Report No: ACS9351 . Republic of Kosovo Kosovo Public Finance Review Fiscal Policies for a Young Nation June 2014 Poverty Reduction and Economic Management Unit Europe and Central Asia Region . . . Standard Disclaimer: . This volume is a product of the staff of the International Bank for Reconstruction and Development/ The World Bank. The findings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. . Copyright Statement: . The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development/ The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, http://www.copyright.com/. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org. CURRENCY AND EQUIVALENT UNITS Currency Unit= Euro € 1.00 = US$ 1.37 (As of March 3, 2014) FISCAL YEAR January 1 – December 31 WEIGHTS AND MEASURES Metric System ACRONYMS AND ABBREVIATIONS BoP Balance of Payments MEST Ministry of Education Science and Technology CBK Central Bank of the Republic of Kosovo MIHIL Model of the Impact of Health Insurance Law CCS Carbon Capture and Storage MoF Ministry of Finance CIS Commonwealth of Independent States MoH Ministry of Health CIT Corporate Income tax MTEF Medium Term Expenditure Framework CPI Consumer Price Index MW Mega Watt DH District Heating NEEAP National Energy Efficiency Action Plan ECA Europe an Central Asia NGO Non-Government Organizations ECE Early Childhood Education NPISH Non-Profit institutions Serving Households EE Energy Efficiency NREAP National Renewable Energy Action Plan ELF Emergency Liquidity Fund OECD Organization for Economic Cooperation and Development EMIS Education Management Information OOP Out Of Pocket System ENS Energy Not Supplied OSR Own Source Revenues ERO Energy Regulatory Office PCF Per Capita Funding ESG Education Specific Grants PER Public Expenditure Review EU European Union PFR Public Finance Review FDI Foreign Direct Investments PHC Primary Health Care FYR Former Yugoslav Republic PISA Program for International Student Assessment GDP Gross Domestic Product PIT Personal Income Tax GER Gross Enrollment Rates PoTI Perceptions of Tax Law Implementation GoK Government of Kosovo PPP Purchasing Power Parity GPS Global Positioning System PTK Post and Telecommunications Company of Kosovo GWh Giga Watt hour PTR Pupil Teacher Ratio HBS Household Budget Survey RE Renewable Energy HIF Health Insurance Fund RfP Request for Proposals HIL Health Insurance Law SAA Stabilization Association Agreement HIV Human Immunodeficiency Virus SBA Stand-By Arrangement HPS Higher Pedagogical School SDR Standard Death Rate IBRD International Bank for Reconstruction SEE South East Europe and Development IED Integrated Education Database SFRJ Socialist Federal Republic of Yugoslavia IMF International Monetary Fund SME Small and Medium Enterprises IT Information Technology SMP Staff Monitored Program KAS Kosovo Agency of Statistics SOE Socially Owned Enterprise KEDS Kosovo Electricity Distribution and TAK Tax Administration of Kosovo Supply KEK Kosovo Electricity Corporation TB Tuberculosis KESIT Kosovo Energy Sector Impact Tool TPE Total Public Expenditure KFOS Kosovo Foundation for Open Society UCCK University Clinical Center of Kosovo KfW Kreditanstalt für Wiederaufbau UK United Kingdom KOSTT Kosovo Transmission, System and UN United Nations Market Operator KPC Kosovo Protection Corps UNICEF The United Nations Children’s Fund KPST Kosovo Pensions Savings Trust UNMIK United Nations Interim Administration Mission in Kosovo KRPP Kosova e Re Power Plant USAID United States Agency for International Development KWh Kilo Watt hour VAT Value Added Tax LFS Labor Force Survey WBI World Bank Institute LPFMA Law on Public Finance Management and WEO World Economic Outlook Accountability LTU Large Taxpayer Unit WHO World Health Organization Vice President: Laura Tuck, ECAVP Country Director: Ellen Goldstein, ECCU4 Acting Sector Director: Roumeen Islam, ECSPE Sector Manager: Satu Kahkonen, ECSPE Task Team Leaders: Agim Demukaj, ECSPE Simon Davies, ECSPE AKNOWLEDGMENTS This report was prepared by a World Bank team working in Kosovo during 2013. The principal authors were Agim Demukaj and Simon Davies, supported by a thematic team consisting of Munawer Khwaja and Sandeep Bhattacharya (revenue); Rhedon Begolli and Karuna Phillips (energy); Baher El-Hifnawi and Robin Carruthers (transport); Lucas Gortazar and Flora Kelmendi (education); Karen Stephanie Coulibaly, Aneesa Arur and Marvin Ploetz (health); Anita Schwarz and Aylin Isik-Dikmelik (social protection); and Agim Demukaj (Introduction, and Economic Context and the Fiscal Challenge). Kim Murrell’s editing helped greatly to produce a better report. The World Bank team is sincerely grateful to the Kosovo Government for its close collaboration during preparation of the report. In particular, thanks are due to senior officials in the Ministries of Finance; Ministry of Health, Education; Labor and Social Welfare, Infrastructure; and Energy Regulatory Office. Special thanks also go to Satu Kahkonen, Yvonne Tsikata, Abebe Adugna and Roumeen Islam of the World Bank, who provided continuous support, guidance, and technical advice. Useful comments were received in September 2013 at a day long workshop, jointly organized by the Ministry of Finance, Heath, Education, Labor and Social Welfare and the World Bank. The team would also like to thank the following reviewers for helpful comments and advice: Raymond Bourdeaux (Sector Leader, World Bank); William Dillinger (World Bank Consultant); Borko Handjinksi (Senior Economist, World Bank); Ranjit Lamech (Sector Manager, World Bank); Dino Merotto (Lead Economist, World Bank); Jacques Miniane (Kosovo Mission Chief, IMF); Valmira Rexhebeqaj (Director, Macroeconomic Policy and Management Unit of the Kosovo Ministry of Finance); Heather Worley (Senior Communications Officer, World Bank); Katelijn Van den Berg (Senior Environmental Specialist, World Bank); and UNICEF Kosovo. Finally, the team wishes to thank Ellen Goldstein, country director for Southeast Europe, and Jan-Peter Olters, country manager for Kosovo, for their valuable advice and guidance, and Mismake Galatis for their support in preparation of this report. CONTENTS EXECUTIVE SUMMARY ......................................................................................................................................... I A. CONTEXT ........................................................................................................................................................... I B. CHALLENGES .................................................................................................................................................... II Unemployment and Poverty................................................................................................................ ii A Young Demographic ......................................................................................................................iii Revenue and Energy ..........................................................................................................................iii C. INCREASING THE EFFICIENCY OF PUBLIC SPENDING.......................................................................................... V D. IMPROVING PUBLIC EXPENDITURE ALLOCATIONS ........................................................................................... VI E. CONCLUSION ................................................................................................................................................... IX F. TABLE OF TOP THREE RECOMMENDATIONS BY SECTOR .................................................................................... XI CHAPTER 1. INTRODUCTION ......................................................................................................................... 1 A. CONTEXT ...........................................................................................................................................................1 B. KOSOVO’S CHALLENGES AND OBSTACLES TO GROWTH......................................................................................1 Unemployment, Poverty and Shared Prosperity ................................................................................ 2 A Young Demographic ....................................................................................................................... 3 A Looming Energy Crunch ................................................................................................................. 4 Declining “Border Revenues”............................................................................................................ 4 C. SPENDING PRESSURES DO NOT MATCH THE COUNTRY’S NEEDS ..........................................................................5 D. OBJECTIVES AND SCOPE OF THE PUBLIC FINANCE REVIEW ................................................................................6 CHAPTER 2. ECONOMIC CONTEXT AND THE FISCAL CHALLENGE ................................................ 9 A. KEY MESSAGES ..................................................................................................................................................9 B. ECONOMIC CONTEXT .........................................................................................................................................9 The Macroeconomic Basics are Right ................................................................................................ 9 C. FISCAL PERFORMANCE 2008 – 2013.................................................................................................................12 The Headline Fiscal Numbers are Sound ......................................................................................... 12 Revenue Growth was Strong, but Has Slowed ................................................................................. 14 Expenditure Efficiency Could be improved ...................................................................................... 16 CHAPTER 3. ENERGY ..................................................................................................................................... 21 A. KEY MESSAGES ................................................................................................................................................21 B. OVERVIEW OF KOSOVO’S POWER SECTOR: GENERATION, DISTRIBUTION AND LOSSES .....................................22 C. PAYING THE PRICE FOR AN IMPERFECT POWER SECTOR ....................................................................................23 D. MEETING THE GROWING DEMAND: BUILDING NEW GENERATION CAPACITY ....................................................27 E. IMPACT OF POWER GENERATION OPTIONS: THREE SCENARIOS ........................................................................29 F. PROTECTING THE POOR FROM RISING ENERGY COSTS.....................................................................................35 G. RECOMMENDATIONS ........................................................................................................................................37 CHAPTER 4. TRANSPORT .............................................................................................................................. 41 A. KEY MESSAGES ...............................................................................................................................................41 B. TRANSPORT SECTOR OVERVIEW ......................................................................................................................41 C. TRANSPORT SECTOR INVESTMENT ...................................................................................................................44 Investment in Transport Infrastructure has Grown Fast but Not Been Aligned with Needs ............ 44 The Cost Efficiency of Transport Investments Has Been Low .......................................................... 47 D. TRANSPORT INVESTMENT NEEDS .....................................................................................................................48 Roads ................................................................................................................................................ 48 Railways ........................................................................................................................................... 50 E. SOURCES OF ROAD MAINTENANCE FUNDING ....................................................................................................52 F. RECOMMENDATIONS ........................................................................................................................................55 CHAPTER 5. EDUCATION .............................................................................................................................. 57 A. KEY MESSAGES ................................................................................................................................................57 Education Matters ............................................................................................................................ 58 B. ENROLLMENT TRENDS ARE MOVING IN THE RIGHT DIRECTION .........................................................................61 C. EDUCATION EXPENDITURES ARE NOT STUDENT-FOCUSED................................................................................64 D. TEACHER SALARIES HAVE DRIVEN SPENDING INCREASES IN EDUCATION .........................................................66 E. CAPITATION FINANCING AND MUNICIPAL COMPARISONS .................................................................................69 F. EDUCATION OUTCOMES ...................................................................................................................................72 G. DEMOGRAPHIC CHANGES COULD BRING COST SAVINGS ...................................................................................74 H. RECOMMENDATIONS ........................................................................................................................................74 CHAPTER 6. HEALTH FINANCING, SECTOR REFORMS, EFFICIENCY AND EQUITY ................. 79 A. KEY MESSAGES ................................................................................................................................................79 B. OVERVIEW OF HEALTH SECTOR........................................................................................................................80 C. HEALTH OUTCOMES .........................................................................................................................................81 D. OVERVIEW OF HEALTH SPENDING IN KOSOVO .................................................................................................82 E. EFFICIENCY OF HEALTH SPENDING ...................................................................................................................85 F. HEALTH SPENDING AND FINANCIAL PROTECTION.............................................................................................87 G. FISCAL IMPACT OF HEALTH REFORM: A FOCUS ON HEALTH INSURANCE ..........................................................90 H. RECOMMENDATIONS ........................................................................................................................................95 CHAPTER 7. SOCIAL PROTECTION ........................................................................................................... 97 A. KEY MESSAGES ................................................................................................................................................97 B. OVERVIEW OF SOCIAL PROTECTION MECHANISMS ..........................................................................................97 C. FINANCING SOCIAL PROTECTION ......................................................................................................................99 D. MAIN SOCIAL PROTECTION TRANSFERS ..........................................................................................................100 Pensions.......................................................................................................................................... 100 War Related Benefits ...................................................................................................................... 101 Social Assistance Scheme ............................................................................................................... 102 E. FUTURE SOCIAL PROTECTION NEEDS ..............................................................................................................105 Social Protection and Energy and Health Modernization ............................................................. 105 War-Related Benefits ...................................................................................................................... 106 Pensions.......................................................................................................................................... 107 F. RECOMMENDATIONS ......................................................................................................................................108 CHAPTER 8. TAX AND REVENUE POLICY.............................................................................................. 109 A. KEY MESSAGES .............................................................................................................................................109 B. TAX POLICY IS JUDICIOUS BUT THERE ARE RISKS AHEAD ...............................................................................110 C. STRUCTURE OF PUBLIC REVENUES IN KOSOVO – A HIGH DEPENDENCY ON BORDER TAXES ............................111 D. METHODOLOGIES APPLIED TO MEASURE INFORMAL SECTOR AND REVENUE POTENTIAL .............................113 Size of the Informal Economy ......................................................................................................... 113 Tax Productivity ............................................................................................................................. 113 Tax Gap Analysis............................................................................................................................ 114 Tax Evasion .................................................................................................................................... 115 Revenue Potential ........................................................................................................................... 115 E. A TAX POLICY REFORM FRAMEWORK FOR KOSOVO .....................................................................................116 Augmenting Revenues from Corporate and Personal Income Taxes ............................................. 116 Enhancing Collection of Property Tax ........................................................................................... 118 Strengthening Domestic VAT and Excise Collection...................................................................... 120 F. RECOMMENDATIONS ......................................................................................................................................121 CHAPTER 9. TOWARD A BUDGET FOCUSED ON NEEDS ................................................................... 123 A. WHERE IS THE BUDGET HEADING? .................................................................................................................124 Baseline scenario – the Status Quo ................................................................................................ 124 Adaptive scenario ........................................................................................................................... 126 Energy Crisis Scenario ................................................................................................................... 129 B. CONCLUSIONS ................................................................................................................................................132 ANNEX 1: DATA ON MUNICIPALITIES .......................................................................................................... 133 ANNEX 2: TRANSPORT ....................................................................................................................................... 135 ANNEX 3: EDUCATION ....................................................................................................................................... 143 ANNEX 4: REVENUE ............................................................................................................................................ 155 ANNEX 5: MACROECONOMIC INDICATORS ............................................................................................... 156 REFERENCES……………………………………………………………………………………………………. 157 TABLES Table 2.1. Key Macroeconomic Indicators 2008-2015................................................................................................ 12 Table 2.2. Revenue Performance 2008 – 2013 ............................................................................................................ 16 Table 2.3. Expenditure by Economic Category 2008-2013, % GDP ........................................................................... 16 Table 2.4. Expenditure by Functional Category 2008-2013, % GDP .......................................................................... 20 Table 3.1. Government and Donor Transfers to KEK 2008 – 2012 in ‘000 of Euro .................................................... 26 Table 3.2. Comparison of Unmetered .......................................................................................................................... 36 Table 4.1. Transport Infrastructure Road Density Comparison ................................................................................... 42 Table 4.2. Development of accidents 2002-2012 ........................................................................................................ 43 Table 4.3. Kosovo Investments in Transport Infrastructure ........................................................................................ 45 Table 4.4. Benchmarks for Transport Infrastructure Investment in Kosovo ............................................................... 46 Table 4.5. Comparison of Actual and Desirable Transport Investment, 2007 to 2012 ................................................ 46 Table 4.6. Comparison of R6 and R7 Construction Costs with Other Highways ........................................................ 47 Table 4.7. Potential Reduced Cost of Prioritized Local Roads Activities ................................................................... 49 Table 4.8. Sensitivity Tests for Main Project in Revised Economic Evaluation of R6 ............................................... 50 Table 4.9. Kosovo Average Fuel Price Structure, 2012............................................................................................... 52 Table 4.10. Gasoline and Diesel Pump Prices, US$ Cents Per Liter, As of November 2012 ...................................... 53 Table 6.1. Progressivity of OOP Payments ................................................................................................................. 88 Table 6.2. OOP Payments as a Share of Total Consumption ...................................................................................... 88 Table 6.3. Incidence and Intensity of Catastrophic OOP Health Payments, 2009 And 2011 ...................................... 89 Table 6.4. Incidence of Catastrophic Payment Using 40 Percent Nonfood Consumption Threshold ......................... 89 Table 6.5. Poverty Impact of Out-Of-Pocket Health Expenditures, 2009 And 2011................................................... 90 Table 7.1. War-Related Benefits ............................................................................................................................... 101 Table 8.1. Regional Comparison of Income Taxes Performance (Excluding Presumptive Taxes) 2011 .................. 117 Table 8.2. Property Tax Rates in SEE (% Property Value) ....................................................................................... 119 Table 9.1. Key Assumptions for each Scenario ......................................................................................................... 131 FIGURES Figure 1.1:Poverty Rates in Kosovo, 2009-2011* .........................................................................................................3 Figure 1.2. Extreme Poverty Rates in Kosovo, 2009-2011* .........................................................................................3 Figure 1.3. Population Pyramid, 2011 ...........................................................................................................................4 Figure 2.1. Contribution to Growth 2008-2012 ........................................................................................................... 10 Figure 2.2. Growth Incidence Curve, 2006-2011 ........................................................................................................ 10 Figure 2.3. Loans and Deposits, 2008–12 ................................................................................................................... 11 Figure 2.4. Indicators of Financial Health ................................................................................................................... 11 Figure 2.5. Sources of Deficit Financing ..................................................................................................................... 12 Figure 2.6. Revenue Performance 2004-2013 ............................................................................................................. 14 Figure 2.7. Total Tax Revenue in SEE, % GDP .......................................................................................................... 15 Figure 2.8. Collection of Taxes By Place, % Total Tax Revenue ............................................................................... 15 Figure 2.9. Structure Expenditures in Southeast Europe ............................................................................................. 18 Figure 2.10. Selected Expenditures by ........................................................................................................................ 18 Figure 3.1. Power Generation, 2012 ............................................................................................................................ 22 Figure 3.2. Power Generation and Consumption, GWh, 2008-12 ............................................................................... 22 Figure 3.3. Power Flows, 2012 .................................................................................................................................... 23 Figure 3.4. Distribution Losses as % ........................................................................................................................... 23 Figure 3.5. System Interruption Frequency and Duration, 2012.................................................................................. 24 Figure 3.6. Average Hourly Supply and Demand of Power, MWh 2012 .................................................................... 24 Figure 3.7. Estimated Transmission and ...................................................................................................................... 25 Figure 3.8. Cost and Quantity of Imports, ................................................................................................................... 25 Figure 3.9. Supply and Demand of Power, GWh ........................................................................................................ 31 Figure 3.10. Estimated Imports and Exports, Gwh ...................................................................................................... 32 Figure 3.11. Estimated Costs, € Cents Per Kwh .......................................................................................................... 32 Figure 3.12. Supply and Demand of Power, Gwh ....................................................................................................... 33 Figure 3.13. Estimated Average Cost, euro cent Per Kwh, 2012 Euro ........................................................................ 33 Figure 3.14. Pre-tax residential costs, 2012, euro cent per KWh................................................................................. 33 Figure 3.15. Annual Budget Costs for Construction, € Mil. ........................................................................................ 34 Figure 3.16. Additional Public Debt Stock, ................................................................................................................. 35 Figure 3.17. Additional Annual Public Interest Payments, € Mil. ............................................................................... 35 Figure 3.18. Distribution of the Implicit ...................................................................................................................... 36 Figure 3.19. Shares of Energy Expenditure in ............................................................................................................. 37 Figure 3.20. Energy Poverty and Energy ..................................................................................................................... 37 Figure 4.1. Rail Traffic Intensity (ton km plus passenger km per km of Line, 2009 (EU =100) ................................. 44 Figure 4.2. Expenditure on National Road Maintenance, €/km ................................................................................... 46 Figure 4.3. SEETO Comprehensive South East Europe Rail Network ....................................................................... 51 Figure 4.4. Access Charges (€ Per Train Km) for a Typical 960 Gross Ton Freight Train (2008) ............................. 54 Figure 5.1. Unemployment by Level of Education 2012 ............................................................................................. 59 Figure 5.2. Education Levels in Excess Supply: Surplus of Workers as % of Unemployed ....................................... 60 Figure 5.3. Occupations in Excess Supply, Surplus of Workers as % of Unemployed ............................................... 60 Figure 5.4. Non-Enrollment Rates by Household Poverty Status: Ages 10-15 ........................................................... 63 Figure 5.5. Non-Enrollment Rates by Household Poverty Status: Ages 16-20 ........................................................... 63 Figure 5.6. Expenditures on Education and Young Population ................................................................................... 64 Figure 5.7. Total Public Expenditure and Public Expenditure on Education, %GDP ................................................. 64 Figure 5.8. Public Expenditures in Education as a Percent of GDP for Kosovo, European, and Other Comparator Countries (2011 or Latest) ........................................................................................................................................... 65 Figure 5.9. Public Expenditures Per Pupil as a % of GDP Per Capita, Pre-University (2011 or Latest) ..................... 65 Figure 5.10. Expenditures by Level of Education, % GDP ......................................................................................... 66 Figure 5.11. Education Expenditures by economic Category, % GDP ....................................................................... 66 Figure 5.12. Distribution of Teachers by Pre-Service Qualifications, 2008 and 2013 ................................................ 68 Figure 5.13. School funding ........................................................................................................................................ 70 Figure 5.14. Pupil to Teacher Ratio and Population Density, by Municipality (2012)................................................ 70 Figure 5.15. Fifth Grade Reading Scores by Municipality .......................................................................................... 73 Figure 6.1. Evolution of Government of Kosovo Public Health Spending, 2007-2012 .............................................. 82 Figure 6.2. Public Health Spending and Per Capita Income, 2011 .............................................................................. 82 Figure 6.3. Regional Variation In Per Capita (Domestically-Financed) Public Health Spending, 2012 ..................... 83 Figure 6.4. Evolution of Public Sector Spending by Functional Category .................................................................. 83 Figure 6.5. Primary Care Public Spending by Economic Category ............................................................................. 84 Figure 6.6. Public Hospital Care Spending by Economic Category ............................................................................ 84 Figure 6.7. Number of Hospital Beds Per 100,000 in Catchment Area ....................................................................... 85 Figure 6.8. Physicians Per 1,000 Population, 2011 ..................................................................................................... 86 Figure 6.9. Nurses Per 1,000 Population, 2011 ........................................................................................................... 86 Figure 6.10. Primary Health Care Visits Per Capita, 2011 .......................................................................................... 86 Figure 6.11. Primary Health Care Visits/GP/ Day and Population Profile .................................................................. 86 Figure 6.12. Number of Discharged Patients Per Doctor in Regional Hospitals and UCCK, 2012 ............................ 87 Figure 6.13. Average Annual Per Capita OOP Health Spending by Consumption Quintile, 2009 and 2011 .............. 88 Figure 6.14. Pen’s Parade of Per Capita Expenditures Gross and Net Of OOP Payments, 2009 and 2011 ................ 90 Figure 6.15. First Year Revenue Scenarios, € ............................................................................................................. 92 Figure 6.16. First Year Spending ................................................................................................................................. 92 Figure 6.17. Health Insurance Net Revenue ................................................................................................................ 93 Figure 6.18. Health Insurance Net Spending, .............................................................................................................. 93 Figure 6.19. Health Insurance Net Revenue Sources, € Mil. Low Scenario ................................................................ 95 Figure 6.20. Health Insurance Net Spending, € Mil. Low Scenario ........................................................................... 95 Figure 7.1. Transfer Expenditure in € and % GDP ...................................................................................................... 99 Figure 7.2 Transfers as Share of Total Public Expenditure ......................................................................................... 99 Figure 7.3 Composition of Social Security Expenditures, EUR, 2008-12 ................................................................. 100 Figure 7.4. Contributions to Increase in Social Protection Budget 2008-12.............................................................. 100 Figure 7.5. Number of Basic Pension Recipients (Excluding Contributors to Ex-Yugoslav System) ....................... 100 Figure 7.6. Distribution of Spending on War-Related Benefits in Bosnia and Herzegovina ..................................... 101 Figure 7.7. Number of Households Receiving Social Assistance Benefits ............................................................... 102 Figure 7.8. Income from Social Assistance as % Household Consumption for Beneficiary Households in Poorest Quintile ...................................................................................................................................................................... 102 Figure 7.9. Coverage: % Benefitting Households in Top and Bottom Quintiles ....................................................... 103 Figure 7.10. Targeting: % Benefits Going to Households in Top and Bottom Quintiles .......................................... 103 Figure 7.11. Pensioner Population and Aged Dependency Ratio, 2010-2050 ........................................................... 107 Figure 7.12. Annual Costs Pensions , % GDP ........................................................................................................... 107 Figure 8.1. Trends in Tax Revenues and Pension Contribution, % GDP .................................................................. 110 Figure 8.2. Structure of GDP, %GDP ........................................................................................................................ 112 Figure 8.3. Imports and VAT on Imports, % GDP .................................................................................................... 112 Figure 8.4. Structure of Taxes in 2012 (% of Total) .................................................................................................. 112 Figure 8.5. Dynamics of Tax Structure (2004-12)..................................................................................................... 112 Figure 9.1. Revenues, Excluding Grants, Status Quo % GDP ................................................................................... 124 Figure 9.2. Spending Composition in 2012 Excluding Grants, Status Quo % GDP.................................................. 125 Figure 9.3. Spending Composition Excluding Grants in 2018, Status Quo % GDP.................................................. 125 Figure 9.4. Expenditures Excluding Grants, Status Quo % GDP .............................................................................. 126 Figure 9.5. Primary Deficits Excluding Net Lending and Grants, Status Quo % GDP ............................................ 126 Figure 9.6. Revenues, Excluding Grants, Adaptive Scenario, % GDP ...................................................................... 126 Figure 9.7. Evolution of Spending Composition, Adaptive Scenario % GDP ........................................................... 128 Figure 9.8. Spending Composition Excluding Grants in 2018, Adaptive Scenario % GDP ...................................... 128 Figure 9.9. Expenditures Excluding Grants, Adaptive and Status Quo Scenarios % GDP ...................................... 129 Figure 9.10. Primary Deficits Excluding Net Lending and Grants, Adaptive and Status Quo Scenarios % GDP ... 129 Figure 9.11. Evolution of Spending Composition Under Energy Crisis Scenario, % GDP....................................... 130 Figure 9.12. Potential Resources Available for KRPP, € Mil. ................................................................................... 130 BOXES Box 1.1. Analytical Tools Used in the PFR and Beyond the PFR .................................................................................7 Box 2.1. Public Financial Management and the Fiscal Rule ....................................................................................... 13 Box 3.1. District Heating ............................................................................................................................................. 25 Box 3.2. Energy Efficiency ......................................................................................................................................... 28 Box 3.3. Kosovo Energy Sector Impact Tool .............................................................................................................. 30 Box 4.1. Economic Evaluation of Route 6 .................................................................................................................. 50 Box 4.2. East-West Railway Rehabilitation Proposal.................................................................................................. 51 Box 4.3. Options for Funding Rail Infrastructure Maintenance .................................................................................. 54 Box 5.1. Overview of the Education System in Kosovo ............................................................................................. 61 Box 5.2. Pre-School and Early Childhood Education .................................................................................................. 64 Box 5.3. Improving Education Outcomes by Improving Teacher Selection ............................................................... 67 Box 5.4. Education Financing Reforms ....................................................................................................................... 69 Box 6.1. Modeling the Impact Health Insurance Law ................................................................................................. 93 Box 7.1. Energy Subsidies and Costs ........................................................................................................................ 104 EXECUTIVE SUMMARY A. CONTEXT Kosovo is Europe’s youngest country, both in terms of history and demographics. As part of the former Yugoslavia, Kosovo became a separate territory under United Nations administration in 1999, and declared its independence in 2008. By February 2014 it had been recognized by 106 UN member states including 23 out of 28 EU members. Kosovo is a potential candidate for European Union (EU) membership and is currently negotiating a Stabilization and Association Agreement (SAA) with the European Commission. From 1999 Kosovo has been using the Deutsche Mark and, since 2002, its successor currency, the Euro, as legal tender. Kosovo is a landlocked country in South East Europe (SEE) with about 1.8 million inhabitants and a large migrant population based mainly in Western European countries. Kosovo has taken great strides to rebuild an economy destroyed by the collapse of Yugoslavia and the 1998-99 war, with sound fiscal numbers and budgets focused on capital expenditure. In general, Kosovo’s headline macroeconomic indicators are also relatively sound. Growth has averaged over 4 percent since 2000 and remained positive after 2008 during the global crisis years. Household survey evidence for 2006-11 suggests that economic growth benefitted all sections of society but the poorest 40 percent of the population saw consumption rise faster than wealthier groups. Public and private investments have made large contributions toward economic growth over the last five years as efforts to rebuild the economy continued. Headline fiscal policy has been counter-cyclical and implemented with the intention of supporting economic growth. The budget recorded surpluses prior to the crisis, which increased cash reserves. These reserves helped to finance low fiscal deficits – averaging 1.8 percent of GDP – between 2008 and 2013, supporting economic growth during the crisis. Low public debt – less than 10 percent of GDP at end- 20131 – also provided room to finance the fiscal deficits. Over the last five years, public spending has been heavily focused on capital expenditure. Capital expenditure averaged 39 percent of total public spending or 11 percent of GDP between 2008 and 2012, over 2.5 times the amount allocated in neighboring countries. Capital spending has been heavily focused on transport projects. Between 2010 and 2013, 40 percent of capital spending was on a new highway. Road improvements have helped to increase regional connections and access to markets, including through ports in Albania. They have also contributed to the development of a nascent construction sector with over half of material inputs and 80 percent of subcontractors for the highway to Albania being sourced locally (although this includes fuel). Investment has also been high in the health sector – averaging 13 percent of total public health spending (or 0.3 percent of GDP) between 2010 and 2012 – with the aim of improving medical equipment at the country’s single tertiary healthcare facility. Kosovo recently introduced a new Fiscal Rule, which aims to ensure that fiscal deficits and public debt are maintained at sustainable levels into the future. The Fiscal Rule requires deficits to be limited to 2 percent of GDP except under specific circumstances. For example, privatization proceeds can be re-invested without counting toward the 2 percent deficit limit. Despite its positive aspects, fiscal policy faces both efficiency and allocation issues. There is a tendency to over-invest in new capital stock and under-invest in maintaining existing capital stock, a particular issue in the case of roads, but many schools also require improvements and some new health facilities have been poorly constructed. In addition, Kosovo has under-invested in education and health. With a young and growing population, this could harm economic growth and public finances in the long 1 Excluding Paris and London Club debt serviced by Serbia until its recognition of Kosovo’s independence. i run. The efficiency of public spending could also be enhanced. For example, health resources could be refocused on primary care to achieve better health outcomes and education funds could follow pupils to ensure that schools have the resources they need to cope with rising enrollment. In some areas there may be scope to reduce spending while achieving the same results, but careful analysis should be undertaken to ensure essential public services are not reduced. B. CHALLENGES Despite its steady progress, Kosovo still faces numerous challenges. At around €2,900 in 2013 (or approximately €6,000 when adjusting for price levels), Kosovo has the second lowest GDP per capita in Europe after Moldova. As a lower-middle-income country, Kosovo’s main challenges are related to its level of development, and high rates of unemployment and poverty. Kosovo also has a young and growing population, posing challenges for youth employment and skills. Finding ways to invest in its young population could help to overcome some of Kosovo’s main challenges. Unemployment and Poverty Kosovo has high unemployment and low labor market participation, especially among youth and females. Unemployment was estimated at 30.9 percent in 20122, the highest in SEE. Continued economic growth, as well as emigration, seems to have helped to reduce unemployment in recent years, but the lack of comparable data prior to 2012 makes it difficult to estimate this effect. Youth unemployment is especially high, at over 55 percent. Every year approximately 30,000 new jobseekers enter the labor market, while current economic growth generates around 15,000 new jobs per year. As a result, only about 10 percent of the active youth population is working, and this falls to less than 5 percent for females. Worse, over a third of young Kosovars are neither in education nor in employment risking a “lost generation”. Even for those with higher education, prospects appear to be worsening, with some evidence suggesting that increasing numbers of those with university degrees are unemployed. In 2012, 15.6 percent of degree-holders were unemployed, possibly reflecting a mismatch between education and demand for skills in the labor market.3 At 63 percent of the population (including discouraged jobseekers) the total inactivity rate is the highest in in Europe, with the female inactivity rate standing at 82 percent. Poverty remains widespread in Kosovo, compounded by the fact that social protection and other government services fail to reach many of the poor. In 2011, 29.7 percent of the population was living in poverty and 10.2 percent was living in extreme poverty. With just one third of households in the bottom quintile receiving social assistance, improvements in the last-resort social assistance scheme could help to reduce extreme poverty. Reducing poverty remains one of the main challenges to overcome. There is a negative association between poverty and level of education. In 2011, more than one third of individuals who did not complete primary education were poor compared with 12 percent of individuals with university degrees. Almost two thirds of the poor possess less than secondary education. Enhancing education in terms of coverage and quality will be a key factor in reducing poverty. Evidence suggests that early childhood education is especially effective in helping individuals break out of the poverty trap. In addition, health care utilization rates by the poor are low, suggesting that they may be underserved. 2 Labor Force Survey 2012, Statistics Agency of Kosovo. 3 Another problem, not discussed much so far, is underemployment. Many young professionals end up in jobs below their education qualifications - many university diploma holders work as waiters in bars and restaurants owing to lack of alternative opportunities. ii A Young Demographic Kosovo’s young population presents some opportunities, challenges and risks. Kosovo’s population is the youngest in Europe, with an average age of about 26.4 Kosovo does not face the fiscal and economic pressures resulting from an aging population, a problem faced by many other eastern European countries.5, 6 A well-trained labor force could be a comparative advantage and a key resource for economic growth. This requires a policy focus to ensure that public spending is in line with the needs of a youthful population and the firms operating in Kosovo. Kosovo can only benefit from the “demographic dividend” that comes with a youthful population if it makes education a priority. Using this window of opportunity to invest in proper education and training will help workers to find employment, increase their tax contributions, and give firms access to the skills they need. It will also boost the productivity of the workforce and help Kosovo to move into the modern economic sectors that will increase growth. However, if that is not done, the ‘demographic dividend’ may well turn into a ‘demographic curse’.7 To address these challenges, two key obstacles to economic growth must be confronted. First, revenues collected at the border have begun to decline and will likely fall further. These “border revenues” – made up of trade taxes and the value-added tax (VAT) charged on imported goods8 – comprised 71 percent of total revenue in 2012. With slowly expanding domestic production substituting for imports, the recent signing of a trade agreement with Turkey, and increasing ties with the EU, border revenues will likely decline further over the medium term. Finding domestic substitutes for border revenues will be essential to finance the public investment needed for sustained economic growth and to address Kosovo’s many challenges. Second, Kosovo faces a potential energy crunch. With growing energy demand and an aging power station slated for closure, Kosovo may not be able to meet the energy needs of its firms and households. Economic growth, already harmed by power cuts, risks being prejudiced further. Revenue and Energy Revenues collected at the border make up over 70 percent of total revenues but have already shown signs of a decline that will continue. Customs tariff revenue and other revenue collected at the border began slowing during 2012 and continued to decline in 2013.9 A problem of dependency on import- related tax revenues has become more visible in the last two years. A Stabilization Association 4 In Kosovo, 38.1 percent of the population is aged less than 19, much higher than neighboring countries such as Albania (32.5 percent), Bosnia (24.6 percent), Serbia (23.4 percent), Slovenia (19.1 percent) or Montenegro (26.7 percent). 5 See World Bank “From Red to Gray: The Third Transition of Aging Population in Eastern Europe and the Former Soviet Union”. 6 In Eastern Europe, Bulgaria in particular faces large challenges due to its aging population. See World Bank, 2013, “Mitigating the impact of an aging population: Options for Bulgaria”. 7 A potentially interesting, though imperfect, comparison could be with Tunisia. See, for example, Boughzala (2013), “Youth Employment and Economic Transition in Tunisia”, Brookings Global Economy and Development Working Papers 57. 8 VAT is usually not considered a border tax because it is meant to be collected from the end consumer. However, in Kosovo about 81 percent of VAT is collected at the border, and VAT makes up 50 percent of border revenues. Only about 7.7 percent of VAT collected at the border is refunded internally. These facts mean that most VAT can be considered as “border revenue” in Kosovo. 9 Kosovo Customs reported that 2013 revenues collected at the border fell short of the plan by 8.8 percent and were by 1 percent lower than in 2012 but improvements have been seen in the first few months of 2014. iii Agreement (SAA) with the EU10 and a trade agreement with Turkey11 will almost certainly result in further declines in border revenues over the medium term. When other SEE countries signed SAAs, border revenues declined steeply as behaviors adjusted in anticipation of their implementation. The Ministry of Finance (MoF) has estimated that the SAA with the EU will have a negative impact on revenues of about €60 million annually, or 1.2 percent of GDP, while a trade agreement with Turkey is estimated to reduce revenues by €20 million or 0.4 percent of GDP annually. These two agreements are expected to become effective in 2016 and 2015, respectively. While revenue collection has recently begun to slow, there are options to compensate for declining border revenues. Revenue collection did not achieve its 2012 plan and worsened further in 2013, though collections have performed better in the first quarter of 2014. It is too early to assess whether this represents a full recovery. The main shortfall came from the revenues collected at the border due to import substitution, lower import prices and a decline in domestic demand. Efforts could be made in several areas to compensate for declining border revenues. A large tax gap combined with poor corruption indicators are indicative of high tax evasion, which could be clamped down on. Increasing collections at the local level are also an option, notably through property taxes, which is the lowest in the SEE, and user-fees, for example for roads. In addition, the introduction of universal health insurance could help to expand the tax base and raise new revenues for an under-funded health system. A second potential obstacle to growth comes from the fact that Kosovo’s energy sector fails to produce enough energy for the country’s needs and the resulting “power gap” is a barrier to growth and doing business in the country. According to a recent USAID study12, unreliable power supply costs Kosovar businesses about €260 million or 5 percent of GDP per year. The 2009 Enterprise Survey found that the problem of electricity supply was the single largest obstacle to doing business in the country. Small and medium enterprises are the most affected group. With the scheduled closure of Kosovo A in the near future, one of the country’s two power plants generating about one third of its electricity, the power gap will widen further unless significant investments in energy efficiency and new power generation facilities are realized. Energy is closely linked to fiscal sustainability. The two are closely linked through loans and grants from the government to the energy sector, through the need to protect the poor from energy price increases, and the potentially costly need to ensure sufficient energy supply for the country:  Between 2008 and 2012 debt owed by companies (mostly state-owned) in the energy sector, increased ten-fold. By the end of 2012, energy companies’ debt reached 5 percent of GDP or over three quarters of their annual income. In addition, they depended on grants from donors and loans from the GoK for around 20 percent of their total income.  If sufficient resources cannot be raised from the private sector to safeguard future energy supply, the GoK will need to find alternative ways to finance the costs of constructing a new power plant, which are estimated at 20 to 30 percent of GDP. Energy efficiency measures and renewable energy will also be important. Some of these costs may have to be covered by the budget.  These large investments will increase energy generation costs. In addition, current tariffs are not at cost-recovery levels. If tariffs are raised to reflect costs, then steps will need to be taken to protect poor and vulnerable groups through the social protection system. 10 Kosovo started Negotiations for Stabilization Association Agreement with EU on October 28 th 2013. The agreement is expected to be signed during the fall of 2014. 11 A trade agreement with Turkey was signed in October 2013 and will be effective starting from 2015. 12 The Effect of Unreliable Power Supply and Quality on Kosovar Businesses, USAID 2012. iv C. INCREASING THE EFFICIENCY OF PUBLIC SPENDING Despite Kosovo’s success focusing public spending on capital investments, resources have not always been allocated to best meet these challenges. In particular, while Kosovo has invested in new physical capital stock, it has not paid enough attention to its existing physical capital and its human capital. It is important to strike a balance between investing in a new capital stock and maintaining existing facilities. While Kosovo has a good record in investing in new capital stock, some basic maintenance has been neglected. Alongside the new highways, existing roads are in poor condition. This stems from a road maintenance budget that is less than half of neighboring Serbia or Bosnia and Herzegovina on a per kilometer basis. The road rehabilitation budget has been crowded out by new investments and, by 2012, reached just a quarter of its 2009 (pre-highway construction) levels. Similar declines occurred in budgets to finance local and national roads. A recent study identified a need for 39 local road improvements with economic rates of return of up to 81 percent. The required average annual investments to cover these improvements would be less than a quarter of the cost of the highway construction but these have barely been touched. Similarly, in the health sector, investments have been made in new medical equipment and buildings while there are deficiencies in basic sewage and water supply systems to health institutions. Decisions on current expenditures have not consistently followed an economic rationale. Across-the- board salary increases for public servants have often been politically-motivated and have rewarded good and bad performers equally and put pressure on ministries with high wage bills. For example, between 2008 and 2012 salaries rose from 76 to 83 percent of the education budget and from 67 to 70 percent of the health budget, crowding out other spending. Further, the social protection budget has been progressively re-oriented away from protecting the poor and toward special interest groups. War-related benefits and pensions for favored groups swallowed almost all increases in social protection spending between 2008 and 2012. At the same time, spending on the last resort social assistance scheme, which targets the poorest, and the universal basic pension decreased by 0.2 percent of GDP. Kosovo under-invests in one of its main assets: its people. With 38 percent of its population less than 19 years of age and one of Europe’s few growing populations, Kosovo has an opportunity to benefit from a “demographic dividend” if it invests in its youth. Instead, the country invests 4.1 percent of GDP in education – below its regional and income group peers, especially when accounting for its youthful population. Few children benefit from pre-primary education despite the mounting evidence showing its importance for lifetime outcomes and its role in escaping inter-generational poverty. Public health spending is similarly under-funded, at just 2.6 percent of GDP. It fails to meet the most basic needs of its population resulting in high rates of preventable maternal and child mortality. The poor suffer the brunt of under-investment in health and education. Kosovo’s educational enrollment rates compare well with neighboring countries but, at 2.9 percent, non-enrollment rates are over twice as high for children from poor households as for others and worse still for the extreme poor. At the same time, there is evidence that the bottom quintile lacks sufficient access to healthcare due to high out-of-pocket payments for medicines. Simulations using household data show that the poverty rate is around 7 percent higher because of spending on medicine, and the middle class is becoming increasingly vulnerable to impoverishing health expenditures. Some health services that poorer groups are more likely to consume are especially under-funded. For example, the share of public health expenditure on primary care fell from 21 percent in 2007 to just 13 percent by 2012. Pressure to continue to spend in non-urgent areas continues to be strong. Spending pressures today stem largely from politically powerful lobby groups. Pressures include (i) pensions; (ii) veteran benefits; (iii) benefits to former political prisoners; (iv) public salaries; and (v) low-value transport infrastructure. v These expenditures do not contribute to ensuring sustained growth, providing employment, and improving the living conditions of Kosovo’s citizens. They risk squeezing investments in health, education and other capital projects, which could boost growth and reduce poverty. The 2014 budget and pre-election promises since then illustrate these pressures. Political prisoners and new war veterans’ benefits together have been allocated €15 million13 in 2014 budget, while public sector wage increases have been allocated €45 million (0.9 percent of GDP). Wage allocation of €45 million covers the implementation of the pay and grading reform14, inclusion of Kosovo Serb structures of the northern municipalities into the Kosovo system15 and wage increases. Pre-election promises raised further the costs of salary increases to over 2 percent of GDP per year. While only €15 million has been allocated in the budget for newly registered war veterans, the benefit package looks to be large with cash payments alone reaching an estimated €40 million (0.8 percent of GDP) per year or more. Additional health, education, housing, business investment and tax benefits, among others, will likely further increase spending over time. However, these costs are difficult to estimate since the interpretation of the new law on war veterans is still unclear and the process of verifying applicants is ongoing so the number of beneficiaries is not clear.16 Full fiscal costs could be high. A new highway connecting Prishtina with the FYR Macedonian border risks crowding out other public investments. A recently approved17 highway between Prishtina and FYR Macedonia – Route 6 – is estimated to cost approximately €750 million18 or 14.4 percent of GDP. It will crowd out other high priority capital investments, as did the nearly-completed Route 7 to Albania. D. IMPROVING PUBLIC EXPENDITURE ALLOCATIONS Without revenue and spending restructuring, Kosovo’s mix of declining revenue and increasing entitlements would result in either a rapid rise in the fiscal deficit or the crowding out of growth and development-focused spending. Simulations of this “status quo scenario” suggest that non-grant revenues would decline from 25.8 percent of GDP in 2013 to 24.7 percent by 2018 due to reduced border revenues narrowing public resources. At the same time non-grant spending would increase slightly from 28.7 percent in 2013 to 29 percent by 2018, driven by increases in veterans’ benefits and increased public service salaries. The deficit would hit 4.3 percent of GDP by 2018 – significantly above the 2 percent limit set by the new Fiscal Rule19. Kosovo’s alternative would be to further squeeze investment in its 13 In December 2013 Government of Kosovo (GoK) decided to grant political prisoners’ scheme a lump sum of €55 million as compensation for their time in prison, to be paid over four years. 14 The implementation of the pay-and-grading reform is estimated to cost about €18 million though there are no guarantees that the reform will take place on time and that this money will not be used for wage increase during this election year. 15 The political agreement between the Prime Minister of Kosovo and the Prime Minister of Serbia on April 19th 2013, foresees dismantling Serbian parallel structures in northern municipalities of Kosovo and the inclusion of some of their personnel into the Kosovo government system. Though there is no clear cost estimate, the cost is expected to be about €13 million in wages and salaries – which is included in the current allocation of €45 million. 16 Over 67,000 people applied for the status of war veterans. The verification process has been slow and is expected to be completed by June 2014. Of the number of verified applications about 35 percent is considered not legitimate. A simple calculation would suggest that the number of legitimate veterans will be about 40,000. 17 This new highway (Route 6) will connect Prishtina with Macedonia at the Hani i Elezit Border Crossing Point. It was approved by an inter-ministerial Steering Committee on January 10th, 2014, and the contract was awarded to the US-Turkish consortium Bechtel-ENKA, which also built the Route 7 Highway connecting Prishtina and Tirana, Albania. 18 €600 million for the project plus an estimated €150 million for expropriations. 