) Property tax Other taxes and fees 6. Capital value tax Capital gains 7. Estate duty Agriculture income tax 8. Mineral oil, minerals, natural gas Stamp duty 9. Tax on production capacity Registration fee 10. Terminal taxes on goods transport & passengers Mutation fee 11. User charges on federal subjects Natural gas excise duty 12. Net hydro profits 13. Electricity duty 14. User charges Figure 9.4: Revenue Composition in Punjab and NWFP Punjab's Revenue Composition (%) NWFP's Revenue Composition (%) 100% 100% 90% 90% 80% 80% 70% 70% Federal Federal Transfers, 60% Transfers, 60% Grants Grants and Loans and Loans 50% 50% 40% 40% 30% 30% 20% Other 20% Provincial Other Revenues Provincial Revenues 10% 10% Provincial Provincial Taxes Taxes 0% 0% 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 Figure 9.5: Government Spending and Own-Source Revenues in Punjab and NWFP Provincial Tax Collection as Percent of Provincial Revenues Per Capita Federal Transfers and Provincial Taxes (2007/08 Prices) 14 3,600 Punjab 3,300 12 NWFP Federal 3,000 2,700 10 2,400 8 2,100 Punjab Federal 1,800 6 1,500 NWFP 1,200 4 900 600 NWFP Taxes 2 Punjab Taxes 300 0 0 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 119 Low Revenue Mobilization 9.10. A number of reasons account for the weak provincial performance in tax collection. First, provincial taxable capacity is low and the tax base is hard to reach. NWFP is the poorest province in Pakistan and has a high concentration of poverty. Per capita GDP in Punjab is higher, but there also is a high concentration of poverty and a large informal sector. Even so, many would argue that the economic base is strong enough to raise more provincial tax revenues than the about 0.4 percent of provincial GDP that were collected in Punjab and NWFP. 9.11. Second, the tax administration machinery is ineffective. Both provinces are plagued by incomplete and out of date records, suggesting that there is not a good sense of the true tax base. Most tax subjects have not been recently surveyed, hence tax bases are understated, and most of the recordkeeping systems are manual. In NWFP, revenue collections, except for some excises, take place only in urban areas (Khan 2004). For example, there are 24 districts in NWFP but 70 percent of all property tax collections are from Peshawar. Over 50 percent of excises come from 7 districts. In addition, about one-third of the area of NWFP is made up of provincially administered tribal areas. Only a few taxes (e.g. tax on the transfer of property, stamp duty and the local rate) are collected in these districts. The urban immobile property tax is collected in only 16 of 24 districts. In Punjab, the tax administration does not tap effectively the significant taxable capacity in urban areas. For example, property values have grown but property tax collections have not; the number of motor vehicles has grown but motor vehicle tax revenues have not kept pace. The province has given away much of its tax base in the form of preferential treatment. In addition, underassessment is considerable and collection rates are low. The same applies in Punjab's rural areas. Both provinces are not able to reach agriculture for collecting agricultural income tax or property transfer taxes. 9.12. Third, the structure of taxes prevents increases in response to a growing economy. The tax structure is partly based on specific rates, and governments do not take discretionary action to increase the nominal rates or revaluate property. In addition, some faster growing components of the tax base are not taxed or are given exempted or given preferential treatment, such as owner-occupied property, industrial property and vacant land, and the consumption of services. 9.13. Fourth, increases in intergovernmental transfers from the center have been large enough to allow a slowing of the effort exerted to collect provincial taxes. The increase in central government assistance has been significant in both provinces while the growth in own source revenues has been nearly flat. This pattern should come as no surprise. There is no incentive built into the transfer formula that would reward provinces for increasing their tax effort, or penalize them for not doing so. We will take up this issue at the end of this chapter. 9.14. Fifth, the federal government has encroached on provincial tax bases. Many provincial officials argue that the federal government has restricted too far the provincial revenue space. Some of these limits are constitutional, but there are also limits imposed by federal government policy. Examples include the following: · The only tax that provinces are exclusively empowered to levy is the tax on professions, trades and callings. · The urban immovable property tax is a provincial government tax, but most of the revenue collected is assigned to local governments. · The federal government requires a mandatory collection of a federal presumptive income tax at the time of vehicle registration, which lowers the compliance for the motor vehicle taxes. · The federal government imposes a 2 percent capital value tax on property transfers, raising the total rate on each transfer and arguably reducing the rate of compliance with provincial stamp duty and registration taxes. 9.15. Finally, elected local officials are hesitant to increase the tax effort for fear of losing political support. There have been no increases in tax rates or expansions in tax bases for 8 years in NWFP and 5 years in Punjab. The provincial government in Punjab postponed in 2007 the introduction of the new property tax valuation roll, in part because of upcoming elections. Enforcement is lax in both provinces. Politics has been perhaps the major reason why provincial tax structures have not developed. Powerful provincial interest groups direct politicians not to increase taxes, and instead lobby the federal government to provide increased federal transfer payments. 120 9.3 Sales Tax on Services Base and Rate 9.16. Provincial governments levy taxes on a broadly uniform list of services. The constitution provides for federal taxation in the case of "taxes on the sales and purchases of goods imported, exported, produced, manufactured or consumed". This leaves open the possibility for provincial governments to levy a sales tax on services. While the constitution allows each province to determine which services it will tax, each province's law were drafted by the federal government, adopted as ordinances, and include a more or less uniform list of services to be taxed. 9.17. The base is the gross amount of sales, and the tax rate is set by the federal government at the national rate of 16 percent. Service providers are entitled to claim input tax credit for the tax paid on taxable purchases, inputs, and utilities. The sales tax on services yielded in 2005/06 7.5 and 9.2 percent of own revenues in Punjab and NWFP, respectively (Figure 9.6). This makes it a considerably more important revenue source than the property tax, but a considerably less important source than either taxes on property transfers or motor vehicle taxes. Sales tax collections are concentrated on four services: hotels and restaurants, travel and shipping agents, advertisements, and courier services. Punjab (40 percent) and Sindh (58 percent) account for almost the whole collection of sales tax on services (Figure 9.7). Figure 9.6: Sales Tax Collection on Services in Punjab and NWFP in 2005/06 Provincial Sales Tax Collection in 2005/06 (Rupees Per Capita) Sales Tax Collection in 2005/06 in Punjab and NWFP 16 28 10 Punjab % of Own-Source 14 Rupees Per Capita Revenue 9 24 12 8 20 7 10 6 8 16 5 6 12 4 4 NWFP 3 8 2 2 4 0 1 Hotels, restaurants, Travel agens, TV and radio Couriers Services Other 9 Services fast food, catering stevedores, advertisements shipping agents, 0 0 ship chandlers Punjab NWFP Figure 9.7: Provincial Composition of Sales Tax Collection from Service in 2005/06 Provincial Shares in Sales Tax Collection from Services in 2005/06 (%) NWFP Balochistan Punjab Sindh 121 Issues 9.18. There are five main issues. First, the revenue potential from a sales tax is not tapped due to lack of political will. The national accounts in Punjab classify 10 percent of provincial GDP is classified as "other services", which excludes transport, storage and communications. If this rough estimate of the broad version of the base for the sales tax on services is indicative, we can say that the effective tax rate at present is well less than one percent. However, provincial governments have to be willing to expand their existing tax base to include additional services. Yet, there is little incentive for political leadership to increase tax efforts as a result of growing federal transfers and political opposition to tax increases. 9.19. Second, the headquarters problem complicates a provincial service tax. The tax may be paid at the headquarter location rather than at the point of consumption. For example, Punjab accounts for close to 60 percent of Pakistan's GDP but only 40 percent of total provincial sales tax collections on services. Sindh, home of the national gateway of Karachi port and commercial hub, accounts for 30 percent of national GDP but close to 60 percent of provincial collections of the sales tax on services. While the headquarters problem cannot be fully addressed, it is accentuated in systems where provinces determine the base and rate of the sales tax on services. 9.20. Third, the current tax administration arrangement also hinders a more revenue productive sales tax. The services sector is notoriously hard to tax because it is composed of many small firms that are difficult to identify and maintain often weak records and accounts. Sales tax assessment and collection is the responsibility of the federal government, which retains a 2 percent collection fee. In view of its low revenue share, the central government might find it too costly to aggressively pursue tax collection. The deeper the province decides to go into the services base, the harder to assess will be the base, and the more costly to the federal government will be the administrative effort required. For example, the Punjab government proposed to expand the tax base to include marriage halls and beauty parlors but the central government was unwilling to collect from these sectors. The split between administrative responsibility and base determination is not likely to work effectively unless there is significant revenue sharing. The levying government (province) will accept the political risk only if the revenue rewards are great enough, and the collecting government (federal) will not allocate the resources necessary to administer the tax properly if the revenue benefits do not justify taking on the additional costs. While the correct split of revenues to accomplish the goal of "adequate" revenues for provincial and federal governments has not been worked out, it seems clear that the present 98 percent/2 percent division is not the answer. A separate administration problem is that there is insufficient sharing of information about tax assessment and collection from the federal government with the provincial governments. This makes it difficult for provincial governments to move forward with base expansion, and to use their comparative advantage of "familiarity" to identify new firms for inclusion in the net. 9.21. Third, the federal government also levies a sales tax on selected services, in the form of excise taxes. While the federal government sees this as a matter of national interest, the provincial governments see it as encroachment on their constitutional base. In 2005/06, the total yield from federal taxation of 10 services was Rs 28.2 billion compared to only Rs 4.2 billion from provincial taxation of 13 services. Most of the federal revenue (95 percent) is collected from telecommunications services. Paradoxically, the provinces would be better off in revenue terms if the sales tax on services was a federal tax with revenues included in the divisible sharing pool for federal transfers. 9.22. Finally, some observers argue that a 16 percent tax on services could hurt part of the economy. For example, some hotels and places of entertainment might see the tax as significantly cutting into their profit margin, or consumers might find their consumption of services more expensive to the extent the tax can be shifted to them (Box 9.2). Should these concerns cause a rethinking of the proposals to extend the sales tax on services? The answer here is that it should not. Services should be taxed, just as other consumption is taxed. Services have enjoyed a preferential tax treatment in Pakistan for many decades at the cost of higher taxation of goods. Why should the sales tax treat the purchase of an appliance any differently than an expenditure on a restaurant meal? Leveling the playing field across goods and services could boost efficiency and horizontal equity by lowering the distortions of the tax system on the allocation of resources among sectors. 122 Reform Options ­Provincial versus Federal Assignment 9.23. A comprehensive reform of the sales tax on services could take several directions, ranging from decentralization to centralization. The incremental solution is to expand the base by bringing more services into the tax net. This extension should begin by including those services that could most easily be reached under current administrative practice (Box 9.3). However, continuing with the present system does not address the disconnect between who administers the tax and who gets the revenue. Nor does it allow a province to easily ratchet up collections if it should choose to do so. The likely result is that there would not be much revenue enhancement under this reform option, especially if this proposal is not accompanied by a national incentive package to stimulate or reward provincial governments for increasing revenue efforts. If the incentive strategy was successful, Punjab and Sindh, with their more developed service sectors, would likely emerge as revenue gainers. 9.24. The second option is to transfer the ten services that are presently taxed by the federal government to the provinces, and make provinces in charge of sales tax collection on services. Assuming a similar revenue effort as today, this would increase provincial revenues from the service tax by a factor of more than seven ­ assuming the provinces administer the tax as well as FBR. The next step would be for the provinces to begin expanding their tax base selectively by bringing the easier to reach services into tax. The revenue yield from this option depends on how quickly the provinces learn sales tax administration, and how quickly additional services are brought into the tax net. Nevertheless, this approach has a number of drawbacks. One disadvantage is that the "easier to tax" services with a broad base are likely to come away with a much heavier tax burden than other services. Second, the proper accounting for cross-province sales of inputs would be particularly difficult for the province tax authorities given their limited experience with the sales tax. Third, the federal government would almost certainly resist surrendering this taxing power, in particular if it might reduce the domestic resource mobilization compared to the current system. Finally, as discussed below, most importantly, moving tax administration from federal to provincial level might incur revenue losses due to the weaker administrative capacity at the provincial level. 9.25. The third reform option is to convert the sales tax on services to a shared tax with the federal level. The federal and provincial governments would tax the same services, with assessment and collection carried out by the federal government. There would be a separate federal and provincial government tax rate (not to exceed 16 percent). The services to be taxed would have to be the same under such a regime, and the federal rate would be the same in all provinces. The provincial tax rate could vary according to the decision of the provincial government. This autonomy in provincial rate setting would preserve an important fiscal decentralization feature, even though it could introduce some administrative complication. There might be restrictions here, e.g., the federal rate might be set at 8 percent and the provincial rate would be limited to 8 percent. This approach combines the rate setting discretion of provincial governments with the superior administration capacity of the federal government. One drawback is that it could accentuate the headquarters problem. Particularly Sindh could follow a strategy of taxing services heavily and exporting its tax burden to other provinces. 9.26. The fourth option is to fold the sales tax on services into the federal general sales tax on goods and services and share the revenue gains between federal and provincial governments. As discussed in Chapter 2, this is the preferred tax reform option, as it would deliver the largest revenue gains, while also bolstering the efficiency and equity of the national tax system. The revenue sharing could be resolved in different ways. In particular, the tax revenues from services could be handled in the same manner as from goods. The federal collections on goods are part of the NFC revenue sharing pool, of which currently 43.5 percent accrue to the provinces. This outcome could also be achieved by taxing services through federal excise duties in VAT mode. But, of course, a different revenue sharing formula could be agreed in view of the potential increase in the revenue base. By comparison with the present system, the province would do better in terms of revenues. At the same time, the federal government would have a greater incentive to collect this tax than under the current system, as over half of the receipts would accrue to the center. The federal tax administration could also exploit synergies in the sales taxes on goods and services in order to strengthen overall GST collection. With both provinces and the federal government gaining revenues, this could be a win-win situation. The main drawback of this approach from the provincial point of view is that provinces give up the formal exclusive claim on services as provincial tax base. 123 Reform Options - Administration 9.27. An important consideration for all reform options is that the tax base will become increasingly difficult to reach as additional services are added. Under the current arrangements, the central government might be unwilling to expand its administrative efforts so as to reach those smaller firms and self-employed individuals who provide services. One solution might be to transfer administrative responsibility for the services tax to the provincial government. Provincial governments have some comparative advantage in administering the sales tax on services. The province is well placed for the identification of liable tax payers, and the maintenance of the tax roll because of its greater familiarity with the local economy. Since the sales tax on services is a provincial revenue source, there would be more incentive to assess and collect the tax than is the case under the present centrally administered system. There might also be some significant advantage to a coordinated collection of the sales tax on services, the professions tax and the urban property tax at the provincial level. However, a centrally administered system also has some key advantages. There are strong synergies between in the production and delivery of goods and services at the company level, which would make the whole chain of registration, filing, assessments, audit, and litigation more effective. In addition, federal tax administrators are already familiar with sales tax concepts. The central government might also find it easier than provincial government to enforce the tax where powerful local interests are involved. On balance, the advantages of a centralized administration clearly outweigh those of a decentralized administration, at least until the provincial tax administration have developed adequate capacity to handle sales tax collection effectively. Until provincial administrative capacity matches the federal standards, devolving administrative powers to the provinces is likely to entail lower revenue gains or even outright revenue losses. 