Turkey - Public expenditure review

Report No. 36764-TR Turkey Public Expenditure Review December 21, 2006 Poverty Reduction and Economic Management Unit Europe and Central Asia Region Document of the World Bank  ACKNOWLEDGEMENTS This report was prepared by a core team led by Aristomene Varoudakis and composed of: Mediha Agar, Mukesh Chawla, Robin Horn, Mark Lundell, Zafer Mustafaoglu, Kamer Ozdemir, Gary Reid; Anita Schwarz, and Sanjay Vani. Contributions were provided by: Cecilia Briceno-Garmendia; Sarbani Chakraborty, Ferhat Emil (consultant); Graham Glenday (consultant); Ranjit Lamech; Brian Levy; Seema Manghee, Koshy Mathai (IMF), Devesh Chandra Mishra; Seda Aroymak; Amithaba Mukherjee; Ebru Ocek (consultant); Panagiota Panopoulou, Abuzer Pinar (consultant); Ferda Sahmali, Sameer Shukla; Emilia Skrok, and Zhicheng Li Swift. Research assistance was provided by Olga Vybornaia, and Izzet Yildiz. Peer reviewers were: William Dorotinsky and Anand Rajaram. Guidance was provided by Cheryl Gray (Sector Director), Pradeep Mitra (Chief Economist, ECA Region), and Felipe Jaramillo (Sector manager). Comments were provided by Richard Allen (IMF), Rodrigo Chaves, Marianne Fay, Ali Mansoor, Maureen McLaughlin, Samuel Otoo and Andrew Vorkink. Administrative support: Lalani Dammika Somasundaram and Pinar Baydar Valuable support from the counterpart team, organized by the Undersecretariat o f the Treasury, i s gratefblly acknowledged, and especially from the services o f State Planning Organization and the Ministry o f Finance, which provided information on many parts o f the study.  Table o f Contents EXECUTIVE SUMMARY ........................................................................ VI OVERVIEW ................................................................................................ IX CHAPTER I SOURCES OF FISCAL CONSOLIDATION AND THE QUALITY OF FISCAL ADJUSTMENT A. ................................................. PATTERNS OF FISCAL ADJUSTMENT IN TURKEY ................................................................. 30 30 B. THE SIZE OF GOVERNMENT AND COMPOSITIONOF EXPENDITURES IN INTERNATIONAL COMPARISON................................................................................................ 36 B.l Public expenditure by economic category ......................................................................................... 37 B.2 Public Expenditure by Government finction ..................................................................................... 40 C. THE REVENUE SIDE ...................................................................................................................... 43 D. STRUCTURAL AND CYCLICAL COMPONENTS OF THE BUDGET BALANCE ................... 50 CHAPTER I1 OPTIONS FOR EFFICIENCY GAINS IN PUBLIC EXPENDITURES-HORIZONTAL I SSUES A. ......................................... PUBLIC SECTOR EMPLOYMENT AND THE WAGE BILL ....................................................... 60 60 B. OBJECTIVES AND REFORM OPTIONS INTHE MODERNIZATION OF CIVIL SERVICE .... 67 C. ISSUES IN MANAGING AND FINANCING THE PUBLIC INVESTMENT PROGRAM ..........73 D. ESTABLISHING TAX EXPENDITURE ACCOUNTS-TOWARDS BETTER FISCAL ACCOUNTABILITY AND TRANSPARENCY .............................................................................. 85 CHAPTER I11 OPTIONS FOR EFFICIENCY GAINS IN PUBLICEXPENDITURES-SECTOR ISSUES A. ..................................... PUBLIC EXPENDITURE IN THE EDUCATION SECTOR........................................................... 94 94 A1. Educational Outcomes ....................................................................................................................... 95 A2 . Structure o f Education Expenditures in Turkey ................................................................................. 98 A3 . Key reform directions ...................................................................................................................... 115 B. THE PENSION SYSTEM AND ITS REFORM ............................................................................. 117 B1. Evolutiono f the Pension System ..................................................................................................... 117 B2. Comparison of 2006 Pension Law with 1999 Law .......................................................................... 119 B3. Impact of 2006 Law and remaining challenges ............................................................................... 122 C. PUBLIC EXPENDITURE IN THE HEALTH SECTOR................................................................ 127 c 1. Health outcomes .............................................................................................................................. 127 c2. Main characteristics of the health sector .......................................................................................... 130 c3 * Public expenditures on health .......................................................................................................... 135 c4. Short-and medium-ternmeasures to improve efficiency and enhance equity o f public spending .. 141 D. RURAL PUBLIC EXPENDITURES .............................................................................................. 146 D1. Rural Investments ............................................................................................................................ 146 D2. Non Investment Spending for Rural Areas ...................................................................................... 148 E. SUMMARY OF EXPENDITURE REFORM OPTIONS IN KEY SECTORS............................... 152 CHAPTER I V MANAGING MEDIUM TERM EXPENDITURE PRESSURES A. .............................................................................................. THE GOVERNMET’S MEDIUM-TERM FISCAL STRATEGY ................................................. 154 153 B. EXPENDITURE PRESSURES IN KEY SECTORS AND MEDIUM-TERM CHALLENGES.... 161 B 1. New and emerging education sector policies in Turkey ................................................................. 162 B2. The Medium Term Education Sector Fiscal Framework for 2007-2009 ......................................... 163 B3. Projecting social protection expenditures ........................................................................................ 167 B4. Projections for Rural Expenditures.................................................................................................. 170 C. REQUIRED FISCAL SPACE TO MEET MEDIUM-TERM EXPENDITUREPRESSURES ...... 171 C.1. A sustainable medium-term expenditure envelope .......................................................................... 172 C.2. An assessment o f required fiscal space ............................................................................................ 174 CHAPTER V STRENGTHENING BUDGETARY INSTITUTIONS FOR EFFECTIVE GOVERNMENT A. ...................................................... STATUS OF PUBLIC EXPENDITUREMANAGEMENT REFORMS........................................ 182 180 B. BUDGET EXECUTION. ACCOUNTING. AND REPORTING ................................................... 191 C. INTERNAL CONTROL AND INTERNAL AUDIT ...................................................................... 196 D. EXTERNAL AUDIT ....................................................................................................................... 199 E. LEGISLATIVE OVERSIGHT ........................................................................................................ 203 F. PUBLIC PROCUREMENT MANAGEMENT ............................................................................... 204 G. COMMON REFORM DIRECTIONS............................................................................................. 209 Annexes Annex I Data and Methodology for Consolidated General Government .................................... 212 Classification o f Expenditures Annex I1 Classification o f Central Government Expenditures across ........................................... 216 Functional and Economic Categories. 2004-2008 Annex 1 11 Estimating Structural and Cyclical Budget Balances..................................................... 222 Annex I V Tax performance: A RegressionModel ......................................................................... 236 Annex V Draft Tax Benchmarks for Tax Expenditures ............................................................... 240 Annex V I Towards Estimates of Tax Expenditures in PIT and CIT .............................................. 243 Annex VI1 Classification of Budget Institutions .............................................................................. 251 Annex VI11 Baseline Indicator System (BIS) for Public Procurement.............................................. 253 .. 11 Tables Table 1.1 Institutional Breakdown o f Consolidated General Government Primary Balance 33 Table 1.2 Fiscal Consolidation in Turkey, 1999-2004 and sub-periods 35 Table 1.3 Revenue Breakdown for Selected I E C D Countries 46 Table 1.4 Outut Elasticity o f Taxes 51 Table 1.5 Fiscal Stance and the Output Gap 55 Table 1.6 Output Elasticity o f Social Security Contributions 56 Table 1.7 Fiscal Stance and the Optput Gap 58 Table 1.8 General Government Cyclical Budget 59 Table 2.1 Personnel Expenditures by government function and level o f General Government 61 Table 2.2 Total employment in Turkey’s public sector 63 Table 2.3 Infrastructure Investment Portfolio 74 Table 2.4 Institutional Breakdown o f Public Investments 75 Table 2.5 Public Sector Investment, 1999-2006 76 Table 2.6 Annual Investment Spending (% o f GNP) 76 Table 2.7 Effect o f Economic returns f r o m Delayed Project Completion 78 Table 2.8 Selected tax expenditures for Turkey in 2003 91 Table 3.1 Overall Public Education Expenditures by Institution 99 Table 3.2 Overall Public Education Expenditures by Economic Classification 100 Table 3.3 Puublic education Expenditures in Turkey by Institution 101 Table 3.4 Expenditure o n educational institutions as a percentage o f GDP f r o m Public and private sources by source o f fund 110 Table 3.5 Per Student Education Expenditure by Education Level 111 Table 3.6 Comparison o f the pension system parameters before and after the 2006 reform 121 Table 3.7 Health indicators for selected countries 128 Table 3.8 Health-related MDG indicators for Turkey f r o m various sources 128 Table 3.9 In128nt Mortality and Under-five Mortality Rates by area o f residence and region 128 Table 3.10 Co-payments for health care services and pharmaceuticals 131 Table 3.1 1 Number o f insured individuals by type o f health insurance coverage 2002 132 Table 3.12 N o o f hospitals and hospital beds by type o f institution 133 Table 3.13 Number o f Hospital Beds and Utilization 134 Table 3.14 Health care service utilization 135 Table 3.15 MOH Hospital revolving Funds 137 Table 3.16 Public Expenditure o n health 137 Table 3.17 Preventive activities 139 Table 3.18 Total Public Investment and Rural Investment 147 Table 3.19 Investment by Rural Agencies 148 Table 3.20 Total Agricultural Transfer Budget 149 Table 3.21 Agricultural Support Instruments’ Shares o f Transfer Budget 150 Table 3.22 Consolidated Budget Expenditures Including Recurrent Costs 151 Table 4.1 Institutional breakdown o f the primary surplus 154 Table 4.2 Central Government Budget Revenues in the M T F S 157 Table 4.3 Adjusted Primary expenditures o f the central government 158 Table 4.4 General Government Revenues and Expenditures 160 Table 4.5 General Government Primary Surplus 161 Table 4.6 Breakdown o f Additional Expenditures Needed for Financing N e w Policies 166 Table 4.7 Total Education Expenditures by Institutions 167 Table 4.8 Social Security System Deficit 167 Table 4.9 Projected Public Expenditures o n Health 169 Table 4.10 Estimated Revenue and Expenditure Savings f r o m Co-payments 170 Table 4.1 1 Rural Expenditure Allocations and Projections 171 Table 4.12 General Government Revenue Projections 172 Table 4.13 Public Debt Forecasts 173 Table 4.14 Total Spending Envelope for Primary Expenditures 174 ... 111 Table 4.15 Expenditure Pressures in Key Areas 175 Table 4.16 Medium t e r m expenditure Pressures, Baseline Scenario 176 Table 4.17 Primary Surplus Needed for Achieving Pre-turmoil Debt Rations 177 Table 4.18 Required fiscal space to meet medium-term expenditure pressures 177 Table 5.1 Summary o f Recent Laws to Strengthen Public Financial Management 181 Table 5.2 Importance o f the Revolving Funds 190 Table 5.3 Extra Budgetary Funds 191 Table 5.4 Action Plan as Provided in CPAR 2001 206 Figures Figure 1.1 Fiscal Balances and primary expenditures and revenues o f the General Government 31 Figure 1.2 General Government Expenditure as % o f GDP 37 Figure 1.3 Compensation o f General Government employees 38 Figure 1.4 General Government Wage Bill in percent o f Primary Public Expenditure 38 Figure 1.5 Expenditure by Economic Category 39 Figure 1.6 General Government Expenditure on Social Benefits and Social Transfers 39 Figure 1.7 Consolidated General Government Gross Fixed capital Formation 40 Figure 1.8 Functional Allocation o f General Government Expenditures 41 Figure 1.9 General Government Revenue in % o f GDP vs. per capita GDP in PPP 43 Figure 1.10 Tax effort indicators 44 Figure 1.11 Tax effort ratios in Turkey 45 Figure 1.12 Actual and Structural Primary Budget Balances for Consolidated Budget 2 1 52 Figure 1.13 Actual and Structural Budget Balances for Consolidated Budget 53 Figure 1.14 Cyclical Budget Balance for Consolidated Budget 54 Figure 1.15 Actual and Structural Primary Budget Balances for General Government 57 Figure 1.16 Actual and Structural Budget Balances for General Government 57 Figure 2.1 Personnel expenditures in percent o f GDP 61 Figure 2.2 Composition o f Wage Bill by Function 62 Figure 2.3 Employment i n the Public Sector 62 Figure 2.4 Employment and average wage cost in the public sector 64 Figure 2.5 Wage bill in consolidated budget agencies actual and alternative scenario 65 Figure 2.6 Compensations, Rewards, Allowances and overtime pay in percent o f total personnel expenditures o f consolidated budget agencies 66 Figure 2.7 Gross Capital Formation by Function 74 Figure 2.8 Maintenance o f Expenditures (% o f NP) 80 Figure 3.1 Percent o f 20-24 year olds with a secondary education diploma 96 Figure 3.2 Distribution o f Students by PISA Proficiency Level in Turkey and the EU, 2003 97 Figure 3.3 Education expenditures in Turkey in 2002 by Source 104 Figure 3.4 Total Education Spending in Turkey in 2002 by Provider 105 Figure 3.5 Financing o f Education, 2002 106 Figure 3.6 Out o f Pocket Spending in Turkey in 2002 by Providers 106 Figure 3.7 Total Education Spending per Student 107 Figure 3.8 Education Spending per Student at Public and Private Schools 108 Figure 3.9 Achievement Across School Types 108 Figure 3.10 Distribution o f Spending on educational Inputs in Turkey in 2002 109 Figure 3.1 1 Expenditures on Pre-Primary Education 112 Figure 3.12 Expenditures on Primary Education 112 Figure 3.13 Expenditures on Secondary Education 113 Figure 3.14 Expenditures on Tertiary Education 113 Figure 3.15 PISA Results by Secondary Expenditure per Student 114 Figure 3.16 Percentage o f the Population over Age 65 in Turkey and Selected OECD Countries 118 Figure 3.17 Pension System Deficits Before and After Passage o f the 2006 L a w 122 Figure 3.18 Contribution Rates in Selected OECD Countries 123 Figure 3.19 Retirement ages in OECD Countries 124 Figure 3.20 Pension Accrual rates in OECD Countries 124 iv Figure 3.2 1 Average Benefits Relative to Average Wage in SSK, post reform 126 Figure 3.22 Inpatient admissions for selected countries 135 Figure 3.23 Outpatient contracts for countries 135 Figure 3.24 Functional classification o f recurrent public expenditures on health 138 Figure 4.1 Unadjusted and Unadjusted SSIs deficit 159 Figure 4.2 (a-d) Total Education E x p e n d i k e Framework Without and With N e w Education Policies 166 Figure 5.1 Schematic Diagram o f the Internal Financial Control Framework for the general Budget Institutions as Defined by the 2003 PFMC Law 196 Boxes B o x 1.1 Conversion between SPO definition o f CGG primary surplus to IMF’s program definition o f public sector primary surplus 32 B o x 2.1 Guidelines to Identify Normative Tax Structure 86 B o x 2.2 Tax expenditure by type o f tax 87 B o x 3.1 Principal sources o f financing for the education sector 101 Box 3.2 Expenditure Procedures at the Provincial Level 102 B o x 3.3 The system o f financial protection for health expenditures 130 B o x 3.4 The role o f the Revolving Funds in the Turkish Health Care System 136 Box 3.5 Examples f i o m Rural Agencies, Issue o f H i g h Personnel Expenditures 151 B o x 4.1 2007 Budget and Medium-TermFiscal Strategy 155 B o x 4.2 The Methodological changes in the Reporting o f Health Expenditures and Social Security Transfers 159 B o x 4.3 Data sources for the medium-termeducation expenditure framework 164 B o x 4.4 Projecting the budgetary impact o f new education policies and proposals 164 B o x 5.1 Towards Full Implementation o f PFMC 185 B o x 5.2 Amendments to the PFMC Law 187 V  EXECUTIVE SUMMARY 1. Turkey’s ambitious fiscal adjustment has facilitated a substantial decline in the public debt ratio and underpinned the strong economic performance since 2001 but existing vulnerabilities underscore the need for continuous fiscal discipline. Resolute fiscal consolidation-primarily based o n the revenue side so far-has been the cornerstone o f the economic program that led to Turkey’s strong economic performance, with annual real GDP growth averaging 7.4 percent during 2002-2005; inflation reduced to single digits in 2004 and 2005; and a significant reduction in the public debt ratio. However, despite improved resilience to shocks, vulnerabilities remain, as confirmed by the strong impact o f recent global market volatility on Turkey. In the current environment, maintaining continuous fiscal discipline, with a high primary fiscal surplus, wiIl be necessary in order to stem pressures o n the current account, help keep disinflation o n track, and achieve a reduction o f the public debt ratio so as to further improve resilience to external shocks. 2. Turkey faces two simultaneous fiscal challenges-maintaining fiscal discipline, while creating the fiscal space needed to meet pressing development challenges and sustain a fast pace o f medium-term growth. A s strong fiscal discipline needs to be maintained in the years ahead, there i s little room for Turkey to further increase total expenditure in order to meet pressing development challenges. Structural fiscal reforms, aimed at improving the quality o f fiscal consolidation, are the only viable means o f sustaining the adjustment, while making appropriate fiscal space for growth-enhancing expenditures and lower taxes in the future. 3. Fiscal space to meet expenditure pressures should be created by a combination of structural public expenditure reforms, expenditure reallocations, and continuing tax reform efforts. 0 Sector-specific reforms to improve the efficiency o f expenditure programs would be required in areas where expenditure pressures are being felt; 0 Trade-offs in expenditure allocations would need to be considered, with the aim o f shifting resources to priority programs for growth and social development; 0 Sector-specific expenditure reforms should be underpinned by horizontal reforms that improve the efficiency o f expenditure programs across sectors; 0 Initiatives to k r t h e r rationalize the tax system-including the tax expenditures- would help create fiscal space by broadening the tax bases. 