19 The main provisions of the Fiscal Rule are: (i) Maintain the overall fiscal deficit at 2 percent of projected GDP in the current budget year. This deficit level applies to the consolidated central and municipal budget; (ii) The unspent part of appropriations from the previous year can be used as a source of financing new budget appropriations in vi people and/or expenditure on maintaining its infrastructure. Worse, the investments in a non-urgent highway, spending on veterans’ benefits and on salary increases would leave little room for maneuver if public resources are needed to ensure sufficient power supply. Kosovo’s record of budgetary discipline shows that the country is capable of taking difficult decisions necessary to meet its challenges, and it now has an opportunity to reorient spending toward areas that address these real challenges. Aligning spending with the country’s real needs will require more than ensuring that the headline fiscal numbers are sound, as the Fiscal Rule aims at ensuring. Although there are genuine needs in many areas, resources being limited means that they should be reallocated toward areas that unlock economic growth potential and help reduce joblessness and poverty. In addition, improving public services will require efficiency improvements, notably in health and education. Addressing Kosovo’s main challenges will require efforts to (i) increase domestic revenue collections to compensate for falling border revenues; and (ii) focus the budget on health, education and on protecting the poor, while maintaining and improving existing infrastructure. These investments can be funded by delaying the construction of the Route 6 Highway, reducing subsidies to the energy sector, diverting funds away from healthy, working-age war veterans toward the poor, and by adjusting public service salaries at a slower pace that is delinked from the political cycle and with more emphasis on rewarding performance than across-the-board increases. However, the value of investments in other areas will suffer if Kosovo is unable to provide sufficient energy for its households and firms. Therefore, significant public funds may need to be diverted toward the construction of new power generation facilities. There is also ample room to increase efficiency within sectors. In education, spending should follow pupils as the demographic composition changes and as families migrate from rural to urban areas. In health, improved outcomes could be achieved by focusing on primary care to meet basic needs rather than hospitals. In transport, the quality of the total road stock could be improved by reallocating resources away from the construction of some new roads and toward better maintenance of existing ones. In social protection, support should focus on the poor rather on special interest groups. In all sectors, improved procurement practices could save significant sums. In particular, there is overpayment for pharmaceuticals and schools and new road construction is expensive given the terrain. capital expenditure, and will not be considered for purposes of compliance with the fiscal deficit ceiling; (iii) Other revenue receipts additional to budgeted revenues from the previous year, may be transferred to the following year for new budget appropriations in the category of capital expenditure, and will not be considered for purposes of compliance with the fiscal deficit ceiling; (iv) Privatization proceeds and their spending will not be subject to the fiscal deficit ceiling. These revenues can be used only for capital investment projects, if the level of usable government bank balance amounts to at least 4.5 percent of projected GDP; (v) any excess of the projected fiscal deficit of more than 0.5 percent of GDP should be compensated within the next three fiscal years. In addition, the Fiscal Rule specifies situations in which the rule may be temporarily suspended, such as: (i) Economic recession, when nominal tax revenues are equal to or lower than the tax revenues collected during the same period of prior fiscal year, excluding the impact of policies and one-off tax revenues; (ii) An emergency such as a natural disaster; (iii) In the event of a crisis in the banking system identified by both the Minister of Finance and the Governor of the CBK, on the proposal of the CBK; or (iv) A call on a state guarantee that will have to be paid by the government and that has an impact on overall expenditures over 1.5 percent of projected GDP. vii Specific recommendations for each sector point both to the need to reallocate resources between sectors and within sectors with the aim of achieving better value-for-money.  Energy: (i) ensure that the energy demand is met, through a well-coordinated energy strategy combining the promotion of energy efficiency, renewable energy and a new thermal power plant, even if public resources need to be found by delaying other investments; (ii) ensure that a tariff trajectory is set so that over time tariffs reflect costs and provide further incentive to reduce loses and encourage private sector participation while guaranteeing social protection mechanisms that sufficiently protect the vulnerable from tariff increases.  Transport: (i) delay the construction of Route 6 and divert these resources to better maintaining existing roads and improving or upgrading rural and regional roads (with higher economic rates of return than Route 6), as well as to priority sectors like energy; (ii) increase road user charges and vehicle registration fees to ensure sufficient funds for road maintenance in the long term.  Education: (i) use some of the resources saved through the delay of Route 6 to increase resources available for pre-university education. These resources should be used to increase enrollment in pre-primary education, which has high economic and social rates of return, and to ensure sufficient resources in upper-secondary education to cope with demographic changes; (ii) increase the efficiency of existing resources by identifying savings that could be made through the rationalization of the school network – especially in rural areas with small classrooms – and by taking measures to improve the school financing formula such as simplifying it and providing sufficient training to school-level financial managers; (iii) increase teacher training opportunities and depoliticize wage increases so that they are based partly on training and performance.  Health: (i) achieve better health outcomes and expand health coverage to poorer groups by refocusing spending on primary health care and improving maintenance of existing facilities rather than spending on new facilities or health equipment that benefit only small groups of people; (ii) improve the procurement system for pharmaceutical products to achieve better value-for-money and expand their provision, particularly for poorer households that are under- served; (iii) introduce a contributory health insurance scheme to raise additional resources for the health sector and help to raise additional domestic revenue while expanding the tax base. This would raise additional resources for the under-financed health sector, regardless of whether Route 6 is delayed or not.  Social protection: (i) use the social protection budget to protect the poor rather than provide benefits for special interest groups. In particular, rethink the proposed new Law on Veterans’ Benefits, which risks costing upward of 0.8 percent of GDP per year in cash benefits alone and may cost significantly more in other benefits and reduced tax revenues; (ii) expand the social protection system to increase coverage of the poor and vulnerable including from the effects of increased energy prices; (iii) maintain the current pension system, which is already sustainable and brings pensioners’ income up to the poverty threshold, and refrain from unsustainable increases. Even if Route 6 is not delayed, the social protection budget could receive more resources by reallocating energy subsidies to protect the most vulnerable.  Revenue: In addition to expanding revenues through health insurance and transport sector charges: (i) strengthen tax administration to reduce tax evasion and expand the tax base; (ii) increase efforts to expand the collection of property tax, one of the least distortionary taxes; (iii) consider changes to the personal income tax to reduce taxes at lower levels and increase rates at higher levels. This would also encourage labor market formality. viii Implementing these recommendations would reduce fiscal deficits while refocusing the budget in a way that better addresses Kosovo’s needs. An illustrative simulation of this “adaptive scenario” shows how revenues may perform with such reforms and how spending within a similar budget envelope as today could be refocused. Introducing obligatory public health insurance is estimated to raise around 0.8 percent of GDP in revenue by 2018, and property tax receipts could double to 0.5 percent of GDP (though remaining low by regional standards) thanks to efforts to increase land and building registration and enhanced enforcement, going some way to compensate for declining border revenues. Increased domestic VAT collections – both as a result of improved tax administration and economic growth – will also help. Between them, these and other measures could help to boost non-grant revenue from 25.8 percent of GDP in 2013 to 26.2 percent by 2018. At the same time, spending on road construction would be reduced from 5.7 percent of GDP in 2012 to 2.1 percent by 2018. As energy prices begin to better reflect costs, public spending on the sector would fall from 0.8 to 0.3 percent of GDP. Other areas would see expansions: education spending would increase from 3.9 to 4.8 percent of GDP; and health – funded through health insurance contributions – would rise from 2.6 to 3.7 percent of GDP (both excluding grants). Social protection spending would need to increase both to combat poverty and to help cushion the poor and vulnerable from the shock of increasing energy prices from 3.9 to 4.7 percent of GDP. Other economic affairs, including road maintenance, would increase from 2.1 to 2.7 percent of GDP. In total, under this adaptive scenario, non-grant expenditures would fall from 28.7 percent of GDP in 2013 to 28.2 percent in 2018 with a large initial fall. Between them, the increased revenue and reduced but refocused spending would ensure that the fiscal deficit would be contained to 2 percent or below each year between 2014 and 2018. Public resources may need to be diverted from Route 6 to the energy sector if required to ensure sufficient energy supply. Within the same fiscal envelope and same deficit levels as in the previous scenario, around €550 million could be made available between 2014 and 2018 to help construct new power generation facilities, if required, by delaying the construction of Route 6. This might not be sufficient to build the full 2×300 MW20 capacity power plant from only public sources, but (i) could support a public private partnership for the 2×300MW plant with public and private financing or (ii) could complement funding from international borrowing. However, since Kosovo does not have experience in international debt markets and has never issued long-term debt, this option would present strong challenges but may be an option with the support of international financial institutions. Other options could also be considered to bridge a financing gap, such as phasing the construction over time by providing the necessary funds to construct just one of the planned units – a 300 MW capacity power plant. This would be sufficient to close the power gap in the immediate future but would most likely increase the cost per unit of capacity due to reduced economies of scale compared to a single step construction. E. CONCLUSION Kosovo has invested heavily in new physical capital, helping to support economic growth and compensating for damage due to the breakup of Yugoslavia and the war, but it has neglected to maintain its existing capital stock and has not invested sufficiently in its human capital. Kosovo’s efforts to maintain high investment is commendable but efforts should also be made to ensure that the existing capital stock is well maintained. Long-term economic growth also depends on investing in areas that are not classified as capital expenditure, such as investment in human capital through education and health. Kosovo’s high poverty and low employment can be partly tackled by investing in its people and this can also boost economic growth. But instead of addressing poverty and unemployment, a proposed expansion of veterans’ benefits would further orient the social protection budget toward special interest 20 The proposed plan is to construct 600 MW of generation capacity in two units, each of 300 MW. ix groups rather than the poor, and it even contains incentives for unemployment or to exit the formal labor market. This is occurring at the same time as potential energy price increases due to the lack of increase in the last two years and to the need for implementing energy efficiency investments, renewable energy investments and replacing Kosovo A with a new power plant. Potential energy price increases mean that the poor and vulnerable will require more targeted social protection services. As well as investing in its people, growth will require keeping the lights on, and the GoK should keep in mind that public resources may be required to ensure sufficient power for its households and firms. Meeting all of these challenges will be tough in an environment with falling border revenues, which make up over 70 percent of total revenues. Policy and administrative efforts will need to be made to increase domestic revenue to compensate. Kosovo has recently passed a Fiscal Rule to limit fiscal deficits and public debt to sustainable levels in the future, while allowing additional investment from privatization receipts or the previous year’s savings. But while the headline fiscal numbers look good, the challenge will be to refocus spending to better address the country’s needs. The Public Finance Review (PFR) analyzes the efficiency of spending between 2008 and 2013 and proposes ways allocations can be adjusted both within and between sectors to better address Kosovo’s main challenges. x F. TABLE OF TOP THREE RECOMMENDATIONS BY SECTOR Main recommendations Continue to maintain sustainable fiscal deficits and debt levels by adhering to the recently passed Fiscal Rule. Ensure that energy demand is met to maintain economic growth, while ensuring tariffs move toward a cost- recovery trajectory Reallocate resources away from the proposed Route 6 toward infrastructure maintenance and investment in social sectors. Depoliticize public servant wage increases by subjecting them to reviews by an independent panel, performance based increases or by subjugating them to regular increases according to a formula, depending upon the requirements of each sector. Reallocate resources away from the proposed expansion of veterans’ benefits and toward the poor and vulnerable, including to shield them from the impact of increased energy prices. Introduce a public health insurance scheme to raise additional resources for the under-funded health sector. Make efforts to achieve better value-for-money in public procurement where the cost of pharmaceuticals and road maintenance notably appear to be expensive. Take steps to increase domestic revenue collection to compensate for declining border revenues by, inter alia, abandoning some tax exemptions, pursuing evaders through the legal system, and considering raising some tax rates. Improve capacities of Tax Administration and of local authorities for property tax collection in municipalities. Energy Ensure that energy demand is met rapidly by facilitating investments and support policies that will improve energy efficiency in the public and private sectors to reduce energy use and generate fiscal savings. Reallocate public resources to energy sector investments as necessary by reprioritizing other public investments. Encourage the supply of renewable energy by simplifying the complex application for renewable energy sources, promoting sustainable and renewable solutions for space and water heating. Ensure that the tariff trajectory is on a path to fully reflect costs and provide incentives to save energy. Guarantee the complete autonomy of the Energy Regulatory Office to maximize the potential for private sector participation in energy and minimize energy subsidies, while using the social protection mechanism to protect poor and vulnerable households from tariff increases. Transport The resources that would be used to construct Route 6 could be put to better use both within the transport sector (notably by maintaining better or upgrading existing roads) and other sectors. Increase road user charges and vehicle registration fees to increase resources available for the long-term maintenance of roads. Find ways to make the railway network financially sustainable such as by reducing services and focusing on exporting industries. Education Expand access to education, particularly pre-primary, financed partly through cost savings resulting from demographic changes, and understand better the needs in upper-secondary education that will result from demographic changes, increased enrollment and rural-to-urban migration. xi Improve the functioning of the municipal-to-school funding formula by simplifying it, providing sufficient training to implement it for school-level management, and encouraging a strong political commitment at the municipal level to rationalize the school network. Ensure that the right incentives are in place to attract and retain highly effective teachers including by ensuring that teachers receive the mandated training and that remedial measures be taken in cases of poor teaching performance. Health Refocus spending on primary health care while restricting further capital investments to prioritize maintenance of existing facilities, and curtailing non-performance-based wage increases for medical staff. Increase the value-for money of pharmaceutical purchases by improving procurement systems, and increase spending on medication and supplies, particularly by using some of the new funds raised through the introduction of health insurance. Introduce a public health insurance scheme with the following conditions: (i) the contribution rate should be set at 5 rather than 7 percent; (ii) only poor groups should be exempt from premium payments and co-payments; (iii) health insurance should be implemented in a fiscally neutral fashion to safeguard the sustainability of the budget and guarantee that other sectors do not suffer an undue burden; (iv) Build adequate capacity to administer a health insurance system and ensure a smooth transition from the Health Financing Agency to Health Insurance Fund. Social Protection Begin efforts to introduce a system that better identifies the poor and vulnerable to enable an expanded coverage of the last-resort social assistance scheme while continuing to minimize leakage to wealthier groups and ensuring that the social assistance scheme provides necessary support for households to climb out of poverty, for example, by providing training and by allowing benefits to “phase out” when a beneficiary finds employment rather than losing benefits instantly. Continue to implement the current pension system without large changes; there is no need to “fix” something that is not broken. The proposed new Veterans’ Law should be fully costed and rethought before it is approved and implemented otherwise this law that could have serious negative fiscal, economic and social consequences. Tax and revenue policies Strengthen Tax administration including by enhancing the quality and quantity of tax inspectors, improving risk assessment modules, adding the practice of random audits to assess areas for improvements in risk assessments, and increasing efforts to fight tax evasion. Consider making some changes to the tax system such as increasing the CIT rate to 15 percent and streamlining the rate structure for PIT to make it more progressive and equitable, and remove some of the exemptions that have no place in a good tax system such as: (i) dividends owned by foreign shareholders; (ii) interest on financial instruments issued by public authorities; and (iii) fringe benefits provided by employers. Take steps to increase collection from property tax through (i) expanding the property tax base by starting to tax agriculture land and other types of property; and (ii) considering increasing the rate of property tax collected by municipalities; and (iii) using central government grants in a way that rewards municipalities’ own-source generation of revenues. xii CHAPTER 1. INTRODUCTION A. CONTEXT 1.1. The Republic of Kosovo is the youngest country in Europe. The Socialist Federal Republic of Yugoslavia (SFRJ) disintegrated during the 1990s, and in June 1999 Kosovo became a separate territory under the United Nations Interim Administration Mission (UNMIK), a situation that lasted until February 2008. On February 17th 2008 Kosovo unilaterally declared its independence and by February 2014 it had been recognized by 106 UN member states, including 23 out of 28 EU members. Kosovo is a landlocked country in South East Europe21 (SEE) with about 1.8 million inhabitants, and a large migrant population mainly based in Western European countries. The country has taken strides to rebuild an economy destroyed by the collapse of SFRJ in early 1990s and then the 1998-99 war. 1.2. A potential candidate for European Union (EU) membership, the Republic of Kosovo is a lower-middle-income country with a solid economic growth performance since the end of the war in 1999. It is one of only four countries in Europe that recorded positive growth rates in every year during the crisis period 2008–12, averaging 4.2 percent. The resilience of Kosovo’s economy reflected: (i) limited international integration into the global economy; (ii) the success of its diaspora in the labor markets of, especially, the German-speaking countries of Central Europe, providing a steady flow of remittances; (iii) a generally pro-growth composition of the budget, with approximately 40 percent of public expenditures used for investments; and (iv) a steady inflow of donor support. 1.3. Despite its efforts, Kosovo still lags its neighbors and other European countries on most aspects of development. Kosovo has the lowest GDP per capita in Europe after Moldova, with a per- capita gross domestic product (GDP) of €2,858 in 2013. It also has the highest rate of poverty in SEE, with 29.7 percent of the population living below the poverty line in 2011, and it suffers from the highest rate of unemployment in SEE, estimated at 30.9 percent in 2012. Worse, 63 percent of its population is economically inactive. 1.4. A key reason why Kosovo is behind – besides its history – is its low institutional, physical and human capital and only sustained economic growth can help the country to address these challenges and catch up with its European neighbors. Closing these gaps will require additional investments, but Kosovo’s resources are insufficient to address all needs immediately. Thus, spending should be prioritized on those needs that are most important for economic growth. Second, resources should be used efficiently in order to maximize their impact in reducing extreme poverty and improving shared prosperity. B. KOSOVO’S CHALLENGES AND OBSTACLES TO GROWTH 1.5. Kosovo faces numerous challenges, many of which can be partly addressed through fiscal policy. As a lower-middle-income country, Kosovo’s main challenges are related to its level of development. In particular, its high rates of unemployment and poverty need to be addressed. Kosovo also has a young and growing population, posing challenges for youth employment and skills. Finding ways to invest in its young population could help to combat some of Kosovo’s main challenges. 21 Southeast Europe refers to the six countries of the Western Balkans: Albania, Bosnia and Herzegovina, Kosovo, FYR Macedonia, Montenegro, and Serbia. 1 1.6. To address these challenges two key obstacles to economic growth must be confronted. First, Kosovo faces a looming energy crunch. With growing energy demand and aging power stations slated for closure, Kosovo may not be able to meet the energy requirements of its firms and households. Economic growth, already harmed by power cuts, risks being prejudiced further. Second, revenues collected at the border have begun to decline and will likely fall further. These “border revenues” – made up of trade taxes and value-added tax (VAT) charged on imported goods22 – comprised 71 percent of total revenue in 2012. With slowly expanding domestic production substituting for imports, the recent signing of a trade agreement with Turkey, and increasing ties with the EU, border revenues will likely decline further over the medium term. Finding domestic substitutes for border revenues will be essential to finance the public investment needed for sustained economic growth, and to address Kosovo’s many challenges. This report evaluates Kosovo’s challenges and obstacles to growth, and provides policy recommendations to address them. Unemployment, Poverty and Shared Prosperity 1.7. High unemployment and low labor market participation, especially among youth and females, are key challenges for Kosovo. Unemployment was estimated at 30.9 percent in 201223, the highest in SEE. Continued economic growth, as well as emigration, has likely helped to reduce unemployment in recent years but lack of comparable data prior to 2012 makes it difficult to estimate this effect. Youth unemployment is especially high at over 55 percent. Every year approximately 30,000 new jobseekers enter the labor market but current economic growth generates no more than around 15,000 new jobs per year. As a result, only about 10 percent of the active youth population is working, and this falls to less than 5 percent for females. Worse, over a third of young Kosovars are neither in education nor in employment – a category that could be called “the lost generation”. Even for those with higher education, prospects appear to be worsening, with some evidence suggesting that increasing numbers of those with university degrees are unemployed. In 2012, 15.6 percent of degree-holders were unemployed, possibly reflecting a mismatch between education and demand for skills in the labor market.24 At 63 percent of the population (including discouraged jobseekers), the total inactivity rate is the highest in Europe, with female inactivity rate standing at 82 percent. 1.8. Poverty remains a scourge in Kosovo but social protection fails to reach many of the poor. In 2011, 29.7 percent of the population was living in poverty and 10.2 percent was living in extreme poverty (Figure 1.1 and Figure 1.2). Both poverty and extreme poverty worsened in 2011. With just one third of households in the bottom quintile receiving social assistance, improvements in the last-resort social assistance scheme could help reduce extreme poverty. Reducing poverty remains one of the main challenges. 22 VAT is usually not considered a border tax because it is meant to be collected from the end consumer. However, in Kosovo about 81 percent of VAT is collected at the border, and VAT makes up 50 percent of border revenues. Only about 7.7 percent of VAT collected at the border is refunded internally. These facts mean that most VAT can be considered as “border revenue” in Kosovo. 23 Labor Force Survey 2012, Statistics Agency of Kosovo. 24 Another problem, not discussed much so far, is underemployment, which seems to be quite evident since many young professionals end up in trivial jobs (many university diploma holders work as waiters in bars and restaurants) since the market does not produce enough high quality jobs. 2 Figure 1.1:Poverty Rates in Kosovo, 2009-2011* Figure 1.2. Extreme Poverty Rates in Kosovo, 2009-2011* 40 14 13.2 35.3 34.5 12.1 12.5 35 31.5 31.1 30.7 29.7 12 10.7 29.2 10.2 30 26.7 9.4 26.7 10 8.5 8.5 25 7.8 8 20 6 15 10 4 5 2 0 0 Urban Rural Total Urban Rural Total 2009 2010 2011 2009 2010 2011 Source: KAS and World Bank (2013) Source: KAS and World Bank (2013) *Poverty thresholds (€ /day): 2009: 1.55; 2010: 1.61; *Extreme poverty thresholds (€ /day): 2009: 1.02; 2010: 2011: 1.72 1.07; 2011: 1.20 1.9. Economic growth between 2006 and 2011 favored the poor and the bottom 40 percent of the population. Household data show that between 2006 and 2011, economic growth in Kosovo has been shared widely, leading to improvements in both poverty and “shared prosperity” (defined as the share of wealth going to the poorest 40 percent of the population). Over this period, both the poor and the poorest 40 percent of the population saw average annual increases in consumption of around 4 percent, while the wealthiest 20 percent saw their consumption rise by between one and three percent per year during the period. With a poverty rate of 29.7 percent in 2011, poverty and shared prosperity are closely connected in Kosovo. Therefore by focusing on poverty as one of the key issues in the country, this report also covers shared prosperity. 1.10. There is a negative association between poverty and the level of education, and the poor may not receive sufficient healthcare. In 2011, more than one third of individuals who did not complete primary education were poor compared with 12 percent of individuals with university degrees. Almost two thirds of the poor possessed less than secondary education. Enhancing education in terms of coverage and quality will be a key factor in reducing poverty. Evidence suggests that early childhood education is especially effective in helping individuals break out of the poverty trap. In addition, health care utilization rates by the poor are low, suggesting that they may be underserved. A Young Demographic 1.11. Kosovo’s young population presents opportunities and challenges. Kosovo’s population is the youngest in Europe with an average age of about 26 (Figure 1.3).25 It therefore does not face the fiscal and economic pressures resulting from an aging population, a problem faced by many other eastern European countries. 26, 27 A well-trained labor force could be a key resource for economic growth. 25 In Kosovo, 38.1 percent of the population is aged less than 19, much higher compared to neighbor countries like Albania (32.5 percent), Bosnia (24.6 percent), Serbia (23.4 percent), Slovenia (19.1 percent) or Montenegro (26.7 percent). 26 See World Bank “From Red to Gray: The Third Transition of Aging Population i n Eastern Europe and the Former Soviet Union”. 27 In Eastern Europe, Bulgaria in particular faces large challenges due to its aging population. See World Bank, 2013, “Mitigating the impact of an aging population: Options for Bulgaria”. 3 However, ensuring that public spending is in line with the needs of a youthful population presents a challenge. 1.12. Kosovo can only benefit from the ‘demographic dividend’ that comes with a youthful population if it makes education a priority. Using this window of opportunity to invest in proper education and training will help workers find Figure 1.3. Population Pyramid, 2011 employment, increase their tax contributions, and give firms access to the skills they need. It will also boost the productivity of the workforce and help Kosovo to move into the modern economic sectors that will increase growth. However, if that is not done, the ‘demographic dividend’ may well turn into a ‘demographic curse’. A Looming Energy Crunch 1.13. Kosovo’s energy sector fails to produce enough energy for the country’s needs and the resulting “power gap” is one of the major barriers to growth and to doing business in the country. According to a recent USAID study28 unreliable power supply costs Kosovar businesses about €260 million or 5 percent of GDP per year. Source: World Bank Health, Nutrition and Population The 2009 Enterprise Survey (see Energy Chapter) Statistics. found that the problem of electricity supply was the single largest obstacle to doing business in the country. Small and medium enterprises are the most affected group. With the closure of Kosovo A, one of the country’s two power plants, the power gap will widen further. 1.14. There will be a significant power gap after Kosovo A is closed if no new generation facilities are built. Despite improvement in domestic supply in recent years, due to improvements in electricity generation and assisted by government subsidies for imported electricity, the situation remains fragile. Economic growth will continue to lead to increased energy demand. Closing the power gap requires building a new power station which is estimated to cost between €1 billion and €1.5 billion – or between 20 and 30 percent of GDP – and will take at least around four years to complete. Declining “Border Revenues” 1.15. Revenues collected at the border make up 70 percent of total revenues but have already shown signs of a decline that will continue. Customs tariffs revenue and other revenue collected at the border began slowing during 2012 and continued to decline in 2013.29 The revenue decline in 2013 was partially driven by price effects and partially due to an import substitution effect. A Stabilization Association Agreement (SAA) with the EU30 and a trade agreement with Turkey31 will almost certainly result in further declines in border revenues over the medium term. The Ministry of Finance (MoF) has 28 The Effect of Unreliable Power Supply and Quality on Kosovar Businesses, USAID 2012. 29 Kosovo Customs reported that 2013 revenues collected at the border missed the plan by 8.8 percent and were by 1 percent lower than in 2012. 30 Kosovo on October 28th 2013 started Negotiations for Stabilization Association Agreement with EU, which are expected to conclude with an agreement signed during fall 2014. 31 A trade agreement with Turkey was signed in October 2013 and will be effective starting from 2015. 4 estimated that the SAA with the EU will have a negative impact on revenues of about €60 million annually, or 1.2 percent of GDP, while a trade agreement with Turkey is estimated to reduce revenues by €20 million or 0.4 percent of GDP annually. These two agreements will become effective in 2016 and 2015, respectively, but when other Western Balkan countries signed SAAs, border revenues declined steeply as behaviors adjusted in anticipation of their implementation. C. SPENDING PRESSURES DO NOT MATCH THE COUNTRY’S NEEDS 1.16. Pressure to spend in non-urgent areas is strong. Spending pressures stem largely from politically powerful constituents representing: (i) pensions; (ii) veteran benefits; (iii) benefits to former political prisoners; (iv) public salaries; and (v) transport infrastructure. Unlike the areas highlighted earlier, these expenditures are not critical to ensure sustained growth, provide employment, and improve the living conditions of its citizens. By contrast, this spending risks squeezing investments in health, education and other capital projects, which could boost growth and reduce poverty. The 2014 budget illustrates these pressures. Political prisoners and new war veterans’ benefits have jointly been allocated €15 million32 in 2014 budget, while public sector wages have been allocated €45 million. This €45 million covers the implementation of pay and grading reform33, inclusion of Kosovo Serbs structures of the northern municipalities into Kosovo system34 and wage increases. While only €15 million has been allocated in the 2014 budget for newly registered war veterans, the benefit package looks to be large with cash payments alone reaching an estimated €40 million per year. Additional health, educa tion, housing, business investment and tax benefits, among others, will likely increase spending on this over time though costs are difficult to estimate since the interpretation of the new law on war veterans still unclear and the process of verifying applicants is ongoing so the number of beneficiaries is not clear.35 Full fiscal costs could be high. 1.17. In addition, a new highway connecting Prishtina with the FYR Macedonian border risks squeezing other public investments. The recently approved36 Route 6 Highway between Prishtina and FYR Macedonia is estimated to cost approximately €750 million37 or 14.4 percent of GDP. The implementation contract is still under preparation and the length of the project is open, though early indications suggest that it is likely to be constructed over five years. However, it will likely squeeze other high priority capital investments, as did the nearly-completed Route 7 to Albania. In particular, it will 32 In December 2013 Government of Kosovo (GoK) decided to grant political prisoners’ scheme a lump sum of €55 million as compensation for their time in prison, to be paid over four years. 33 Implementation of pay-and-grading reform is planned to cost about €18 million though there are no guarantees that the reform will take place on time and that this money will not be used for wage increase during this election year. 34 Political agreement between PM of Kosovo and PM of Serbia on April 19 th 2013, foresees dismantling Serbian parallel structures in northern municipalities of Kosovo and inclusion of some of their personnel into Kosovo government system. Though there is no clear cost estimation for this, it is foreseen that it will cost about €13 million in wages and salaries – which is included in the current allocation of €45 million. 35 About 67,000 people applied for war veteran status. The verification process has been slow so far and is expected to finish by June 2014. From the number of verified applications about 35 percent are not considered legitimate. A simple calculation puts the number of legitimate veterans at about 47,000, and depending on the level of benefits a total cost for the scheme will be calculated. 36 This new highway (Route 6) will connect Prishtina with Macedonia at the Hani i Elezit Border Crossing Point. It was approved at an inter-ministerial Steering Committee on January 10th, 2014, and the contract was awarded to the US-Turkish consortium Bechtel-ENKA, which also built the Route 7 Highway connecting Prishtina and Tirana, Albania. 37 €600 million for the project plus an estimated €150 million for expropriations. 5 deplete resources needed to invest in energy generation capacities, should sufficient private funds not become available. 1.18. Kosovo has an opportunity to re-orientate spending toward areas that address its real challenges. Although there are genuine needs in many areas, limited resources mean that funds should be re-orientated toward areas that would unlock economic growth potential and help reduce joblessness and poverty. D. OBJECTIVES AND SCOPE OF THE PUBLIC FINANCE REVIEW 1.19. The objectives of this Public Finance Review (PFR) are: (i) to identify the main spending needs in Kosovo for boosting economic growth, protecting the poor and vulnerable, and finding ways to benefit from the country’s youthful population, and to compare these needs with the current spending pressures; (ii) to identify options to accommodate these objectives through budget reallocations, and to find ways in which the efficiency of public spending could be improved to achieve better value-for-money; and (iii) to identify measures that help adapt to declining border revenues by increasing domestic revenue collections in ways that promote economic growth and labor market formality. To achieve these objectives the PFR is organized in four thematic parts. 1.20. The first part of the report presents the macroeconomic context and the past fiscal performance. It provides a brief analysis of fiscal policies, revenue and expenditure performance between 2008 and 2013. The revenue section reviews tax revenue performance, while the expenditure section evaluates expenditure trends by economic and functional categories, and provides an overview of: (i) "hard" investments in energy and transport – the areas of priority for public infrastructure investment; and (ii) "soft" investments in education, health and social protection. 1.21. The second part analyzes in detail current spending efficiency and the upcoming spending needs and options in: (i) energy, (ii) transport, (iii) education, (iv) health, and (v) social Protection. Each of these chapters presents the past sectoral performance and the medium term fiscal needs, and evaluates the ongoing and future reforms and their fiscal implications. 1.22. The third part reviews revenue performance. Keeping in mind that the public sector will suffer from declining border revenues over the short and medium term, the analysis identifies opportunities for further domestic revenue mobilization through tax policy adjustments and strengthening of tax administration. 1.23. The final part provides a brief overview of how public finances could evolve under different scenarios. In particular, it provides a “business as usual” scenario, and a scenario tha t shows how public finances could be adapted, even within a limited resource envelope, to better address Kosovo’s challenges. 6 Box 1.1. Analytical Tools Used in the PFR and Beyond the PFR The work undertaken under this PFR has sought not only to analyze public revenues and expenditures but also to provide sectoral insights and tools for further continued analysis. Examples include:  The Kosovo BOOST uses Treasury data in an Excel spreadsheet to easily obtain highly disaggregated spending data between 2007 and 2012 from an easy-to-use pivot table.  The Kosovo Energy Sector Impact Tool (KESIT) shows the impact on consumer tariffs, subsidies and public debt of different investment choices and demand scenarios in the energy sector. The fiscal and tariff impacts of main energy scenarios are estimated using this model. The KESIT can be used to provide information on the likely outcomes of different options for energy sector investment. In particular, the tariff impact of reforms can help to design social protection responses to higher energy prices.  An Education Database integrates four different, previously uncombined databases on pupils, teachers, schools and exam results enables the Ministry of Education staff to understand better the inputs and outputs of the education system. This database contributes to analysis of the efficiency of the inputs in education.  Population data in Kosovo does not cover well the Serb minority. Work under the PFR used other sources of data to obtain a better approximation of population by municipality. See Annex 1 for details.  A Model of the Impact of Health Insurance Law shows the immediate and medium-term fiscal impacts of the introduction of health insurance under different scenarios, such as the premium rates, uptake rates and exemptions. The impact of the HIL is estimated using this model.  A Tax Gap Analysis was conducted using a wide range of data to estimate the difference between revenue collections and revenue potential. The results of this study are presented in the Revenue Chapter.  A survey on the Perceptions of Tax Law Implementation (PoTI) among large and medium-sized firms makes an effort to show the cost of corruption for the tax system. The results of this study are presented in the Revenue Chapter. 7 CHAPTER 2. ECONOMIC CONTEXT AND THE FISCAL CHALLENGE38 A. KEY MESSAGES  Kosovo’s headline fiscal and macroeconomic indicators are relatively sound. Kosovo continued to grow, even during the economic crisis, public debt is low, inflation is held in check, and banking sector indicators are sound. Thanks to fast-growing fiscal receipts and previously accumulated reserves, when the crisis hit, Kosovo was able to adopt a counter-cyclical fiscal policy with investments in infrastructure projects, helping to support economic growth.  For Kosovo, forging an economic growth path that sustainably creates more and higher quality jobs for its youthful population and reduces poverty is a central development challenge. Fiscal policy is the main macroeconomic instrument to do that as the country has unilaterally adopted the euro as its currency. Consequently, fiscal policy plays a crucial role for macroeconomic stability and economic growth as well as in addressing the challenges of unemployment, poverty and a young population.  However, the current allocation of public resources combined with inefficient spending will make it difficult to address the key challenges faced by Kosovo and boost economic growth. There is evidence that allocation of resources within sectors could be adjusted to achieve better outcomes (in terms of the provision of public services, growth and poverty reduction), notably in health, education and transport. In addition to intra-sector reallocation, inter-sector reallocations could help Kosovo meet some of its key challenges, notably in the energy sector. In addition to the reallocation of resources, there is some evidence that corruption in both revenue and public procurement harms the public sector. B. ECONOMIC CONTEXT 2.1. The macroeconomic basics of Kosovo are relatively good with solid positive growth driven by investments. Unilateral adoption of euro provided for some price stability and the financial sector remained sound during the crisis. This macroeconomic stability should be maintained as a good starting point in drafting policies that can bring higher growth. The Macroeconomic Basics are Right 2.2. Kosovo’s economic growth has averaged over 4 percent per annum since the end of the war in 1999 and remained positive throughout the global economic crisis. Economic growth slowed from its peak of 7.2 percent39 in 2008 but remained positive during the global financial crisis, shielded by the country’s limited integration with the global economy and the Government of Kosovo’s (GoK) ability to run a counter-cyclical fiscal policy by boosting public investments during the crisis. See Table A.21 in Annex 4 for macroeconomic indicators. 2.3. Investments, both public and private, have played a crucial role in boosting economic growth since 2008, contributing positively even in 2012 and 2013. Public investments doubled in 2008, providing a significant boost to growth in that year, and off-setting a real contraction of private 38 Chapter prepared by Agim Demukaj. 39 GDP by expenditure approach 2004-2012, Statistics Agency of Kosovo. 9 investment. Public investment continued to grow over the following years, albeit at a slower rate than in 2008. At the same time, private investment accelerated after 2009, contributing more than 2 percentage points to growth in 2010 and 2011 and slightly less in 2012. While public investments in roads crowded out other possible public investments, private investments were, to some extent, “crowded in” by public ones as private firms established to supply goods and services to the transport construction sector. 2.4. Net exports contributed negatively to growth. Total imports were in the range of 52-57 percent of GDP while total exports fluctuated between 16 and 21 percent of GDP during 2008-2012. Exports of minerals more than doubled between 2008 and 2011 and exports of other goods increased somewhat, as the domestic production base grew slowly. But the euro crisis dampened the external demand and exports fell in 2012. Exports of services, which were highly related to spending by the diaspora visiting Kosovo, remained robust despite the poor external economic climate. At the same time, imports of investment goods rose and imports of consumables stagnated. 2.5. Overall, while Kosovo’s growth rates have been fairly stable, the main drivers of growth have shifted over the past few years (Figure 2.1). Growth between 2008 and 2011 was driven by high levels of public investment and booming domestic consumption, the latter facilitated by credit growth and stable inflows of remittances. GDP growth has slowed significantly thereafter due to external (Euro zone crisis) and domestic reasons (slowdown in lending and domestic demand). Consumption has also slowed, turning from a positive to a negative contributor to economic growth in 2012. 2.6. Economic growth has benefited all sections of society but the poorest have benefitted most. Household Budget Surveys (HBS) reveals that economic growth in Kosovo has been beneficial for both poverty and shared prosperity.40 Between 2006 and 2011, the poorest 40 percent of the population saw average annual consumption increase by close to 4 percent (Figure 2.2). Growth rates were slightly higher for the 30 percent of the population defined as poor in 2011. Given the close correlation between poverty and shared prosperity, the rest of this report will, where warranted, discuss poverty in general. Figure 2.1. Contribution to Growth 2008-2012 Figure 2.2. Growth Incidence Curve, 2006-2011 15% 10% 5% 0% -5% -10% 2008 2009 2010 2011 2012 Consumption Investments Net Exports Real GDP growth Source: MoF and World Bank staff calculations. Source: World Bank, Regular Economic Report, Dec. 2013. 2.7. Thanks partially to the unilateral adoption of the euro, inflation has remained in check (Table 2.1). After the 1998-99 war, Kosovo adopted the Deutsche Mark as its currency, and in 2002 40 Defined as the share of income going to the poorest 40 percent of the population. 10 converted to the euro, thereby limiting its ability to utilize monetary policy instruments. The adoption of the euro has played a crucial role in keeping inflation under control. But, as a small economy Kosovo is a price taker and therefore subject to international price fluctuations of its main trading commodities – primarily food and energy. The Consumer Price Index (CPI) tends to reflect global prices, which shift on factors related to weather, conflicts and global demand. It therefore moved higher in the second half of 2010 and the first half 2011 and again in 2013 in response to drought-related food price hikes. In 2010 and 2012 energy prices increases similarly impacted the CPI. 2.8. The banking sector has proven resilient to the deterioration of the external environment, but credit growth rates have declined. The banking sector performed well during the crisis period thanks largely to its conservative outlook and risk-averse lending decisions before, during, and after the crisis. The largely foreign-owned banking system has remained well-capitalized, liquid, and profitable throughout this period. However, growth rates for both deposits and lending have remained positive but slowed. Deposits grew by 8 percent and credit by 3.8 percent by end 2012 on an annual basis, down from 10.1 and 16.5 percent respectively in 2011 (Figure 2.3). Nonperforming loans increased to 8.7 percent in 2013 but remained low compared to the region (Figure 2.4). The weaker financial performance has impacted negatively economic growth in 2012 and 2013. Figure 2.3. Loans and Deposits, 2008–12 Figure 2.4. Indicators of Financial Health Sources: Central Bank of the Republic of Kosovo; and Sources: Central Bank of the Republic of Kosovo; and World Bank staff estimates. World Bank staff estimates. 2.9. The Central Bank of the Republic of Kosovo (CBK) and Ministry of Finance (MoF) have focused on preserving and strengthening banking-sector stability. Reforms designed to achieve this included: (i) the establishment of an Emergency Liquidity Fund (ELF) in 2012; (ii) the establishment of a Deposit Insurance Fund; and (iii) a new law on banking, microfinance, and the establishment of non-bank financial institutions. The ELF was created during the Euro-zone crisis to provide needed liquidity to the financial sector in Kosovo. It consisted of €92 million, of which half were provided by the GoK and half by the CBK. The ELF has not been utilized but the funds remain available if needed. 11 Table 2.1. Key Macroeconomic Indicators 2008-2015 2008 2009 2010 2011 2012 2013 2014 2015 Estimate Projections Real growth rates GDP real growth 6.9 2.9 3.9 4.5 2.7 3.0 3.5 3.5 CPI, annual average 9.4 -2.4 3.5 7.3 2.5 1.9 1.0 1.0 Share of GDP Revenue (including budget grants) 23.9 28.3 26.7 27.2 26.7 26.5 26.8 26.3 Expenditure (and net lending) 24.1 29.0 29.3 29.1 29.2 29.5 28.8 28.3 Primary deficit -0.2 -0.7 -2.4 -1.7 -2.3 -2.6 -1.7 -1.7 Overall deficit -0.2 -0.7 -2.6 -1.9 -2.6 -3.0 -2.0 -2.0 Gov. bank balances (usable) 10.8 8.7 5.8 3.5 4.3 3.2 2.8 2.6 Public Debt 0.0 6.1 6.4 6.0 8.7 9.7 11.4 13.0 Investments 28.6 32.3 33.9 33.2 32.6 33.2 32.3 32.1 FDI 8.7 6.8 7.6 7.9 4.3 6.0 6.5 6.5 Current Account Balance -11.7 -9.2 -11.9 -13.8 -7.6 -10.9 -9.5 -8.2 GDP (€, Mil) 3,940 4,046 4,334 4,770 5,017 5,213 5,442 5,644 GDP per capita (€) 2,255 2,297 2,441 2,665 2,779 2,858 2,955 3,034 Population (Mil.) 1.75 1.76 1.78 1.79 1.81 1.82 1.84 1.86 Source: MoF, Statistics Agency of Kosovo and WB staff calculations. C. FISCAL PERFORMANCE 2008 – 2013 2.10. This PFR focuses on analyzing fiscal performance between 2008 and 2013, including both revenue and expenditure sides. The headline fiscal numbers such as deficit and debt levels are sound and show some degree of stability. However revenue collection has started to stagnate lately and expenditure requires efficiency improvements to justify further increases. This section will discuss general fiscal numbers, followed by revenue performance and expenditure efficiency. The Headline Fiscal Numbers are Sound 2.11. Because of the use of Euro as the currency, fiscal policy is the main tool Figure 2.5. Sources of Deficit Financing 2009 2010 2011 for macroeconomic stability and economic growth in Kosovo. Fiscal policy has been lately relatively well managed. Kosovo ran successive budget surpluses between 2000 and 2008, accumulating a cash buffer of over €420 2012 2013 Drawings, incl. IMF million (4.8 percent of GDP). This allowed Carry-over OSR One-off Financing-Privatization fiscal policy to turn expansionary in 2008, T-Bills when economic growth was at risk due to Use of Previously Accumulated Cash Balance the effects of the global financial crisis. The previously accumulated cash buffer Source: Macroeconomic Department MoF. financed fiscal deficits between 2009 and 2011, debt financed the fiscal deficit in 2012 and debt and the cash buffer in 2013 (Figure 2.5). A recently passed Fiscal Rule has reinforced the intentions of the GoK to maintain sustainable deficits and debt levels (see Box 2.1). 12 2.12. Kosovo’s public debt stock is low but increasing slowly. Kosovo has the lowest public debt and among the lowest fiscal deficit in SEE. Total public debt was estimated at 9.7 percent of GDP at end- 2013. Kosovo inherited a debt of former SFRJ to IBRD, which consisted of three loans, one of which was repaid in September 2009 (partially financed by donor money). The government issued €74 million domestic debt in 2012 and another €80 million in 2013. However, the government is planning to increase its domestic debt in the coming years, and roll-over the existing short term domestic debt by refinancing it with a longer maturity domestic debt. A debt sustainability analysis, which includes potential debt owed to the Paris and London Clubs (they become due once Serbia recognizes Kosovo’s independence, at which point it would add another 9 percentage points of GDP to the stock of public debt), suggests that public debt indicators are sustainable. Box 2.1. Public Financial Management and the Fiscal Rule To safeguard fiscal sustainability, Kosovo has adopted tight constitutional and legal frameworks to prevent unsustainable fiscal deficits and debt policies. Under the constitution, all new external debt needs to be ratified by a two-thirds majority in Parliament. In addition the Law on Public Financial Management and Accountability (LPFMA) requires that supplementary budgets have to be deficit neutral. In 2013 Kosovo adopted a new fiscal rule which was incorporated into its main fiscal law. The Fiscal Rule established policy restraints on the fiscal deficit, public debt level, general expenses and other indicators to control macro-fiscal sustainability. It sets the overall fiscal deficit at 2 percent of GDP, reducing further debt risks. The key points of the Fiscal Rule in supplementing/amending the LPFM were as follows: a) Maintain the overall fiscal deficit at 2 percent of projected GDP in the current budget year. This deficit level applies to the consolidated central and municipal budget; b) The unspent part of appropriations from the previous year can be used as a source of financing new budget appropriations in capital expenditure, and will not be considered for purposes of compliance with the fiscal deficit ceiling; c) Other revenue receipts additional to budgeted revenues from the previous year, may be transferred to the following year for new budget appropriations in the category of capital expenditure, and will not be considered for purposes of compliance with the fiscal deficit ceiling; d) Privatization proceeds and their spending will not be subject to the fiscal deficit ceiling. These revenues can be used only for capital investment projects, if the level of usable government bank balance amounts to at least 4.5 percent of projected GDP; e) Any excess of the projected fiscal deficit of more than 0.5 percent of GDP should be compensated within the next three fiscal years. The Fiscal Rule also specifies situations when the rule may be temporarily suspended, such as: a) Economic recession, when nominal tax revenues are equal to or lower than the tax revenues collected during the same period of prior fiscal year, excluding the impact of policies and one-off tax revenues; or b) An emergency such as a natural disaster; or c) In the event of a crisis in the banking system identified by both the Minister of Finance and the Governor of the CBK, on the proposal of the CBK; or d) A call on a state guarantee that will have to be paid by the government and that has an impact on overall expenditures over 1.5 percent of projected GDP. Source: World Bank staff and MTEF, Ministry of Finance of the Republic of Kosovo, April 2013. 13 Revenue Growth was Strong, but Has Slowed 2.13. Revenue growth of about 40 percent between 2008 and 2012 and earlier accumulated bank balances of 10.5 percent of GDP enabled expenditure growth during the crisis. However, the recent revenue stagnation and the decline of the cash buffer to 3.2 percent of GDP at end-2013 have introduced a budget constraint. The failure to privatize 75 percent of Post Telecommunication of Kosovo (PTK) in 2013 has produced a situation where stricter expenditure prioritization will be needed and efficiency of spending improved. 2.14. Revenue performance was strong until 2011 but has slowed thereafter (Figure 2.6). Total revenues increased by 44 percent between 2008 and 2012, with tax revenue growth at 27 percent. Two thirds of revenues came from indirect taxes. Due to continued economic growth, revenue performance in Kosovo was strong even during the crisis41 and marked a record growth of 18 percent in 2011. Taxes on goods and services grew by 33.6 percent over the period, of which VAT growth was most rapid (53.2 percent), followed by excises (20 percent). However, overall revenue growth began to stagnate in 2013, with border revenues declining by 1 percent. Figure 2.6. Revenue Performance 2004-2013 Total revenues Tax revenues Direct taxes Indirect taxes Non-tax revenues 1500 1000 500 0 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Source: Treasury Department, Ministry of Finance. 2.15. Despite the growth in tax revenues in the last five years, Kosovo has one of the lowest overall tax-to-GDP ratios in SEE and relies heavily on border revenue. In the SEE, only Albania collects less tax revenues than Kosovo as a percent of GDP (Figure 2.7). Being among the poorest countries in SEE with high unemployment and poverty rates, makes it critical to choose the right revenue mix for poverty alleviation and boosting employment. Therefore, low and simple taxation focused on indirect taxes is considered appropriate. Due to reliance on imports, most taxes (about 71 percent in 2012, Figure 2.8) are collected at the border. This is comprised of the value added tax (VAT)42, which accounted for 48 percent of total taxes in 2012, followed by excises with a 26 percent share. Other SEE countries collect on average about 10 percent of GDP as social security contributions. Kosovo has a pillar two pension system, and 10 percent from gross salary is collected at Kosovo Pension Savings Trust (KPST) of which 5 percent is paid from employer and 5 percent from employee. KPST is a private trust fund and is not included in the budget revenues. 41 Revenue grew by 22 percent in 2009 including dividends or 0.5 percent without dividends, and 2.7 percent in 2012. 42 VAT is usually not a border tax because it is meant to be collected from the end consumer, but in Kosovo about 81 percent of VAT is collected at the border, or VAT is 50 percent of border revenues. Only about 7.7 percent of VAT collected at the border is refunded internally. 14 Figure 2.7. Total Tax Revenue in SEE, % GDP Figure 2.8. Collection of Taxes By Place, % Total Tax Revenue Tax revenue collected internally 29% Tax revenue collected at border 71% Source: World Bank staff estimates. Source: Ministry of Finance. 2.16. Property taxes do not play an important role in Kosovo. Property taxes provide about 0.6 percent of GDP on average in SEE, about twice the share that Kosovo collects from this source. A new property tax on agricultural land is being drafted and may be implemented during 2015. That would enable an increase in property taxes revenues. 2.17. Non-tax revenues and municipal own source revenues (OSR) have performed quite well if dividends from publicly owned enterprise (POE) are included, but they grew only 5.5 percent between 2008 and 2012 excluding dividends. Growth in main categories of non-tax revenues has not been equal. Non-tax central level revenues grew by 67 percent between 2008 and 2012, while own source43 central revenues declined by 38 percent during the same period. Municipal own source revenues increased by 31 percent, of which property tax collection increased by 24 percent. A dividend from the POE Post and Telecommunications Company of Kosovo (PTK) (introduced in 2009) played an important role in boosting revenues (Table 2.2). 2.18. Customs tariffs revenue and other revenue collected at the border have shown signs of slowing during 2012 and decline in 2013. The revenue decline in 2013 was partially driven by price effects and import substitution effects. A Stabilization Association Agreement (SAA) with the EU and a trade agreement with Turkey mean that border revenues will decline further in the medium term (see Chapter 1). When other SEE countries signed SAAs, border revenues declined steeply already in anticipation of the agreement and further during its implementation. For example, in both Albania and Bosnia and Herzegovina, which signed SAAs in 2009 and 2008, respectively, tariff revenue fell from 6 percent of total revenue in 2006 to 2 percent in 2012. In Serbia, tariff revenue fell from 5.2 percent of total revenues in 2006 to 2.6 percent in 2012 in anticipation of the SAA, which was signed in 2013. 43 Own source central level revenues include revenues from: fees and charges for different services offered by central level government, licenses, fines and penalties, inspection, rent, etc. 15 Table 2.2. Revenue Performance 2008 – 2013 2008 2009 2010 2011 2012 2013 Preliminary Total revenue (including grants) 23.9 28.3 26.6 27.2 26.4 25.8 Tax Revenue 20.4 19.3 20.5 22.2 21.6 21.2 Value Added Tax (VAT) 8.7 9.0 9.9 10.8 10.3 10.0 Customs tariffs 2.4 2.3 2.4 2.5 2.3 2.3 Excise 5.2 4.9 5.3 6.0 6.0 5.8 Profit tax 3.0 2.2 1.9 1.7 1.8 1.9 Personal income tax (PIT) 1.1 0.9 1.0 1.2 1.2 1.3 Non-tax revenue 3.5 8.2 5.4 4.6 4.1 2.7 Non-tax central level revenue 2.4 7.0 4.2 3.5 3.2 1.6 o/w Dividend from PTK 0 4.9 2 1.3 0.9 0.8 Municipal Own source revenues 1.1 1.2 1.2 1.2 1.2 1.1 Grants for budget support 0.0 0.8 0.7 0.4 0.7 0.0 Source: Treasury department, MoF. Expenditure Efficiency Could be improved 2.19. The share of capital investments in total expenditure is very large at 40 percent (Table 2.3), but prioritization towards pro-growth sectors and expenditure efficiency should improve. Kosovo spends more in capital investments than other SEE countries. While this may be needed to create missing basic infrastructure, the investment does not always address well the main challenges facing Kosovo. Therefore allocation among sectors and economic categories could be better prioritized and efficiency improvements within sectors are needed. Table 2.3. Expenditure by Economic Category 2008-2013, % GDP 2008 2009 2010 2011 2012 2013 Preliminary Total Expenditure 24.1 29.0 29.3 29.1 29.2 28.8 Current Expenditure 15.0 17.1 17.2 17.2 17.5 18.1 Wages 5.8 6.5 7.2 8.1 8.1 8.0 Goods and Services 4.0 4.2 4.2 3.7 3.8 4.1 Subsidies 5.3 6.3 5.8 5.4 5.6 6.0 Capital Expenditure 8.8 9.9 10.5 11.1 11.0 10.2 Other expenditure (net lending) 0.2 2.0 1.6 0.9 0.7 0.5 Source: Treasury Department, Ministry of Finance. 2.20. Expenditure execution improved between 2008 and 2012, notably in capital investments. Budget execution increased from 80 percent in 2008 to a peak of 93 percent in 2010 but has remained around this level since then. Execution of capital projects noted major improvement from a low 68 percent in 2008 to 93 percent in 2010. That was due mainly to a single large transport investment (Route 7). Also, better institutional capacities helped, including improvements in procurement law that shortened tender procedures and provided higher transparency therefore reducing time consuming complaints of tender participants. However, fiscal cuts used as fiscal adjustment measures designed to preserve fiscal deficits44, have lowered budget executions in 2012 and 2013. While wages were executed at 98-99 44 Kosovo was under an IMF program since 2010, which put strict benchmarks – the key one being fiscal deficit – in place. Therefore depending on the performance of revenues (which were lower than planned during 2012 and 2013) GoK cut expenditures to maintain agreed fiscal deficits. This resulted in lower expenditures compared to original 16 percent, the execution of goods and services budget was always low – between 85 percent in 2008 and 91 percent in 2011. 2.21. Kosovo spends a higher share on capital investment than any other SEE country. About 40 percent of its budget is spent on capital investment, which is double the share in the second ranked country – Albania. In contrast, the share of social transfers in total expenditures is only one third to one half of what other countries spend (Figure 2.9). Capital investments may support economic growth. But approximately 40 percent of investments went to a single project, the highway connecting with the Albanian border (Route 7), which squeezed other capital investments during 2010-13.45 This was followed by subsidies to POEs46, largely in the energy sector (see Energy Chapter), and Socially Owned Enterprises (SOE).47 A new highway to Macedonia (Route 6) was introduced in the 2014 budget at a total cost of €750 million or 14.4 percent of GDP and is likely to crowd out other investments. 48 Further, wages of public employees have continued increasing, with a substantial increase (in response to political pressure) granted during the electoral campaign in 2010-11 and were repeated again during 2014 campaign.49 Pensions have also increased significantly, but social assistance has not. 2.22. Spending on “economic affairs” which includes transport infrastructure is double the SEE average (Figure 2.10). During 2008-2010 a large proportion of transport investments were directed in rehabilitation and expansion of main roads of Prishtina-Peja and Prishtina-Mitrovica that enable internal road connection of Kosovo. During the same period there has been some expansion in construction of local roads, financed by the municipal budgets and supported by grants from Ministry of Transport. This resulted in expansion of road network paved with asphalt all over Kosovo. However, the biggest share of transport budget since 2010 was taken by Route 7, which brought other road construction projects to a halt. By 2012 the road rehabilitation budget was just a quarter of its 2009 (pre-highway construction) levels. This stems from a road maintenance budget that is less than half of neighboring Serbia or Bosnia and Herzegovina on a per kilometer basis. The result is that alongside the new highways, existing roads sit in poor condition. Worse, the rates of economic return on highways are comparatively low and call into question the choice to prioritize this spending over investments in energy, education and health as well as maintenance of existing roads. The Transport Chapter of this report discusses these developments and potential ways forward to a greater detail. budgets and therefore lower execution rates. However if execution would be measured against finally agreed budgets, then they have continuously improved. 