9.28. The conversion of the sales tax on services into some form of a shared tax, while retaining the collection responsibility with the federal administration, could ensure better collection. If federal revenue retention from collections was high enough, this is likely to result in a more aggressive administrative effort. This proposal would also tackle the issue of federal encroachment on the provincial sales tax base leads to a competition for the same revenue base. The provinces feel that they have a constitutional right to the tax base, and would like to have the federal government give it an exclusive on taxing services. The federal government sees its right to tax services when in the national interest, and when the headquarters problem makes provincial taxation infeasible. The middle ground is a shared tax, where each side has a claim on this tax base. 9.29. A federally administered general GST on goods and services that is fairly shared with and within provinces would also resolve to a large extent the headquarters problem. Countries have tried various approaches to the interprovincial competition issue that surrounds the headquarters problem. One step that might be taken is to allocate all final sales for a company across provinces. If this can be done, the proration can be used to deal with the headquarters problem. However, this is administratively difficult, involves a good deal of subjectivity in establishing an allocation formula, and increases the compliance costs on companies. This approach is used, for example, in the allocation of the state government corporate income tax base across states in the US. Second, each company could be asked to report its taxable sales by province, but this could significantly increase compliance costs and administrative costs. Third, those components of the service tax base where the headquarters problem is particularly serious, could be assigned to the federal tax base, and a revenue sharing scheme could be worked out. A fourth approach is to introduce a tax effort component into the NFC award, but to eliminate "headquarters taxes" from the tax effort computation. In this case, the federal government would be responsible for identifying "headquarters taxes", i.e., provincial taxes whose burdens are largely exported. None of these proposals could be implemented without a great deal of arbitrariness. However, with a unified general GST on goods and services and a fair revenue sharing formula, the headquarters problem would be largely controlled. 9.30. Finally, regardless of which reform option is chosen, it will be important to harmonize provincial and federal sales tax legislation. This includes, among others, clarifying rules regarding the place of supply (Box 9.4). 124 Box 9.2: Incidence of Sales Tax on Services Would revenue increases from the sales tax on services be feasible in terms of the burden it would impose on taxpayers in the provinces? To get a rough indication of the likely tax incidence, we estimate of the distribution of the burden of the sales tax on services using 2004/05 PSLM data on the consumption of personal and professional services by households at various income levels. The figures below compare the distribution of income across population deciles with the distribution of consumption of personal services. For example, the top 10 percent of the households in NWFP account for about 44 percent of all household income, but only about 36 percent of consumption of personal services. The distribution shows that the lower income groups allocate a greater share of their budgets to services. This suggests that the GST on services is a mildly regressive tax. The same holds in Punjab. However, poorer households are more likely than richer households to consume personal services provided from retailers that are either exempt (due to the GST threshold) or provided in the informal sector. Although these effects could not be modeled, they could make the effective tax burden either neutral or even progressive. Distribution of Household Consumption and Expenditures on Personal Services (%, 2004/05) 35 30 25 20 NWFP Personal Punjab Household Services Consumption NWFP Household Punjab Personal 15 Consumption Services 10 5 0 Bottom 2 3 4 5 6 7 8 9 Highest Box 9.3: Expanding the Provincial Sales Tax Base While there is no easy way to identify the "best" candidates for inclusion in the tax base for the sales tax on services, one could follow a five-step procedure which would require in all likelihood new provincial ordinances. · Charge the federal government with the responsibility for compiling a comprehensive list of services, perhaps by reference to the categories listed in the standard industrial classification. · Charge the provinces with responsibility for identifying those services to be brought into tax and assisting the federal sales tax department in building the tax roll. Provinces might start this process by determining whether a useful establishment survey exists or could be carried out. Other files that might help in identifying firms for the sales tax base are UIPT records, electricity duty payments, and professional registries. An issue to be resolved is whether all provinces would need to tax the same set of services. Under a provincial administration, each province would be able to choose its own base from among the list of eligible services. Under a federal administration, there might be more pressure for uniformity. · Each province should work with the federal sales tax administration to identify a size threshold for inclusion in the services tax net, and then estimate revenue yield. · The federal sales tax department in each province should develop comprehensive reports on assessment and collection detail, and delinquency lists. These reports should be shared with the provincial governments. Pasha (1995) carried out an exercise of identifying sectors for inclusion in the sales tax base for all provinces, and Malik (2004) more recently for NWFP. Box 9.4: Place of Supply and the Taxation of Services Services taxed under provincial ordinances are intended to be `located' in a particular province on an origin basis, where the origin is the place where the services are `rendered or performed.' However, exactly what is meant by place of performance, or place where rendered is not clear. For example, television and radio advertisements are taxed, but is the place of taxation determined by the location of the supplier, the location of the recipient, or the place where the services are physically performed? If the place of performance is the determinant, where are television advertisements considered to be performed? These issues need consideration: · FBR distributes the provincial sales tax to the provinces after deduction of a 2 percent fee. Does FBR distribute to each province the revenue collected on sales tax collected on services performed in that province or according to another formula? · The provincial sales tax as input tax can be credited against federal sales tax as output tax; and the federal sales tax as input tax be credited against provincial sales tax as output tax. How do these tax credits impact the revenue distribution to the provinces? · Were the provinces to take back administrative responsibility for sales tax on services, would the collection be based on origin or destination basis? 125 9.4 Motor Vehicle Taxes Base and Rate 9.31. Revenues from motor vehicle taxes are low. There are two kinds of taxes in the present system: a one- time registration charge and an annual token tax. In both NWFP and Punjab, about two-thirds of motor vehicle revenues are collected from the token tax. In terms of per capita revenue generation, motor vehicle tax receipts are small: in 2005/06, they equaled only about Rs. 32 and Rs. 46 per capita in Punjab and NWFP, respectively, and accounted for about 14 percent of own source revenues and no more than 0.1 percent of provincial GDP in both provinces (Figure 9.8). Revenue collection in Punjab has picked up since 2001/02 on the back of rising car ownership and upgrading of cars in terms of value and engine capacity. However, it has stagnated in NWFP due to a combination of weak tax administration, registration in other provinces and widespread exemptions. 9.32. The registration tax is a relatively simple levy that is collected only once at the time of initial purchase, and the tax rate varies by type and age. It is an ad valorem sales tax that is assessed according to the value of the new motor vehicle and its engine capacity. No other additional registration tax is levied, whether at the time of sale or move to another province. First time purchasers are free to pay the registration charge in any province they choose. There are no tax exemptions, not even for government vehicles. Tax rates have remained unchanged for 4 years in Punjab and 8 years in NWFP. The tax can be paid in installments, with a 10 percent discount for full payment at the time of registration. Assessment and collection are relatively straightforward. NWFP is trying to expand the computerization of motor vehicle registration but to date no tax data has been included in this database. Moreover, it is impossible to link the numbering on each record with other tax information. 9.33. The annual token tax on automobiles is a fixed rupee amount, depending on the engine capacity. The average car (1,000 to 5,000 cc) pays a flat amount of Rs 1,500 per year. The token tax rate has remained unchanged since 2000. Trucks pay according to axle weight, and buses according to seating capacity. Vehicles older than 10 years receive a 20 percent discount. Vehicles used for agricultural purposes are exempt. The token tax for motor cars can be paid where the vehicle was originally registered or can be paid in another district if the owner applies for a change of residence. For "tied" vehicles, which include commercial vehicles, buses and trucks, the tax can only be paid in the office that maintains the registration record. Otherwise, taxes are paid either at the post office or at the Excise and Taxation offices. Police squads are the major form of enforcement, using random road stops. Proof of compliance is a registration card that must be carried in the car. Neither number plates nor windscreen stickers are used in enforcement. Figure 9.8: Level and Trends in Motor Vehicle Tax Collection Motor Vehicle Tax Collection in Punjab and NWFP in 2005/06 Motor Vehicle Tax (Rupees Per Capita, 2005/06 Prices) 35 50 18 Rupees Per Capita NWFP 45 30 % of Own-Source 15 40 Revenue 25 35 12 30 20 25 9 Punjab 15 20 6 15 10 10 3 5 5 0 0 0 Punjab NWFP 1999/2000 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 126 Issues 9.34. The major policy issue regarding motor vehicle taxation is whether the structure of the tax matches with the government's goals. Provincial governments might have multiple goals for the tax on motor vehicles. The most obvious is to raise some target amount of revenue for financing public services. However, there are other possible goals that the government might have for motor vehicle taxes. The government might want to charge road users for the upkeep of the road system, or it may want to impose an extra tax on those who generate congestion and pollution. It may even want to discourage the use of roadways to reduce congestion. All of this needs to be considered in light of the demand for motor vehicle transport, and of the economic and political sensitivity of taxes on this sector of the economy. 9.35. The current tax regime for motor vehicles suffers from a number of structural problems. First, the level of taxation is very low. At a time when the provinces face a shortfall in the quality of public services provided, there would seem to be a premium on revenue mobilization, and motor vehicle taxation would appear to offer some potentially significant fiscal space. One question to be answered is "how much revenue should be raised from these taxes"? One approach is to view motor vehicle taxes as a kind of benefit charge for using roadways. Yet, in 2005/06, taxes on motor vehicles were equivalent to only about 65 percent of provincial government expenditures on road maintenance and construction in Punjab, and 53 percent in NWFP. In addition, actual roadway expenditures have been insufficient for several years, leading to deteriorating road conditions and poor quality of public services at a time when the numbers of registered cars rose from 3 million in 2003 to almost 5 million in 2007 in Punjab alone. Clearly, from the perspective of a benefit charge for the use of public roads, significant increases in the motor vehicle tax seem justified. Such increases would also seem justified from an equity point of view, as ownership of motor vehicles, especially of larger and more expensive vehicles, is concentrated in the highest household income deciles. However, the present token and registration taxes do not allocate burdens according to the amount of road use, which makes them an imperfect levy from the benefit perspective. 9.36. A second problem is that there are significant leakages from the tax base. With respect to the registration charge, some owners just do not pay the tax. Some register in other provinces as provided under this tax. For example, many commercial vehicles appear to be registered in Balochistan where the tax rate is lower but operate primarily in NWFP. Some use fake documents to under-declare taxable value. In addition, significant exemptions are a problem for NWFP. One study estimated that less than 50 percent of the vehicle traffic in NWFP was registered in the province (Malik 2004). Khan (2004) also notes the problem associated with a large number of vehicles brought into the country without payment of custom duty. He estimates about 20,000 such vehicles in FATA/PATA areas in NWFP, of which less than half are registered. 9.37. Finally, there are other administration problems. The annual token tax for private cars is paid at the post office. Because there is not good coordination between the post office and the excise and taxation department, record keeping on compliance with motor vehicle taxes is not very good. Commercial vehicles can pay only at the district excise taxation department offices, where recordkeeping is thought to be better. The motor vehicle tax payment records are not computerized. All of this suggests some unfairness in the levy of motor vehicle taxes. Some pay it fully, while some are outside the tax net. This differential taxation erodes taxpayer confidence and further stiffens resistance to tax increases. 127 Reform Options ­ Piecemeal 9.38. The rational for increases in motor vehicle taxation is well founded. A properly structured tax regime could provide a needed boost to revenue collection, and can potentially address two other needs: (a) act as a benefit charge on road users to partially offset the cost of road construction and maintenance, and (b) compensate society for the external costs of motor vehicle (i.e., pollution and congestion). However, the present system does not address these goals. Reforming the existing motor vehicle tax structure is going to be essential to increase the revenue yield. The challenge is how to change the revenue structure so as to hit the collection target and achieve the goals outlined above. The provinces can take either of two approaches: a piecemeal reform or a comprehensive reform approach. 9.39. A piecemeal reform would leave the basic structure intact and instead iron out some of the problems with the component registration and token taxes. A number of reforms may be considered for both the registration and token taxes. For the registration tax, one possibility would be to transfer the tax to the central government with agreement to share the revenue on the basis of the size of the road network. While the registration charge is intended as a onetime payment for the privilege of using provincial roads, it has certainly failed this test given the option of choosing any province as the place of registration irrespective of place of residence or road use in each province. However, the registration charge is an important provincial tax and provinces will most likely resist its conversion to an intergovernmental transfer. Hence, in case the provinces would like to retain this tax, the provinces could introduce a residency requirement that will be enforced with provincial number plates and windscreen stickers. This has a potential revenue enhancing impact, but compliance enforcement and policing will require new techniques that will be administratively difficult. It could also pose a problem for commercial vehicles that operate across provinces. Another option is that the provinces could double the tax rate for each class of motor vehicle. This has the potential to increase revenues by 12 and 16 percent in Punjab and NWFP, respectively. The effective rate on automobiles would rise from 1 or 2 percent of value to 2 or 4 percent. While this may not seem an inordinately large increase, it comes on top of a number of other taxes on new purchases, such as the presumptive income tax for the federal government, and could face stiff political resistance. 9.40. For the token tax, one option would be to increase the tax rate on the present levy through inflation indexing. For example, such inflation indexing would have roughly doubled revenues from the token tax in both Punjab and NWFP between 2003 and 2008. However, such a reform would simply be a revenue measure and not remove its fundamental flaw: the token tax has failed the test of a good roadways use charge because the amount of the tax does not vary with motor vehicle use or, by proxy, the place of residence. At any rate, because it is a flat charge per type of vehicle, it does not discriminate according to miles driven. Thus, any rate increase would simply be a revenue measure and does not link the tax burden to road use. Linking the tax to road use would require to taxing fuel usage. 9.41. Supporting administrative reform for both the registration charge and the token tax is to computerize the records to allow easier tracking of registrations and comparisons with other information. There is much to recommend automation, e.g., those with tax liability can be tracked and enforcement can become more efficient, a reliable defaulter list can be generated, the work of the "spot check" squads can become more efficient, the impact of tax policy changes can be better estimated, and revenue projections for budget purposes can become more accurate. In fact, such an automated system is now being developed in Punjab, but is still in the pilot stage. A related improvement has to do with human resources. The collection rate could be strengthened if more manpower were available to enforce the taxes. This could be assisted by windscreen or number plate certificates that would enable easier detection of nonpayment. Still, these administrative improvements will fall well short of objectives if there is no residence requirement for registration, because leakages from the taxed base will continue. 128 Reform Options ­ Comprehensive 9.42. One option for comprehensive reform is to replace the existing registration and token taxes with a unified annual license tax based on vehicle type as a rough proxy for engine capacity. All motor vehicles could be grouped into three classes, and each could be subjected to a specific rate that would be in indexed to the general rate of inflation to allow for built-in revenue growth. The specific rates chosen would cover the cost of roadway expenditures at present levels in the provinces. This new structure would be more simple to administer and more revenue productive than the present system, and would not heavily burden low income households. 9.43. A more far-reaching comprehensive reform would be the adoption of a provincial motor fuel tax. While the token tax is meant to be a user charge for road use, it does not reflect the amount of road use, and could be introduced alongside an unified annual license tax. A tax on motor fuels would better serve this objective and would mobilize significant revenues. Motor fuels are taxed at the central government level in virtually all countries. In the United States, the state governments have autonomy in setting the rate of tax on gasoline. The levy of motor fuel taxes by subnational governments in developing economies is not universally practiced, but is certainly not unknown. In fact, a motor fuels levy has been proposed and evaluated as a revenue source for provincial governments in Pakistan. One fiscal analysis of Punjab recommended the adoption of a tax on motor fuels, on grounds of revenue productivity, ease of collection and the benefit charge aspect (Government of Punjab 2007). The report cautioned, however, that the tax should be phased in rather than implemented fully in one year, and that it should be proposed as a user charge to protect its legal status. Though there is much detail to be worked out, a provincial tax on motor fuels would seem to be a workable proposition (Box 9.5). The following is a general outline of a proposal for a provincial motor fuels tax (motor fuels charge): · The tax would be levied on an ad valorem basis with the rate set at the discretion of the province. This means that fuel prices would vary from province to province. · The tax rate also could vary by type of fuel, at the discretion of the provincial government. · The tax would be collected by the oil marketing companies based on fuel sold in the province. The tax would be remitted by the oil marketing companies directly to the provincial governments or to the federal government. In the latter case, the federal government would then transfer the revenues to the provincial treasury. 