4. Significant fiscal space to meet expenditure pressures will be required during 2007-2009 under a baseline and alternative scenarios. In a baseline scenario taking into vi account the expenditure pressures in key sectors such as health and education, the fiscal space needed to meet these expenditures i s estimated at 1.1 percent o f GDP in 2007, reaching 1.7 percent in 2008 and 2 percent in 2009. With a stronger effort in education, and assuming a larger primary fiscal surplus-as needed in order to reduce the public debt ratio to the levels projected before the recent market turmoil-the needed fiscal space would increase to 2.4 percent o f GDP in 2007, up to 3.3 percent by 2009. However, the fiscal space required to reduce the debt ratio would be lower if expected privatization revenues were used to repay debt. 5. Improving education outcomes will require additional financial resources, together with increased institutional efficiency. Existing educational gaps suggest the need for additional fiscal space for educational expenditures with the aim of: increasing the share o f non personnel expenditures to levels similar to private schools and universities; improving the regional balance o f educational services across the country; increasing per student expenditure on pre primary education. The tight fiscal framework in the years ahead makes it necessary to improve the efficiency o f public spending on education-by encouraging schools to fill their classrooms up to acceptable capacity levels; providing financial resource autonomy to schools while introducing accountability for results; granting higher education institutions greater autonomy over financial resources, with incentives for efficient management. 6. The combination o f poor health outcomes and relatively high expenditures suggests that important room exists for improving the efficiency o f the health care system. The introduction o f Universal Health Insurance (UHI) will improve access and equity but, to remain fiscally viable, i t should go in tandem with cost savings and marked improvements in efficiency o f public spending. The principal measure to control costs and dissuade indiscriminate use o f health care services i s the introduction of copuyments for outpatient visits and for drugs. But cost containment measures should go in tandem with more comprehensive structural reform-such as, full expansion o f family medicine; reforming the patient referral system; introduction o f payment mechanisms for providers which will discourage unnecessary use o f physician and diagnostic services; consolidation o f hospitals and greater management autonomy. 7. The 2006 reform will bring fiscal balance to the pension system in the long term, as well as the unification o f a previously fragmented system, but future challenges should be kept in sight. With adequate implementation o f the reform, the pension system deficit is expected to be absorbed by 2040, with a fiscal gain o f 7 percent o f GDP compared to a “no reform scenario”. However, long-run financial sustainability will be secured under a high level o f payroll taxes by international standards-a key impediment to j o b creation in the formal sector. The results o f the 2006 reform should be closely monitored, and if lowering contribution rates i s a desired objective, more changes may be required in the future to generate sufficient surpluses in the pension system. 8. Sector-specific expenditure reforms will provide a better pay-off when combined with “horizontal ” reforms that improve the efficiency o f expenditure programs across sectors. Horizontal reforms should include initiatives aimed at strengthening incentives vii in the c i v i l service towards the achievement o f results, while containing pressures o n the wage bill. Adoption o f efficient project evaluation methods, effectiye multi-year programming, and adequate provisioning for long-term operation and maintenance o f public capital will be key to improving the quality o f the investment program across functional expenditure areas. A framework conducive to a greater role o f the private sector in the financing, development, and operation o f infrastructure should be created- requiring a predictable policy and regulatory environment, together with careful design o f Government commitments to private operators to minimize the risk o f contingent liabilities. 9. On the revenue side, reform efforts should focus on tax rationalization and base broadening. This will improve tax efficiency while laying the groundwork for lowering tax rates in the future. Establishing an adequate tax expenditure framework would promote fiscal accountability and transparency and help rationalize the tax system. Preliminary estimations suggest that, at around 5 percent o f GDP in 2003, tax expenditures in Turkey were close to the high end o f the OECD country range. A sustained effort is required over the medium term to rationalize tax expenditures, and make their selection and associated trade offs with direct spending programs an integral part o f the budgeting process. 10. Despite progress, addressing the unfinished agenda in public financial management reform represents a major challenge and requires strong coordination and monitoring. Far-reaching public financial management reforms-further to the enactment o f the Public Financial Management and Control (PFMC) Law in 2003 and several other laws-have improved budget coverage, formulation, execution, accounting, audit, and procurement, providing a new legal framework for modem public expenditure management and accountability. The main challenge ahead i s the implementation o f the reform agenda throughout the entire general government, including extra-budgetary funds and revolving funds. The key areas that require immediate attention are: (i) establishment o f the internal audit structures; (ii) enactment o f the Turkish Court o f Account legislation consistent with the PFMC law; and, (iii) bringing the remaining Extra Budgetary Funds and Revolving Funds under the budgetary and financial control structures. Effective leadership and coordination among agencies would be required for the effective implementation o f the overall Public Financial Management reform agenda. viii  OVERVIEW* 1. Turkey’s ambitious fiscal adjustment has paved the way for the strong economic performance since the 2000/2001 crisis. Following the 2001 crisis, the Turkish economy rebounded very strongly, with annual real GDP growth averaging 7.4 percent during 2002-2005. Inflation came down to single digits in 2004 and 2005 for the first time in the last 35 years. Resolute fiscal consolidation-primarily based on the revenue side so far-has been the cornerstone o f the economic program, with the public sector primary balance moving to a surplus o f 6.9 percent o f GDP in 2005, from a deficit o f 1.6 percent in 1999. Gross public debt was thus reduced from 106 percent o f GDP in 2001 to 71 percent in 2005. Resolute fiscal consolidation relieved the conduct o f monetary policy from fiscal dominance, thus enhancing the credibility o f the commitment to l o w inflation and facilitating the implementation o f inflation targeting. 2. Despite progress, the need for continuous fiscal discipline has been recently underscored by existing vulnerabilities. Vulnerabilities have also emerged because, despite fiscal consolidation, strong domestic demand growth, exchange rate appreciation during 2002-2005, and an imported energy bill inflated by record-high o i l prices, have all resulted in a widening deficit o f the external current account that reached 6.4 percent o f GDP in 2005. As a result o f recent global financial market volatility, the lira has depreciated by 20 percent and domestic government bond yields have significantly increased. The disinflation, whose pace had already slowed in the late months o f 2005, will likely be further prejudiced in the short term by the pass-through o f depreciation o f the currency. Owing to higher interest rates and currency depreciation, the debt ratio will also be reduced at a slower pace. In the current environment, maintaining fiscal discipline and a high primary fiscal surplus will be necessary in order to stem pressures on the current account, while an increase in private savings will also work in the same direction. Continuous fiscal discipline will help keep disinflation on track, sustain market confidence, and achieve a reduction o f the public debt ratio so as to improve resilience to external shocks. 3. Structural public expenditure reforms-reviewed in this study-would be the main option to create the fiscal space needed tu meet Turkey’s pressing development challenges and sustain a fast pace of medium-term growth. As strong fiscal discipline will have to be maintained in the years ahead, there i s little room for Turkey to further increase total expenditure in order to meet pressing development challenges. Structural fiscal reform, aimed at improving the quality o f fiscal consolidation, i s the only viable means o f sustaining the adjustment in the future while making appropriate fiscal space for growth-enhancing expenditures and lower taxes. The focus o f efforts should now shift to the expenditure side in critical sectors, with the aim o f achieving efficiency gains and sustained cost savings in areas where expenditure pressures are being felt. At the same time, horizontal trade-offs in expenditure allocations would need to be considered, with the aim o f shifting resources to priority programs for growth and social development. * The Analysis in this study draws on information and policy initiatives until mid-November 2006 ix Expenditure reforms should be combined with initiatives to further rationalize the tax system and broaden the tax bases so as to create room for a sustained reduction in Turkey’s high tax burden in the future. The quality offiscal adjustment 4. At the General Government level, Turkey’s ambitious fiscal consolidation has relied mainly on revenue-increasing measures rather than expenditure rationalization. Over 1999-2005 the primary fiscal surplus o f the Consolidated General Government increased by 5.7 percentage points o f GDP. This improvement was largely driven by primary fiscal revenues, which increased by 6.1 percentage points o f GDP, while primary expenditures slightly increased by 0.4 percentage points. However, although more steps are clearly needed, initiatives have been taken to prevent growth o f low-productivity expenditures and improve priority setting, which have supported a better control o f overall expenditure growth than in the past. 5. Outside the General Government, expenditure downsizing came mainly through adjustment in State-Owned Enterprises (SOEs). The SOEs covered under the IMF program were generating a deficit o f 2.1 percent o f GDP in 1999, which was turned into a surplus o f 0.15 percent o f GDP in 2005. These SOEs thus added the equivalent o f almost 3 percentage points o f GDP to fiscal consolidation o f the public sector as a whole. This improvement was generated by a close to 2 percentage points o f GDP reduction in their investment expenditures, mainly for infrastructure, and a cut in personnel compensation equivalent to 1 percent o f GDP-as a result o f a 45 percent downsizing o f employment since 1999 (including the privatized SOEs). 6. Fiscal adjustment has been impeded by a growing social security deficit. Cuts in investment expenditures have merely offset a growing deficit o f the social security institutions, which-despite Turkey’s favorable demographics-reached 4.8 percent o f GDP in 2005, from 1.9 percent o f GDP in 2000. In the absence o f the social security reform enacted in June 2006, the projected deficit o f the system would have further increased to reach 7 percent o f GDP in the long term. 7. Most o f the fiscal adjustment since 2002 came from cyclical revenue improvements, reflecting robust growth after the 2001 crisis. From 2002 to 2005 the primary fiscal surplus o f the consolidated general government (excluding SOEs) increased by 2 percentage points o f GDP. Over this period, buoyant fiscal revenues, due to robust growth, contributed an estimated 3.4 percentage points o f GDP to fiscal consolidation-more than the improvement in the primary fiscal balance. Thus, since 2002 the structural primary fiscal surplus-a better measure o f policy-induced changes in the fiscal stance-has actually declined (by an estimated 1.2 percent o f potential GDP). X 8. The decline in the structural primary surplus is o f concern both for the sustainability of fiscal adjustment and the management o f the fiscal stance. From a sustainability perspective, the positive cyclical contribution to the budget may wane if growth were to slow down in the future, triggering a decline in the primary fiscal surplus and hindering the reduction o f public debt. From a fiscal policy perspective, the declining structural primary surplus, especially in 2004-05, has added fiscal stimulus to domestic demand and indicates that the fiscal stance has turned pro-cyclical-a potentially unwelcome development at a time o f broadening external current account imbalances and slowing disinflation. To avoid a pro-cyclical fiscal stance, the structural primary surplus should be maintained unchanged during years o f robust growth. This would call for saving fiscal revenue over performance due to robust growth-a policy that would be reflected in an increasing actual primary surplus in proportion to GDP. An alternative option would be to introduce an expenditure cap together with a target for the actual (as opposed to the structural) primary fiscal surplus. 9. Turkey’s primary expenditures are still at the low end of EU countries but much higher than in other fast-growing emerging economies. At 33 percent o f GDP in 2005, primary expenditures o f the general government were lower than in European Union (EU) member countries, but about 10 percentage points higher than in a comparison group o f emerging economies. The difference mainly reflects a larger general government wage bill, which, at 9 percent o f GDP in 2005, has remained largely unaffected by fiscal consolidation. It also reflects large expenditures for current transfers, mainly comprising social benefits, despite Turkey’s favorable demographics. Despite significant downsizing since 2001, public investment remains comparable to levels seen in other EU.members, and higher than in some other emerging economies. However, infrastructure gaps in Turkey are larger than in some new EU members. 10. Functional allocations of public expenditures reveal potential for trade offs and room for improved efficiency and cost containment. Expenditures in general public services and in defense and public order and safety are above the average o f comparator countries by the equivalent o f 0.8 and 1.2 percentage points o f GDP respectively. By contrast, expenditure on environmental protection i s comparatively low and would need to increase on the way to the EU. Expenditure on health care i s relatively high in international comparison, while outcomes are below the average o f comparator countries. Similarly, expenditure on social protection i s high by international comparison in view o f Turkey’s favorable demographics. Expenditure on education i s in line with comparator countries but educational attainment remains weak. Horizontal public expenditure reforms 11. Sector-specific expenditure reforms will provide a betterpay-off when combined with “horizontal ” reforms that improve the efficiency o f expenditure programs across sectors. Horizontal reforms should include initiatives aimed at strengthening incentives in the civil service towards the achievement o f results, while containing pressures o n the wage bill. A larger share o f expenditures could thus be spent on materials and equipment necessary for the provision o f high-quality public services. Adoption o f efficient project xi evaluation methods, effective multi-year programming, and adequate provisioning for long-term operation and maintenance o f public capital will also help improve the quality o f the investment program across functional expenditure areas. 12. Managing personnel costs would merit attention as the Government wage bill has been bypassed by fiscal consolidation. From 1999 to 2004 total compensation for General Government employees hovered at around 10 percent o f GDP, before declining to 9 percent o f GDP in 2005. However, the wage bill could increase again as a result o f the additional pay rise for civil servants granted for 2006, the estimated cost o f which is 0.35 percent o f GDP. W h i l e the wage bill has remained rigid, employment in consolidated budget agencies and municipalities has declined by 3 percent since 1999, resulting in fast growth o f average compensation. 13. The share o f personnel expenditures on economic affairs and security is particularly high in international comparison. Personnel expenditures are concentrated o n education, security, and health care, which absorb more than two-thirds o f the wage bill. The share o f personnel compensation in economic affairs-for the provision o f infrastructure services, rural development, and services to agriculture-is twice as high as the average o f comparator countries, revealing a possible room for savings. 14. Slower growth in average compensation would create room for fiscal savings. Average compensation in consolidated budget agencies, when measured in proportion to per capita GDP, is higher than in comparator countries for which similar information is available. Over 2000-2004, had the ratio o f average compensation to per capita GDP remained constant at i t s average level in 1999-2000, the wage bill o f consolidated budget agencies would have been reduced by about 1 percentage point o f GDP. Linking wage , increases to inflation has not proven enough to contain growth in average compensation, probably because other forms o f compensation have provided room for increases in personnel costs. 15. Moving to a simpler and more transparent compensation system would contribute to containing pressures on the wage bill. About 32 percent o f personnel compensation i s composed o f various allowances, compensations, and rewards. Many o f these side benefits are exempt from the income tax, with estimated foregone tax revenue o f around 0.5 percent o f GDP. The system o f allowances and overtime pay is complex and offers substantial opportunities for discretion. 