45 The Route 7 (Merdare – Morine highway) is a national priority aiming to improve regional and European integration. However, the efficiency of the project is questionable. Route 7 costs have halted investments in other capital investments in Kosovo, such as the rehabilitation and expansion of main roads connecting Prishtina with Peja and Prishtina with Mitrovica, two major towns. The road to Peja has the heaviest traffic as it connects two main regions of Kosovo. Its rehabilitation and expansion stopped in 2010 due to relocation of budget resources to Route 7. 46 These are referred to as Publically Owned Enterprises or POEs in Kosovo. This was mainly subsidies to support investments in Kosovo Electricity Corporation (KEK). 47 This refers to subsidies to support important Socially Owned Enterprises such as Trepça mines. 48 Although in 2014 the Route 6 project has been alloca ted only €102 million, which compares to €244 million allocated for Route 7 in2013, the project is expected to be completed in 4 years; therefore the annual budget will climb and even double starting from 2015, which will likely come at the cost of squeezing other capital investment projects. 49 The wage bill increased by 27 percent in 2011 driven by electoral promises. The wage increase was granted to all public employees, with some differentiation between sectors. This triggered termination of the IMF SBA program in spring 2011. A “wage freeze” was applied in the following years. However, that policy created cyclicality and mounting pressures resulted in new wage increase by 25 percent in the next electoral cycle in 2014. 17 Figure 2.9. Structure Expenditures in Southeast Figure 2.10. Selected Expenditures by Europe Functional Classification in 2012, % GDP 100% 9.4 12.7 13 9.7 17.1 80% 39.5 Economic affairs 33 25 35 32 36 60% Education 23 40% 31.5 42.7 29.3 30.3 29.3 13 Social protection 20% Health 0% ALB BIH KOS MKD MNE SRB 0.0 5.0 10.0 15.0 Wages and salaries Social transfers Other current spending Capital spending Small CIS Other SEE* EU11 Kosovo Note: ALB=Albania, BIH=Bosnia and Herzegovina, Notes: ‘Small CIS’ grouping includes Armenia, Georgia, and KOS=Kosovo, MKD=FYR Macedonia, MNE=Montenegro, Moldova. ‘Other SEE’ excludes BiH and Montenegro due to and SRB=Serbia. Source: World Bank. data unavailability. 2.23. The energy sector is currently getting significant public sector financial support provided on an ad hoc basis and, while this is inefficient, significant new public investment in energy may be warranted if the private sector is not willing or able to invest in new power generation capacity. Large subsidies from the GoK and donors, focused on capital investments for rehabilitation and maintenance of existing generation and distribution capacities, have helped keep energy prices low. In addition, the GoK has provided a special energy grant for households on social assistance and subsidized energy imports50 for everyone. Kosovo is also facing an energy shortage and it needs to build a new power plant to respond to the demand while closing Kosovo A, an old and polluting plant. The GoK needs to consider other options of financing to avoid significant blackouts once Kosovo A is closed. A more detailed analysis of energy sector, with past developments and current situation, as well as potential forward looking scenarios and their outcomes are discussed in the Energy Chapter of this PFR. 2.24. Education expenditure, as a percent of GDP51, is slightly higher than in the rest of SEE, but lower than in EU-1152 and Commonwealth of Independent States (CIS) countries. But despite the fact that education expenditure increased by half53 between 2008 and 2012, spending per pupil is lower than in the EU-11, CIS or SEE. Kosovo spends 14 percent of its total public expenditure on education, with the largest share of education spending going to wages and salaries. While high wage spending is justified due to staff-centric nature of teaching inputs, large politically-motivated wage increases have prevented merit-based wage increases54 from functioning effectively. Another important area of education spending was investment in school premises designed to eliminate “three shift” schools, which was successful by 2010. While there remain “two shift” schools, these can largely be eliminated through demographic changes and so should not be seen as the main priority. Instead, there are more pressing priorities in education. In particular, efforts could be made to expand pre-school education – an 50 Large energy import subsidies were introduced in Kosovo budget since 2008 (€23 million), increased in 2009 and 2010 (€42 and €40 million respectively), slightly declined in 2011 (€27 million) and declined in the last two years (about €12 million). 51 In 2011 education related expenditure reached 4 percent of GDP. 52 EU-11 consists of Poland, Slovenia, Czech Republic, Slovakia, Hungary, Lithuania, Latvia, Estonia, Bulgaria, Romania, and Croatia. 53 Education expenditure increased from €130 million in 2008 to €197 million in 2012. 54 Education sector in Kosovo is undergoing large reforms and these large wage increases to everyone undermine the reforms. 18 investment with high individual and social returns, especially for children from poorer households, and resources could be allocated toward rehabilitating and maintaining old schools. Unfortunately, there are again pressures to offer a politically-motivated wage increase55 which would further undermine merit- based wage increases and divert resources from more urgent needs in education. The Education Chapter elaborates further. 2.25. Spending on healthcare – a critical expenditure necessary for sustainable development – is lower, as a percent of GDP56, than in other countries in SEE as well as in the small CIS countries.57 Public health expenditures account for only about 60 percent of total health expenditures and the share of out-of-pocket payments is relatively high. There is evidence that the poor significantly under-consume medication. Public health expenditures58 have more than doubled59 in real terms over the past five years, but Household Budget Survey (HBS) data suggest that health expenditures are still impoverishing a number of households in Kosovo. A new public health insurance scheme appears to be popular and could help to raise additional resources for this under-funded sector, while compensating some of the border revenue shortfall (discussed in the Revenue Chapter). The success of the scheme will depend crucially on whether the new resources are used and are seen to be used wisely. In addition, it will be important to protect the poorest groups from the resulting increased expenses. Issues of health spending are elaborated further in the Health Chapter of this report. 2.26. Social protection spending is marginal compared with other European countries, which partly explains the higher incidence of poverty. The basic pension represents the largest component of social protection spending in Kosovo. Over 115,000 individuals benefited from the basic (pillar 1) pension in 2013. They receive an average of €60 per month, financed directly from the state budget. Another 35,830 beneficiaries who could prove over 15 years of contributions to the old Yugoslav system receive on top of basic pension also a monthly supplement of €52. This supplement has seen the fastest spending growth of any area of social protection since 2008. There is growing demand to expand coverage of this despite the fact that the basic pension lifts pensioners to the poverty threshold60. War- related benefits have also grown fast and special interest groups continue to press for their expansion. Social assistance, which tackles the extreme poor, however, has seen a decline in spending in real terms between 2008 and 2012, and the number of beneficiaries has declined by 21 percent despite the continued high rates of poverty. The Social Protection chapter of this report expands further on these issues. Detailed functional spending data are provided in Table 2.4 below. 55 Prime Minister of Kosovo in September 2013 promised to education workers that another significant wage increase will be granted to them soon without specifying the timing. This promises increase expectations which lead to budget pressures. In April 2014 wages were increased by 25 percent across the board. 56 Health expenditure was 2.6 percent of GDP in 2012. 57 Armenia, Georgia, and Moldova. 58 Details of donors’ off-budget support are largely undocumented. The in-depth analysis in this chapter therefore takes domestically-financed spending as public health sector expenditure. 59 Public health expenditure increased from €68.2 million in 2007 to €128.4 million in 2012. 60 In May 2014 the parliament of Kosovo decided to recognize pension contributions of former employees in the parallel system during the 90s and to include them in the former contributor pension scheme if they have more than 15 years of joint contributions. That will certainly increase the number of beneficiaries and the overall costs. 19 Table 2.4. Expenditure by Functional Category 2008-2013, % GDP 2008 2009 2010 2011 2012 2013 Actual Preliminary Economic affairs 7.0 7.2 7.6 8.8 8.6 7.4 Education 3.3 3.6 3.7 4.0 3.9 3.9 Health 2.0 2.5 2.5 2.5 2.6 2.6 Social protection 3.6 3.9 3.7 3.6 3.9 4.2 General services 4.7 5.9 6.1 5.5 5.3 6.0 Public order and safety 2.0 2.2 2.3 2.2 2.3 2.2 Defence 0.5 0.5 0.6 0.6 0.6 0.6 Other sectors 0.8 1.1 1.1 1.0 1.2 1.3 Source: Treasury Department, MoF. This table includes primary expenditure and does not include on- lending or debt repayment. Therefore, if you add these categories it does not provide the overall expenditure as in other tables. 2.27. Prioritization among sectors, given the budget constraint, has been a challenge for Kosovo in the past and is becoming even more relevant in the near future. Most resources were allocated to a few favored projects, which did not always offer the highest value for money. Going forward Kosovo needs to better address spending needs and prepare viable options on how to finance them. The recent stagnation of overall revenue growth, driven by declining border revenues, and a failure to generate one- off revenues from privatization of PTK have created a tight budget constraint and the need for better prioritization of budget spending. A detailed sectoral analysis focused on energy, transport, education, health and social protection, will discuss further the current situation, the needs and options for a better allocation of budget financing among sectors that will be needed to sustain economic growth, and prosperity. 2.28. The development outcomes in each functional category depend not only on the amount but also on the efficiency and effectiveness of spending. Unfortunately, available data point to weak spending efficiency across all major spending categories, and in particular in sectors such as health, education, and transport. This report therefore analyzes the efficiency of spending within each sector. A detailed expenditure efficiency analysis on health, education and transport sector is presented in the sector specific chapters that follow. 20 CHAPTER 3. ENERGY61 A. KEY MESSAGES  Kosovo’s fragile electricity system cannot meet the current demand, jeopardizing future economic growth. Although power generation has increased in recent years, uninterrupted economic growth has led to a surge in power demand, and Kosovo’s aging plants are unable to produce enough energy. This has resulted in costly blackouts: a recent USAID study 62 calculated that Kosovo’s unreliable power supply was costing businesses approximately €260 million (5 percent of GDP) per year. Moreover, the 2009 Enterprise Survey identified lack of reliable electricity as the No.1 obstacle to doing business in Kosovo. Small and medium enterprises were found to be the most affected. Kosovo A, an aging power plant producing nearly a third of the country’s energy, is slated for closure and, without new power generation facilities, the gap between power supply and demand (the “power gap”) will grow still further.  Closing the power gap requires a comprehensive energy strategy, including implementing energy efficiency measures (loss reduction and energy efficiency in buildings), supporting the development of renewable energy and building a new power station to meet sizable base-load demand. The cost of construction is estimated to exceed 20 percent of GDP, according to the 2011 World Bank Options Study63. Given these costs and the need for state-of- the-art technology, private sector participation in the financing and management of the facility is desirable.  Measures to close the power gap include reducing distribution losses and investing in renewables and energy efficiency: o Renewable energy is responsible for just 3 percent of total power generation in Kosovo. Although feed-in tariffs are designed to attract investors, lengthy approval times for new renewable projects have meant that few applications had been approved. This needs to change for rapid growth of renewable sources of energy. o In 2012, total distribution losses amounted to 1,859 gigawatt hours (GWh), or about 37 percent of domestically generated electricity. While commercial losses (essentially electricity theft) account for close to half of these losses, the bulk of the inefficiencies in the system stem from political constraints to improved collection in the north of the country and a dilapidated distribution infrastructure. The Kosovo Electricity Distribution Company (KEDS), privatized in 2013, is committed to realizing significant rehabilitation investments and reducing total losses by 3 percentage points annually. o In 2011, the Ministry of Economic Development approved the National Energy Efficiency Action Plan64, which forecasts energy savings of 9 percent across the whole economy (including in the public sector) between 2010 and 2018, compared to the baseline increase in energy demand. 61 Chapter prepared by Rhedon Begoli (principal author) and Karuna Phillips. 62 The Effect of Unreliable Power Supply and Quality on Kosovar Businesses, USAID 2012. 63 World Bank (2011), “Development and evaluation of power supply options for Kosovo”. Report available from: http://siteresources.worldbank.org/INTENERGY2/Resources/Kosovo_generation_options_report_12312011.pdf 64 For the period 2010-2018 21  Investments in improving energy infrastructure are required to ensure the security of supply and avoid an otherwise inevitable energy crisis. Such a situation would be triggered by lack of dependable capacity. Given the need to recoup investment costs, the cost of supply would need to increase. New investments will increase cost of supply and protecting consumers from the full effects of these increased costs would be challenging from a fiscal perspective, although necessary for the poorest segment of the population. Subsidies have been reduced successfully in recent years, and reversing this progress is neither desirable nor fiscally sustainable. Tariffs are currently at the lowest level in Europe except for Serbia. Maintaining these tariffs in the face of rising costs would require unaffordable fiscal subsidies. The authorities should ensure cost recovery via a medium-term tariff trajectory with gradual adjustments over time. Tactical steps should be taken in a carefully designed and targeted manner to protect the poor and vulnerable groups through active social safety net policies. B. OVERVIEW OF KOSOVO’S POWER SECTOR: GENERATION, DISTRIBUTION AND LOSSES 3.1. Two aging coal-fired power plants dominate power generation. Approximately 97 percent of power generation in 2012 came from two aging coal-fired power plants, Kosovo A and Kosovo B. These power plants provided 5,218 GWh of generation in 2012 (Figure 3.1). Hydropower makes up most of the remaining 3 percent of generation, with one medium-sized hydropower plant (Ujman/Gazivoda) contributing about 65 GWh and a number of smaller installations totaling about 31 GWh. Recently, the country’s first wind turbines were installed but these are not in operation due to licensing issues. 3.2. Power generation has increased in recent years but has failed to meet the demand. Power generation has increased steadily from 4,527 GWh in 2008 to 5,314 GWh in 2012, an increase of 17 percent (Figure 3.2). This increase has come from technical improvements at the existing power plants. Thanks partly to the increase in generation, the gap between production and consumption have been kept at around 10 percent over most of the period. It fell to just 3 percent in 2012 due largely to slower demand arising from a milder winter than the previous year and slowed economic growth. Current assessments suggest that power generation has reached its limit given the state and age of the current facilities. Figure 3.1. Power Generation, 2012 Figure 3.2. Power Generation and Consumption, GWh, 2008-12 5800 5600 5400 5200 5000 4800 4600 4400 4200 4000 2008 2009 2010 2011 2012 Production Consumption Source: World Bank calculation based on Energy Source: Energy Regulatory Office, Annual Report 2012. Regulatory Office, Annual Report 2012. 22 3.3. The power sector is relatively simple and, consequently inflexible. Power from Kosovo A, Kosovo B, and the large hydropower plants feed into the transmission system, managed by KOSTT, the public sector transmissions operator (Figure 3.3). Between them, these sources contributed 5,283 GWh in 2012. Imports enter and exports leave the transmission system, adding net inflows of 154 GWh in 2012. Transmission system losses, at less than 2 percent of power received, are low. The largest industrial consumers are connected directly to the transmission system65, consuming 473 GWh between them in 2012. The generation company KEK and the lignite mine that supplies it consume an additional 118 GWh. Remaining power is received by the – now privately owned - distribution company, KEDS. In 2012, KEDS received 4,736 GWh from the transmission network and another 31 GWh from small hydro power plants connected directly to the distribution grid. Of this, 3,019 GWh reached consumers while about 1,750 GWh (37 percent) were lost as technical (physical losses of energy, potentially through unknown illegal connections) and commercial losses (largely non-payment of bills). 3.4. Distribution losses have been reduced in recent years but they remain high. Distribution losses fell from 43 percent of power entering the distribution network in 2008 to 37 percent in 2012 (Figure 3.4). In 2012 losses comprised about 20 percent commercial and 17 percent technical losses. Almost all of the improvement is due to improved collection efforts, with distribution losses falling from 26 to 20 percent during the period. Less than 1 percentage point of the reduction came from an improvement in technical losses. Figure 3.3. Power Flows, 2012 Figure 3.4. Distribution Losses as % Generation, 2008-12 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% 2008 2009 2010 2011 2012 Commercial Losses Technical Losses Source: World Bank calculation based on Energy Source: Energy Regulatory Office, Annual Report Regulatory Office, Annual Report 2012. 2012. C. PAYING THE PRICE FOR AN IMPERFECT POWER SECTOR 3.5. The cost of losses is high and paid by energy consumers or falls on the public purse. In total, power worth an estimated €57.6 million was lost in the transmission and distribution network in 2012, a bill that has to be paid by the taxpayers or compliant bill payers. Losses from the transmission network, at wholesale prices, were estimated at €4.4 million in 2012 (Figure 3.7). Distribution losses were calculated 65 These were Ferronikeli – an iron smelter – and the Trepça mine are supplied by 110 KWh separate lines. 23 to be worth a €53.2 million in 2012, of which €6.6 million were unbilled/unpaid supplies going to the north of Kosovo. 3.6. General electricity sector fragility and large distribution losses, together with the power gap, have resulted in many unplanned power outages, with large costs to business. Reliability of energy supply continues to be lower than the standards requested by the energy regulator. In 2012, Energy Not Supplied (ENS) to distribution customers was estimated at 68.5 GWh, while the average annual number of interruptions per customers was 59.5, of which 2.6 were planned (Figure 3.5). The average interruption duration of 114.5 hours, of which, 4.9 hours were planned. At some times of day, power produced tended to be lower than power demanded during 2012 (Figure 3.6). A recent USAID study showed that the annual total costs and losses of Kosovar businesses were about €260 million due to unreliable power supply, and small and medium enterprises were the most affected. Frequent power cuts are a major obstacle to day-to-day operations and a constraint both to investment in new equipment and business expansion, in turn affecting job and employment creation and investments. Figure 3.5. System Interruption Frequency Figure 3.6. Average Hourly Supply and Demand and Duration, 2012 of Power, MWh 2012 900 800 4.9 700 600 500 2.6 400 109.6 300 56.9 200 100 0 Frequency (SAIFI) Duration in hours (SAIDI) 1 3 5 7 9 11 13 15 17 19 21 23 Hour of day Unplanned Planned Consumption Production Source: Energy Regulatory Office, Annual Report Source: Energy Regulatory Office, Annual Report 2012. 2012. 3.7. The large gap between supply and demand creates the need to import from an energy- starved region at high prices. Despite the increase in production, increasing demand led to higher levels of imports between 2008 and 2011 (Figure 3.8). They fell in 2012, largely due to lower domestic demand. The cost of imports fluctuates widely with international prices and weather conditions. The latter affects both demand and supply in the region. During winter 2012, import prices rose to around €120 per MWh – over double the average domestic retail cost – as demand for energy for heating increased due to an especially cold winter and supply faltered due to lack of rainfall at hydropower plants in the region (for more on heating, see Box 3.1). 24 Figure 3.7. Estimated Transmission and Figure 3.8. Cost and Quantity of Imports, Distribution Losses, € Mil., 2012 2008-12 Northern Transmission Kosovo 4.4 6.6 Technical losses 20.7 Commercial losses 25.9 Source: World Bank calculation based on Energy Source: Energy Regulatory Office, Annual Report 2012. Regulatory Office, Annual Report 2012. Box 3.1. District Heating Heating is neither financially nor environmentally sustainable in Kosovo with District Heating (DH) playing a small role in the energy sector. The total installed capacity of 183.5 MW only produces 130 GWh/p.a. (thermal). Only 3 percent of Kosovo’s households are connected to DH. The main energy sources for both space and water heating in Kosovo are biomass (mainly firewood) and electricity, together accounting for over 80 percent of heating consumption. The high consumption of unmanaged and unregulated firewood results in forest degradation, giving rise to adverse environmental, economic and health impacts. Electricity generated from lignite is used inefficiently and there are wide seasonal variations in demand. These exacerbate power supply interruptions and create the need for electricity imports, especially during the heating (winter) season. Only two of the four district heating companies are functional – DH Termokos in Prishtina and DH Gjakova in Gjakova. Both DH companies face serious problems as the heat demand exceeds supply, and thermal losses are very high, accounting for as much as 18 percent at DH Termokos. In addition, the isolated DH systems are running on expensive heavy fuel, which combined with significant losses in the distribution network and low bill payment, make these public companies strongly dependent on municipal and central government subsidies. Germany’s Kreditanstalt für Wiederafbau (KfW) has started a cogeneration project to support DH Termokos obtain heat supply from the steam released by Kosovo B turbines. The project should provide a more reliable and affordable supply of thermal energy for DH Termokos, therefore improving the supply of space heating for its Prishtina customers. The €28 million project is expected to be completed by the end of 2014 financed by KfW, the Municipality of Prishtina, as well as grants from EC, Swedish and Luxembourg development agencies. The Government of Kosovo needs to promote an enabling environment that would incentivize further growth of District Heating systems as well as sustainable biomass as options for the replacement of electricity as a source of heating. 3.8. The energy sector has received large subsidies but these have fallen in recent years. Energy sector subsidies have fallen from a high of €69 million (1.7 percent of GDP) in 2009 to €18 milli on (0.4 percent of GDP) in 2012. To protect consumers from high import prices, over €150 million (0.7 percent of GDP) was spent in total on import subsidies between 2008 and 2012. Import subsidies accounted for 78 percent of total subsidies in the sector. The government also subsidized energy costs for recipients of 25 the last-resort social assistance program (discussed in more detail in the Social Protection chapter of this report). These subsidies amounted to €4.5 million (approximately 0.1 percent of GDP) per year between 2008 and 2012. 3.9. In addition to cash subsidies, firms in the power sector as well have received grants and loans from donors or the public sector to fund capital repairs or improvements. KEK, the publically owned energy company that was vertically integrated until 2010, has received significant grants from both the budget and donors, and successor companies have continued to be supported (Table 3.1). These grants largely financed capital repairs or improvements. In 2009 alone, the Government of Kosovo (GoK) transferred 3.8 percent of GDP to KEK in the form of grants and loans totaling €141 million. Much of this went to the Sibovc Mine though large amounts were also spent on rehabilitating Kosovo A and upgrading the transmission network.66 Capital grants to the sector moderated after 2008 but remained high: for example, €16.5 million were spent directly from the budget improving the transmission network between 2010 and 2012. Grants and loans from the GoK and donors decreased from nearly 60 percent of total income for the sector in 2009 but still made up over a fifth of reported income in 2012. Audit reports revealed that KEK’s accumulated debt increased every year between 2008 and 2012. In addition, no audit reports during this period showed any sign of the company making interest payments on their debt stock. The result is that outstanding debt owned by the sector increased from around €21 million in 2008 to over €200 million by the end of 2012, or about 4 percent of GDP. Table 3.1. Government and Donor Transfers to KEK 2008– 2012 in ‘000 of Euro Year 2008 2009 2010 2011 2012 Liabilities Accumulated debt (loans) 20,827 97,540 157,540 187,540 206,274 Cash flows from financing activities Grants received during the year from GoK 87,744 64,197 51,961 28,024 13,682 Net loans received during the year from GoK 487 76,713 60,000 13,682 10,802 Income Income from grants (GoK + donors) 47,695 58,472 44,055 38,033 46,470 Energy revenues (sales) 163,043 174,392 192,757 215,960 222,591 Total income 210,738 232,864 236,812 253,993 269,061 Income from grants and net GoK loans as percentage of 23% 58% 44% 20% 21% total income (%) 3.10. These subsidies, loans and grants have kept energy prices artificially low. Despite the large distribution losses, there are few countries in Europe whose consumers pay cheaper electricity prices than Kosovo. At over 12 cents per KWh in 2012, tariffs in both Croatia and Turkey were double those of Kosovo.67 Such low tariffs in Kosovo arise partly because Kosovo A and B were able to generate power at wholesale prices of less than 2.5 cents per KWh thanks to the generous grants and (unrepaid) loans to fund capital repairs and improvements. 66 In 2008-09, direct expenditures from the budget included nearly €21 million on the Sibovc Mine, €4.6 million on the rehabilitation of Kosovo A and €5.2 million improving the transmission network. 67 In absolute rather than PPP terms. 26 D. MEETING THE GROWING DEMAND: BUILDING NEW GENERATION CAPACITY 3.11. Kosovo will need to improve energy efficiency, encourage use of renewables, finance new generation capacity, and improve regional transmission connections to meet ever-growing demand for power. Meeting growing energy demand will require effort on several fronts. First, the construction of a new power station is estimated to cost in excess to 20 percent of GDP, according to the World Bank Energy Options Study68 though there are uncertainties estimated at plus or minus around 25 percent. For that reason, the cost in this report is considered to between €1 and €1.5 billion. The new plant will take around four years to complete. Given the large costs and need for state-of-the-art technology, it will be desirable for the private sector to take a lead role in financing and managing the power plant. Constraints on the availability of private sector financing will determine the need for residual public funding from various sources. . This could mean that other capital projects might need to be delayed as resources are reallocated. Second, given the power gap that exists even in the short term, it will be essential to ensure sufficient regional inter-connectivity to be able to import energy when necessary. Third, investment in renewable energy (RE) sources can help to close some of the power gap, diversify Kosovo’s energy sources, and support the country to meet European environmental obligations. Pump storage hydropower can also be useful to meet peak power demand. 3.12. Plans have already begun to improve Kosovo’s international transmission connections (though there is limited capacity available for imports in the region). Kosovo’s transmission grid is in good shape but awaiting an important investment for a 400 kV interconnection line with Albania. Currently, Kosovo is linked to the regional system through 400 kV interconnections with Serbia, FYR Macedonia, and Montenegro, and a 220 kV line to Albania. The new interconnection project with Albania, which is financed by KfW, will facilitate power exchanges between the complementary power systems of both countries. In December 2013, transmission operators from Kosovo and Albania signed contracts for the construction of the 400kV line. Total investment cost for the Kosovo side is €33.5 million (€16.5 million grant and €17 million loan from KfW) financed by the German government based on an agreement signed between KfW, Kosovo’s Ministry of Finance, and KOSTT. The contract specifies that the construction needs to be completed in 24 months. 3.13. Developing a dynamic mix of resources, including ambitious renewable energy (RE) development for Kosovo, could achieve a well-balanced energy system. In October 2012, Kosovo set an ambitious RE target of 25 percent of its total energy consumption by 2020. Private sector interest in Kosovo’s RE sector has increased substantially. By the end of 2013, the Energy Regulatory Office (ERO) had received 26 license applications for hydropower and wind projects totaling over 350 MW of generation capacity. Around 157 MW of generating capacity from wind power and about 201 MW from small hydro plants have been submitted to the ERO. As is typically the case in renewable projects, the applications may not be fully realized. 3.14. Use of renewable energies is not expanding as fast as it could, partly due to administrative bottle-necks. The growing number of applications from the private sector for renewable energy generation licenses has increased pressure on the ERO to effectively fulfill its mandate with regard to renewable energy application processing. However, so far, only two applications for hydro plants totaling 30 MW of installed capacity and two wind power applications for a total of over 60 MW have received final authorizations from the ERO. Other applications have received a preliminary authorization. In addition to generation for the power grid, there is also significant potential to incorporate more renewable energy applications in space and water heating, since current practices rely on fossil fuels, such as lignite and oil, electricity, or unsustainable biomass. Water heating, for example, represents 15 percent of heating 68 World Bank (2011), “Development and evaluation of power supply options for Kosovo”. Report available f rom: http://siteresources.worldbank.org/INTENERGY2/Resources/Kosovo_generation_options_report_12312011.pdf 27 demand and primarily uses electricity, which is inefficient. The use of sustainable biomass to replace electric heating could be an attractive policy option if it can be done in a more sustainable manner. Box 3.2. Energy Efficiency Kosovo has relatively high energy intensities that are approximately 2.5 times higher than the average for European countries. The residential sector represents the largest portion of energy consumption at 38 percent. Promoting energy efficiency in the public, residential and private sectors could help save money and reduce energy demand, narrowing the power gap. Energy efficiency (EE) can help address issues related to energy security (current account deficits and availability of imports), while reducing public expenditures on energy and the environmental impacts of energy use. The Government of Kosovo has implemented a National Energy Efficiency Action Plan (NEEAP) for the period 2010-18 that calls for a cumulative energy savings of 9 percent over three periods of three years each. There were 38 EE measures included in the first NEEAP period from 2010-12 which resulted in meeting the target of achieving energy savings of 3 percent of total energy consumption (360 GWh). The EE savings from measures during this period were split almost evenly between the residential and services sectors at roughly 10 kilo ton of oil equivalent (ktoe) each. The measures were concentrated in the residential and services sectors (public sector only), but select measures in the industrial, transport, agricultural sectors, which collectively account for 53 percent of energy use, where also included. The World Bank Institute (WBI) conducted a preliminary assessment of market potential for EE investments covering the entire building stock of Kosovo in 2013. The building sector (residential, public, and commercial buildings) accounted for 47.5 percent of final energy consumption in 2011, with the residential sector accounting for the bulk of energy consumption. The energy savings potential across the building sector was estimated at more than 44 percent. As a percentage of primary energy supply, energy savings potential for the building sector in Kosovo totals approximately 20 percent. To implement the necessary measures, an investment of €1.37 billion in the building sector would be required. Investment needs were primarily concentrated in the residential sector, representing 80 percent of total investment needs. But unlocking the required investments will be difficult, potentially requiring subsidized energy-efficiency measures. A high savings potential exists for public buildings. For central government buildings, the energy savings were estimated to be about 49 percent. For municipal buildings, health buildings could save 47 percent, schools 38 percent and other municipal buildings 46 percent, even with modest improvements in current comfort levels. Such savings offer substantial budgetary savings. Estimates indicate that GoK spent some €41 million in 2011 on energy in its buildings and could save as much as €18.9 million annually through cost -effective EE measures. Based on data from similar projects in neighboring countries, simple payback periods for investments in public buildings are in the 6-8 year range. 3.15. The cost of renewable energy will add to the overall cost of supply and will require integration into a sustainable tariff path. In November, 2013, the GoK adopted its National Renewable Energy Action Plan (NREAP) for 2011–2020, committing the country to a 25 percent RE target (of final energy consumption) by 2020 with a voluntary target of just over 29 percent. With regard to the electricity sector, the NREAP foresees increased RE generation from an additional 240 MW of small hydro, 305 MW of large hydro, 150 MW from wind, 14 MW from biomass and 10 MW from solar PV. To encourage investment in renewables, the ERO has set the feed-in tariff for wind generators at €85/MWh, while for small hydro plants ERO set a feed-in tariff of €63.3/MWh. A feed-in-tariff of €71.3/MWh has been set for biomass/biogas and a photovoltaic solar feed-in-tariff is currently under development. Renewable energy will not be able to meet the full growth in demand and, in particular, will not be able to meet peak demand (because it is not dependable capacity). Rationalization of energy prices will be fundamental to the sector because when low, they make RE options wholly uncompetitive. 28 E. IMPACT OF POWER GENERATION OPTIONS: THREE SCENARIOS 3.16. Replacing Kosovo A with a state-of-the-art power plant would have immediate environmental, economic and social benefits. The health costs from air pollution were estimated at 2.3 percent of GDP in 201069 – almost as much as the entire public health budget. Although there are several key sources of air pollution, including wood and lignite use for household heating, the 50 year old power plant remains a significant source of pollution70. Lost days of work and school add to the social and economic costs. Other (non-environmental) long-term benefits of replacing Kosovo A with a new facility according to EU-compliant standards71, would be even larger. The authorities are acutely aware of the importance of building new generation facilities and have begun the process of seeking private investors to build a new power plant. 3.17. The rehabilitation of Kosovo B also will be necessary to prolong its life and meet environmental standards. Rehabilitating Kosovo B would increase its net output capacity from the current 550 MW to 600 MW, while improving pollution controls to meet the relevant European Union environmental Directives. A USAID pre-feasibility study placed the cost of rehabilitation at about €350 million over a period of four years from the start of the project. The rehabilitation of Kosovo B would extend its life, but would not be sufficient on its own to meet growing demand. 3.18. For the purposes of this report, a detailed modeling tool was developed to estimate the financial implications of several energy scenarios for the government and consumers. A sectoral modeling tool – the Kosovo Energy Sector Impact Tool (KESIT) – was developed based largely on the cost calculation model used by the ERO. The tool also draws on a financial model developed to estimate the wholesale cost impact of different possible cost and financing scenarios likely for the construction of the new power plant. A detailed analysis was possible using these two models, as well as other available reports, and provides an understanding of the financial implications for end-user costs, subsidies, imports/exports, power generation, and supply and demand. 3.19. The KESIT model is presented for illustrative purpose only. It is designed to provide an estimate of potential impacts of various policy choices or other outcomes. Users are able to change parameters to view possible outcomes of over a hundred different policy mixtures. Out of many possible scenarios that could be presented using the sector modeling tool, the three selected for presentation in this report are: (i) inaction; (ii) Energy Strategy; and (iii) last resort. These scenarios are hypothetical and designed to serve as a useful starting point for analyzing potential outcomes with different policy choices. 69 World Bank (2012), Country Environmental Analysis : Kosovo http://documents.worldbank.org/curated/en/2013/01/17485553/kosovo-country-environmental-analysis-kosovo- country-environmental-analysis-cea 70 In 2013, KEK completed installation of electrostatic precipitators (dust filters) in all working units of Kosovo A power plant (units A3, A4 and A5). In addition KEK has installed wet ash transport replacing old belt conveyer system. 71 Kosova e Re Power Plant (KRPP) will be compliant with the new EU Directive 2001/80/EC on limitation of emission of certain pollutants into the air from large combustion plants, EU Directive 2008/1/EC concerning integrated pollution prevention and control and EU Directive 2010/75/EC on industrial emissions which came into force in 2012, and the associated sector specific Best Available Techniques Reference (BREF) documents (latest BREF published in 2006). 29 Box 3.3. Kosovo Energy Sector Impact Tool The Kosovo Energy Sector Impact Tool (KESIT) was developed to model the impact of energy sector options on power supply and demand, average costs, public subsidies, public debt, and imports and exports. It was based on a model used by the Energy Regulatory Office to determine average costs. The KESIT models all stages of the power sector in Kosovo between 2012 (actual data) to 2030. It therefore considers the costs (including recurrent costs, investment/depreciation and allowed profit) and revenues for the public or private companies that operate the mine, power generation, transmission and distribution. It also takes into account the small amount of power generation connected directly to the distribution network and the large industrial consumers that receive their power directly from the transmission network. The parameters can be modified within that model. In the KESIT, imports and exports are used as residuals to balance the supply and demand. The baseline model considers that, with the exception of subsidies for poor consumers, all costs are passed through to the final consumer. KESIT is very user-friendly. Users can easily adjust parameters through drop-down menus on a dashboard and the effects of their changes (compared with the baseline) on power supply and demand, average costs, public subsidies, public debt, and imports and exports are shown in graphic form on the same page. By combining different changes, the user is able to easily generate energy scenarios such as the ones presented in this report. The key parameters in the KESIT are presented below: Energy demand parameters:  Under the baseline, energy demand increases by around 4.5 percent per year for both large industrial consumers and all others (households, most private sector firms). Energy use by the power plants themselves and related (e.g. mine) remains a constant function of their generation throughout the period. The user can alter demand growth to between 2.5 and 6 percent per year. The changes affect imports and exports, and since these are not traded at the same price as domestic production, this alters the average domestic cost. Energy supply parameters:  The user is able to modify the project cost, interest rate, grace period, financing mix (including public financing options), and years of the construction.  Under the baseline, the user can select which year Kosovo A is decommissioned.  Under the baseline, Kosovo B is renovated over three years when KRPP comes online, financed by the public sector and at a cost provided by a 2012 USAID report. It is more efficient (and meets European emissions standards following renovation). The user can select the financing source (private, public debt or directly from the budget), the year or renovation or can choose to allow the power plant to continue generating in its current state over the medium term.  The tool also contains numerous other parameters that the user can adjust. These include: distribution losses, which fall steadily from 37 percent in 2012 to 11 percent by 2030 in the baseline (this is responsible for large savings, which can be passed on to consumers); new renewable energy generation; wind and hydro capacity factors; export and import price shocks; the import price elasticity (important in some scenarios since Kosovo could require large imports from an energy starved region and could therefore influence regional power prices); and losses from unbilled supplies. *Users of the KESIT without access to this model are able to use a simplified version within the KESIT, which generates similar but not identical results. 30 i) Scenario 1: Inaction 3.20. Kosovo does not build the KRPP Figure 3.9. Supply and Demand of Power, GWh and continues to run its old power plants for the maximum time allowed by the Energy Community Treaty agreements. Kosovo A was originally planned to be decommissioned by the end of 2017. Kosovo B will continue to operate indefinitely Power gap although it is not refurbished under this scenario. There will be some renewable energy uptake by private sector operators. At the same time, domestic demand will continue to grow. 3.21. Under this scenario, a large gap would open between supply and demand requiring a growing amount of imports that would need to be paid for through increased costs or subsidies. The reduction of output of Kosovo A and then its closure would leave a large gap in generation capacity Source: KESIT. (Figure 3.10). Some new renewables are expected to come online in this time and these will help the country to reduce required imports but overall there will remain some power gap even in the short to medium term. After this period, the power gap will grow significantly. By 2030, generation will be less than 60 percent of domestic demand, resulting in large and costly imports. 3.22. Replacing Kosovo A with imports is not an option. Even if it were, imports would impose heavy financial burdens on consumers and/or the budget. The KESIT assumes that it will be possible to import sufficient quantities of energy to meet demand, albeit at higher prices. It is not clear that this will be possible since import availability relies on uncertain relations with neighboring countries, unknown future generation capacity in the region, and weather conditions, which affect both demand and supply in the region. A large proportion of power in the region (98 percent in Albania and around 40 percent in Bosnia and Herzegovina for example) is produced by hydropower stations, whose generation will likely become more volatile as the impact of climate change is felt. Even if a large volume of imports were possible, the costs would be high. Imports would increase from the 300 to 700 GWh per year imported over the last decade to over 4,500 GWh (Figure 3.11) at a cost of over €480 million per year or 5 percent of GDP by 2030.72 This could be higher or lower depending upon (unknown and difficult-to-estimate) price elasticities. If these full costs are passed through to consumers, the average cost can be expected to increase to around 10.3 euro cent per KWh even excluding inflation, compared with less than 6 euro cent today, or an increase of 72 percent (Figure 3.12). 72 In 2012 Euro terms (i.e. excluding inflation). 31 Figure 3.10. Estimated Imports and Exports, Figure 3.11. Estimated Costs, € Cents Per Kwh Gwh Source: KESIT. Source: KESIT. ii) Scenario 2: Energy Strategy 3.23. A new thermal power plant presents the best value-for-money option even when including the costs of externalities (if carbon taxes were to be introduced in Kosovo). Several alternative energy strategies have been considered including those that have focused on renewables and on building new thermal generation capacity. The authorities elaborated an Energy Strategy which includes new renewable energy and construction of a new thermal power plant (KRPP) to fill the energy gap created by growing demand and closure of Kosovo A. The World Bank (2011) Options Study examined several alternatives for closing the power gap including renewables, imports, a new natural gas power plant and a new thermal power plant. 3.24. Under the Energy Strategy scenario, new renewables and a privately constructed and operated KRPP replace Kosovo A. The Energy Strategy includes more renewables than in the Inaction scenario with the aim of meeting the National Renewable Energy Action Plan (NREAP)73 targets. Construction of the KRPP would begin in 2015 at a total cost estimated at €1.2 billion by the Options Study (the total costs are currently uncertain and could be between 20 and 30 percent of GDP), and is anticipated to be fully financed by private sector investors. The plant will begin operation at the same time Kosovo A is decommissioned (Figure 3.12). Since the KRPP will more than fully close the power gap in the short term, this will also be the ideal time to refurbish Kosovo B, whose production will therefore decline over the short term. 3.25. Feed-in tariffs for renewable energy (of between €63 and €85 per MWh) and the costs of investing in energy efficiency combined with the cost of building the KRPP will increase energy costs. Total project costs of €1.2 billion for the KRPP, including interest, would increase the average cost to an estimated 8.55 euro cent per KWh before subsequently declining as distribution losses are cut, if all costs are passed through to consumers (Figure 3.13). However, by smoothing the cost increase, average tariff adjustments could be smoothed over time. The estimations presented do not integrate the costs of 73 Available from: http://www.energy-community.org/pls/portal/docs/2570177.pdf 32 repaying the €206 million (presented in Table 3.1 above) owed by the energy sector to government. Integrating these repayments would raise costs for the consumer even further. Figure 3.12. Supply and Demand of Power, Figure 3.13. Estimated Average Cost, euro cent Gwh Per Kwh, 2012 Euro Source: KESIT. Source: Source: KESIT. Note: Baseline assumes construction costs of €0.95bn and the Simulation assumes construction costs of €1.28bn. 3.26. Even with adjustments, costs would Figure 3.14. Pre-tax residential costs, 2012, remain low compared with most SEE and Eastern euro cent per KWh74 European countries. Kosovo currently has the 16 lowest pre-tax average residential costs among SEE and comparator Eastern European countries, with the 14 exception of Serbia (Figure 3.14)75. Kosovo’s 12 comparatively low costs are due to the cheap cost of generation from its aging power plants, where 10 construction costs have been fully amortized. While 8 maintaining such low tariffs will be difficult following the construction of a new power plant, 6 smoothed cost adjustments could help to ensure that 4 costs remain among the lowest in Eastern Europe. It should be noted that these costs do not include 2 externality payments such as carbon taxes, and it is 0 SRB EE LT LV TU ALB BiH HR PL HU SK not currently planned to introduce these in Kosovo. KOS BG MKD RO MDA MNE 3.27. Current low energy costs could not be Source: Energy Regulations Regional Association: maintained following large investments in energy http://www.erranet.org/ efficiency, renewable energy or a new power 74 ALB=Albania; BiH=Bosnia and Herzegovina; BG=Bulgaria; HR=Croatia; EE=Estonia; LV=Latvia; LT=Lithuania; MKD=FYR Macedonia; MDA=Moldova; MNE=Montenegro; PL=Poland; RKS=Kosovo; RO=Romania; SRB=Serbia; SK=Slovakia; TU=Turkey. 75 Post-tax costs are also the lowest except for those in Serbia. 33 plant. Maintaining 2013 prices following large-scale investments of any kind would require significant government subsidies to the sector. With non-smoothed costs, these subsidies are estimated to peak at around €240 million annually in 2020, which would be difficult from a fiscal perspective. An alternative option to minimize the negative social and economic consequences would be to smooth cost adjustments. 3.28. The impact of large-scale investment on retail costs should be carefully managed, partly by smoothing cost adjustments for consumers. With an estimated value of more than 20 percent of GDP, investment in the KRPP – as well as those in energy efficiency and renewable energy – will need to be recouped. If all costs are passed through to consumers, the increase could have economic and social consequences as households and firms see bills rise. Particular attention should be paid to the poor and vulnerable, such as those receiving the last-resort social assistance. Another way to reduce the impact will be to give households and firms time to adjust. Changes will need to be made to the tariff adjustment mechanisms to allow for smoothing over a longer time period. iii) Scenario 3: Last Resort 3.29. If sufficient private resources are not available for the KRPP, the authorities will need to consider financing construction by reallocating budget resources. The Energy Strategy assumes that the KRPP will be developed using private resources. As with every power plant tendered to the private sector, uncertainty remains over the ability or willingness of the private sector to undertake such a project, especially in the context of tight financing conditions. Should it prove impossible for the private sector to finance such project, ensuring energy supply in Kosovo may require turning the KRPP into a publicly financed project. This would not be a preferred option, and should be considered a last-resort option only. Nonetheless, it would be preferable to the Inaction scenario detailed above. The financing options available would be (i) using budget financing reallocated from other large investment projects, notably in the transport sector; (ii) issuing debt specifically to finance the construction of the power plant assuming that this could fit within the framework of the Fiscal Rule, and/or finding significant privatization receipts to finance the construction; (iii) reallocating a part of the transport capital budget to co-finance the construction of the KRPP jointly with private sector investors; and (iv) a combination of these options. In all cases, the authorities would need to ensure that tariffs remain cost reflective to recoup their investment or repay debt. Here, we show the estimated Figure 3.15. Annual Budget Costs for impact on government spending, public debt, and Construction, € Mil. interest payments if the public sector finances the KRPP with construction costing €950 million, using a mix of 50 percent budget payments and 50 percent debt. 3.30. This option would have significant fiscal and public debt impacts but may be necessary to ensure sustained economic growth. Under this scenario, annual budget costs would be between €107 million and €160 million (2 and 3 percent of GDP) per year over the first four years of construction (Figure 3.15). Public debt would increase during the construction period as new borrowing was issued each year. At its peak, additional public debt would be worth €591 million or 12 percent of 2012 GDP, more than doubling the current 11 percent (Figure 3.16). Resources would be needed not only to fund the short- Source: KESIT. term budget costs, but also to service the debt (Figure 3.17). These should be found by ensuring costs and tariffs remain aligned, with the adjustment of tariffs 34 presenting the same social and economic issues as posed by private sector financing of the KRPP. Debt service would begin when the power station opens. Interest payments alone would total €48 million that year. While this scenario may become unavoidable, the authorities and their partners should make every effort to secure private financing for the KRPP and avoid pressure on the budget. To avoid such high debt and interest payments, the GoK may need to consider reallocating resources from other capital spending, such as the proposed construction of Route 6. Figure 3.16. Additional Public Debt Stock, Figure 3.17. Additional Annual Public Interest € Mil. Payments, € Mil. Source: KESIT. Source: KESIT. F. PROTECTING THE POOR FROM RISING ENERGY COSTS 3.31. Regardless of the energy solutions chosen, poor households will be disproportionally affected by price adjustments. Current efforts to protect the poor are inadequate. Electricity represents a larger share of household spending for the poor than for the non-poor. Although there would likely be some behavioral adjustment, the cost increase projected under the Energy Strategy scenario would be difficult for some households to absorb. In addition to the €4.5 million spent every year between 2008 and 2012 subsidizing energy for recipients of social assistance, the energy tariff structure is meant to provide cheaper energy for poorer consumers. 3.32. The tariff structure has been designed partly to provide an implicit subsidy from wealthier to poorer consumers. The energy tariff structure is complex with 61 different rates76 depending upon connection type, meter type, season and time of day, plus an additional six fixed connection charges. Households can be subject to one of 21 different tariffs plus a fixed connection charge depending upon the same factors. This structure was designed to provide an implicit subsidy from larger consumers (assumed wealthier) to smaller (assumed poorer) consumers. For example, from June 2012, a household with a single tariff meter paid 3.22 euro cent per kilowatt hour (KWh)77 for the first 200 KWh consumed in the low (summer) season, while consumption greater than 600 KWh attracted a rate of 6.49 euro cent/KWh. 3.33. The current tariff structure has not been an efficient way to help the poor. A recent analysis conducted for the World Bank showed that the current tariff mechanism is actually regressive. Wealthier 76 Excluding street lighting, which attracts its own specific rates. 77 Excluding VAT. 35 consumers benefited more than poorer ones in 2011, with the wealthiest quintile receiving 27 percent of the subsidies and the poorest just 11 percent (Figure 3.18). The failure of implicit subsidies appears to have come from three main sources: first, wealthier households were better able to capitalize on the subsidies than poorer ones because wealthier households used more energy-consuming utilities – 11 percent of the wealthiest quintile reported having central heating78 in the 2011 HBS compared with fewer than 2 percent of those in the poorest quintile – they tended to consume more energy on average. This means that they were able to benefit fully from the subsidized energy, whereas poorer households were not. Second, the system allowed consumers on more complicated meters to adjust their energy consumption profiles to take advantage of lower energy prices at certain times of day (for example, by installing storage heaters). Third, for unmetered households, prices depended on estimated energy consumption, and they paid significantly more for their energy than other consumers. For example, an unmetered household estimated to have consumed between 200 and 600 KWh of energy per month has paid a monthly charge of €42.37 since June 2012. On the other hand, a consumer on a single tariff meter could have consumed 670 KWh of energy for the same price in the peak (winter) season, and paid even less in the low demand season (Table 3.2). If poorer consumers were more likely to be unmetered, the effectiveness of the implicit energy subsidies will have been lower than hoped for. Due to these defects, the implicit energy subsidies compared poorly with Kosovo’s social assistance sch eme, which saw 82 percent of spending go to the poorest 40 percent in 2011 (and which included some energy subsidies). Simplifying the tariff and ensuring its neutrality with regards to income could be considered. Figure 3.18. Distribution of the Implicit Table 3.2. Comparison of Unmetered Electricity Subsidy Across the Population, Households and Single-Tariff Metered Kosovo 2011 Household Energy Costs 90 Monthly charge for Energy consumption at Proportion of the subsidies 80 received by each group Estimated consumption estimated equivalent high season 70 bracket, unmetered consumption, €, cost for single tariff 60 households unmetered households meter households 50 <200kWh 23.41 360 KWh 40 200-600 42.37 670 KWh 30 <600kWh 71.40 990 KWh 20 10 Source: Energy Regulatory Office and World Bank 0 calculations. Poor Non Poor quintile 2 quintile 3 quintile 4 quintile 5 quintile 1 Source: HBS Kosovo 2011. 3.34. In addition, despite the implicit subsidies, the poor spent a larger share of total expenditures on electricity than the non-poor in 2011 and “energy poverty” was high. On average poor households spent 8.4 percent of their expenditures on electricity in 2011, compared with 5.9 percent for non-poor households (Figure 3.19), according to the 2011 Household Budget Survey (HBS)79. Energy poverty, defined as spending more than 10 percent of household expenditure on energy, was 23 percent in 78 Note that central heating is distinct from district heating, the former being provided to a single home, usually using electricity in Kosovo. 79 This can be compared with some selected regional comparators: In Bulgaria, for example, the wealthiest quintile spends an estimated 16 percent of its total income on energy and the poorest quintile spends 9.5 percent. Hungary: Wealthiest quintile spends 12.6 percent and the poorest 20.9 percent; in Poland the wealthiest quintile spends 12.3 percent and the poorest 13.5 percent. See World Bank (2013), Balancing Act: Cutting Subsidies, Protecting Affordability, and Investing in the Energy Sector in Eastern Europe and Central Asia. Available from: https://openknowledge.worldbank.org/bitstream/handle/10986/12296/9780821397893.pdf?sequence=7 36 2011 though it has fallen from 38 percent in 2009 despite an increase in costs from 6.13 euro cent/KWh in that year (Figure 3.20). 52 percent of households in the bottom quintile reported no electricity spending in 2011, indicating that they are either not accessing electricity, or not paying for it. Figure 3.19. Shares of Energy Expenditure in Figure 3.20. Energy Poverty and Energy Kosovo, 2011, by Quintile Expenditures in Kosovo, 2009-2011 12 45% Percentage of total expenditures 40% 10 35% 8 30% 6 25% 20% 4 15% 2 10% 5% 0 Poor Non Poor 0% electricity share gas share LPG share 2009 2010 2011 wood share oil share other fuel share Energy poor no energy expenditures collected wood Source: HBS Kosovo 2011. Source: Household Budget Surveys. G. RECOMMENDATIONS 3.35. The energy sector is crucial for fiscal sustainability and growth in Kosovo. Although significant private sector participation is to be strongly encouraged, the extent of its eventual involvement remains uncertain. Yet, this will be the most important factor in determining public spending in the sector over the medium term. Regardless of the extent of private sector participation, reforms can make the sector more sustainable and reduce ad hoc public sector subsidies. 3.36. Among Kosovo’s many competing needs, the most important aim for fiscal and energy policy should be to ensure that energy demand is met in order to maintain economic growth. Without this, the resources to make long-term investments in other sectors may fail to materialize. This report therefore recommends:  Ensuring that energy demand is met, even if public resources need to be “freed” by delaying other investments. Although demand forecasts for energy consumption vary, clearly there will be increased demand. The aging power plants will not be able to meet the growing demand and need to be decommissioned. So far, Kosovo has not built new generation facilities as planned, and it is uncertain how energy will be supplied to meet future demand, a particular concern given that most neighboring countries are also in energy deficit. . Relying on external supply is not realistic from both a supply and a cost perspective. The GoK should ensure that demand growth can be met through a mix of EE, RE and thermal power generation, through the construction of the KRPP. Every effort should be made to attract sufficient private sector financing. However ensuring adequate power supply may entail a difficult decision to postpone other large public sector investments, including the new Route 6 highway to FYR Macedonia. While both investments could be seen as important, the adverse impact of failing to meet power demand would be especially detrimental to economic growth. 37  Streamlining complex application for renewable energy sources to increase supply of renewable energy. With just 15 percent of applications for hydropower capacity approved and all new applications facing slow progress through the regulatory approval process, Kosovo could be missing a chance to attract renewable energy sources that would increase its energy independence and slow the growing gap between supply and demand. To attract private sector financing for renewable energy, the ERO needs to continue to pursue the development of a robust regulatory framework, including the streamlining of the licensing process. 3.37. Other policies in the energy sector could also have beneficial impacts on the fiscal stance, reduce the negative impacts of new investments on poverty, and increase the likelihood of increased private sector participation in the energy sector, thereby reducing the fiscal burden. These policies include:  Ensuring that tariffs fully reflect costs and guaranteeing the complete autonomy of the ERO to maximize the potential for private sector participation in energy and minimize the requirement for energy subsidies. Since the GoK has planned a significant role for the private sector in the development of the Kosovo energy sector, it is crucial that investors benefit from a predictable and fair market environment. Investors need confidence that investments can be fully recovered through cost reflective tariffs. In addition, investors need to be persuaded that the ERO will independently regulate the energy market without favoring particular market participants, including public sector companies. Without these guarantees, it will be difficult to ensure private sector participation. Cost reflective energy tariffs will also reduce the need for the public sector to provide ad hoc subsidies and (unrepaid) loans that currently absorb a significant share of the budget.  Considering affordability issues when deciding how to address future energy shortfalls. New investments will have to be paid by either the end consumers or government subsidies. The GoK needs to analyze how these different investments will impact the end-user costs and whether all consumers can afford such a price increase. At the same time, realism is needed on whether these costs can be borne by government budgets. A “middle path” can be found by raising costs gradually to give households and firms time to adjust, and by compensating poor and vulnerable consumers for the price adjustments through the social protection mechanism.  Eliminating the use of energy tariffs as a social protection mechanism. This mechanism is currently inefficient and regressive. The complex tariff mechanism could also be simplified. Electricity meters could be installed in all households since they allow for cheaper energy, as well as provide both the incentive and opportunity to adjust behavior to reduce consumption (at peak times), thus easing the burden of energy costs on poorer households. At the same time, this would help the energy sector to charge cost-reflective prices, reducing the need for government subsidies. Instead, the social assistance system should be used to protect the poor from rising energy prices (see Social Protection Chapter). 