9.44. The revenue yield from the tax would depend on the revenue target of the provincial government, and subsequently on the rate that it chose. Revenues actually raised from motor vehicle taxes in 2005/06 could be covered by a specific tax rate of Rs. 0.52 per litre of diesel and Rs. 1.00 per litre of motor fuel in Punjab, and a tax rate of Rs. 0.33 per litre of diesel and Rs. 0.75 per litre of motor fuel in NWFP (Table 9.3). The necessary ad valorem rates would have been 1.47 percent in Punjab and 0.93 percent in NWFP. Table 9.3: Revenue Potential of a Motor Fuel Tax Punjab NWFP Motor Vehicle Revenues in 2005/06 (in Rs millions) 4,154 677 Consumption of Petroleum (Litres in millions) Diesel 6,049 1,840 Motor Fue l 1,013 89 Necessary Tax Rate Spec ific Rate per litre (in Rs) Diesel 0.52 0.33 Motor Fue l 1 0.75 Ad valorem rate (% of total Rupee value of consumed fuel) 1.47 0.91 129 Box 9.5: Overcoming Impediments to the Adoption of a Provincial Motor Fuel Tax There would be numerous impediments to the adoption of a provincial motor fuel tax. The following are notable questions and some possible responses. (1) The federal government is committed to a uniform national price for motor fuels. This proposal would lead to a variable price if provinces chose different tax rates. In actuality the proposed fuel tax is a user charge for provincial services. Since the cost and quality of provincial road services varies across provinces, differences in the fuel tax rate are justified. Such variations exist for example in both India and Brazil. (2) Some provinces might intentionally set their fuel tax rate low to attract consumers. This is possible, but it should be a phenomenon affecting provincial border areas only and these provinces would lose revenue if they chose a lower rate. The federal government might provide an incentive for levying this charge at higher rates by reducing federal transfers by the estimated amount of fuel tax that could have been raised at "normal" rates, or by mandating a minimum tax rate as is done in India. (3) Would certain uses of petroleum be exempt from the tax, and would certain uses be taxed at a different rate than others? In principle, exemptions and preferential rates might be justified based on use by low income households. Most often mentioned in this regard are home heating and cooking, public transport and agriculture. In practice, much of such preferential treatment is not feasible on administrative grounds, and would open the door to abuse (as has happened in the case of other provincial level taxes). The best route is to have no preferential treatments, with the possible exception of kerosene for domestic use. Utility production and pricing decisions should be taken in recognition of the market price of petroleum (including taxes) as should public transport and agricultural use. If it is desired to lower the tax burden on low income households, there are better ways to do this than with price reductions for petroleum products. While this principle seems an appropriate basis for structuring a tax on petroleum products, the present analysis is focused only on motor fuels. (4) Commercial vehicles (e.g., trucks) may buy gas in one province but use the roadways in another, thereby compromising the benefit justification of a provincial motor fuel tax. To some extent this will happen. The use of an apportionment formula, as is done in the US (McLure et. al. 2007), is not administratively feasible in Pakistan. To some extent the effect of this leakage will be to narrow the inter-provincial differences in motor fuel tax differentials. (5) How to prevent tax base erosion due to smuggling? There is a particular concern from Balochistan that they will lose revenue from reduced motor vehicle fees and not recapture this revenue from motor fuel taxes because of smuggled petroleum from Iran. The proposed size of the provincial fuel charge might not be great enough to significantly increase smuggling. In any case, the first line of defense against fuel smuggling should be provincial and federal controls. 130 9.5 Urban Immovable Property Tax Base and Rate 9.45. The urban immovable property tax (UIPT) is an old tax in Pakistan, but it has never generated significant amounts of revenue. For instance, the national property base was estimated to be Rs. 70 billion in 1996 and the national collection would have been Rs. 7 billion in that year with a statutory tax rate of 10 percent (Ghaus-Pasha et. at. 1998). Property tax collection in all four provinces was only Rs. 506 million in 1995/96, and only about Rs. 1 billion in 2005/06. Punjab collected in 2005/06 no more than Rs. 13 per capita, equivalent to only 4 percent of own source revenues and 0.03 percent of GDP. This level of collections is low by comparison with other developing economies, where the average is about 0.5 percent of GDP (Figure 9.9). The property tax in Punjab is based on an annual rental value of a flat rate of 20 percent on properties valued at Rs. 20,000 or less, and 25 percent on properties valued at more than Rs. 20,000. The law defines annual rental value as the amount of rent that could be obtained in an unencumbered market transaction, less an allowance for maintenance and repairs. A valuation table is developed using a combination of market data and expert judgment about rental values. This table serves as the basis for valuing all properties in the province. 9.46. Following the Local Government Ordinance of 2001, Punjab's urban immovable property tax is designated a local government tax. The city districts and tehsil municipal administrations (TMAs) have the authority to set the rate, while assessment and collection can be at the district level. In actual practice, however, the UIPT is a provincial level tax subject to revenue sharing with the city districts and TMAs. The tax rate and base are set by the province, and tax administration is carried out under the leadership of the Province. The revenues are mandated for distribution to TMAs as specified in the Local Government Ordinance 2001. The province retains 15 percent, and the TMAs receive 85 percent. The provincial government has launched reforms to reconcile the property tax legislation with the Local Government Ordinance (Box 9.6) 9.47. Just as in Punjab, the property tax in NWFP is effectively a provincial government levy. TMAs have property taxing powers, including powers to set the tax rate and to administer the tax, but they do not use these powers. As in Punjab, property tax revenues are allocated 85 percent to the TMAs on a basis of origin of collections, and 15 percent is retained by the provincial government. While NWFP collects in per capita terms more UIPT than Punjab, the level of property taxation is still well below the international average (Figure 9.9). NWFP uses an area-based system of taxation. The base of the tax is the sum of land area in square yards and covered area in square feet. This implies that more intensively used land will be taxed at a heavier rate than less intensively used land. The area-based system combines the tax rate and the tax base into a location coefficient, of which there are eight in the province. The size of these coefficients depends on the location of the property and whether it is residential or commercial. The province is divided into four location classes based on the desirability of the location, availability of amenities, etc. The determination of location class is done by field inspectors who possess knowledge of the areas, and this determination is reviewed and eventually validated by the provincial government. The last exercise of classification of each property according to location and use was carried out in 2001. The actual tax rates are notional, but they attempt to reflect both relative values and a policy choice that commercial land use should pay more property tax than residential land use. Figure 9.9: UIPT Collection in Punjab and NWFP in International Comparison UIPT Collection in Punjab and NWFP in 2005/06 International Comparison of Property Tax Collection (Percent of GDP) 16 7 Source: Bahl and Martinez-Vazquez (forthcoming), IMF (2005) % of Own-Source 1.2 Revenue 1.1 14 6 1.0 Rupees Per Capita 0.9 12 5 0.8 0.7 10 4 0.6 8 0.5 0.4 3 6 0.3 0.2 2 4 0.1 0.0 Hungary NWFP Punjab Poland Sri Lanka Thailand Indonesia Argentina Mexico Croatia Ethiopia Slovak Republic South Africa 1 Chile 2 0 0 Punjab NWFP 131 Issues ­ Low Revenues 9.48. UIPT in Punjab and NWFP is beset by a number of problems. Some are structural and some are administrative, but none are insurmountable with present levels of technology and with some upgrading of staffing. 9.49. Low revenue yield and revenue growth. The property tax yields very little revenue in both Punjab and NWFP. The low revenue yield might be attributed to a number of factors: (i) taxpayers do not see adequate value in the services they receive from local governments, hence they are not willing to tax themselves at a higher rate; (ii) the property tax is an inherently unpopular way to raise revenues and elected politicians are unwilling to enforce it; (iii) those who receive property tax preferences are keen to keep such privileges; (iv) intergovernmental transfers, which provide provincial governments with adequate revenues at low political cost, have become a disincentive for increased property tax effort; and (v) enforcement is generally weak. To move property tax revenues to a higher level, all of these barriers will need to be addressed. Increases in taxable property values are not captured because both Punjab and NWFP use outdated valuation tables. In 2007, Punjab postponed the introduction of a new valuation as elections were near. Newly developed properties are not being brought into the tax roll and there is no annual revenue pickup due to expanding economic activity. Punjab province has not changed nominal tax rates since 2002, and NWFP province has not adjusted location values since 2000. TMA councils can assign tax rates to newly developed properties through a resolution, but have been unwilling to do so. The low yield and growth of revenues are related to a narrow tax base, tax preferences, and properties' evaluation. 9.50. Tax base - Punjab. Punjab's rental value system leaves a number of properties out of the tax system. Owner-occupied residential properties and vacant properties are not taxable under the current system because they do not have a rental value assigned to them. The valuation of industrial properties under the current system is also uncertain because capitalization rates must be assumed. These issues may be remedied by adopting a capital value base system. However, this system also has its share of problems, mostly related to valuation. Depending on the kind of capital valuation approach adopted ­ estimating the market value of either the total property or the land and structure separately ­ it can be a very expensive process that is hampered by lack of accurate data on sales of property and a cadre of skilled property assessors or it could end up being alike to what is practiced under the current system. Whether the government of Punjab chooses to stay with the rental value system or adopt a capital value system, it needs to develop a method for properly valuing residential, commercial, and vacant properties. 9.51. Tax base - NWFP. The UIPT in NWFP is an area-based system of taxation which replaced an annual rental value system in 1997. A problematic feature of the area-based system used in NWFP is that there is no distinction between the tax rate and the tax base, as both a combined in the location coefficient. There is hence no mechanism to allow automatic growth in the tax base. 9.52. Tax preferences. The tax base have been significantly narrowed by exemptions and other preferential treatments (Box 9.7). These preferential treatments create great inequities in the distribution of property tax burdens, contribute to noncompliance, and reinforce resistance to increasing the statutory rate and to improving valuation. The government does not keep record of the revenue cost of exemptions, but they certainly large: for Punjab, we estimate the revenue cost of preferential treatments in 2006/07 to be Rs. 3.7 billion or 130 percent of the level of demand for collections; for NWFP the responding figures for 2005/06 are Rs. 610 million or over 220 percent. For Punjab, brining in the new valuation roll designed in 2007 and eliminating preferential treatments alone could lead to more than a tripling of UIPT revenues (Box 9.8). 9.53. Undervaluation. Properties that are in the tax base are significantly undervalued. It is difficult to make an objective estimate of the revenue cost of undervaluation because there are no independent estimates of property or rental values, and the government does not carry out a sales-assessment ratio study. Some available data suggest a large degree of undervaluation. For example, in Punjab, the new but not yet applied 2007 valuation table would result in a fivefold increase in values of properties compared to the present 2002 valuation table. In NWFP, there is at present no provision for regular revaluation of properties or location values. The only revenue growth comes from new properties that are being brought on to the tax rolls and from collections of arrears. None of the increase in values for properties already on the roll has been captured since 2000, the year of the last recalculation of location values, and even that was done only in Peshawar. In other words, the measurements of built-up areas, land use, and the condition of the surrounding infrastructure are all based on outdated information. 132 Box 9.6: Medium Term Reforms of the Urban Immovable Property Tax in Punjab According to the Local Government Ordinance (2001), the property tax is to be fully devolved to the local governments. In theory, the TMAs have the authority to set the tax rate and to administer the tax. In fact, neither the province nor the TMAs have been willing to use their taxing power. Under the medium term policy framework approved in December 2007, the provincial government will prepare new legislation to reconcile the property tax legislation and the Local Government Ordinance. Tax policy and valuation will be given to five city district governments initially. In time, the TMAs will be allowed to choose tax rates from a range set by the province. Implementation will rely on a "readiness" criterion. The province will continue to perform an oversight function, lay down acceptable ranges of tax rates, and review valuation and tax collection procedures. After decentralization, local governments will continue to use uniform methodologies for valuation and tax collection. Box 9.7: Exemptions and Preferential Reductions of UIPT in Punjab and NWFP Exemptions and preferential reductions in Punjab's UIPT include: · A 10 percent reduction in taxable value for all properties; · significantly low assessment of owner-occupied units compared to rented units; · no taxes for owner-occupied residential properties and vacant properties; · no taxes for retired government employees that are owner occupiers and live on plots of 500 square yards of land or less; · deductions of Rs 48,000 in annual rental value for widows, orphans and the disabled; and · no taxes for government properties and cantonments. Exemptions and preferential reductions in NWFP's UIPT include: · 10 percent reduction in tax liability if property is more 10 years old, 20 percent reduction if more than 20 years old, and 30 percent reduction if more than 30 years old. The age test is independent of location; · 50 percent reduction for owner-occupiers receive a 50 percent reduction, plus above mentioned age based deduction; · no tax for properties with lot sizes less than 5 marla; · no tax on property owned by widows; · no tax on government properties, except on local government property; and · cantonment to levy their own property tax and pay 15 percent of revenue to provincial government. Box 9.8: Estimates of the Revenue Cost of Preferential Treatments in Punjab and NWFP For Punjab, our estimates of the revenue cost of preferential treatments in 2006/07 is Rs 3.7 billion, an amount equivalent to 132 percent of the current level of demand for collections. If the new tax roll were brought on line, the revenue cost of preferential treatments would rise to Rs 7.6 billion. These estimates are made with the help of some simplifying assumptions, but the results suggest that even other reasonable assumptions would yield a similar result. This is an important finding if one wants to estimate the revenue potential of the property tax in Punjab. Bringing in a new valuation roll and eliminating preferential treatments alone could lead to more than a tripling of revenues. For NWFP, our estimate is that the loss due to all tax preferences is about Rs. 610 million, which is equivalent to more than 220 percent of actual collections in 2005/06. While NWFP has not developed a new roll, we assume an increase in current values by a factor of 1.5 for updating the valuation roll. Together with the elimination of tax preferences, this would generate additional revenues of about Rs. 1.1 billion, or about four-times the current collection. Figure: Revenue Cost of Preferential Treatments Punjab NWFP Outdated Valuation Roll (183%) Outdated Valuation Roll (178%) Tax Preferences (223%) Tax Preferences (128%) Actual (100%) Actual (100%) 133 Issues ­ Land Use, Devolution, and Administration 9.54. Inefficiencies in land use. In Punjab, the tax burden under the area-based system of taxation falls more heavily on improvements than on land. This is because the same tax rate applies to land area in square yards and covered area in square feet. Under this system, more intensively used land is taxed at a heavier rate than less intensively used land, which encourages less intensive use of land. This disincentive to a more efficient use of land is reinforced by the full exemption of vacant properties. At current low levels of property tax, these disincentives to more intensive land use may not matter very much, but under an upgraded property tax they would be of greater importance. 9.55. Devolution. The provincial governments have yet to take a clear decision on whether the property tax should be a provincial levy or a local levy. Punjab's medium term policy framework attempts to remove this ambivalence by committing to pass this responsibility to local governments by 2012. The legal authority is in place, but past provincial actions suggest a reluctance to release this authority: TMAs are not regularly notified of their revenue entitlements; property tax revenues have until recently been intercepted for payment of local utility bills; and no significant fiscal incentives have been put in place to encourage local governments to impose higher rates. Local governments have also not shown a willingness to take on the property tax as a major local revenue source. They have not used the rate setting power that they have been given due to a combination of significant inflows of intergovernmental transfers from the province, concern that the inflows would reduce if they raise rates, and close relationship of elected officials with tax payers. Perhaps unsurprisingly then, in NWFP, local government have lobbied the provincial government on behalf of their constituents for lower property valuation classifications (Khan 2004). As the implementation of Punjab's medium term policy framework proceeds, the provincial government plans to closely monitor the property tax actions of the local governments. Presumably this will include an assessment of the extent to which they are prepared to assume their new administrative responsibilities. 