16. The Government is considering plans for civil service reform to address a number of major objectives: 0 improving efficiency-by making i t easier to adjust staffing composition when required; enhancing sources o f performance motivation; facilitating opportunities for human capital improvement o f public servants; 0 eliminating discrepancies in remuneration across similar positions within the public sector-by consolidating the legal framework governing salary setting; concentrating salary in basic duty wage and duty difference allocation; xii  0 making i t easier to ensure a fiscally sound wage bill-with the consolidation o f legal frameworks expected to facilitate monitoring o f employment and pay policy changes; 0 providing competitive opportunities for entry into the public administration-by further strengthening requirements for standardized examination-based competitive recruitment and promotions 17. All of the proposed changes in the legal framework represent important steps towards achieving those objectives but some changes are likely to pose implementation challenges. These include in particular: (i) establishing a clear legal distinction between civil servants and contracted personnel; (ii) mandating dismissal o f c i v i l servants for repeated unsatisfactory performance rating; ( iii)introducing performance bonuses; (iv) concentrating salary in basic duty and “duty difference” components; (v) managing the fiscal cost o f compensation harmonization by setting a new structure o f basic salary and wage scale index; and (vi) monitoring the reliability o f recruitment and promotion procedures. 18. Public investment was cut significantly as part o f the fiscal adjustment efforts during 2001-2004 at a time when rationalization o f the public investment portfolio was overdue. Total public sector investment fell to 4.2 percent o f GNP in 2004, from 6.8 percent in 2000, before recovering to 5.1 percent in 2005, and the brunt o f adjustment was borne by investment in infrastructure. In the past, unclear criteria and processing rules had resulted in “over programming” o f the public investment portfolio (PIP), with the average project completion time as high as 12 years in 2001. Rationalization o f the PIP has been quite effective in reducing the number o f projects and the average completion time to 6.7 years by 2005. However, as a result o f the downsizing o f the PIP, the number o f projects put on hold has significantly increased, with “trace allocation” projects accounting for about 10 percent o f the total value o f the PIP in 2003-04, a significant rise from the 1999-2000 average o f only 3 percent. 19. Meeting financing requirements o f infrastructure is a key challenge for sustained growth. I n order to emulate EU standards o f infrastructure service provision, Turkey must systematically improve the quality o f infrastructure investments in the medium term. The quantity o f infrastructure investments will also need to increase in some key areas for EU accession. Key issues in infrastructure financing in Turkey include (i) protecting high-priority investment projects from fiscal adjustments, and ensuring that key multi-year projects are carried on and completed on time, (ii) long-term provisioning for operation and maintenance o f existing assets, (iii) ensuring adequate security and reliability o f supply o f infrastructure services, and (iv) ensuring that infrastructure investment addresses economic development objectives. More effective utilization o f the recently introduced medium-term expenditure framework and enhancing the participation o f operating agencies in actual budget allocation decisions will be important in addressing these challenges. 20. An important challenge is to create a framework conducive to a greater role of the private sector in the jinancing, development, and operation of infrastructure. In xiii view o f the large financing requirements in infrastructure, achieving the Government’s objectives o f secure and reliable infrastructure services in a sustainable way will require sufficient support from private investment. A predictable policy and regulatory environment will be required to encourage Private-Public Partnerships (PPPs), together with careful design o f Government commitments to private operators to minimize the risk o f contingent liabilities. In the energy sector private investment n o w makes up about 50 percent o f electricity generation, but this has been made possible only by significant contingent liabilities o n the Government in the form o f guarantees and off-take agreements. Letting the regulatory framework operate independently, and allowing prices to fully reflect costs would make i t possible to attract private investment without significant guarantees. 21. The quality o f the infrastructure portfolio needs to be assessed in order to determine which projects are worth pursuing further, dropped, or redefined in their scopes. This involves carrying o n a project-level evaluation and revising the rationalization policies and criteria, while establishing a proper prioritization o f investments aligned with the Government’s development objectives. Projects that are considered economically unviable due to cost escalation, changing government priorities, or even an initial deficient project appraisal, should be individually assessed and either redefined or dropped as the case may be. Proper economic and financial appraisals o f projects and improved incentives and pricingkubsidy policies are necessary for enhancing efficiency. 22. Appropriate allocations for operation and maintenance o f public capital need to be ensured. This is a requirement not only for improving the quality o f services, but also for restricting the cost o f the projects as accelerated decay o f capital goods may unduly burden the investment program in the future. Appropriations for maintenance declined from 0.4 percent o f GNP in 1999-2000 to 0.2 percent in 2003-04, in particular in the transportation sector, health sector, and other public services. Such reductions in O&M expenses should be avoided since they are not sustainable while they reduce the contribution o f public capital to efficiency and growth. 23. The medium-term challenge is to create fiscal space for public investment while preserving fiscal discipline. In order to complete all on-going projects, Turkey would require about YTL 82 billion over the next 5 years: an annual average investment o f roughly 2.5 percent o f GNP for 2006-10. Funding for infrastructure would thus need to substantially increase compared to current levels. Due to the overarching requirement to preserve fiscal sustainability, the level o f investments that can be made is constrained not so much by the availability o f financing but by the need to meet the primary balance targets in the face o f strong expenditure pressures in some other functional areas. Creating the needed fiscal space through structural expenditure reforms and cost containment in these areas would be the key requirement for stepped up investment in infrastructure in the future. Well designed Private-Public Partnerships for the provision o f infrastructure services will also help alleviate existing constraints. xiv Structural expenditure reforms in key sectors 24. Public education expenditures have been on a downward trend since the 1999 crisis. At approximately 4 percent o f GNP o n average over the past decade, public education expenditures have reached an estimated 4.17 percent o f GNP in the 2006 budget-less than the 4.33 percent o f GNP spent in 2002. Including private financing, Turkey spends on education the equivalent o f 7 percent o f GDP. There i s significant private financing o f courses for high school and college entrance exams, and private financing also significantly contributes to the operation o f public schools as a consequence o f the continuing financial constraints faced by the public schools. 25. Total education expenditures in proportion to GDP are higher in Turkey than in most OECD countries, but expenditures per student vary widely by level o f education. Turkey’s per student spending on pre-primary schools i s very low, while spending in primary schools is consistent with expectations accounting for differences in per capita income. For secondary and tertiary education, Turkey spends somewhat more per student than what would be predicted on the basis o f its per capita income. 26. Even though per student expenditures in secondary education are comparatively high, student achievement i s lagging behind comparator countries. Turkey’s educational achievements, as measured by the learning scores in the PISA study, remain relatively poor. Importantly, Turkey’s score remains below the level that could have been expected in view o f the relatively high per student expenditures in secondary education. Although student achievement, as measured by PISA indicators, i s expected to increase with per capita spending, i t would be counterproductive to further increase expenditure without, at the same time, taking steps to ensure that the impact o f spending has improved. 27. The differences in educational expenditures at public and private schools are significant. Per-student spending in public primary schools i s almost two-thirds lower than in private primary schools, while per-student expenditure in public general secondary schools is almost h a l f the spending at private schools offering the same education. Public schools, regardless o f the level o f education provided, spend more o n personnel relative to private schools, while use o f books, supporting educational materials, and school support services is more intensive in private schools. Thus, educational outcomes vary by school type, with primary school students studying in private primary schools performing about twice as well as those in public primary schools. 28. Improving education outcomes will require additional financial resources together with increased institutional efficiency. New education policies to be introduced by the Government will have a significant medium-term impact on the budget, while existing educational gaps suggest the need for additional fiscal space for educational expenditures: xv  a improving the quality o f educational achievement in public schools would require increasing the share o f n o n personnel expenditures to levels similar t o private schools and universities; a improving the regional balance o f educational services across the country would require enhanced public education service provision in most backward provinces to compensate for possible lack o f private sector initiative-while taking advantage o f opportunities for public school retrenchment where private sector initiative i s strong; a increasing per student expenditure o n pre primary education-an area where progress i s critical to encourage greater participation o f women in the labor force. 29. The tight fiscal framework in the years ahead makes it necessary to improve the efficiency of public spending on education. A first priority is to implement policies that encourage or force schools to fill their classrooms up to acceptable capacity levels. This would increase the efficiency o f spending by reducing the number o f teachers employed for the same level o f enrollment. Second, schools need to be given autonomy over their resources and held accountable for their results. A system o f per-student financing could be considered, with funding allocated in proportion to the number o f students a school enrolls, up to its capacity. Schools would be responsible for using these hnds to cover their personnel, material, and professional development costs, and would be held accountable for results. 30. Higher education institutions need to be given greater autonomy over financial resources, with incentives to manage these resources efficiently. To become more efficient and provide better results, universities need to move funds between items and activities during the course o f the year. Within a financially autonomous system, the amount o f funding allocated to each university should not be based o n inputs, but o n performance, in terms o f outputs for teaching, research and service development. Increases in financial autonomy should be matched by a commensurate increase in accountability for internal accounting, reporting to independent boards, and producing annual reports. A comprehensive approach need to be designed to achieve administrative and financial autonomy o f higher education institutions. In a complementary way, the university entrance examination system should be modernized in such a way as to eliminate the advantage o f the widespread private examination preparation courses. 3 1. The Turkish pension system has been a perennial drain on public finances as it continues to run large deficits. Turkey i s a young country and should not be experiencing pension system deficits o f the current magnitude. However, because o f very l o w retirement ages, high benefit accrual rates, and other disincentives to remain in the formal labor force, pension deficits have persisted, with the exception o f a temporary improvement in 1999, when parameters improved with the passage o f a pension reform law. 32. The 2006 reform will bring fiscal balance o f the pension system in the long term as well as the unification o f a previously fragmented system. With adequate implementation o f the reform, the pension system deficit i s expected to decline by 1 xvi percent o f GDP below a “no reform” projection by 2020. The deficit will be absorbed by 2040, with a fiscal gain o f 7 percent o f GDP compared to a “no reform scenario”. However, the beneficial impact o f these measures is expected to be rather limited over the medium run, thus creating little fiscal space for expenditure reallocations towards more productive uses. Moreover, long-run financial sustainability will be secured only thanks to a high level o f payroll taxes by international standards. High contribution rates hamper j o b creation in the formal sector, and eventually deprive the budget from significant revenues. If lowering contribution rates is a desired objective, more changes will be required to achieve that goal by generating sufficient surpluses in the pension system. 33. Despite the strong ambition o f the 2006 reform, the pension system still offers options for further fiscal savings that would help reduce payroll taxes in the future. Retirement ages remain substantially below those in other OECD countries. As the 2006 law will increase retirement ages only for new entrants to the labor force, low retirement ages will prevail for a very long time, becoming comparable to the OECD average only by 2048. Accrual rates o f benefits per year o f contribution are coming closer to OECD levels, but will not reach OECD averages even with the new pension law. High accrual rates undermine incentives for longer careers into the formal labor force. This works against the formalization o f the labor force-a key requirement for a better fiscal performance o f the pension system. The new law stipulates that the full working career will be 25 years, but this is being phased in extremely slowly so that only those who begin work in 2026 will face the 25 year constraint. Furthermore, the new system continues to offer the possibility o f a partial pension with only 15 years o f contributions. 34. The combination o f poor health outcomes and relatively high expenditures suggests that important room exists for improving the efficiency o f the health care system. Turkey spent approximately 6.6 percent o f i t s GDP on health in 2004-a higher level than comparator countries. Public health care expenditures accounted for 68 percent o f total, up from 61 percent in 1999. Although the health status o f the population has considerably improved, health outcomes still compare unfavorably with countries that have similar or lower level o f GDP per capita and approximately the same share o f public spending on health. Improvements in health indicators are not spread evenly across urban and rural areas and across regions, with infant and under-five mortality rates in rural areas and in the Northern and Eastern regions much higher than in the rest o f the country. High levels o f infant and under-five mortality, as well as high maternal mortality, are related to the poor usage o f preventive care. 35. Public expenditures on health care are on a strong upward drift. Public expenditures have exceeded 5 percent o f GNP in 2004 and 2005, up from 3.3 percent in 1999. The increase in health expenditures in recent years i s almost entirely explained by fast growing spending on hospitals and pharmaceuticals, while the share o f expenditures on preventive services has actually decreased (from 11 percent in 1999 to 7 percent in 2005). The increase in the quantity o f medical services partly reflects expanded coverage and access-even before the official enactment o f Universal Health Insurance (UHI).The change o f incentives facing medical providers in Ministry o f Health hospitals and xvii outsourcing o f services may have led to supplier-induced demand and an increase in production, provision and utilization o f health services. 36. Efficiency and equity should be simultaneous objectives o f health care reform. The introduction o f UHI will improve access and equity but will undoubtedly generate greater demand for public expenditures o n health. T o remain fiscally viable, the introduction o f UHI should go in tandem with marked improvements in efficiency o f public spending. Turkey should also consider devoting a higher share o f public expenditures to preventive activities, while targeting resources where they are most needed, to address the persistent variation o f health indicators between urban and rural areas and among regions. 37. Consolidation o f hospitals and greater management autonomy hold the promise of efficiency gains. M a n y MOH hospitals are too small in size to allow for efficient operation and provision o f care, and have significantly lower utilization rates compared to University hospitals. The introduction o f UHI provides a good opportunity to further strengthen the gains f i o m the merger o f MOH and SSK hospitals under MOH ownership and management. Further gains will come about by improving efficiency in the use o f hospital resources and overall management and accountability, which would be facilitated by granting financial and administrative autonomy to public hospitals. 3 8. The challenge for cost-containing measures i s to manage q u a n t i w s p e c i a l l y to contain pharmaceutical expenditures-without adversely affecting access, utilization and effectiveness o f services. The principal measure suggested to control quantity and dissuade indiscriminate use is the introduction o f copaymentsfor outpatient visits and for drugs. Even nominal levels o f copayments would help rationalize the consumption o f health services and result in savings, not only in the production o f health services but also in the consumption o f drugs. 