3.38. Energy efficiency measures play a critical role in moderating energy demand. Energy efficiency investments are win-win investments, providing energy bill savings for end users, reduced investments and imports for the government, and reduced environmental impacts. If EE measures are implemented in the building sector, a total of 239.055 ktoe energy savings, or a third of projected electricity imports for 2014 (729.87 ktoe), could result in €198 million in total energy savings per year 80 80 WBI Building Stock Study, page 109; Table 10-22. 38 from reduced consumption. This does not include the monetary value of reduced imports. There may also be cost savings for the public purse. Ways to do this include:  Continuing to promote energy efficiency in the public, residential, and private sectors to reduce energy use and generate fiscal savings. Energy use in public buildings could be reduced by over 40 percent without loss of comfort due to the current large energy wastage. Investments of €79 million in energy saving measures in municipal and central public buildings could save close to €19 million per year from the public sector energy bill. The GoK has requested to make such investments81. Although the residential sector represents the bulk of energy consumption in the buildings sector and therefore holds the most potential, energy savings in the residential sector would be more difficult to unlock, possibly requiring subsidized energy-efficiency measures. A careful cost-benefit study and pilot studies could be undertaken.  Promoting the growth of sustainable and renewable solutions for space and water heating. With increasing costs and challenges meeting the demand for electricity, ratepayers that use electricity for space and water heating must be incentivized and informed to switch to more affordable and sustainable heating solutions such as district heating, sustainable biomass (e.g. wood, pellet), heat pumps and solar water heating. Central and local governments need to promote an enabling environment that would make sustainable heating solutions available for households that use electricity for heating while also incentivizing them to switch. 81 World Bank is supporting GoK in implementing energy efficiency measures in public buildings through a $31 million project, Kosovo Energy Efficiency and Renewable Energy Project (KEEREP). 39 CHAPTER 4. TRANSPORT82 A. KEY MESSAGES  Compared with other countries in South East Europe (SEE), Kosovo has a large number of roads per square kilometer and per US Dollar of GDP, but the condition of many of them needs improvement. Among ten comparator countries in Eastern Europe, Kosovo ranks fairly high on road density which would argue for shifting the focus in the transport sector from the construction of new motorways to asset preservation of the existing capital road assets.  Kosovo has spent a large proportion of its capital budget on constructing a single highway even though other smaller road improvements would have had higher economic rates of return. Transport investment averaged 4 percent of GDP between 2007 and 2012 but almost 70 percent of this was used to construct the Route 7 Motorway to Albania. At the same time, investments with rates of return of up to 81 percent were ignored and the road maintenance budget was crowded out.  More attention needs to be given to optimizing investment in new roads so as to maximize their economic and social return. Consideration of new large road investments as single entities, rather than breaking them into their main segments, is incompatible with sound economic and financial management.  Resources for road maintenance could be increased through road-user and registration fees. Kosovo has low registration fees compared with most countries in Europe and there is scope to increase them. Increasing these charges would raise resources to improve roads, and also help to compensate for declining border revenues.  Kosovo’s railway network requires rationalization. It is not clear that rail is the most efficient way for people to travel in a small country particularly when investments in new roads will further reduce rail’s competitiveness. In addition, without a large change in the way the railway is operated, its subsidies will be an increasing burden on the state. Actions to improve its financial performance include carrying out route cost benefit analyses and closing passenger services (when alternative bus transport provides a better quality of service at a lower cost), and non-commercially viable parts of the network. For the remainder of the network there is potential for greater rail transport of minerals, containers, petroleum and cement. B. TRANSPORT SECTOR OVERVIEW 4.1. Kosovo’s main modes of transport– roads and rail – were in poor condition at independence due to the effects of war and years of low investment and maintenance, and needs are still great . Roads are the main mode of transport in Kosovo, comprising more than 90 percent of transport infrastructure by length and 98 percent by service delivery (ton/passenger kilometers). Despite a 82 Chapter prepared by Robin Carruthers and Baher El-Hifnawi. 41 rebuilding of main roads and railway facilities since the war and investments in new motorways, resources have not focused on providing an adequate level of maintenance for the infrastructure. 4.2. Overall, Kosovo’s road transport infrastructure is comparable with other Eastern European countries. Among a group of ten comparator countries in Eastern Europe83, Kosovo ranks high on measures of “road transport density”, defined as the average of road length per square kilometer, per dollar of GDP and per capita (Table 4.1). Kosovo’s density ranking is average when measured as the average road length per thousand of population. Table 4.1. Transport Infrastructure Road Density Comparison Kilometers per: Country Thousand population Thousand square km US Dollar of GDP Pave Tota Pave Pave Unpaved Unpaved Total Unpaved Total d l d d Albania 2.0 3.1 5.1 244 382 626 559 875 1,43 4 1,32 Bosnia & Herzegovina 5.3 0.5 5.8 413 35 448 1,219 104 3 Bulgaria 5.2 0.1 5.3 358 5 363 777 12 789 Croatia 5.9 0.6 6.5 471 48 519 471 49 520 1,33 Kosovo 2.9 2.1 5.0 461 327 788 780 553 3 1,28 Moldova 1.8 0.3 2.1 238 38 276 1,111 177 8 1,80 Montenegro 8.4 3.8 12.2 381 171 552 1,250 559 9 Romania 2.4 1.9 4.3 195 151 346 275 211 486 1,08 Serbia 3.5 2.1 5.6 300 174 474 688 400 8 Slovenia 19.0 0.0 19.0 1,920 0 1,920 857 0 857 Average 5.5 1.6 7.1 476 155 631 762 331 1,09 3 Kosovo % average 27 215 70 51 352 125 54 278 122 Kosovo rank 6 3 6 2 2 2 5 3 3 Source: Authors estimates using data from CIA Fact book (2012). 4.3. Kosovo ranks high among the comparator countries on densities estimated on the basis of unpaved roads. The proportion of roads that are paved varies significantly in the sample, from 100 percent in Slovenia to 39 percent in Albania (58 percent in Kosovo). A high ranking on total road density is an indication of the level of connectivity within the country, so long as the unpaved roads are in good condition. If they are not, improving their condition would be a cost effective way of increasing connectivity without having to invest in paved roads.84 A high ranking on paved road density is an indication of how well the major population centers are connected to each other, and taken with the length per euro of GDP, it is also an indication of how well the road network can deal with traffic demands. 83 Albania; Bosnia & Herzegovina; Bulgaria; Croatia; Kosovo; Moldova; FYR Montenegro; Romania; Serbia; and Slovenia. 84 Indicators that include unpaved roads (and so also total roads) need to be treated with some caution as the specification of what is included as an unpaved road varies significantly between countries. At one extreme, some include every discernible track while at the other extreme only roads that have some engineering foundations are included. 42 4.4. Recent efforts to improve Kosovo’s road network concentrated on motorways while many national, regional and local roads did not receive much attention. The majority of high-level main roads has been developed and upgraded – several sections to motorway standards. Of particular note, is Route 7, a new highway to Albania, completed in November 2013. Since 2008, the GoK has also cooperated with the municipalities to upgrade around 1,100km of local and regional roads from earth and gravel to paved surface at an estimated cost of €200 million. Despite these efforts, many regional or local roads remain in poor condition, limiting access to businesses, schools and health facilities. 4.5. The recent trend in the number of road fatalities and fatal accidents is positive but traffic accidents and injuries have been on the rise. The number of fatal accidents increased sharply from 107 in 2002 to 156 in 2006, dropped to 118 in 2008 only to rise again back to 158 in 2010 (Table 4.2). Fatal accidents though have been on a downward trend from 2011 to 2013. The trend in number of fatalities has been generally in line with the number of fatal accidents. Higher penalties for infractions, stricter enforcement and a large road construction program (which slowed down traffic) were the main reasons for the reduction in the numbers of fatal accidents and fatalities. However, while fatal accidents and fatalities have been on a downward trend, the number of injured persons has more than quadrupled during the past decade rising from 1,983 in 2002 to 8,561 in 2012. Although the fatality rate, around 6.7 fatalities per 100,000 populations in 2012, is comparable to the figures in many EU countries, fatalities per 10,000 vehicles (around 4.2 in 2012) is 7-8 times higher than best performing EU countries, and more than 4 times the average imposing a significant cost on the economy.85 Table 4.2. Development of accidents 2002-2012 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Accidents with material 8,126 4,019 5,097 11,266 11,413 12,978 11,313 14,330 12,594 13,338 14,043 damages Accident with 1,153 1,415 1,326 2,506 3,013 3,901 3,850 4,730 4,327 4,462 4,555 Injuries Accidents with 107 107 141 145 156 127 118 152 158 128 116 fatalities Number of 132 130 170 155 178 139 133 176 175 155 121 fatalities Number of Injured 1,983 2,012 2,053 4,206 4,789 6,264 6,427 7,984 7,731 8,324 8,561 Persons Sources: (1) Data for 2002 – 2010 from the Road Safety Strategy and Action Plan for Kosovo, Egis International March 2012; (2) Data for 2011 – 2012 from Aksidentet rrugore për pesë vitet e fundit 2007 -2012. 4.6. It is important not only to sustain the reduction in fatal accidents that occurred between 2011 and 2013, but also to reduce the number of road traffic injuries. In order to meet the road safety targets set in the 2012 Road Safety Strategy and Action Plan, attention to the implementation of the action plan needs to be stepped up. The estimated cost the implementation of the Action Plan through 2020 is 2.5 million Euros. A careful analysis of the reasons behind the continuing increase in traffic accidents and injuries needs to be carried out in order to enable a better understanding and consequently a more targeted response to the problem. 85 Road Accidents for the Last Five Years (2007-2012), IRTAD Road Safety Annual Report 2013 43 4.7. Kosovo’s railway network is small compared to most SEE countries, even when considering its size. Kosovo has 330 kilometers of railway comprising two lines. The North-South line goes from Serbia at Polje through Prishtina to FYR Macedonia at Hani i Elezit, with both freight and passenger services and is part of Corridor X of the South-East Europe Transport Observatory SEETO Comprehensive Railway Network. The East-West line extends to the west for about 140km from Prishtina to Peja with a southern branch to Prizren, while in the east it reaches the Serbian border at Podujeva, with passenger services as far as Bardhosh. The combined freight ton km and passenger km per kilometer of track (railway intensity) is the second lowest in SEE and is only 7 percent that of the average of the railways in the European Union (Figure 4.1). However, geography needs to be considered when looking at railway density. Given Kosovo’s small geographic area most transport distances are too short for rail transports to be a viable alternative to roads. 4.8. Like roads, the railway infrastructure has suffered from decades of neglect and war and is proving expensive and difficult to repair. Many of its stations, bridges, tunnels and viaducts were destroyed. Neither freight nor passenger revenues are able to cover operating costs and therefore do not provide sufficient revenue for infrastructure Figure 4.1. Rail Traffic Intensity (ton km plus maintenance and upgrading. However, there may be passenger km per km of Line, 2009 (EU =100) a few bright spots on the horizon. For example, increased mineral production and exports could 180 provide opportunities to mining companies to 160 contract some portion of freight operations. 140 Traffic unitrs per km 120 100 C. TRANSPORT SECTOR INVESTMENT 80 60 Investment in Transport Infrastructure has Grown Fast but Not Been Aligned with Needs 40 20 4.9. Between 2007 and 2012, Kosovo’s 0 EU Slovenia BH ZFBH BH ZFBH Montenegro Croatia FYROM Serbia Germany Poland Kosovo Turkey Romania Bulgaria BH ZRS Albania transport investment has focused on new road construction. There was a large change in the scale and composition of transport investment measured both in euro and as a share of GDP between 2007 and 2012 (Table 4.3). As a share of GDP, the Source: World Bank (2011), Railway Reform in average over the six year period was 4.0 percent of South East Europe and Turkey. GDP, but fluctuated from a minimum of 0.4 percent of GDP in 2007 to an unsustainably high 6 percent of GDP in 2011 and 2012. Almost all of this increase in investment was due to the construction of Route 7, the new highway to Albania86. Road maintenance/rehabilitation investment also increased sharply until 2009 before falling back to less than 0.5 percent of GDP, as Route 7 expenditures began to crowd out other transport investments. 86 See Annex 2 for a discussion of trade and international transport infrastructure. 44 Table 4.3. Kosovo Investments in Transport Infrastructure Annual Expenditure in € million Activity Average 2007 2008 2009 2010 2011 2012 Average as % of GDP Road Rehabilitation 6.0 8.4 9.7 11.2 8.6 6.6 8.4 0.20% Road Maintenance 10.3 50.0 64.1 48.1 20.2 13.4 34.3 0.81% New road construction 1.0 30.6 37.8 143.7 270.3 258.8 123.7 2.91% Railway maintenance 3.0 3.4 3.4 3.7 4.1 4.3 3.6 0.08% Total investment 20.4 92.4 114.9 206.6 303.2 283.1 170.1 4.00% GDP (€m) 3,461 3,940 4,008 4,291 4,776 5,012 4,248 100.0% Transport investment % GDP 0.6% 2.3% 2.9% 4.8% 6.3% 5.6% 4.0% Sources: Authors’ analysis of Budget data and IMF Global Prospects, October 2013 4.10. Despite a low economic rate of return, the construction of Route 7 has had some positive economic effects. The construction of Route 7 generated temporary employment, helped to develop some local supply chains, and generated tax revenue. Between the start of the project and mid-2013, an estimated 11,084 jobs were created, of which 73 percent of project personnel were locally hired. Locally sourced material purchases for the project were around 52 percent, helping to sustain local business, although about three quarters of this was imported fuel. In 2012, the main contracting company, Bechtel, contributed approximately 19.5 percent of Value Added Tax (VAT) and 3.9 percent of Corporate Income Tax (CIT) received by the Large Taxpayers Unit (LTU). These benefits need to be weighed against both the costs and the alternative uses of the resources. 4.11. Delaying the investment in Route 7 would have increased the economic returns to investment. The estimated economic internal rate of return for Route 7 was around 8 percent.87 It is not clear that the construction of Route 7 had the highest possible rate of return on investment: outside the transport sector, there appear to be both current and potential investments needs that are greater than the need for Route 7. Even within the transport sector, a recent analysis showed economic rates of returns of up to 81 percent for improving more minor roads. Actual investments in transport therefore appear to differ from those that would provide the highest economic returns. 4.12. Kosovo has over-invested in new roads at the expense of other areas. In some cases a high level of transport investment can be justified by previous periods of under-investment that resulted in an infrastructure condition too poor to meet the demands of an expanding economy. But even where this is the case, the investment needs to be at an affordable level. Table 4.4 provides some international benchmarks for investment and maintenance. Compared with international benchmarks Kosovo has over- invested in new roads and underinvested in maintenance and rehabilitation in the last few years, as well as in railway maintenance (Table 4.5) (See Annex 2 for greater detail on international benchmarking of transport investment.) 87 Egis International, Economic Evaluation of Road Projects: Example 2: Route 7 Motorway from Morine to Merdare in Kosovo, June 2011. 45 Table 4.4. Benchmarks for Transport Infrastructure Investment in Kosovo ITF MED11/AICD Average Kosovo Maintenance of new and current 1.0% 1.0% 1.0% 1.0% New and upgraded roads 0.6% 1.8% 1.2% 2.9% Rail maintenance 0.4% 1.2% 0.8% 0.1% Total 2.0% 4.0% 3.0% 4.0% Sources: International Transport Forum, Statistics Brief, 2012 and Transport Infrastructure for MED11 Countries, CASE Network Reports 108, 2012. Table 4.5. Comparison of Actual and Desirable Transport Investment, 2007 to 2012 Actual Activity Average % GDP Needed % GDP Maintenance and rehabilitation 42.7.0 1.0% 42.7 1.0% New road construction 123.2 2.9% 51.0 1.2% Railways maintenance 4.3 0.1% 34.0 0.8% Total investment 167.5 4.0% 127.4 3.0% Source: Data from previous tables. 4.13. Kosovo’s average investment in maintenance and rehabilitation between 2007 and 2012 has been in line with international benchmarks, but it has dropped significantly in the past few years. Kosovo’s Figure 4.2. Expenditure on National Road road maintenance expenditure per kilometer is low Maintenance, €/km compared with other countries in SEE (Figure 4.2)88. 10,000 Road Maintenance €/km 89 9,000 While there are significant caveats to this comparison , 8,000 7,000 the results indicate an underfunding of maintenance in 6,000 5,000 Kosovo, however the costs are measured. 4,000 3,000 2,000 4.14. Rail investment has also been low and 1,000 0 dependent on support from the budget. In 2011 Serbia (2012) Kosovo(2007-2012) Romania (2011) Albania (2012) Bosnia (RS, 2011) Bosnia (FBH, 2011) Kosovo’s railway was split into two companies to bring operations in line with EU requirements: infrastructure is now the responsibility of INFRAKOS and operations that of TRAINKOS. The former is responsible for maintaining the 333 km rail network, while the latter’s responsibility is to provide freight and passenger train Source: Albania Public Financial Review, 2013. services. 88 Unit prices for maintenance in Serbia are set administratively, and not competitively, which is likely to result in higher-than-average prices. 89 Each national authority and international institution uses unique definitions of road maintenance and includes different types of roads. For example, in the comparison presented in this chapter, the national roads in Kosovo are the only roads included. However, in other countries, maintenance may include only routine maintenance or may, include in addition the reconstruction of paved roads. Finally, activities may be funded in different ways leading them to be included or excluded from the estimated maintenance expenditure. For example, if reconstruction is funded by IFIs the cost might be lost in the budget line on loan amortization. Comparisons that include all roads and/or varied expenditures, might give a different indication. 46 4.15. Between them, INFRAKOS and TRAINKOS planned to receive over €6 million or 0.12 percent of GDP in subsidies and other grants from the GoK in 2013.90 INFRAKOS’s business plan91 for 2013 included a subsidy worth €2.3 million and an additional payment of €1.9 million as compensation for a Ministry of Infrastructure request to reduce track charges below what would normally be required to maintain the track in a safe condition without imposing stricter speed requirements or axle loads. Between them these revenues, worth 0.08 percent of GDP, accounted for 71 percent of INFRAKOS revenues in 2013. TRAINKOS’s business plan for 2013 included subsidies worth an additional €1.9 million or 0.04 percent of GDP, accounting for around 30 percent of its total revenues. In total though, the railway companies use less than 2 percent of GoK transport investment. The Cost Efficiency of Transport Investments Has Been Low 4.16. In addition to not being fully aligned with needs, some recent transport investments may have been over-priced. Route 7 construction costs appear to be high compared with the average cost per kilometer for four-lane divided highways in developing countries, but about the same cost per kilometer as found in EU countries (Table 4.6).92 Kosovo’s high costs could be partly attributable to higher than average civil structures costs (for example the length of tunnels, embankments and cuttings, and bridges). To determine if the costs are reasonable, the total would need to be broken down into major components, including for these types of civil structures.93 Bidding inefficiencies may also have increased costs. Table 4.6. Comparison of R6 and R7 Construction Costs with Other Highways Country Project €m/km Argentina 2 bypass projects, total length 17.7km 3.0 India 7,000km of new expressways 4.5 4-lane divided highways in rural and United States mountainous regions 10.4 to 14.8 3.1 to 14.1 with an average European Union Expressways projects in 14 countries of 8.1 Kosovo Route 7 10.6 12.5 (as high as 23 for the Kosovo Route 6 Southernmost 15 km) Sources: COWI Economic Evaluations and Government Economic Viability Assessment of R6 and R7, various for others. 4.17. However, the upgrading of Highway 9 (M9) has been cost-efficient. Although some sections of the M9 were upgraded in 2000 and the section to Prishtina airport in 2009, several remaining sections still need to be improved. The original upgrading of 76km included reconstruction and a new surface, but no capacity expansion, at a cost of €1.9 million (2012 prices), or about €25,000 per kilometer. 90 All calculations exclude donor resources, for which there are no good data. 91 See Table A.2 in Annex 2 for summary Business Plans. 92 The comparison is based on costs adjusted to end-2012 prices. Some of the roads were constructed in the early 2000s and since then there has been a large cost escalation in the costs of expressway projects. 93 Construction costs for new highways have increased dramatically in the last 5 years, so comparisons with earlier projects are not helpful. 47 D. TRANSPORT INVESTMENT NEEDS Roads 4.18. Identified transport infrastructure needs remain large but meeting them would require a focus on road maintenance rather than new roads. Two studies detailed road maintenance needs over the medium term but many of these investments have not been undertaken due to the crowding out effect of Route 7. Investment needs may have increased with worsening road conditions; however the road condition database does not appear to have been updated so it is difficult to assess changes. Nonetheless, it is likely that similar needs remain today and there may scope to make cost savings on the estimated costs of road maintenance included in the database, particularly on local road maintenance. 4.19. Two widely divergent local road maintenance estimates exist. The first, the Local Roads Inventory undertaken in 2010, indicated that there are 39 road sections with “acceptable returns on investment” requiring investment of approximately €234m, or approximately €46.8m per annum spread over a five-year period. The second study, also made in 2010, estimated local road annual investment requirements of around €25.4 million.94 This estimate was designed to be compatible with the Medium Term Expenditure Framework (MTEF) for 2009 to 2011 and was therefore severely budget constrained. The higher estimate of the Local Roads Inventory made no attempt to reduce the unit costs by using alternative proven technologies, and it prioritized road sections with Economic Internal Rate of Return (EIRR) as high as 81 percent. 4.20. It may be possible to reduce the costs of local road maintenance. The Local Roads Inventory identified three types of activity that would be required for the maintenance of local roads: (i) sealing a surface on an earth or gravel road (which is classified as new investment rather than maintenance); (ii) asphalting roads (often currently earth or gravel); and (iii) overlaying an existing asphalt road with a new layer. Cost savings can be identified in each of these activities:  Potential cost savings for sealing earth or gravel roads: The average traffic level on the earth and gravel roads included in the study was only 100 vehicles per day – below the minimum 250/day normally required for new road investment. When sealing a road is considered desirable on non-economic grounds there are less-costly and longer-lasting methods to achieve this aim. In fact, sealed roads could cost approximately a quarter of the €370,000 per km95 quoted in this estimate.  Cost savings for overlaying existing asphalt roads and newly asphalting roads: The high costs of these estimates combined with the low traffic volumes produce very low and even negative economic outcomes. Overlays and surface dressings are more plausible interventions, given traffic levels and unit costs. A possible explanation is that the unit costs are about one third higher than the average of other countries.96 4.21. Overall, it may be feasible to reduce the investment cost of the asphalt projects to about one quarter of current estimates, and reduce costs of overlays and surface dressings by at least one third. The impact of these cost reductions would be to reduce the overall cost of the local roads activities from €234m to €125m and increase the economic rate of return from 21.2 percent to 26. 9 percent (Table 4.7). The average annual investment cost would reduce from €46.8m per year to €24.9m. 94 €23 million in 2010 prices. 95 The Low Volume Sealed Roads approach used in many developing countries can provide a long life sealed road (for example with an Otta Seal and Cape Seal combination for about €80,000 per km. 96 The unit costs shown in € per km are about the same numeric values as found elsewhere in U$ per km. 48 Table 4.7. Potential Reduced Cost of Prioritized Local Roads Activities Original Revised Cost/km ERR total cost Cost ERR Activity Length Km € % €m €m % Asphalt construction 206 372,538 1.6 76.8 19.2 6.3 Asphalt overlay 818 138,000 21.8 112.9 75.6 29.0 Surface dressing 531 84,000 23.7 44.6 29.9 31.5 Total/Average 1,555 138,000 21.3 234.3 124.7 26.9 Source: Authors analysis of data in Inventory of Local Roads in Kosovo, Final Report, 2010. 4.22. A system for the maintenance of local roads needs to be established. The National Government has made significant investments in the local road network since 2008, but the maintenance of these roads taken place in an ad hoc way. It is therefore important to establish a system for the maintenance of local roads in order not to lose the investment that was made. 4.23. Further investments will also be required in national and regional roads. The estimate designed to be consistent with the MTEF indicated an annual investment need of about €46.4 million97 for national and regional roads. 4.24. The pattern of demand for transport infrastructure will also influence where investments should be directed. Countries such as Kosovo with large agricultural and mining sectors, have high freight transport intensities of demand (ton-kilometers per € of GDP) and need an extensive and well maintained transport infrastructure. Whereas, with respect to passenger transport, largely rural countries (latest data for Kosovo urbanization is as low as 35 percent of the total population) have relatively low transport intensities (number of trips per capita) and have difficulty in sustaining fixed mode (rail) transport and rely far more on bus and informal transport. 4.25. Transport demand and the need to maintain existing infrastructure indicate that Kosovo requires an extensive but low capacity transport infrastructure. There are important exceptions to these generalities, such as the need for rail lines for mining activities and for transport infrastructure in key export corridors but they show where Kosovo’s future transport infrastructure investments would be most beneficial: (i) upgrading and maintenance of regional and local roads; (ii) upgrading and maintenance of viable sections of the railway network; and (iii) expansion of trade corridors for exported products. 4.26. Like its past investments, Kosovo’s future investment plans do not appear to be calibrated to meet the country’s needs. The authorities have recently accepted a bid to construct a large highway from Prishtina to the FYR Macedonian border at an estimated cost of around €750 million over three to five years98 (see Box 4.1). While this will contribute toward improving trade corridors, it would likely crowd out the investments in other transport infrastructure needed by exporters (as did Route 7 investments). It may also crowd out investments in railways that could support mining exports. Given that transport investments use scarce budget resources, investments in human capital – much needed in a young country – could also be crowded out. 97 €42 million at 2010 prices. 98 €600 million for the project and €150 million for expropriation. 49 Box 4.1. Economic Evaluation of Route 6 The estimated cost of the Route 6 highway is € 750m (including expropriation), or about €12.5m per km, which is about 47 percent more expensive than the high-cost Route 7. €150 million is estimated to be required for expropriation. Route 6 will eventually link Prishtina via Skopje to the port of Thessalonika through mountainous terrain in the south of Kosovo and on through Macedonia. The current travel time between Thessalonika and Prishtina (312km) is 4 hours 15 minutes. Distance between the port of Durres and Prishtina (355km) was over 7hr 43min, but was reduced to about 3hrs 30min using Route 7. The Route 6 highway was originally (in 2008) considered in two sections, one north of Prishtina and the other south. The EIRR on the northern part was estimated at 11.9 percent but was highly sensitive to traffic projections, with a plausible 20 percent reduction in traffic giving an EIRR of only 3.3 percent. The evaluation did not include any combination of sensitivity tests nor did it provide any switching values 99 for the key parameters. The economic evaluation of the southern section showed negative results even with the base parameter values. The sensitivity tests included a range of construction costs from plus 20 percent to minus 50 percent, even though there are no recorded instances of road projects having an outcome cost so much lower than the base value. A subsequent revised economic evaluation provided a positive result (EIRR of 8.25) for the two sections considered together. However, it is contrary to sound financial and economic practice to combine two non- interconnected projects in this way as it is possible to hide a poor investment by including it as a part of an otherwise sound investment. However, an alternative estimate can be found using the results of the latter Table 4.8. Sensitivity Tests for Main Project in Revised study by removing the assumption that Economic Evaluation of R6 vehicle operating costs and vehicle time costs would be reduced by 10 percent (Table 4.8). Taken together these could be considered equivalent to assuming 10 percent fewer vehicles use the road than forecast. This would reduce the EIRR by about 17 percent – that is from 8.25 to about 6.9 percent. If the impact of this 20 percent reduction in traffic is combined with Source: Update of the Traffic and Economic Evaluation of the Route the value provided for a 20 percent 6 Motorway, Draft Report, December 2012. increase in construction costs, the result is an EIRR of 5.8 percent.100 Given the impact of these outcomes on the credibility of the project, more analysis before the project was implemented could have resulted in a worthwhile investment, particularly when other evaluations show significantly higher economic returns to other transport projects and Kosovo needs to meet so many needs in other areas. Options could have included delaying the construction of Route 6 or building only sections with high EIRR. Railways 4.27. Maintaining Kosovo’s rail infrastructure will likely require investments worth between €14 and €21 million per year but given the scope for rail use, it is not clear this would be a good use of limited resources. Approximately €21,600 per km is required to prevent the deterioration of a railway infrastructure and keep it in good condition, including overheads. With 333 km of railway, around €7.2 99 A switching value is that which would change a positive to a negative outcome. Provision of switching values and sensitivity tests that include a combination of parameters are standard requirement for economic evaluations of WB transport investment projects. See WB Transport Note TRN7, January 2005. 100 There is an inconsistency in the revised Economic Evaluation report between the text describing the sensitivity tests and table that shows the results. The text appears to be more plausible so we use those estimates. 50 million would be required per year in Kosovo. However, Kosovo’s railway infrastructure is in poor condition and international experience suggests that these costs could double or t riple to between €14 and €21 million annually. Some suggested renovations could prove costly (Box 4.2). 4.28. In addition, most of the aging rolling stock operated by TRAINKOS will need to be replaced over the medium term. Refurbished locomotives are estimated to cost around €0.6 million each and TRAINKOS will need to purchase six at a total cost of €3.6 million if it is to maintain the current levels of operation. Box 4.2. East-West Railway Rehabilitation Proposal In addition to the investments required to maintain the current levels of service, a project is under preparation with support from the EU through the Western Balkans Investment Framework (WBIF) for the rehabilitation and upgrading of the East-West railway, including the railway sections between Podujevo (Serbian border) and Peja together with its southern branch lines: from Klina to Prizren and from Fushe Kosove and to Prishtina airport (Figure 4.3). The total length to be rehabilitated is around 190 km at an estimated c ost of €104m or about €0.54m per km – rather low considering the number of structures and stations that need reconstruction. For the airport connection about 1 km of line will need to be constructed to connect to the existing railway which passes the edge of the airport property. A study of this link reportedly indicated a positive economic outcome. International experience suggests that it usually requires at least 5 million airport passengers per year (and at rail share of at least 35 percent) to generate enough revenue to cover the operating costs of an airport railway.101 The current passenger level at the airport is about 1.5m passengers per year and increasing at an average annual average rate of about 10 percent since 2008. At this rate, the passenger level would reach 5 million in 2025 so this would appear to be the optimum time to open the airport rail link. Funding for the rehabilitation has not yet been found, neither has the way in which the railway might be operated as a public/private venture. Even with substantial public funding, operation of the railway will not be financially feasible to private operators with traffic growth based on current traffic levels. A careful study of the airport connection is necessary Figure 4.3. SEETO Comprehensive South East Europe to determine the optimal time for Rail Network investment. For investment in major rehabilitation of railway infrastructure a traffic intensity equivalent to at least 5 million tons per km per year (which could be a combination of freight and passengers) would be required. With less than 1 million tons of freight and 0.3 million passengers on the whole rail network, it is difficult to see how this rehabilitation was found to be economically feasible unless a major new input of freight traffic was considered. The vision for the future includes links from Prizren to the Albanian border and on to Durres port. Source: SEETO The only part of the SEE Core Railway network that would transit Kosovo is Route 10 which would link Corridor 10 transiting Serbia to its intersection with Route 11 and Route VIII near Skpoje. However, a map presented by Kosovo Railways in 2011 showed instead a link from Prizren (the terminal of the southern branch of the East West railway) connecting to Albanian railways and the port of Durres. A route analysis should be carried out for each route to determine its financial and economic viability. If the route is providing a social service and requires subsidization, then it is important to assess alternatives, particularly road transport. 101 Without access to the Feasibility study it is not possible to be more precise on the outcome of the economic and financial evaluations. 51 E. SOURCES OF ROAD MAINTENANCE FUNDING 4.29. Funds for road maintenance may be raised partly through road user charges to reduce the burden on the budget and the crowding out that appears to have resulted from the new highway construction. Currently Kosovo imposes vehicle registration fees (paid at the time of purchase of a vehicle), vehicle license fees (paid annually), and fuel excise duty but this revenue is not earmarked for road maintenance. There could be space to increase these charges and to allocate the resources in a way that will motivate collection efforts and ensure funds are used to help maintain local and regional roads. 4.30. Additional revenues could be raised from vehicle registration fees. The estimated revenue from vehicle registration fees was just under €16m in 2012, and given the projected increases in numbers of vehicles (about 6 percent per year) this would increase to about €25m by 2020102. If Kosovo’s average registration fee were to increase by 100 percent (to €50 for cars and €100 for trucks – still less than a quarter of the average EU rates103) the revenue would reach approximately €50m by 2020. It would be difficult to implement a 100 percent increase in one stage, so one option would be to increase fees by 20 percent each year from 2014 to 2018, resulting in an eventual doubling of the fee to an average of €65 per vehicle. The cumulative increased revenue between 2014 and 2020 would be around €170 million. 4.31. The most productive use of the registration fee revenue would be to fund road maintenance. Since most road maintenance is a charge on the municipalities for maintaining local roads, the revenue would best be allocated to them.104 There are several examples from countries that operate Road Maintenance Funds on how the revenue from vehicle registration fees can be allocated to funding maintenance of local roads and be made conditional on compliance with technical standards.105 4.32. As of November 2012, prices at the pump were lower than prices in the region, suggesting that there Table 4.9. Kosovo Average Fuel Price may be scope to raise transport fuel excise duties. The Structure, 2012 pump price of transport fuels in Kosovo is made up of the Price component € per liter, 2012 import price, the excise duty, the cost and profit Import 0.68 component of the distributor and the VAT on the final Excise 0.35 price (Table 4.9). It is not possible to distinguish gasoline VAT 0.16 from diesel consumption, so the same structure of prices Distribution 0.10 and average price has been assumed for both. Given that Total 1.29 median gasoline and diesel prices in SEE region are about Source: Authors analysis based on data from 10-15 US$ cent higher than the ones in Kosovo, there is Ministry of Economic Development. scope for raising prices Kosovo above the pump price (Table 4.10). To make any implementation of an increase in fuel excise duty acceptable, it is best done when volatile fuel prices are low. 102 See Annex 2 for a discussion of Road and Vehicle Registration Fees in the EU. 103 The average rates for the two smaller categories of car are about €220 per year, for the larger cars about €400 per year and for trucks about €1,100. 104 One option would be to make municipal allocations conditional on their providing a Road Maintenance Plan (RMP) and each year’s allocation being subject to a satisfactory technical (to ensure compliance with the RMP) and financial audit of compliance with the RMP in the allocation of the previous year. 105 For example, procedures for dividing funds between different road agencies, World Bank, http://www.worldbank.org/transport/roads/rf_docs/annex15.pdf 52 Table 4.10. Gasoline and Diesel Pump Prices, US$ Cents Per Liter, As of November 2012106 Country Gasoline Country Diesel Slovenia 198 Serbia 180 Montenegro 182 Albania 179 Albania 181 Slovenia 177 Serbia 178 Montenegro 175 Croatia 177 Romania 173 MEDIAN 176 MEDIAN 171.5 Macedonia 175 Croatia 170 Romania 170 Bulgaria 168 Bulgaria 169 Bosnia 162 Kosovo 165 Kosovo 160 Bosnia 159 Macedonia 155 4.33. Additional revenue of about €50 million could be raised. Assuming the same 6 percent increase in vehicles as in the estimate of increase in registration fee revenues and the same offsetting of increased kilometers per vehicle by reduced fuel consumption an additional revenue of €50 million could be raised 2014 rising to about €71 million by 2020, 4.34. Revenues from an increase in excise duty on fuel would best be used by the Ministry of Infrastructure for the maintenance of highways and regional roads. This is based on two premises. First, that revenue from user charges should as far as possible is allocated to the level of government that levies the charge. So given that vehicle registration and use fees are collected at the local level of government, their most appropriate use is for maintenance of local roads. Similarly, fuel excise duty is levied as a national government charge so is best used for the maintenance and development of highways and regional roads. The second is that user charge revenue is most appropriately used for current expenditures (maintenance) since it is collected from current users. Investments in new roads are mostly to expand capacity for the benefit of future road users and so are best funded by national taxes and allocated by the national government that has the social responsibility to provide for future as well as the current generation of road users. Box 4.3 discusses funding options for railway infrastructure maintenance. 106 International Fuel Prices, 2012-2013, GIZ, 2013 http://www.giz.de/expertise/downloads/Fachexpertise/giz2013- en-ifp2013.pdf 53 Box 4.3. Options for Funding Rail Infrastructure Maintenance It would be desirable to cover a higher proportion of the track maintenance costs from the revenue of track access charges. This would make it possible to use the current subsidy for track maintenance towards investment in improving the quality of the infrastructure for those sections of line where there is a business case to do so. The four main options for increasing track charge revenue are to: (i) increase rail traffic; (ii) increase the track user charges; (iii) reduce maintenance costs; and (iv) transfer maintenance responsibility of parts of the network to their sole user. With estimated costs of around €300 million to upgrade the whole network it seems unlikely that some combination of all four measures could be avoided. Increase in traffic: Low revenues, and share of revenues from access charges reflect the very low density of traffic on the network. Eliminating INFRAKOS’ deficit at current access charge rates 107 would require a level of traffic 3.5 times higher than it is now (about 3.5m tons per year and 1m passengers per year), or some combination that arrives at the same total. To increase maintenance expenditure to the level that would prevent further deterioration (at least €14m per year) the track charge revenue would need to increase by a fu rther 2 to 3 times. To cover the deficit on current maintenance expenditure would require a total traffic level of about 7m tons per year (or the equivalent in tons and passengers) and for the necessary maintenance about 14m tons. It is highly improbable that these levels of traffic can be realized in the foreseeable future, even with a completely rehabilitated railway, particularly given that the two major highways under construction will compete directly with the railway. Therefore, expecting to cover all infrastructure maintenance costs through revenues from higher traffic will not be feasible. Increase in track user charges: Without knowing what the current charges are it is difficult to know what increase might be possible. There are many different bases on which track charges can be assessed108, but most have some combination of costs per train kilometer and costs per gross ton kilometer of the trains. Whichever system is used, some allocation of the costs between passengers and freight is also made, since passenger trains incur quite different costs to freight trains. The Figure 4.5 shows the track Figure 4.4. Access Charges (€ Per Train Km) for a access charges in EU countries in 2008. Typical 960 Gross Ton Freight Train (2008) The railways with the higher charges are thought to be using average cost recovery as the basis of their charges. The average of these higher charges in 2008 was about €6.5 per train kilometer, and for the same countries in 2012 it was about €9.4 per train kilometer. Given the very low density of trains in Kosovo compared to the countries using average cost charging - the Czech Republic, Bulgaria, Lithuania, Estonia and Latvia - an average charge for Kosovo might be in Source: Charges for the use of rail infrastructure 2008, the order of €13 per train km. To International Transport Forum. generate revenue of €14m, this would require an average of about 20 trains per day.109 Marginal cost pricing does not seem realistic for Kosovo. Reduce maintenance costs: One way to reduce maintenance costs by about 9 percent would be to reduce high overhead share from 25 percent of the total to the more usual 18 percent. This would produce a cost saving of about 9 percent in actual costs. While a worthwhile reduction, it would not by itself deal with the difference between maintenance costs and revenues. A more significant way that railways often use to reduce their total maintenance costs is to suspend services and maintenance on parts of the network that have least traffic, especially where there are feasible alternatives for both freight and passengers. For many railways the decision 107 These charges are not available and cannot be estimated from the little traffic data available. As a minimum the number of traffic units (the sum of net tonne kilometers and net passenger kilometers) would be needed. 108 Charges for the use of rail infrastructure 2008, International Transport Forum 109 Although no data is available it is unlikely that the average freight train in Kosovo has 960 gross tons. The estimate of number of trains assumes each train travels 150km. 54 on closing a low traffic line is made when a large infrastructure investment is needed to keep its operations commercially viable (this is without imposing excessive axle load limits and speed restrictions). The East-West line is approaching this condition, not only with axle load constraints and speed restrictions but also with suspensions of passenger services over much of its length. If the Kosovo railway were reduced to just the North- South line, perhaps as the only one that is part of the SEE Core Railway network, the length of network to be maintained would be reduced to about 43 percent of the current total. If the maintenance costs are proportional to route length, then the average annual maintenance cost would decline from €14 million to a more manageable €6 million. Contract maintenance of parts of the network to their users: Some parts of the network only carry minerals from local mines. Since the cost of running these sections of the infrastructure is only incurred for the benefit of the mines, their full maintenance cost could be recovered as a charge to the mine, rather than cross subsidizing the cost from other users of the railway. This could be done through an operating concession, or more practically through an operating contract. The cost saving from this measure cannot be estimated with the data currently available, but the route length involved is unlikely to be more than 20 percent of the total, and much of this would be on the East-West line that would be closed under the previous cost saving proposal. Although the financial separation of railway infrastructure from operations helps to clarify which parts of the network are financially sustainable, implementation of any strategy to apply any of the suggested methods of increasing the recovery of maintenance would be advantageous. F. RECOMMENDATIONS 4.35. The resources that will likely be used to construct Route 6 could be put to better use. The €750 million estimated construction costs of Route 6 to FYR Macedonia are high per kilometer compared with similar roads in other middle-income countries. The economic rates of return are also much lower than alternative road construction and maintenance projects identified and other sectors discussed in this report have needs that are at least as great as those for a new highway. This report therefore recommends delaying the construction of Route 6 and using some of the resources saved to improving and better maintaining existing roads as well as investing it in other areas. 4.36. To guarantee the long-term maintenance of the road stock, road user charges and vehicle registration fees can be increased. To motivate collection efforts, road user charges could be collected and spent on road maintenance by municipalities. At the same time, centrally collected vehicle registration fees could be doubled between 2014 and 2020, raising an additional €170 million in total over the period and still remain among the lowest in Europe. 4.37. Comprehensive and consistent interventions to ensure sustainable road safety should be put into effect. Apart from fragmented applications and low enforcement of action plans, GoK should put necessary human and financial resources into action. 4.38. Serious consideration should be given to finding ways to make the railway network sustainable. This will likely entail a mix of reduced services and a focus on servicing companies that export heavy mining products. 55 CHAPTER 5. EDUCATION110 A. KEY MESSAGES  Kosovo can benefit from the ‘demographic dividend’ that comes with a youthful population only if it raises the skills of its workforce. Education is an essential part of Kosovo’s efforts to boost economic growth, increase productivity and wages and reduce its high unemployment. Large regional differences in education outcomes – even after controlling for socio-economic characteristics – suggest that further improvements are needed in order to reduce disparities of quality of education between municipalities. Partly as a result, skills shortages have become evident in some sectors of the economy. At the same time, the largest population cohort in Kosovo is about to join the labor market in the next few years. To help combat these challenges, the authorities could do much more to raise educational standards across the whole country both by allocating some additional resources to education and by using existing resources more efficiently.  Public expenditure on education has grown as a share of GDP over recent years but is still low compared with the average for Europe and Central Asia (ECA) and for that of middle income countries. On a per-pupil basis, Kosovo is even further behind other countries in the region and in the same income group. The large gap between current needs and resources means that well-targeted spending increases could be warranted. Fiscal constraints combined with projected demographic changes may also require the government to adjust its spending patterns within the sector to achieve efficient and equitable outcomes. The demographic change can also help Kosovo meet its challenges with smart investments.  Demographic change could make improving access to education easier to address with less new capital investment than would otherwise be required over the medium term. Although enrollment in public education is universal in primary and lower secondary education, there is a need to increase opportunities and access in pre-primary and upper secondary schools. The GoK has committed to achieve full coverage of pre-primary education (grade 0) by 2016. Although this is an ambitious objective, spending pressures may be eased as fewer new pupils enter the system due to demographic changes. Some classrooms will be freed as primary aged pupils are not all replaced by new pupils. Given that pre-primary and primary levels tend to be integrated in the same building, there will be limited need to invest in new classrooms, especially in rural areas where average class sizes are low. Fewer children will also make it easier to achieve full coverage in upper secondary schools. However, given that upper secondary institutions are not integrated with other education levels, there will be a need to make investments that will decrease average class size and improve the learning environment.  Policies that enhance teacher development rather than flat increases of wages are necessary. The government has raised salaries of all teachers and school management staff considerably in recent years. This expansion of the wage bill may have crowded out other expenditures that are critical to quality improvements. A preliminary analysis suggests that the wage level and structure is adequate to attract high performing graduates to the teaching 110 Chapter prepared by Lucas Gortazar and Flora Kelmendi with input on pre-primary education from UNICEF. 57 profession111. Therefore, efforts should now be focused on implementing a new teacher career system, and directing resources toward increased opportunities for teacher training and professional development, especially in light of the new curriculum. Wage increases should be de-politicized but targeted wage increases, linked to professional development and teacher evaluation - requirements for teacher’s climb up the career ladder should be considered.  While the improved capitation funding formula from central government to the municipal authorities has been implemented since 2008, the uneven implementation of the per-capita formula from the municipalities to schools leads to some inefficiency in spending. To improve expenditure efficiency by ensuring that schools with more pupils receive more resources, funding should follow the pupils. This is especially important in Kosovo with its high rates of rural-to-urban migration. Municipality-to-school funding or Per Capita Funding (PCF) was therefore piloted in 2009 and rolled out in 2011, with the intention of incentivizing an efficient use of resources. However, the implementation of PCF has been uneven. While increased school autonomy and transparency have been realized in some municipalities, the efficiency gains have been less visible. This is partly because the formula parameters have not always been correctly applied at the local level. Better monitoring by the Ministry of Education, Science and Technology (MEST) to ensure that formula parameters are respected by the municipalities could help to improve PCF implementation. At the same time, MEST can continue to support the municipalities in its implementation and facilitate exchange of experiences between the better performing municipalities with those that still face challenges using the formula.  Improved policy-making will require better data gathering and availability, particularly for higher education. At the higher education level, information on both education inputs and outputs is difficult to obtain. While the Treasury records spending information, data on staffing, enrollment by course, graduation rates, or employment outcomes are difficult to obtain. The data are more complete for pre-university education but there is still room for improvements in expenditure reporting at the local level by level and type of education, integrating more automatically assessment data and completion rates. While there are good data on the condition of school buildings, and teacher and pupil characteristics, as well as some data on exam results, these data are not linked. Part of the reason lies in the fact that different school identification codes are used in each database. Education Matters 5.1. Kosovo can benefit from the ‘demographic dividend’ that comes with a youthful population if it makes education a priority. Kosovo has a young population, with 38 percent of the population aged under 19, much higher than neighboring countries such as Albania (33 percent), Bosnia (25 percent), Serbia (23 percent), Slovenia (19 percent) or Montenegro (27 percent). This youthful population provides an opportunity to build a large, educated workforce with few of the fiscal and economic pressures that burden countries with aging populations – a problem faced by many eastern European countries.112 111 This analysis for the chapter was completed in 2013 prior to the announcement of flat wage increases for civil and public servants by the Government in early 2014. For teachers, the base salaries were increased by 25 percent as of April 1, 2014. Due to data and time constrains, this chapter does not include an analysis of the implications of these salary increases. 112 See World Bank, 2013, “Mitigating the impact of an aging population: Options for Bulgaria” and World Bank 2007 “From Red to Gray: The Third Transition of Aging Population in Eastern Europe and the Former Soviet Union”. 58 However, if Kosovo does not take advantage of its youthful population, the demographic dividend may become a demographic curse. 5.2. Improving and investing in education early and for all will be essential to develop a more productive economy and can disproportionately benefit poorer families. There is strong evidence that education and worker skills more broadly are essential for economic growth.113 One study found that education was the single most important and robust factor in explaining GDP per capita growth.114 There is also increasing evidence that investing in the first thousand days after birth, including early childhood education, provides larger returns to both individuals and society than any other level of education. Heckman (2008) finds that pupils from poor households benefit more from this type of intervention than those from wealthier households, implying that investment in early stage education is one of the most cost-effective and pro-equitable educational strategies that government can provide. Expanding access to early childhood education therefore not only boosts growth but also helps children from poorer families to break out of the “poverty trap”. 5.3. Raising educational attainment will also benefit Kosovo’s medium to long term fiscal position. Although there are short-term costs associated with investing in education and training, these expenditures should be viewed as human capital investment which will ultimately benefit the country. A better-trained workforce will generate higher tax revenues through increased personal income tax (PIT) and social security contributions. Moreover, a higher-skilled labor force can more easily adapt to changing economic trends and technologies, thereby reducing the risk of unemployment and allowing people to work for longer. Extending the years of workforce participation and increasing employment more broadly will not only increase tax collections but will also reduce government spending on social programs such as those associated with unemployment, poverty and retirement. Figure 5.1. Unemployment by Level of Education 2012 5.4. Skills shortages and unemployment are 70% key challenges that better education can help to National 60% Unemployment tackle. The World Bank 2009 Enterprise Survey Rate found that an inadequately educated workforce was a 50% major or very severe obstacle for over 14 percent of 40% firms surveyed.115 Large and fast-growing firms116 30% were particularly likely (around 25 percent) to report 20% this to be an obstacle, suggesting that skills shortages may constrain private sector growth. At the same 10% time, unemployment remains a severe problem for 0% Tertiary Secondary Vocational No School Basic Education Secondary General many segments of the population. Over 60 percent of workers without any schooling were unemployed according to the 2012 Labor Force Survey (LFS), over double the national rate (Figure 5.1). Workers with levels of education at upper secondary or below also had unemployment levels above the national average. Those with vocational training to the secondary level fared better in the labor market than Source: Kosovo Labor Force Survey 2012. 113 For example: Hanushek, Woessmann and Zhang (2011); Jakubowski et al., (2011); Krueger and Lindahl (2001). 114 Sala-i-Marti et al. (2004). 115 World Bank Business Environment and Enterprise Performance Survey (BEEPS) for Kosovo, 2009. Available from: http://www.enterprisesurveys.org/ 116 Defined, respectively, as having over 20 employees and have increased employment by more than 5 workers (the average) over the previous three years. 59 their colleagues with standard secondary education. 5.5. Demand for skilled labor is strong. Analyzing supply and demand using the 2012 LFS117 revealed that technicians and professional occupations, and tertiary and vocational education, are both highly demanded in the labor market (Figure 5.2 and Figure 5.3). There is a shortage of workers with university or ‘Higher Pedagogical Schools’ (HPS) credentials118, but a surplus of workers with general secondary education or only basic education. Regarding occupations, there is a lack of machine operators in the labor market, while occupations such as sales services or clerks are in excess supply. The most demanded skills tend to correspond to workers with tertiary education. Therefore, these groups may have a better chance of employment. Figure 5.2. Education Levels in Excess Supply: Figure 5.3. Occupations in Excess Supply, Surplus of Workers as % of Unemployed Surplus of Workers as % of Unemployed Tertiary University -9.26% Operators -20.19% High School (Tertiary) -5.63% Technicians -3.10% Vocational Secondary (4 -2.52% years) Professionals -0.69% Vocational Secondary (2 -2.47% Agriculture 1.29% years) Post university / Master -1.74% Managers 1.84% Primary 0.52% Craft 2.12% No school 1.60% Elementary 5.26% Upper Secondary (general) 5.03% Sales Services 5.65% Lower Secondary 14.80% Clerks 9.92% -15%-10% -5% 0% 5% 10% 15% 20% -30% -20% -10% 0% 10% 20% % of Unemployment % of Unemployment Source: Kosovo Labor Force Survey 2012 and WB Source: Kosovo Labor Force Survey 2012 and WB staff staff calculations. calculations. 117 Rutkowki (2009). This analysis uses the 2012 Kosovo Labor Force Survey to assess the education levels and occupations in excess supply by comparing a theoretical structure of unemployed (based on employment patterns) with the actual unemployment occupations and skills among workers. 118 Until 2003 Kosovo offered teacher trainings through higher pedagogical schools, a form of higher education lasting 2 years. From 2003 Higher Pedagogical Schools were transformed into three year University degree programs instead of the previous two year program. 60 Box 5.1. Overview of the Education System in Kosovo The education system in Kosovo follows standard international practice. It consists of nine years of compulsory basic education, which includes five years of primary education (normally starting from age six and lasts from grades 1 to 5) and four years of lower secondary (grades 6-9), supplemented by three years of non- compulsory upper secondary education (four years for some types of vocational upper secondary schools). The government also funds some pre-primary education through pre-primary classes in primary public schools and public (non-compulsory) kindergarten serving children from ages one to five and pre-primary school (grade 0) for children aged five to six years. Tertiary education is publicly provided at the University of Prishtina, University of Prizren (established in 2010) and University of Peja (formerly branch of University of Prishtina established in 2012). In 2013, the GoK established the University of Gjakova, Gjilan and Mitrovica – all three former branches of University of Prishtina. There is also the University of Prishtina in North Mitrovica, an institution serving Serb students in Kosovo and largely funded by the Serbian Government, which is not discussed in this chapter. Pre-primary education is expanding. A new Law on Pre-University Education was passed in 2011. It provides for compulsory pre-primary (grade 0 for children aged 5 to 6) education from school year 2015-16 provided that fiscal implications are evaluated and financial conditions are created. It also mandates the creation of a new curriculum and learning standards for both pre-school and pre-primary pupils. Finally it allows for pre-school education to be provided by public and private kindergartens, private pre-schools, and for pre-primary classes to be organized within primary schools or community based early childhood centers, licensed according to the provisions of the Law. Municipal authorities are responsible for pre-university education delivery in Kosovo, while the Ministry of Education, Science and Technology (MEST) play a policy setting oversight role. Municipalities receive earmarked funding to cover minimum standards for provision of pre-primary (including kindergartens), basic and upper secondary education. This budget allocation to municipalities is known as the Education Specific Grant (ESG) and provides funding for recurrent expenditures. It is determined by the pupil-teacher ratios, levels of education and types of schools in the municipality. After the budget ceilings are approved by the Grants Commission119, the municipal authorities are responsible for the allocation of lump sum budgets to schools following a municipal-to-school formula based on per capita (per pupil) funding. Public pre-school institutions and kindergartens are also financed by parents’ contributions. Furthermore, munic ipalities can use own source revenues to top-up education spending. The capital investments budget is managed by the MEST. B. ENROLLMENT TRENDS ARE MOVING IN THE RIGHT DIRECTION 5.6. Nearly a quarter of Kosovo’s population is enrolled in school. Kosovo’s education system serves 468,000 pupils enrolled in all levels of education in 2012/2013, nearly a quarter of its population (see Table A12 and Figure A9 in Annex 3). Enrollment in primary and secondary schools was around 392,000 in 2012, down from its peak of 419,000 in 2008 but in line with demographic trends. Following a baby boom (late 1980s through early 2000s), there was a slight decrease of births after 2002, resulting in a 20 percent decline in the number of primary school pupils between 2004 and 2012. Upper secondary 119 The Grants Commission is a key institution in the system of intergovernmental transfers in Kosovo, and has been part of legislative and institutional arrangements since 2001, according to the Law on Public Financial Management and Responsibilities (LPFMA). The Grants Commission currently consists of the Prime Minister or his representative, the Minister of Economy, the Minister of Finance, the Minister of Local Government, a minister from a line ministry (designated by the government), the Chair of the Committee for Budget and Finance in the Kosovo Assembly, and four mayors, one of whom is from a municipality with a minority population. The functions, mandate and composition of the Grants Commission were further elaborated in the 2008 Law on Local Government Finances. The Commission has clearly defined responsibilities in the areas of municipal finance, particularly regarding the government grants. It is a forum that facilitates communication and dialogue between central and local governments. 