9.56. Administration. Even with a more rational structure of UIPT, revenue enhancement will be held back because of the presently poor state of property tax administration. One issue is the major need is to upgrade the skills and size of the staff that assess and collect UIPT. For example, Peshawar's over 115,000 property units cannot be managed by the present excise and taxation department staff (Khan 2004). A second issue is to begin modernizing recordkeeping toward an automated system. Property tax records are manually recorded, in Urdu, and there is no automation in billing or in tracking collection rates. This makes an efficient collection process very costly and likely dampens the collection rate. Payment records are said to be out of date, and the matching of property tax payment with motor vehicle registration and other third party information cannot be done. A World Bank financed project in Punjab is now undertaking computerization of rural land records, but it will take time to complete this work. Finally, on the assessment side, there is little or no evaluation of the effectiveness of the process as it now operates. Three points in particular require special note. First, there are no sales-ratio studies hence it is not possible to grade the effectiveness of the valuation process in estimating market value. Only anecdotal evidence is available regarding underassessment, and about whether the process is fair in assigning the same values to similarly situated properties. Second, the government does not systematically evaluate the impacts of the various exemptions that have been given. For example, estimates of the revenue cost of the important 5 marla exemption have not been made by the Government. Third, there are no surveys of all properties in urban areas. Therefore, there are doubts that the tax roll is comprehensive and up-to-date. 134 Reform Options ­ Punjab 9.57. Punjab has to resolve two basic issues in property tax reform. The first issue is whether this tax will become an important source of financing for government services. If a significantly higher level of revenue is to be reached, then the base and rate structure, the administration, and the intergovernmental arrangement all should be changed. Coordination of these changes and a phasing in plan will necessarily be part of the reform program. There are a number of reform options open to the Punjab government to make the property tax more revenue productive and fairer. Some, but not all, of what we propose here has been discussed in other research reports (The Urban Unit, 2006 and 2007) and is also contained in the medium term policy framework. The second issue is whether it will be left to local governments or to the provincial government to govern the tax so as to reach the target level of revenue. This question has been answered largely with the approval of the medium term policy framework. 9.58. The first reform decision to be taken is the revenue target for the property tax. Ideally, this will be determined with reference to the amount of expenditures to be covered by the tax, or by all own source revenue. For purposes of illustration, we arbitrarily set the level of target revenue at the international average for less developed countries of approximately 0.5 percent of GDP. The implied tenfold increase in UIPT collection to Rs. 25 billion may seem more than a little ambitious. Still, this target can be achieved with a combination of structural reforms, rate increases and administrative improvements (Table 9.4). 9.59. The most obvious revenue-raising measure is the introduction of a new valuation roll. If fully implemented under the present rate structure, it could lead to a significant one-time increase in revenues. This could be perceived to cause enough "tax shock" that it might need to be phased in, but it should be introduced as soon as possible. As welcome as revaluation would be, its introduction without other changes would cover only about 12 percent of the gap between the current level of property taxes and the target level. 9.60. The present package of exemptions and preferential treatments could be significantly reduced. The reform argued here is to tax all property at the same rate, except that which is exempt by normal convention (e.g., places of worship, charities) and that which is exempt to protect low income families. Eliminating five types of preferential treatments (5 marla exemption, owner-occupied preferences, vacant property exemption, taxation of industrial property at residential rates, and exemption of provincial government properties (Box 9.8) would bridge about one-third of the amount needed to reach the Rs. 25 billion target. In addition, the taxation of land and improvements should be brought to the same basis, i.e., the tax base should be the number of square feet of covered area plus the number of square feet of land area. Since this involves increasing the tax on land relative to that on buildings, it will lead to an additional revenue increase although there are no data to estimate its magnitude. 9.61. The revenue growth issue might be addressed in one or both of two ways. First, the government might introduce the practice of indexing the tax rates in the period between revaluations. This would generate more revenue between revaluation periods, would cushion the fifth year shock associated with the mandated five-year revaluation, and would raise the GDP elasticity of the property tax. In our calculations, indexing would contribute about 4 percent of the revenue gap. Indexing is not without problems, primarily because it treats all properties as if they grow in value at the same rate between revaluation periods. A second approach is to shorten the time between revaluation periods from five to three years. A three-year cycle captures increases in the tax base than better does a five-year cycle, provides a greater revenue flow in the long run, and reduces the shock of introducing a new tax roll; at the same time, updating valuation rolls is costly and there is a risk of postponement. 9.62. Increasing the nominal tax rate from the present average of about 22 percent to 40 percent could cover the remaining gap of Rs 11.2 billion. A 40 percent statutory tax rate may sound exorbitant or approaching confiscation. However, what really matters is the effective rate, i.e., the ratio of tax paid to market rental value. A 40 percent nominal rate translates into a much lower effective rate, depending on the assessment ratio. 9.63. Finally, there are two organizational issues. The responsibility for valuation should be separated from the political decisions regarding rate levels and exemptions. In addition, all properties, whether exempted or not, should be valued and the total potential base should be reported. Such reporting, and taking them into account as offsets to the Provincial Finance Commission award, might curb their adoption. 135 Reform Options ­ NWFP 9.64. NWFP should rethink the rate and base structure of its property tax. The tax liability per area unit of a property is all that is reported in the valuation table, and this coefficient is the product of a nominal tax rate and a valuation. A first necessary reform is to make explicit the separation between tax rate and tax base. This separation needs to be made explicit for several reasons: · Responsibility for rate setting could be assumed by city district government and TMAs before valuation responsibilities are handed over. This could not be done under the present regime. · Good property tax practice requires a separation between valuation (a technical matter) and rate setting (a political matter). · Monitoring of the accuracy and fairness of property tax assessment practices requires regular sales- assessment ratio studies. These would require separate data on property values per area unit. 9.65. If we use the international average for less developed countries of 0.5 percent of GDP, the target for 2005/06 would have been Rs 3.6 billion, or an implied increase of about 10 times the present level. There are reform options available to upgrade the property tax to this extent (Table 9.5). 9.66. The first problem to be addressed is to build into the system a method to capture property value increases. This should be followed by a general revaluation, or in the NWFP case, a revision of the valuation table. At present, the NWFP government has no plans to construct a new table. There is no solid empirical basis for estimating the revenue increase that would follow a revaluation, e.g., data that show increments in market value over the past decade. We can offer only a hypothetical example of how revaluation might affect the property tax base. We do this by noting that the increase in assessed value in Punjab was estimated to be by a factor of 1.8, suggesting an 80 percent increase in property values between 2002 and 2007. To illustrate the potential effects of revaluation in NWFP, we have used a factor of 1.5 --- the lower factor to suggest the lower level of income in NWFP. Such upgrading of the valuation table (by 50 percent) would lead to a revenue increase of Rs 150 million, an amount equivalent to only 4.5 percent of the targeted revenue increase. 9.67. A second set of reforms would focus on base broadening. We estimate the revenue gain from the following changes in the present exemption package, using exactly the same methodology that was followed in the case of Punjab. These changes would yield an increase of Rs. 1,541 million at new valuation levels. This is more than three times the level of current collections. Base broadening in NWFP would cover about one-third of the gap between the present level of property taxes and the target level. 9.68. To address the problem of building some natural growth into the property tax base, we evaluate two reform options. First, we have indexed the tax rate by the (one-year lagged) general rate of inflation. Over the 2002/03 to 2005/06 period, the province would receive a total of Rs. 127 (3 year total) million in additional revenues (83 percent of the 2002/03 total). We can also note that in the event of a general revaluation in 2006/07 without indexing, the tax shock would have been 98 percent of 2002/03 levels, but under indexing, the shock would be reduced to 47 percent. The second possibility for growth enhancement is the introduction of a three year valuation cycle. Using historical data and assuming general revaluations in 2002/03 and 2005/06, we show that between general revaluations the government would have gained Rs. 60 million in additional revenues. 9.69. The estimated effect of these changes in property tax structure is to cover about Rs 2 billion of the estimated revenue gap of about 3 billion. How could the remainder be made up? In the NWFP system, this might be done by further increasing the location values (above the 150 percent built in to the revaluation). Assuming a 75 percent collection ratio, we estimate that the `location factors' will need to be increased by 2 times to bridge the revenue gap. 136 Table 9.4: Illustrative Property Tax Reform Program for Punjab Revenue Impacts of Reform Package Amount (Rs. Million) 1. Revenue Target 2006/07a 25,478 2. PT Collection 2006/07 2,311 3. Gap 23,167 4. Revenue impact of introducing a new valuation roll on the existing 2,889 tax baseb 5. Revenue impact of base broadening measures after introduction of 7,551 the new rollc 6. Indexation (Additional Revenue) 827 7. Payment in lieu of tax 611 8. Revenue gap [3-(4+5+6+7)] 11,289 9. Property tax rate required to cover the gapd 0.40 Notes: a Revenue target is 0.5 percent of the 2006/07 provincial GDP, as estimated by Punjab Bureau of Statistics. b Revenue impact is based on the assumption that a new roll is introduced for 2006/07. This is computed by subtracting the actual collection in 2005/06 from the new demand worked out by ETD after revaluation. c Net impact of removal of exemption for 5 marla properties, vacant properties and provincial government properties, and removing preferential treatment for owner-occupied properties and industrial properties. d We use a rate of 22 percent (an average statutory tax rate, considering the nominal rates of 20 percent and 25 percent) to calculate the annual rental value of the assessed property from the actual collection. Then we add the projected revenue from reform to the baseline collection and calculate the new annual rental value, using an effective tax rate of 22 percent. To conclude we use the calculated annual rental value and the total revenue target to calculate the effective tax rate required to bridge the gap. Table 9.5: Illustrative Property Tax Reform Program for NWFP Revenue Impacts of Reform Package Amount (Rs. Million) 1. Revenue Target 2005/06a 3,631 2. PT Collection 2005/06 300 3. Gap 3,331 4. Revenue impact of introducing a new valuation roll on the existing 150 tax baseb 5. Revenue impact of base broadening measures after introduction of 1,030 the new rollc 6. Indexation (Additional Revenue)d 78 7. Payment in lieu of taxe 511 8. Revenue gap [3-(4+5+6+7)] 2,073 9. Percent increase in location coefficient required to cover the gap 200 Notes: a Revenue target is 0.5 percent of the provincial GDP for 2005/06 from Government of NWFP (2005). b NWFP does not have a projected demand based on a new roll. We assume it will be 1.5 times the 2005/06 demand. c Net impact of removal of exemptions for owner occupied units, higher rates for industrial property, taxation of vacant plot. d Using 2002/03 as the base line year. e Payment in lieu of taxes calculated as 12 percent of annual rental value (government expenditure is 12 percent of GDP). 137 Reform Options ­ Intergovernmental Arrangements 9.70. The provincial government in Punjab has adopted a medium term framework that will move it toward the assignment of property tax policy and administration to local governments. In the interim it should implement the legal tax sharing arrangement with local governments. There are strong arguments to make the property tax a local government revenue source, with the TMAs having rate setting powers and some degree of administrative control, as announced in the policy framework. While this is the present legal arrangement for the property tax, de facto it is not yet the practice in either province. It would be useful to revisit the benefits of the legal arrangements laid down in the 2001 local government ordinance, and to begin implementation, for the following reasons: · This will permit a stronger link between property taxes paid and local public services received. · Taxing powers will make elected local officials more accountable to the voters for the quality of local public services delivered. · Local officials have greater familiarity with the local economy hence a comparative advantage in some areas of tax administration, e.g., identification of properties for the tax roll. · There are also disadvantages to adopting this intergovernmental approach, and concerns that might be raised about the potential success of the medium term framework. First, local governments already have the authority to set the tax rate and to administer the tax, and they have chosen to do neither. Why will things be any different under Punjab's medium term framework, and why might things change in NWFP? Second, some local governments may not be technically up to the job of property tax administration. This could result in a long transition period. The first challenge in structuring the reform program and implementing it would be to get around these disadvantages. 9.71. As a first step, the provincial governments could move toward a redefinition of certain intergovernmental practices. · So long as the provincial government remains involved in collections, it should notify each local government of its revenue entitlement (the amounts collected in its area). This should be done on a timely basis so as to assist local government budgeting process and cash flow planning during the fiscal year. · No intercept from property taxes should be allowed for utility payments. If there are to be intercepts, they should come from the general grant to local governments. The intercept from property taxes weakens the perception of the linkage between property taxes paid and local public services received. In fact, the practice of intercepts for utility payments has recently been discontinued. 9.72. While increasing the effective rate of property taxation is a reasonable goal in view of the shortage of funds for local public services, local governments have not been willing to use the rate setting powers that they now have. The alternative, transfers from the province, are "easier" money in that they have little political cost to local politicians. This is the mindset that must be broken if decentralization is to succeed. One way to do this is to provide a significant incentive for local governments to increase the level of property taxation. The province might attach a tax effort feature to its grant program to try and coax this increased property tax effort out of the local governments. 9.73. The administration of the property tax needs to be aligned with the local capacity. There is a strong a priori case for local administration and there is a legal basis for this. This is the long term plan under the new framework in Punjab and presumably it is the plan in NWFP. However, there is a question of administrative capacity at the district and TMA levels. This leaves three options, all of which would increase the revenue importance of the UIPT: · Divide responsibilities along functional lines, for example, leave preparation of the valuation table and recordkeeping with the province, but let the local government be responsible for collections and for identification of new properties to be added to the tax rolls. · Let the local governments assume further administrative responsibility when they demonstrate readiness, as measured by some objective benchmarks. · Let the tax administration be led by the province, but allow the local governments to set the tax rates (perhaps above some minimum). 138 9.6 Land and Property Transfer Taxes Base and Rate 9.74. NWFP and Punjab collect little in the way of land and property transfer taxes. The Board of Revenue (BOR) is responsible for rural land taxes and charges and all property transfer taxes. Taxes on property transfers also include mutation or registration fees. NWFP, but not Punjab, collects an annual tax on land. In both provinces, local governments also collect an additional tax on property transfers. BOR's revenue collection from these taxes stagnated since the late 1990s and accounts for no more 0.2 percent of GDP in Punjab and 0.1 percent of GDP in NWFP, even though agriculture accounts for about 30 percent of GDP in both Punjab and NWFP (Figure 9.10). Stamp duties, in particular on property transfers, are the major revenue component of BOR taxes. Punjab collected Rs. 11.4 billion land and property transfer taxes in 2005/06. Stamp duty contributed half the receipts, registration one-fifth, and other land revenue taxes the rest. NWFP received Rs. 660 million, with stamp duty contributing 45 percent, registration 6 percent, and other land revenue taxes, including the land tax, almost half. The principal tax base of the property transfer taxes is the gross selling price, irrespective of any other factors, including whether or not any gain since the purchase has been made. \ 9.75. The base for stamp duty includes virtually all kinds of transfers and legal documents. For example, more than 120 items are subject to stamp duty in NWFP. The stamp duty rates vary by type of document, but most revenue is collected from stamps on property transfers, where NWFP levies a 3 percent rate. A new schedule of stamp duty rates is under discussion. In Punjab the stamp duty rate is 2 percent on the value of property transfers. Stamp duty reform has been under discussion in Punjab, including a reduction in the tax rate on property transfers. \ 9.76. There are two forms of property transfer: mutation and registration, and property transfers are subject to an array of tax rates and special fees. A mutation involves a transfer of property when there is a record of ownership but not necessarily a deed (Box 9.9). This form of transfer used in rural areas where most land has been inherited and patwaris maintain mutation register (Box 9.10). Registration transfer is the only option for most urban purchasers, as organized land records typically do not exist and mutation register are not maintained. Whenever land records do exist, as in part of Peshawar old city, registration is preferred, since a registered deed comes with a presumption of truth under the evidence laws. Due to deterioration of land records, the courts do not always accept a certified mutation as a basis for establishing ownership. Transfers by inheritance and those transferred under court decree pay only a Rs. 200 fee that is applied to any mutation. Otherwise, mutation transfers and registration transfers are subject to a tax on the value of the transfer of 1 percent in Punjab and 2 percent in NWFP, a stamp duty of 2 in Punjab and 3 percent in NWFP; a local tax of 1 percent in Punjab and 0.5 percent in NWFP, and a federal capital value tax of 2 percent. Hence the total rate paid on a property transfer is around 6 percent in Punjab, and 7.5 percent in NWFP. 9.77. While the system of mutation, registration, and stamp duty is much the same in the two provinces, there are some differences in terms of additional taxation of land imposed in NWFP. This includes Ushr, a tax based on the volume of farm produce, which BOR collects and turns over to the district Zakat account to fund welfare projects and assistance to the needy. NWFP also imposes a land tax, based on the area of cultivated land owned. This tax is part of the general revenue in the provincial government budget, but is a minor revenue source. Figure 9.10: Trends in Land and Property Transfer Collection in Punjab and NWFP Land and Property Transfer Collection in Punjab and NWFP in 2005/06 Land and Property Transfer Taxes (Rupees Per Capita, 2005/06 Prices) 140 40 200 Rupees Per Capita 36 180 120 Punjab 32 160 100 28 140 24 120 80 20 100 60 % of Own-Source 16 80 Revenue 40 12 60 8 40 NWFP 20 4 20 0 0 0 Punjab NWFP 1999/2000 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 139 Issues 9.78. Revenue collection is low and has not increased in line with rising property values. With both urban and rural property values increased considerably for most of the 2000s, revenues from the various land taxes have barely kept pace with inflation (Figure 9.10, right panel). 