39. Changes in provider payment systems should introduce incentives for physicians to provide quality care at lowest costs. Key options involve paying family physicians o n the basis o f capitation fee per enrollee, so as to encourage conservative use o f health care services. Such changes could be complemented with prospective payment mechanisms, introduced at the hospital level-according to which hospitals get reimbursed according to a pre-fixed rate per bundle o f services associated with a particular treatment. To make such changes work, hospital care providers and managers would need the flexibility and tools to actively manage their resources and redirect their use. 40. Redeploying o f rural public expenditures would be needed to improve their efficiency. At 1.2 per cent o f GDP, the level o f public rural investments i s l o w compared to the size o f the agricultural sector and its modernization needs-although rural investments have recently been o n an upward trend. Greater emphasis would be needed in areas such as land consolidation, storage and pre-marketing facilities, and alignment with the EU acquis o n safety and phytosanitary standards-with particular focus o n small fanners with weak l i n k s to markets-and away from irrigation projects. Agricultural xviii transfers have been significantly reduced and restructured by switching the main focus o f agricultural transfers to the Direct Income Support (DIS) program, with reduced artificial incentives for inputs and particular crops. As recognized by the Government, the D I S Program needs to be accompanied by transfer payments with the aim o f promoting productivity and market development more directly. But with more than a third o f rural expenditures on salaries o f government employees, there i s little room in the budget for investments as well as other operation and maintenance expenses. With greater emphasis o n participatory approaches and private sector involvement in service provision, more effective and efficient use o f funds could be achieved. 41. Horizontal and sectoral expenditure reforms to contain costs should be combined with expenditure reallocations to maximize the expected efficiency gains. Tables A and B, at the end o f the overview section, summarize some key reform options in expenditure programs discussed in this study with a view to promoting expenditure efficiency and realizing fiscal costs over the medium term. Some o f the proposed measures aim to directly contain short-term expenditure pressures, especially in health care, and it would be advisable to consider implementation in 2007. Other reforms should be planed over the medium term, during 2008-09, but preparatory steps would need to be taken earlier. The tables provide an illustrative implementation time frame o f the proposed reform options. In some sectors, some proposed measures would entail expenditure reallocations or an increase in expenditure levels over the medium term- such as, for example, measures to strengthen pre-primary education or preventive health care. In these cases, increased expenditures could improve the functioning o f a sector and thus facilitate the realization o f cost savings-such as, for example, an increase in preventive health care expenditures that would help contain pharmaceutical and hospital costs over the long run. Tax revenues andprogress in tax reform 42. The composition o f revenues has shifted markedly toward indirect taxation over the last few years. Hikes in excise rates along with buoyant VAT collections in the post- crisis years combined to increase indirect tax revenues from 11.6 percent o f GDP in 1999 to 16.7 percent o f GDP in 2005. During the same period, personal income tax revenue f e l l from 6.5 percent o f GDP in 2001 to only 4.9 percent GDP in 2005, while corporate income tax revenue remained broadly stable. As a result o f the reliance o n revenue- increasing measures, tax revenues in Turkey (including social security contributions) reached nearly 32 percent o f GDP in 2005, with an estimated tax effort o f about 30 percent higher than expected in international comparison. Primary fiscal revenues stand at 40 percent o f GDP, a level similar to EU countries despite Turkey’s smaller welfare state. 43. Reliance on indirect taxation may have a regressive incidence on income distribution but offers some advantages that should be taken into account in the design of the tax system. Given pervasive informality, in the short term, reliance on indirect taxes appears to be a reasonable way o f ensuring some degree o f compliance o f those who would have otherwise escaped taxes altogether. In addition, indirect taxes uniformly xix levied on consumption offer advantages on efficiency grounds as they are relatively neutral towards saving and investment decisions while they do not distort incentives to work. However, greater reliance on indirect taxes should not be accompanied by proliferation o f special regimes and exemptions. In particular, reduced VAT rates introduced recently for textile and clothing products are o f questionable effectiveness as a means o f assisting domestic producers to withstand competition and tackle the problem o f fraudulent refunds. Multiple V A T rates and exemptions increase the complexity o f the tax system, create more opportunities for misreporting and misclassification for tax purposes, and thus work against efforts o f strengthening tax compliance. 44. Because there is little fiscal space for further lowering the tax rates in the short term, reform efforts should focus on tax rationalization and base broadening. This will improve tax efficiency while laying the groundwork for lowering tax rates in the future. Important reforms o f the Corporate and Personal Income Taxes have been initiated, with a reduction in the CIT rate and tax base broadening, and a simplification o f the rate structure o f the PIT, together with the harmonization o f the taxation o f wage and non- wage income. 0 Further reform options for the Personal Income Tax (PIT) could include: replacing the costly and potentially inequitable VAT-based consumption credit with a simpler tax credit; introducing a reduction for dependents so as to reduce the high tax wedge o n married workers; maintaining neutrality in the taxation o f interest income across asset classes and avoiding discrimination among domestic and foreign investors. 0 Further reform options for the Corporate Income Tax (CIT) could include: avoiding regional and sectoral tax incentives; and incorporating dividends into a new regime for capital income taxation, with a view to progressively phasing out the double taxation o f dividends. 45. Establishing an adequate tax expenditure framework would promote fiscal accountability and transparency. Aware o f the challenges, the Government has initiated work to analyze and estimate tax expenditures. Based on draft tax benchmarks, initially some 186 tax provisions were identified as generating potential tax expenditures. Some o f the major tax expenditures are granted with the aim o f promoting investment, employment, regional development, R&D, and trade incentives (primarily through the CIT). A range o f tax expenditures aims at assisting low-income tax payers, in support o f social policy. These are delivered primarily through the PIT and C I T and, to a lesser extent, the VAT. Some major tax expenditures in this latter group include the deductibility (for PIT purposes) o f contributions to the social security system for pensions, disability, unemployment and health benefits as well as contributions to private pension and insurance company-based pension and health policies. 46. The estimation o f tax expenditures is still in its early stages but preliminary estimations point to a significant fiscal cost. As part o f the PFMC Law, the authorities have estimated that tax expenditures for 15 tax provisions will cost 1.6 percent o f GDP per year over 2006-08-including, however, the investment tax allowance, since phased out. According to very approximate estimations, including part o f omitted tax xx expenditures would raise total tax expenditures to at least 5 percent o f GDP per year, or about 18 percent o f government revenues in 2003. This level o f tax expenditures would place Turkey towards the high end o f the OECD country range. A sustained effort i s required over the medium term to rationalize tax expenditures, and make their selection and associated trade-offs with direct spending programs an integral part o f the budgeting process. Required fiscal space to meet medium-term expenditure pressures 47. Expenditure pressures are growing in key sectors while the development challenges on the way to the EU are amplifiing expenditure needs. Several new policies have been initiated or proposed in education that will have a signiJicant impact on the medium-term expenditure program. Simulated new or proposed policies would increasingly boost annual expenditures, by up to 0.5 percent o f GDP in 2009. Health expenditures are projected to increasefrom 5.2 percent o f GDP in 2006 to 6.1 percent in 2009-assuming no effective cost containment through co- payments for outpatient visits and drugs. To sustain income convergence and alignment with the EU, investment in infrastructure, operation and maintenance, and environmental expenditures will all need to increase. Combined annual general government investment in infrastructure and environmental expenditures would need to increase by at least 0.6-07 percent o f GNP during 2007-2009. 48. The social security system deficit is set to decline, generating some modest fiscal space. With the enactment o f the new social security laws, revenues o f the system are projected to increase by 0.3 percent o f GDP until 2009, while expenditures would decline by 0.2 percent o f GDP. As a result, the social security system deficit i s estimated to decrease by h a l f a percentage point o f GDP until 2009-with hrther savings expected in the long term. 49. A sustainable budget envelope for the financing o f primary expenditures at the general government level i s estimated at around 33 percent o f GDP until 2009. A sustainable expenditure envelope i s calculated based on a revenue projection and the assumption that the primary surplus for the whole public sector would be kept at 6.5 percent o f GDP. Maintaining a high primary surplus will be needed to further reduce the public debt ratio. Despite the substantial improvements in the level and composition o f public debt, Turkey’s debt ratio i s relatively high and its maturity short. The recent turmoil in international markets and i t s transmission to domestic markets justify the need for a lower debt ratio to ensure macroeconomic stability. 50. Significant fiscal space to meet expenditure pressures will be required during 2007-2009 under the baseline and alternative scenarios. The fiscal space needed to meet expenditure pressures i s calculated as the difference between the projected primary expenditures and the sustainable expenditure envelope. In the baseline projection, the needed fiscal space i s estimated at 1.1 percent o f GDP in 2007, reaching 1.7 percent in xxi 2008 and 2 percent in 2009. In a scenario with a stronger reform initiative in education, the expenditure pressures would be larger, by 0.3 percent o f GDP each year. In an alternative scenario, assuming a larger primary fiscal surplus-as needed in order to reduce the public debt ratio to the levels projected before the recent market turmoil-the sustainable expenditure envelope would be reduced by an estimated 1 percent o f GDP. The required fiscal space would thus increase by an equivalent amount. With a stronger effort in education and more ambitious reduction o f the debt ratio the fiscal space needed would increase from 2.4 percent o f GDP in 2007 to 3.3 percent by 2009. Using expected privatization revenues to repay debt would significantly reduce the fiscal space needed to reduce the debt ratio. 5 1. Fiscal space to meet expenditure pressures should be created by a combination of structural public expenditure reforms and expenditure reallocations. Structural fiscal reforms, aimed at improving the quality o f fiscal consolidation, are the only viable means o f sustaining the adjustment in the future while making appropriate fiscal space for growth-enhancing expenditures and lower taxes in the future. Reforms that improve the efficiency o f expenditure programs would be required in sectors where expenditure pressures are being felt. Trade-offs in expenditure allocations would also need to be considered, with the aim o f shifting resources to priority programs for growth and social development. Sector-specific expenditure reforms should be underpinned by horizontal reforms that improve the efficiency o f expenditure programs across sectors. Initiatives to fbrther rationalize the tax system-including the tax expenditures-would also create fiscal space by broadening the tax bases. Strength ening budgetary institutions 52. Far-reaching public financial management reforms have improved budget coverage, formulation, execution, accounting, audit, and procurement. Enactment o f the Public Financial Management and Control (PFMC) Law in 2003 marked a defining moment, providing a new legal framework for modern public expenditure management and accountability. Several other laws have been or are being enacted, including the Special Provincial Administration Law, the Municipalities Law, the Metropolitan Municipalities Law, The Turkish Court o f Accounts (TCA) Law. The law for redefining the revenue sharing between the central and local administrations is also in Parliament. In 2004, a GFS-compliant budget classification system was adopted. An online networked accounting system for the central government and an automated tax management system have also been implemented. The new laws are inspired in part by the prospect o f EU accession and by the government’s desire to introduce greater performance orientation in budget management. 53. The coverage o f the 2006 budget and medium-term fiscal framework has been expanded to all central government institutions but some steps are still pending to broaden the comprehensiveness o f the budget. As defined by the PFMC, the budget approved by the Parliament covers the general budget, special budget, and regulatory and supervisory institutions. Moreover, several extrabudgetary funds have been closed down, creating a framework more conducive to fiscal discipline. However, even though the xxii PFMC L a w required incorporation o f extrabudgetary funds into the budget o f the related administration, they continue to operate outside o f budget discipline, with insufficient scrutiny o f their operations. Similarly, revolving funds continue to operate outside the budget coverage at the risk o f distorting budget discipline, further to a decision to restructure the revolving funds by the end o f 2007 with an amendment made to the PFMC in December 2005. The Government’s preferred approach i s to restructure remaining revolving funds, rather than including them in the budget. 54. Properly phasing-in and completing the implementation o f the accrual accounting system requires addressing complex issues. K e y issues include valuation o f assets and inventories, depreciation rates, and gains and losses from foreign exchange. As shown by the experience o f countries that have switched to accrual accounting, implementation generally works best when done in phases. Extensive training i s necessary in how to use and interpret accrual accounting statements. Implementing full commitment accounting is also needed as, until recently, only commitments for large multiyear capital expenditure projects were fully captured. Recording o f all commitments will ensure budget discipline and avoid the possible accumulation o f arrears. 55. The PFMC Law has introduced a modern internal audit framework- The PFMC L a w requires each public administration to establish an internal audit unit. The law clearly articulates the accountability o f ministers and other high-level c i v i l servants. I t also requires every l i n e ministry to prepare an annual accountability report-also including performance data and details in the implementation o f the strategic plan and budget. 56. -but implementation o f the law itselfposes a major challenge as it envisions a substantial change in the public internal _financial control framework The law envisages a uniform implementation schedule across the general government, irrespective o f the size or capacity o f the agencies concerned. This raises an implementation challenge, as smaller municipalities may either ignore the provisions that they may consider too expensive, o r they may implement the provisions by focusing o n form rather than substance. A transition strategy needs to be devised, outlining interim measures while the new institutional structures become fully operational. 57. The Turkish Court o f Account’s (TCA) drap law creates a sound basis.for the development o f government audit and should be complemented by additional initiatives. In particular, the audit side o f T C A is not yet fully developed as the draft law i s still insufficient in terms o f structures and responsibilities. Even after enactment o f the law, T C A will s t i l l be dominated by its judicial side. Developing capacity for financial statement audits would be necessary since the government will be producing accrual- based financial statements in addition to cash-based budget execution reports. Moreover, the envisioned increase in the role for local administration prompts the need for reconsideration o f audit policies for local government. TCA’s role might be redefined in several ways-for example, with different strategies for local administrations o f differing sizes; creating a new local government audit service; using private sector auditors to audit local administrations. xxiii 58. Since 2001 the public procurement system has made good progress, yet some significant issues require attention. A new Public Procurement L a w (PPL) and secondary legislation was prepared. However, other laws have made several amendments to the PPL, mainly to provide exemptions to various agencies and to exclude procurement o f certain services from the scope o f the law. Internal consistency and integrity o f the original law should be sought before making amendments. All draft laws amending the PPL should be prepared and finalized in close consultation with the Public Procurement Authority. Improvements are also required as part o f the alignment with the EU ucquis- in particular because: (i)the 15 percent national preference for public contracts above the threshold i s incompatible with the ucquis; (ii) there i s lack o f sector (utilities) legislation; ( iii)a horizontal legal framework for concessions i s lacking; and (iv) a platform for Electronic Procurement i s s t i l l to be developed. 59. Despite progress, addressing the unfinished agenda in public financial management laws represents a major challenge and requires strong coordination and monitoring. In some instances implementation o f the reform agenda has been hesitant, and some regressive tendencies in the implementation o f the P F M C law tend to perpetuate budgety fragmentation. Although some aspects o f the new legal framework are being piloted intermittently, some o f the secondary regulations related to practical implementation are yet to be issued. The government’s decision to undertake fkndamental changes in the responsibilities o f provincial and municipal governments also injects new complexity and considerable uncertainty into the P F M C implementation schedule. Enactment o f the Omnibus L a w has created concern that its provisions run counter to the intention o f the PFMD and P F M C L a w as approved by the Parliament. Without addressing the unfinished elements o f the agenda and arrangements to ensure effective leadership and coordination among agencies, implementation o f the overall PFM reform agenda will remain at risk. xxiv 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 . . 2 .3 X . . . e e e n m L m % 3 f? $ . Y I 3 E L a2 Y B 1 3 .I e e e e e n L m % E 0 W m a 0 YI . 3 E L a2 Y f: 0 E m  c . . . . . ... . . . . . . .... e e e e . .