61 school enrollments increased by 50 percent during the same period, due partly to demographic trends and partly to the fact that gross enrollment rates have increased since 2009120, and remained at high levels for basic education.121 Although the number of students in Kosovo’s schools has increased substantially (especially in lower and upper secondary), it will need to keep increasing to reach universal coverage. 5.7. Although Kosovo lags behind the economic development of most countries in SEE, enrollment rates are similar to comparator countries for all educational levels. Kosovo’s gross enrollment rates are similar to other SEE countries: in primary and lower secondary education, these rates are between 95 and 100 percent - higher than in Bosnia and Herzegovina or Macedonia (see Table A15 in Annex 3). Kosovo has reached full gender parity in basic education, but needs to increase female enrollment in upper secondary. In recent years, gross enrollment rates (GER) have been virtually identical and almost universal for boys and girls in basic education, ranging from 97 to 101 percent (Figure A.10 in Appendix 3). However, enrollments in upper secondary are neither universal nor gender-balanced, although there have been recent improvements. The GER for girls in grades 10-12 increased from 81 percent to 88 percent between 2010 and 2012, while for boys it grew from 88 percent to 92 percent 122. Girls’ rates of transition from one grade to the next drop significantly in the 10 th grade, where the transition rate was 84.9 percent for girls and 96.8 percent for boys in 2012123. However, completion rates for girls once in upper secondary education are higher than for boys (percentage of girls in upper secondary in the last years was on average 44.5 in the 10th grade, but 46.5 percent in the 11th grade), which implies a selection of girls to carry on their studies in secondary schools. This suggests that there is a need to increase transition rates of girls from lower secondary to upper secondary and improve completion rates of boys in upper secondary. 5.8. Children from poor and extremely poor households are less likely to be enrolled in education according to the 2011 Household Budget Survey. Non-enrollment of children aged 10 to 15 years (who are expected be in basic education) in households defined as poor, according to the national poverty line, is 2.9 percent, and rises to 3.5 percent for children in extremely poor households (Figure 5.4 and Figure 5.5). Differences are even larger for higher levels of education. While around a quarter of all 16 to 20 year olds are not enrolled in higher education this increased to over 30 percent for children from poor households and 40 percent for children from extremely poor households. This suggests that efforts to increase enrollment would benefit from targeting to offer children from poorer households a way to break out of the cycle of poverty. 120 Note that part of the changes may have been related to inaccuracy of data until 2009. Anecdotal evidence suggests that municipalities are known to have increased the numbers of students for purposes of financing, which could imply that enrollment rates in primary have not increased as much as suggested by data. This inflation may have been alleviated in the last years with the new improvements of financing formula and increase of reliability of data (EMIS). 121 From now on, basic education is defined as the aggregation of primary and lower secondary education. 122 Source: UNICEF-MEST Report for 2012-2013. 123 Source: EMIS 2013, MEST. 62 Figure 5.4. Non-Enrollment Rates by Figure 5.5. Non-Enrollment Rates by Household Household Poverty Status: Ages 10-15 Poverty Status: Ages 16-20 4.0% 45% 3.5% 40% 3.0% 35% 30% 2.5% 25% 2.0% 20% 1.5% 15% 1.0% 10% 0.5% 5% 0.0% 0% Population Poor Extreme poor Population Poor Extreme poor Source: 2011 Household Budget Survey and WB staff Source: 2011 Household Budget Survey and WB staff calculations. calculations. 5.9. Ethnic minority pupils are only partially reflected in official data, and are less likely to be enrolled. According to the official data system, EMIS124, the school population consists mainly of ethnic Albanian pupils (96 percent), and other ethnic groups such as Serbs, Turkish, Bosniaks, Roma, Ashkali, Egyptians, and Gorani. Due to the fact that some municipalities with Serb majorities report little or no information to EMIS, the data from EMIS shows a Serbian pupil population only at 0.12 percent. However, population data indicate that 5.7 percent of the total population is Serbian.125 This disparity regarding Serb minorities is likely explained by underreporting of data to EMIS. However, it should be noted that there is no evidence of underreporting for other minority groups. Around 11.3 percent of non- Serb minority children aged 10-15 years were not enrolled in school according to the 2011 Household Budget Survey (HBS), compared with a non-enrollment rate of 1.3 percent for the total population and just 0.7 percent for Albanian children (Figure A.12 and Table A.18 in Annex 3). All ethnic Serb children in the sample were enrolled. There were also large differences at higher levels of education. Nearly 60 percent of non-Serb minority children aged 16-20 years were not enrolled in education compared with around a quarter for the whole population and 23 percent for Albanians. A better understanding of the impediments ethnic minorities face in education would help the authorities to develop policies to increase opportunities for all groups.126 124 Education Management Information System, MEST. 125 Population distribution has been constructed using data from 2011 Census and OSCE (Organization for Security and Co-operation in Europe Mission in Kosovo 2010). See Annex 1 for more details. 126 The survey classifies together “Turkish, Bosnian, Montenegrin, Ashkali, Roma and Other” so it is not possible to distinguish between these groups. The sample size for Serbian children is problematic with a total of just 23 observations, meaning these results should be treated cautiously. 63 Box 5.2. Pre-School and Early Childhood Education The social and economic returns for investing in early childhood education (ECE) are highest at young ages. In addition, investment in ECE is particularly beneficial for children from poorer families, as they are less likely to receive sufficient support in the home with respect to social, communication and cognitive development. Despite this, at 71.2, the GER for pre-primary (Grade 0) is low compared with close to 100 for primary education, although it is higher than other SEE countries (see Table A.15 in Annex 3). Efforts to expand pre- primary education would lead to societal-wide benefits in the medium to long term in terms of higher employment and worker productivity and contribute to a positive fiscal dynamic. Recognizing the importance of ECE, the GoK has committed to achieve full coverage of pre-primary education (grade 0, ages 5-6) by school year 2015-2016, provided that fiscal implications are evaluated and financial conditions are created. Since many pre-primary classes are housed in existing primary schools, and changing demographics will likely free some classrooms, pre-primary education could be expanded with minimal capital investments in some areas. A recent UNICEF report127 puts the recurrent costs of making pre-primary education universal at around €600,000 per year (0.01 percent of GDP) .128 C. EDUCATION EXPENDITURES ARE NOT STUDENT-FOCUSED 5.10. Public expenditures in education have grown in recent years and are catching up with the regional averages. Public spending in education grew steadily from 3.3 percent of GDP in 2007 to 4.1 percent in 2012. Despite this growth, Kosovo still spends less than the ECA average (4.6 percent) or the upper middle income country average (5 percent). Among SEE countries, Kosovo spends more than Albania (3.3 percent) and Macedonia (3.5 percent) but less than Croatia (4.4 percent) and Serbia (4.7 percent). Countries with younger populations, such as Kosovo, tend to spend more on education as a proportion of total public expenditure than those with older populations. Kosovo spends around 14 percent of total public expenditure (TPE) on education. This puts Kosovo in line with other middle income countries with similar age profiles129. Figure 5.6. Expenditures on Education and Figure 5.7. Total Public Expenditure and Public Young Population Expenditure on Education, %GDP 25% Total Public Expenditure as a percent Thailand 50% Serbia Hungary Public Expenditure in Education as 20% Guatemala Ukraine Croatia EcuadorPoland Macedonia, 40% FYR Lithuania percent of TPE Romania Belarus Kosovo Ecuador Turkey Bulgaria of GDP 15% Macedonia Jamaica Slovenia Lithuania 30% Albania Armenia El Salvador Kosovo Bulgaria Poland Jamaica Paraguay Armenia Georgia Romania Albania Thailand 10% Croatia Dominican Serbia 20% Dominican Hungary El Salvador Paraguay Republic Republic Georgia Guatemala 5% 10% 10% 20% 30% 40% 50% 1% 3% 5% 7% Percentage of Population under 19 Public Expenditure on Education as a percent of GDP Source: World Bank EdStats, IMF WEO, Boost Kosovo, Kosovo Census. 127 UNICEF. Joining Hands: Better Childhood. 2011. Prishtina, Kosovo. 128 With an average class size of 20 pupils. See Table A.13 and Table A.14 in Annex 3 for more details of kindergarten and pre-school enrollment. 129 See Table A.8 in Annex 3 for a summary of comparator countries. 64 5.11. However, due partly to its enrollment success, Kosovo spends less per pupil than most comparator countries (and less than all other countries in SEE), both in primary and secondary education (Figure 5.10 and Figure 5.11). Another measure of expenditure shows that, due to its (comparative) success in enrollment, expenditure on education may still be insufficient. Pre-university spending per pupil is only 13 percent of per capita income in both primary and secondary education, lower than the average for upper-middle income countries, which are at 16.2 and 17.8 percent, respectively. This gap is higher still when compared to the average EU15+3130 countries (19.5 and 22.0 percent) or other EU member states such as Poland (23.5 and 24.6 percent) or Slovenia (26.9 percent for both primary and secondary education). Figure 5.8. Public Expenditures in Education as Figure 5.9. Public Expenditures Per Pupil as a a Percent of GDP for Kosovo, European, and % of GDP Per Capita, Pre-University (2011 or Other Comparator Countries (2011 or Latest) Latest) 10% 40% 9% 35% 8% 30% 7% 25% 6% 20% 5% 15% 4% 10% 3% 5% 2% 0% Georgia Armenia Latvia Dominican Rep. Kosovo Lithuania Poland Guatemala El Salvador Romania Bulgaria Ecuador Upper-Middle Income Slovakia Jamaica Serbia Paraguay Czech Republic EU 15+3 Hungary Slovenia Thailand 1% 0% Paraguay Guatemala Russia Romania Albania Croatia Dominican Rep. Bulgaria Serbia Moldova Ecuador Jamaica Georgia Slovakia Poland Armenia Slovenia Macedonia Thailand Kosovo El Salvador Upper Middle Income Source: World Bank EdStats and Kosovo BOOST. Source: Kosovo BOOST, World Bank EdStats, and World Bank (2013). 5.12. In tertiary education the spending composition seems to be more aligned with OECD averages. Available data show that wages comprise 67 percent of spending, which is aligned with the OECD average. However, capital expenditures comprised only 3.2 percent in 2012, lower that the OECD average of 9 percent in 2009. Further analysis of tertiary education budget expenditures, including those from the private sector, was not possible due to data limitations. 5.13. Expenditure growth at different levels of education has been similar; however expenditures on teacher wages have driven the increases. Increases in education expenditures between 2007 and 2011 benefited primary, secondary and tertiary education to a similar extent (Figure 5.12). However, the large increase in teacher wages in 2011 drove education spending from 2.8 to 3.3 percent of GDP (Figure 5.13). Further, once the government eliminated three-shift schools in 2010, capital expenditures dropped. Overall, in the education budget, capital expenditures, subsidies, utilities, and goods and services accounted for spending worth approximately 0.8-1 percent of GDP between 2007 and 2012. 130 Austria, Belgium, Denmark, Estonia, Finland, France, Germany, Greece, Ireland, Italy, Latvia, Lithuania, Luxembourg, Netherlands, Portugal, Spain, Sweden, United Kingdom. 65 Figure 5.10. Expenditures by Level of Figure 5.11. Education Expenditures by Education, % GDP economic Category, % GDP 4.5% 4.5% 4.0% 4.0% 0.3% 0.2% Capital outlays 0.4% 0.4% Other 0.1% 0.3% 0.4% Expenditures 3.5% 0.3% 0.4% 0.4% 0.1% 3.5% 0.4% 0.7% 0.7% 0.1% 0.2% 0.1% 0.1% 0.7% 3.0% 0.1% 0.1% 0.4% 0.5% Subsidies and 3.0% 0.7% Tertiary 0.5% 0.5% transfers 0.8% 0.8% education 2.5% 2.5% 0.6% 0.6% Utilities 2.0% 2.0% Grades 10- 3.3% 3.3% 1.5% 12/13 1.5% 2.9% 2.8% 2.5% 2.6% Goods and 2.1% 2.1% 2.3% 2.3% 1.0% services 1.0% Grades 0-9 0.5% 0.5% Wages and salaries 0.0% 0.0% 2009 2010 2011 2012 2007 2008 2009 2010 2011 2012 Source: Kosovo BOOST. Source: Kosovo BOOST. 5.14. Teacher salary increases may have crowded out other spending.131 Spending on wages under the education budget increased by over 25 percent in real terms between 2009 and 2012, taking wages from 85 percent of total spending on basic education (grades 0-9) in 2009 to 92 percent by 2012.132 At the same time, spending on all other items fell. Capital spending fell by over 60 percent in real terms, declining from 8 to 3 percent of the total. Goods and services spending remained constant in nominal terms but declined from 6 to 5 percent of the total. Secondary education suffered from a similar budget squeeze. Spending on non-salary recurrent items was low in 2012 compared to OECD or regional countries. On average, OECD countries spent 22 percent of education budget on non-salary recurrent items, and about 8.7 percent on capital expenditures.133 In Europe, Slovenia spent 19 percent of total expenditures on non-salary items and 8 percent on capital expenses, while Bulgaria and Romania spent 26 percent on non-salary items, and 6 and 4 percent on capital expenditures respectively.134 In Kosovo, increases granted to the education sector have been devoted almost entirely to salary increases, and for the most part have not been directed to other quality enhancement investments. D. TEACHER SALARIES HAVE DRIVEN SPENDING INCREASES IN EDUCATION 5.15. Recent changes to the teacher salary structure aimed to improve teacher quality. In October 2008, the pre-university “teachers’ differentiated salary system” came into effect. This was part of a comprehensive teacher licensing and professional development effort intended to improve the teaching system’s ability to attract and retain qualified staff. Prior to the reform, teacher salaries were uniform, differentiated only by the grade taught (i.e., with different salaries for teachers at pre-primary, primary and secondary grades). Under the reform, differentiated pay scales were introduced based on qualifications, grade level and experience. As a result, all teaching staff, other than those who were unqualified and those with less than one year of experience, received a pay increase.135 The 2009 budget 131 See Table A.19 in Annex 3 for further details of the teacher salary structure. 132 As a percentage of recurrent budgets the figures are 94 percent and 92 percent for primary and secondary respectively. 133 OECD, Education at a Glance 2012, Table B6.2. 134 World Bank EdStats. 135 The base salary increased by 10 percent to 35 percent depending on qualifications: teachers with (i) 5-year secondary teacher school received a 10 increase; (ii) 2-year Higher Pedagogical School received 19 percent; (iii) 66 for education grants (which finances teachers’ salaries at the municipal level) went up by a little over 20 percent to accommodate the salary increase alone. 5.16. The reform of teachers’ career development and remuneration was a welcome development; however, the politically-motivated increases that followed in 2011 distorted the reform. In keeping with electoral promises, the government increased the “base salaries”136 of all teachers by 50 percent in 2011. While the differentiated salary structure was kept, the decision was not aligned with the reform principles of future salary increases being linked with performance and professional development. Moreover, increases were granted to unqualified or beginner teachers, which was not the case with the 2008 increases. As a result, the difference in base salary for an unqualified/beginner teacher and a teacher with bachelor’s degree137 was reduced to 18 percent compared with 27 percent in 2008. Box 5.3. Improving Education Outcomes by Improving Teacher Selection Rethinking teacher selection policies as part of future reforms could increase the quality of education. In the last few years a strong consensus has developed stressing the importance of high quality teachers. In quantifiable terms, research by Hanushek (2010) showed that a good teacher can teach 1.5 years of educational content in one year, while a bad one teaches only 0.5 years. Moreover, Hanushek (2009) estimated that the impact on overall student learning of replacing the worst 10 percent of teachers with teachers of average quality would increase student achievement (measured by exam results) by 0.5 standard deviations*. Although Kosovo has recently made substantial improvements in terms of teacher accreditation, there are still many teachers that lack adequate certification. Therefore, to increase average teacher quality, a range of policies need to be developed from selecting and retaining the best professionals, to dismissing less effective teachers in the system through early retirement and other policies. * Teacher quality is measured on the basis of teacher performance in obtaining gains in student achievement. This is done through distinguishing the effects of teachers from the selection of schools by teachers and students and the matching of teachers and students in the classroom. For this piece of research, a teacher at the mean of the quality distribution is compared to a teacher 1 standard deviation higher in the quality distribution (84 th percentile), labeled as a “good teacher”. A similar calculation is performed to define and label “bad teachers”. See figure A.13 in Annex 3. Bachelor degree holders received 27 percent; and (iv) Master’s or Ph.D. degree holders received 35 percent. Additionally, for each year of working experience teachers received 0.003 percent of their base salary. 136 “Base salaries” are the salaries that teachers received in 2007, before an y adjustments for inflation or other increases followed. Salary adjustments are not always made with relation to teachers’ current salaries but to their base salary. 137 As per the Law on Pre-University Education (2011), all new teachers must now complete at minimum a Bachelor’s Degree to be able to join the teaching force. 67 5.17. In the short run though, it appears that teachers were incentivized to upgrade their qualifications, albeit through a government subsidized program. When comparing the composition of the teaching force by qualifications in 2008 with data available in 2013, it appears that Figure 5.12. Distribution of Teachers by Pre- teachers have invested in upgrading their pre- Service Qualifications, 2008 and 2013 service qualifications (Figure 5.14). Over this 12,000 time, there was a large reduction in teachers with Higher Pedagogical School (HPS) 10,000 education and a corresponding increase in 2008 those with a Bachelor degree. This shift can be 8,000 attributed to the initiative to fund the upgrade 2013 of pre-service qualifications of all teachers 6,000 with HPS to a Bachelor degree by the end of 2015. The number of unqualified or beginner 4,000 teachers fell by around 20 percent. Due to 2,000 limited data it is not possible to assess separately the changes in the numbers of 0 teachers that are unqualified (defined as those Unqualified 5-year 2-year BA (3-year Masters and who do not have an appropriate pre-service / Beginner secondary Higher & 4-year) Ph.D teachers teacher Pedagogical degrees degrees degree), and beginner teachers (defined as school School those with an appropriate pre-service qualifications but without the minimum Source: Kosovo EMIS experience). Data for 2013 show that out of the total number of teaching staff without qualifications (2,474), about 1,178 have two or more years of experience – these are likely to be unqualified but continue to teach.138 5.18. Following the reform process and the 2011 increases, teachers’ salaries became comparable with other sectors. The average teacher’s net salary was €347 per month139. This was only about 5.7 percent below the average net salary in the public sector, and about 3 percent below the average salary in health sector.140 5.19. The best use for any additional funds for teacher salaries would be to provide incentives for improved performance and professional development rather than politically motivated increases.141 The National Teacher Licensing Council has developed a professional development and performance evaluation mechanism that will provide teachers with an opportunity to strengthen their qualifications and move up the career ladder. The system, which is already in place, grants a temporary license and a regular license to teachers. Those on temporary licenses (about 14 percent of the current teaching force that are beginner or unqualified) will need to meet qualification and training criteria to receive a regular license or risk losing the right to teach.142 At the same time, teachers on a regular license can be promoted through 138 The figure does not include Serbian teachers since no data are available on their qualifications. 139 This was in 2013 before the wage increase of 25 percent granted in April 2014. 140 Kosovo Agency of Statistics, General Statistics, Quarterly Bulletin, Average Monthly Paid Net Wages in the Budget sector by year, April 2013. 141 In fall 2013, the government had promised another wage increase by 50 percent. In March 2014 GoK decided to increase wages by 25 percent applicable from April 1 st 2014, and the decision has had a negative impact in implementation of reform for career development of teachers. Furthermore, because the wage increase was done in a similar way as in 2011, the wage structure has pressed further the difference by qualifications pre-service. 142 In March 2014, GoK decided to increase wages of all civil and public servants by 25 percent, applicable from April 1st 2014. As noted earlier, due to time and data constrains, further analysis on the implications of these increases could not be included in this report. 68 five career grades. Teachers need to take training and receive at least one positive performance evaluation in a period of five years to move from one grade to another.143 Additionally, the teacher career reform has made it mandatory for teachers to attend at least 70 percent of “core trainings”. Linking future salary increases to training and education of the teaching force would reinforce these reforms designed to enhance quality. Granting politically-motivated across-the-board increases, on the other hand, would risk undermining them. E. CAPITATION FINANCING AND MUNICIPAL COMPARISONS 5.20. Capitation funding was designed to increase the adequacy and efficiency of spending on education. Since 2011 per capita funding to individual schools (the municipal-to-school formula) has been extended to thirty municipalities144 (see Figure 5.15 for a schematic of school funding). The aim was to incentivize more efficient and transparent use of resources by municipalities. Since there are a large number of schools with few pupils due to demographic changes and rural to urban migration, by ensuring that municipalities fund schools in proportion to the number of pupils, some smaller schools would become unviable. These schools, which were draining resources from the majority of pupils, could be rationalized by municipalities, which have this responsibility. The reform was also accompanied by an increase in school autonomy that delegated financial management authority to schools, though implementation is lacking. Box 5.4. Education Financing Reforms In 2009, the government piloted a pre-university school financing reform which included the revision of the state formula for pre-university education, the introduction of a municipal-to-school formula and the extension of school financial autonomy. The implementation of the municipal-to-school formula started with three pilot municipalities in 2009, another ten municipalities in 2010 and was rolled out nationally in 2011. Also in 2009, the parameters of the state formula for education specific grant to municipalities were revised to improve the allocation of resources to pre-university education. The revised formula lowered the pupil-teacher ratios (PTRs) for pre-school and vocational education, and provided additional funds for schools in mountainous areas. It also provided norms for administrative support staff and additional funding for maternity and sick leave. In calculating the funding, enrollment figures are converted into full-time equivalent teachers using fixed PTRs norms for pre- primary, primary and secondary education, and for majority and minority students. The introduction of capitation based funding to schools was aimed at increasing the role of municipalities in management and allocation of resources, and encouraging individual school directors to be more autonomous accountable for their management of the school predetermined budget. Through this process, the municipal-to-school formula should be used to determine the ‘lump sum’ allocation of school budgets, which should be dependent on two factors: (i) the number of students in each grade and other characteristics of the school, such as the school locality, type of heating, etc.; and (ii) the values of the parameters of the formula which the municipality selects, such as the pupil teacher ratio for each category of pupils.145 After the lump sum school budget is allocated, the School Governing Board and School Director plan the school budget in categories, and submit them to the municipality for approval. 143 According to Administrative Instruction No. 5/2010, a teacher career license is valid for 5 years. To extend the license, teachers need, at minimum, to have a satisfactory performance evaluation and to have completed at least 100 training hours, of which 70 percent in core in-service training programs and 30 percent in optional courses, in a period of five years. The criteria for advancement to a higher license require at least 300 hours of teacher training over a period of five years. 144 For reasons already explained, municipalities with majority Serbian population were not part of the process. 145 See Annex 3 for a description of the capitation parameters. 69 5.21. The implementation of the municipal-to-school formula Figure 5.13. School funding and the school financial autonomy to date has been uneven, however. In 2011 only some municipalities used the formula to determine the school budgets for 2012 while all thirty municipalities were reported to have used the formula in 2013 to determine their school budgets for 2014.146 However, despite basing school allocations on the formula, implementation monitoring carried out by the MEST in 2012 showed that in only three municipalities was the school budget compliant with the formula allocation for 2012. The evaluation found that implementation issues relating to municipalities’ limited capacities to use the formula included: short time-frames for planning school budgets, and in some instances, limited political will on the part of municipalities to close down financially unviable schools.147 MEST monitoring of the Per-Capita Financing (PCF) implementation showed that sixteen out of thirty municipalities had implemented elements of the school financing autonomy. In these places, the allocation of budgets to schools combined with school autonomy has provided a greater budgetary autonomy to schools and more responsibilities to school management to manage resources, and to work closely with communities on school development planning, budgeting and implementation. 5.22. An analysis of expenditure data by Figure 5.14. Pupil to Teacher Ratio and Population municipalities suggest that per pupil Density, by Municipality (2012) spending is higher than the allocation received by the central government. Per pupil spending on basic education (grades 0 to 9) by municipality was on average €50 more than the education grant allocated based on the formula.148 Municipalities appear to use their own source revenues for education, as they are entitled to provided that they adhere to the standards set by MEST in, for example, the class size. Anecdotal evidence suggests that municipalities do not always obtain the amount prescribed by the education specific grant, therefore they additionally use more own-source revenue to top-up financing for education. Given that the education specific grant formula uses the number of pupils enrolled in the previous school year, rather than the estimated Source: EMIS and 2011 Census. number of pupils, municipalities may need to 146 MEST commissioned software dedicated to the municipal formula to facilitate the work of municipalities in allocating the resources to schools. 147 A MEST official reported that only three municipalities initiated the process of closing small schools. 148 Per student final spending was calculated based on financial executed data from BOOST by municipalities, and EMIS data for students, while Education Grants used data from EMIS for number of students and followed the capitation parameters. The average was computed only for the municipalities in Figure A8 in Annex 3 which excludes from the sample Serb majority populated municipalities. 70 allocate additional resources to cover the cost of new pupils, especially for Grade 0 which has seen an expansion in recent years. 5.23. There is a large variation in pupil-teacher ratios across the municipalities.149 On average, Pupil-Teacher Ratios (PTRs) in urban schools are above the national average while in rural areas they are lower than national average of 17.2 for pre-university education. Part of these variations can be explained by geographical conditions such as population density (Figure 5.16). The remaining variations could be explained either by the presence of a second language in the school, requiring additional teachers, or by remoteness (additional budget is allocated to smaller schools at an altitude above 700 meters, a proxy for remoteness). However, this does not explain all variation and it appears that municipalities choose to over-fund under-enrolled rural schools at the cost of under-funding overcrowded urban schools, a risk that was noted in the previous World Bank Public Expenditure Review.150 The implementation has therefore been patchy and limited progress has been made by the municipalities to make the school networks more cost effective, and rationalize resources. Additionally, there is great reluctance to close small village schools and use alternative approaches, such as bussing students to nearby schools. 5.24. There are likely intra-municipality subsidies, with small class sizes in some schools being maintained at the cost of higher class sizes within the same municipality. Around 369 schools (or 30 percent of total 1,202 schools) had fewer than 50 pupils enrolled in 2010, of which 114 (or 31 percent) are at high altitude.151 Additionally, another 111 schools (or 9 percent of total) enrolled from 51 to 100 pupils, of which 39 (or 35 percent) were at high altitude. There could therefore be room for school network optimization, especially in basic education. A more in-depth analysis of school level data for each municipality would be required to identify potential areas for efficiency gains and municipalities would need to ensure continued access to education for children in remote areas by providing school transport. The financing formula could add distance of travel for pupils as a funding parameter to ensure that additional resources focus more on the remoteness. 5.25. Class sizes fell slightly after the introduction of the formula, though they continued to be high. An analysis of the data for the thirteen municipalities that were the first to use the municipal-to- school formula152 showed that average class sizes continued to be high even after the introduction of the formula. In school years 2011/12 and 2012/13, the average class size was respectively 22.4 and 22.2 for Grades 1-9, and 32.3 and 31.1 for Grades 10-13. These are much higher than the EU average of 20 and 21.9 respectively, and much higher than Slovenia (18.5 and 19.8) and Hungary (20.8 and 21.8). 5.26. In total, the ratio between pupils and teachers has decreased slightly and now compares with the average ratio for upper-middle income countries. From 2008 to 2011, pupil-teacher ratios decreased in Kosovo from 18.1 to 16.9 in primary education and from 20.1 to 19.7 in secondary education. In 2011, the ratio was above upper-middle income countries’ average for upper secondary (15.4) and below the average for basic education (18.5). 149 See Table A.10, Table A.11, Table A.17 and Figure A.11 in Annex 3. 150 World Bank, 2010, Kosovo Public Expenditure Review. 151 A proxy for ‘remoteness’. Defined as being above 700 me ters above sea level. 152 See discussion on Education Grants and Per Capita Funding in Annex 3. 71 F. EDUCATION OUTCOMES 5.27. Measuring educational outcomes in Kosovo is a challenge due to a lack of data. Kosovo does not participate in international student assessments such as PISA. There is also a lack of data on tertiary education graduation rates. Exams taken during pre-university education can be accessed but it is difficult to match this data with other school characteristics. School coding for exams at grades 9 and 11 does not match school coding used elsewhere in the MEST, and exam data may not be consistent throughout years. In addition, school codes are determined by municipalities, which do not necessarily use consistent coding over years (making it difficult to track school level exam grades over time) and do not provide the names of schools. This makes it problematic to link pupils’ grades with regional, school, pupil and teacher characteristics. Nonetheless, a 5th grade assessment was undertaken in 2010 as a pre-cursor to entry in PISA, provides some indication of education outcomes such as age, gender or ethnicity. 5.28. A more accessible and functional higher education management and information system is needed. For exam results, school codes are determined at the municipal level and the MEST is not always able to identify the schools consistently across years, making it almost impossible to link regional, school, pupil and teacher characteristics with exam results. This report makes, for the first time, an effort to link these. However, since this information will be critical for sector planning and direction, better efforts should be made to systematically link and analyze these data. At the higher education level, even dispersed data are unavailable. An accessible and functional higher education management and information system that incorporates both public and private sector data and regular reporting on higher education institutions could be put in place. Additionally, pupil level data (including exam results) should be collected for all levels to better assess the impact of policies as well to provide better information to all the relevant stakeholders. 5.29. Evaluating pupils’ core skills will enable the government to foc us on developing and implementing policies to raise educational and skill standards. To this end Kosovo’s first assessment of core competencies took place in April 2010. Modeled on the PISA methodology, a sample of 6,000 (19 percent of the total cohort) 5th grade pupils throughout Kosovo sat a test, which evaluated knowledge of and skills with their mother tongue and mathematics.153 The methodology of this assessment was aligned to international exams154, and consisted of determining not only the achievement level of fifth grade pupils, but also set out to track the factors that affect the differences in achievement, such as gender, municipality, or personal characteristics. 5.30. Exam performance differed significantly by municipality and gender (Figure 5.15). Results showed that girls and urban pupils performed better in reading while no such major differences were registered in math scores (Figure A.6, Figure A.7 and Table A.9 in Annex 3).155 Controlling for individual and school factors such as income, parental education, or geographic settlement, and peer effects, girls obtained scores on average 4.8 points higher (equivalent to 0.25 standard deviations) than boys in reading. Urban pupils scored on average four points more than their rural peers (equivalent to 0.2 standard deviations), after controlling for individual and school factors. In the case of the math test, the differences 153 The exam was coordinated by the Ministry of Education, Science and Training (MEST) under the World Bank Institutional Development Education Project. See: http://www.worldbank.org/projects/P102174/institutional-development-education- project?lang=en&tab=overview 154 In particular, the assessment items, type of questions and assessment methodology were inspired in the international assessment PISA, a competence based exam which asks students to apply their knowledge to solve problems in real-world contexts with skills such as creativity, critical thinking, and practical use of knowledge. 155 Although the exam was taken also by students from Turk and Bosnian minorities, WB staff has had no access to data on the results for these minorities. 72 are smaller (1 point for girls and 1.5 points for urban pupils) and are insignificant in regression models. In the math exam, there were differences of almost 30 points (almost one and a half standard deviations) between municipalities (on a 100 points scale the average score in Gjilan was 56 points while the average score in Novobërdë was 29). The differences are lower but still considerable in the reading exam, with Gjilan and Novobërdë being the best and worst municipalities. Taking into account individual and school characteristics reduces these differences but not by much suggesting differences in quality of school resources and teachers, and socio-economic environment by municipality. Although inter-municipality differences are reduced when controlling for socio-economic factors, they remain rather high and suggest the large discrepancies of the efficiency in education between municipalities. Figure 5.15. Fifth Grade Reading Scores by Municipality 15 Without individual and school factors 10 With Individual and school factors 5 0 -5 -10 -15 Fushë Kosovë Shtërpcë Vushtrri Suharekë Klinë Kamenicë Rahovec Lipjan Novobërdë Podujevë Prishtina Dragash Ferizaj Kaçanik Mamush Deçan Istog Gllogoc Junik Malishevë Pejë Obilic Prizren Viti Gjakovë Shtime Mitrovicë Gjilan Hani i Elezit Source: WB Staff calculations based on 5th grade exams database. Note: Results reported are the regression coefficients of reading scores and municipality fixed effects, with and without individual controls on socioeconomic background and learning conditions. 5.31. Kosovo plans to participate in the 2015 PISA exam. The Program for International Pupil Assessment (PISA) is the most widely available and used tool. PISA is a worldwide study by the OECD that takes place every three years in member and non-member states and assesses 15-year old pupils’ capacity to deal with critical thinking problems in reading, math, and science. Hanushek and Woessmann (2009) showed that an increase of 50 points in PISA scores (measured on a 1000 point scale), which is equivalent to more than one year of schooling, raises a country’s growth rate by one percentage point per year for the decades to come. To understand Kosovo’s potential in the 2015 PISA exams a non -linear regression of PISA scores and GDP per capita (PPP) was conducted. The predicted score for Kosovo in reading category was 393 (OECD average is 493), similar to scores in other Western Balkan countries. The participation in PISA will allow to extend the results obtained in the 2010 national 5th grade assessment (which followed PISA’s methodology) to a more comprehensive analysis of the quality and equity of the system throughout differences between municipalities, ethnic groups, gender disparities, or types of schools. This will be possible given the high quality of the data extracted from the student and school questionnaires in PISA. The other main contribution of the PISA assessment will be to compare Kosovo’s education system performance internationally, identifying the strengths and weaknesses of the system and pointing towards potential areas of improvement. 73 G. DEMOGRAPHIC CHANGES COULD BRING COST SAVINGS 5.32. Enrollment projections suggest that there will be a decline in total enrollments . The population structure of Kosovo will change over the medium term due to the end of the baby boom in the early 2000s. While the number of children under age 19 was 655,000 in 2010, World Bank population projections suggest that this number will decrease from approximately 634,000 potential pupils aged 0-19 in 2015 to 604,000 in 2025, an estimated decrease of 8 percent. On the other hand, enrollment rates will likely increase for pre-primary (Grade 0) by 10,000 new enrollments for the period to 2017, and upper secondary education enrollment area also likely to increase.156 Overall, the combined outcome of both trends will lead to an expected enrollment decline across all levels of education from 440,000 to 381,000 pupils by 2025. 5.33. Given the evolution of the population structure and assuming full enrollment by 2018, there could be potential savings of 4 percent of public expenditures on education. Based on the demographic scenario described above (including expansion of pre-primary and upper secondary), there could be an average annual savings of around 4 percent for the education grant (€6.2 million on average per year) - provided that the government adjusts the use of educational resources in line with these changes.157 Although this estimate is made under the assumption that there will be expansion of access on pre-primary and upper-secondary, capital expenditures need to be taken into account for this to occur, given that upper secondary pupil teacher ratios are now very high and that gymnasiums and other secondary institutions are not physically integrated with basic education institutions.158 Additionally, expansion in access may require and expansion of teachers in these levels. 5.34. Internal migration will also change the demographics of the population differently in urban and rural areas, requiring school network optimization, especially in basic education where smaller satellite rural schools and lower PTRs are more prevalent159. With a decline in the school age population, comes the challenge of too many teachers relative to the number of school-aged children. Countries in this situation have to reduce costs by dismissing redundant teachers, consolidating classes and rationalizing schools. In Kosovo, gains from the latter may be limited because average class sizes are high, especially in urban areas. However, savings from consolidation of under-enrolled schools could be offset – at least in the short-term by efforts to reduce overcrowding in others, or by increasing access to education in pre-primary and upper secondary. Unfortunately, a high percentage of schools could still be considered small and there could be ongoing intra-municipality subsidies (i.e. smaller classes in some schools are maintained at the cost of larger classes in others). H. RECOMMENDATIONS 5.35. Pre-university education could benefit from both more resource and improved efficiency of existing resources. Demographic changes – resulting from both a lower birth rate and migration – are partly responsible for future savings and required future spending increases. Other savings and spending adjustments can be made by adjusting policy. Some overall spending increases are also warranted, 156 The model used to estimate future enrollment assumes that by 2025, there will be an increase of enrollment rates in upper secondary from 90 to 98 percent. 157 Given that the financial data from pre-primary classes is integrated in schools with basic education, the projection estimates assumes that the cost per student is equal in both educational stages. By dividing total expenditures of municipalities by total number of students in grades 0-12, we find a per student cost of €377 in 2012/2013. 158 According to the infrastructure database from MEST, there are no upper secondary institutions integrated with other education levels. 159 See Table A.16 in Appendix 3 for more details of enrollment by school size, including satellite schools. 74 especially to ensure increased access to pre-school and pre-primary education, which has high individual and social returns, especially for poorer groups. An assessment of needs in upper-secondary education may also point to increased spending also in this area. 5.36. Expanding access to pre-school and pre-primary education is key.160 Based on data from MEST, the combined GER for pre-school (ages 0-5) and pre-primary (ages 5-6) is 15.5 percent161, while the pre-primary enrollment is 75.2 percent. While pre-primary enrollment rates compare favorably to other SEE countries, access to pre-school opportunities is extremely low, meaning that most children in Kosovo do not have the opportunity to gain a set of emotional, behavioral and cognitive skills that lead to better school preparedness. This report therefore recommends:  Steadily expanding access to pre-primary education, financed partly through cost savings resulting from demographic changes. The system needs to absorb an estimated 8,531 new entrants in the next three years to achieve the government’s priority of reaching universal coverage by 2016. The average operating cost per pupil is estimated to be at least €377 per school year162, implying an additional allocation of €3.2 million or 0.07 percent of 2012 GDP on average to absorb the new entrants163. However, pre-primary classes are mainly integrated within primary schools and are half day programs. There will be savings from a decrease of enrollment in basic and secondary education, which could be used to free space for new entrants in pre-primary without the need for new investments, especially in rural areas where class sizes are lower. A detailed review of existing school space in basic education schools would need to be undertaken to better assess the implications for infrastructure investments needs, if any, to expand access in pre-primary.  Steadily expanding access to pre-school education is key. The government’s stated target is to achieve 50 percent coverage by 2016 may be ambitious given the current very low levels of publicly-funded kindergartens. Nevertheless, there is a need to review in detail the current full-time kindergarten model financed publicly. Currently, these kindergartens tend to be located in urban areas and seem to be inefficient and inequitable.164,165 Therefore there is a need to consider cost-sharing options that facilitate more equitable enrollment increases and expansion of provision accessible for all members of society. To create a clearer picture for investment needs, the existing capacities within municipalities in terms of building facilities, 160 This section is partly based on the report and analysis provided by UNICEF Kosovo. 161 The figure for pre-school shows enrollment in publicly funded institutions, normally referred to as kindergarten (44 of them in 22 municipalities) and pre-primary classes operating in existing schools. It does not include information on an estimated 55 licensed private kindergartens or pre-school facilities that do not report data on enrollments or children age groups. 162 Due to data limitations, it is not possible to extract the average cost per child for pre-primary. For the purposes of this analysis, the average cost per student on pre-primary has been extracted from the overall analysis presented in section 5.31. Data from BOOST database and EMIS has been use to perform this estimate. Given that there are some other expenditures financed by central government which are devoted to municipalities, it is likely that this number is an overestimate. 163 Assuming an equal distribution to today, where enrolment will increase in equal annual lots, in three coming years 8,531 students will need to be absorbed by the system to reach full coverage. The annual increase in resource requirements would be €1.07 million would be needed annually (€377x2844 pupils). 164 The unit cost per child in kindergarten in Kosovo is three times higher (at about 70 percent of GDP per capita) compared to countries that have similar kindergartens in Macedonia, Armenia, Poland and Romania. This fact points to considerable inefficiencies and over-resourcing of public fund going towards a very limited number of children. UNICEF Joined Hands (2011). 165 Anecdotal evidence suggest that one of the criteria for enrollment is public kindergartens is employment of at least one parent in public sector, which suggests further inequalities. 75 teaching staff and children transportation needs should be reviewed. To facilitate the process, clear standards should be developed to foster a licensing process of existing and new private kindergartens in order to increase the quality and access to these institutions. 5.37. Understand better the needs in upper-secondary education that will result from demographic changes, increased enrollment and rural to urban migration. Although projections show that the decline of population may allow increasing enrollment rates in upper-secondary education, two things need to be kept in mind. First, most gymnasiums and vocational education schools are in different facilities than basic education schools, so that adjustments in upper secondary institutions will not be able to access vacated classrooms as lower levels of education will be able to do. Second, the average class size in upper secondary schools is high. Although population decline will allow the system to absorb more students, there is a need to rethink the existing supply of schooling in upper secondary schools. A more comprehensive understanding will be needed taking into account potential policies such as decreasing the average class size (especially in urban areas), decreasing the number of shifts, increasing the quality of infrastructures, and investing in creating a better learning environment. 5.38. Enhancing Teacher Performance through professional development is critical in improving the quality of education, in addition to ensuring that the number of teachers is adequate. Making continuing professional development a requirement for maintaining licensing or re-certification is a way to increase quality, and take steps to ensure that teaching standards be maintained. This report therefore recommends:  Ensuring that teachers receive the mandated minimum training and that remedial measures be taken in cases of poor teaching performance. MEST has estimated the average cost of one hour of training at €1.5. To meet only the minimum requirement of 100 training hours per year, the GoK will need to allocate a total of €3.7 million or 0.08 percent of GDP over the next five years (or on average €750,000 per year).166 The major challenge with this requirement, in addition to limited resources, is the provision of high quality training programs that are aligned with the new curriculum, and the creation of mechanisms to support equitable access to these opportunities. 5.39. The municipal-to-school financing formula could be simplified and management of resources at the school level could be improved. While there have been some improvements, the municipal-to-school formula should have led to a more optimal use of resources than it seems to have done. However, to reap the full benefits of the municipal-to-school financing, further reforms are necessary including:  Providing sufficient training for school-level financial managers. Some managers at the school level appear to have been unprepared for managing the resources, and this has reduced their effectiveness.  Simplifying the municipal-to-school formula and monitoring its correct implementation. There has been feedback from municipalities that the municipal-to-school formula is complicated and the Ministry is undertaking steps to review the formula as well as develop monitoring criteria.  Encouraging a strong political commitment at the municipal level to close down small schools. This municipal-to-school formula is a key element in encouraging the optimization 166 The estimate is based on total number of teachers (24,930) and does not discount for teachers near retirement age who may not need to renew their license. 76 of the school network by closing smaller schools (often rural with small classes) and diverting resources toward larger ones (often urban with large classes). 5.40. The wide disparity in class sizes between schools means that the school network could be further optimized. Since municipalities determine which schools remain open and which ones can close, the central government can only provide information and encouragement. Improved implementation of the PCF formula should encourage municipalities to make school network changes. However, to assist them this report recommends that the MEST undertakes relevant studies:  Identifying areas for network rationalization and assessing the scale of potential efficiency gains, including class consolidation opportunities within schools. In addition to altitude, the study should also take into account distance between schools and remoteness factors. The study should also assess to what extent bus programs and existing road conditions allow for rationalization.  Identifying the geographical areas in which the saved resources are most needed. Resulting investments could be used to expand schools in areas with large classes and fast population growth. 5.41. Significant investments are likely to be required to improve existing dilapidated school infrastructure and to ensure sufficient provision of education where demographic changes will require it. Due to demographic changes, some types of school in some areas will require further investment. This will likely include urban schools and upper-secondary schools. In addition, a recent school infrastructure database (school map) points to significant investment needs. In particular, there are about 450 schools that have safety issues (unsafe and old structures, asbestos roof or stone foundations), another 400 schools either that have no water supply or no drinking water, and about 580 schools are without proper toilets. Resources saved from the closure of small schools can be used to help fund some of this required investment but is unlikely to be sufficient implying that additional resources may need to be found. This report recommends:  Ensuring that the provision of education keeps up with demographic changes. This includes ensuring that funds follow the pupils as they move geographically and as they advance in the school system. This will in particular imply increasing provision of upper secondary education as the pupil bulge moves from primary to secondary education, and as enrollment rates increase.  Costing and then taking steps to address the large deficiencies in schools. This will need to include prioritizing expenditures and ensuring that safety issues and provision of basic services, such as water, are met first. 77 CHAPTER 6. HEALTH FINANCING, SECTOR REFORMS, EFFICIENCY AND EQUITY167 A. KEY MESSAGES  Kosovo has made significant progress on a number of health indicators but outcomes are still comparatively poor, particularly with respect to maternal and infant mortality. Life expectancy at birth in Kosovo has increased over time (from 69 years in 2007 to 70.1 years in 2011), but remains the lowest in South East Europe. It is also lower than in Turkey (73.9 years) and 10 years lower than the EU average (80.2 years). Deaths associated with readily preventable perinatal conditions and maternal deaths are less frequent. Yet, infant and maternal mortality rates are still high compared to the EU averages. Circulatory diseases and malignant neoplasms death rates (though still significantly lower than EU averages) have risen rapidly in recent years. Circulatory diseases are now the primary cause of death in Kosovo, followed by neoplasms, perinatal conditions and respiratory diseases.  The GoK allocates a below average share of public sector resources to the health sector and fails to meet WHO’s macro criterion for financial protection: the share of out -of- pocket payments in total health spending exceeds the recommended upper limit of 15-20 percent. In 2012, the Government of Kosovo spent 9 percent of total spending and 2.6 percent of GDP on health, significantly less than both the EU (12 percent of general government spending and 4.4 percent of GDP in 2012) and SEE countries (13 percent of general government spending and about 5.3 percent of GDP in 2012). Private spending, typically in the form of out- of-pocket (OOP) payments at the point of service, contributes an estimated 40 percent of total spending in the sector. The largest share of OOP spending is on medication and reached 85 percent of OOP in 2011, possibly because GoK spending is directed toward other areas, leaving health care consumers to pick up the bill for medication.  Public health spending is heavily focused on capital investment, hospital care and wages and salaries. The GoK spent a comparatively large share of the public health budget on capital expenditures in 2007-2012 (11.8 percent), compared with the OECD average of 4-5 percent. Unfortunately, construction standards are below average and newly constructed facilities are often connected to inadequate water and drainage systems that have not been updated in years. Hospital spending accounted for 43.7 percent of total public health spending on average over 2007-2012 and 44.6 percent of total public sector spending in 2012 (compared to the EU and OECD averages, which were of the order of 30 percent in 2011) and hospital shares of spending on wages and salaries have significantly increased between 2007 and 2012.  Kosovo’s most recent key sector reforms were initiated under the new Health Law adopted by Parliament in December 2012. They included the establishment of medical chambers (to promote high and unified practice standards), a purchaser-provider separation (with a new health financing agency tasked to purchase services from facilities/providers), the introduction of service provider performance-based incentive schemes, and the extension of basic services co- payment exemptions to new categories of the population. The GoK has also recently stepped up its efforts in the direction of more rational pharmaceutical procurement, particularly with more active management of the positive drug list. 167 Chapter prepared by Karen Coulibaly (principal author), Marvin Ploetz and Aneesa Arur. 79  The introduction of mandatory health insurance financed through a combination of payroll tax revenues and citizen premiums is currently under discussion. The scheme is expected to raise additional revenues for the health sector and increase financial protection (particularly of the most vulnerable) through risk pooling. At the proposed premium rates, projections suggest that health insurance would raise an additional €66 million of revenue for the health sector. The initial benefit package would likely include provision of medicines and medical consumables through an outpatient drug benefit.  Improvements in efficiency of spending and quality of care can be made. For example, efforts can be made to refocus spending on primary and preventive care by channeling some of the revenues from the new Health Insurance Fund (HIF) towards primary and preventive care. Further capital investments can be restricted to use funds to prioritize maintenance. Improved procurement practices can save money on pharmaceuticals. Improved service can be incentivized by awarding wage increases based on performance.  The success of the health insurance scheme could be enhanced by adhering to several principles. For example, potential negative labor market impacts from health insurance premiums can be reduced by initially introducing health insurance at a comparatively low level – for example 5 percent rather than 7 percent and by ensuring that only poor groups are exempt. In addition, implementing the Health Insurance Law (HIL) in a manner that ensures that the overall fiscal deficit does not increase and that guarantees that other sectors (especially personnel intensive ones such as education) do not suffer as a result of increased public sector premium contributions will help. At the same time, implementing and be seen to implement well the benefits provided under the health insurance, including the initial out-patient drug benefit, notably by building adequate capacity to administer a health insurance system and ensuring a smooth transition from the Health Financing Agency to Health Insurance Fund. This will increase uptake rates among flat-rate contributors and minimize non-compliance. B. OVERVIEW OF HEALTH SECTOR 6.1. The health service delivery system in Kosovo consists of a network of public health facilities, including small primary health care facilities, larger primary health care centers with diagnostic facilities and specialists, regional hospitals, and the University Clinical Center of Kosovo (UCCK). The UCCK and the seven regional hospitals are owned by the Ministry of Health, while primary health care facilities are owned by the municipalities. 6.2. The health system reform is necessary for financial sustainability, and to increase financial protection for citizens, and has been at the top of Kosovo’s health policy agenda for nearly a decade.168 The health care financing reform is also one of the five objectives outlined in the Health Sector Strategy 2010-2014. The four other objectives are to reduce morbidity and mortality; reorganize and complete construction of the existing health infrastructure; improve the management of existing resources and the quality of services; and develop, pilot and use a health information system for evidence-based decision making. 6.3. Kosovo’s most recent reforms were initiated under the new Health Law adopted in 2012 and include both organizational and financing reforms. For example, the establishment of medical 168 The 2004 Health Law includes direct references to the introduction of a health insurance scheme. 80 chambers to promote high and unified practice standards, a purchaser-provider separation with a new health financing agency (tasked with purchasing services from facilities/providers), the introduction of service provider performance-based incentive schemes, and the extension of co-payment exemptions for basic services to new segments of the population. The introduction of mandatory health insurance financed through a combination of payroll tax revenues and citizen premiums is currently under discussion. 6.4. This chapter describes public health spending patterns focusing on domestically-financed public expenditures, and reviews and analyzes the efficiency and equity of current spending. It also discusses the potential fiscal impact of the introduction of health insurance, and makes recommendations for improved fiscal sustainability, efficiency, and financial protection. C. HEALTH OUTCOMES 6.5. Kosovo has made progress on a number of health indicators in recent years, but outcomes are still comparatively poor. Life expectancy at birth in Kosovo has, for instance, increased over time from 69 years in 2007 to 70.1 years in 2011, but remains the lowest in SEE. Life expectancy at birth was 77 years in Albania and approximately 75 years in Bosnia and Herzegovina, Serbia, FYR Macedonia and Montenegro in 2011. Life expectancy at birth in Kosovo is also lower than in Turkey (73.9 years) and 10 years lower than the EU average (80.2 years). 6.6. Maternal and child mortality rates are still cause for concern. The infant mortality rate has decreased from 12 per 1,000 live births in 2006 to 9.8 per 1,000 in 2011.169 The maternal mortality rate is highly volatile and more difficult to assess, in part due to the small number of deliveries each year: it was officially 7.2 per 100,000 live births in 2011, but as high as 43.3 per 100,000 in 2009.170 Deaths associated with readily preventable perinatal conditions and maternal deaths are less frequent. Yet, infant and maternal mortality rates are still high compared to the EU averages of 4.1 per 1,000 and 5.8 per 100,000 in 2011. The most frequent causes of infant mortality in Kosovo are lower respiratory tract infections, acute infective diarrhea and congenital malformations. 6.7. Circulatory diseases are now the primary cause of death in Kosovo, followed by neoplasms, perinatal conditions and respiratory diseases. Kosovo’s standard death rate (SDR) for circulatory diseases (187.5 per 100,000 in 2011) is 13 percent lower than the EU average and 56 percent lower than Bosnia and Herzegovina’s rate, but has increased in recent years, up from 137.1 per 100,000 in 2006. Circulatory disease mortality varies both by gender and age group: more men than women die from circulatory diseases in Kosovo and people over 65 are especially vulnerable. The SDR for malignant neoplasms is 70 percent lower than the EU average, but is rapidly increasing (26 per 100,000 in 2006 compared to 47.6 per 100,000 in 2011). Mortality associated with respiratory diseases is also on the rise, and rose from 9.7 per 100,000 in 2006 to 15.2 per 100,000 in 2011.171 169 NIPHK based on registrar records. 