9.79. While the tax base of the property transfer tax is the market value of the property transfer, property transfers appear to be often undervaluated in rural areas. In rural areas in Punjab, the buyer and seller of a property transfer transaction make a declaration of the value in consultation with the patwari. The law provides for a right of preemption, i.e., neighbors may buy the land for a price higher than the stated declaration. While these checks are supposed to ensure a proper self-declaration of land value, they are typically ineffective and the declared selling price is usually accepted as the base of taxation. In NWFP, the base for the tax is not declared value but a notional value that is drawn from a table of declared values in the past year by type of land. While this table is certified by a district committee, the table is generally considered to be out of date. There are no sales ratio studies for independent appraisals of market value to enable an estimate of undervaluation. The only supporting evidence for rural areas in Punjab and all areas in NWFP is the ausat yaksala prepared by the patwari which is an average sales value for different types of land in a recent period. In compulsory land acquisition cases, the courts have routinely refused to accept the annual average value worked out from patwari records and have awarded higher compensation to land owners. A Supreme Court judgment in the 1990s ordered that government should pay on the basis of market values instead of on the basis of the patwari's ausat yaksala. 9.80. Land in urban areas appears to be equally undervalued. Both provinces use a valuation table in urban areas. The valuation tables are prepared by the district collector, who is head of the land revenue department. These values are based on a combination of subjective, expert judgment and objective price information, and are to be updated regularly. In practice, the approach is more ad hoc. In Punjab, the land values in the table are reported to be the same as those laid down in the (repealed) wealth tax act of 1963. Moreover, the table does not include the value of structures. In neither province is there any coordination between the Board of Revenue and the excise and taxation department. Hence, the transfer taxes yield only a fraction of their revenue potential at current rates. 9.81. A second problem is that the tax base has been narrowed by exemptions given for political or administrative reasons. Officials in NWFP offered the following list of issues, but much of this would be applicable to Punjab as well. · Large tracts of land in some districts have not been `settled', which means that detailed records for them have not been prepared. They remain outside the tax base. · PATA districts in NWFP are largely outside the land tax base because of their legal status. The tax laws have not yet been extended to cover them. · Small farms are exempt. Due to subdivision of land, through inheritance, the base is shrinking. · In rural areas, the Patwari's office is overburdened with duties which do not relate to tax collection. Without staff, office or mobility, the patwari is dependent on local support for performance of his official duties. The result is a significant delay in recording property transfers and collecting taxes due. 9.82. Third, the record keeping is poor. An updating of the "record of rights" is required every fourth year, but they are often not updated on time. The records in Lahore appear to be 20 years out of date. While the patwari maintains detailed records of land ownership, land characteristics and land use in rural areas, they are less common in urban areas; and even if they exist, they are not reliable for the physical characteristics of properties because these properties have rarely been surveyed. The low quality of these records has led significantly to civil litigation. In the absence of any formal property records, the PT-1 register of the excise and taxation department is the only record of these properties that might give evidence of ownership, although this fiscal register does not give any proof of ownership. 140 Box 9.9: Mutations and Registrations Mutation is applicable primarily to the transfer of agriculture land. For a mutation, the transfer is reported to the patwari who enters it in the mutations register where he notes the reference in the record-of-rights and the number of the parcel. The mutation is then certified in a public gathering by the Tehsildar when he visits the patwar circle. In general this happens once a month. Upon certification by the Tehsildar in the mutation register, the transfer becomes final. The patwari then makes a red ink entry in the records-of-rights register. This entry forms the basis of changes in the ownership column of this parcel of land when the next edition of the record-of-rights is prepared. The editions are prepared every four years. Mutations are carried out for all forms of transfer, including inheritance, sale or in compliance with a civil court decree. The tax treatment varies by type of transfer. Registration, applies mostly to urban property, for which the records-of-rights do not exist. For registration, the transfer deed is prepared on a stamped paper and attested to by the registrar of properties. The registrar maintains a record of the deeds but there is no comprehensive record of properties in most urban areas. If a person wishes to carry out registration for agricultural land, he is allowed to do so as an additional step after the mutation. Box 9.10: The Patwari The patwari is the junior most official of the BOR tax collection administration. He is underpaid, does not have any staff or transport and in many cases does not have a government-provided office. The bundle of registers he is required to maintain mostly is wrapped in the cadastral map or latha, when he needs to carry it around. In this sense he has a mobile office. A patwari's importance belies his junior standing. He has numerous important duties. (1) He is the custodian of land records. In rural lands, and some urban areas, a presumption of truth is attached to the records of rights in the courts. Invoking these records, titles are determined. (2) The record-of-rights is updated every four years under the law. The patwari writes the record with his pen, specifying the titles, occupancy and other rights in land. The four yearly editions are maintained by him. (3) He keeps a `Mutations Register', which is linked with the four yearly record-of-rights register. The process of mutation is initiated and completed with the patwari. He enters the details of property, records the transaction and upon attestation by a Revenue Officer, makes a cross entry in the record-of-rights register. (4) He assigns a value to the property when recording the transaction and determines the liability for taxes on property transfers. (5) The patwari maintains a Crop Register or Khasra Girdawari. In this register he records the results of his six monthly crop inspections for each parcel of land, also recording the ownership and occupancy. (6) For Valuation Tables prepared for any land-related taxes he provides the estimates--the ausat yaksala. (7) Estimates of crop yields, jhar padawar, are prepared with inputs from patwari. (8) He issues the ownership certificate, fard malkiyat, required for many basic citizenship documents and mortgage and banking requirements. (9) He certifies ownership for agriculture loans. (9) In many cases he issues a "no-dues certificate". (10) With the assistance of the village lambardar, he collects land related dues. (11) In cases of natural disaster, he is dispatched as the vanguard of assistance to assess damages and thereby recommend compensation. (12) A patwari prepares records and notifications for acquisition of land for government or large private projects, determines titles, prepares valuation estimates and certifies payment to owners. (13) Informally, the local officials count on him for miscellaneous functions including public meetings, arrangements of some official functions and data gathering. (14) Due to his role in land administration, the patwari is sought after by local elite and he in turn can count on their support. His accountability to the department leadership is limited. (15) He maintains land records in Urdu which are based on antiquated survey techniques. A patwari will determine the boundaries of a certain property if there is a dispute. (16) He is called to attend courts and render evidence in cases disputed title, possession, partition and compensation. As the number of land-related litigation cases has multiplied, increasing amounts of his time is spent in court. This has the result of his spending less time on maintaining records and carrying out field inspections, setting up a vicious circle. 141 Reform Options 9.83. The reform of the land and property transfer taxes should be part of the rationalizing the overall tax system. The changes would aim to ensure that the revenue flow is adequate to meet the revenue target, the tax burdens are fairly spread between urban and rural residents, and undesirable behavioral effects are avoided (Box 9.11). In the short run, some will argue for an increase in property transfer tax rates as a revenue enhancement measure. This is not a good solution to the revenue problem, and it is not good tax policy, because it ignores or even exacerbates the underlying valuation and compliance problem. The alternative route, indexing the existing rates to some indicator of average property value growth, would penalize properties experiencing slower growth in favor of those experiencing faster growth, and may thwart the real estate market dynamics. 9.84. A crucial element to any sustainable reform is to improve the method of valuing property transactions so as to better approximate true market value. There are a number of necessary actions that must be taken to eliminate or at least reduce the gap between assessed and market value (Box 9.12). In addition, simplification in the tax structure should be sought, especially given the limited administrative capacity of the provincial government. An obvious change is to collapse the three different provincial property transfer taxes into one levy. This would simplify the tax regime and make it more transparent and manageable. While there are no reliable sales-ratio studies available to estimate the revenue impact of better valuation, anecdotal evidence suggests that a 50 percent undervaluation of transferred property might be a reasonable, even though perhaps conservative, estimate. In this case, a 15 percent increase in taxable values, and tax collections, would be well within the reach of provincial governments if they chose to make the necessary administrative improvements. 9.85. A more comprehensive reform of the taxation of the rural sector might pull back from taxing property transfers altogether. The rationale for this proposition is straight forward. Provincial taxes might be seen as a payment for public services received. It seems more reasonable to extract this payment yearly rather than at the time of a transfer. Moreover, there might be some administrative economies since UIPT and the agricultural income tax are annual taxes, and since a cadastral survey is required in any case. The present regime of taxing only transfers could be replaced with an annual tax on land in rural areas, perhaps imposed as a presumptive income tax on rural land. The umbrella for this new tax would be the agricultural income tax in the case of lands used for agricultural purposes, as discussed further in the next section. For rural lands not used in farming, it would be a tax on annual rental value at highest use. Valuation would be carried out following the steps noted in Box 9.12. Urban land would continue to be taxed under the UIPT. 9.86. The restructuring could be done in three steps. First, the mutation fee, registration fee and stamp duty on property transfers would be combined into a single land tax, levied on the basis of the market value of land. The rate would be set according to revenue considerations. Second, in order to move to a unified annual tax, it will be necessary to prepare a cadastre showing the ownership and physical characteristics of every parcel. After this, a valuation table would be prepared for the province, and updated on a periodic basis. The revenue yield from a unified annual tax on urban and rural property will depend on the nominal tax rate that is chosen. The elasticity will depend on the accuracy and timelines of revaluation and the indexation of property values. A unified, annual land tax with "reasonable" rates and proper valuation could easily more than double rural land tax revenues. 9.87. A capital gains tax on land could be introduced in addition to the annual land tax proposed above. The tax liability would be taken on the difference between the buying and selling price of the property, indexed for inflation. The buying price would be set according to historical records of purchase price. Owners could petition to have this base increased. The selling price would be verified by the valuation staff in the capital gains tax office. The nominal gain would be adjusted for inflation and for the cost of allowable improvements to the property (e.g., irrigation). A major drawback is administration, especially establishing the basis for a capital gains tax, and developing a method for indexing for inflationary increases and adjusting for qualified investments in real property. As complicated as these administrative requirements seem, they are not any more daunting than the present requirement of establishing a true market values for every real estate transaction. Once a capital gains tax becomes established, the unified administration would be more efficient at collecting revenues than is the case under the present regime. The incentives for under-declaration of property prices at the time of sale would be lessened. The better flow of information about land prices would help to develop a property market. 142 Box 9.11: International Experience with Land and Property Transfer Taxes Is Pakistan's present regime of land and property transfer taxes suitable in light of international practice (Bahl 2004)? Stamp duties are a part of the tax system in most countries, but they are widely criticized. The fundamental question to be addressed in Pakistan (and elsewhere) is "why have a stamp duty?" Clearly there is need to legalize documents and assure that they are properly filed, and a government stamp is one way to do this. To levy a service charge that would cover the stamping and verification cost would seem a reasonable justification. There might be some justification for differentiating the rate of charge by type of document, given the different degree of examination required for various types of documents. The problems arise with respect to property transfers, such as the regime that includes stamp duty, mutation fee and registration fee in Pakistan. There are a number of reasons why many developing countries rely on real estate transfer taxes. First, it is an easy tax handle in that most buyers and sellers desire to legally record the transfer and therefore will voluntarily comply. Second is the revenue motivation and what might appear to be a very low cost of collection. When property values escalate, as happens periodically in most countries, the revenue take can be quite significant. Third, if property ownership is concentrated in the higher income classes, the distribution of tax burdens may be progressive. Fourth, the number of people in the taxpaying population in any given year is much smaller than in the case of more general taxes, hence the opposition to the tax may not be as great as would be the case if, say, an increase in the value added tax were proposed. Finally, a property transfer tax might reach that part of the population that ordinarily avoids payment of most income tax and value added taxes. There are major disadvantages to the property transfer tax. First, it raises the cost of property transactions thereby reducing the volume of transactions, hence slowing the development of the real estate market. Second, if the tax is properly assessed, administrative costs could be high, at least because of the need to check the self-reported property values and revalue when necessary. Third, a property transfer tax gives property owners an incentive to understate taxable value, hence weakening the database that is called on for assessment of the urban property tax. The problems with the property transfer tax are dependent on the level of the nominal tax rate chosen. At very low rates, these problems may be of less consequence. But when the tax rate is high, the implications of these problems are magnified. In fact, countries choose very different rates of taxation on the value of property transfers. South Africa, for example, taxes property transfers and also subjects these sales to a 14 percent VAT. But in many countries, the rates are below those levied in Pakistan. The reform options most often seen for property transfer taxes in recent years are reductions in the rate to mitigate the problems described above. Recent examples include Czech Republic, Portugal, Slovakia, Taiwan, and Dominican Republic. Box 9.12: Tackling the Problem of Undervaluation 1. Carry out a survey of all properties so that a full cadastre is available for compiling the tax roll. 2. Subject all declared transfer values to independent valuations carried out by the BOR. 3. Increase the staff of qualified property assessors used in updating the tax roll and in making the assessments. If the goal is to subject all property transfers, or even all suspect declarations, to desk and possibly field verification, a significant (and costly) increase in staff will be called for. 4. Settle on a method of valuation and develop the appropriate procedural manual. Given the present state of valuation in the two provinces, this method will likely involve making use of a combination of expert judgment and objective information (e.g., realtor opinions and the valuation data established in urban areas for UIPT). 5. Carry out regular sales-assessment ratio studies to monitor valuations and to update the valuation tables in both urban and rural areas. This could be done by the increased staff, and with the use of expert opinions. 6. Enact a substantial surcharge on tax due in the case of declarations that are found to be well below market levels based on independent valuation. 7. A potential check on valuation might be made by coordination with the valuations placed on urban properties by the excise tax department for purposes of UIPT. At present, there is no such coordination. 143 9.7 Agricultural Income Tax Base and Rate 9.88. While agricultural income is exempt from federal income tax, both provinces impose the tax, although as a presumptive tax rather than as a tax on agricultural income. NWFP started levying a land tax and agriculture income tax (AIT) in July 2001, and is credited with some of the early policy development of the tax. According to the provincial laws, agricultural income is defined as rents received from property used for agricultural purposes, income derived from cultivation, or income derived from owner-occupied buildings on the property (including small businesses such as dairies). The AIT raises little revenues in Punjab and NWFP. Punjab obtained in 2005/06 only 2.1 percent of total own source revenue. This is an average of Rs. 170 per farm, or Rs. 1,669 per farm greater than 12.5 acres in size, based on the 2000 Census of Agriculture. Real per capita collections fell from 1999/2000 to 2005/06 by two-thirds (Figure 9.11). NWFP collected only 1.5 percent of total own source revenue through AIT, implying an average burden of Rs. 52 per farm, or Rs. 241 per farm greater than 5 acres. 9.89. Even though the laws provides for two alternative calculations of the tax, one based on land area and the other one on agricultural income, the land area method is the one actually used in both provinces. For the income method, the land revenue staff would have to establish an average yield per acre for all crops cultivated in each revenue unit of the district, which is then certified by the district collector. Input prices and market prices for crops would be based on surveys and analysis conducted as part of the Census of Agriculture. However, the lack of accurate, timely, and detailed data on ownership of land, crop yield, and input prices leads to a situation where there is little basis to establish the AIT liability based on income. Moreover, the land owners do not typically file an income tax return which reduces the ability to track tax liability. Hence, in practice, the acreage method is used in both provinces. The laws provide detailed schedules for unirrigated and irrigated land. In Punjab, irrigated land of less than 12.5 acres is exempt, and non-irrigated land with less than 25 acres is exempt. In NWFP, there is no exemption level for the land area method. The tax is based on sown land area, but the valuation process is more focused on land tenure than on crop inspections. While the patwari is supposed to perform crop inspections every 6 months, negotiation between the land owner and the patwari can substitute the inspection. 