- X X X  CHAPTER I SOURCES OF FISCAL CONSOLIDATION AND THE QUALITY OF FISCAL ADJUSTMENT 1.1 Improving the quality o f fiscal adjustment is a key challenge. The Turkish economy in the 1990s has been characterized by persistent fiscal imbalances and fiscal dominance, a pattern that contributed to growth volatility, chronic inflation, and continuous macroeconomic instability. A s high public sector borrowing requirements became unsustainable and structural problems mounted, it became inevitable for Turkey to embark o n an ambitious reform program aiming at macroeconomic stability and fundamental restructuring o f the economy following the 2001 financial crisis. The analysis in this chapter focuses o n the quality and sustainability o f the strong fiscal adjustment achieved so far, since maintaining hard-won macroeconomic stability is a pre- requisite for promoting faster growth o n a sustained basis. At the same time, the analysis benchmarks Turkey’s size and composition o f public expenditures and fiscal revenues in international comparison, as achieving an efficient allocation o f expenditures and an efficient distribution o f the tax burden are key preconditions for fast growth. 1.2 The first section reviews the patterns o f fiscal consolidation since 1999 and examines shifts in the focus o f fiscal adjustment. Section two benchmarks public expenditure allocations at the general government level in international comparison. Section three reviews trends in tax revenue mobilization and contribution to fiscal consolidation, as well as major tax initiatives taken recently. The fourth section completes the analysis o f fiscal adjustment by estimating discretionary changes in the primary budget balance separately from changes due to cyclical automatic stabilizers. Despite possible limitations-due to the difficulty in estimating tax elasticities in the face o f frequent changes in taxes-this analysis provides a more accurate view o f the fiscal consolidation effort. A. PATTERNS OF FISCAL ADJUSTMENT IN TURKEY 1.3 Turkey has gone through an impressive fiscal adjustment during 1999-2005- In the aftermath o f the 2000 and 2001 crises, Turkey had to generate a sizeable primary surplus to reduce its public debt stock. Turkey’s fiscal position improved dramatically between 1999 and 2005, as reflected in the improvements in the primary balance o f the Consolidated General Government (CGG), fkom a negligible primary surplus o f 0.5  (See Box 1.1 for percent o f GDP in 1999 to a surplus o f 6.1 percent in 2005 (Figure l.l).' the conversion from the SPO-based consolidated general government primary surplus to IMF program definition o f public sector primary surplus). The C G G borrowing requirement excluding privatization proceeds decreased by 11.8 percentage points o f GDP between 1999 and 2005, driven by the increase in the primary surplus and the decline o f interest payments on public debt. Turkey's public sector-comprising the consolidated general government plus the State Owned Enterprises (S0Es)-achieved an even larger fiscal consolidation in 1999-2005, from a primary deficit o f 1.6 percent o f GDP in 1999 to a primary surplus o f 6.9 percent in 2005.2 1.4 -which was largely revenue-driven. Fiscal adjustment at the C G G level was mainly driven by a substantial tax effort achieved from 1999 to 2001 and maintained thereafter (Figure 1.1). By contrast, primary expenditure has hovered at around 33-34 percent, without any noticeable reduction during the period. ~ Figure 1.1 Fiscal balances and primary expenditures and revenues of the General Government (1999-2005, in %of GDP) 10.0 l- - -I^- 41.0 ~ 1999 2000 2001 2002 2003 2004 2005 Source: World Bank staff calculations 1.5 Reflecting a sound fiscal framework, interest payments on public debt have declined significantly. Strong fiscal adjustment helped Turkey reduce its gross public debt in proportion to GDP ratio to 72 percent in 2005 from 108 percent in 2001 (with the net debt ratio declining, during the same period, to 55.8 percent from 90.5 percent). Declining debt ratios and lower interest rates-thanks to receding inflation and stronger confidence in the ability o f the Government to restore macroeconomic stability-have led ' While Annex I presents detailed data sources and methodology, it should be noted that the definition o f general government used in t h i s report includes the (i) consolidated budget, (ii) social security institutions (SSI), ( iii)a subset o f budgetary and extra budgetary funds (EBFs), (iv) local administrations, and (v) off- budget revolving funds created by public entities. The analysis does not include the revenues and expenditures o f the Central Bank and other public depository institutions, 36 non-financial state economic enterprises (SOEs), nine regulatory and supervisory agencies, and 45 out o f a total o f 50 special budget institutions listed in the Public Financial Management and Control L a w (PFMC). Special budget institutions refer to 50 public entities established as affiliated or related t o a m i n i s t r y t o provide certain public services. These special budget institutions receive revenues and are authorized t o spend them. The analysis includes the net subsidies and transfers between these institutions and the consolidated budget. ' The I M F ' s Stand-By Arrangement (SBA) program target o f 6.5 percent o f G N P primary surplus for 2005- 2007, includes primary surplus o f a subset o f SOEs together with the general government. The primary surplus data for the SOEs provided in Table 1.1 are f r o m IMF staff reports. 31 to sharp declines in interest payments on public debt: Interest payments, which had peaked at 25.4 percent o f GDP in 2001, up from 14.6 percent in 1999, declined to 9.6 percent o f GDP in 2005. However, interest payments s t i l l represented the equivalent o f 28 percent o f total primary expenditures o f CGG in 2005. Box 1.1. Conversion between SPO definition of CGG primary surplus to IMF’s program definition of public sector primary surplus: The backbone of the Government’s Stand-By Arrangements (SBA) with the IMF has been continued and strong fiscal discipline monitored through program definition o f the public sector primary balance. A detailed definition of the program primary surplus methodology can be found in the Fiscal Targets annex of the April 2005 SBA with Turkey. The objective of the IMF methodology i s not to monitor the consolidated public sector revenues and expenditures according to Government Financial Statistics (GFS) definition and the underlying primary surplus, but rather to focus on the fiscal balance measured through the primary surplus. Moreover, the program definition also does not account the revenues and expenditures that are not of continues and regular nature. In order to reach the program definition of the public sector primary surplus the following adjustments should be made to the SPO definition of the primary surplus. Table: Public Sector Primary Surplus; conversion fiom the SPO to the IMF Program definition (% of GDP) 1999 2000 2001 2002 2003 2004 2005 Primary Surplus (SPO definition o f CGG) 0.5 4.9 5.8 4.1 5.4 6.2 6.1 Adjustments on revenues (-) CB profit 0.1 0.2 0.3 1.1 0.0 (-) Transfer o f special revenues (-) Dividend revenues from state banks 0.3 0.3 (+) Tax arrears interest 0.1 0.1 (-) Revaluation difference 0.5 0.0 0.0 (-) Minting coins revenues 0.1 (f) Support Price Stabilization Fund repayment 0.0 0.0 0.0 0.0 (+) Interest revenues o f SSIs 0.0 0.0 0.1 0.0 0.0 0.0 0.0 Adjustments on expenditures (-) Retirement Bonus 0.1 (+) Mandatory Savings 0.7 0.5 (-) expenditures due to privatization activities 0.0 0.0 0.0 0.0 0.0 0.0 0.0 (+) PrivatizationFund net lending 0.1 0.1 0.1 0.1 0.1 0.0 0.1 (+) Defense Fund net lending (0.0) 0.0 (0.0) 0.1 0.0 (0.0) 0.0 (+) M ass Housing Fund net lending (0.0) 0.0 (0.0) (+) Interest expenditures of SSIs 0.1 0.1 0.0 (-) Risk Account 0.3 0.1 0.0 Program Adjusted General Government primary Surplus (0.4) 4.1 5.0 3.0 5.6 6.0 5.8 SOEs primary surplus (program definition) (2.1) (1.5) 0.1 1.1 0.7 1.1 0.8 Program adjusted public sector primary surplus (2.5) 2.6 5.1 42 6.3 7 . .1 _ 6.6 Source: World Bank staff 1.6 At the same time, fiscal transparency and accountability have considerably improved (see chapter 5). Turkey has aligned, to a great extent, i t s legislation o f public procurement, financial management and financial control with international standards. A public procurement agency and a risk management unit for public debt (under the 32 General Directorate o f Public Finance in the Treasury) have been established; most o f the extra budgetary funds have been integrated into the budget; the so called “duty losses” o f state banks (obligations to state banks and SOEs due to quasi-fiscal policies) have been accounted in a more transparent manner while generation o f new duty losses without an appropriation in the budget have been legally prohibited. More importantly, a new Public Financial Management and Control Law (PFMC) has been enacted as o f December 2003. The PFMC L a w aims at establishing a public financial management and control system compatible with international standards and EU norms. It extends the scope o f the budget; provides budgetary unity, increase effectiveness, fiscal transparency, and accountability during the process o f preparation and implementations o f budgets; ensures transparency in financial management; and restores the balance between authorizations and responsibilities in the spending process by establishing an efficient accountability mechanism. Therefore, a more transparent and close management o f the fiscal stance i s now achieved and quasi-fiscal activities are much harder to undertake. 1.7 Actions reflected in the consolidated budget (CB) and extra-budgetary funds (EBFs) contributed most to fiscal adjustment at the General Government level. Actions reflected in the consolidated budget (Central Government budget, accounting for 57 percent o f CGG expenditures in 2005) by itself created a fiscal adjustment o f more than 6 percentage point o f GDP (Table 1.1). With the primary surplus o f 9.9 percent o f GDP, the C B more than doubled its primary surplus compared to 1999. Controlling the extensive off-budget activities o f the past by eliminating the earmarked revenue system through the closure o f numerous Extra-budgetary Funds (EBFs) contributed to fiscal consolidation. Abolishment o f all budgetary funds, with the exception o f the Support Price Stabilization Fund (DFIF), and all but five extra budgetary Funds (Le., the Social Solidarity Fund, the Defense Fund, the Promotion Fund, SDIF and the Privatization Fund) improved fiscal discipline and brought the deficit o f the EBFs to a small surplus in 2005.3 Table 1.1: Institutional breakdown of Consolidated General Government Primary Balance (Yo of GDP) 1999 2000 2001 2002 2003 2004 2005 Consolidated Budget 3.8 6.2 6.2 6.7 8.9 9.2 9.9 SSIS -3.0 -1.9 -2.6 -3.1 -3.8 -3.8 -4.1 Local Administrations 0.3 0.5 0.7 0.1 -0.2 0.1 -0.3 Revolvingfinds 0.0 0.1 0.1 0.2 0.3 0.4 0.3 EBFs -0.6 -0.3 0.8 -0.2 -0.2 -0.2 0.0 Unemployment Insurance 0.0 0.3 0.6 0.3 0.4 0.4 0.3 Total Primary Surplus 0.5 4.9 5.8 4.1 5.4 6.2 6.1 Memo Item: SOEs -2.1 -1.5 0.1 1.1 0.7 1.1 0.2 Source: World Bank staff calculations based on SPO data The number o f EBFs has been reduced in 2000 and 200 1, and the related earmarked revenue system was abolished in 2004. 33 1.8 Further fiscal adjustment has been impeded by a growing social security deficit, which reached 4.1 percent o f GDP in 200X4 The fiscal situation would have been much better if Turkey had managed to contain the social security deficit after the 1999 reform. The overall deficit o f the Social Security Institutions (SSIs) increased from 1.9 percent o f GDP in 2000 to 4.1 percent in 2005. Given the favorable demographic profile o f Turkey, the currently high deficits pose a bigger challenge in the long run considering that the number o f elderly will start to increase as the demographic shift starts to kick in. As further explained in chapter 3, in the absence o f the social security reform enacted in June 2006, the projected deficit o f the system could have reached 6.7 percent o f GDP over the long term. 1.9 Recently enacted social security reform will help contain the deficits over the long term. With the help o f the recent reform the pension system deficit is expected to decline by 1 percent below the baseline projection o f “no reform” by 2020 (chapter 3). However, the beneficial impact o f these measures i s expected to be rather limited over the medium run, thus creating l i t t l e fiscal space for growth-enhancing expenditures. As further explained in Chapter 3, more ambitious steps will be needed to ensure financial sustainability if payroll taxes were to be lowered in the h t u r e in order to facilitate formalization and j o b creation. 1.10 The primary balance o f State-Owned Enterprises has significantly improved since 1999. The SOEs covered under the IMF program generated a deficit o f 2.1 percent o f GDP in 1999 but their fiscal situation improved substantially, resulting in a surplus o f 0.15 percent o f GDP in 2005 (Table l.l).’ This improvement was mainly an outcome o f reductions o f close to 2 percentage points o f GDP in their investment expenditures, and 1 percentage point cut in their personnel compensation. The personnel retrenchment o f the SOEs has been supported by the PFPSAL program. 1.11 With the privatization o f profitable SOEs, like Turk Telekom and Tupras, the contribution o f the SOEs to the public sector primary surplus will decrease, but foregone revenues may be offset from other sources such as increased tax receipts. These SOEs were also helping the fiscal position o f the consolidated budget through their n2005, the net transfer from the SOEs to the budget was Y T L net transfer to the budget. I 0.8 billion or 0.17 percent o f GDP. Revenue losses for the consolidated budget will be mitigated by corporate tax receipts from privatized SOEs. International experience suggests that additional tax receipts may offset foregone fiscal revenues from profitable SOEs as a result o f growing business activity in liberalized sectors. I n addition, with The deficit o f the social security institutions does not include payments made on behalf o f the consolidated budget since those expenditures are treated as part o f the consolidated budget expenditures. Therefore the deficit o f 4.1 percent o f GDP i s consistent with the more commonly used definition, (including payments made on behalf o f the consolidated budget) which corresponds to 4.8 percent o f GNP in 2005. The number o f SOEs covered in the program was eight in 1999, and t h i s number increased to 26 in 2005. The transfers from SOEs to the budget include: Dividend payments o f YTL 2.13 billion and Treasury levy o f YTL 0.17 billion. Transfers from the budget to SOEs include capital transfers o f YTL 0.9 billion and “duty loss” payments o f YTL 0.57 billion. Net transfers from SOEs to the budget amount thus to YTL 0.8 billion. 34 privatization revenues used to pay down public debt, there should be room for expenditure savings as a result o f lower interest payments on the public debt stock. 1.12 Fiscal adjustment at the general government level was mainly revenue-driven in 1999-2001. As shown in Figure 1.1 and Tables 1.1 and 1.2, the effort o f fiscal consolidation until 2001 reflects predominantly an increase in general government revenues, excluding SSI, and an improvement in the primary balance o f the SOEs-the latter up until 2002.7 During 1999-2001 primary revenues increased by 5.9 percentage points o f GDP, driven by an increase in indirect taxes o f 3.2 percentage points o f GDP (Table 1.2). By contrast, primary expenditure o f the general government remained almost constant in percent o f GDP. Tal ' 1.2: Fiscal Consolidation in Turkey, 1999-2005 and sub-per ods' (differences between years in % of GDP) 1999-2005 1999-2001 2001-2005 Taxes 3.1 3.9 -0.8 Direct -2.4 0.7 -3.1 Indirect 5.1 3.2 1.9 Wealth 0.4 0.0 0.4 Non-Tax Revenues 0.4 -0.2 0.6 Factor Incomes 2.2 1.5 0.7 Interest revenues 1.1 0.3 0.8 Social Funds 1.6 1.o 0.5 Total Revenues 7.2 6.2 1.o Primary Revenues 6.1 5.9 0.2 Current Expenditures -0.3 -0.2 0.0 o f which personnel -1.7 -0.5 -1.2 Investment Expenditures -0.6 0.3 -0.9 Transfer Expenditures -3.6 11.3 -14.9 Current Transfers -3.0 11.4 -14.4 of which interest payments -5.0 10.8 -15.8 Capital Transfers -0.6 -0.1 -0.6 Primary Expenditures 0.4 0.5 -0.1 Borrowing Requirement -12.5 4.3 -16.8 Primary Surplus 5.7 5.3 0.3 Source: SPO and World Bank Staff estimates 1.13 Starting in 2001 fiscal consolidation was more balanced on the revenue and expenditure sides. In 2001-2005 total primary revenues decreased by 0.8 percentage 'Although tax revenues in percent o f GDP have peaked in 2001, for the whole 1999-2004 period, this partly reflects the sharp drop in GDP in this crisis year. For this same reason, expenditure items appear also artificially inflated in 2001 when measured in percent o f GDP. 35 points which was mainly due to a decrease in the direct tax collection. A 3.1 percentage point decrease in direct taxes was partially off-set by close to 2 percentage points increase in indirect taxes. Primary expenditures, on the other hand, stayed almost constant although there has been a significant reduction in investment expenditures. A reduction in personnel expenditures was also noted in 2005. The fiscal space created in these two expenditure areas was completely offset by the increase in the social security deficit. 1.14 Overall, fiscal consolidation since 1999 has come primarily from the revenue side, with heavy reliance on indirect taxes, although initiatives have been taken to improve the quality and limit the growth of spending. Despite the partial shift in the adjustment pattern since 2002, Turkey’s focus for achieving fiscal consolidation has been o n revenue-increasing measures rather than expenditure rationalization. From 1999 to 2005, almost 100 percent o f the 6.1 percentage points o f GDP increase in the C G G primary surplus came from revenue-increasing measures-especially from higher indirect taxes. Indirect tax revenues increased from 11.7 percent o f GDP in 1999 to 16.7 percent o f GDP in 2005. In contrast to primary revenues, non-interest expenditures reflect no major changes during the period due to increase in social security transfers, despite the downsizing o f investment and reduced expenditures for personnel in 2005. I t should be noted, however, that initiatives have been taken to prevent l o w productivity expenditures, improve spending quality, and set expenditure priorities-for example, annual budget ceilings are set to recruitments in the public sector to better control the wage bill (with priority to education, health care, and justice), while non-priority investment projects have been excluded from the investment program. Although more steps are needed, as further explained below, these measures have supported a better control o f expenditure growth than in the past. B. THE SIZE OF GOVERNMENT AND COMPOSITION OF EXPENDITURES IN INTERNATIONAL COMPARIO S N* 1.15 Turkey’s primary expenditures are still at the low end o f EU countries- Turkey’s public expenditure allocations have been benchmarked against a set o f countries comprising cohesion EU countries, new EU members, and emerging market economies. Reflecting still high interest payments on public debt, total expenditures o f the Consolidated General Government (CGG) are among the highest o f the comparison group (Figure 1.2). However, at 32.7 percent o f GDP in 2004, primary expenditure i s much lower, and compares favorably with the average size o f government seen in cohesion countries o f the EU and new EU members-although Turkey’s primary expenditure in percent o f GDP i s at par with Ireland’s and higher than Romania’s. 1.16 -but much higher than in other emerging economies. Primary expenditure in the comparison group o f emerging economies (excluding Mexico where available data do In this section expenditure data for Turkey refer t o 2004 rather than 2005 as comparable international data could not be collected for more recent years than 2003 or 2004. The changes in expenditure shares between 2004 and 2005 are not significant. 