170 Public health facility and registrar records. 171 NIPHK. 81 D. OVERVIEW OF HEALTH SPENDING IN KOSOVO 6.8. Public health spending is largely tax-funded but there is also additional donor support that is not reflected in the general budget. The GoK finances health expenditure from its budget through transfers. On average, 93 percent of GoK funds were from general revenues between 2007 and 2012. Transfers to hospitals are based on inputs (e.g. beds and staffing) while funding for primary healthcare facilities is provided to municipalities on a capitation basis.172 About 5 percent of on-budget financing is financed by health institutions’ own-source revenues (e.g. health service user fees and drug co-payments). The remainder is financed by on-budget domestic and foreign donations. Donor contributions that do not appear in the general budget amount to an estimated €5.5 million in cash and €4.5 million in -kind in 2013. Off-budget donor support for the period 2014-2016 (including Government of Luxemburg support to the development of the health information system, Global Fund TB/HIV prevention and treatment project support, and the supply of drugs and consumables through the “Hope” project) is projected to amount to €12.3 million. 6.9. Public health sector expenditures173 have more than doubled in real terms over the past five years. Yet, aggregate levels of spending have remained more or less constant, both as a percent of GDP and of total government expenditures, and Kosovo continues to allocate a below average share of public sector resources to the health sector. In 2012, the GoK spent 9 percent of total spending on health: spending on health has remained roughly constant as a percent of GDP in recent years, rising from 2.5 percent of GDP in 2009 to 2.6 percent in 2012 (Figure 6.1). Both the share of GDP and the share of general government expenditures dedicated to health are lower than the SEE regional average (12 percent of general government spending and 4.4 percent of GDP in 2012), and the EU average (13 percent of general government spending and about 5.3 percent of GDP in 2012).174 The Government of Kosovo’s spending on health is also below spending averages reported by countries at similar levels of GDP outside the region (Figure 6.2). Figure 6.1. Evolution of Government of Figure 6.2. Public Health Spending and Per Kosovo Public Health Spending, 2007-2012 Capita Income, 2011 12.0 18 16 Public health spending (as a % of 10.0 14 8.0 12 6.0 10 GDP) 4.0 FRA DEU 8 SWE BIH ESP CHE 2.0 6 SRB MNE CZE MDA TUR SVK ROMPOL EST 0.0 4 KGZ MKD UKR BLR RUS LVA 2007 2008 2009 2010 2011 2012 ALB KSV KAZ 2 GEO ARM AZE Public health spending (share of total expenditures) 0 Public health spending (share of GDP) 1,000 100,000 GDP per capita (current US$) Source: Kosovo BOOST v1.0 based on MEF data. Source: WDI, 2013. Note: x-axis log scale. 172 The primary health care transfer formula officially accounts for the number of visits and services delivered and has a performance element, but it applies the same average number of visits and services (and standard prices) to all municipalities. Performance payments were made in 2009 only and discontinued ever since because facility-specific performance data was either unreliable or not transmitted to the relevant authorities. 173 Details of donors’ off-budget support are largely undocumented. The in-depth analysis in this chapter therefore takes domestically-financed spending as public health sector expenditure. 174 ECA fiscal database (FACE) 82 6.10. Regional and municipality-level variation in public spending is low. Own-source revenue financed spending made up only about 7 percent of municipality spending on primary care in 2012, and government grants are on a capitation basis. Regional and municipality-level variation in spending is therefore low: the coefficient of variation in per capita spending at municipality level was 0.36 in 2012, excluding Graqanica which is an outlier (Figure 6.3). 6.11. Hospital care consumes the largest share of public sector spending on health. Kosovo allocates a higher share of total public health spending to hospitals than EU countries. Hospital sub-sector spending accounted for 43.7 percent of total public health spending on average over 2007-2012 and 44.6 percent of total public sector spending in 2012, compared to the EU and OECD averages, which were of the order of 30 percent in 2011 (Figure 6.4). Figure 6.3. Regional Variation In Per Capita Figure 6.4. Evolution of Public Sector Spending (Domestically-Financed) Public Health by Functional Category175 Spending, 2012 40 180 Per capita public health spending 160 Millions euros (2007 prices) 35 140 30 (in euro, current) 120 25 100 20 80 15 60 10 40 5 20 0 0 2007 2008 2009 2010 2011 2012 administration and other medical products and equipment hospital care Kosovo average outpatient Source: Kosovo BOOST v1.0 based on MEF data. Source: Kosovo BOOST v1.0 based on MEF data. 6.12. There has been a shift in the composition of both primary care and hospital expenditures between 2007 and 2012, with a decrease in the share of resources spent on goods and services (particularly medical supplies), and increases in the shares of spending on wages and salaries and capital. The GoK spent 21.2 percent of total Primary Health Care (PHC) spending and 33.8 percent of hospital spending on goods and services in 2007. By 2012, these proportions had shrunk to 13.2 and 17.6 percent (Figure 6.5 and Figure 6.6) respectively. 175 Outpatient spending category includes spending on primary care and public health as well as spending classified as outpatient expenditures within the MEF treasury system; hospital care includes secondary care expenditures. 83 Figure 6.5. Primary Care Public Spending by Figure 6.6. Public Hospital Care Spending by Economic Category Economic Category 100% 100% 5.1 12.9 Capital Capital 90% expenditures 90% expenditures 80% 80% 21.2 13.1 17.6 Subsidies and Subsidies and 70% transfers 70% transfers 33.8 60% 60% Municipal Municipal 50% expenditures/ 50% expenditures/ Utilities Utilities 40% 40% 66.6 69.5 Goods and Goods and services 30% services 65.7 30% 20% 47.5 Wages and 20% Wages and salaries 10% salaries 10% 0% 0% 2007 2012 2007 2012 Source: Kosovo BOOST v1.0 based on MEF data. Source: Kosovo BOOST v1.0 based on MEF data. 6.13. A comparatively high percentage of the health budget is allocated to capital expenditures, but construction standards are below average. Kosovo spent 11.8 percent of the public health budget on capital expenditures on average over 2007-2012 while the OECD average was 4-5 percent. Serbia spent 2.5 percent of public health expenditures in 2010 and FYR Macedonia 4.8 percent of total public health spending on capital in 2011. Renovations and construction work undertaken between 2007 and 2009 was unfortunately of poor quality, with newly constructed facilities often connected to inadequate water and drainage systems.176 Yet, little has been spent on facilities’ sewage and water supply systems in recent years. 6.14. Spending on medical equipment is now mostly at hospital level. In 2012, 83 percent of public spending on equipment was done by hospitals, an increase from 58 percent in 2007. Expenditures on specialized medical equipment at UCCK accounted for as much as 45 percent of all spending on equipment in 2011. 6.15. Hospital shares of spending on wages and salaries have significantly increased. The share of spending dedicated to wages and salaries at the primary level rose from 66.6 percent in 2007 to 69.5 percent in 2012; the share of wages and salaries’ expenditures has increased even more dramatically at hospital level from 47.5 in 2007 to 65.7 percent of total sub-sector spending in 2012. This growth in the wage bill can mostly be attributed to increases in average salaries. 6.16. Kosovo may be paying more than necessary for pharmaceuticals. Public spending on pharmaceuticals and equipment made up 13.6 percent of total public health spending in 2012 and 15.5 percent of total public health spending on average over 2007-2012 (with a peak at 18.7 percent in 2009). Such levels of spending are on par with the OECD average, which is of the order of 15 percent. However, in absolute per capita terms, Kosovo spent around US$16 compared with US$110 in Turkey, which has fairly good pharmaceutical coverage. Yet, limited competition through open tenders, the small number of wholesale providers and the absence of reference pricing suggest there is room for better value for money. 6.17. Oncology drugs have been the major cost driver of the public sector pharmaceutical budget although only about 1,500 patients benefit from treatment every year. Oncology drugs accounted for 176 World Bank Kosovo Health System Network Master Plan project, Data report, 2009. 84 about half the increase in spending on pharmaceuticals between 2007 and 2012. Cancer drugs are also the biggest product category in the 2013 budget, with an allocation superior to amounts provisioned for spending for hemodialysis, vaccines and insulin put together.177 6.18. The GoK has recently stepped up its efforts to rationalize pharmaceutical procurement, particularly with more active management of the positive drug list. The Ministry of Health (MoH) is also developing protocols and clinical guidelines to establish and standardize criteria for selecting and using drugs to treat specific diseases or perform a given health procedure. The MoH recently (in April 2013) updated its classification of drugs and medicinal consumables (such as gauze and bandages, orthopedic and dialysis materials) into cost-impact categories according to a “Vital, Essential, Necessary” and “ABC” system.178 Yet, more needs to be done to improve the process for defining the essential medicines’ list and guidelines are yet to be drafted. There are currently no administrative instructions for pricing of medicinal products. In the absence of a mature and functional health information system179, authorities’ ability to closely monitor transactions (prescription practices and dispensing at pharmacies) to intervene to curb overuse of medicines and/or by-passing of the generic substitution policy is very limited. E. EFFICIENCY OF HEALTH SPENDING 6.19. Physical access to facilities does not appear to be a major issue in Kosovo. As of 2008, 80 percent of Figure 6.7. Number of Hospital Beds Per the population had access to some level of primary care 100,000 in Catchment Area within 10 minutes, over 80 percent of the population had access within 45 minutes to secondary coronary care units, neonatology and pulmonology departments. 6.20. At aggregate level, bed capacity does not seem excessive, with 223.9 hospital beds per 100,000 populations in Kosovo in 2011, compared to the EU average of 539.9 beds per 100,000, and 251.7 hospital beds per 100,000 in Turkey.180 There is, however, considerable variation in the number of hospital beds in different hospital catchment areas, with six times more beds in Prizren’s than in Ferizaj’s hospital (Figure 6.7).181 Source: Kosovo Statistics Agency. 6.21. There is a deficit of health personnel throughout the country. There were 1.2 doctors and 3.5 nurses per 1,000 population in Kosovo in 2011 which is low compared to the average of 3.5 doctors and 8.3 nurses in EU countries. Medical staffing levels are also lower across all regions and municipalities (Figure 6.8 and Figure 6.9) than in other SEE countries. 6.22. Though they have declined in recent years, the non-medical staff remains excessive. Non- medical staff in regional hospitals accounted for about 20 percent of total staff in 2012 compared to 40 177 MoH data and budget estimates. 178 ABC analysis classifies drugs from the most to the least expensive, category A drugs being the most expensive. 179 The health information system is currently at the pilot stage and pilot implementation is expected to begin in 2014 in Prishtina and Prizren. 180 NIPHK, 2013 and WHO HFA-DB. 181 Very low numbers of beds in Vushtrii, Mitrovica and Ferizaj would suggest reorganization and/or merger with a larger hospital is needed, but these options have to be considered along with potential impact on access time. 85 percent in 2009. This is high compared with many EU countries such as the UK, where the share of non- medical staff is of the order of 13 percent. Figure 6.8. Physicians Per 1,000 Population, Figure 6.9. Nurses Per 1,000 Population, 2011 2011 4 3.46 3.5 3 2.5 2 1.5 1.15 1 0.5 0 Source: NIPHK, HFA-DB (WHO). Source: NIPHK, HFA-DB (WHO). 6.23. Primary health care (PHC) facility utilization rates are low and there are marked municipality-level variations and substantial inter-facility variation in productivity of PHC. There were on average 2.4 visits per capita and 18.2 visits per physician per day in 2012, while there were 1.9 visits per capita and on average 18.4 visits per physician per day in 2006. In comparison, among SEE countries, only Albania reports lower visit rates (2.2 visits per capita in 2011). The EU average is 7 visits per capita and Turkey’s is 8.2 visits per capita, but given Kosovo’s demogra phic profile (and its greater proportion of youth), the gap may be justified (Figure 6.10). While the variation at the regional level is limited, variation in primary health care facilities’ utilization rates is significant at the municipality level. The number of visits per physician per day varied between a low 2.9 in the municipality of Podujeva to 61.6 in Ranillug in 2012.182 There are virtually no regional differences in the distribution of medical staff. Variations in municipalities’ demographic profiles only partially account for municipality-level variations in utilization and productivity, implying that there may be room to improve efficiency (Figure 6.11). Figure 6.10. Primary Health Care Visits Per Figure 6.11. Primary Health Care Visits/GP/ Day Capita, 2011 and Population Profile 9 8.1 8.2 70 Ranillug PHC visits per physiscian per day 8 6.98 7 60 Gllogovc 6.1 6 5.3 50 5 Fushe Novoberdo 40 Kosova Istog R² = 0.3507 4 Graqanica 2.5 Shtime 3 2.15 30 KaqanikDeqan Shterpce Kllokot Partesh 2 Malisheva Gjilan Kamenica 20 Gjakova Hani i Elezit Klina 1 Lipjan Ferizaj Mamusha 0 10 Podujeva Dragash 0 5.0 10.0 15.0 20.0 25.0 Share of population aged 60 and over Source: HFA-DB (WHO), Kosovo Statistics Agency Source: MoH, National Statistics Agency. Note: Serbia 2010 data. 182 2006 and 2012 data extrapolated from Statistics Agency and MoH data assuming 250 working days per year per physician. 86 6.24. Variations in productivity are also Figure 6.12. Number of Discharged Patients Per common at the hospital level where low Doctor in Regional Hospitals and UCCK, 2012 occupancy rates reflect low utilization rates. Hospital care utilization rates, measured as 400 350 hospital admission rates are low. There were on 300 average across the country only 9.6 admissions per 250 200 100 inhabitants in 2012 (with the EU average at 150 100 17.9). The average daily bed occupancy rate in 50 Kosovo’s regional hospitals was 54 percent in 0 2012; UCCK’s occupancy rate was higher at 65.2 percent. The average length of stays (5.4 days in 2012 in Kosovo in comparison with the EU average of 9.1 days in 2011) and variations in Source: Kosovo Statistics Agency. indicators of productivity such as the number of discharged patients per doctor across hospitals suggest the extent of inefficiency in some facilities (Figure 6.12). F. HEALTH SPENDING AND FINANCIAL PROTECTION 6.25. Public health expenditures account for only about 60 percent of total health expenditures and the share of out-of-pocket payments is relatively high. Private spending, typically in the form of out-of-pocket (OOP) payments at the point of service, contributes an estimated183 40 percent (38 percent in 2011 to be exact) of total spending in the sector. This proportion of private expenditure is on par with SEE countries, but high in comparison with the EU average: in 2011, SEE countries’ households OOP payments’ share of total health spending was 38.2 percent on average and significantly higher than the 16.1 percent share in Turkey and 16.2 percent share on average in the EU.184 6.26. Kosovo does not protect its citizens from the financial costs associated with out-of-pocket (OOP) payments on health according to the World Health Organization (WHO) guidelines. Financial protection is usually measured in terms of OOP payments against a household budget threshold. If out-of-pocket health spending exceeds this certain threshold of total household or total nonfood consumption, then it is considered “catastrophic”. Health spending large enough to make a household cross the poverty line is considered “impoverishing”. Kosovo fails to meet the WHO’s criterion for financial protection, as its share of OOP payments in total health spending exceeds the recommended upper limit of 15-20 percent. 6.27. OOP per capita health spending has slightly risen in nominal terms, but has remained roughly constant in real terms in recent years. In 2009, individuals spent €29.4 on average on health services compared to €31.2 in 2011. Correcting for inflation (using CPI), this translates into a slight decrease in spending with 2011 levels equivalent to €28.1 in 2009 constant prices. 6.28. The largest share of OOP payments covers the cost of pharmaceuticals and other medical goods. The share of out-of-pocket spending for drugs was as high as 67 percent of total OOP spending in 2008; it was even higher in 2011 and reached 85 percent of total OOP spending.185 183 Estimates based on household budget survey data. 184 HFA-DB (WHO). 185 HBS estimates. 87 6.29. The rich spend more out-of-pocket on Figure 6.13. Average Annual Per Capita OOP health than the poor. In 2011, individuals in the Health Spending by Consumption Quintile, richest quintile spent on average €74.9 compared to 2009 and 2011 €10.7 spent by individuals in the poorest quintile 80 (Figure 6.13). Individuals below the poverty line Average per capita OOP health spent about €20 on health out-of-pocket in 2011 70 compared with €37 on average for individuals 60 spending (euro) above the poverty line. 50 40 6.30. Out-of-pocket payments are moderately 30 progressive, but this may reflect self-rationing by the poor. The Kakwani index measures 20 financing progressivity as the difference between 10 the concentration index and the gross consumption 0 Gini coefficient. The Kakwani index for OOP Q1 Q2 Q3 Q4 Q5 payments was effectively zero in 2009, showing (lowest) (highest) 2009 2011 that OOP spending was equally spread out across income levels. In 2011, however, the Kakwani Source: Kosovo HBS 2009 and 2011. index was approximately 0.1, indicating that the rich spent a slightly greater share of their income on OOP payments for health than the poor. OOP payments accounted for 4.3 percent of the highest quintile’s total consumption, yet only 2.6 percent of the lowest quintile’s total consumption (Tables 6.1 and 6.2). This may mean that the poor forgo treatment (due to limited payment capabilities). Table 6.1. Progressivity of OOP Payments Table 6.2. OOP Payments as a Share of Total Per capita consumption 2009 2011 Consumption quintile Per capita consumption 2009 2011 Q1 (lowest) 6.5 6.9 quintile Q2 10.5 11.0 Q1 (lowest) 4.0 2.6 Q3 16.6 15.2 Q2 3.7 2.8 Q4 18.3 18.8 Q3 4.2 3.0 Q5 (highest) 48.1 48.1 Q4 3.4 2.8 Concentration Index 0.340* 0.384* Q5 (highest) 4.4 4.3 * * Total 4.0 3.4 Kakwani Index 0.028 0.096* Source: Authors’ estimates using ADePT and HBS * 2009 and 2011 data. Source: Authors’ estimations using ADePT and data from Kosovo HBS 2009 and 2011 Note: *Index is significant at 5%; **index is significant at 1%. 6.31. Catastrophic spending incidences remained relatively stable between 2009 and 2011, and were more common among the poor. In 2009, 11.5 percent of households exceeded the 10 percent threshold of total household consumption on health compared with 7.3 percent of households in 2011. Relative to nonfood consumption, the shares were 25.3 percent in both in 2009 and in 2011 (Table 6.3). The concentration index indicates how concentrated catastrophic spending was between different wealth groups. In 2009, catastrophic payments as a share of total household consumption were more common among the poorest but in 2011 catastrophic payments as a share of total household consumption became more common among wealthier groups. As a share of nonfood consumption, however, catastrophic payments were generally more concentrated among the poor both in 2009 and 2011 (Table 6.4). 88 Table 6.3. Incidence and Intensity of Catastrophic OOP Health Payments, 2009 And 2011 Threshold share of total household consumption 5% 10% 15% 25% 30% 40% 2009 Headcount 26.6 11.5 6.3 1.5 1.1 0.3 Concentration index 0.013 -0.060 -0.105* -0.105 -0.129 -0.189 2011 Headcount 19 7.3 3.1 1.0 0.7 0.4 Concentration index 0.097** 0.116* -0.012 0.090 0.141 0.322 Threshold share of nonfood consumption 5% 10% 15% 25% 30% 40% 2009 Headcount 40.4 25.3 16.9 7.3 4.9 2.8 Concentration index 0.005 -0.054* -0.107** -0.221** -0.221** -0.216** 2011 Headcount 38.9 25.3 15.8 7.7 6.0 3.2 Concentration index 0.019 -0.033 -0.086** -0.180** -0.236** -0.233** Source: Authors’ estimates using Adept and HBS data. Note: *CI is significant at 5 %; **CI is significant at 1%. Table 6.4. Incidence of Catastrophic Payment Using 40 Percent Nonfood Consumption Threshold Per capita consumption quintile 2009 2011 Q1 (lowest) 5.0 5.8 Q2 2.1 4.2 Q3 3.3 3.1 Q4 1.8 1.6 Q5 (highest) 1.7 2.3 Total 2.8 3.4 Concentration Index -0.216** -0.233** Source: Authors’ estimations using ADePT and data from Kosovo 2009 and 2011 HBS data. Note: *CI is significant at 5 percent, **CI is significant at 1 percent. 6.32. Health expenditures are impoverishing a number of households. In 2011, 29.7 percent of the population was below the poverty line compared with 34.5 percent in 2009. However, if health expenditures are considered as essential and the poverty threshold is adjusted accordingly, poverty increases to 31.3 percent in 2011 and 37.7 percent in 2009. This indicates that about 2 percent of the population in 2011 is not counted as living in poverty but would actually be considered poor if health payments were taken into account. This represents a 7 percent increase in the poverty headcount for 2011. Both in 2009 and 2011, the estimated poverty gap increased by about 10 percent when health payments were taken into account. When health payments were deducted from non-food expenditures, the normalized mean positive gap increased by about 1 percent – from 32.6 percent to 32.9 percent in 2009, and by 2.8 percent in 2011 – from 25.1 percent to 25.9 percent in 2011 (Table 6.5). This suggests that the rise in the poverty gap is due mainly to more households being brought into poverty and not to a deepening of the poverty of the already poor. 89 Table 6.5. Poverty Impact of Out-Of-Pocket Health Expenditures, 2009 And 2011 Gross of health payments Net of health payments Change % change 2009 Poverty headcount 34.5 37.7 3.1 8.9% Poverty gap 50.8 55.8 5.0 9.9% Normalized poverty gap 11.3 12.4 1.1 9.9% Normalized mean positive poverty gap 32.6 32.9 0.3 0.9% 2011 Poverty headcount 29.7 31.3 2.1 7.0% Poverty gap 46.2 50.9 4.6 10.1% Normalized poverty gap 7.4 8.1 0.7 10.1% Normalized mean positive poverty gap 25.1 25.9 0.7 2.8% Source: Authors’ estimates using HBS 2009 and 2011 data. 6.33. Middle-class households are the most vulnerable to impoverishing health expenditures. While health payments are largest at higher values of total consumption, it is the households in the middle that are brought below the poverty line as a result of health payments. The situation has slightly worsened over time (Figure 6.14).186 Figure 6.14. Pen’s Parade of Per Capita Expenditures Gross and Net Of OOP Payments, 2009 and 2011 2009 2011 5 pre-OOP consumption pre-OOP consumption 8 post-OOP consumption post-OOP consumption 4 Consumption as multiple of PL 6 3 4 2 2 1 0 0 0 .2 .4 .6 .8 1 0 .2 .4 .6 .8 1 Cumulative proportion of population, ranked from poorest to richest Cumulative proportion of population, ranked from poorest to richest Source: Kosovo HBS 2009 and 2011. G. FISCAL IMPACT OF HEALTH REFORM: A FOCUS ON HEALTH INSURANCE 6.34. The authorities have recently drafted and approved a new Health Insurance Law (HIL) which would, whenever enacted, provides the legal basis for the creation of a social insurance scheme and the associated Health Insurance Fund (HIF). The scheme is expected to raise additional revenues for the health sector and increase financial protection (particularly of the most vulnerable) through risk pooling. It would be financed through a combination of payroll tax revenues and flat-rate citizen premiums. The coverage would be mandatory, with penalties for those who do not contribute, though households receiving last resort social assistance (see Social Protection Chapter) and other poor 186 For each household, the vertical bar (or “paint drip”) shows the extent to which health payments reduce consumption. 90 groups would be exempt from premiums and co-payments. The initial benefit package would include provision of medicines and medical consumables through an outpatient drug benefit. 6.35. At the approved premium rates, projections suggest that health insurance would raise an additional €72 million of revenue for the health sector. The HIL foresees a formal sector contribution rate of 7 percent applied to both the public sector and VAT-registered businesses.187 This would be paid equally by the employee and employer. All members of households with at least one formal sector worker would be covered through the formal employee’s contributions. Other households would need to pay a flat-rate premium of €3.50 per month per individual, for those who are not exempt. While revenues from payroll tax contributions are relatively certain, the total revenues raised will depend on the uptake rate among flat-rate contributors in the informal sector which may increase over time. Assuming that the uptake rate among flat rate contributors begins at around 40 percent, a model built by the World Bank projects that, at these premium levels (the “baseline” case), gross revenues would equal €91 million but reimbursements to the budget for exemptions and public sector employer contributions of €25 million (to ensure that the HIL does not lead to increased fiscal deficit188) would bring down net revenues to €66 million189 or 1.3 percent of GDP (Figure 6.15). An alternative scenario posits a formal sector premium of 5 percent, a slower uptake for flat-rate contributors, and increased exemptions. Net revenue under this scenario is projected at €45 million, or 0.9 percent of GDP. Both cases provide for significant additional resources for the health sector, with an increase of approximately 50 percent under the baseline or 35 percent under the alternative scenario. The ability of the HIF to spend and be seen to spend these resources wisely will be the single most important factor in the success of the scheme. 6.36. A new outpatient drug benefit would likely be the first sign of progress achieved under the HIL. As shown earlier, out-of-pocket expenses can be large – and sometimes catastrophic – for households. In addition, there is some evidence that poor households under-spend on medication, with both individual and social costs. An outpatient drug benefit is believed to be administratively achievable in a short time period as well as highly visible and pro-poor. The cost of this is difficult to estimate since (i) while consumers currently spend an estimated €30 million per year on OOP for pharmaceuticals, this will likely rise when the cost reduces following the introduction of the benefit; (ii) the MoH may choose to introduce this benefit in steps rather than for all covered pharmaceuticals at once to ensure that the system functions well; (ii) centralized procurement could lead to savings compared with private sector provisions, including through increased use of generic medications. The eventual cost would ultimately depend on the medications covered, consumer behavior (price elasticity), and the co-payment rate (assumed, in the model to be 10 percent). Under the baseline scenario, “running costs” during the first year of the scheme are projected to be: €7.2 million to build up a reserve fund190; €1.2 million for administration; €25.3 million to reimburse the budget to ensure the HIL does not crowd out spending in other areas or increase the fiscal deficit. There would be €63.8 million for increased expenditures in the health system, including the new out-patient drug benefit (Figure 6.16). Under the more conservative 187 With an annual turnover of over €50,000. 188 That is, the HIL is designed to be fiscally neutral so that all new spending is matched with new revenues. One risk is that spending in other sectors, such as education, could have to be reduced to fund the new health insurance contributions. In a sector such as education, where 80 percent of spending is on wages, financing a 3.5 percent employer contribution would mean a cut of 14.3 percent from non-wage spending (based on 2012 outturn). This would not be affordable at a time when there is a pressing need to expand pre-school education (see Education Chapter). Therefore, it will be important to find a mechanism to ensure that public spending health premiums do not crowd out other spending. 189 In constant terms 2012 euro. 190 The eventual size of the reserve fund should be between three and six months of spending. However, the baseline scenario assumes that the reserve fund is built up over a period of five years to reach €41.2 million, or close to five months’ of projected expenditures in the fifth year. Clearly as spending changes, the reserve fund required will also change. The simulation assumes reserves are built up over a three year period. 91 alternative scenario, with lower premiums and a faster build-up of reserve funds, there would be €34 million for additional health care spending. The additional revenues under the baseline scenario would bring health spending in Kosovo up to 4 percent of GDP, which is closer to the EU (4.4 percent) and SEE country average (5.3 percent). Figure 6.15. First Year Revenue Scenarios, € Figure 6.16. First Year Spending Mil. Implications, € Mil. 100 100 90 12.4 90 Outpatient Drug 80 80 30 Rest of Benefit 5.5 population Quality 70 70 42.2 VAT sector Enhancement 60 60 30 29.0 Administration 50 Public sector- 50 34 employee 6 Reserve fund 40 40 17.6 12.2 Public sector- 9 30 employer 30 7.2 Exemptions 12.2 Exemptions 8 17 20 20 17.6 Public sector 10 16.9 10 18 premiums 7.7 12 0 0 Baseline Simulation Baseline Simulation Source: World Bank staff calculations. Source: World Bank staff calculations. 6.37. Revenue from the health insurance premiums would increase over time if contributions from the informal sector increase. Under the baseline scenario, total net revenues191 would increase from €72 million in the first year of collections (assumed to be 2015) to just over €100 million by the fourth year (2018) (Figure 6.17) if efforts to increase contributions from the informal sector are successful. Around two thirds of this increase would need to come from flat-rate contributors as uptake rates increase from 40 percent in the first year to close to 100 percent after five years. Most of the remaining increase is projected to come from increased contributions from the formal sector as the workforce increases. The increased revenue will allow for higher spending, first likely to be on the outpatient drugs benefit and then on other areas elsewhere in the health system (Figure 6.18). 191 Net after reimbursement for public sector employer and exemption contributions. 92 Box 6.1. Modeling the Impact Health Insurance Law The World Bank developed a model with adjustable parameters to understand better the impact of the HIL on revenue and spending under different scenarios. The baseline scenario uses the parameters in the proposed law as submitted to parliament. These include: (i) formal sector contribution rates of 7 percent of gross salaries (full compliance is expected – 40 percent is public sector), split evenly between the employer and employee; (ii) €3.50 per month contribution rates for all other non-exempt individuals: efforts will be made to make payment as easy as possible through the tax office and commercial bank branches; and (iii) exemptions for those on social assistance (see Social Protection Chapter), wards of state such as orphans and prisoners, and several other groups – including pensioners and war veterans – if they can demonstrate that they are “poor” (definition will need to be determined in secondary legislation). Other baseline assumptions include: (i) population growth of 0.9 percent per year (in line with World Bank, 2012 forecasts); (ii) economic growth of 4.5-5 percent per year, inflation of 1.5 percent per year; (iii) wage growth conservatively estimated as equal to inflation plus half of economic growth; and (iv) public sector employment remaining stable but private sector employment growing in line with the working age population such that the labor force participation rate remains constant. Figure 6.17. Health Insurance Net Revenue Figure 6.18. Health Insurance Net Spending, Sources, € Mil. Baseline € Mil. Baseline 120 120 100 100 80 80 60 60 40 40 20 20 0 0 2016 2017 2018 2019 2016 2017 2018 2019 Flat-rate premium payers VAT sector New health expenditure Administration Public sector-employee Reserve fund accumulation Source: World Bank staff calculations. Source: World Bank staff calculations. 6.38. Despite the health insurance scheme’s risk pooling potential there would be risks in the implementation of the HIL. First, the health insurance will effectively represent an increase in labor taxes for formal workers. Net contributions are projected at 1.3 percent of GDP, of which 1.1 percent is from the formal sector. This compares with 1.8 percent collected from Personal Income Tax and second pillar pension contributions in 2012 (0.6 and 1.2 percent, respectively). A 7 percent contribution rate would increase labor tax collections by over 60 percent. As discussed in the Revenue Chapter, social insurance rates are lower in Kosovo than in neighboring countries.192 Health insurance represents one way to increase these collections. Nonetheless this would represent a large increase for the formal labor 192 Social insurance rates are 10 percent in Kosovo combined for employer and employee contributions (for pensions only). Montenegro has similar rates while other countries collect between 18 and 27 percent (excluding Bosnia and Herzegovina). A recent paper (Khwaja and Iyer, forthcoming) suggests that many countries in the region are already harming growth due to over-taxation of the formal economy. 93 market to absorb in the country with the region’s highest unemployment rates.193 Second, the uptake rate among flat-rate payers may be lower than expected. Flat-rate payers are expected to contribute €12.4 million or 17 percent of total net collections under the baseline. Third, experience with the social protection sector suggests that there may be political pressures to increase exemptions, particularly for war-veterans and other groups that are not necessarily poor. Fourth, although the HIF is not permitted to run arrears, health facilities could run arrears (as has happened in a number of neighboring countries) in the absence of good management capacity and flexibility over labor costs. Fifth, the authorities may over- promise benefits, leading to disappointment in the scheme. Sixth, the benefits under the scheme may be – or, equally importantly, may be perceived to be – poorly implemented, resulting in a negative public view of the scheme and discouraging contributions. 6.39. However, the authorities can rise to these challenges and, indeed, have already taken steps to mitigate some of these risks. It will be essential that the benefits of health insurance are well implemented and visible. This would increase the uptake rate for flat-rate contributors and reduce non- compliance among formal sector workers. Indeed, if the scheme is successful, informal employees may encourage their employers to register them formally. Successful implementation would also increase public support of the scheme and increase “tax morale” – that is, the voluntary payment of tax (see Taxation Chapter). The authorities already appear to be taking a conservative approach to the benefits promised by the scheme, focusing to begin with on a visible, pro-poor and administratively feasible out- patient drugs benefit, which can be expanded or contracted to match revenues available with relative flexibility. 6.40. Benefits from the health insurance scheme could also be sustained with a lower formal sector premium rate and a low-uptake rate among flat-rate payers. Simulations show that the health insurance scheme would provide additional resources for the health sector even if formal sector premiums were to be set at 5 percent and the flat-rate uptake rates reaches only 40 percent by 2018. Under this “low scenario”, net revenues are still projected to reach €55 million per year (compared with over €100 million under the baseline) (Figure 6.19). This scenario would still allow for the provision of the outpatient drugs benefit and some additional quality enhancement in the rest of the health system (though less than under the baseline scenario) (Figure 6.20). 6.41. The HIF would not be able to run a deficit, helping to ensure fiscal sustainability. The HIL contains tight fiscal controls on the HIF including a provision to establish a reserve fund and ensure that spending stays within revenues. Spending through the HIF will be determined each year as part of the budget process. To ensure the long-term financial integrity of the HIF, exemptions should be strictly limited to the poor only. 193 A recent paper (Khwaja and Iyer, forthcoming) suggests that, while the whole of the economy in Kosovo could support additional taxation, the small number of existing taxpayers are already over-taxed. It will therefore be important to ensure that the health premiums are as broad based as possible. Koettl and Weber (2012) found that a 10 percent increase in labor taxation led to an increase in informality of between 0.8 and 1.6 percent, with lower earners being more affected among new EU member states and Kosovo is unlikely to be different in this respect. Therefore, too large an initial shock when introducing he tax may have negative labor market impacts. 94 Figure 6.19. Health Insurance Net Revenue Figure 6.20. Health Insurance Net Spending, € Sources, € Mil. Low Scenario Mil. Low Scenario 120 120 100 100 80 80 60 60 40 40 20 20 0 0 2016 2017 2018 2019 2016 2017 2018 2019 Flat-rate premium payers VAT sector New health expenditure Administration Public sector-employee Reserve fund accumulation Source: World Bank staff calculations. Source: World Bank staff calculations. H. RECOMMENDATIONS 6.42. While there have been some improvements in recent years, there remain many potential efficiency gains in health spending. In particular, improvements in efficiency of spending and quality of care could be made by:  Refocusing spending on primary care. The GoK share of health care spending on PHC and hospital goods and services declined between 2007 and 2012 from 21.2 and 33.8 percent respectively to 13.2 and 17.6 percent. While health expenditures have been steady, this shift from primary care and hospital goods and services to wages and salaries, the spending mix could be further improved, for example by channeling some of the net revenues from the new Health Insurance Fund (HIF) towards primary care.  Restricting further capital investments, and prioritizing maintenance. The GoK spent a comparatively large share of the public health budget on capital expenditures in 2007-2012 (11.8 percent), compared with OECD average of 4-5 percent. Unfortunately, construction standards are below average and newly constructed facilities are often connected to inadequate water and drainage systems that haven’t been updated in years.  Improving the value-for money for pharmaceutical purchases, and increasing spending on medication and supplies, particularly through using some of the new funds raised through the introduction of health insurance for this purpose. Poor and middle-class households are vulnerable to impoverishing health expenditures. The largest share of OOP spending is on medications and reached 85 percent of OOP in 2011. The new HIF will raise substantial revenue that will provide a drug benefit for OOP which will help the poor. In addition, Kosovo may be paying more than necessary for pharmaceuticals, and recent efforts to rationalize the procurement process should be continued to achieve better value-for- money.  Curtailing non-performance-based wage increases for medical staff. Hospital shares of spending on wages and salaries have significantly increased between 2007 and 2012194. 194 And again in April 2014. 95 Future increases should be focused on improving incentives for quality and efficiency through performance-linked payments. 6.43. Because of the low health care spending and the impoverishing impact of private health spending on poor groups, this report recommends the introduction of health insurance. In addition, health insurance would help to compensate for declining border revenues (discussed in the Revenue Chapter). The success of the HIL could be enhanced by:  Setting the contribution rate at 5 rather than 7 percent. Although it is not clear which effect would dominate a priori, there are clearly labor market risks, which should be considered in a country with the highest unemployment rate in the region. If the downside risks dominate, it would reduce private sector collections, including to the HIF itself. The risks could be minimized by introducing health insurance at a comparatively low level – for example 5 percent rather than 7 percent. This would reduce the pressure on the taxation and allow the health sector time to absorb the funds successfully.  Ensuring that only poor groups are exempt from premium payments and co-payments.  Ensuring that the HIL is implemented in a fiscally neutral fashion to safeguard the sustainability of the budget and guarantee that other sectors (especially personnel heavy ones such as education) do not suffer an undue burden.  Implementing and be seen to implement well the benefits provided under the health insurance, including the initial out-patient drug benefit, notably by building adequate capacity to administer a health insurance system and ensuring a smooth transition from the Health Financing Agency to Health Insurance Fund. This will increase uptake rates among flat-rate contributors and minimize non-compliance in the formal sector. 96 CHAPTER 7. SOCIAL PROTECTION195 A. KEY MESSAGES  Policy or political choices, rather than demographics or the economic cycle, have driven social protection spending. This has led to some “unfortunate” outcomes. For example, the last-resort social assistance scheme, which primarily targets the poor, declined in real terms over recent years while social protection programs for the more politically vocal groups, such as war veterans and non-poor pensioners, increased.  Pensions comprise over 80 percent of all social protection spending and have driven most of the social protection spending increase between 2008 and 2010. With favorable demographics, Kosovo does not face a looming unaffordable pensions bill, unlike many countries in Eastern Europe. Kosovo’s current pension system is working well and should be maintained. However, policies that increase pensions significantly above the rate of inflation could come back to bite as the current favorable demographics change in the longer-term.  Future large-scale investments in energy and health care reforms will raise costs and premiums which will adversely impact vulnerable households. Planned spending on social protection does not match the needs, and addressing the impacts of energy and health reforms should be the focus of social protection. Going forward, social protection should focus on the remaining challenges: (i) protecting the poor and vulnerable and enabling them to climb out of poverty; and (ii) mitigating the impact of modernization in other sectors – notably energy and health.  The draft Law on Veterans’ Benefits could have many negative fiscal and economic consequences, therefore its costs should be fully calculated and it should be rethought. The draft law proposes costly cash benefits, incentives for labor market informality, risks for the success of the Health Insurance Law and potential tax loopholes. It also risks crowding out other more urgent spending needs or increasing the fiscal deficit. B. OVERVIEW OF SOCIAL PROTECTION MECHANISMS 7.1. The social protection system in Kosovo comprises income transfer and other support. Income transfers that fall under the social protection budget include: (i) pensions, which include a universal tax-financed basic pension or old-age pension payable to all those over the age of 65, with a supplement for those who can prove that they contributed at least 15 years to the old Yugoslav pension system; a defined contribution pension established in 2002, which is fully funded and consists of mandatory contributions by employers and employees; and early retirement benefits for a very small number of beneficiaries196; (ii) war-related benefits, which include a plethora of benefits for civilians and soldiers who became disabled during the 1998-99 military conflict; and protection of surviving relatives of fallen soldiers and civilians, as well as for families of those missing; (iii) social assistance scheme, which is a tax-financed last-resort income support program, but which provides assistance only to a limited set of beneficiaries; and (iv) other transfers, which include a non-contributory disability pension; 195 Chapter prepared by Anita Schwartz (lead author) and Aylin Isik-Dikmelik. 196 This is for former Trepça Mine employees and former Kosovo Protection Corps (KPC) members who lost their jobs after the transformation of KPC to Kosovo Security Force (KSF) in 2008. 97 and benefits for children with disabilities. In addition to these, the social protection budget also finances other services including: social care services, provided by decentralized centers for social work, and focused mostly on child protection services; and employment services for those registered as unemployed but willing to work. Finally, though not part of the social protection budget, there are energy subsidies for recipients of the social assistance scheme and energy costs, which are designed to provide implicit subsidies from wealthier to poorer consumers through energy tariff differentiation; This chapter focuses on the largest ‘income transfer’ component of social protection - namely (i) to (iii) above – and also discusses briefly other measures, such as energy subsidies for poor households, which are not part of the formal social protection net. 7.2. Pensions represent the largest component of social protection spending in Kosovo. In 2013, 116,512 individuals benefited from the basic pension (pillar 1 pensions). They received €60 per month, financed directly from the state budget. Another 35,830 beneficiaries received a supplement of €52 per month in 2013 (making the total pension received €112 per month). This supplement was introduced in January 2008 for those who could prove over 15 years of contributions to the old Yugoslav pay-as-you-go system. Finally, a fully funded, contributory system (pillar 2 pensions) became mandatory for all formal sector employees who were under the age of 55 in 2002, when the system began. Contributions are paid by the employer and employee (5 percent of earnings each, paid through the taxation system). The funds are held in individual accounts managed by the Kosovo Pensions Savings Trust (KPST).197 Payments of €100 per month are made to beneficiaries until their accounts are empty. Since this second pillar is fully funded (with funds managed by a professional body and largely invested abroad), it is considered sustainable and does not require budgetary financing. Pensions are also paid to disabled persons, and early retirement schemes exist for the Trepça miners and for those individuals who were downsized when the Kosovo Protection Corps (KPC) was transformed in the Kosovo Security Force. In 2013, 18,891 individuals received disability pensions, down from 21,200 in 2011 and around 3,800 former Trepça miners received early retirement benefits in 2013, down from 4,130 in 2011. Finally, 798 persons received early retirement benefits from the KPC in 2013, reduced from around 1,600 in 2011. 7.3. A plethora of war-related benefits cover those who were injured during the 1998-99 war and the families of those killed or missing. War-related benefits were introduced in 2000 by UNMIK and additional categories were added in 2011. They covered around 11,000 people in 2013, in six main categories: (i) soldiers who became disabled; (ii) families of soldiers who died; (iii) families of soldiers who are missing; (iv) civilians who became disabled; (v) families of civilians who died; and (vi) families of missing civilians. In addition to these main categories, there are also benefits for caregivers to the disabled who are more than 80 percent disabled and payments made to the widows of those who were disabled and subsequently died. The total benefits received by households depend on how many dead or missing there are in a particular family and on the degree of disability for the injured. 7.4. Around 35,000 households benefited from the last-resort social assistance scheme in 2011 though this declined to 29,633 by end-2013. Beneficiaries fall into two categories identified through filters supposed to test poverty. Category 1 beneficiaries need to live in a household in which every adult is disabled, dependent or a care-giver for these beneficiaries. Category 2 beneficiaries need to be registered as unemployed and the household must have a minimum of one child under the age of six or an orphan less than 15 years old. Since 2013, there may be a maximum of one able-bodied person in the household. Both categories also carry additional exclusion filters, such as owning a car. Recipients of the social assistance scheme receive an average of around €63 per month, conditional upon meeting education and health-related conditions. 197 At the end of 2013, KPST held assets of around €914 million, of which €651 million is invested abroad and the remainder held in Kosovo either in cash or T-Bills. 98 7.5. Energy subsidies are delivered through the social assistance scheme and through a tariff structure that is designed partly to provide an implicit subsidy from wealthier to poorer consumers.198 All households that receive social assistance also benefit from energy subsidies at a total cost of around €4.5 million per year. At the same time, energy tariffs are designed to provide an implicit subsidy from larger (assumed wealthier) consumers to smaller (assumed poorer) consumers by charging tariffs differentiated by quantity consumed. C. FINANCING SOCIAL PROTECTION 7.6. Despite an increase in 2009, social protection has not been a spending priority over recent years. Transfer spending (of which around 95 percent is social protection) increased from 3.6 percent of GDP in 2008 to 4 percent in 2009 (Figure 7.1), largely as a result of an increase in pensions for those who contributed toward the old Yugoslav pay-as-you-go system. It has remained relatively stable as a percent of GDP since then. However, as a share of total spending, transfers declined from 14.3 percent in 2009 to 13.9 percent in 2012 (Figure 7.2) as the authorities chose to focus on other priority spending areas, notably transport infrastructure. 7.7. Policy choices, rather than demographics or the economic cycle, have driven social protection spending. The largest policy-driven increase in non-pension social security spending between 2008 and 2012 came from the war-related benefits, some of which were paid regardless of the beneficiaries’ age or ability to work, though some are determined by level of disability. Increases in the war-related benefits have been responsible for nearly a quarter of the total increase in the social protection bill, and almost all non-pension increases, over this period. The basic pension contributed an additional 6.4 percent of the increase, also as a result of policy choices. Sharp relative jumps of 12 to 13 percent in basic pension spending in both 2009 and 2012 as a result of increases in pensions in these years – first from €40 to €45, and then to €50. Figure 7.1. Transfer Expenditure in € and % Figure 7.2 Transfers as Share of Total Public GDP Expenditure 250 EUR (Mil.) 5.0 15.0 % GDP (rhs) 4.8 14.5 200 4.6 4.4 14.0 150 4.2 4.0 13.5 100 3.8 13.0 3.6 50 3.4 12.5 3.2 0 3.0 12.0 2008 2009 2010 2011 2012 2008 2009 2010 2011 2012 Source: BOOST and staff calculations. Source: BOOST and staff calculations. 198 As well as from less to more efficient consumers. 99 Figure 7.3 Composition of Social Security Figure 7.4. Contributions to Increase in Social Expenditures, EUR, 2008-12 Protection Budget 2008-12 Maternity leave 250 100% Maternity leave 6.8% Social services 90% 3.1% Social services 200 80% Pensions Social assistance Social assistance scheme scheme 22.9% 70% War invalid benefits 150 Kosovo Police Corps -- 60% Early retirement pensions Disability benefits 50% Early retirement scheme (Trepça) 100 40% Kosovo Police Disability benefits Corps - Early 55.6% retirement pensions 30% Early retirement War invalid benefits 50 scheme (Trepça) 20% Basic pensions - Basic pensions - Yugoslav Yugoslav system 10% system 0 6.4% Basic pensions 0% Basic pensions 2008 2009 2010 2011 2012 Contributions Source: BOOST and staff calculations. Source: BOOST and staff calculations. D. MAIN SOCIAL PROTECTION TRANSFERS Pensions 7.8. Pensions comprised over 80 percent of all social protection spending in 2012. There were four pension or retirement schemes funded through the budget in 2012 (Figure 7.3). Their Figure 7.5. Number of Basic Pension Recipients share of social protection spending increased from (Excluding Contributors to Ex-Yugoslav System) 79.9 percent 2008 to 83.3 percent in 2012. At €50 140,000 per month, the universal basic pension almost 130,000 brought pensioners up to the poverty threshold, which stood at €1.72 per day in 2011 (or €51.60 120,000 per 30-day month). Pensioners who also benefited from the supplement for contributors to the ex- 110,000 Yugoslav scheme received more, and these 100,000 supplements drove most of the pension spending increase between 2008 and 2010. 90,000 80,000 7.9. The introduction and expansion of 2008 2009 2010 2011 payments to contributors to the ex-Yugoslav pension system has driven over half of the Source: KAS, "Social Welfare in Kosovo". increase in social protection spending since Note: In 2009, pensioners who benefited from the ex- 2008 (Figure 7.4). Spending on pensions for those Yugoslav top-up were no longer classified as receiving who could prove a minimum of 15 years of the ‘basic pension’, although they did still receive it, resulting in the large decline between 2008 and 2009. contributions to the ex-Yugoslav system increased steadily from €18.4 million in 2008 – when these payments were introduced – to €42 million in 2012 - a jump from 14 to 36 percent of total pension spending. These increases came partly as a result of more people demonstrating eligibility and partly from an increase of supplement value from €35 to €51 in 2011 (Figure 7.5). However, this increase also partly derived from a redefinition of basic pensions. Initially, basic pensions were for all individuals aged 65 years and older. This was changed in 2009, when 100 only those who could not prove contributions toward the ex-Yugoslav pension scheme remained in the ‘basic pension’ category. Those who could prove these contributions were included in the new classification for both their basic pension and supplementary benefit. War Related Benefits 7.10. Benefits to each of the categories of war-related benefits have increased steadily since 2009. Total spending on war-related benefits increased from €18.4 to €31.1 million between 2008 and 2012. Nearly a quarter of the increase in social protection spending over this period went toward these benefits, which have increased sharply in value over this short period (Table 7.1). For example, monthly benefits for families of the war deceased increased by over 50 percent in just two years between 2009 and 2011. Interestingly, the benefits have increased slowest for those least able to participate in the labor market. War invalids with disability levels of over 80 percent received a 97 percent increase in monthly stipends – from €182 in 2009 to €358 in 2011 – while those with a disability level of 20 percent received increases of 230 percent. Table 7.1. War-Related Benefits 2009 2010 2011 % Increase 2009-11 Monthly benefits for war invalids with different disability levels, € Full invalidity 182 239 358.5 97% 71-80% 145.6 220 330 127% 61-70% 127.4 210 315 147% 51-60% 109.2 200 300 175% 41-50% 91 190 285 213% 31-40% 72.8 170 255 250% 20-30% 54.6 120 180 230% Monthly benefits for families of war deceased, €* Martyr Families 234 239 358.5 53% 2 martyrs 280 285.8 428.7 53% 3 martyrs 304.2 309.2 463.8 52% 4 martyrs 327.6 332.6 498.9 52% >4 martyrs 351 356 534 52% *Full invalids only. Benefits reduced for those with more minor invalidities. 7.11. War-related benefits are not designed Figure 7.6. Distribution of Spending on War- to cover the poor. Poverty alleviation may not Related Benefits in Bosnia and Herzegovina be the objective of war-related benefits and there is little firm evidence on the proportion of war- related benefits that go to different income- groups in Kosovo. However, anecdotal evidence and experience from other countries in the region suggest that, despite being part of the social protection budget, they tend not to reach the poor. In Bosnia and Herzegovina, for example, over 18 percent of war-related benefits go to the wealthiest quintile of the population while just over a third goes to the poorest quintile (Figure 7.6). Source: Europe Central Asia Region Social Protection Database. 101 Social Assistance Scheme 7.12. The last-resort social assistance scheme did not benefit from the same increase in spending granted to pensions and war-related benefits. Between 2008 and 2012, spending on the last-resort social assistance scheme fell by 8.6 percent in real terms, taking it from 19 percent of total social protection spending to 14 percent or from 0.7 to 0.6 percent of GDP. This decline reflected policy choices and priorities, which focused social protection on more politically vocal groups – such as war veterans and select groups of pensioners – rather than the poor. During this period, the government tightened the eligibility criteria while making small increases in the benefit. This reduced the number of beneficiaries (Figure 7.7) but kept costs stable in nominal terms. It also helped to ensure that social assistance was high as a percentage of beneficiaries’ total consumption. Indeed, at 44 percent, social security represented a higher share of consumption for beneficiaries in the poorest quintile in Kosovo than in other SEE countries (Figure 7.8). However, this is partly due to the very low consumption by the poorest groups in Kosovo – indeed, average individual receipts were worth just 7.2 percent of the minimum wage in 2011.199 In addition, the strict eligibility criteria have ensured that the social assistance scheme is one of the best targeted in the region. Figure 7.7. Number of Households Receiving Figure 7.8. Income from Social Assistance as % Social Assistance Benefits Household Consumption for Beneficiary Households in Poorest Quintile 39,000 50.0 45.0 37,000 40.0 35,000 35.0 30.0 33,000 25.0 31,000 20.0 29,000 15.0 10.0 27,000 5.0 25,000 0.0 2008 2009 2010 2011 2012 2013 ALB BiH MNE MKD RKS SRB Sources: KAS, "Social Welfare in Kosovo" and Source: National authorities and Europe and Central MoLSW. Asia Social Protection expenditures and evaluation Database, World Bank. Note: Data are from the following years ALB: 2008; BiH: 2007; KOS: 2009; MKD: 2010; MNE: 2011; SRB: 2010. 7.13. Kosovo’s last-resort social assistance scheme covers the highest proportion of the poor and has one of the lowest levels of leakage to the wealthy in the SEE. An estimated one third of households in the poorest quintile received income from the social assistance scheme in 2011 compared with an average of 19 percent for the region (Figure 7.9). Less than one percent of households in the top quintile 199 World Bank (2013), “Activation and smart safety nets in Kosovo: constraints in beneficiary profile, benefit design, and institutional capacity. 102 received these benefits – about average for the region. Targeting also performed well: over 70 percent of spending on the social assistance scheme went to households in the bottom quintile, while less than 3 percent was “leaked” to households in the top quintile. This compares favorably with regional averages of 62 and 4 percent, respectively (Figure 7.10). Although the coverage and targeting were good, there has been little effort to expand the social assistance scheme since 2008. In fact, the authorities have aimed to refine the exclusionary filters supposed to test for poverty to reduce the number of beneficiaries including through reducing the number of able-bodied people beneficiary households are able to have and introducing binary conditions to receive benefits such that people are excluded if they own a car. At the same time, almost all other social protection benefits – including those not targeted toward the poor – have increased. This reveals that the government’s priorities lie in rewarding those who fought in the war and increasing pensioners’ income at the expense of poverty alleviation. The reduction in the number of beneficiaries comes despite a high poverty rate. Figure 7.9. Coverage: % Benefitting Households Figure 7.10. Targeting: % Benefits Going to in Top and Bottom Quintiles Households in Top and Bottom Quintiles 35.0 9.0 90.0 14.0 8.0 80.0 30.0 12.0 7.0 70.0 25.0 10.0 6.0 60.0 20.0 5.0 50.0 8.0 15.0 4.0 40.0 6.0 3.0 30.0 10.0 4.0 2.0 20.0 5.0 2.0 1.0 10.0 0.0 0.0 0.0 0.0 ALB BiH KOS MKD MNE SRB ALB BiH KOS MKD MNE SRB Bottom quintile (lhs) Top quintile (rhs) Bottom quintile (lhs) Top quintile (rhs) Source: National authorities and Europe and Central Source: National authorities and Europe and Central Asia Social Protection expenditures and evaluation Asia Social Protection expenditures and evaluation Database, World Bank. Database, World Bank. Note: Data are from the following years ALB: 2008; Note: Data are from the following years ALB: 2008; BiH: 2007; KOS: 2009; MKD: 2010; MNE: 2011; SRB: BiH: 2007; KOS: 2009; MKD: 2010; MNE: 2011; SRB: 2010. 2010. 103 Box 7.1. Energy Subsidies and Costs Although not part of the Social Protection budget, poor Kosovars benefit from explicit and implicit energy subsidies. These aim to protect poor groups from rising energy prices, to reduce “energy poverty” – defined as spending more than 10 percent of household expenditure on energy – and ensure sufficient access to energy for poorer households. Figure B7.1.1. Distribution of the Implicit As well as the monthly stipend of €63 (on average), social Electricity Subsidy Across the Population, assistance beneficiaries received explicit annual energy Kosovo 2011 subsidies in the range of €120-130 per household between 2008 100 and 2012. These subsidies are distributed at a cost of around Proportion of the subsidies 80 €4.5 million per year or 0.1 percent of GDP. In addition, the received by each group 60 energy tariff structure is designed to provide an implicit 40 subsidies from large (assumed wealthiest) consumers to smaller 20 consumers by charging lower prices per kilowatt hours for 0 smaller consumers than for larger ones. While the explicit Poor Non Poor quintile 2 quintile 3 quintile 4 quintile 5 quintile 1 energy subsidies are well targeted, implicit cross-subsidies do little to help the poor. A recent analysis shows that this mechanism was actually regressive with wealthier consumers receiving more (27 percent for the wealthiest quintile) than Source: HBS Kosovo 2011. poorer ones in 2011 (the poorest received just 11 percent) (Figure B7.1.1). Figure B.7.1.2. Electricity Price and Electricity Figure B.7.1.3. Shares of Energy Share of Total Household Expenditures in ECA Expenditure in Kosovo, 2011, by Quintile Region 18 10 Percentage of total 16 Hungary average electricity price in expenditures 14 Latvia Turkey Poland cents of US$/KWh 12 Croatia Montenegro Romania 5 Lithuania Estonia Moldova 10 Albania 8 Bosnia Georgia and Bulgaria Belarus Armenia Herzegovina Kosovo 6 Azerbaijan Macedonia Serbia Kazakhstan Russia 4 Ukraine 0 2 Tajikistan Kyrgyz Republic Poor Non Poor 0 electricity share gas share LPG share wood share 0% 2% 4% 6% 8% Electricity as a share of total household expenditures oil share other fuel share collected wood Source: World Bank 2013, updated with Kosovo data Source: HBS Kosovo 2011. from ERRA database and from the HBS. Electricity prices are low compared with other countries in the region but electricity represents a larger share of spending in Kosovo than other countries, especially for the poor. In 2012, the average consumer tariff was 6.88 euro cent/ KWh after taxes, significantly below the regional average of 7.97 euro cent. Despite this, Kosovo’s households spent a larger share of their household expenditure on energy than almost any other country in Europe or Central Asia (Figure B7.1.2).On average poor households spent 8.4 percent of their expenditures on electricity in 2011, compared with 5.9 percent for non-poor households (Figure B.7.1.3). Future energy cost increases will be necessary but will likely pose risks for poor or vulnerable households. Although there would likely be some behavioral adjustment, the 47 percent cost increase projected under the Energy Strategy scenario (see Energy Chapter) would hit some households hard. Part of the solution lies in smoothing the cost increase to give time to adjust. While there are already both explicit subsidies (given to recipients of social assistance) and (ineffective) implicit subsidies, social protection for the poorest groups may need to be expanded to help them cope with the increases. The alternative would be increased poverty and increased non-payment of energy bills. With 52 percent of the poorest quintile reporting no spending on energy in 2011 (compared with 6 percent for the top quintile), the situation would risk worsening without corrective social protection measures. Kosovo should therefore take steps to identify the most vulnerable groups and, if necessary, modify social assistance criteria to ensure the poorest and most vulnerable households are able to access sufficient 104 E. FUTURE SOCIAL PROTECTION NEEDS 7.14. Policy reforms and high poverty point toward clear social protection needs. Two likely modernizations in key sectors will have consequences for household spending and may require a social protection response. The large investments required in the energy sector will increase energy costs (see Energy Chapter and Box 7.1). At the same time, the introduction of health insurance will likely have benefits for poorer households but will also add to the monthly bills of households that are not exempt from premium payments. For both health and energy, identifying vulnerable households and exempting them from payment (in the case of health insurance) or protecting them against energy price rises will be important. Without careful identification and indemnification of the poor, the (necessary) energy and health modernizations risk having a negative impact on the poor and vulnerable. 