9.90. In Punjab, the tax rates have been constant since 1996 but the exemption level has changed over time. The per acre tax rate structure is not a graduated marginal rate structure. Instead, all acreage of farm land in a rate class is taxed based on the same slab rate. For example, for a parcel of 50 irrigated acres, the tax rate is Rs. 250 on all 50 acres. This leads to a notch problem with large changes in the tax rate at each size boundary and may in fact encourage individuals to understate the size of their properties. The acreage-based structure in NWFP does not have this problem. The tax rate rises from Rs. 50 per acre for less than 5 acres, to Rs. 72 per acre with 5 to 12.5 acres, and to Rs. 100 per acre with more than 12.5 acres. The reference acres are doubled for unirrigated land. 9.91. The AIT is collected by the Board of Revenue in both NWFP and Punjab. To the extent they exist, records on land holding and crops are maintained in duplicate; one copy is with the patwari and the second is kept at the district office. The patwari is in charge of all of the land records at the local level. The district collector is a registrar for the district as well as a tax collector, but in effect the patwari carries out most of the record keeping. Figure 9.11: Trends in Agricultural Income Tax Collection Agricultural Income Tax Collection in Punjab and NWFP in 2005/06 Agricultural Income Tax (Rupees Per Capita, 2005/06 Prices) 24 8 2.2 Rupees Per Capita 22 2.0 7 20 1.8 6 18 % of Own-Source 1.6 Revenue 16 5 1.4 14 1.2 4 12 Punjab 1.0 10 3 0.8 8 0.6 2 6 NWFP 0.4 4 1 0.2 2 0 0.0 0 Punjab NWFP 1999/2000 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 144 Issues ­ Low Revenues 9.92. The effective rate of agricultural income taxation is very low. The AIT provides very little support to the provincial budgets despite the fact that agriculture is one of the largest sectors in the provincial economy. The revenue potential of the agricultural income tax is much greater than its current yield. We have made estimates of this revenue gap using agricultural census data. In Punjab, we estimate the potential revenue under current law to be Rs. 2.9 billion or more than 4 times the actual collections in 2000/01. For NWFP, we estimate potential collections under current law of Rs. 249 million, or 10 times the 2000/01 level of collections (Figure 9.12). Malik (2004) reports much the same result for NWFP in an earlier study. While these estimates are rough, they do suggest that the orders of magnitude of evasion, compliance, and administrative problems are very large. 9.93. One major reason for these revenue shortfalls is that the tax base has been significantly eroded by exemptions in the case of Punjab. The level of exemption in Punjab (12.5 acres) significantly narrows the tax base. Based on data from the Agriculture Census 2000, about 85 percent of all properties are outside the tax base. While some provincial level officials in Punjab took the position that 12.5 acres of irrigated land was well above subsistence and there was room for taxation, others said this was not the case. They also pointed out the political difficulties inherent with increasing this tax. We simulate the potential revenue associated with reducing the exemption in Punjab from 12.5 acres to 5 acres. This would yield potential AIT revenue of Rs. 4 billion but would tax some farmers with incomes less than Rs. 80,000. If the relative compliance rate and administration costs were similar for large and small farms before and after this change, this would increase actual AIT revenue in 2005/06 from an estimated Rs. 658 million to nearly Rs. 910 million in Punjab. 9.94. Another problem with the present structure is that it makes no allowance for the fact that some crops are more lucrative than others. For example, sugar cane yields 15,000 net rupees per acre, while many other crops net only 8,000-9,000 rupees per acre (Agricultural Policy Institute 2004). This makes the tax system unfair. 9.95. The system of administration, while well documented in law, is not enforced in either province. Observers point out that neither the revenue officials nor the taxpayers are adequately familiar with the legal requirements. The required records are not maintained so that holdings and crop values are not updated. While there is information available to estimate crop value, gross income is typically determined on an ad hoc basis. Figure 9.12: Estimated Additional Potential for AIT Collection in 2000/01 based on the 2000 Agricultural Census Punjab NWFP Additional Potential Additional Potential (983%) (332%) Actual (100%) Actual (100%) 145 Issues ­ Is Agriculture Undertaxed? 9.96. There is a broader question of whether the agricultural sector is "undertaxed" in Pakistan. While agriculture contributes about one-fifth to national income, the sector "pays" a very small percentage of tax revenue. Chaudhry (1999) refutes the underpayment argument and points out that there are implicit taxes on agriculture such as those associated with price controls for many farm products. Khan (1999) points out that the incidence of all taxes--implicit or explicit--is on individuals and the focus should be the incidence of the tax on consumers, labor, or capital and not on sectors. By this argument, the concept of the "tax burden of the agricultural sector" does not have a great deal of meaning. Nevertheless, it can be indicative of inequities and inefficiencies in taxation. 9.97. The implicit taxation of agriculture in Pakistan has come down over the last decades, following worldwide trends. These "taxes" come via macroeconomic policies as well as explicit taxes on the agricultural sector. Many countries protect their industrial sector--artificially raising prices of certain agricultural inputs, while controlling output prices for the agricultural sector below world market prices. In addition, in many countries, protectionism has led to increases in the exchange rate which in turn reduced the competitiveness from agricultural exports. Chaudhry and Kayani (1991) calculated the implicit taxes on Pakistan's agriculture between 1970 and 1990 and found effective rates as high as 75 percent, with wide variation among crops and years. However, the net taxation of agriculture has fallen sharply in most countries during the last twenty years (World Bank 2007), and the effective tax rate on agriculture in Pakistan has also come down (Box 9.13) (Dorosh and Salam 2007, World Bank 2007). The latest developments in the wheat sector, which brought domestic prices more in line with international prices, the rise in agricultural commodity prices, as well as the depreciation of the Rupee, suggest the net taxation of agriculture has further declined in the last year. 9.98. The case for increasing the level of taxes raised from agricultural income relies both on equity and efficiency. From a horizontal equity perspective, there seems to be some justification for a higher level of agricultural income tax. Agricultural income is exempt from income tax, so non-farming households at any given income level face a higher tax burden on their labor income than do agricultural households. Secondly, because agriculture is such a large part of Pakistan's economy, not taxing agricultural income means that effective tax rates on other sectors of the economy must be higher. This decreases the economic efficiency of the tax system as taxpayers in these other sectors seek ways to avoid or evade those taxes. Third, the agriculture income tax may be justified on the principle that government infrastructure provides benefits to agriculture. This sector also benefits from low interest loans, fertilizer subsidies, and public research and development. Though a fully-fledged cost- benefit analysis would be required to firm up the argument, these flows of benefits could also justify increasing the agricultural income tax. Box 9.13: Changes in the Net Taxation of Agriculture World Bank (2007) and Anderson (2007) report that in three groupings of countries (agriculture-based, transforming, and urbanized) macro and public finance policies reduced the effective rate of taxation on agriculture in 24 out of 28 countries. In Anderson's analysis, higher explicit and implicit taxes yield a negative "nominal rate of assistance", where subsidies or protection of the sector yields a positive nominal rate of assistance. Most urbanized countries now provide a positive rate of assistance while the rate of assistance in most transforming countries is negative: Estimates of Nominal Rate of Assistance Country 1980-84 2000-04 Indonesia +10 to 15 +35 India +0 to 5 +15 Thailand -5 to -1 +5 to +10 Malaysia -10 to -5 +3 to +5 China -50 +1 Pakistan -18 to -15 -5 to -2 Sri Lanka -10 to -8 -7 to -3 Egypt -20 to -15 -15 to -10 Senegal -30 -18 to -15 Zimbabwe -45 -70 or less 146 Reform Options 9.99. Proposals to strengthen the AIT revenue collection and to make the AIT income-based are not new and not uncontroversial (NTRC 1986, Coopers and Lybrand 1989, Chaudhry 1999, Khan 1999, and Malik 2004). One major change is that the provinces could retain the basic approach of taxing the farm size, but augmented with a presumptive assessment based on crop type and expected net income. In addition, rental income of sharecroppers could be taxed, but there is little evidence that information on rental values and tenants is readily available. Since there is a mechanism to measure crop output and values (as flawed as it might be), it may be administratively easier to estimate potential value of output. 9.100. The main two types of options for a reform program are land-based presumptive options or income based presumptive options. Land-based presumptive reforms have the advantage of simplicity, but they may perpetuate a presumptive system and prevent the AIT from growing into a "true" income-based tax in the long-run. Three land-based reform components could be: · Reduce the exemption level in Punjab which currently exempts 85 percent of all land. An alternative would be to have no exemption and to charge a nominal (flat) amount for small farms. This would raise administrative and compliance costs but would expand the tax base to cover the entire sector. · Second, the tax rates could be increased. The new rates could be set according to revenue targets, or to redress sectoral imbalances, or to gain parity with taxpayers in other sectors of the provincial economy or among those with similar levels of income. · Third, a more progressive rate schedule could be developed. 9.101. Income-based presumptive reforms have the benefit of keeping the AIT a direct tax on income. This would allow such a tax to be ultimately integrated with the general income tax. However, given the constitutional issue related to the taxation of agricultural income, it might be difficult to treat agricultural income like other forms of income for tax purposes. A reform could look as follows: · Define a presumptive net income amount by farm, based on farm size and crop type. The rates could be flat or variable. A subsistence level of income could be exempt from tax, similar to the personal income tax exemption of 100,000 rupees. 9.102. Using data from the Agricultural Census, Board of Revenue, and Agriculture and Crop Reporting Center, we can estimate the revenue impact of these proposals under the assumption of full collection. In the three simulations, the revenue increases range from 11 percent to 27 percent in NWFP, and from 10 percent to 37 percent in Punjab (Box 9.14). 9.103. Whatever structural reform is envisioned, a major reform risk is that no reform will yield substantial increases in revenue if the administration of the system is not enhanced. In particular, self-assessment should be considered as part of the reform process, but, until the administration is upgraded, it seems unlikely that self assessment would reduce the inequities in the system or increase tax revenues. Administration reform options include the following: · Creation of a withholding system for the AIT, where tax is withheld on the purchase of agricultural inputs, with an exemption limit for small farmers; or alternatively for the sale of cash crop outputs · Inclusion of estimates of costs of inputs and gross receipts by crops, as the produce index unit system, into the records (Khasra Girdawari). 147 Box 9.14: Revenue Impact of Three Scenarios for the Reform of Agricultural Income Tax We consider two land-based reform scenarios (options a and b) and one income-based reform scenario (option c). Option a. Progressive rate structure (independent of crop): 7.5 acre exemption for irrigated or unirrigated land; 7.5 to 25 acres taxed at 100 rupee per acre, 25 to 50 acres at 150 rupees per acre, 50 to 150 acres at 200 rupees per acre, and Rs 300 per acre for farms greater than 150 acres. Unirrigated land would be taxed at half the rate of irrigated land and orchards would be taxed at 300 rupees per acre. Our simulations by farm size give a result of an 11 percent revenue increase in NWFP and a 21 percent increase in Punjab. Option b. Flat rate structure by crop type: 7.5 acre exemption for all farms (except orchards); wheat taxed at 100 rupees per acre, cotton taxed at 200 rupees per acre, rice taxed at 300 rupees per acre, and sugarcane taxed at 350 rupees per acre. These amounts were chosen based on the relative level of profitability of these crops. Unirrigated land taxed at 150 rupees per acre regardless of the crop. Orchards would be taxed at 300 per acre. Our simulations show that revenues would increase by 27 percent in NWFP and 37 percent in Punjab. Option c. Tax presumptive net income based on crop yield and profitability, with exemption of 100,000 rupees and progressive rates from 5 to 15 percent. Our simulations show that revenues would increase by 20 percent in NWFP and 10 percent in Punjab. Relative to current law, option a increases revenue by 11 percent in NWFP and by 21 percent in Punjab. The resulting distribution of tax burden is more progressive than the present system under this reform scenario. Option b imposes a tax structure based on farm size, but adjusted for crop type. Adjusting the land-based rates by type of crop is admittedly a more complicated way to go. However, this type of reform brings the AIT closer to a tax on potential income than the tax structures based on land alone. We estimate this approach considering four crop types: wheat, cotton, rice, and sugarcane. For these purposes, we collapse other grains and fruits and vegetables into the wheat category. These categories can easily be changed. The revenue impact for NWFP is an increase in AIT of 27 percent and the increase in Punjab is 37 percent under this structure. Finally, under option c, we estimate the revenue potential of a tax based on potential income, where income is estimated based on the average profit/loss including land rent per crop, per acre. To simplify the analysis, along the lines proposed in Malik (2004), we estimate the potential revenue assuming an average net profit per acre of Rs. 4,000 (lower bound of all crops) and Rs. 8,000 per acre (an upper bound estimate of the average crop yield). With a threshold of Rs. 100,000 for tax purposes, on average, only farms of 25 acres or more would be taxable. The revenue analysis suggests that this structure could yield Rs. 280 million in NWFP (a 20 percent increase) and approximately Rs. 2.9 billion in Punjab (a 10 percent increase), assuming that such a tax could be fully administered. The smaller increase in Punjab is due to the relatively large number of farmers that would be exempted due to the threshold set at Rs. 100,000. Revenue Impact of Illustrative AIT Reforms (% Increase in 2005/06 AIT Collection) 40 Punjab 35 30 NWFP 25 20 15 10 5 0 Progressive, crop-independent Flat rate structure by crop type Presumtive income tax (option c) rate structure (option a) (option b) 148 9.8 Professional Tax Description and Issues 9.104. The professional tax, or the "tax on professions, trades, and callings, is levied as a flat amount but this "rate" varies by type of profession. In both provinces, the tax accounts for only a minimal share of the budget in 2005/06: 1.6 percent of own source revenues in NWFP and 0.8 percent in Punjab. Real per capita collections increased in NWFP due to rate increases, while collections in Punjab declined (Figure 9.13). 9.105. The tax may fall on the corporate entity ("companies registered"), on self-employed professionals (doctors, lawyers, marriage hall entrepreneurs, etc.), and on employees. This definition of the base gives the professions tax some characteristics of a business license fee for companies and self-employed, and an additional wage tax for individuals employed in these sectors. A large proportion of the tax is paid by salaried government workers. Each province has the authority to develop its own list of professions, but the list of professions is similar in NWFP and Punjab. In Punjab, there are six general categories of "profession", with a fixed amount of tax assessed to each category. Each category has multiple subcategories so that in total 47 "professions" are taxed. NWFP lists 11 main categories. The specific rates range from Rs. 200 to Rs. 100,000 in Punjab and from Rs. 100 to Rs. 50,000 in NWFP. The tax rates were last changed in 2002/03 in both provinces. An analysis of Lahore data can give some indication of the relative importance of the different categories of profession in current revenue collection. Based on 2006/07 data from the Punjab Excise and Taxation Department for Lahore District demonstrates that four-fifths of the revenue take is from employed individuals (some of whom are employed by "paying" companies) and directly from companies, and about half of the revenue from persons engaged in a profession, trade, calling or other employment subject to the income tax. 9.106. In both provinces, the excise and taxation department is in charge of the collection of the professional tax. It builds up the tax rolls with the help of data sources such as income tax statistics, limited company directories, chamber of commerce, and professional association lists. 9.107. The revenue potential of the professions tax is far larger than what is collected. The Pakistan Labor Force Survey suggests that for four occupational categories only (professionals, technicians and associate professionals, and legislators, senior officials and managers), the revenue potential of the tax would be Rs. 1.3 billion in NWFP and Rs. 6.3 billion in Punjab. In both cases, the analysis suggests that current collections are less than 4 percent of revenue potential. In addition, the tax is levied at such low rates that it would likely fail on the revenue raising objective, even if administration were efficient. For example, a physician might earn Rs. 20,000 per day, but pay only Rs. 1,000 per year in professions tax. In such a case, the tax is little more than a nuisance. 9.108. The rationale of the professional tax is unclear. The professions tax might be justified on the basis of benefits received from provincial government services, and as a tax paid by a business, employee or self-employed person for the privilege of doing business in the province. A benefits tax approach would indicate transforming the professions tax into a business levy, or into a piggyback on the federal corporate income tax. In this case, it is not obvious that there should be an additional fee levied on individuals employed in these businesses, while currently some professionals are taxed twice--once through an official employer and then again as a professional employee. Alternatively, the professional tax might be an attempt to pursue equity objectives by taxing higher paid professions at higher rates. However, the variable tax rates across professions do not align with the ability to pay and the individual income tax is much better equipped to pursue equity objectives. 9.109. Administration is very weak. In Punjab, about 35 lower level officers are working on the labor-intensive assessment and collection of this tax, yet there are no designated professions tax officers. In Lahore district alone, there are over 250,000 taxpayers on the tax rolls. In Peshawar district, there are only 7 inspectors assigned to cover the tax on professions, implying that each inspector has the impossible task of supervising about 10,000 tax units (Khan 2004). Professionals not tied to an incorporated business are notoriously difficult to find for purposes of tax administration. The sheer number of professionals (2.7 million in NWFP and 12.5 million in Punjab) make an efficient administration of a stand-alone tax a costly proposition, especially without an integrated data base that captures professional registration, motor vehicle registration, government contracting, and other such activities. 