36 not capture properly spending by sub national governments) represented 23 percent o f GDP in 2003, about ten percentage points o f GDP lower than in Turkey. As primary expenditure was only marginally reduced in percent o f GDP between 1999 and 2004, Turkey's relative position remained stable within the benchmark group. I:;. Figure 1.2 : General Government Expenditure as % o f GDP 60 IOPrirnaiy Expenditure Interest 1 Source: World Bank staff calculations B.1. Public expenditure by economic category' 1.17 The Government wage bill is high compared to other emerging economies and some new EU members. At 10 per cent o f GDP in 2004, Turkey's expenditure on wages and salaries at the level o f the General Government is relatively high. I t i s s t i l l lower than in Southern EU countries (Greece, Portugal, Spain) or in some new EU members with a tradition o f bloated public service (Poland, Hungary-see Figure 1.3). However, personnel compensation is significantly higher than in other fast-growing emerging economies (Chile, Korea, Malaysia, Thailand) or in some new and acceding EU countries that have streamlined the public sector (Czech Republic, Bulgaria, Romania). At 30 percent o f total General Government primary expenditure in 2004, up from 28.6 percent in 2000, the wage bill represents a large fraction o f public spending by international 9 This section briefly examines, in international comparison, Turkey's expenditures for personnel compensation, current transfers, and public investment at the CGG level. These three categories comprised the bulk (75.4 percent) o f primary CGG expenditures in 2004, o r the equivalent o f 24.6 percent o f GDP. Remaining primary expenditures (the equivalent o f 8 percent o f GDP) are allocated for purchases o f goods and services (7.3 percent o f GDP) and capital transfers. The bulk o f purchases o f goods and services are allocated to health (3.8 percent o f GDP) and Defense, Public order and safety (1.8 percent o f GDP combined). A detailed cross-classification (i.e., economic and functional) o f CGG expenditures i s provided in Annex I . 37 comparison (Figure 1.4). This imparts rigidity in the budget at a time when more fiscal space i s needed. Figure 1.3: Compensationof General Government employees (2003-04; in % of GDP) ................. ................................................. 14 12 35 10 30 8 25 20 6 15 4 10 5 2 0 n Source: World Bank staff calculations 1.18 Expenditures for current transfers, mainly comprising social benefits, are oversized by international standards, in view o f Turkey's favorable demographics. At 11.7 percent o f GDP in 2004, expenditure on current transfers i s lower than in most EU comparator countries (15 percent and up, with the exception o f Ireland), but higher than emerging market economies such as Chile, Korea and Malaysia (Figure 1.5). Current transfers in their majority (8.5 percent o f GDP) are for social protection, so that spending in this category is better analyzed in the context o f the age distribution o f population. Given that Turkey has a very young population and the lowest old age dependency ratio among a selected group o f comparators, its social protection expenditures are oversized by international comparison (Figure 1.6). In addition, the proportion o f Turkey's population aged 65 and over to the labor force will increase by 10 percentage points between 2000 and 2020, according to OECD projections." Therefore, despite the current favorable demographic situation, in the future, Turkey i s likely to face an upward pressure on social spending in order to expand the social safety net. loOECD Fact book 2005 - ISBN 92-64-01869-7 - 0 OECD 2005. See Section "Population and migration - demographic trends - ageing societies". 38  F i g u r e 1.5: E x p e n d i t u r e by E c o n o m i c C a t e g o r y : C u r r e n t transfers--Turkey (2004) a n d c o m p a r a t o r s (2003), in % o f GDP ---- - - - - *51---- I ____I___-_--_-_ 20 - 15 - - 10 ~ 5 - 0 - Source: World Bank staff calculations Figure 1.6: General Government Expenditureon Social Benefitsand Social Transfers (as X of GDP) M. Old Age DependencyRatio WL . CZE' HUh.I SW _. 0 00 0 05 0 10 0 15 0 20 0 25 0 30 Old Age Depndmsy Ratio Source: World Bank staff calculations 1.19 Turkey's domestic investment level is close to the average o f comparator countries. Despite significant downsizing since 2001, public investment remains comparable to levels seen in other EU members, and higher than in other emerging economies (with the exception o f Malaysia and Korea-see Figure 1.7). Although infrastructure gaps in Turkey may be larger than in cohesion countries or in some new EU members, carefil consideration should be given to increasing the size o f the PIP based o n sound costhenefit assessment o f investment projects. The large portfolio o f frozen projects suggests that the PIP has been oversized in the past, including with projects o f questionable priority. Issues in the management o f the public investment program are reviewed in chapter 2. 39  , Figure 1.7: Comolidated General Govenunent Gross Fixed Capital Fomtion- Turkey (2004) and comparators (2003), in % of GDP 1 1 I l 6 5 4 i2 3 1 I I O ~~ Source: World Bank staff calculations B.2. Public Expenditure by Government function 1.20 Implementation o f functional classijkation o f expenditures is very recent in Turkey. Turkey initiated the GFS 2001 consistent functional classification for the consolidated budget institutions in 2004, and rest o f the general government institutions started implementing the GFS classification with the 2006 budget. Although there have been some partial efforts to estimate the functional classification o f social sector expenditures, an official functional breakdown o f the general government expenditures is not available for 1999-2005 .l The cross functional and economic classification for 2003 and 2004 used in this study i s a first attempt to estimate the general government functional expenditures. Completing a hnctional classification o f expenditures for past years will help analyze more accurately the sectoral impact o f the fiscal adjustment. Turkey's functional distribution o f consolidated general government expenditure has been benchmarked against a set o f comparators for which a similar functional classification is available.12 The main findings are summarized below (also see Figure 1.8) and further detailed in chapter 3 for specific sectors such as education, social protection, and health care: I' General Government social expenditures first initiated under the PEIR 2001 study o f the W o r l d Bank. The coverage has been expanded under the PFPSAL program. Comparators include: Russia, Portugal, Greece, Czech Republic, Spain, Ireland, S. Africa, Romania, Malaysia, Mexico, Thailand, Brazil, and Chile. Turkey's size o f government, as measured by primary expenditure in percent o f GDP (32.6 percent in 2004), was broadly comparable t o the average o f these countries (30 percent o f GDP in 2003). 40  Figure 1.8: Functional Allocation o f General Government Expenditures -Turkey (2004) and Comparators (2003); in % of GDP 12 7 io J 1 ! Turkey I E Comparators 1 a 6 4 2 Gen. Pub. Defense and Economic Enwonment Housing and Health Recreation, Education Social SeMces Public order affairs protechon community culture and protection I and safeiy amenities religion Source: World Bank staff calculations 1.2 1 Expenditure for the provision o f public goods (general public services, defense, public order and safety) is relatively high in international comparison. Spending levels in general public services and in defense and public order and safety are above the average o f comparator countries by 0.8 and 1.2 percentage points o f GDP respectively. 1.22 Expenditure on economic affairs (comprising mainly investment in infrastructure and various subsidies) is broadly in line with average levels in comparator countries. As already mentioned, this mainly reflects drastic downsizing o f the investment budget and continuing reform o f subsidies to agriculture since 1999. 1.23 Expenditure on health care is relatively high in international comparison. At 5.9 percent o f GDP in 2004 (including Hospital Revolving Funds), CGG expenditure on health care i s 2.2 percentage points o f GDP higher than in the average o f comparator countries. Public expenditure on health care depends o f course on many factors, such as the structure o f the health care system, demographics, and per capita GDP. Further analysis in this area i s provided in chapter 3. 1.24 Expenditure on education is in line with comparators but educational attainment remains weak. Turkey spends more on education than other emerging economies with similar educational attainment. Public expenditure on education in percent o f GDP i s similar to other EU countries and comparable to emerging economies, but educational attainment in Turkey is s t i l l lagging behind with respect to these comparators. As in other countries experiencing educational gaps, public expenditure on education would need to increase in the future to promote the closing o f these gaps. This would be a pre-requisite for faster long-term growth (World Bank, 2006 CEM). However, in view o f current levels o f public spending, existing educational gaps also 41 suggest potential for efficiency improvements that should go in tandem with increases in the level o f education expenditures. Options in this direction are reviewed in chapter 3. 1.25 Expenditure on social protection is high by international comparison in view of Turkey’s favorable demographics. A s noted earlier, although Turkey’s expenditure in percent o f GDP i s on par with the average o f comparator countries, favorable demographics suggest considerable room for savings through more ambitious pension reform. 1.26 Expenditure on environmental protection is comparatively low and may need to increase on the way to the EU. Alignment with the EU Acquis on environmental protection will place a burden on public expenditure for control o f air pollution, quality o f water supply, wastewater and waste management. Although considerable uncertainty surrounds the cost estimates o f necessary investments, i t should be noted that costs o f meeting the Acquis requirements will also depend on policy choices and reforms in infrastructure services with environmental externalities, while there i s considerable room for participation o f the private sector. l3 1.27 Downsizing low-value expenditure programs, to make room for growth- enhancing expenditures and lower taxes, would deserve particular attention in the years ahead. Turkey has accomplished a remarkable effort o f fiscal consolidation since 1999 but improving the quality o f fiscal adjustment will be a key challenge o n the way to the EU. This i s important for a number o f reasons: (i) international experience suggests that the quality o f fiscal adjustment i s a key factor for sustainability and i s closely associated with stronger growth performance; l4( ii)expenditure pressures are likely to be felt over the medium term, either as a result o f past policy commitments-such as in education and access to universal health insurance (see chapter 4), or owing to a still pending reform agenda (such as in civil service pay); ( iii)infrastructure investment may need to increase in less developed regions to alleviate persistence o f regional disparities, while maintenance needs to be scaled up (see chapter 2); (iv) Government will have to make room in later stages o f EU accession for expenditures in areas such as environmental protection where harmonization with the EU Acquis i s costly-even though the cost will be, to some extent, offset by EU pre-accession funds; (v) civil service pay for highly qualified personnel may also need to increase to support the EU accession process. 1.28 Trade offs in expenditure allocations would need to be considered to meet Turkey’s development challenges. As public expenditures are already high compared to other emerging economies and some new EU members there i s little room for Turkey to further increase expenditure in order to meet pressing development challenges. l3Markandya, A. 2005. “Turkey on the Path o f to the Accession: The Environmental Acquis” p. 295 - 309 in “Turkey. Economic Reforms & Accession to the European Union“, co publication o f the World Bank and the Centrefor Economic Policy Research l4A. Alesina and R. Peroti, “Fiscal Adjustments in OECD Countries: Composition and Macroeconomic Effects”, IMF Staff Papers, 1997; A. Alesina and S. Ardagna, “Tales o f Fiscal Adjustment”, Economic Policy, 1998; R. Peroti, “Fiscal Consolidation in Europe: Composition Matters”, American Economic Review, 1996. 42 Expenditure should also be contained in order to make room for lower taxes in the long run while preserving a sound fiscal framework. Policy would thus need to focus o n trade- offs in expenditure allocations, possibly by reducing spending in functional areas where it appears to be oversized in international comparison (such as general public services and defense, public order and safety). At the same time, reforms should be implemented with the aim o f improving the efficiency o f expenditure programs in areas where expenditure pressures are being felt-such as health care, education, social protection (see chapter 3). Horizontal reforms, focused on the modernization o f c i v i l service pay and employment system, and the rationalization o f the investment program would also help contain expenditure pressures o n the wage bill and investment spending and thus better control public expenditure across hnctional areas. Incorporation o f tax expenditures into the budget allocation process would also support their rationalization, help avoid duplication with direct expenditure programs, and support tax base broadening. Reform options in these horizontal areas are reviewed in chapter 2. C. THE REVENUE SIDE 1.29 As a result o f revenue-driven fiscal adjustment, general government revenues have increased to levels comparable to EU countries and much higher than in emerging market economies. I n most EU cohesion countries and new EU members the fiscal burden has remained constant, or has even decreased in percentage o f GDP, as a result o f initiatives to contain the increase in the size o f Government (Figure 1.9). But contrary to this trend, and as a result o f the pattern o f fiscal adjustment since 1999, general government revenues in Turkey stood at 40 percent o f GDP in 2005, a level that ~ increase in the fiscal burden in 1999-2005 was the i s now similar to EU c ~ u n t r i e s . 'The largest among the comparator countries-similar only to Brazil. Flours 1.9: General Government Revenue in % o f GDP VI per capita GDP In PPP 60 , , IRE 9 P M E X I lo i , 0 0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000 I GDP per capita, PPP Source: World Bank staff calculations ~ 15 General government revenues include tax revenues, social security contributions, factor income (mainly revenues f r o m SOE's), and interest income. 43 1.30 The tax effort in Turkey is largely above the average o f comparator countries. Tax revenues (including social security contributions) have increased to 3 1.7 percent o f GDP in 2004, from 27.1 percent in 1999 and 19.7 percent in 1995. Whether or not this fiscal performance is above the expected norm can be judged by accounting for factors that affect tax performance across countries. A formal analysis o f tax performance across countries in the period 1995-2004 confirms that differences in par capita income, the share o f agriculture in GDP, and openness to trade, are consistent factors accounting for international variations in tax ratios (see Annex IV). Accounting for such determinants o f tax performance, the estimated “tax effort ratio” in Turkey was 1.3 in 2004, indicating an actual tax ratio o f about 30 percent higher than expected based on international experience.I6 Turkey’s tax effort was one o f the highest among comparators, mainly OECD countries and emerging economies, equal to that o f Moldova and surpassed only by Croatia (Figure 1-10). 1.3 1 The tax effort has been on the rise, with recently opposing trends between direct and indirect taxes. The tax effort ratio was less than potential in 1995, at an estimated level o f 0.9 (Figure 1.10). I t has steadily increased since then and reached its peak in 2001, initially boosted by above-potential direct tax collections and, later, by a strong indirect tax effort (Figure 1.1 1). The overall tax effort has stabilized since 2002 at around 1.3, on the back o f a slightly lower direct tax effort while, at the same time, the indirect tax effort has kept increasing reflecting the introduction o f special consumption taxes and a heavy reliance on excises. 1 Figure 1.10: Tax effort indicators (1995 and 2004) l6Tax effort i s measured by comparing the actual tax ratio o f a country with that predicted by using a panel regression. An index o f one means the country’s tax effort is at the expected level, given the structural factors o f the country. In that case the country i s using its taxable capacity at level consistent with the average o f the other countries in the sample. See annex I V for the calculations. A tax effort ratio above one potentially indicates excessive taxation, which m a y have a negative incidence o n efficiency and growth for many reasons (capital is mobile; labor and sales could go to the informal market, etc). 44  Figure 1.11: Tax effort ratios in Turkey (1995-2004) 0.8 1 ..- . 0.6 - -owall -dir - - - - .- . indir 0.4 - 0.2 - 0 , , I , I 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 Source: World Bank staff calculations 1.32 High actual tax ratios and estimated tax effort suggest limited room for further revenue mobilization in the coming years. The high estimated tax effort ratios should not be interpreted as indicating limited room for broadening tax bases, strengthening tax compliance, and improving tax administration-all priority areas o f fiscal policy in Turkey, especially in view o f the extensive informal economy. However, in most comparator countries the tax effort ratio is close to o n e - o n average, an estimated 0.9 in the 8 new EU members in 2004. The high tax effort ratio tends thus to suggest that the tax burden in Turkey i s overstretched, most likely reflecting high and multiple tax rates, given Turkey’s level o f development and economic structure. There i s thus very limited room for further revenue mobilization through higher tax rates or new taxes in the years ahead. With a fiscal burden higher than in other emerging economies, further increases in tax rates would risk hindering competitiveness, bolstering informality, and eventually compromising growth. There is, however, considerable room for broadening the tax bases, as explained in World Bank 2006 (CEM) and in chapter 2 (see section on tax expenditures). Tax base broadening would create room for lowering the tax rates in the future. 