7.15. Planned spending on social protection does not match needs. The social assistance scheme covered just one third of the poorest 20 percent of the population in 2011 and a tightening of eligibility criteria will further restrict coverage. At the same time, the authorities plan to further increase war-related benefits and political pressures have expanded coverage of the ex-Yugoslav pension supplement200. Since the basic pension is already sufficient to bring pensioners to the poverty threshold and war-related benefits are, in many cases, very generous, these steps would do little to combat poverty. In addition, expanding war-related benefits and pensions would fail to address the challenges posed by rising energy prices or health insurance. The government should thus re-think priorities. Social Protection and Energy and Health Modernization 7.16. Going forward, policy considerations should focus on the impacts of energy and health reforms for the social protection system. The impact of necessary modernizations in energy and health are likely to affect poor and vulnerable households unless the social protection system is able to rise to the challenges the reforms pose. 7.17. Energy: Energy investments worth between 20 and 30 percent of GDP will need to be paid for through cost increases, which are estimated to increase from an average of 5.80 euro cent/KWh in 2013 to between 8.55 and 10.80 euro cent/KWh201 in 2019. This would be an increase of 47 to 86 percent in real terms. Since the poor already spend 8.4 percent of expenditures on energy, such an increase would be a significant burden. In addition, there is a risk that non-payment could increase. Commercial losses were equal to 20 percent of energy distribution in 2012 and 52 percent of households in the bottom quintile reported no electricity spending in 2011. This is classified as ‘non-payment’ or commercial losses by the energy companies. Using the social assistance scheme with an expanded net to “catch” vulnerable households could help to reduce poverty, minimize the negative impact of the cost increases and reduce non-payment of energy bills. 7.18. Health: The Health Insurance Law (HIL) provides for the non-formally employed under the poverty threshold to be exempted from premium payments. These would otherwise be €3.50 per month per family member. For an average sized household of five (according to the 2011 HBS) this would reach €17.50 per household per month. Put differently, a household on the poverty line, would need to find an additional 6.8 percent of household spending to fund obligatory insurance premiums. There is however evidence that medication comprised a large proportion of the budget of many poor household in 2011 and, in addition, many poor likely under-consumed medication (see Health Chapter). Therefore the premiums will be offset by lower health care costs for many households. Nonetheless, there are likely to 200 Parliament of Kosovo decided on May 2014 to recognize the right as ex-contributors to the employees in parallel system during the 90s, although the potential costs from this decision are not known. 201 Before taxes and at 2012 prices. 105 be poor households for whom this increase in spending would leave them in a vulnerable position. It will be essential therefore to identify the poor and vulnerable to ensure that the exemptions to which they are entitled in the HIL are applied to minimize this risk. 7.19. To combat the impact on the poor of both energy and health, a better effort to identify the poor will be required. The current social assistance scheme uses strict eligibility criteria to identify the poor including binary filters such that a single characteristic can exclude potential benefits, regardless of other characteristics. However, only around one third of households in the poorest quintile are covered due to the ineffective targeting mechanism. Expanding coverage for poor and vulnerable households, while preventing leakage to wealthy households should be a priority to both tackle poverty and mitigate the negative impact of energy price increases. This can only be achieved with a better targeting mechanism. This will also be important for health insurance, which will require better identification of the poor than currently exists to ensure that they are exempted from health insurance premiums. Identifying the poor would be the first step to tackle poverty and mitigate the risks associated with necessary modernizations in energy and health. This, combined with a willingness to expand the social assistance scheme to meet these challenges, should be the policy priority in social protection. War-Related Benefits 7.20. New war-related benefits included in the 2014 budget risk increasing the social protection budget while failing to address the current or upcoming challenges facing the country. The authorities are considering expanding and introducing new categorical benefits rather than focusing on benefits that target the poorest. Potential new war-related benefits include war veterans’ benefits and one- off payment for political prisoners:  One-off payment for political prisoners: Proposals to award a one-off payment to political prisoners under the Serbian regime following efforts to further the aim of the independence of Kosovo could prove costly. The law that would allow for these payments has already been passed by parliament (though it was not a government initiative). The GoK decided in December 2013 that the total amount for the political prisoners will be €55 million (1.1 percent of GDP) to be paid over four years. For the same amount, spending on health could be increased 10.7 percent per year, spending on education by 7 percent per year, or spending on social protection for the poorest and most vulnerable could increase by half.  War veterans’ benefits could prove extremely harmful for fiscal sustainability, have negative labor market consequences and crowd out more urgent spending. The process that will allow war veterans to receive benefits is already advanced. The relevant law has been adopted and close to 67,000 people registered. A process of verification is under way and expected to be completed in mid-2014. The 2014 budget has allocated €15 million – equal to 54 percent of the amount budgeted for the social assistance scheme. However, the total costs could be much larger. The proposed minimum cash benefit is equal to the minimum wage but employed veterans would not be eligible. This could have negative labor market consequences as potential beneficiaries move into the informal sector to receive the benefits. If there are just 20,000 eligible beneficiaries, the cost of the cash benefits alone would be around €40 million per year – close to 1 percent of GDP. 70 percent of these benefits could also be bequeathed to family members. Additional benefits for war veterans are also included in the draft law including those linked to health, education, housing, transport, and undefined business subsidies and tax and customs benefits. All ministries would be obliged under the law to propose what they can offer to eligible veterans. The final costs are unclear but it could result in: lost public revenues; increased tax loopholes; incentives for labor market informalities; business subsidies focused on the less viable businesses; an increasing share of the social protection budget going to the non-poor; and a 106 combination of increased fiscal deficits or crowding out of other more urgent spending, such as in health and education. In the health sector, the veterans’ law could undermine the new HIL as veterans and their families find that they are able to benefit from it while avoiding contributions and other members of society see fewer benefits as resources are diverted toward veterans. Pensions 7.21. Kosovo’s pension system is working well and with a young population, demographic pressures are not likely to be a challenge for Kosovo’s pension scheme over the medium term. Kosovo provides a universal but affordable basic pension and a defined contribution second pillar pension scheme. Between them these help to ensure pensioners are brought automatically up to close to the poverty line while providing formal workers with larger pensions with little or no budgetary risk. Nonetheless, there are both longer-term demographic challenges and policy risks. 7.22. In the longer term, as the currently youthful population ages, pressures on pension spending will grow. World Bank population projections foresee an increase in the population aged over 65 (the age at which the universal basic pension is received) from around 119,000 in 2010 to 356,000 by 2050 (Figure 7.11). At the same time, the number of working-age adults to support each pensioner will decrease from ten today to just four by 2050. The current pension system is well able to cope with this change if the economy continues to grow. Holding the basic pension at €60 per month in real terms would reduce pension spending as a share of GDP over the long term, reaching an estimated 1.2 percent of GDP by 2050, despite the growing retired population (Figure 7.12). 7.23. However, pressures to grant pensions increases have recently pushed pension spending onto a higher trajectory. Increasing numbers of pensioners are petitioning for inclusion in the ex-Yugoslav contributors’ supplement.202 Many argue that they would have achieved the requisite 15 years of contributions if they had not lost their employment during years of latent or overt conflict. Teachers and those who worked in closed factories notably fall into this category. If the ex-Yugoslav supplement continues to expand, more pensioners will shift away from the basic pension of €50 per month and toward a pension of €112 per month203, paid directly from the budget. If all pensioners shift begin to receive €112 per month (the €60 basic plus the €52 supplement), spending on pensions would jump to 3.2 percent of GDP by 2015, up from 2.3 percent in 2012. Figure 7.11. Pensioner Population and Aged Figure 7.12. Annual Costs Pensions , % GDP Dependency Ratio, 2010-2050 400 30 5.0 25 4.0 300 20 200 15 3.0 10 2.0 100 5 0 0 1.0 2010 2015 2020 2025 2030 2035 2040 2045 2050 0.0 2010 2015 2020 2025 2030 2035 2040 2045 2050 Total aged 65+, 000s Aged dependency ratio € 60 € 112 € 150 Source: World Bank Population Projections. Source: WB Population Projections and estimations. 202 As explained earlier such inclusion is already approved in spring 2014 although the costs to it are not known. 203 This amount was before increase of pensions by 25 percent starting from April 2014 driven by elections. 107 F. RECOMMENDATIONS 7.24. Social protection spending could be better aligned with Kosovo’s needs. In particular, any increases in social protection should be sustainable and targeted toward areas of need rather than special interest groups. Significant steps have already been taken to protect those who are not able to work, such as pensioners and the war-disabled. Therefore, going forward, social protection should focus on the remaining challenges: (i) protecting the poor and vulnerable and enabling them to climb out of poverty; and (ii) mitigating the impact of modernization in other sectors – notably energy and health. Social protection should be calibrated to encourage rather than minimize labor market participation and should therefore attempt to avoid leakages to wealthier groups who are able to work. This report therefore recommends:  Using the resources that are proposed to fund additional benefits to able-bodied war veterans and political prisoners instead to: (i) begin efforts to introduce a system that better identifies the poor and vulnerable. This will help both to identify “at -risk” groups following reforms, such as the new HIL or increased energy prices, and combat poverty by allowing improved social benefit coverage; (ii) expand coverage of the last-resort social assistance scheme while continuing to minimize leakage to wealthier groups. This will help to provide a safety net for poorer households who are affected by energy-price increases; and (iii) Ensure that the social assistance scheme provides necessary support for households to climb out of poverty, for example, by providing training and by allowing benefits to “phase out” when a beneficiary finds employment rather than losing benefits instantly.  Continuing to implement the current pension system without large changes. The current pension system is working well and there is no need to “fix” it. In particular, careful analysis should be undertaken before expanding the scope of the supplementary benefit for ex-Yugoslav pension scheme contributors and any increases in budget-financed pensions should be carefully considered in light of the long-run budgetary risks. 7.25. The new Veterans’ Law should be fully costed and rethought. The potential negative consequences of the law cannot be ignored. The law contains incentives for labor market informalities, elements that could reduce the effectiveness and the chances of success of the HIL and increased tax exemptions or loopholes. It could cost one percent of GDP per year in cash benefits alone, crowding out other needed expenditures and increasing the fiscal deficit. The full costs are likely to be much higher when other benefits and tax incentives are considered. The authorities should therefore seriously re-think this law that could have serious negative fiscal, economic and social consequences. 108 CHAPTER 8. TAX AND REVENUE POLICY204 A. KEY MESSAGES  Kosovo has a simple tax system and relatively low tax rates. Value Added Tax (VAT) and Corporate Income Tax (CIT) rates are flat at 16 percent and 10 percent respectively and Personal Income Tax (PIT) is progressive at 0, 4, 8 and 10 percent for income brackets of €0- 80, €81-250, €251-450 and over €450 per month. The simple tax system could make it easier for the Tax Administration of Kosovo (TAK) to adjust direct taxes to encourage labor market formality while increasing equity and collections. The tax legislation in Kosovo with respect to PIT, VAT and CIT are broadly aligned with international and EU standards.  A risky feature of Kosovo’s tax system is the high dependence on “border taxes”. In 2012, 71 percent of revenue was collected at the border in the form of trade taxes and VAT on imports. The current structure of Kosovo’s economy explains the current tax base profile. Over the recent past, imports have been worth between 52 and 57 percent of GDP. A combination of import substitution, lower prices of fuel and construction materials205 and slowing domestic demand in 2013 resulted in 8.8 percent lower than planned tax collection at the border, and 1 percent decline year on year for the first time during the last decade.  Shifting from border to domestic revenue collection is needed. The solution for the excessive reliance on indirect taxes and border revenues partly lies in bolstering domestic production and reducing its excessive reliance on imports. But strengthening the TAK and improving the tax policy framework by broadening the tax base, is also required. A gradual shift to revenues from direct taxes – personal income tax, corporate income tax, social insurance, and property tax – and from domestically generated VAT and excise, are necessary for sustainable fiscal development. Preliminary estimations suggest that improved compliance and an expanded domestic production base would likely need to be the main source of new revenues, with increased collections from domestic VAT, PIT, CIT and excises making positive contributions of 0.5, 0.5, 0.4 and 0.3 percent of GDP between 2012 and 2018. A smaller positive contribution – estimated at 0.2 percent – can also be made from property tax.  Low domestic revenue collection suggests the presence of a large tax gap in Kosovo, which has been estimated at about 35 percent of actual collections. Undertaking a tax gap analysis is challenging in an environment with limited or questionable data, such as Kosovo. Nonetheless, a tax gap analysis was conducted and it revealed that there is a significant tax gap: a VAT gap of about 34 percent was attributed largely to low VAT collection from domestic suppliers; a PIT gap is about three times what is collected now; and a CIT gap estimated at about 17 percent. A tax evasion perceptions survey206 conducted by the World Bank in 2013 also supports these findings. These results should be treated with caution due to data quality – particularly for PIT – but the results nonetheless indicate the presence of a large tax gap that could be narrowed through enhanced enforcement. 204 Chapter prepared by Munawer Khwaja and Sandeep Bhattacharya 205 Fuel and construction materials present a large share of overall imports and any price changes for these commodities is reflected in total border revenues. 206 Perceptions of Tax Implementation (PoTI) Survey, was conducted by World Band in June 2013. 109  A recent World Bank survey207 revealed a perception among firms that many of their competitors evade taxes. While not a surprising finding, given that Kosovo ranked alongside Ethiopia and Tanzania in Transparency International’s 2013 Corruptions Perceptions Index, it does however, have large costs to the tax system. Tax evasion is estimated at 5 percent of domestically collected VAT and CIT revenues, and 12 percent of PIT revenues. A high degree of evasion risks creating a downward spiral of non-compliance as firms evade taxes to ensure a level playing field with their competitors, whom they perceive as evading taxes.  Direct taxes collected at the center and municipal-level taxes are potential domestic revenue sources. Income taxes (CIT and PIT) account for about 2.5 percent of GDP, which is the lowest in the region. Kosovo follows good practice with progressive PIT rates which provide vertical equity – individuals with low incomes pay lower tax rates on their income – though there is scope to adjust them to both increase equity and encourage labor market formality. Income below the minimum wage should be exempt from tax, both to provide greater equity and to draw people out of shadow employment. At the same time, the introduction of health insurance could help to broaden the tax base and, if well-implemented, encourage tax compliance. Kosovo could consider increasing the CIT rate to 15 percent. Municipal level taxes can stimulate accountability, fiscal sustainability and efficient public services. Property taxes have several advantages over other taxes and are a potential source of domestic revenue. They are also likely to be below their optimal rate, and are currently below neighboring country rates. B. TAX POLICY IS JUDICIOUS BUT THERE ARE RISKS AHEAD 8.1. The right revenue mix is critical to enhancing economic growth, alleviating poverty and ensuring sustainable revenue streams to finance much-needed public investment. Since Kosovo adopted the euro as its official currency, it has had no control over monetary policy. Tax policy and budget therefore are the only instruments available to the GoK for achieving sustainable economic growth and macroeconomic stability. With large Figure 8.1. Trends in Tax Revenues and Pension investment needs (due to run down physical Contribution, % GDP infrastructure), a fiscally sustainable flow of revenues is necessary for Kosovo to build and maintain its physical infrastructure and address its key challenges. As such, the tax system needs to: (i) raise sufficient revenues to fund public spending that enhances growth and reduces poverty in a sustainable way without over- burdening taxpayers; and (ii) achieve the right balance between equity and maximizing economic incentives to work and invest. In this chapter Kosovo’s tax policy framework and revenue administration performance is appraised with a view to identifying policy options that would encourage additional revenues to cover spending Source: World Bank staff estimates. needs in a sustainable way. 207 The same survey as the above. 110 8.2. Kosovo achieved early successes in fiscal policy. Tax policy and administration reforms together with the introduction of new taxes and collection methods, contributed to a five-fold208 increase in domestic revenues between 2000 and 2004. The United Nations Interim Administration Mission in Kosovo (UNMIK) introduced a customs administration in September 1999, when it established border controls to collect taxes on all goods imported into Kosovo. Soon thereafter a tax administration was created. 8.3. The reforms helped Kosovo maintain low deficits and increase revenue collections, though they remain lower than other SEE countries. As a result of the reforms, Kosovo managed to maintain fiscal deficit below 3 percent in the last 7 years. Tax revenues grew from 18.4 percent of GDP to 21.9 percent of GDP in 2012; revenues growth was 18 percent in 2011 and 5 percent in 2012 (Figure 8.1). Pension contributions for the Kosovo Pensions Savings Trust (KPST) (a defined contributions mandatory second pillar) collected by the tax administration grew from 2.2 percent of GDP to 2.7 percent of GDP during the same period. Despite the growth in tax revenues, Kosovo has one of the lowest overall tax-to- GDP ratios in the SEE region. C. STRUCTURE OF PUBLIC REVENUES IN KOSOVO – A HIGH DEPENDENCY ON BORDER TAXES 8.4. A potential risk in Kosovo’s tax system is the high dependence on tax revenue from imports (71 percent) compared with the rather low figure of only 29 percent from domestic revenue . Although this is an improvement from 2004 when only 23 percent was collected domestically, this is clearly not a viable long term situation. The country cannot rely so heavily on import revenues and will need to build strong domestic revenue resources that are fiscally sustainable. 8.5. The current structure of the economy impacts the way taxes are collected. Household consumption averaged 89 percent of GDP over the last decade, most of which was imported (Figure 8.2). In 2012, imports accounted for 67 percent of household consumption, and import VAT receipts were 81 percent of total VAT revenue (Figure 8.3). In 2012 a significant part of household consumption and FDI was financed from remittances, (about 15 percent of GDP) and international aid (about 8 percent of GDP). The industrial sector is weak and most economic activities in the private sector – trade, retail, restaurants/bars and construction -are mainly small-scale, typically hard to tax, and many are not taxed. 8.6. The problem of border revenues has become more visible in the last two years. Revenue collection did not achieve its 2012 target and worsened in 2013, mainly as a result of a shortfall from revenues collected at the border. A €30 million shortfall or 0.6 percent of GDP in 2012, widened in 2013 to €50 million, or 1 percent of GDP. In 2013, for the first time, border revenue collections were lower than in previous year. This weak result was driven by a combination of a price decrease in Kosovo’s main import commodities, an import substitution effect (reducing taxes collected at the borders) and a decline in domestic demand for imports. Domestically collected VAT compensated 1.8 times the shortfall of border VAT. This is a very important development toward creating tax stability by compensating shortfalls of tax revenues collected at the border with more sustainable tax collected domestically, especially as imports decline. 208 This great result is also because revenue collection started from very low levels in 2000. TAK was established only in second half of 1999 after the 1998-99 war in Kosovo. 111 Figure 8.2. Structure of GDP, %GDP Figure 8.3. Imports and VAT on Imports, % GDP Source: Statistical Agency of Kosovo. Source: World Bank staff estimates. 8.7. A narrow tax base and low labor participation are also partly responsible for the high dependence on consumption tax revenues. Only 15 percent of tax revenue comes from direct taxation (including 1 percent from property tax) while the rest mainly derives from taxes on consumption (VAT and excise) and customs duties (Figure 8.4 and Figure 8.5). In part the limited coverage of direct taxes is constrained by the narrow tax base, a very low labor participation rate (37 percent) among the working age population209 and a high unemployment rate of 30.9 percent. In part this also reflects very low direct taxation rates: the highest rate of income tax is 10 percent; and in 2009 corporation tax was halved from 20 percent to 10 percent. Figure 8.4. Structure of Taxes in 2012 (% of Figure 8.5. Dynamics of Tax Structure (2004-12) Total) 100% 90% Other taxes 80% 70% Property tax 60% Customs duties 50% 40% Excise 30% VAT 20% 10% CIT 0% PIT 2004 2005 2006 2007 2008 2009 2010 2011 2012 Source: Ministry of Finance. Source: World Bank staff estimates. 8.8. VAT more than doubled with a growth of 125 percent between 2004 and 2012 (Table A.20 in Annex 4). VAT revenues grew from 42 percent of total tax revenue in 2004-08 to 48 percent in 2012. Excise duties remained steady at 26 percent and customs tariffs and duties at 12 percent of total tax revenues between 2005 and 2012 (Figure 8.5). PIT also grew but remained around 7 percent of collections and corporate income tax at around 6 percent of total tax revenue over the same period. PIT generated only 1.7 percent of GDP during the period, which is about half of the SEE average of 3.5 209 This level of labor participation is substantially below the average among all transition economies (65 percent) 112 percent of GDP.210 As a result of halving the corporate income tax (CIT) rate revenues fell from 1.7 percent of GDP in 2008 to 1.3 percent in 2012, which is lower than the average of 2.0 percent in SEE countries. Property tax provided only 1 percent of total tax revenue (or about 0.3 percent of GDP), compared to an average of about 0.6 percent of GDP for the SEE region, despite the post-conflict construction boom. Property tax rates remain exceedingly low by regional standards. D. METHODOLOGIES APPLIED TO MEASURE INFORMAL SECTOR AND REVENUE POTENTIAL 8.9. Analysis of the informal sector and the revenue potential can be approached from different angles and can be calculated using different methodologies. From the tax policy perspective, what matters are the taxable base and the scope of exemptions, loopholes and tax holidays that narrow the tax base by creating revenue gaps that could have been collected under a more simple tax regime. One way to look at the revenue performance in a country is using the top-down approach where the productivity of different tax revenues is calculated. Productivity measures how much of the tax base is taxed and how much is left untaxed. In an ideal world, productivity would be 100 percent if the entire base is taxed, there are no exemptions and there is no leakage due to tax evasion or collusion. Another method is the bottom- up approach of drilling into the bases of each of the main taxes to examine the potential tax base and the tax gap. Yet another measure is to estimate evasion based on currency demand to show the size of the underground economy. Unlike productivity, the last two methods demonstrate the gap from the perspective of leakage due to tax evasion or collusion rather than from the perspective of tax policy loopholes or exemptions. Size of the Informal Economy 8.10. Low revenue collection from domestic sources suggests the presence of a large informal economy. Using the Vito Tanzi model based on currency demand, the overall size of the underground disposable income as a proportion of disclosed disposable income211 averaged 17 percent over the period 2004 to 2011. The model takes into consideration currency in circulation, monetary base to liabilities ratio of the central bank, GDP, consumer price inflation, marginal lending rate, number of bank outlets (ATMs, branches), money velocity, disposable income, and currency/deposit ratio. The underground disposable income is likely to be an underestimate to the extent that the cash in circulation is underestimated with reference to measured GDP. With remittances estimated at 15 percent of GDP and a large errors and emissions line in the balance of payments, estimates of cash in circulation are likely to be on the low side. Tax Productivity212 8.11. The productivity of VAT in Kosovo at 65 percent is close to the average for ECA countries. This means that in Kosovo, for every 1 percent of VAT tax rate, VAT collection captures 65 percent of the VAT tax base (which is consumption), leaving 35 percent of consumption untaxed. The untaxed base could be a result of exemptions for certain part of the VAT base, or because tax evasion and/or weaknesses in the tax administration prevent the full base to be taxed. The performance of VAT collection seems adequate since it is close to the average for Europe and Central Asia. 210 Grote, M, D. Benedek and Selcuk Caner. Kosovo: Reform Proposals Toward a More Balanced and Revenue Productive Tax System, November 2011, IMF, Washington, D.C. 211 Disposable income without time adjustment. 212 Tax productivity is the actual tax collected as a percent of the tax base divided by the standard tax rate. For VAT, consumption is usually taken as the tax base since GDP excludes imports (which are subject to VAT) and GDP includes export (zero-rated under VAT) and investment (for which input credit is given under VAT). For income taxes, GDP is typically used at the tax base. 113 8.12. The productivity of revenues from income taxes is very low. For CIT the productivity is 14 percent, which means that only 14 percent of the CIT base is actually taxed. For PIT the situation is even worse. The productivity of PIT revenue is just 12 percent, meaning that only 12 percent of the PIT base is taxed. Clearly there is much scope for improvement in revenue performance in both PIT and CIT. This can be done by reducing tax loopholes and exemptions and strengthening the capacity of the tax administration to effectively fight tax evasion. Given that the tax system is already simple, with few loopholes (and should remain this way), the focus should be on the tax administration. It is clear from the productivity analysis that opportunities for improving the revenue performance are much greater for PIT and CIT than for VAT. Tax Gap Analysis213 8.13. The World Bank conducted a detailed tax gap analysis for VAT, PIT and CIT. Key findings are presented below and details of estimation methodology and the data used are presented in detail in a companion paper. It is important to note that the results presented below should be seen as indicative due to both availability of data and the quality of some data. Two issues in particular stand out: income data in the Household Budget Survey is not believed to be reliable as it totals around a quarter of consumption; and in absence of detailed firm-level data and random company tax audits, the study needed to simulate the range of the likely CIT gap based on non-random audits. These issues are discussed in detail in the companion paper, which also presents corroborating evidence from a separate firm survey of perceptions of tax implementation as well as other indicators. Despite the data issues, the results presented indicate the presence of a large amount of tax evasion. 8.14. The VAT gap is calculated to be about 34 percent. A tax gap analysis of the VAT gap was conducted using data from 2011 for household consumption with breakdown of expenditure and details of categories and exempt consumption. Similar calculations were made for the government and Non- Government Organizations (Non-Profit Institutions Serving Households, NPISH). Based on the analysis, the total VAT base, after taking into account exempt sectors, was estimated at €4.5 billion. This is close to 2011 GDP of €4.77 billion. At the rate of 16 percent the potential VAT revenue would be €719 million. In 2011, the Tax Administration of Kosovo (TAK) reported revenue of €121 million and the Customs Authority collected €415 million in VAT, a total of €536 million. This implies that the VAT gap is approximately 34 percent. This number is consistent with the figure for VAT productivity. 8.15. The PIT gap is estimated to be approximately three times what is collected now. Data for the PIT base was computed from a combination of the household budget survey (HBS) and the household consumption data in the absence of other sources.214 The data on household income are perhaps the most problematic in the estimation of the tax gaps. For the PIT gap the following were included: (i) wage tax, (ii) SME tax, (iii) taxes on interest, dividend, rent and gambling, (iv) individual profit tax and (v) other taxes. In 2012, €91 million were collected from these taxes. It is estimated that only about one-third of the potential PIT revenue is collected. The tax potential is in the range of three times what is collected now, i.e., about €270 million and the tax gap would be about €180 million. There seems to be a large potential for collecting additional revenue from PIT, partly through increasing formality and partly by reducing tax cheating. 8.16. The CIT gap is estimated at about 17 percent but better data are needed to be more accurate. In the absence of detailed information the tax gap for CIT was computed based on audit 213 The tax gap analysis was done by the Bank team together with Sandeep Bhattacharya, Associate Professor at the Sanford School of Public Policy at Duke University. 214 In Kosovo national accounts are not estimated on the income side. 114 detection. Data from audit detection show that the tax gap is very high for micro and small firms. In the baseline case, potential evasion is estimated at €10.6 million or 17 percent of the CIT receipt of €62.7 million for 2012. Further refinement of data – notably fully random audits to compare with risk-based audits – would be required to get a more accurate estimate. Tax Evasion215 8.17. A recent World Bank survey revealed that firms perceive that a significant proportion of their competitors evade taxes. The Perceptions of Tax Implementation (PoTI) survey questioned nearly a quarter of firms in the Large Taxpayers’ Unit (LTU) as well as some medium sized firms about their views on their competitors’ tax paying behavior. If perceptions reflect reality, CIT and domestic VAT evasion from large, formal firms alone cost the government at least 5 percent of collections annually, and around 12 percent of PIT collections. These results imply that tax evasion is responsible for between 15 and 25 percent of the tax gap discussed above. 8.18. The wider social and economic consequences are even higher. For the private sector, the implications are large: (i) corruption adds additional – and possibly unpredictable – costs to doing business, which may discourage investment; (ii) corruption may encourage firms to put effort into evading taxes and escaping penalties when caught rather than becoming more productive, harming economic development; and (iii) cheating by some creates an uneven playing field in the private sector with firms with better connections, rather than those that are most productive, more likely to be able to benefit. In addition, tax evasion by some firms reduces tax collection, which has to be compensated by higher taxes for firms that are tax compliant. Worse, tax compliant firms may become more reluctant to pay taxes if they perceive that their competitors are avoiding taxes. Ultimately, perceptions are capable of reversing the current culture which accepts taxes as a “moral” imperative. 8.19. The survey revealed that political connections were considered helpful in escaping severe penalties for tax evasion. Around 31 percent of respondents believed that “at least one of the owners being a politician” is the single most helpful way to escape a severe penalty for tax evasion. Another 30 percent of respondents believed that having “an owner or top-level manager who has personal friends in TAK” was the single most helpful way to avoid a penalty. These results point to significant high-level collusion in tax evasion, which would require robust mechanisms of enforcement to counter and to avoid a downward spiral into non-compliance. Revenue Potential 8.20. A recent study by Khwaja and Iyer216 shows that Kosovo’s revenue potential (including social contributions) is 25.9 percent of GDP compared to 21.9 percent actually collected The study uses a mix of structural variables, institutional variables and policy variables, using panel data from 61 countries, to analyze empirically the revenue potential. The results indicate that there is significant potential for improving the revenue performance by making the right policy choices. This could be used as an opportunity to correct the imbalance between revenues generated domestically and at the border. However, this result assumed that the informal sector can be brought into the tax net. Excluding the informal sector suggested that the existing small formal tax-base is already over-taxed. Therefore care should be taken not to further burden compliant taxpayers but to broaden the tax base or to replace existing taxes – such as those collected at the border – with new ones collected domestically. 215 This section is based on: World Bank (2014). “How my competitors avoid paying their taxes: Perceptions of t ax evasion among firms in Kosovo”. 216 Khwaja, Munawer S. and Indira Iyer (2013) “Revenue Potential, Tax Space and Tax Gap in ECA Countries: A Comparative Analysis, World Bank, Working Paper (draft). 115 E. A TAX POLICY REFORM FRAMEWORK FOR KOSOVO 8.21. Because Kosovo’s heavy reliance on indirect taxes collected at the border risks long-term fiscal stability, the objective of this review is to identify features in the current tax structure that could be altered in ways that would help bring greater balance between direct taxes and indirect taxes and between revenues collected domestically and at the border. Indirect taxes are considered to be less elastic – less responsive to changes in economic growth – and more regressive than direct taxes. A good tax policy framework should reflect the compliance environment, the size of the shadow economy and the capacity of the tax administration using international best practice as appropriate to the country situation. 8.22. Given the weak tax administration, the shift from taxes collected at the border should be done in a gradual and phased manner to optimize stability. The mix of tax policy measures should include: (i) augmenting revenues from personal income tax and corporate income tax; (ii) enhancing the collection from property tax; (iii) strengthening domestic VAT and excise collection through better compliance management; (iv) rationalizing SME tax regimes; and (v) improving tax administration. Augmenting Revenues from Corporate and Personal Income Taxes 8.23. Income taxes (CIT and PIT) account for about 2.5 percent of GDP which is the lowest in the region (Table 8.1). The combined productivity of CIT and PIT in Kosovo is also among the lowest in SEE. In Kosovo, CIT was originally applied at the rate of 20 percent on the accounting profit after adjustment for tax purpose. This was reduced to 10 percent from January 2009. The reduction in the CIT rate led to a drop in revenue from 1.7 percent of GDP to 1.4 percent of GDP. Between 2009 and 2012 it has averaged 1.3 percent of GDP. Albania, Bosnia and Herzegovina, Bulgaria, Kosovo, FYR Macedonia, Montenegro and Serbia all had CIT and PIT rates of 10 percent or less in 2011. The financial crisis forced some countries to increase tax rates. In 2012, Serbia raised CIT rate from 10 percent to 15 percent while the PIT rate for income exceeding 6 times the annual average wage is now 15 percent. In 2013, Montenegro raised the PIT rate for income above €720 to 15 percent. 8.24. Kosovo follows the good practice of having progressive rates for PIT which provides vertical equity - individuals with low incomes pay lower rates of tax on their income. Other Eastern European countries that have progressive rates include Croatia, Moldova, Montenegro (from 2013), Serbia, Slovenia and Turkey. In Kosovo, PIT rates start with a rate of 0 percent for annual income below €960, followed by 4 percent for annual income above €960, and a marginal rate of 8 percent above annual income of €3,000. The top marginal rate is 10 percent for annual income above €5,400. However, the progressivity is thin because the difference between the top and the bottom marginal rates is small and the range of income brackets between €960 and €5,400 is narrow.217 8.25. There is potential for Kosovo to raise additional domestic revenue from PIT. The tax gap analysis discussed earlier showed that Kosovo’s revenue potential from PIT is three times what it currently collected. Kosovo has one of the lowest PIT revenue productivity in the region (Table 8.1). FYR Macedonia, Bosnia and Herzegovina and Bulgaria collect about twice as much PIT revenues as a share of GDP with similar tax rates. This is a result of a combination of several effects: (i) exemption from PIT of several categories of income which narrows the tax base; (ii) weak tax administration; and (iii) large shadow employment that is not adequately captured. While serious efforts have been made by the Tax Administration of Kosovo (TAK) to improve performance much more need to be done. 217 The monthly minimum wage in Kosovo is €170 (for people a ged over 35), which is an annual income of a little over €2,000. 116 Table 8.1. Regional Comparison of Income Taxes Performance (Excluding Presumptive Taxes) 2011 Statutory Top marginal CIT Revenue PIT Revenue CIT+PIT yield CIT PIT CIT Rate PIT Rate (% of GDP) (% of GDP) (% of GDP) Productivity Productivity Kosovo 10% 10% 1.3 1.2 2.5 0.13 0.12 Macedonia 10% 10% 0.8 2.1 2.9 0.08 0.21 Moldova 12% 18% 0.7 2.2 2.9 0.06 0.31 Bosnia 10% 10% 0.8 2.3 3.1 0.08 0.23 Herzegovina Tajikistan 25% 13% 1.4 1.9 3.3 0.06 0.15 Albania 10% 10% 1.5 2.1 3.6 0.15 0.21 Bulgaria 10% 10% 1.9 2.9 4.8 0.19 0.29 Armenia 20% 25% 2.6 2.2 4.8 0.13 0.09 Montenegro 9% 9% 2.0 3.3 5.3 0.22 0.37 Turkey 20% 35% 1.9 3.6 5.5 0.1 0.10 Azerbaijan 20% 30% 4.3 1.4 5.7 0.22 0.05 Serbia 15% 15% 1.2 4.7 5.8 0.12 0.31 Croatia 20% 40% 2.4 3.5 5.9 0.12 0.09 Kyrgyz 10% 10% 4.2 2.0 6.2 0.42 0.20 Republic Belarus 24% 12% 3.1 3.2 6.3 0.13 0.27 Russian 20% 35% 4.2 3.7 7.9 0.21 0.28 Federation Georgia 15% 20% 2.8 5.9 8.7 0.19 0.30 Ukraine 25% 17% 4.1 4.7 8.8 0.16 0.31 Kazakhstan 20% 10% 8.2 1.4 9.6 0.41 0.14 Source: Tax at a Glance in ECA Countries, World Bank, 2013. 8.26. Policy makers should consider reorganizing the rate structure for PIT to make it more equitable and encourage formality in the labor market, and removing some of the exemptions that have no place in a good tax system. The income brackets range from half the minimum wage to 2.5 times the minimum wage which minimizes vertical equity. Income below the minimum wage should be exempt from tax, both to provide greater equity and to draw people out of shadow employment. Also the income brackets should be broadened and the rates should be raised for higher income brackets. A recent IMF report218 illustrated a simulation of a more equitable rate structure using 2010 data, with rates of 10 percent, 15 percent and 20 percent for income brackets of €2,000 to 4,500, €4,500 to 8,000 and over €8,000 respectively. According to the report this would increase PIT revenue by €20 to 22 million, an increase of 42 percent, raising PIT as a ratio of GDP from 1.1 percent to 1.6 percent. 8.27. Care should be taken when amending tax brackets to ensure that it does indeed encourage formality and does not over-burden the small group of existing formal taxpayers. If it done correctly, a reform such as this would lower the burden for low income tax payers and collect more revenue from high income taxpayers. One option would be to expand social insurance to introduce a widely-demanded health insurance scheme (see Health Chapter). According to data from the KPST, there were 225,000 employees and 30,684 self-employed who contributed to the second pillar pension in 2012. Since the second pillar is collected by the TAK, these numbers also reflect the approximate number of PIT 218 Grote, Martin, D. Benedek and S. Caner. “Kosovo: Reform Proposals Toward a More Balanced and revenue - productive Tax System”. IMF, Washington, November 2011. 117 taxpayers.219 This is a mere 15 percent of the total population. There is obviously a large shadow sector, some of which could be reduced if the tax burden for the lower levels of wage earners is lowered. Likewise, the policymakers should consider removing the following tax exemptions which may have been useful during post-war reconstruction but have outlived their utility: (i) dividends owned by foreign shareholders; (ii) interest on financial instruments issued by public authorities; and (iii) fringe benefits provided by employers. 8.28. An inclusive health insurance scheme would be one way to increase direct revenues and broaden the tax base. With one of Europe’s most under-funded healthcare systems as a share of GDP that is defined by a lack of access and significant catastrophic spending for the poor (see Health Chapter), the introduction of a health insurance scheme appears to be popular. Such a scheme could help to address these issues, posed by the chronic under-funding of the sector, while replacing falling border revenues with internally-generated ones. In addition, the scheme could be more inclusive than just the formal sector by collecting premiums from all citizens who are not exempt (e.g. the poor). This would help to expand the revenue base. Care would, of course, need to be taken to minimize negative impacts on the labor market, to protect the poorest groups from payments (see Social Protection Chapter), and to ensure that health insurance does not increase the overall fiscal deficit nor place additional spending burdens on other ministries (see Health Chapter). In addition, systems would need to be in place to ensure that payments could be made easily by those who are not formally employed, for example through the commercial banking system. However, if a health insurance scheme were seen to deliver value-for-money, it could increase demand for the service, thereby making a positive contribution toward tax compliance. The benefits of a well-functioning system could therefore potentially be large making positive contributions to state revenues (partly offsetting the decline in border revenues), increasing tax compliance, and broadening the tax base. 8.29. The rationale given for reducing the corporate rate from 20 percent to 10 percent is questionable. The logic has been to make Kosovo a favored destination for attracting foreign investments, reduce informality and corruption. As far as attracting foreign investment is concerned, it is a misplaced notion that reduced CIT rate will attract investment. First, if all countries in the immediate neighborhood reduce their rates, as was the case in SEE, then the effect of the reduced rate is neutralized and investment inflow would depend on the business environment and available physical and social infrastructure (healthy and skilled manpower). Second, a reduction in the rate in the host country (e.g., Kosovo) does not benefit the investing foreign company, since it pays tax on its global income in the home country (usually a developed country) and gets a deduction of the tax paid in the host country, to avoid double taxation. The smaller the tax rate in the host country, the smaller the deduction and the larger the share of the tax paid in the home country of the foreign country. Thus, in effect, by having a low CIT rate, Kosovo is transferring the benefit of its scarce revenue to the treasury of the developed home country, not to the investor. Nonetheless, halving the CIT rate from 20 to 10 percent in 2009 did reduce tax revenue but by less than 20 percent, suggesting that it was partly successful in encouraging payment and formality, although there were many confounding factors. Kosovo could consider raising its CIT rate to 15 percent to cover some of the tax gap mentioned earlier, but would have to ensure that compliance was maintained evenly across firms to reap the full benefits. Enhancing Collection of Property Tax 8.30. Taxation at the subnational level is a key revenue source which can stimulate accountability, fiscal sustainability and efficient public services. Kosovo has a fairly well-designed system of budgetary grants which unfortunately leaves little incentives for municipalities to raise own- 219 This number has gradually grown from 165,000 employees and 18,675 self-employed in 2004. 118 source revenue, resulting in exceptionally low municipal tax rates. The minimum rate on property tax collected by municipalities should be raised and central government grants should reward own-source generation of revenues. This would require careful dialogue and revision of the grants system. In addition, property taxes can be difficult politically and so would have to be managed in a way that is perceived to be fair. 8.31. Property taxes have several advantages over other taxes and are a potential source of internal revenue. Property taxes tend to be easier to collect than other forms of taxation and are relatively less distortive. They are difficult to evade (property is visible); are less likely to affect behavior than taxes on income; and taxes on land cannot distort supply since this is fixed. Because of these benefits, property taxes are more conducive to economic growth than other taxes, especially in emerging economies (OECD, 2008). In addition, revenues tend to be more stable than most sources of tax (Norregaard, 2013). Several countries (notably Ireland and Portugal), facing recent revenue shortfalls from more mobile or volatile sources, chose to ramp up their property tax collections. Despite the potential benefits of property taxes, Kosovo has not, so far, made property taxes a priority. 8.32. Collections from property taxes are likely below their potential. Kosovo raised around 0.3 percent of GDP from property taxes in 2011. This compared with an average of 0.4 percent for middle income countries and was significantly below the 0.9 percent average achieved the top-five middle income performers (Norregaard, 2013). Among SEE countries, both Montenegro and Serbia out-perform Kosovo with collections of 0.4 and 0.7 percent of GDP, respectively. One of the reasons may be Kosovo’s comparatively low property tax rates and/or narrow tax base. The other reason could be a large proportion of non-collection of what is already billed property tax.220 8.33. Kosovo’s property tax rates are lower Table 8.2. Property Tax Rates in SEE (% than some of their neighbors. Property taxes in Property Value) Kosovo are charged at an annual rate of between Minimum Maximum 0.05 and 1 percent of the value of the property with reductions for main residences (Table 8.2). Rates are Albania* 1.00 3.00 Bosnia and higher in Montenegro (0.1 to 1 percent of the Herzegovina** 0.05 0.50 property value), Serbia (0.4 to 2 percent), and Macedonia 0.10 0.20 Albania (1 to 3 percent outside of Tirana). Rates are lower in FYR Macedonia (0.1 to 0.2 percent of the Kosovo 0.05 1.00 property value) and Bosnia’s Republika Srpska (0.05 Montenegro 0.10 1.00 to 0.5 percent). As in Kosovo, all countries offer Serbia 0.40 2.00 potential reductions based on factors such as Source: Deloitte Tax Highlights, 2013. occupancy. Property taxes are levied by local Notes: *Outside of Tirana (Tirana applies rates of authorities in Serbia, Albania, FYR Macedonia and between 2 and 4 percent); **Republika Srpska only. Bosnia as well as in Kosovo. 8.34. Recognizing that improvements could be made, Kosovo has increased its property tax collection capacities. Considerable efforts have been made to improve cadaster through increased focus on registration and greater use of land titles. This has resulted in a doubling in the number of properties registered to around 80,000 in 2012. Similar efforts were made to construct a land registry (with nearly 2.3 million lots). The Property Tax Department of the Ministry of Finance (MoF) has upgraded its Information Technology (IT) systems so that all properties and lots can be identified with GPS 220 Data suggest that during a local election year the collection of property tax is weakest. Anecdotal evidence suggests that local governments are not making enough efforts to collect taxes during those years to avoid/reduce tax pressure to potential voters. 119 coordinates and photographs. The Department is now developing a system to categorize properties based on use, size and quality as well as environmental efficiency. 8.35. These efforts are expected to boost property tax revenues for municipal authorities. The expanded cadaster registry will help to broaden the tax base and raise additional revenues. The land registry alone is estimated by the MoF to raise around €2 million per year for municipal authorities. In total, the MoF expects that property taxes will increase from around 35 percent of municipal revenues to around 50 percent, once the full impact of the improvements is felt. Strengthening Domestic VAT and Excise Collection 8.36. The basic VAT structure in Kosovo is sound and designed according to EU norms. Although the VAT productivity is among the highest in the SEE region, much of the efficiency is gained from the fact that more than three-quarters of it is collected at the border where controls are easier and evasion less likely. While imports as a percent of household consumption is 67 percent, import VAT as a percent of total VAT collected is 81 percent. This shows that a significant part of domestic consumption remains untaxed under VAT. The low level of VAT collected from domestic suppliers indicates that there is significant tax evasion. Some of this is the result of the weaknesses in the tax administration to combat evasion of domestic VAT revenue. The VAT registration threshold of €50,000 is among the highest in SEE. Businesses below the threshold are subject to a presumptive tax. While the high threshold reduces compliance cost for taxpayers and administrative costs for the TAK, it also risks many large and medium sized businesses to abuse the system by hiding as small businesses. Designing an SME tax regime that reduces the risk of abuse and encourages SMEs to graduate to the normal VAT regime would help erosion of the tax base from domestic value added. 8.37. Kosovo collects about 5.7 percent of GDP in excises which is among the highest in the region. However, almost all of it is collected at the border with domestic excise collections being miniscule. The current level of excise does not require much modification. The excise system has specific rates which is appropriate for Kosovo. Improving Tax Administration 8.38. Created in the year 2000 under the auspice of UNMIK, and arguably the youngest in Europe, the Tax Administration of Kosovo (TAK) has had a reasonably successful short history. TAK has full operational autonomy within the Ministry of Finance. It administers personal income tax, corporate income tax, VAT, as well as performs the revenue collection functions for the second pillar pension fund on behalf of the Kosovo Pension Savings Trust. Kosovo Customs performs customs functions including collection of VAT on imports. Because it is new, it has been fortunate to avoid dealing with legacy systems and mindsets that exist in many other transition countries. As a result, over the last 13 years, TAK has taken significant steps to improve its capacity and service to taxpayers, thanks largely to support provided by international donors, primarily USAID, IMF, the World Bank and the European Commission. Kosovo ranks fairly high at 44th out of 186 countries on the Doing Business Paying Taxes Indicators. The productivity of VAT revenue at around 66 percent is close to the ECA average although the productivity of PIT and CIT revenues are still below the ECA average. The mandatory use of fiscal registers has helped improve VAT collection. 8.39. The TAK has developed an organizational structure that is function-based, and efforts to improve its capacities are evident. It has a relatively independent appeals unit which provides balanced dispute resolution in tax matters. As a result, eight decisions were decided in favor of taxpayers, 19 decisions were rejected. It has an internal audit department that examines the quality of tax functions and staff performance is in place. The IT department within TAK has upgraded continuously its capacities and 120 recently developed and implemented e-filing, integrated system for business registration, an electronic taxpayer accounts system, and has a functioning call center. Training of personnel and public outreach of taxpayers is ongoing. Most of its inspectors, and almost all senior management, are highly educated and relatively well experienced. A strategic management department focuses on actions for future reforms and has developed the TAK Strategic Plan 2010-2015. Also annual compliance plans are produced in-house and greater ownership has produced better collection results. The efficiency of tax administration has been improving in terms of the cost of collections 8.40. A new IT solution, with assistance from the SIDA, has helped to complete a survey of all property in Kosovo and consolidate property data, taking in additional property characteristics. It is hoped that this will help improve property tax collection by municipalities. 8.41. Severe challenges remain: the large shadow economy and tax gap point towards serious weaknesses in the ability and willingness of the tax administration to deal with tax evasion and tax avoidance. TAK lacks the capacity to effectively monitor private sector activities. In 2010, a tax investigation unit was created within TAK to deal directly with tax crimes, which it is hoped will, in the medium term, help reduce the tax gap. The effectiveness and sustainability of TAK is severely impeded by the lack of professional, transparent and effective service to taxpayers. Although the situation has recently improved221 through public outreach, still there are taxpayers who do not understand their taxpaying obligations. According to Global Integrity 2011222, tax laws are not always enforced uniformly and without discrimination, and corruption seems to be a major concern. Falling short on fair and uniform application and enforcement of tax laws to collect revenues is a critical impediment for establishing credibility and creating an atmosphere conducive to voluntary compliance. Arrears collection has been a significant problem but recently TAK has become more active in enforced collection actions including seizure of property and vehicles to the extent of € 500,000. The total amount of old arrears has consistently been falling. 8.42. Deficiencies in core areas of tax administration require greater focus on improving management effectiveness, staff training, and completing identification and registration of non- filers. The tax administration IT system, SIGTAS, needs to be upgraded to improve processing capacity, risk management and audit functions and enforced collection. Development of a reliable database and information matching system will be critical for reducing the number of non-filers and stop-filers. F. RECOMMENDATIONS 8.43. A good tax policy framework should reflect the compliance environment, the size of the shadow economy and the capacity of the tax administration using international best practice as appropriate to the country situation. For sustainable fiscal development, there is need for a gradual shift from revenues collected at the border to revenue generated internally, both from direct taxes and domestically generated VAT and excise. Several steps can be taken to help achieve this within the existing tax structure: 221 The PoTI survey found that 90 percent of firms agreed that “all the information I need about tax compliance is easily available” and 84 percent agreed that “tax procedures are clear and simple”. Three quarters of firms agreed with the statement “tax legislation is sufficiently clear and precise to be easily understood by taxpayers and to avoid legal ambiguities in its application”. 222 Global Integrity Report 2011, Washington, DC. 121  Strengthening the tax administration. This could including enhancing the quality and quantity of tax inspectors, improving risk assessment modules, including by adding the practice of random audits to assess areas for improvements in risk assessments.  Taking steps to fight tax evasion. Analyses undertaken for this chapter and the PoTI survey suggest large tax evasion. Anecdotal evidence suggests that even in cases in which the TAK attempts to pursue evaders, they face obstacles in the judicial system. Efforts should be made to curb these illegal practices. 8.44. Changes could be made to the existing tax structure without making the tax system more complicated. These changes could also help to encourage formality, broaden the tax base and increase the perceived fairness of the tax system. Efforts could include:  Designing an SME tax regime that reduces the risk of abuse and encourages SMEs to graduate to the normal VAT regime. This would help erosion of the tax base from domestic value added.  Considering raising the CIT rate to 15 percent while reinforcing enforcement. The rationale given for the past reduction in the corporate rate from 20 percent to 10 percent is questionable. It did not produce any effects in attracting FDI as the same reduction of CIT happened elsewhere in the region. However, at the same time as increasing the rate, efforts should be made to strengthen enforcement of higher CIT. The halving of the CIT rate from 20 to 10 percent resulted in a decrease in tax revenues by just 20 percent suggesting that they may have been some benefits in terms of compliance. Any increase could have the opposite effect. Therefore, compliance capabilities would need to be ensured both to guarantee the maximum revenue benefits and to guarantee that all companies are equally burdened so as to avoid distortions in the private sector.  Considering streamlining the rate structure for PIT to make it more progressive and equitable, and remove some of the exemptions that have no place in a good tax system. The following tax exemptions in particular, which may have been useful during post-war reconstruction, have outlived their utility: (i) dividends owned by foreign shareholders; (ii) interest on financial instruments issued by public authorities; and (iii) fringe benefits provided by employers.  Introducing a health insurance scheme that expands the tax base. It appears to be demanded by a large proportion of the population and, if the scheme is well-implemented, could increase tax compliance.  Increasing efforts to expand collection of property tax. The minimum rate on property tax collected by municipalities should be raised and central government grants should reward own-source generation of revenues. Also government should continue with its efforts to expand the property tax base by starting to tax agriculture land and other types of property. 122 CHAPTER 9. TOWARD A BUDGET FOCUSED ON NEEDS 9.1. This report has shown how Kosovo benefited from a budget focused on investment in physical assets over the past decade. Although the government’s steerage of the economy in recent years has been good, looking to the future it is clear that Kosovo will continue to face many challenges, particularly with respect to investing in and developing its human capital. While the continued planned investments in transport infrastructure will be beneficial for the country, shifting some of those resources away from non-urgent highways and toward more humble but also more urgent areas – such as local roads, education and protecting the poor – could prove to be better investments in the future, with higher economic and social returns. Overshadowing all of Kosovo’s policy options is a looming energy crunch and the issue of new power generation financing. Private sector investment is considered the best option but the government cannot rule out the possibility that significant public investment will be necessary. 9.2. Kosovo will need to meet these challenges while facing declining border revenues. The authorities are well aware of the need to compensate for declining border revenues that will be squeezed by trade agreements and import substitution. Without firm measures to compensate for falling revenues, the country’s ability to meet future spending needs will be curtailed. 9.3. Kosovo is essentially faced with two choices: (i) to stay on the current course (a status quo scenario), or (ii) to take the incremental reform path (an adaptive scenario).  The first is a passive option that would lead to falling revenues and either fiscal deficits that fail to meet the Fiscal Rule or a crowding out of urgent spending needs. In addition, it would mean: little efforts to reduce the costs of public procurement in areas such as pharmaceuticals or road maintenance, where potential savings have been identified; wage increases that continue to be ad hoc and politically-motivated, failing to motivate improved performance in health and education; few new resources to allow the education system to respond to demographic changes or to expand pre-school; little effort to combat poverty or incentivize formal employment; and insufficient resources to step in to ensure power supply for the country, if this becomes required.  The second option would entail compensating for declining border revenues with domestic revenues, and adjusting spending priorities to better meet Kosovo’s needs. That would mean addressing low spending on health and road maintenance, refocusing the social protection budget on social protection, incentivizing good service delivery in health and education by awarding salary increases based partly on performance and training, making some additional resources available to increase pre-school spaces and allow the secondary system to adjust to demographic changes.  In either scenario, Kosovo may be faced with the need to find public funds to invest in a new power plant if private investment is not forthcoming. With insufficient power availability, economic growth would be harmed, reducing the resources available to finance other investments and harming employment growth. Addressing power availability challenges may mean making difficult spending choices and adjusting priorities. 