149 Reform Options 9.110. There are several ways to improve the practice of taxing professionals at the provincial level. The first option is to consider the professions tax as unproductive nuisance and to abolish it as provincial tax. It might then be turned over to the TMAs or union councils and be treated as a kind of general charge for the privilege of doing business in the local area. However, weak administrative capacity would remain a crucial bottleneck. 9.111. The second route is to take the view that there is need for a more broad-based provincial tax to reach the growing fiscal capacity, particularly in urban areas. The professions tax could be converted into a piggyback levy on the federal income tax with collection by the federal government. Provincial governments could choose a piggyback rate within a prescribed range and the federal government could collect the income tax as it does now. For example, a three percent piggyback on collections for individuals and self employed in Punjab could yield Rs. 487 million per year and Rs. 83 million in NWFP--both substantially more than current collections. Objections to this option are that it could be struck down as a "double tax" particularly on workers in compliant enterprises, and unconstitutional. Also, increasing the overall income tax rate, even by this small amount, might reduce compliance. Another objection would be that under this scheme the self-employed would escape the local tax net altogether since they often evade the federal income tax. The self-employed are after all notoriously hard to tax. To deal with this, the piggyback income tax might be supplemented by a presumptive income tax administered by the province. Even if the presumptive tax on the self-employed was not effective, it would be no worse than the present professions tax administration where 80 percent of collections come from those already in the income tax base. This option has the most potential given current administrative capacity. 9.112. A third reform route would convert the current professions tax into a sales tax on services. Many professionals provide some type of service although they are not all covered under the current tax on services. A drawback with this option is the headquarter issues, as the headquarter office of the firm that delivered services in a particular provinces might be in another province. In addition, services are difficult to tax, although the collection rate of a sales tax on services probably would still be higher than that for the current professions tax. At the end, it would be preferable to tax all service under a comprehensive GST on goods and services. 9.113. A fourth option would be to increase the specific rates and bring them more in line with expected levels of income for the various professionals, and to eliminate the tax on salaried individuals who are employed in companies. While this option has its merits (Khan 2004)--it is an income-based tax that has its roots in the current professions tax--the provincial authorities would have to commit more administrative resources, including training, performance evaluation, and data base development. First steps would be a survey to enumerate all of those professions liable for the tax, and an automation of this data base. 9.114. A fifth option would be to restructure the tax as a benefits charge for services provided by the provincial government (e.g., road infrastructure, certain utilities). In this case, the tax might be levied on businesses and not on employees of taxpaying entities. A provincial level business registration fee (flat fee or progressive fee based on type or size of business) could be assessed annually as a presumptive tax. The main concern would be weaknesses in compliance, record keeping and administration that may lower revenue collection. Figure 9.13: Trends in Professional Tax Collection Professionals Tax Collection in Punjab and NWFP in 2005/06 Professionals Tax (Rupees Per Capita, 2005/06 Prices) 4.0 1.8 12 % of Own-Source Revenue 3.5 1.6 10 1.4 Punjab 3.0 Rupees Per Capita 1.2 8 2.5 1.0 2.0 6 0.8 1.5 0.6 4 NWFP 1.0 0.4 2 0.5 0.2 0.0 0.0 0 Punjab NWFP 1999/2000 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 150 9.9 Other Taxes 9.115. Both provinces levy more or less the same set of "smaller" taxes (Figure 9.14). In a few cases, the tax base calls for some differentiation, e.g., tobacco in NWFP and cotton in Punjab. Notwithstanding that these are minor levies, their potential is not being fully exploited in part due to their narrow base and in part due to weak administrative efforts. The sum of hotel tax, entertainment tax, excises, and tobacco and cotton levies and other agricultural cesses, yielded an amount equivalent to only 1 percent of total revenues in Punjab and about 1.5 percent in NWFP. The administrative costs of efficiently collecting these taxes are high. In the case of entertainment tax and hotel tax, the assessment is notional and exemptions have compromised fairness. 9.116. The following reform package would result in only small revenue losses while freeing up provincial administrative resources to be devoted to taxes with higher tax potential: · The hotel tax could be folded into the sales tax on services. This could lead to a revenue gain, even if the smaller hotels are subjected only to a flat charge. The electricity duty might also be moved to the sales tax on services. · The entertainment tax could be passed down to the local governments, since they are in a better position to administer the tax efficiently. · The excise duties also might be abolished, on grounds that revenue yield does not justify the administrative effort. · The agricultural cesses and excises could be folded into the agricultural income tax. This would increase fairness and revenue flow as the reform of the agricultural income tax proceeds. · The tobacco cess in NWFP could be abolished in favor of a surcharge on the federal excise duty on tobacco products. The additional revenue could be returned to the province. Figure 9.14: Collection of Other Taxes in Punjab and NWFP in 2005/06? Collection of Other Taxes in Punjab and NWFP in 2005/06 Other Taxes, (Rupees Per Capita, 2005/06 Prices) (hotel tax, entertainment tax, excises, and tobacco and cotton levies and other agricultural cesses) (hotel tax, entertainment tax, excises, and tobacco and cotton levies and other agricultural cesses) 18 20 5.5 18 5.0 16 Rupees Per Capita 4.5 NWFP 16 14 % of Own-Source Revenue 4.0 14 12 3.5 12 10 3.0 10 2.5 8 8 Punjab 2.0 6 6 1.5 4 4 1.0 2 2 0.5 0 0.0 0 Punjab NWFP 1999/2000 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 151 9.10 Vertical Equity 9.117. The provincial taxes are somewhat progressive. Figure 9.15 shows the distribution of the tax burden for provincial taxes in 2004/05 for Pakistan overall, covering all provinces. It shows the effective tax rates and tax shares by household consumption quintiles, based on the tax incidence model introduced in Chapter 3 and under the incidence assumptions listed in Table 9.6. The top consumption decile accounts for about two-fifth of the overall tax provincial collections. The effective tax rate increases moderately from 0.32 percent of GDP for the bottom decile to 0.36 percent of GDP for the sixth decile, and then rises more sharply to reach 0.54 percent of GDP for the top decile. This result is remarkable, as about 70 percent of the provincial tax collection comes from indirect taxes (excises, stamp duty, motor vehicle tax, and others) rather than direct taxes (the agricultural income tax and tax on professions, trades and callings, property tax, and land revenue) (Figure 9.3). It suggests that an expansion of provincial taxation, while allowing for appropriate thresholds and exemptions to protect poorer households, does not have to worsen the vertical equity of the tax system. Figure 9.15: Vertical Equity of Provincial Taxes (2004/05) Incidence of Provincial Taxes by Household Consumption Deciles 45 0.6 40 0.5 35 Effective Tax Rate (% of GDP) 30 0.4 25 0.3 20 15 0.2 10 0.1 Share in Total Taxes 5 (%) 0 0 Bottom 2 3 4 5 6 7 8 9 Highest Table 9.6: Incidence Assumption for Provincial Taxes Tax Incidence Assumption Allocation (Data) Urban immovable Personal portion: 50% on owners of Value of owner occupied housing, property tax housing, 50% on renters; Commerical rent payment portion 50% on consumption, 50% on owners of capital Agricultural income tax Agricultural workers Income from agriculture (>Rs.100,000) Professions tax Borne by salaried labor Wages of salaried workers Provincial excise taxes Borne by consumers Based on share of consumption of major items Stamp duties Borne by owners of capital Income from property Motor vehicle tax Borne by owners of motor vehicles Purchase of new motor vehicles Others Borne by general population Based on share of income 152 9.11 Policy Reform Revenue Targets 9.118. There are good opportunities for reforming the system of provincial level taxation and fiscal decentralization in Pakistan. A comprehensive reform will require both federal and provincial government involvement. The leadership in program design for this activity might come from FBR and the Ministry of Finance, but any reform of the intergovernmental system must leave significant room for provincial level input and discretion. A properly designed reform could at once allow the federal government to satisfy its objective of increasing provincial level taxes in a framework of good tax policy, and the provincial government to satisfy its objective of enhancing both public service levels and fiscal autonomy. In keeping with the spirit of fiscal decentralization one would not expect all provinces to choose exactly the same reform program. It is imperative that there is clarity on what is inside and what is outside the legal boundaries for provincial revenue mobilization. 9.119. Perhaps most important of all considerations about the proper reform package is its political acceptability. The reform program laid out here is fairer than the present system in that it moves toward treating all taxpayers in the same way by removing tax preferences from many households and businesses. In addition, the reform package includes a net revenue increase of significant amount. Tax reform coupled with tax increase will require a strong political will on the part of the leadership in the provinces. 9.120. The design of the tax reform might center on four elements. Taking the expenditure allocations across federal and provincial governments as given, the first is to determine appropriate targets for provincial tax revenues, i.e., by how much should revenues be increased? The second is to identify structural reforms that would make the provincial tax system operate in a more efficient way, would hit these revenue targets, and would further the decentralization goals that the government has adopted. The third is to put appropriate incentives in place to encourage provinces to implement tax reforms that will generate increased revenues. The fourth is to lay out the elements of a feasible implementation program for these structural reforms. 9.121. The basic goal of this reform is to achieve a significant increase in revenue mobilization by provincial governments. For example, the federal government has called for an increase in provincial taxes to reach a level equivalent to one percent of GDP (Pakistan Economic Survey 2007), which is close to the international norm. There are other systematic approaches to fixing on provincial revenue targets (Box 9.15), but ultimately each target remains to some degree subjective. In developing the revenue profile for a comprehensive reform program, we will work with two targets: scenario "A" is based on the ambitious target of eliminating the structural deficit (the uncovered deficit displayed in the right panel of Figure 9.2) of the provinces, and scenario "B" is based on the more moderate target of meeting an international average of about 1.1 percent of GDP (Figure 9.16). These targets imply increases by 110 and 200 percent for Punjab compared to a 2005/06 collection of Rs. 22.2 billion, and by 175 and 330 percent for NWFP compared to a 2005/06 collection of Rs. 2.8 billion. These revenue targets can be reached with what many would see as "reasonable" structural reforms. Moreover, the reforms outlined here have other beneficial properties in terms of improving the equity of the system, reducing the administrative cost, and rationalizing tax instruments. But it is very important to realize that these policy changes cannot be realized without a significant upgrading of the tax assessment and collection efforts of the provincial governments. Figure 9.16: Revenue Targets for Provincial Tax Collection (% of 2005/06 Provincial Tax Collection) Punjab NWFP Structural Deficit (84%) Structural Deficit (157%) 1.1% of GDP (111%) 1.1% of GDP (175%) Actual (100%) Actual (100%) 153 Structural Reform 9.122. The package of comprehensive reforms can cover about 60 percent of the structural deficit in Punjab, and about 45 percent in NWFP. It comprises measures for all major provincial taxes with revenue estimates drawn from the individual analyses presented in the previous sections (Figure 9.17 and Table 9.7). This package of reforms enables Punjab to raise taxes to the international average, but in case of NWFP, the poorer province of the two, it falls short of the target by about 10 percent. What to make of these results? Even this ambitious program does not allow Punjab to cover its structural deficit. However, the administrative reforms that accompany this set of rate and base changes may cover the remaining gap ­ equivalent to about 25 percent of the present level of taxes. If not, a further increase in the effective tax rates proposed here will be necessary. In the case of NWFP, we might make the same observation, but note that even with administrative improvements, a significant gap will remain. Part of the solution for NWFP will include increased equalization transfers. 9.123. The reform would also change the relative importance of individual taxes. The UIPT reforms would generate revenue increases of about 1,100 percent in Punjab, and about 600 percent in NWFP (Figure 9.18). This would make UIPT post-reform their most important tax revenue source, compared to pre-reform land taxes in Punjab and motor vehicle tax in NWFP. An increase of this magnitude is rarely observed, which might suggest that it is nearly impossible, although even with this increase, property tax revenues in both provinces would be lower than the international average for developing countries. Another way to look at it is that the present levels of property tax in Punjab and NWFP are so low that this reform is tantamount to introducing a new tax. Even so, the full "reform" would need to be phased in, perhaps over a period of three years. 9.124. Revenue gains will be realized only if provincial governments are willing to enact and implement these reform measures. Unfortunately, there has not been much interest in revenue mobilization on the part of the provincial governments in the past. There are three ways in which this reluctance might be overcome enough to reform the tax system. The first is to convince taxpayers that a result of the reform will be improved public services. The second is to convince taxpayers that a comprehensive reform will bring about a fairer tax system. It also can be argued that such a tax system will be friendly to economic development because of its fairness and because the higher rate of revenue mobilization could lead to infrastructure improvements. The third is to offer an incentive that is lucrative enough to help provincial governments overcome their reluctance to increase taxes. Figure 9.17: Overall Indicative Revenue Impact of Comprehensive Tax Policy Reform in Punjab and NWFP Punjab NWFP Remaining Structural Gap (108%) Remaining Structural Gap (189%) Actual with Reform (115%) Actual with Reform (143%) Actua l (100%) Actual (100%) 154 Box 9.15: Setting Provincial Tax Revenue Targets There are systematic approaches to fixing on provincial revenue targets, and arguably this is where the formation of an intergovernmental fiscal reform program should begin. First, probably the best approach is to start with expenditure targets that reflect minimum acceptable service levels. After accounting for financing from intergovernmental transfers, the remainder of the cost of providing minimum service levels would be covered by own source revenues. This normative service level approach to determining the minimum needed level of tax effort is what each provincial government should do, i.e., revenue needs should be based on an expenditure plan. However, the major difficulty with this approach is that "minimum service level" is itself a subjective concept. Determining minimum service levels, and costing them out, becomes a balancing act between affordability and what a province considers to be its most pressing needs for upgrading services. Sometimes the balancing act is driven more by affordability (and political) considerations than by service upgrading. This would seem to be the case in Pakistan. Politics often has driven higher investments in the expansion of service delivery networks. In the 1990s, costly cash development loans were used to carry on infrastructure development programs despite revenue shortfalls. The provinces recognize that the services they provide are greatly inadequate, but they have not been willing to increase taxes to cover this gap. In fact, it is more likely that expenditure levels are driven primarily by the amount of transfers (and loans) received from the center. The solutions to this problem are either to leave it to local voters to push elected officials for increases in service levels that would lead to tax increases, or to put in place a system of incentives/penalties for higher/lower rates of provincial revenue mobilization. Either of these solutions is consistent with the fiscal decentralization approach to governance. However, it does not appear that any of Pakistan's provinces have established minimum spending or service levels. A second approach is to set the provincial revenue target as the amount necessary to eliminate the structural budget deficit for each province. In our analysis of Punjab, the shortfall is equivalent to 137 percent of total provincial taxes in 2006/07. The analogous number for NWFP is about 300 percent of total provincial taxes. This "fiscal discipline" approach to setting budget targets is driven more by goals of budget balance than by public service needs. Finally, the revenue target might also be set by using international comparisons. The average level of subnational government taxes in developing countries is equivalent to about 10 percent of total taxes raised by central and subnational governments (Bahl and Wallace 2005). Thus each province might take the target of raising its ratio of taxes to GDP to a level that is equivalent to 10 percent of the central government's tax to GDP ratio. Using this approach, we estimate that Punjab's target would require an increase of 111 percent in taxes and NWFP's would require a 179 percent increase in taxes. The fault with the international standards target is that it is calculated without reference to expenditure needs or to expenditure responsibilities assigned to the provincial and local governments. International subnational fiscal data is also fraught with comparability problems. Figure 9.18: Indicative Revenue Estimation for Comprehensive Tax Policy Reform in Punjab and NWFP 2005/06 Tax Collection in Punjab: Actual and Post-Reform (Rs Million) 2005/06 Tax Collection in NWFP: Actual and Post-Reform (Rs Million) 16,000 2,200 2,000 14,000 Post-Reform 1,800 12,000 Actual 1,600 10,000 1,400 1,200 8,000 1,000 6,000 Post-Reform 800 Actual 4,000 600 400 2,000 200 0 0 UIPT Motor Motor Fuel Land Taxes Agriculture Professions Sales Tax Other Taxes UIPT Motor Vehicle Motor Fuel Land Taxes Agriculture Professions Sales Tax on Other Taxes Vehicle Tax Tax Income Tax Tax on Services Tax Tax Income Tax Tax Services 155 Incentives 9.125. In 2007/08, Punjab financed some 82 percent, and NWFP 70 percent, of its expenditures from intergovernmental transfers and from loans and grants. Provincial taxes accounted only for 6 percent and 3 percent, respectively. This state of affairs raises two problems. First, it weakens the link between the expenditure benefits enjoyed by local residents and the burden associated with paying for those expenditures. This results in weak accountability of subnational officials to their voting constituency. Second, subnational governments have a comparative advantage in raising certain types of revenue. Their failure to aggressively collect these revenues imposes a "cost" in terms of an overall lower level of revenue mobilization and lower level of public services in the country than otherwise would be the case. 9.126. The standard answer as to why provincial government taxes are not higher is not entirely satisfactory. Weak taxable capacity, inadequate tax sources, deficient tax administration, and voter resistance all have some role to play, but the failure of provincial governments to mobilize more revenue relates also to the lack of incentive to do so. The central government provides significant and growing revenue through transfers and loan funds (Box 9.16), so why incur the wrath of voters and take on the local elite by raising taxes? Arguably, political leaders in provincial governments would rather be accountable for the present levels of public services than spend their political capital raising taxes. A related issue is that provincial governments are not very good at spending for public services. In fact, provincial spending patterns are characterized by unfilled positions and delays in capital project disbursements. If provinces cannot spend what they have available now, why raise more in taxes? 