1.33 The composition o f revenues has shifted markedly toward indirect taxation over the last few years. Hikes in excise rates along with buoyant VAT collections in the post- crisis years combined to increase indirect tax revenues from 11.6 percent o f GNP in 1999 to 16.7 percent o f GNP in 2005. During the same period, personal income tax (PIT) revenue f e l l from 6.5 percent o f GNP in 2001 to only 4.9 percent GNP in 2005, while corporate income tax (CIT) revenue remained broadly stable. As a result o f these trends, Turkey is one o f the countries with the heaviest reliance on indirect taxes among OECD countries, with indirect taxes representing 52 percent o f tax revenues, against 36.5 percent on average in the OECD (Table 1.3). 45 Table 1.3. Revenue h e a k d o w n for selected OECD Countries 2003 General Government Taxes Income Taxes PIT CIT VAT Excises ImpSS (in percent of GDP) & ss Tax Duty Austria 43.0 28.3 12.9 10.4 2.3 9.5 2.7 0.0 14.7 Belgium 45.2 30.6 16.5 13.4 3.0 12.2 2.2 0.3 14.6 Canada 33.6 28.6 15.6 ... ... 11.8 0.3 0.2 4.9 Czech Republic 36.5 21.3 9.6 4.9 4.7 8.4 2.5 1.3 15.2 Denmark 48.7 47.1 28.8 26.0 2.8 17.8 4.0 ... 1.7 Finland 44.8 32.6 17.5 14.0 3.5 14.0 4.3 0.0 12.2 France 43.4 26.8 10.3 8.1 2.2 14.2 2.5 0.0 16.6 Germany 40.3 22.8 10.3 ... ... 7.7 2.6 0.7 17.5 Greece 36.1 23.2 8.3 4.9 3.3 11.9 3.2 0.0 12.9 Hungary 39.1 26.4 9.3 7.1 2.2 13.2 3.3 1.1 12.7 Ireland 29.6 25.0 11.9 7.0 3.8 9.1 2.0 1.4 4.6 Italy 42.6 29.8 13.2 10.8 2.3 11.7 2.5 0.0 12.9 Netherlands 38.7 24.2 10.1 6.9 3.2 10.7 1.3 1.3 14.5 Norway 43.4 33.4 19.3 11.2 4.3 14.4 1.7 0.1 9.9 Poland 36.6 22.5 6.3 4.1 2.2 12.8 3.9 1.0 14.1 Portugal 36.9 25.1 9.1 5.9 3.2 14.8 3.3 0.2 11.8 Slovak Republic 31.2 18.8 7.0 3.4 2.8 9.3 ... 1.4 12.4 Spain 35.8 23.0 10.2 6.9 3.3 10.3 2.6 0.0 12.8 Sweden 50.2 35.8 18.1 16.0 2.1 17.1 3.3 0.0 14.4 United K i n g d o m 36.1 28.9 13.2 10.4 2.8 10.5 3.5 0.0 7.3 United States 25.5 18.5 10.8 8.9 2.0 10.0 1.6 0.2 7.0 Turkey (2004) 30.3 23.3 7.3 5.0 2.3 8.0 7.7 0.3 7.0 OECD average: 39.1 27.4 12.9 9.5 2.9 11.8 2.5 0.7 11.7 Source: OECD and World Bank staff calculations 1.34 High excises, narrow tax bases, and high labor income taxes are main features of the tax system. While international comparisons o f tax ratios need to be interpreted with care, three features o f the tax structure in Turkey are especially noteworthy: 0 A very heavy reliance on excises, which account for nearly one-third o f all tax receipts and raise at least twice as much, relative to GDP, as in any other country in the ~ a m p 1 e . lHeavy ~ reliance on excises i s o f concern as i t may encourage activity in the informal sector, which would hrther undermine the tax bases and impede efficiency for the economy as a whole. Reducing this high level o f excise taxation does not appear, however, to be seen as a priority: the 2005 budget included several excise increases in order to attain the primary surplus target while meeting spending needs.’* *’This appears to mainly reflect high rates o n ‘traditional’ excisable goods: around h a l f i s f r o m petroleum products alone-indeed, at about $8/gallon, pump prices for gasoline are about the highest in Europe. In addition, the taxes o n “luxury goods” raise revenue o f about 0.1 percent o f GDP. ’* O f course excises need t o be raised t o ensure that negative externalities are fully offset. While a systematic international comparison o f specific excises would be appropriate, cursory evidence suggests that excises in Turkey are comparatively high. F o r example, super gasoline prices in 2005 were 20 percent higher than in Poland and 50 percent higher than in Romania. In the 2006 budget, while total tax revenues 46  Low receipts from, and reliance on, income taxation. Although the P I T yield in percent o f GDP is significantly lower than the OECD average, the top marginal rate was a substantial 40 percent in 2005 for non-wage earners (planned to be reduced to 35 percent in 2006). One reason for this pattern m a y be the existence o f a large informal sector. Similarly, Corporate Income Tax (CIT) revenues are somewhat lower than the OECD average, despite the fact that the C I T rate in 2004 was a substantial 33 percent (though reduced in 2005 to 30 percent).” This indicates that the base o f the tax is narrow, owing to substantial tax expenditures as further discussed below. A high tax wedge on labor income. When the PIT, employer and employee social security contributions, and the stamp duty o n wages are combined, total payments to the government out o f labor income can approach 43 percent o f the labor cost for some workers (see World Bank 2006). Indeed, according to the OECD’s Taxing Wages study, Turkey emerges as the country with the highest labor tax wedge for the married “average production worker” with a non-working spouse and two children. High labor income taxation impedes employment generation, particularly in the formal labor market. 1.35 Reliance on indirect taxation may have a regressive incidence on income distribution but offers some advantages that should be taken into account in the design of the tax system. Reliance o n indirect taxes is often criticized on equity grounds, as the tax burden is indiscriminately distributed across all income groups. Nevertheless, given pervasive informality, in the short term, reliance on indirect taxes appears to be a reasonable way o f ensuring some degree o f compliance o f those who would have otherwise escaped taxes altogether. Moreover, indirect taxes offer a more symmetrical treatment o f labor, transfer, and capital income, thus meeting some criteria for horizontal taxpayer equity better than income taxes that discriminate against some types o f income or are subject to considerable evasion. These factors tend to alleviate the distributional consequences o f indirect taxes. In addition, indirect taxes uniformly levied on consumption offer advantages o n efficiency grounds: (i) they are relatively neutral towards saving and investment decisions; (ii) they distort relatively less incentives to work. Hence, all else equal, a tax m i x relying more o n indirect taxes may be more conducive to growth. Caveats apply, however, to these considerations, if greater reliance o n indirect taxes i s accompanied by proliferation o f special regimes and exemptions, which may boost tax evasion and feed informality, thus being counterproductive for growth (see chapter 2, section D for more information about tax expenditures related to the VAT regime in Turkey). 1.36 Reducing the burden o f indirect taxes would call for careful study and long- term design. Although the balance between direct and indirect taxation in Turkey i s considerably out o f l i n e with that in comparator countries, options are limited in the short to medium term because revenue cuts in one area must-given the continued are estimated t o increase by 11 percent, the increase in SCT o n Petroleum and Natural Gas i s lower, estimated at 3 percent. 19 Further reduction in the CIT rate t o 20% has been initiated starting f r o m January 2006. 47 macroeconomic need to achieve large primary surpluses-be offset by gains in another. If excises and VAT collections are cut, then income taxes will likely need to be raised. However, such a substitution o f direct for indirect taxes would not be advisable in view o f the above-mentioned considerations regarding the efficiency benefits o f indirect taxes and their equity-improving incidence when significant informality i s present, as in the case o f Turkey. Appropriate fiscal space for lower indirect taxation will thus have to be created over the long term. Fiscal space could come from three main sources: (i) Continuous fiscal discipline to reduce the level o f the debt ratio and create room for a lower primary surplus in percent o f GDP; (ii) better control o f primary expenditures in low-value programs; (iii) tax base broadening and reduction in the size o f the informal sector, The analysis that follows presents options for income tax reform with the aim o f improving efficiency and broadening the tax base. Eventually, these initiatives will also facilitate a shift in the balance o f the tax burden between direct and indirect taxation. 1.37 The tax landscape is changing fast- The authorities are currently in the process o f reviewing and rewriting each o f their major tax laws, also following commitments taken in the last Stand-By Arrangement with the IMF. The authorities eliminated stamp duties o n financial transactions in 2004, and in April 2005 announced their intention to eliminate the two remaining financial intermediation taxes-the Banking and Insurance Transaction Tax (BITT) in 2006 and Resource Utilization Support Fund (RUSF)-if conditions permitted, within the program period. A new regime for capital income taxation was to be introduced as from 2006, subjecting interest and capital gains, including o n previously exempt government securities, to final withholding at 15 percent. This would move Turkey closer to an explicit “dual income tax” (DIT) model o f the sort employed in the Nordic countries. Labor would remain subject to progressive taxation, while capital income would be subject to a flat and relatively modest rate-in recognition o f the futility o f trying to tax a highly mobile base l i k e capital. 1.38 -but sometimes without sufficient attention to tax neutrality and tax equity considerations. Further to the recent financial market turmoil, in an attempt to improve the attractiveness o f domestic Government bonds to investors, the authorities waved the 15 percent withholding tax for non-resident investors, while for domestic investors the rate was lowered to 10 percent. For domestic investors, the 15 percent withholding tax o n deposits and repos will remain in place. The tax exemption for all derivative instruments will remain in place indefinitely. These initiatives will undo the simplification and elimination o f asset-specific distortions that the withholding tax regime was supposed t o achieve. A s for offering different tax treatment to non-residents, despite the temporary boost to the attractiveness o f government bonds, this will also create tax loopholes and incentives for residents to misreport their investments, thus putting the domestic tax base at risk. A better option, perhaps over the longer run, would be to apply a common tax across investors and then work o n building a broader array o f tax treaties that would carve out particular exceptions for non-residents. 1.39 The 2006 budget and drap legislation to amend the existing PIT and CIT laws initiated significant reforms o f the Corporate Income Tax and the Personal Income Tax. They include: (i) a reduction o f the C I T rate from 30 to 20 percent, with a parallel 48 elimination o f the Investment Tax Allowance (ITA); and, (ii) a reduction o f the top personal income tax rate from 40 to 35 percent for non-wage earners, accompanied by a reduction in the number o f tax brackets from 5 to 4 and a unification o f the tax regimes for wage and non-wage income.20 A s a result o f tax base broadening, due to the elimination o f the ITA, ongoing tax administration reforms, and broader strengthening o f the C I T code (regarding, for example, transfer pricing and thin capitalization),21 the Government does not expect a significant impact o f these measures o n C I T collections, and stands ready to take offsetting expenditure measures if necessary to keep the primary surplus o n target. The C I T tax reforms have the potential to promote investment o n a sound basis and stimulate j o b creation in the formal sector. The P I T reform will improve the efficiency o f the tax system, as the dual tax system for wages and n o n wage income created complication and incentives for income recharacterization, and was unduly generous to highly paid wage earners. Unifying the rate schedules is consistent with movement towards a Dual Income Tax system. The C I T legislation has been put into force, while only the P I T legislation concerning the tax brackets and tax rates has been implemented. 1.40 Reduced VAT rates introduced recently are o f questionable effectiveness and not supportive of base broadening efforts. As part o f its tax initiatives the Government announced in March 2006 reduced VAT rates o n textiles, clothing, and some leather products. These measures are intended to help domestic producers face competition from cheaper imports and also to provide incentives for better formalization in these sectors where informal activity is particularly important.22 However, the instrument used i s not appropriate for the goal pursued, as reduced V A T rates on textiles and clothing, if anything, will boost consumption and imports and thus intensify competition faced by domestic producers. In addition, multiple VAT rates and exemptions (see chapter 2, section D) increase the complexity o f the tax system, create more opportunities for misreporting and misclassification for tax purposes, and thus work against efforts o f strengthening tax compliance. 1.41 Initiatives in the years ahead should focus on tax-base broadening. T w o goals should be pursued with appropriate prioritization: (i) In the short term, promoting tax rationalization, with the aim o f reducing the distortions associated with a given level o f the tax burden and broadening the tax bases; (ii) Over the medium term, making fiscal room for lower taxes, by taking advantage o f the reduction in the debt ratio and pursuing public expenditure reforms aimed at containing expenditure pressures. 2o By the new reform, it i s also planned t o reduce the lowest tax rate for income components other than wages f r o m 20 percent to 15. The highest and the lowest tax rates o n wages will be maintained at 35 f,ercent and 15 percent respectively, thus effectively unifying the two tax schedules. “Thin capitalization” occurs when multinational firms are able to s h i f t profit across jurisdictions for tax purposes by increasing the amount o f debt financing in jurisdictions with high tax rates in order to benefit f r o m the deductibility o f interest f r o m the corporate tax base. 22 Authorities argue that the rate reduction will also tame incentives t o use forged documents t o get fictitious rebates in VAT refund system, thus helping improve compliance in the short term. 49 D. STRUCTURAL AND CYCLICAL COMPONENTS OF THE BUDGET BALANCE 1.42 Structural budget balances convey useful information about the fiscal stance prevailing in an economy and the extent offiscal adjustment. Government budgets can significantly be affected from growth cycles. For example, during economic contractions, there will be a reduction in tax revenues and an increase in unemployment benefits and other social protection expenditures. By contrast, during economic expansions, there will be an increase in tax revenues and a decrease in unemployment benefits and related expenditures. Therefore, an analysis aiming at measuring fiscal stance and the extent o f fiscal adjustment has to distinguish between structural and cyclical parts o f the budget. 1.43 Changes in the structural deficit can contribute to the assessment o f the sustainability of fiscal adjustment in the medium term by identibing the impact o f the growth cycle on the budget. Short-term improvements in the budget, due to robust economic growth, m a y be reversed when growth slows down and should therefore not be seen as an underlying improvement in public finances. Relaxing fiscal discipline by interpreting a cyclical improvement in budget balance as permanent at a time when the economy expands may create unsustainable budget deficit and debt stock in the medium term. In a similar way, to interpret a cyclical increase in the budget deficit as permanent when the economy contracts and taking measures to reduce the deficit may exacerbate the slow down. 1.44 The sustainability o f Turkey’s strong fiscal performance depends on how much of the achievement was due to policy-induced changes as opposed to cyclical improvements stemming from robust growth. I t would also depend o n whether the policy-induced changes are sustainable themselves over the medium term. The lack o f major restraint in non-interest expenditures can raise questions as to the sustainability o f the adjustment as discussed earlier. This section aims to address only the first issue above related to the cyclicality o f the observed fiscal adjustment. 1.45 The basic methodologyfor estimating structural budget balance mainly consists of two steps; measuring the potential output and linking the related budget items, revenues and expenditures, to the cycle. The methodology used in this study and its alternatives are discussed in detail in the Annex 1 11. Because long time series data is not available at general government level, estimations relied o n consolidated budget figures. Estimated tax elasticities were then used to adjust tax revenues for the cycles in economic activity. These adjusted tax revenues were then used to calculate structural budget balances both at the consolidated budget level and general government On the expenditure side, n o adjustments were made for the impact o f cycles in economic activity. The main budget expenditure item sensitive to the cycle i s unemployment 23 The consolidated budget accounts for more than 85 percent o f the total tax revenues at the general government level. Therefore using the elasticities estimated for the consolidated budget also for the general government is a reasonable assumption. 50 insurance. I t should be noted that unemployment insurance i s a recent development in Turkey with very little insurance payments to unemployed so far and is also out o f the coverage o f the consolidated budget. 