9.4. This chapter provides illustrative examples of what a budget may look like under the status quo and adaptive scenarios. It is not meant to provide details of exactly how a budget could be composed but instead provide an overview to illustrate how resources could be reassigned under different scenarios, including by implementing some of the recommendations in this report. It first considers both 123 the first and second scenarios described above. Next, it shows how a budget could be designed if significant public sector resources are required to ensure continue power supply for the country. A. WHERE IS THE BUDGET HEADING? Baseline scenario – the Status Quo 9.5. While it is difficult to forecast exactly how a budget would be composed with no policy changes, this section illustrates how the continuation of current policies would affect the budget over the medium term. In particular, with no efforts to collect more domestic revenue and no tax policy changes, revenues are projected to decline as a share of GDP over the medium term. At the same time, although spending on new highway construction will slow, the resources freed will be used to finance across-the-board wage increases and war-related benefits. 9.6. Without action, revenues are projected to decline over the medium term. Border revenues will come under pressure in the medium term as trade agreements with Turkey and the EU come into force. These two agreements together will reduce border revenue (trade taxes and VAT on imports) by 1.6 percent of GDP (see Revenue Chapter). However, imports are also assumed to be substituted by domestically produced goods and services, helping to increase domestically collected VAT, even in the absence of policy change or increased collection Figure 9.1. Revenues, Excluding Grants, Status efforts. However, since it will be more difficult to Quo % GDP collect domestically than at the border, it does not 28.0% fully compensate for the lost border revenue. Domestically-collected VAT are thus assumed to 27.0% increase by 2 percent above nominal the GDP per year. Personal and corporate income tax 26.0% collections are also expected to increase (as they 25.0% have in recent years) by 1 percent above the nominal GPD per year. Property taxes are 24.0% expected to grow in line with the nominal GDP while non-tax revenues are assumed to evolve 23.0% according to the average of the previous four years, without steps to boost, for example, 22.0% transport-related revenues. Even under these 21.0% optimistic assumptions, domestic revenue will not 2010 2011 2012 2013 2014 2015 2016 2017 2018 compensate for lost border revenues and total revenues (excluding grants) will decline from 25.8 No reforms Slow growth percent of GDP in 2013 to 24.4 percent by 2018. This figure could be even lower if the growth rate Source: World Bank staff estimates. is only half of the 4.5 percent per year forecast in the baseline (Figure 9.1). 9.7. Even with the continuation of current policies, expenditure patterns will moderately change over the medium term. The pace of construction of Route 6 is expected to be slower and its cost lower than that of Route 7 helping to reduce the share of GDP spent on road construction from 5.7 percent of GDP in 2012 to 4.3 percent by 2018 (Figure 9.2 and Figure 9.3). Subsidies for the energy sector are projected to remain constant in nominal terms, thereby falling as a share of GDP from 0.8 to 0.6 percent. Spending on social protection is projected to increase from 3.9 percent of GDP to 4.4 percent as the policy of expanding war veterans’ benefits continues. Public sector wage increases costing €120 million (over 2 percent of GDP) in 2015 and around three quarter of this in 2014 drive the increase in spending 124 projected in most other areas, many of which spend a large proportion of total resources on wages, including health (2.6 to 3 percent of GDP excluding donor grants); education (3.9 to 4.8 percent of GDP excluding donor grants) and public order and defense (3.0 to 3.5 percent of GDP). 9.8. Unfortunately, the general population is unlikely to reap many benefits from these moderate changes in spending patterns. While on the surface, the projected budget composition seems reasonable, in practice it does little to address Kosovo’s underlying challenges. The foreca sted increase in social protection spending is due to the expansion of war-related benefits – which are conservatively estimated to cost around €40 million per year – rather than efforts to protect the poor or vulnerable. The spending increases in health and education are driven by salary increases rather than significant additional resources to improve the quality of services. Figure 9.2. Spending Composition in 2012 Figure 9.3. Spending Composition Excluding Excluding Grants, Status Quo % GDP Grants in 2018, Status Quo % GDP Social Social protection, Other Other protection, 4.4% Economic Economic 3.9% Affairs, Affairs, 2.1% 2.3% Other, Other, 6.6% 6.9% Energy, Energy, 0.8% Public 0.6% Public order and order and defense, defense, 3.0% Road 3.5% Road Health, constructio 2.6% Health, constructio n, 4.3% Education, 3.0% n, 5.7% Education, 4.8% 3.9% Source: World Bank staff estimates. Source: World Bank staff estimates. 9.9. Under this status quo scenario, the combination of falling revenues and spending promises will result in an increasing fiscal deficit or a crowding out of essential spending. Increases in public sector salaries and war-related benefits and continued high spending (albeit lower than in the recent past) on new highways will result in a small non-grant spending increase – from 28.7 percent of GDP in 2013 to 30.1 percent in 2016. This and declining border revenues combine to increase the primary fiscal deficit to over 5 percent of GDP by 2016 (Figure 9.5) in the status quo scenario – significantly higher than the 2 percent mandated by the Fiscal Rule (see Economic and Fiscal Context Chapter). The only way authorities could contain the deficit while continuing to fund Route 6 and new veterans’ benefits would be to cut spending in other essential areas. Indeed, the authorities have already announced that they hope to reduce spending on goods and services by 15 percent in 2014 through “efficiency gains”. Such large efficiency gains in a single year seem hopeful, and could even be harmful if they are not carefully managed to avoid cuts in vital services. 125 Figure 9.4. Expenditures Excluding Grants, Figure 9.5. Primary Deficits Excluding Net Status Quo % GDP Lending and Grants, Status Quo % GDP 30.5% 7.0% 30.0% 6.0% 29.5% 5.0% 29.0% 4.0% 28.5% 3.0% 28.0% 2.0% 27.5% No reforms 27.0% 1.0% No reforms and slow growth Fiscal rule 26.5% 0.0% 2010 2011 2012 2013 2014 2015 2016 2017 2018 2010 2011 2012 2013 2014 2015 2016 2017 2018 Source: World Bank staff estimates. Source: World Bank staff estimates. Adaptive scenario 9.10. An alternative budget is possible. This alternative scenario shows how revenue and spending patterns can be adapted to better meet Kosovo’s Figure 9.6. Revenues, Excluding Grants, current needs, in line with the recommendations in Adaptive Scenario, % GDP this report. In particular it includes increased domestic revenue collections and, on the spending 27.0% side: (i) new war veterans’ benefits are not 26.5% introduced; (ii) large public sector wage increases 26.0% planned in 2014 do not occur; (iii) the construction of Route 6 – much of which had low economic 25.5% rates of return – is delayed beyond the period 25.0% analyzed; and (iv) subsidies to firms in the energy sector are halved. 24.5% 24.0% 9.11. A combination of strengthened tax administration and some policy changes could 23.5% help boost revenues. Under the adaptive scenario, 23.0% the declining border revenues are compensated by 2010 2011 2012 2013 2014 2015 2016 2017 2018 alternative sources of revenue. Health insurance is estimated to raise around 0.8 percent of GDP No reforms Reforms when it is introduced from 2016 (see Health Chapter for details of estimation).223 Efforts to Source: World Bank staff estimates. collect more property tax including registering property and raising the tax rates (which are low compared with most countries in SEE) are estimated to almost double revenue from this source to around 0.5 percent of GDP by 2018. It is difficult to estimate 223 This assumes a contribution rate of 5 percent, following the recommendation in the Health Chapter of this report. Higher revenues would result from higher contribution rates. 126 the impact of efforts to increase compliance (such as pursuing tax evaders in court), and other revenue reforms (such a policy choices to change tax rates or reform the structure of PIT). However, halving the CIT rate in 2009 reduced revenues by only 19.5 percent224, therefore an increase from 10 to 15 percent is unlikely to increase revenues by 50 percent. The elasticity is assumed to be 1.5 times greater than during the 2009 change since many CIT payers will not find it easy to exit the tax system. As a result, CIT collections will increase from 1.3 to 1.6 percent of GDP between 2014 and 2018. Additional enforcement efforts in PIT and domestic VAT increase collections from these sources such that they rise 2 percent faster than the nominal GDP (compared with 1 percent in the status quo). These reforms together help raise revenues from 25.8 percent of GDP in 2013 to 26.2 percent in 2018 – significantly higher than the 24.4 percent projected without reforms (Figure 9.6). 9.12. The savings made from Route 6, veterans’ benefits, reduced energy subsidies and wage increases can be invested in other, more urgent, areas. Funds saved from these measures compared with the status quo are significant: around €40 million (0.8 percent of GDP) less per year would be spent on war veterans; the public sector wage bill would be reduced by around €45 million (0.9 percent of GDP) per year initially and more by eliminating future ad hoc wage increases; reduced energy subsidies are estimated to save €20 million per year (0.4 percent of GDP) but this could easily be higher. Finally, delaying Route 6 would save a total of €750 million between 2014 and 2018 (15 percent of 2013 GDP). The adaptive scenario redistributes these savings with the aim of addressing some of the key challenges discussed in this report and ensuring that the fiscal deficit adheres to the Fiscal Rule. Specific changes are:  Public servants are awarded more regular wage increases based partly on merit. The adaptive scenario assumes €10 million additional is spent on wages each year. This would be particularly beneficial for the education sector, where merit-based pay increases can help to improve the quality of teaching.  Some of the funds saved from Route 6 can be invested in improving local and regional roads where evaluations show high economic returns. Between 2014 and 2018, an additional €380 million can be allocated to road maintenance and the construction of smaller roads. This includes €171 million estimated in the Transport Chapter as the cost of improving the 39 local roads whose upgrades have been identified has having positive economic rates of return and €230 million (€46 million annually) for the maintenance and upgrade of regional and some national roads. An additional €25 million is provided as a “cushion”.  Savings can also be used to improve the education system – notably early childhood education – essential if Kosovo is to utilize its young population to generate future growth. The alternative budget allocates between €21 and €24 million (around 0.4 percent of GDP) per year additional resources for the education budget compared with the status quo. These additional resources are intended to be used to improve the quality of education rather than across-the-board salary increases.  Savings made by foregoing new war benefits can be used to expand and improve the existing social protection mechanisms – especially important in the face of rising energy prices and the introduction of obligatory health insurance. In total, the last-resort social assistance – or a similar program – is increased by between €21 and €24 million compared with the status quo, an increase of around 75 percent.  Additional resources for the health sector – of around €40-50 million (0.4 to 0.5 percent of GDP) per year – are fully funded by the introduction of health insurance, which does not affect spending patterns for other ministries since public sector employer contributions should be reimbursed by the Health Insurance Fund. 224 As shares of GDP. 127 9.13. These changes would adjust the budget composition. Under this adaptive scenario, spending on road construction is projected to be 2.2 percent of GDP by 2018, compared with 4.3 percent in the baseline projections (Figure 9.7 and Figure 9.8). Subsidies to the energy sector would halve as energy tariffs move closer to the cost recovery rate. Other economic affairs such as road maintenance would be slightly higher at 2.7 percent of GDP compared with 2.3 percent under the status quo scenario. Spending on social protection would increase from 4.4 percent to 4.7 percent of GDP to help combat poverty and moderate the impact of energy costs increases on the poor and vulnerable. Spending on education and health would also both be higher than under the status quo (education at 4.8 percent of GDP compared with 4.5 percent in the status quo and health at 3.7 percent compared with 2.9 in the status quo), the latter funded through health insurance premiums. 9.14. This re-prioritization of expenditure should go hand-in-hand with efforts to improve efficiency. Spending increases in health and education in particular should also be met with improvements in the efficiency of spending – as discussed in the relevant chapters – to ensure that real improvements in services are delivered. At the same time, tighter controls over procurement could reduce costs by reducing corruption. The construction sector (including roads) in particular could reap significant benefits. Figure 9.7. Evolution of Spending Figure 9.8. Spending Composition Excluding Composition, Adaptive Scenario % GDP Grants in 2018, Adaptive Scenario % GDP 35% Social protection, 30% 4.7% Other , 6.7% 25% Other 20% Economic Affairs, 15% 2.7% 10% Energy, 0.3% Public 5% Road order and constructi defense, on, 2.1% 3.3% 0% 2010 2011 2012 2013 2014 2015 2016 2017 2018 Other Public order and defense Health Education Education, Health, Road construction Energy 4.8% 3.7% Other Economic Affairs Social protection Source: World Bank staff estimates. Source: World Bank staff estimates. 9.15. Under the adaptive scenario, the fiscal deficit is controlled thanks to the increased revenues and reduced total expenditures. Total non-grant expenditures decline from 28.7 percent of GDP in 2013 to 27.8 percent in 2015 (Figure 9.9) as savings are made on Route 7 and from lower energy subsidies, planned public sector wage increases are introduced more gradually, and ill-advised plans for war veterans’ benefits are shelved. These spending reductions combined with increased revenues help to contain the fiscal deficit, which is maintained at 2 percent or slightly under in every year after 2015 in the adaptive scenario (Figure 9.10). 128 Figure 9.9. Expenditures Excluding Grants, Figure 9.10. Primary Deficits Excluding Net Adaptive and Status Quo Scenarios % GDP Lending and Grants, Adaptive and Status Quo Scenarios % GDP 30.5% 6.0% 30.0% 5.0% 29.5% 29.0% 4.0% 28.5% 3.0% 28.0% 27.5% 2.0% 27.0% 1.0% No reforms Reforms 26.5% 2010 2011 2012 2013 2014 2015 2016 2017 2018 Fiscal rule 0.0% 2010 2011 2012 2013 2014 2015 2016 2017 2018 No reforms Reforms Source: World Bank staff estimates. Source: World Bank staff estimates. Energy Crisis Scenario 9.16. If public sector investment in the power sector is needed to ensure a sufficient energy supply, there will be fewer resources for most other sectors (Figure 9.11). To maintain the same low fiscal deficits as in the adaptive scenario, around €550 million saved from the postponement of the construction of Route 6 could be made available to bolster energy supply (Figure 9.12) , likely through the construction of a new power plant, the KRPP. Some additional funds would still need to be made available to protect the poor from rising energy prices and these could be found by resisting the temptation to increase war-related benefits, as in the adaptive scenario. Many of the other efficiency improvements discussed throughout this report could also be implemented – for example, awarding salary increases based partly on performance. In addition, since the increase in health resources will come from health insurance, health spending will also increase. However, there would be no other areas would see spending increases above those presented in the baseline as resources are diverted toward the construction of a new power plant. A summary of the main assumptions used to estimate each of these three scenarios is presented in Table 9.1 below. 9.17. Even with these additional resources, Kosovo would need to make difficult choices in the energy sector. The available resources may be sufficient or close to sufficient to construct one of the two planned units of the KRPP. This would provide Kosovo with 300 MW of generation capacity instead of the 600 MW desired, sufficient to close the power gap in the short term but at a cost of losing some economies of scale and there would remain a need to construct additional generation facilities in the longer term. Alternatively, the GoK could invest in a company established to build a 600 MW power plant and seek private sector investors and/or donors to invest the remaining required funds. A third choice could be to raise the additional funds through public debt. But, while Kosovo has low public debt it has little experience in debt markets. Kosovo’s public debt is around 10 percent of GDP, and would remain at under 20 percent even if it inherited its share of ex-Yugoslav debt225, and it has only a small and 225 This would occur if Serbia, which is currently servicing this debt, recognizes Kosovo as an independent state. 129 shallow domestic debt market226. Kosovo has still not issued a long-term bond domestically and has had no exposure in international markets. Issuing the long-term bond required to pay for a new power plant would therefore be challenging and may prove to be a more expensive option than it would be if Kosovo had more experience in the debt market. Figure 9.11. Evolution of Spending Composition Figure 9.12. Potential Resources Available for Under Energy Crisis Scenario, % GDP KRPP, € Mil. 35% 160 30% 140 25% 120 20% 100 15% 80 10% 60 5% 40 0% 2010 2011 2012 2013 2014 2015 2016 2017 2018 20 Other Public order and defense Health Education 0 Road construction Energy 2014 2015 2016 2017 2018 Other Economic Affairs Social protection Source: World Bank staff estimates. Source: World Bank staff estimates. 226 Kosovo began to issue T-Bills only in 2011 beginning first with three month issuances and increasing to 12 month issuances in late 2013. 130 Table 9.1. Key Assumptions for each Scenario Status Quo Scenario Adaptive Scenario Energy Crisis Scenario Revenues Border Border revenues fall by a total of 1.6 percent of Border revenues fall as per the Status Quo revenues GDP in 2015 and 2016 as trade agreements with scenario. Turkey and the EU come into force. Domestic Domestically collected VAT increases by 2 percent Domestic VAT collections increase by 2 percent VAT above nominal GDP as domestic production above nominal growth annually thanks to better substitutes for some imports. enforcement. PIT PIT increases by 1 percent above nominal GDP on PIT collections increase by 2 percent above the back of increased domestic production and nominal growth annually thanks to better employment. enforcement. CIT CIT increases by 1 percent above nominal GDP on CIT is increased to 15 percent in 2016 but revenues the back of increased domestic production and go up by around 30 percent (not 50 percent) as a employment. result, 1.5 times the elasticity of the impact on Revenues perform as per the Adaptive Scenario. revenues when CIT was reduced from 20 to 15 percent in 2009. Property Property taxes grow in line with nominal GDP. Property tax revenue nearly doubles between 2013 taxes and 2018 thanks to increased collection efforts. Non-tax Non-tax revenues increase according to the Non-tax revenues increase according to the revenues average of the previous four years. average of the previous four years plus an additional amount for transport related increases. Health No health insurance Health insurance is introduced from 2016 at a rate insurance assumed to be 5 percent of gross salaries, raising around 0.8 percent of GDP per year. Expenditures Savings Energy Energy subsidies are reduced by half, saving subsidies around €20 million per year. War The proposed new veterans' benefits are not veterans' introduced, saving around €40 million per year. benefits Public sector Public sector wage increases planned for 2014 are Savings are made as per the Adaptive Scenario. wage not introduced, saving €45 million per year. increases Route 6 Delaying the construction of Route 6 to after 2018 saves €750 million in total between 2014 and 2018, or an average of €150 million per year. Alternative Spending Energy Subsidies for the energy sector remain constant in Subsidies to energy sector are cut by half. subsidies nominal terms. War War veterans' benefits cost around €40 million per veterans' year from 2015. benefits Public sector Public sector wages increase by €45 million in 2014 Public sector wages are increased by €10 million Public sector wages increase as per the Adaptive wage and by €30 million in 2017. each year, partly based on performance. Scenario. increases Route 6 Route 6 construction costs €750 million between €380 million of the €750 million saved from delaying 2014 and 2018, of which €150 million is for Route 6 is invested in road maintenance and expropriation. €130 million is spent in 2014 and upgrade over the whole period. €140-€160 million in each of the following years. Education An additional €21-€24 million more than the Status Quo scenario per year is allocated to education (and used in goods and services, maintenance and other required investment rather than wage increases, which are taken into account in the performance-based public sector wage increases). Social The last-resort social assistance scheme receives The last resort social assistance scheme receives protection for additional funding of between €21 and €24 million funding as per the Adaptive Scenario. the poor and above the Status Quo scenario per year to expand vulnerable coverage of the poor and help to shield them from energy price increases. Health Health sector spending increases by between €40 Health sector spending increases as per the and €50 million per year above the Status Quo Adaptive Scenario, fully funded by health scenario from 2016, fully funded by health insurance premiums. insurance premiums. Deficit Remaining savings from the spending reductions Deficit is maintained at below 2 percent of GDP as reduction are used to ensure the deficit is below 2 percent of per the Adaptive Scenario GDP as foreseen by the Fiscal Rule. Construction All remaining savings not required to maintain the of new power deficit at 2 percent of GDP are channeled to the generation energy sector. This amounts to around €550 million facilities in total over the period. 131 B. CONCLUSIONS 9.18. Kosovo has an opportunity and a need to produce an alternative budget focused on priorities that will meet its long term growth needs while keeping its fiscal deficit under control. Without revenue-boosting measures and expenditure re-prioritization, there would either be a rapid rise in the fiscal deficit or a crowding out of essential spending. However, there is room both to bolster domestic revenues and cut spending on less urgent areas. The resources saved during this process can be invested in other, more urgent, areas. This re-prioritization of expenditure should go hand-in-hand with efforts to improve efficiency. 9.19. Without measures to bolster domestic revenues, total revenues are projected to decline as a share of GDP over the medium term, but a combination of strengthened tax administration and some policy changes could help to boost them. The major contributors to declining revenues are related to imports and non-tax revenues. Some domestic taxes are expected to partly compensate for these declines but these will not likely be sufficient without additional measures in fighting tax evasion and some policy changes – such as raising CIT, introducing health insurance and enforcing property taxes. 9.20. Continuing with existing policies without change will do little to address Kosovo’s underlying challenges. As such, it will not protect the poor nor will it improve outcomes in health and education as the key sectors for economic growth. A different scenario is needed. 9.21. Restructuring the budget would help contain expenditure and focus it on the most needed areas. The adaptive scenario would see expenditure fall as a share of GDP initially as wage increases are spaced more evenly, new war-related benefits are not introduced, and spending on new highways is reduced. Public sector resources may be required to prevent a looming energy crunch though different financing options should be explored. 9.22. Spending efficiency improvements are necessary across the board. Spending increases in health and education in particular should also be met with improvements in the efficiency of spending to ensure that real improvements in services are delivered. At the same time, tighter controls over procurement could reduce costs by reducing corruption. 132 ANNEX 1: DATA ON MUNICIPALITIES Figure A.1. Municipalities In Kosovo. Source: WB Staff calculations based on Kosovo 2011 Population Census and OSCE. Note: See Code Numbers in Table A., Annex 1 for each municipality 133 Table A.1. Available Data Sources Per Municipality. Code Albanian Serbian BOOST EMIS CENSUS Percentage CENSUS Number name name BOOST name notes notes SOURCE of Serbs notes 1 Deçan Dečane 631 Municipality of Deqan 2011 Census 0.0 percent 2 Dragash Dragaš 621 Municipality of Dragash 2011 Census 0.0 percent 3 Ferizaj Uroševac 656 Municipality of Ferizaj 2011 Census 0.0 percent 4 Fushë Kosovo Polje 612 Municipality of Fushe 2011 Census 0.9 percent 5 Kosovë Gllogoc Glogovac Kosova 611 Municipality of Gllogovc 2011 Census 0.0 percent 6 Gjakovë Djakovica 632 Municipality of Gjakova 2011 Census 0.0 percent 7 Gjilan Gnjilane 651 Municipality of Gjilan 2011 Census 0.7 percent 8 Hani i Elezit General Jankovič 659 Municipality of Hani i Elezit 2011 Census 0.0 percent 9 Istog Istok 633 Municipality of Istog 2011 Census 0.5 percent 10 Junik Junik 636 Municipality of Junik 2011 Census 0.0 percent 11 Kaçanik Kačanik 652 Municipality of Kaqanik 2011 Census 0.0 percent 12 Kamenicë Kamenica 653 Municipality of Kamenica 2011 Census 4.3 percent 13 Klinë Klina 634 Municipality of Klina 2011 Census 0.3 percent 14 Leposaviq Leposavič 641 Municipality of Leposaviq NO SERBS OSCE 96. 0 percent NO SERBS 15 Lipjan Lipjan 613 Municipality of Lipjan 2011 Census 0.9 percent 16 Malishevë Mališevo 625 Municipality of Malisheva 2011 Census 0.0 percent 17 Mamush Mamuša 626 Municipality of Mamusha 2011 Census 0.0 percent 18 Mitrovicë Mitrovica 642 Municipality of Mitrovica 2011 Census 0.0 percent 19 Novobërdë Novo Brdo 654 Municipality of Novoberdo NO SERBS OSCE 60.0 percent 20 Obilic Obilič 614 Municipality of Obiliq 2011 Census 1.3 percent 21 Pejë Peč 635 Municipality of Peja 2011 Census 0.3 percent 22 Podujevë Podujevo 615 Municipality of Podujeva 2011 Census 0.0 percent 23 Prishtina Prishtina 616 Municipality of Prishtina 2011 Census 0.2 percent 24 Prizren Prizren 622 Municipality of Prizren 2011 Census 0.1 percent 25 Rahovec Orahovac 623 Municipality of Rahovec 2011 Census 0.2 percent 26 Skënderaj Srbica 643 Municipality of Skenderaj 2011 Census 0.1 percent 27 Suharekë Suva Reka 624 Municipality of Suhareka 2011 Census 0.0 percent 28 Shtërpcë Štrpce 655 Municipality of Shterpce NO SERBS OSCE 66.7 percent 29 Shtime Štimlje 617 Municipality of Shtime 2011 Census 0.2 percent 30 Viti Vitina 657 Municipality of Vitia 2011 Census 0.2 percent 31 Vushtrri Vučitrn 644 Municipality of Vushtrri 2011 Census 0.5 percent 32 Zubin Potok Zubin Potok 645 Municipality of Zubin Potok NO SERBS OSCE 93.3 percent NO SERBS 33 Zveçan Zvečan 646 Municipality of Zveqan NO SERBS OSCE 96.1 percent NO SERBS 34 Graçanicë Gračanica 618 Municipality of Graqanica 2010 NO OSCE 86.2 percent 35 Kllokoti Klokot 660 Municipality of Kllokot 2010 NO OSCE 72.1 percent 36 ZAMV AKSM 647 Ad. Office North Mitrovica 2012 NO OSCE 76.5 percent NO 37 Partesh Parteš 658 Municipality of Partesh 2012 NO OSCE 100.0 percent 38 Ranillug Ranilug 661 Municipality of Ranillug 2010 NO OSCE 98.5 percent Albanian Majority Serb Majority Turkish Majority Source: EMIS, BOOST, Kosovo Census, OSCE. Note: In non-recognized Serb and Serb majority municipalities, data on census is extracted from OSCE. This has been done given that there is few or no data on Serbian population in the 2011 Census. 134 ANNEX 2: TRANSPORT Table A.2. Railway Finances Summary Business Plans 2013 INFRAKOS Costs € Share Revenues € Share Staff 1,393,125 26.0% Track charges 894,500 15.3% Materials 2,962,100 56.0% Other services 838,600 14.3% Overheads 964,375 18.0% Subsidies 2,255,100 38.6% 100.0% Access fee loss 1,861,000 31.8% Total 5,319,600 100.0% Total 5,849,200 100.0% TRAINKOS Costs € Share Revenues € Share Staff 1,412,220 23.0% Service revenues 2,680,564 43% Materials 3,413,381 55.0% PSO 1,241,308 20% Access fees 1,246,207 20.0% MED subsidy 500,000 8% Other 159,903 3.0% Subsidy 1,400,000 22% Reserves 427,572 7% Total 6,231,711 100.0% Total 6,249,444 100% INFRAKOS and TRAINKOS combined Costs € Share Revenues € Share Staff 2,805,345 27.2% Service revenues 2,680,564 28.7% Materials 6,375,481 61.9% Other services 838,600 9.0% Access fees 0 0.0% PSO 1,241,308 13.3% Overheads 1,124,278 10.9% Subsidies 4,155,100 44.5% From reserves 427,572 4.6% Total 10,305,104 100.0% Total 9,343,144 100.0% Deficit 961,960 10.3% Sources: INFRAKOS and TRAINKOS Business Plans for 2013 International Benchmarks of Transport Investment This section of the Transport Annex presents in detail international benchmarks of transport investment. Kosovo’s investment in transport infrastructure is compared with three benchmarks from the International Transport Forum (ITF)227 and a study of eleven countries in North Africa and the Middle East (including Turkey), collectively known as the MED 11.228  transport investment expenditure as a percentage of GDP;  the share of that investment between modes, and;  the share of that investment allocated to maintenance. 227 International Transport Forum, Statistics Brief, 2012. 228 Transport Infrastructure for MED11 Countries, CASE Network Reports 108, 2012. This assessment undertaken in 2011 for the EU based on a method used in an Africa Infrastructure Country Diagnostic (AICD) undertaken by the World Bank in 2009. 135 The ITF values relate to actual transport sector expenditure, without assessing whether these levels of expenditure are sufficient. The MED11 values on the other hand, are based on the investment that would be needed to achieve specific connectivity standards to provide capacity for the geographic and economic size of the countries. The ITF benchmarks relate to OECD countries where transport infrastructure is mostly in good condition, so the maintenance investment is relatively low. In contrast, the MED11 values relate to infrastructure that has not been so well maintained in the past, so the maintenance benchmark is higher. Both the ITF and MED11 data relates to all modes of transport, but only those relating to roads and railways are used here. Table A.3. Comparison of Kosovo per capita income in benchmark countries GDP per Region/Country Kosovo share capita (U$) Kosovo 3,454 100% AICD 7,862 44% MED11 23,182 15% EU27 34,064 10% OECD 37,010 9% Source: OECD Stat. extracts: http://stats.oecd.org/index.aspx?queryid=558 Total transport investment expenditure as a % of GDP In Western Europe, the transport infrastructure investment share of GDP has declined steadily from 1.5 percent in 1975 to 1.2 percent in 1980 and further to 1.0 percent in 1982 after which it levelled off. An investment level of 1 percent per GDP remained a norm for many years, such that it became de facto political benchmark in the 1980s, though with no theoretical basis. The investment needs for transport infrastructure depend on a number of factors, such as the quality and age of the existing infrastructure, geography of the country and transport-intensity of the country’s productive sector, among other things. The fact that the share of GDP dedicated to transport infrastructure has tended to remain constant across many countries and over time suggests investment levels are affected by factors other than real investment needs, such as institutional budget allocation procedures or budgetary constraints. The share of investment in inland transport infrastructure in Central and Eastern European countries, which until 2002 had also remained at around 1.0 percent of GDP, increased significantly from then on reaching a peak of about 2.0 percent in 2009. The decline to 1.7 percent in 2010 appears to be consequence of the financial crisis. Figure A.2. Transport investment as a % of GDP for global regions Source: International Transport Forum, Statistics Brief, 2012. 136 Investment in different modes of transport The share of rail investment in the transport total for OECD countries increased from 15 percent in 1995 to 23 percent in 2010. In Western Europe the rail share has been higher, increasing from 20 percent in 1995 to 30 percent by 1975 and about 40 percent by 2010 – a reflection of maturity of the road networks and political commitment to development of railways. Whereas Western European countries (WEC) have increasingly directed investment toward rail, Central and Eastern European countries (CEEC) are still investing more heavily in roads to achieve a comparable maturity of their networks. Their roads share increased from 65 percent in 1995 to about 80 percent in 2005 and remained at about this level through 2010. Table A.4. ITF estimates of the mode shares of transport investment Mode Region 1995 2010 Rail OECD 15% 23% WEC 20% 30% Road CEEC 35% 20% Source: Authors analysis of ITF data Investment in road maintenance The balance between road maintenance and new road investment has been relatively constant over time in many regions, with the maintenance share being between 25 percent and 35 percent in both Western Europe and Central and Eastern Europe. However, there are significant differences between individual countries. Those countries that have a higher maintenance share tend to have both a more mature road network and less need to invest in new roads, but also a higher quality of road network. Figure A.3. Maintenance share of road investment Source: International Transport Forum, Statistics Brief, 2012. EU Road and Vehicle Registration Fees Many EU countries (20 out of 27) apply a registration tax/charge/fee/excise duty on at least some share of vehicles (re)entered into the fleet. Estonia, Germany, Lithuania, Luxemburg, the Slovak Republic, Sweden and the UK are the countries not levying a tax on vehicle registration. Many EU countries use this as an opportunity to promote or discourage certain vehicle types. In nine Member States (Austria, France, 137 Latvia, Malta, the Netherlands, Portugal, Romania, Slovenia and Spain) and in the Flemish region in 19 Belgium , CO2 emissions (expressed in g/km) are the main parameter in registration taxes. Combinations of fuel type, important indicators for emissions of local pollutants such as nitric oxides (NOx) and particulate matter, are used by half of the EU countries that have a registration tax. Half of the EU countries also use the value (for new vehicles equivalent to purchase price, for older vehicles the age or accumulated mileage is also considered) of the vehicle as the starting point for the tax calculation. Four vehicle types have been used as the basis for a comparison between the charges for vehicle registration and use; a small car such as a Peugeot 207, a medium car such as Volkswagen Golf; a large car such as a Ford Mondeo, and; a 400hp five axle articulated truck with a gross weight of 40,000kg. In 2012 the average registration fee for the smallest cars was €1,700, for the largest car €7,000 and for a truck €1,400. Excluding Denmark which had registration fees of over €10,000 for all vehicle types, the next highest fees for cars were in the Netherlands and Malta (over €5,000) while Greece had the highest rate for trucks of €6,000. Eight EU countries had fees of over €1,000 for cars and twelve had fees of over €1,000 for the largest group of cars. The comparable rates in Kosovo are €25 for cars and €50 for trucks (plus €30 for plates and other first registration charges). Vehicle ownership and use taxes All EU countries levy periodic ownership or circulation taxes. Contrary to registration taxes, heavy duty vehicles are subject to this form of taxation in nearly all countries, often with a separate regime targeting only this class of vehicles. Many countries are in a transition from one form of taxation to another, usually from engine size/power to CO2 emissions. As such, the system that determines the tax level depends on the year of first registration, with newer vehicles benefitting from modern technologies and the lower ratio CO2/engine size this provides them with. Treatment of diesel vehicles differs significantly. Countries that have a tax based on CO2 emissions only, favor the use of diesel powered vehicles. The list includes France (only company cars), Greece, Ireland, Malta and the UK. The Netherlands and Sweden on the other hand impose much heavier charges on diesel than on gasoline powered vehicles. Finland does both: a pure CO2 based tax for all vehicles, plus an additional tax for diesel vehicles only, which results in a situation similar to that in the Netherlands and Sweden. 138 Figure A.4. Registration Fees for Four Typical Vehicles in EU Countries Source: An Inventory of Measures for Internalizing the External Costs in Transport, DGMOVE, EU, 2012. Taxation of heavy goods vehicles is mainly based on a combination of their gross vehicle weight (GVW), axle configuration and suspension type, targeting trucks that impose high road maintenance costs. Countries using only GVW as the charge base are missing an important part of the correlation between vehicle properties and infrastructure damage, as a truck with more axles will cause less road damage than one with the same GVW but fewer axles. Figure A.5. Use Fees for Four Typical Vehicles in EU Countries Source: An Inventory of Measures for Internalizing the External Costs in Transport, DGMOVE, EU, 2012. 139 Trade and Transport Infrastructure As a land-locked country, Kosovo’s container transport costs are predictably high. As Table A.6 shows, Kosovo’s largest trading partners are nearly all neighboring countries, and with the possible exception of Italy, cargo are moved via roadway rather than rail. A large share of trade with Italy, which is dominated in value and volume by imports, uses maritime transport through Albania (ro-ro service from Durres in Albania to Bari in Italy). Imports from China come on feeder services to regional hub ports such as Rijeka in Croatia and then on to Kosovo via road. The poor quality and limited nature of the transport infrastructure is reflected in the high cost of freight container exports and imports. In comparison with thirteen SEE countries, Kosovo ranks next to last after Moldova, with respect to transport, logistics and trade facilitation costs: it has the highest export costs and the third highest import costs. Table A.6 provides some key regional comparisons from the World Bank’s Doing Business survey. Table A.5. Kosovo’s top 5 export destinations and import origins Export destination Share Import origins Share Albania 15.70% Serbia 12.40% Italy 14.90% Italy 9.60% India 11.70% Germany 8.90% Macedonia 10.20% Macedonia 8.90% Serbia 7.40% Turkey 6.60% Montenegro 6.00% China 6.50% Share of top six destinations 65.90% Share of top six origins 52.90% Total export value (2011) €310m Total import value (2011) €2,400m However, while high transport costs may impact trade, the trade imbalance also contributes to high transport costs. Kosovo has a large trade imbalance with imports between nine and ten times greater that exports in terms of value (but less in terms of volume as the unit value of imports is higher than that of exports). The large trade imbalance prejudices the cost efficiency of the transport and logistics of Kosovo’s international trade. A possible explanation for the large difference between import and export cost might be found in the loading of trucks in each direction – inbound trucks tend to be fully loaded, while export trucks are very lightly loaded. 140 Table A.6. Trading Across Borders, 2013 Cost to Cost to Documents to Time to Documents Time to export import Country Rank export export to import import (US$ per (US$ per (number) (days) (number) (days) container) container) Georgia 43 4 9 1,355 4 10 730 Slovenia 48 5 16 745 7 14 1,810 Greece 52 4 16 1,040 6 15 1,135 Montenegro 53 6 14 985 5 14 1,200 Hungary 70 6 17 885 6 19 1,760 Romania 76 5 13 1,485 6 13 1,480 Bulgaria 79 4 20 1,375 5 17 1,365 Albania 85 7 19 745 8 18 1,595 Serbia 98 6 12 1,455 7 15 2,175 Croatia 99 7 18 1,335 7 15 1,495 Bosnia & Hertz. 107 8 16 1,260 8 13 985 Slovakia 108 7 17 1,500 6 16 1,870 Kosovo 121 8 15 1,775 7 15 1,185 Moldova 150 7 32 1,545 8 35 1,200 Source: Doing Business 2013, World Bank. Table A.7. Kosovo’s Top Five Export and Import Products Export products Share Import products Share Ferro-alloys: 31.4% Petroleum and petroleum 13.8% products Ferrous waste and scrap 8.7% Motor vehicles 4.2% Lead ores and concentrate 5.5% Iron and steel bars 3.3% Copper waste and scrap 4.5% Tobacco products 2.0% Raw hides and skin 3.9% Cement 1.9% Share of top 5 in total 54.0% Share of top 5 in total 25.2% 141 142 ANNEX 3: EDUCATION Table A.8. Comparator Country List: ECA and Non-ECA Countries Public expenditure on Country Share of Population GDP per capita (PPP) in education as percent of Under 19 (2010) 2011 GDP (2011 or latest) Albania 32.55 percent 3.27 percent 7848 Armenia 29.11 percent 3.21 percent 5405 Belarus 21.24 percent 5.41 percent 15056 Bosnia and Herzegovina 24.62 percent 0.00 percent 8115 Bulgaria 18.54 percent 4.58 percent 13812 Croatia 20.96 percent 4.42 percent 17850 Czech Republic 19.99 percent n/a 27112 Georgia 24.95 percent n/a 5502 Hungary 20.65 percent 5.12 percent 19571 Kosovo 38.08 percent 4.19 percent 6700 Latvia n/a 5.03 percent 16717 Lithuania n/a 5.67 percent 20343 Macedonia, FYR n/a n/a n/a Moldova n/a 9.11 percent 3382 Montenegro 26.71 percent n/a 11628 Poland 21.52 percent 5.09 percent 20013 Romania 20.73 percent 4.32 percent 12520 Russian Federation 21.06 percent n/a n/a Serbia 23.71 percent 4.71 percent 10405 Slovak Republic n/a n/a 23366 Slovenia 19.07 percent 5.70 percent 28436 Turkey 35.42 percent 2.86 percent 14543 Ukraine 19.98 percent 5.28 percent 7210 China 26.11 percent n/a 8391 Dominican Republic 41.02 percent n/a 9287 Ecuador 40.47 percent 4.89 percent 9552 El Salvador 43.77 percent 3.21 percent 7222 Guatemala 52.59 percent 2.80 percent 5093 Jamaica 38.76 percent 6.35 percent 9027 Thailand 26.50 percent 3.75 percent 9390 Upper middle income n/a n/a n/a Source: World Bank EdStats, IMF WEO, UN Population Data, Kosovo Census. 143 Figure A.6. Distribution of Reading and Math Scores by Gender Source: WB Staff calculations based on 5th grade Exams database. Figure A.7. Reading and Math Scores by Rural/Urban Divide Source: WB Staff calculations based on 5th grade Exams database. 144 Table A.9. Average scores in 5th Grade Exams 2010, by Municipality (out of 100 points) Achievement Municipality Mathematics Albanian language Deçan 50.4 52.5 Dragash 54.7 52.9 Ferizaj 44.8 54 F. Kosovë 52.2 60.3 Gllogovc 40.4 45.9 Gjakovë 54.5 57.8 Gjilan 55.9 61.4 H. Elezit 39.1 41.1 Istog 49.6 54.7 Junik 41.1 46.9 Kaçanik 51.1 53 Kamenicë 51.6 50.8 Klinë 44.6 50.2 Lipjan 44.3 51.9 Malishevë 44.4 46.3 Mamushë 45.3 56.1 Mitrovicë 50.9 54.3 N. Bërdë 29.2 35.6 Obiliq 42.4 56.4 Pejë 50.6 58.6 Podujevë 47 55.3 Prishtinë 43.9 60.1 Prizren 41.7 52.8 Rahovec 55.3 54.4 Skenderaj 42.4 48.5 Suharekë 41.6 50.1 Shtërpcë 46.5 39.8 Shtime 36.3 40.3 Viti 57.6 56.3 Vushtrri 50.6 52.8 Source: Report of Assessment of 5th Grade Pupils Achievement, MEST. Note: The results of Viti municipality are presented by MEST with validity concerns since there are doubts in the list of names. 145 Table A.10. Average Class Size and Student-Teacher Ratios By Level Of Education, Selected Municipalities Grades 1-9 Grades 10-13 Average class size STR Average class size STR 2011/12 2012/13 2011/12 2012/13 2011/12 2012/13 2011/12 2012/13 Prishtina 24.8 24.6 19.7 19.5 33.3 33.1 19.5 19.4 Podujeva 22.0 22.0 17.4 16.5 35.8 35.0 23.2 22.5 Mitrovica 22.3 22.8 18.1 17.5 31.2 29.6 19.6 18.7 Peja 23.3 23.4 17.4 17.5 25.4 26.1 15.8 16.2 Istog 23.1 22.3 17.7 17.0 34.8 30.0 19.8 19.5 Klina 22.9 23.0 17.1 16.7 28.4 28.0 19.6 19.3 Gjakova 23.0 22.3 16.7 16.1 33.6 33.0 19.8 20.2 Prizren 21.3 21.1 17.4 17.0 35.8 33.5 18.2 19.5 Suhareka 23.0 22.6 16.6 16.5 36.0 34.5 21.7 20.8 Ferizaj 23.5 23.7 17.5 17.5 36.5 35.8 21.5 20.5 Shtime 22.2 22.9 17.3 16.8 30.2 29.4 18.9 19.8 Kacanik 19.3 19.6 15.9 15.2 30.2 29.2 18.9 18.3 Gjilan 20.1 19.9 15.2 14.9 29.9 28.4 17.4 16.6 Total 22.4 22.3 17.2 16.8 32.4 31.2 19.5 19.3 Average Source: WB Staff calculations based on Kosovo EMIS. 146 Education Grants and Per Capita Funding This section of the Education appendix describes how the Education Grant from the central government to municipalities is supposed to work. Education Grants for each municipality = Wages and Salaries + Goods and Services + Capital outlays 1. Wages and Salaries = Total wages and salaries for (a) teachers multiplied by the average teacher salaries; and (b) administrators and support staff, multiplied by the Kosovo-wide average salaries.229 (a) Teachers: based on the pupil-teacher ratios of : Pupils per teacher Pupils per teacher Majority municipalities Minority municipalities Pre-school (Grade 0) 12 12 Basic (Grade 1-9) 21.3 14.2 Upper Secondary Gymnasium (Grade 10-12) 21.3 14.2 Vocational School (Grade 10-13) 17.2 11.5 Special Needs Education 7.9 230 7.9 Mountainous (above 700m) 14.9 14.2 - An additional 3.33 percent of the total number of regular teachers to cover the cost of substitute teachers on sick and maternity leave; - 10 percent of total allocation to cover cost of additional English teachers.231 (b) Administrators and support staff: according to the following criteria: Staff position Principal per school with < 250 pupils 0.5 Principal per schools with > 250 pupils 1.0 Deputy principal per schools > 1000 pupils 1.0 Librarian per school > 1500 pupils 1.0 Pedagogue per school > 1500 pupils 1.0 Administrator for each school > 500 pupils 1.0 Pupil to administrator ratio in basic schools 675 Pupil to administrator ratio in upper sec. schools 640 Pupil to cleaner ratio 170 Number of guards for each schools 1 2. Goods and Services = A fixed amount per school (€1500 for pre-primary and primary and €3,250 for secondary) and a per pupil amount (€23 per the majority and €25 per the minority pupil). 229 MEST Administrative Instruction No. 18/2009 on “Pupil-teacher ratios and non-teaching staff ratios and numbers of pupils per class are defined for the purposes of determining (i) the state education grant to municipalities and (ii) staffing allocations made by municipalities to schools” issued in August 28, 2009. 230 For 2014, the budget allocated to special needs education will be attached to the specific education grant allocation. Until now, the budget was allocated to MEST and then distributed to schools. 231 This is to cover the cost of additional teaching staff needs as per MEST new policy to start teaching English classes from grade 1. The policy was introduced in 2011. The additional allocation is for classes in grades 1 and 2 and the assumption is that each class receives two hours of English classes a week (10 percent of 20 hours). 147 3. Capital Outlays = Capital outlays: €7 per pupil. Optimal, minimum and maximum number of children per class: The ratios presented above are used only for purposes of determining the grant for education to municipalities. Municipalities are not required to implement exactly these ratios when deciding to allocate the budget or staff in schools except in specific circumstances described in the table below: Table A.11. Maximum and minimum number of pupils by formula criteria School type Optimal Maximum Minimum Minimum Minimum class size (all) majority minority mountain areas (majority) (700m above sea level) Pre-school (Grade 0) 24 with two 26 10 10 5 educators Basic (Grade 1-9) 30 35 10 5 5 Gymnasium (Grade 10-12) 32 35 20 10 10 Vocational School (Grade 10-13) - Theoretical classes 32 35 20 10 10 - Practice classes 16 19 10 10 10 Figure A.8. Difference in Actual Expenditures and Education Budgeted Grant Per Pupil by Municipalities (2012) 140 120 100 National Average 80 Excess Euro 60 40 20 0 -20 -40 Source: BOOST Kosovo and Education Grant Data MEST. 148 School Enrollment Table A.12. Education System in Kosovo, 2012/2013 Education Level No. of pupils No. of schools Pre-school and pre-primary (ages 1-6) 25,706 - Kindergartens 5,389 - Pre-primary (age 5-6) 20,317 - Basic (Grades 1-9) 288,378 998 Upper Secondary 104,268 115 General (Gymnasium) (Grades 10-12) 44,852 54 Vocational Education (Grades 10-12/13) 59,416 61 232 Higher Education 49,841 2233 Total 468,193 1,115 Source: Kosovo EMIS. Figure A.9. Total Enrollments and Gross Enrollment Rates Trends in General Education 100% 200,000 180,000 80% 160,000 Primary 140,000 60% 2009/10 120,000 Lower 2010/11 100,000 40% Secondar 2011/12 80,000 y 20% 60,000 Upper 40,000 Secondar y (All) 0% 20,000 0-5 years 5-6 years Primary Lower Upper Secondary Secondary Source: Kosovo EMIS. Table A.13. Number and Percentage of Enrolled Children Age 5-6 Years by Institution Type Age of children and type of Enrolled Enrolled % of Enrolled Enrolled % of institution number of children number of children children children 2011/12 2012/13 Age 5-6 in kindergartens 1,590 7.27% 1,655 7.53% Age 5-6 in schools (pre-primary) 20,284 92.73% 20,317 92.47% Total 21,874 100.00% 21,972 100.00% Source: EMIS MEST. 232 There is no reliable central system of collecting higher education system data for public or private higher education providers, either for number of students or number teaching on non-teaching staff. 233 The numbers are for University of Prishtina and University of Prizren only, for school year 2011/2012. 149 Figure A.10. Girls as Percent Total Enrollment 49% Ratio of 48% Girls aged 0-19 in 2011 47% Primary 46% 45% Lower Secondary 44% Upper 43% Secondary 42% 2008 2009 2010 2011 2012 Source: Kosovo EMIS and Population Census 2011. Figure A.11. Pupil Teacher Ratios for Primary and Upper-Secondary Education, 2011 Basic Education Upper Secondary 35 35 30 30 25 25 20 20 15 15 10 5 10 0 Guatemala Romania Croatia Moldova Albania Serbia Ecuador Dominican R. Georgia Bulgaria Latvia Poland Hungary Montenegro Slovenia Bosnia Macedonia China Thailand Turkey Kosovo Czech Republic El Salvador Slovak Republic 5 Monte… Slovak… Domin… Maced… Czech… Russia… Guate… El… 0 Thailand Latvia Kosovo Poland Serbia Slovenia Jamaica Georgia Ukraine Hungary Ecuador Croatia Romania Bulgaria Albania China Belarus Moldova Source: WB EdStats, latest available data and EMIS Kosovo. Table A.14. Enrollments in Kindergarten and Pre-Primary Education Level of 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 education Kindergarten 5,108 5,051 4,830 5,091 5,066 5.154 5.167 5.389 Pre-primary 20,750 21,089 19,716 19,679 19,589 20,352 20,284 20,317 Source: Kosovo MEST. 150 Table A.15. Gross Enrollment Rate by Level of Education, Comparator Countries (ECA and Non- ECA), 2011 or Latest GER Lower GER Upper GER GDP per GER Pre- GER Secondary Secondary Tertiary capita (PPP) Primary Primary All All (ISCED 5 2011 Programs programs and 6) Albania 7848 57.5 105.2 104 89.2 43.9 Bosnia and Herzegovina 8115 17.4 90.5 92.1 86.4 38.1 Bulgaria 13812 79.4 102.6 83.1 93.6 56.9 Croatia 17850 61.3 93 104.7 87 54.1 Hungary 19571 85.2 101.7 100.1 100.1 60.7 Kosovo 7300 71.2234 98.1 96.4 91.8 n/a Macedonia, FYR n/a 25.5 90.1 90.2 77.8 38.6 Montenegro 11628 40.2 118.8 96 97.6 47.6 Poland 20013 71.1 98.6 97.3 96.7 72.4 Romania 12520 79 95.9 96.3 98 58.8 Serbia 10405 53.2 94.9 98 85.7 50.4 Slovenia 28436 89.7 98.4 95.3 98.6 89.6 Dominican Republic 9287 37.8 106.9 85.5 71.4 34 Ecuador 9553 150.5 120.6 97.2 77.6 39.8 Jamaica 9027 113 88.7 91.4 94.9 26 Thailand 9390 100.3 90.7 92.3 66.5 47.7 Upper middle income 10867 61.7 110.9 96 73.3 34.7 Source: World Bank EdStats, IMF WEO, International Database United Census Bureau, and Kosovo MEST (2012). 234 Enrollment rates in pre-primary education can vary widely as the official pre-primary education age is not equal across countries. In the case of Kosovo, pre-primary education is officially defined only for one grade (grade 0). 151 Figure A.12. Non-Enrollment Rates by Ethnicity Panel A: Ages 10-15 Panel B: Ages 16-20 12% 70% 10% 60% 50% 8% 40% 6% 30% 4% 20% 2% 10% 0% 0% Population Albanian Serb Other Population Albanian Serb Other Source: 2011 Household Budget Survey and WB staff calculations. Table A.16. Number of Schools by Enrollment Size (2010) No. of pupils <50 51-100 101-300 301-500 501-3000 Total > > > > > > Total 700 Total 700 Total 700 Total 700 Total 700 Total 700 m m m m m m Total 369 114 111 39 300 51 176 10 246 1 1202 215 % of schools 100% 31% 100% 35% 100% 17% 100% 6% 100% 0% 100% 18% Main schools 129 89 26 23 191 13 168 0 239 0 753 125 % of main school 100% 41% 100% 47% 100% 6% 100% 0% 100% 0% 100% 14% Satellite schools 240 25 85 16 109 38 8 10 7 1 449 90 % of satellite school 100% 9% 100% 16% 100% 26% 100% 56% 100% 13% 100% 17% 152 Table A.17. Average Pupil-Teacher Ratio (STR) By Municipality Schools Schools Above with Average Urban Rural with 700m second satellites language Decan 14.4 19.3 13.5 N/A 15.5 13.2 Dragash 20.6 17.3 21.2 21.2 17.3 23.2 Drenas 18.7 19.4 18.5 17.6 N/A 16.2 Ferizaj 19.3 19.9 18.9 18.3 N/A 16.9 Frushe Kosove 19.7 24.3 14.8 14.8 25.9 14.8 Gjakova 16.6 21.4 15.0 10.2 N/A 14.7 Gjilan 13.9 17.6 12.8 10.7 13.9 11.8 Hani i Elezit 18.6 22.0 15.2 10.6 N/A 16.2 Istog 17.3 20.0 16.8 N/A 17.3 17.5 Junik 21.1 21.1 N/A N/A N/A 17.3 Kaçanik 19.3 22.4 15.8 12.4 N/A 9.6 Kamenica 10.9 14.3 10.6 7.6 N/A 17.3 Kline 17.5 22.4 15.7 N/A N/A 22.1 Kllokot 19.2 N/A 22.0 N/A N/A 17.1 Lipjan 16.9 18.9 16.7 16.8 16.8 18.5 Malisheva 19.7 20.5 19.6 20.1 17.8 15.2 Mamusha 27.9 27.9 N/A N/A N/A 11.7 Mitrovica 16.4 19.8 13.5 8.2 21.4 15.9 Novobrdo 10.4 9.8 11.8 11.8 N/A 14.8 Obiliq 15.2 16.0 14.8 9.6 20.6 15.5 Peja 16.6 18.2 16.1 3.7 17.4 10.1 Podujeva 17.2 20.9 16.3 15.2 20.6 16.0 Prishtina 17.6 19.9 11.7 5.1 22.5 16.8 Prizren 17.2 20.5 15.4 13.8 18.0 16.9 Rahovec 18.1 23.9 17.5 10.5 N/A 19.1 Shtime 16.5 15.9 16.8 N/A N/A 16.7 Suhareka 18.3 19.5 17.3 9.1 N/A 15.3 Viti 16.9 21.0 16.2 12.9 7.7 15.7 Vushtrri 18.4 20.8 17.1 5.9 N/A 18.4 All 17.2 19.2 16.2 14.3 18.6 17.2 Table A.18. Distribution in Minorities in Kosovo Albanian Serbian Turkish Bosniak Roma Ashkali Egyptian Goran Other Population 89.9% 5.7% 1.1% 1.6% 0.5% 0.9% 0.6% 0.6% 0.1% EMIS 96.2% 0.1% 0.7% 1.2% 0.5% 0.9% 0.2% 0.1% 0.0% Source: World Bank Staff calculations based on Population Census 2011, OSCE, and EMIS. 153 Figure A.13. The Impact of Teacher Deselection on Pupil Performance Gains. Source: Hanushek (2009). Table A.19. Teacher Salary Structure Teachers KS (Euro, Monthly) Base salary based on qualifications Current base salary with 50% increase Actual % Base salary (in 2007) % Increase (in 2008) based on qualifications (in 2011) Increase based based on on qualifications qualifications Pre- Pre- Pre- Primary Secondary Primary Secondary Primary Secondary primary primary (in 2008) primary (in 2011) Unqualified / 201 216 236 201 216 236 0% 302* 324 352 0% Beginner235 5-year secondary 201 216 236 221 238 260 10% 322** 346 377 7% teacher school 2-year Higher Pedagogical 201 216 236 239 257 281 19% 340 365 398 13% School BA (3-year & 4- 201 216 236 255 274 300 27% 356 382 417 18% year) degrees Masters and Ph.D 201 216 236 271 292 319 35% 372 400 436 23% degrees Source: Ministry of Public Administration, Teachers’ Payroll Data July 2013; *Payroll data suggest that unqualified/beginner teachers’ salaries were also raised by 50 percent; ** The 50 percent increase and the percentage increase for each qualification were applied to 2007 base salary. 235 It should be noted that the number of teachers falling in the unqualified category is 14 percent of the total teaching staff, of which about 4 percent are teachers from Serbian community who do not report data on qualifications to MPA therefore are grouped with unqualified. 154 ANNEX 4: REVENUE Table A.20. Breakdown of Total Revenues in Kosovo (EUR Mil.) 2004 2005 2006 2007 2008 2009 2010 2011 2012 Total revenues 628.5 628.5 712.2 898.4 941.5 1145.8 1126.5 1278.0 1313.3 Tax revenues (net of refund) 551.0 553.6 629.2 722.6 817.2 827.4 907.0 1057.4 1081.3 Direct taxes 74.9 82.7 119.7 133.1 163.2 130.9 126.2 138.9 153.5 Income taxes 49.1 58.6 75.4 103.7 127.9 114.2 121.6 133.8 148.7 Personal income tax 19.1 36.2 44.1 50.0 62.4 59.5 69.7 79.2 86.0 Wage tax 19.1 36.2 29.1 34.0 43.6 39.0 44.2 55.7 60.1 SME tax 0.0 0.0 15.0 16.0 18.8 20.5 25.5 23.5 25.9 Corporate income tax 30.0 22.4 31.3 53.7 65.5 54.8 51.9 54.6 62.7 Other direct taxes 25.8 24.1 44.3 29.3 35.3 16.7 4.6 5.1 4.8 Profit tax 12.9 12.7 9.4 3.4 4.4 8.4 0.6 1.8 0.1 Tax on interest, dividend, rent and gambling 0.0 4.7 5.8 2.0 1.2 2.9 1.5 1.4 1.6 PTK monopoly surcharge 0.0 0.0 21.3 16.6 26.5 0.0 0.0 0.0 0.0 Other 12.9 6.7 7.8 7.3 3.2 5.4 2.5 1.9 3.1 Indirect taxes 476.1 470.9 501.0 581.2 641.3 684.9 767.4 905.2 914.2 Consumption taxes 403.2 396.1 416.2 495.3 544.7 584.6 657.6 784.8 796.4 VAT (net of refund) 228.8 239.6 255.1 304.2 341.5 378.0 427.3 506.4 514.1 VAT on import 186.0 195.2 210.8 255.0 304.8 327.7 359.9 415.0 416.5 Domestic VAT 42.8 46.4 47.8 58.8 58.4 75.5 95.0 121.4 130.0 VAT refund 0.0 -2.0 -3.5 -9.6 -21.7 -25.2 -27.6 -30.0 -32.4 Excise (net of drawback) 174.4 156.5 161.1 191.1 203.2 206.6 230.3 278.4 282.3 Excises on import 171.1 164.3 161.9 191.4 203.3 206.9 230.6 278.7 282.6 Domestic excises 3.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Excise refund 0.0 -7.8 -0.8 -0.3 -0.1 -0.3 -0.3 -0.3 -0.3 Tariff and duties 72.9 74.8 84.8 85.9 96.6 100.3 109.8 120.4 117.8 Customs tariff 69.9 71.9 77.7 79.8 94.8 99.5 109.0 119.6 117.0 Other duties 3.0 2.9 7.1 6.1 1.8 0.7 0.8 0.8 0.8 Property tax 0.0 0.0 8.5 8.3 12.7 11.6 13.4 13.3 13.6 Non-tax revenues 77.5 74.9 83.0 175.8 124.3 318.5 219.5 220.6 232.0 GDP per capita 2912.5 3002.9 3120.4 3393.7 3851.4 3912.4 4192.0 4776.1 4946.4 Source: Ministry of Finance, Tax Administration of Kosovo, Statistical Agency of Kosovo, Kosovo Customs and KPST data. Note: Total revenues do not include grants. 155 ANNEX 5: MACROECONOMIC INDICATORS Table A.21. Key Macroeconomic Indicators, 2008-2015 2008 2009 2010 2011 2012 2013 2014 2015 Estimate Projections Real growth rates GDP real growth 6.9 2.9 3.9 4.5 2.7 3.0 3.5 3.5 CPI, annual average 9.4 -2.4 3.5 7.3 2.5 1.9 1.0 1.0 Share of GDP Revenue (including budget grants) 23.9 28.3 26.7 27.2 26.7 26.5 26.8 26.3 Expenditure (and net lending) 24.1 29.0 29.3 29.1 29.2 29.5 28.8 28.3 Primary deficit -0.2 -0.7 -2.4 -1.7 -2.3 -2.6 -1.7 -1.7 Overall deficit -0.2 -0.7 -2.6 -1.9 -2.6 -3.0 -2.0 -2.0 Gov. bank balances (usable) 10.8 8.7 5.8 3.5 4.3 3.2 2.8 2.6 Public Debt 0.0 6.1 6.4 6.0 8.7 9.7 11.4 13.0 Investments 28.6 32.3 33.9 33.2 32.6 33.2 32.3 32.1 FDI 8.7 6.8 7.6 7.9 4.3 6.0 6.5 6.5 Current Account Balance -11.7 -9.2 -11.9 -13.8 -7.6 -10.9 -9.5 -8.2 GDP (€, Mil) 3,940 4,046 4,334 4,770 5,017 5,213 5,442 5,644 GDP per capita (€) 2,255 2,297 2,441 2,665 2,779 2,858 2,955 3,034 Population (Mil.) 1.75 1.76 1.78 1.79 1.81 1.82 1.84 1.86 Source: MoF, Statistics Agency of Kosovo and WB staff calculations. 156 REFERENCES BOOST database for Kosovo. Boughzala, M. (2013), “Youth Employment and Economic Transition in Tunisia”, Brookings Global Economy and Development Working Papers 57. Available from: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.306.4768&rep=rep1&type=pdf Central Bank of the Republic of Kosovo, Time series. http://www.bqk-kos.org/?cid=2,124 Deloitte (2013), Tax Highlights Series. Available from: http://goo.gl/0VWaUO Global Integrity (2011), Global Integrity Report 2011, Washington, D.C. Hanushek, E., Woessmann, L. & Zhang, L. (2011). General Education, Vocational Education, and Labor- Market Outcomes over the Life-Cycle. NBER Working Paper 17504. Hanushek, E. & Woessmann L. Does Educational Tracking affect performance and inequality? Differences-in-differences evidence across countries. NBER Working Paper Nº 11124. February 2005. Hanushek, E. (2009). Teacher Deselection. Creating a New Teaching Profession. Washington, DC. Urban Institute Press, 2009, pp. 165-180. Hanushek, E. (2010). The Economic Value of Higher Teacher Quality. Economics of Education Review 30 466-479. Hanushek, E. and Woessmann, L. (2009). The High Cost of Low Educational Performance: The Long- Run Impact of Improving PISA Outcomes. OECD, Paris. Heckman, James J. (2008). "Schools, Skills and Synapses," Economic Inquiry, 46(3): 289-324. Health Law 2004. http://www.kuvendikosoves.org/common/docs/ligjet/2004_4_en.pdf Jakubowski, M, Patrinos, H.A. Porta E.E, Wisniewski, J. The impact of the 1999 Education Reform in Poland. OECD Education Working Paper Nº49 (2011). Kammen, D., M.Mozafari, and D.Prull (2012), “Sustainable energy options for Kosovo: An analysis of resource availability and cost”. http://rael.berkeley.edu/kosovoenergy Khwaja, Munawer S. and Indira Iyer (2013) “Revenue Potential, Tax Space and Tax Gap in ECA Countries: A Comparative Analysis, World Bank, Working Paper (draft). Koettl, Johannes & Weber, Michael, 2012. "Does Formal Work Pay? The Role of Labor Taxation and Social Benefit Design in the New EU Member States,"IZA Discussion Papers 6313, Institute for the Study of Labor (IZA). Kosovo Agency of Statistics. (2013), Kosovo Population Census 2011, http://esk.rks- gov.net/ENG/pop/tables Kosovo Agency of Statistics. Labor Force Survey 2012, http://esk.rks- gov.net/ENG/publikimet/cat_view/16-labour-market Kosovo Agency of Statistics (2013), GDP by expenditure approach 2004-2012. http://esk.rks- gov.net/ENG/publikimet/cat_view/13-national-account 157 Kosovo Agency of Statistics, General Statistics, Quarterly Bulletin, Average Monthly Paid Net Wages in the Budget sector by year, April 2013. http://esk.rks-gov.net/ENG/publikimet/cat_view/36-general- statistics Kosovo Agency of Statistics, Household Budget Survey data. http://esk.rks- gov.net/ENG/publikimet/cat_view/17-living-conditions- Kosovo Customs (2014), Monthly revenue report. Krueger, A. B. and Lindahl, M. (2001). Education for growth: Why and for whom? Journal of Economic Literature 39(4), 1101–1136. Ministry of Education, Science and Technology, Republic of Kosovo (2012). Kosovo Education Indicators. Ministry of Education, Science and Technology, Republic of Kosovo, Law on Pre-University Education. Ministry of Finance, Republic of Kosovo, Treasury Department, Annual Financial Reports 2008 - 2012. http://mf.rks-gov.net/en- us/ministriaefinancave/raportedhepublikime/raportimimujoritehyravedheshpenzimeve.aspx Ministry of Finance, Republic of Kosovo, Budget Department, Budget 2014. http://mf.rks-gov.net/en- us/ministriaefinancave/buxhetiirepublikessekosoves/buxhetiqendrore.aspx National Renewable Energy Action Plan http://www.energy-community.org/pls/portal/docs/2570177.pdf Norregard, J. (2013), “Taxing immovable property: Revenue potential and implementation challenges”, IMF Working Paper, WP/13/129. http://www.imf.org/external/pubs/ft/wp/2013/wp13129.pdf OECD (2010), Tax Policy Reform and Economic Growth, OECD Tax Policy Studies, No. 20, http://www.oecd.org/ctp/tax-policy/taxpolicystudyno20-taxpolicyreformandeconomicgrowth.htm OECD, Education at a Glance 2012, Table B6.2 http://www.oecdbookshop.org/display.asp?K=5K9CV8N71Z9W&LANG=EN OECD Publishing. http://dx.doi.org/10.1787/9789264091085-en Organization for Security and Co-operation in Europe: Mission to Kosovo (2011). Kosovo Community Profiles. Rutkowski, J. (2009). “Ukraine Labor Demand Study”. World Bank. Sala-i-Martin, X., Doppelhofer, G., and Miller, R. I. (2004). Determinants of long-term growth: A Bayesian averaging of classical estimates (BACE) approach. American Economic Review 94(4), 813–835. UNICEF (2011). Joined Hands: Better Childhood in Kosovo. http://www.unicef.org/kosovoprogramme/Joined_Hands_ENG_web.pdf USAID 2012. The Effect of Unreliable Power Supply and Quality on Kosovar Businesses. World Bank (2010). Kosovo Public Expenditure Review. http://www- wds.worldbank.org/external/default/WDSContentServer/WDSP/IB/2010/07/05/000333038_20100705005 452/Rendered/PDF/537090ESW0P117101Official0Use0Only1.pdf 158 World Bank Business Environment and Enterprise Performance Survey (BEEPS) for Kosovo, 2009. http://www.enterprisesurveys.org/ World Bank (2013). Social Spending: From Boom to Bust in the EU11. World Bank (2011) Background Paper: Development and Evaluation of Power Supply Options for Kosovo http://siteresources.worldbank.org/INTENERGY2/Resources/Kosovo_generation_options_report_12312011.pdf World Bank (2013), “Consumption Poverty in the Republic of Kosovo in 2011”. http://esk.rks- gov.net/ENG/publikimet/cat_view/17-living-conditions- World Bank (2013), “Balancing Act: Cutting Energy Subsidies while Protecting Affordability” http://documents.worldbank.org/curated/en/2013/06/17910085/balancing-act-cutting-energy-subsidies-protecting- affordability World Bank (2013), “Mitigating the impact of an aging population: Options for Bulgaria”. http://documents.worldbank.org/curated/en/2013/09/18262916/mitigating-economic-impact-aging-population- options-bulgaria World Bank (2007) “From Red to Gray: The Third Transition of Aging Population in Eastern Europe and the Former Soviet Union”. http://web.worldbank.org/WBSITE/EXTERNAL/COUNTRIES/ECAEXT/0,,contentMDK:21378474~pagePK:1467 36~piPK:146830~theSitePK:258599,00.html World Bank (2012), Country Environmental Analysis: Kosovo http://documents.worldbank.org/curated/en/2013/01/17485553/kosovo-country-environmental-analysis-kosovo- country-environmental-analysis-cea World Bank (2011), “Development and evaluation of power supply options for Kosovo”. Report available from: http://siteresources.worldbank.org/INTENERGY2/Resources/Kosovo_generation_options_report_12312011.pdf World Bank EdStats. http://web.worldbank.org/WBSITE/EXTERNAL/TOPICS/EXTEDUCATION/EXTDATASTATISTICS/EXTEDST ATS/0,,menuPK:3232818~pagePK:64168427~piPK:64168435~theSitePK:3232764,00.html World Bank, Institutional Development Education Project http://www.worldbank.org/projects/P102174/institutional-development-education-project?lang=en&tab=overview World Bank, Kosovo Health System Network Master Plan project, Data report, 2009. World Bank (2011), “Railway Reform in South East Europe and Turkey On the Right Track?” Report No. 60223-ECA. World Bank (2013), “Activation and smart safety nets in Kosovo: constraints in beneficiary profile, benefit design, and institutional capacity. World Bank. Europe and Central Asia Social Protection expenditures and evaluation Database. 159