9.127. The federal government may want to reconsider the structure of transfers by including a reward for increased revenue mobilization in the distribution formula. At present, the distribution of the NFC award among the four provinces is based on population size and the awards are invariant with respect to tax effort. The goal of an incentive program is to change this state of affairs. For example, the NFC award could be divided into two components: an NFC incentive award and an NFC normal award. The normal award would continue to be distributed according to population shares, or whatever other criteria a future NFC might adopt. The NFC would decide on the relative sizes of the incentive pool and the normal pool. The more weight placed on the goal of increased tax effort, the greater will be the share of the incentive pool. The other dimension of the NFC award, the distribution of the incentive pool among provinces, raises more difficult questions. There are many ways to structure an incentive grant, and this will require a careful empirical analysis. Such a study would include all four provinces, and involve simulation of the possible impacts of various formulae. Box 9.16: The National Finance Commission Award To bridge the gap between large provincial expenditures and low provincial revenues, the federal government makes large fiscal transfers to the provinces based on the National Finance Commission (NFC) award. The NFC includes representations of the federal and each provincial government and takes decisions on consensus basis. While Sindh, home to the commercial hub of Karachi, favors allocation on origin of collection basis, Punjab, the most populous province, supports the population basis, while NWFP and Balochistan argue in favor of accounting for backwardness and population density (Kaiser 2006). In January 2006, the President announced an amended NFC award after failure to reach an agreement on revising the sixth award from 1997. The largest revenue source for all provinces is the divisible pool, which comprises two components. The provincial pool for the NFC award in 2008/09 is 43.75 percent of the federal divisible pool and is scheduled to increase by 1.25 percent per year up to 46.25 percent by 2010/11. The distribution of this pool among provinces is by population shares: NWFP receives 13.82 percent and Punjab receives 57.36 percent. According to the structure of the NFC grant program, the only revenue growth for a province during an award period comes from increases in the rupee amount of the vertical share. This in turn depends on the growth in federal government tax revenues. So, there is stability in the distribution system that helps long term fiscal planning. In addition, one-sixth of the GST collection is distributed to the provinces according to the octroi and zila tax collections in 1998/99, the last year when these two taxes were collected. NWFP receives 9.93 percent and Punjab 50.0 percent of the pool. This transfer is designated for pass-through to local governments. The other two types of NFC transfers are smaller. Straight transfers cover royalties and development surcharges on oil, gas and hydroelectricity and are allocated on derivation basis. They are an important revenue source for Sindh and Balochistan. The federal government also provides annual grants-in-aid to provinces on account of remoteness and backwardness. They amounted to Rs. 27.8 billion in 2006/07 and were growing in size in line with the overall divisible pool. Punjab receives a 10 percent share, and NWFP a 35 percent share. 156 Implementation and Institutional Change 9.128. The reform of provincial taxation must be comprehensive, i.e., it must address policy design, incentives to increase the rate of revenue mobilization, and tax administration. The focus of this analysis is tax policy but the success of a structural reform will depend on whether an improved tax administration can be achieved. Without significant improvements in tax administration, especially updated surveys of taxable subjects, more accurate valuation, and the automation of recordkeeping, better tax structures will come to naught. An important first step toward improving the system of tax administration is to reconsider the fragmented responsibility for tax assessment and collections. With a unified tax administration in place, it will be easier to measure and assess tax bases. 9.129. In Punjab and NWFP, the principal tax collecting agencies are the Board of Revenue (BOR) and the Excise and Taxation Department (ETD). The BOR is mainly concerned with taxes on rural lands. Specifically, it assesses and collects stamp duty, mutation fee, registration fee, land revenue and agriculture income tax. In the case of stamp duty and registration fee, BOR is also responsible for collections in urban areas. The ETD collects UIPT, motor vehicle taxes, professions tax, provincial excise duty and some smaller provincial taxes. In the case of the sales tax on services and electricity duty, the finance departments of the federal and provincial governments, respectively, play a primary role. In all of this work, there appears to be little collaboration between ETD and BOR, or between the federal and provincial government. 9.130. This state of affairs is counterproductive to effective policy and collection. The kind of problems that arise might be illustrated with three examples. First, the valuation tables for the property transfer taxes and the UIPT are separately done, even though both taxes rely on much the same information to establish a base of taxation. Increasingly, the two agencies are said to be coordinating at the field level to ensure reliability of information, but still their records are not aligned with each other. Second, the record keeping function can fall between the two agencies. In areas which are newly urbanized, the patwari does not update his record to reflect the development of housing units on erstwhile agriculture land. If the area is not notified as urban, the ETD does not cover it in their surveys nor do they pick it up on their valuation registers. The fact that BOR is primarily concerned with rural lands means that its record keeping does not extend into urban areas. Nearly all urban areas lack a systematic record of property titles. The systems and expertise of the BOR is not used for urban areas. The result is that nearly all urban areas lack a systematic record of property titles. Third, there is no way to cross check for compliance problems, either within or between agencies. For example, motor vehicle registrations cannot be cross checked against land ownerships records (which could aid compliance and would be a desirable feature of a motor vehicle registration system that had a residence requirement). 9.131. Both the governments of Punjab and Sindh have in the past discussed creating a unified tax authority. A Provincial Board of Revenue might be set up with all provincial tax assessment and collection functions assigned to it. A unified provincial Board of Revenue, organized according to tax functions would give the provinces a better functional capacity to raise the revenues assigned to them. It should be equipped with computerized systems of tax base documentation, qualified tax administrators and adequate resources to manage collection. There are numerous advantages to be had from a unified tax administration at the provincial level. The possible gains from a unified authority are great enough to merit a serious review. · All tax bases and tax payment histories could be cross­referenced. For example, payment of motor vehicle annual tax, UIPT and professional tax could be linked by a unique taxpayer number. · There could be a complete enumeration of all land (rural and urban) in the province, and this enumeration could include physical dimensions, value information, use, ownership and tax payment history. · A unified tax administration would allow the government to capture economies of scale from staff training to collection and enforcement. 9.132. Another dimension of implementation is a timetable for reform. Some of the reforms with significant revenue impact suggested in this review can be adopted quickly, if government chooses to do so, while others would require more time to prepare for implementation (Table 9.8). 157 What Next? 9.133. Each province could appoint a blue ribbon commission to develop a comprehensive reform program, based on the provincial objectives for its tax system and informed by local and international experience in subnational taxation. The commissions would operate in parallel with a federal commission on tax reform, as discussed in Chapter 2. FBR would join these commissions in an advisory capacity. Since the issues of tax policy, tax administration and intergovernmental transfers cannot be separated for a viable reform, they should all be part of the terms of reference for the commissions. Three areas for study and recommendations should be covered. 9.134. First, the structure of taxation should be simplified. The commission could recommend improvements in the tax structure as well as increases in the level of taxes that would be consistent with provincial revenue targets. Of course, the discussion of the tax assignment structure cannot be looked at independently from decisions on the appropriate allocation of tasks among levels of government. Only when the allocation of expenditure responsibilities has been decided, then the right allocation of tax assignment can be determined. Taking the division of tasks across government levels as given, much of the work of the commission would involve debating the merits of alternative tax packages. For example, a radical simplification of the provincial tax structure could be as follows: · An urban property tax and an annual tax on agricultural and other rural land. The former would be devolved to local governments. · A capital gains tax on land. · A federally administered surcharge on the individual income tax and a federally administered and expanded sales tax on services. Both could feature provincial level rate setting. The provincial share of revenues would flow directly to the provincial governments. · An annual tax on motor vehicle use, and a tax on motor fuels. · All other taxes in the system would be abolished or folded into those above. 9.135. The second part of the work would be to identify the kind of tax administration that is desired. In the long run, a unified provincial tax administration, and shared federal and provincial tax administration responsibilities should be in the mix. A stronger, automated recordkeeping system for taxation is essential. Training of tax officials is a high priority in order to establish state of the art methods of assessment and collection. 9.136. The third component of the commission's agenda would be intergovernmental fiscal relations within the province. This would involve a reassessment of the relationship between the province and its junior bodies, i.e., districts, TMAs and union councils. The three questions to be addressed in this analysis are expenditure assignment, revenue assignment and the system of intergovernmental transfers. Questions of the goals for revenue mobilization by local governments, and equalization, will be central to this work. 9.137. The results of the four provincial commissions should be brought together in a consensus report. Such a white paper on provincial government finance could serve as a blueprint for provincial fiscal policy in the next decade. 158 Table 9.7: Piecemeal and Comprehensive Policy Reform Matrix Reform Type Proposal Urban Immovable Property Tax Piecemeal and Comprehensive · Eliminate 5 marla exemption · Eliminate preferential treatment of owner occupiers · Bring in new valuation roll (Punjab) · Upgrade valuation table (NWFP) · Bring land and structures to same basis · Adopt a single rate (Punjab) · Tie TMA rate setting to provincial transfers Motor Vehicles Tax Piecemeal · Abolish registration tax or shift to the federal level · Increase the token tax rate Comprehensive · Introduce unified annual license tax on motor vehicles · Adopt a motor fuels tax Professions Tax Piecemeal and Comprehensive · Introduce a 3 percent piggyback on federal individual income taxes Property Transfer Taxes Piecemeal · Improve valuation methods · Update property records Comprehensive · Introduce unified annual tax on rural land Agricultural Income Tax Piecemeal and Comprehensive · Introduce progressive rate structure by farm size · Reduce exemption from 12.5 to 7.5 acres (Punjab); increase exemption to 7.5 acres (NWFP) Sales Tax on Services Piecemeal and Comprehensive · Introduce a shared federal-provincial tax with federal administration Table 9.8: Sequencing of Tax Policy Reform Short-Term · Prepare a new valuation table in Punjab · Eliminate preferential treatments under UIPT · Adopt a new property tax base in NWFP · Unify motor vehicle taxes into an annual tax; introduce a motor fuel tax · Convert the sales tax on services to a shared federal-provincial tax · Bring the hotel tax and the electricity duty into the sales tax on services · Allocate the entertainment tax to local governments Medium-Term · Unify land taxes as an annual tax on land value, based on a new survey of all properties; eliminate the property transfer tax · Introduce a new valuation roll in NWFP · Expand the base of the Professions Tax · Convert AIT into a presumptive, income-based tax on land area and crop type; bring agricultural cesses into this tax base · Fold the sales tax on services with a federally administered comprehensive GST on goods and services shared fairly with the provinces · Introduce an incentive element into the NFC formula · Adopt a capital gains tax on land · Convert the professions tax to an income tax surcharge ************** 159 REFERENCES · Ahmad, Ehtisham and Nicolas Stern (1991). "The theory of taxation for developing Countries". Cambridge University Press, Cambridge · Alesina, Alberto, Filipe Campante and Guido Tabellini (2007). "Why is fiscal policy often procyclical?" Working Paper ­ 2007. Harvard University and IGIER Bocconi · Alm, James and Janet Rogers (2005). "Do State Fiscal Policies Affect Economic Growth?" Mimeo, Department of Economics, Andrew Young School of Policy Studies, Georgia State University. · Atkinson, A.B. and Stiglitz, J.E. (1976). "The design of tax structure: direct versus indirect taxation". Journal of Public Economics 6, 55-75. · Auerbach, Alan J. (2006). "The Choice Between Income and Consumption Taxes: A Primer". NBER Working Papers 12307. National Bureau of Economic Research Inc. · Bahl, Roy W. and Sally Wallace (2005). "Public Financing in Developing and Transition Countries". Public Budgeting & Finance, Vol. 25, No. 4S, December 2005. Georgia State University · Baunsgaard, Thomas and Michael Keen (2005). "Tax Revenue and (or?) Trade Liberalization". International Monetary Fund Working Papers No. 05/112, June 2005 · Bird, Richard (2008). "Tax Challenges Facing Developing Countries". Working Papers Series 12, revised Mar 2008. Institute for International Business, Joseph L. Rotman School of Management, University of Toronto. · Chaudhry, M. Ghaffar. (1999). "The Theory and Practice of Agricultural Income Tax in Pakistan and a Viable Solution." Pakistan Development Review 38, no. 4: 755-765. Pakistan Institute of Development Economics · Dubner, Stephen J. and Steven D. Levitt (2006). "Filling in the Tax Gap". Mr. Szilagyi's Billion- Dollar Idea, April 2, 2006. · Easterly, William, Timothy Irwin, and Luis Servén (2008). "Walking up the Down Escalator: Public Investment and Fiscal Stability". World Bank Research Observer 2008, Volume 23, Number 1. · Edgeworth, F. (1897), "La teoria pura del monopolio, Giornale Degli Economisti 40"; pp. 13-31. · Ghaus-Pasha, Aisha, et. al. (1998). "Devolution and Fiscal Decentralization". Research paper # 40, Social Policy and Development Center. · Gorodnichenko, Yuriy, Jorge Martinez-Vazquez, and Klara Sabirianova Peter (2008). "Myth and Reality of Flat Tax Reform: Micro Estimates of Tax Evasion Response and Welfare Effects in Russia". NBER Working Papers 13719, National Bureau of Economic Research, Inc. · Government of Pakistan (1991). "Report of Committee on Tax Reform". Ministry of Finance, Government of Pakistan. 160 · Government of Pakistan (2000). "Agriculture Census 2000". Pakistan Agricultural Census Organization · Government of Pakistan (2001). "Reform of Tax Administration in Pakistan"; Report of the Task force on Reform of Tax Administration, April 14, 2001; Central Board of Revenue, Government of Pakistan · Government of Pakistan (2007). "Pakistan Economic Survey-2007". Annual report of Ministry of Finance, Government of Pakistan. · Government of Pakistan (2008). "An Update on Tax Administration Reforms". FBR (Federal Board of Revenue) Reform Program, Quarterly Review, Vol. 7, No. 4, April ­ June 2008 · Keen, Michael (2007). "International Tax and Public Finance; VAT attacks". Volume 14, Number 4, August, 2007 · Khan, Muhammad Aslam (1999). "The Theory and Practice of Agricultural Income Tax in Pakistan and a Viable Solution: Comments". Pakistan Development Review, v. 38, issue. 4, pp. 766-68. Pakistan Institute of Development Economics · Mirrlees, James A (1971). "An Exploration in the Theory of Optimum Income Taxation". Review of Economic Studies, vol. 38(114), April. 1971. Blackwell Publishing. · Norregaard, John and Tehmina S. Khan (2007). "Tax Policy: Recent Trends and Coming Challenges". IMF Working Paper No. 07/274, December 2007 · Pasha, Hafiz.A. (1995). "Political Economy of Tax Reforms: The Pakistan Experience". Pakistan Journal of Applied Economics, 11(1&2), pp. 129-154 · Pasha, Hafiz A., and Aisha Ghaus-Pasha, "Pakistan Development Review". Pakistan Institute of Development Economics, 39,(4) Part-II, pp.981-1011 · Pasha, Hafiz A., and M. Asif Iqbal (1994). "Taxation Reforms in Pakistan". Pakistan Journal of Applied Economics, v. 10, iss. 1-2, pp. 47-75 · Perotti, Roberto (2007). "In Search of the Transmission Mechanism of Fiscal Policy". NBER Working Papers 13143. 2007. National Bureau of Economic Research. · Petit, Patrick, and Emil M. Sunley (2008). "Reform of the Federal Excise Tax on Tobacco in Pakistan". unpublished working paper- 2008 · Ramsey, F.P. (1927). "Facts and Propositions". Aristotelian Society Supplementary, Volume 7, 153­170.1927 · Refaqat, Saadia (2005). "Redistributive Impact of GST Tax Reform: Pakistan, 1990-2001", 44 (4), 2005. Pakistan Development Review, Pakistan Institute of Development Economics 161 · Rubin, Robert E., Peter R. Orszag and Sinai Allen (2004). "Sustained Budget Deficits: Longer-Run U.S. Economic Performance and the Risk of Financial and Fiscal Disarray". Budget Deficit, U.S. Economy, Federal Budget · Silvani, Carlos, Wyatt Grant, and Orlando Reos, and Paulo Dos Santos (2008). "FBR Tax Administration Review". The World Bank · Sinai, Allen, Peter R. Orszag, and Robert E. Rubin. "National Economic and Financial Policies for Growth and Stability". AEA-NAEFA Joint Session, Allied Social Science Associations Annual Meetings, The Andrew Brimmer Policy Forum · Smith, Stephen and Michael Keen (2007). "VAT Fraud and Evasion: What Do We Know, and What Can Be Done?" IMF Working Paper No. 07/31. February 2007. · Talvi, Ernesto and Carlos A. Vegh (2005). "Tax base variability and procyclical fiscal policy in developing countries". Journal of Development Economics, Elsevier, vol. 78(1), October. 2005 · World Bank (2006). "World Development Report". Annual development report on equity and development- 2006 · World Bank (2007). "World Development Report". Annual development report on agriculture- 2007. · Zee, Howell H. (2005). "Personal Income Tax Reform: Concepts, Issues, and Comparative Country Developments". IMF Working Papers 05/87, 2005 · Zee, Howell H., Janet G. Stotsky, and Eduardo Ley (2002). "Tax Incentives for Business Investment: A Primer for Policy Makers in Developing Countries". World Development, Elsevier, vol. 30(9), September. 2002 ********************* 162