1.46 Tax elasticities in Turkey are close to the OECD average. Table 1.4 compares our elasticity estimates for Personal Income Tax (PIT) and Corporate Income Tax (CIT) in Turkey with existing estimates for OECD countries. Following standard practice, the elasticity o f indirect taxes i s taken as unity. Indirect taxes are assumed to increase one to one with the income growth, as they are not expected to be progressive or regressive. Therefore, income elasticities have been estimated only for PIT and C I T (see Annex 1 11. ) The estimated elasticities are close to the OECD averages. The fact that the elasticities are greater than one reflects the progressivity o f taxes in Turkey, as in most OECD countries. Table 1.4: Output Elasticity o f Taxes (l) Income Tax Corporate tax United States 1.3 1.5 Japan 1.2 1.6 Germany 1.6 1.5 France 1.2 1.6 Italy 1.8 1.I United Kingdom 1.2 1.7 Canada 1.I 1.5 Australia 1.o 1.4 Austria 1.3 1.7 Belgium 1.I 1.6 Denmark 1.o 1.6 Finland 0.9 1.6 Greece 1.7 1.I Ireland 1.4 1.3 Netherlands 1.7 1.5 New Zealand 0.9 1.4 Norway 1.o 1.4 Portugal 1.5 1.2 Spain 1.9 1.2 Sweden 0.9 1.8 Average 1.3 1.5 Turkey (*) 1.5 1.6 (1): Elasticities for all countnes except Turkey are taken from the study “Girouard, N. and C. AndrC (2005), Measuring Cyclically-AdjustedBudget Balances for OECD Countries, Economics Department Working Papers, No. 434. (2): World Bank staff estimations 51 1.47 Estimated structural primary bud et balances indicate a strong consolidated budget performance from 2000 0nwards.9~ Using the elasticities given in Table 1.4, structural budget balances were first calculated for the consolidated budget. The structural primary budget balances were close to zero or in deficit o f 1-2 percent o f potential GDP in the early 1990s (Figure 1.12). Although budget performance improved after the crisis in 1994, this did not last long as the government eased fiscal discipline soon after a short-term recovery from the crisis. At the end o f the 1990s, the fiscal imbalances were unsustainable and therefore a substantial fiscal adjustment was a necessity. The government initiated a significant fiscal adjustment in 1998 and this adjustment was hrther strengthened after the crisis in 2001. The structural primary budget surplus for the consolidated budget was consistently above 5 percent during this period with a peak in 2001 where the fiscal measures taken led to very high budget surplus in proportion to GNP. Similar conclusions follow from the analysis o f changes in the structural overall balance o f the consolidated budget (Figure 1.13). Figure: 1.12. Actual and StructuralPrimary Budget Balances for Consolidated Budget - 1.30 T - 1.25 - 1.20 - 1.15 - 1.10 2 - 1.05 * . - 1.00 - 0.95 - 0.90 0Structural PBB -Actual PBB - YIYP I Note: Structural budget balances are percent o f potential GNP, while actual budget balances are percent o f actual GNP. Source: World Bank staff calculations 24 Note: Privatization revenues are excluded from the calculations as they represent one-off asset sale revenues and are n o t steady revenue sources. 52 Figure: 1.13. Actual and StructuralBudget Balances for Consolidated Budget I 1.18 18'0 9.0 1.09 * Ei n 00 1.00 2 t 2 a x -9.0 0.91 -18 0 0.82 1 0 Structural BB -Actual BB - YIYD 1 ote: Structural budget balances are percent o f potential GNP, while actual budget balances are percent o f actual GNP. Source: World Bank staff calculations 1.48 A significant part of the large primary fiscal surplus in 2004 and 2005 reflects a contribution from strong growth. The estimations show that the negative cyclical budget balances due to output being lower than potential vanished and the cyclical component o f the consolidated budget balance was significantly positive, up to 1.9 percent in 2005 (Figure 1.14). The growth rate was at historic highs and, as suggested by the estimations, the output was above potential both in 2004 and 2005. The contribution o f the cycle to the performance o f the primary budget balance has reached substantial levels as o f 2005.25 This indicates that the current fiscal performance can be significantly undermined in case o f a slowdown in growth. 25 This is so despite the fact that automatic stabilizers in Turkey are rather weak compared to other countries. This reflects the large share o f agriculture (which does not p a y taxes) and the informal economy, and also the large share o f indirect taxes w h i c h are less cyclical compared t o direct taxes. 53 Figure: 1.14 Cyclical Budget Balance for Consolidated Budget 5.0 1.15 4.0 1.10 3.0 2.0 1.05 * 1.0 : E 2 0.0 1.00 n 2 x -1.0 0.95 -2.0 -3.0 0.90 -4.0 -5.0 0.85 -Cyclical BB -YIP Source: World Bank staff calculations 1.49 Changes in the structural budget balance are a useful indicator o f whether fiscalpolicy dampens or amplifies the cycle. Table 1.5 compares the changes in out put gap and structural fiscal balances from the previous year. A positive association o f the changes in the structural fiscal balances with the changes in the output gap implies a counter-cyclical fiscal stance: I t shows that fiscal adjustments are taking o f f steam from an overheating economy or, inversely, supporting a flagging economic activity. By contrast, a negative association implies a pro-cyclical fiscal stance with a magnifying effect on the cycle. 1.50 Measured at the level o f the consolidated budget, the fiscal stance has often turnedpro-cyclical. The analysis o f the changes in the output gap and structural fiscal balances in the Table 1.5 shows that the fiscal stance was mostly pro-cyclical before 2001 and this was particularly the case in the first h a l f o f 1990s. In this context, i t can be claimed that pro-cyclical fiscal stimulus has exacerbated the growth cycles. Although the pattern seemed to be changing in 2002-2003, with a positive correlation as the government maintained a surplus throughout the initial years o f the recovery, the weakening in the structural primary balances in 2004-2005 indicates that the fiscal stance has become pro-cyclical again. 54 Table 1.5: Fiscal Stance and the Output Gap'') (Consolidated Budget) Change from the Previous Year (Percentage points) Output gap Structural Primary Structural Budget (actual/potential) Budget Balance Balance (% o f Potential GNP) (% o f Potential GNP) 1990 4.4 -0.4 -0.5 1991 -3.9 -1.4 -1.5 1992 2.3 0.4 0.5 1993 3.7 -0.9 -3.3 1994 -9.8 6.0 4.8 1995 3.8 -0.9 -0.9 1996 3.1 -2.1 -5.1 1997 4.2 -2.5 -0.5 1998 0.4 4.3 0.3 1999 -8.8 -1 .I -2.2 2000 3.1 2.7 -0.3 2001 -11.5 3.2 -1.2 2002 5.1 -3.0 0.0 2003 3.3 0.4 2.2 2004 5.8 -0.7 1.7 2005 3.4 -0.6 2.9 (1): Bold figures in the table shows pro-cyclical fiscal policy periods while unbolded ones show counter-cyclical fiscal policy periods. Source: World Bank staff calculations 1.5 1 The estimated structural primary budget balance confirms that fiscal performance was also strong at the general government Structural budget balances for the general government have also been calculated using the same tax elasticities estimated for the consolidated budget. However, in the case o f general government, social security contributions are also adjusted for the cycle. The estimated elasticity o f 0.82 indicates that social security contributions are regressive. Social security contributions are usually levied at a flat rate up to a ceiling and this makes them moderately regressive. The elasticity o f 0.82 is close to the average o f OECD countries (Table 1.6). The estimated structural primary balance o f the general government was, on average, about 6 percent for the period o f 2000-2005 (Figure 1.15). The overall structural deficit o f the general government has declined substantially after 2001 with the help o f high primary surpluses and declining real interest rates. The structural deficit declined from a peak o f 14 percent in 2001 to about 6 percent in 2005 (Figure 1.16). 26 Note: Privatization revenues are excluded from the calculations as they represent one-off asset sale revenues and are not steady revenue sources. 55 Table 1.6: Output Elasticity o f Social Security Contributions (l) Social Securitv Contributions United States 0.6 Japan 0.5 Germany 0.5 France 0.8 Italy 0.9 United Kingdom 0.9 Canada 0.6 Australia Austria 0.6 Belgium 0.8 Denmark 0.7 Finland 0.6 Greece 0.8 Ireland 0.9 Netherlands 0.6 New Zealand Norway 0.8 Portugal 0.9 Spain 0.7 Sweden 0.7 Average 0.7 Turkey (*) 0.8 (1): Elasticities for all countries except Turkey are taken from the study “Girouard, N. and C. AndrC (2005), Measuring Cyclically-Adjusted Budget Balances for OECD Countries, Economics Department Working Papers, No. 434. (2): World Bank staff estimation 1.52 However, measured at the General Government level, the fiscal stance has not been uniform since 2000 and has turned somewhat pro-cyclical more recently. Despite strong performance over those years, the structural primary surplus increased during 1999-2003, but has declined by 1.8 percentage points o f potential GDP since then. Similar to the case o f the consolidated budget, Table 1.7 compares the annual changes in output gap and structural fiscal balances at the general government level. The analysis confirms that fiscal policy turned pro-cyclical in 2001-2002 and in 2004-2005, with a negative association o f the changes in the structural fiscal balances with changes in the output gap. 56 Figure: 1.15. Actual and Structural Primary Budget Balances for General Government I 10.0 T r 1.30 1.25 7.5 1.20 50 1.15 u c 1.10 5 a 8 8 2.5 1.05 : 00 1.oo 0.95 -2 5 0.90 -5.0 1 L 0.85 m 0 N 0 m d 0 vi 0 0 0 0 Q' 0 N 0 N 0 N N 0 N 0 N 0Structural PBB -Actual PBB -y/yp Note: Structural budget balances are percent o f potential GNP, while actual budget balances are percent o f actual GNP. Source: World Bank staff calculations Figure: 1.16. Actual and Structural Budget Balances for General Government [ 20.0 1.18 10.0 1.09 u n 2 e, 0.0 1.00 ti A a -10.0 0 91 -20.0 0 0 0 N - 0 N 0 N 0 N , rn 0 N 0 w 0 0 N vi 0 0 N 0.82 0 1 Structural BB -Actual BB - YIYP I 'ote: Structural budget balances are percent o f potential GNP, while actual budget balances are percent o f actual GNP. Source: World Bank staff calculations 1.53 The pro-cyclicality o f the fiscal stance during the first period is a reflection of the adjustment to the crisis. In 2001-2002, pro-cyclicality reflected the necessary fiscal adjustment to the crisis, despite a sharp contraction in economic activity. However, as many o f the one-off, revenue raising measures vanished in 2002, structural budget balances deteriorated although growth was recovering. This pattern o f fiscal adjustment contributed to the pro-cyclical stance o f fiscal policy in 2002. 57 1.54 Since 2004, the decline in the structural budget balances indicates an easing in the fiscal stance attributable to fiscal imbalances of a more fundamental nature. This weakening in the structural primary balances mainly reflects the decline in the structural primary balances o f the consolidated budget, social security institutions and local administrations. The structural primary balances o f the consolidated budget declined by 0.7 percent o f the potential GNP in 2004-2005.27 The structural primary balances o f the social security institutions weakened by 1.2 percent o f potential GNP in the period 2004- 2005 and by 0.7 percent o f potential GNP in 2005 alone. The long-awaited amnesty for premium collections must have contributed to this outcome. I t i s expected that the recently enacted social security reform will slowly contribute to the reversal o f this trend over the coming years (see chapter 3). Similarly, the structural primary balances o f the local administrations f e l l by 0.6 percent o f potential GNP in 2005 but the overall decline over 2004-2005 was only 0.4 percent o f potential GNP due to a small improvement in 2004. The pro-cyclicality in the last period i s likely to have had a magnifying effect on the growth cycle. Table 1.7: Fiscal Stance and the Output Gap‘’) (General Government) Change from the Previous Year (Percentage points) Output gap Structural Primary Structural Budget (actual/potential) Budget Balance Balance (% o f Potential GNP) (% o f Potential GNP) 2000 3.1 3.7 0.8 2001 -11.5 5.0 -0.9 2002 5.1 -3.7 2.0 2003 3.3 0.6 1.3 2004 5.8 -0.9 1.7 2005 3.4 -0.9 2.8 (1): B o l d figures in the table shows pro-cyclical fiscal policy periods while unbolded ones show counter-cyclical fiscal policy periods. Source: World Bank staff calculations 1.55 Thefindings are robust to alternative modeling assumptions. The impact o f the growth cycle on the structural balances has also been tested by a comparative analysis due to its importance. The results do not change much with respect to changes in the estimated elasticities or to the calculation method o f potential output. The cyclical part o f the general government budget was recalculated by setting all tax elasticities to unity in order to see if higher tax elasticities overstate the impact o f the cycle and by estimating potential output with a different methodology in order to test robustness o f the results with respect to the estimation o f the cycle.28 The results do not change significantly and even under the most conservative assumptions the estimated cyclical part is substantial in 2004 and 2005 (Table 1.8). 27The consolidated budget figures here are defined within a context o f general government, which requires some adjustments in the data. Therefore the figures presented here may differ from the ones given in the previous section. To estimate potential output, the actual output time series was “filtered” to remove cyclical fluctuations using a Hodrick-Presscott filter. 58 Table 1.8: General Government Cyclical Budget in % o f GDP 1999 2000 2001 2002 2003 2004 2005 Potential output i s estimated by production function C clical Bud et aJT= 1.57 a'= 1S O assc=O. " 82 -0.8 0.2 -4.0 -1.8 -0.8 1.1 2.2 C clical Budget aJT=1.0 aP'T=1.0 aSSC=1.0 -0.7 0.2 -3.5 -1.6 -0.7 1.0 2.1 Potential output i s estimated by alternative method (Hodrick-Presscott filter) C clical Bud et azIT=1.57 ap"=1SO aSSC=0.82 -0.5 0.6 -3.8 -2.0 -1.3 0.8 2.2 C clical Budget aJT=1.0 aP'T=I,O aSSC=1.0 -0.4 0.5 -3.3 -1.8 -1.1 0.7 2.1 Source: World Bank staff calculations 1.56 Moving forward, the observed robust growth rates may weaken, together with the positive contribution of the cyclical component, triggering a decline in the primary surplus. The widening current account deficit, higher interest rates as a result o f higher than expected inflation and exchange rate depreciation, can trigger a slowdown in growth. The operation o f automatic fiscal stabilizers would limit the size andor duration o f any growth slowdown, so that their impact o n the primary fiscal surplus would normally be welcome. However, even under a soft landing scenario, a declining primary surplus may have secondary unwanted effects o n the economy, ranging from a deterioration o f debt dynamics, to even higher real interest rates, and a non-negligible effect o n market sentiment. 1.57 To avoid a pro-cyclical fiscal stance, the structural primary surplus should be maintained unchanged during years o f robust growth. This would call for saving fiscal revenue over performance due to robust growth-a policy that would be reflected in an increasing actual primary surplus in proportion to GDP, owing to the cyclically buoyant fiscal revenues. Maintaining a constant structural primary surplus would provide insurance that the reduction o f the debt ratio would continue even in case o f a temporary growth slowdown. An alternative option would be to introduce an expenditure cap as a complementary policy goal if the actual (as opposed to the structural) primary fiscal surplus were to continue to be targeted. Under this option, robust fiscal revenues in years o f strong growth would be reflected in a higher actual primary fiscal surplus rather than higher expenditure, which would work counter-cyclically. By contrast, the primary fiscal surplus target would s t i l l have to be preserved in cases o f slower growth, thus making the hnctioning o f fiscal stabilizers "asymmetric". In view o f the positive output gap, the decline in the structural primary surplus in 2004 and 2005, the widening deficit o f the external current account, and the need to ensure that the debt ratio continues to fall despite higher interest rates, some withdrawal o f fiscal stimulus would seem appropriate in the 2007 budget. 59  CHAPTER I 1 OPTIONS FOR EFFICIENCY GAINS IN PUBLIC EXPENDITURES-HORIZONTAL ISSUES 2.1 To maximize the efficiency o f expenditure programs, reforms in key functional areas should be underpinned by horizontal public expenditure reforms. The natural focus o f public expenditure reforms is to improve the efficiency-for a given c o s t - o f expenditure programs in priority policy areas or in sectors where pressures are being felt. However, these sector-specific reforms will have a better pay-off when combined with “horizontal” (or cross- cutting) reforms that improve the efficiency o f expenditure programs across sectors. Horizontal reforms include initiatives that improve, for example, incentives in the c i v i l service towards the achievement o f results, or help contain pressures o n the wage bill, so that a larger share o f expenditures can be spent on materials and equipment necessary for the provision o f high-quality public services. These reforms can obviously help achieve superior outcomes in key sectors that absorb large amounts o f public financial resources, such as education or health care. Adoption o f efficient project evaluation methods, multi-year programming, and adequate provisioning for long-term operation and maintenance needs are examples o f initiatives that improve the quality o f the investment program across functional expenditure areas. 2.2 The analysis in this chapter addresses selected issues in horizontal expenditure reforms. The first section reviews trends in public sector employment and pay, and identifies some key challenges for the future, while the next section discusses options in c i v i l service reform. The third section reviews trends in the public investment program across sectors and highlights some priority directions to ensure the sustainability o f the medium-term investment program in infrastructure. The fourth section provides a preliminary analysis o f tax expenditures in Turkey-an important area for the improvement o f fiscal accountability, the transparency o f governance, and the safeguarding o f a strong fiscal framework. Transparency o f tax expenditures would support their rationalization; help avoid duplication with direct expenditure programs in k e y functional areas; and promote tax base broadening. A. PUBLIC SECTOR EMPLOYMENT AND THE WAGE BILL 2.3 The Government wage bill has been bypassed by fiscal consolidation. Since 1999 total compensation for General Government employees has hovered at around 10 percent o f 60 GDP, growing from 9.5 percent to 10 percent between 2000 and 2004 (Figure 2.1).29 The lion’s share o f personnel expenditures, about 80 percent, i s incurred by consolidated budget agencies. Local administrations and revolving funds account for 9 and 7 percent o f total personnel expenditure. These shares have remained stable over time. Figure 2.1: Personnel expnditures in petrent o f GDP (1999-2004) ~ 111 10 l 9 ~a 7 6 I 1999 2000 2001 2002 2003 2004 Consolidated budget SSLS . - 0 Local administrations 0 Revohnng Funds EBFs -. - Source: SPO and World Bank staff calculations 2.4 Personnel expenditures are concentrated on education, security, economic affairs, and health care. These four functional categories absorbed more than two-thirds o f the wage bill in 2004 (Table 2.1).30 The fourth major category o f personnel expenditure is in economic affairs, accounting for the provision o f infrastructure services, rural development, and services to agriculture. The share o f this functional category in total personnel expenditure is equal to that o f public order and safety and higher than defense. Provision o f general public services i s the fifth largest category o f personnel spending for the General Government as a whole. It is relatively oversized at the level o f local administrations, representing the largest component o f their wage bill. Table 2.1 : Personnel expenditures by government function and level of General Government (2004; in percent). I Consolidated Social Securitv Local - Revolving Total budget Institutions administrations Funds General Public Services 7.9 0.1 2.4 0.1 10.4 10.1 ... 10.2 12.7 0.8 ... 13.5 10.5 ... 1.6 1.1 13.2 0.1 0.5 ... 0.7 ... 1.7 ... 1.8 7.7 3.0 0.6 5.3 16.4 2.5 ... 0.6 ... 3.1 27.6 1.o 0.5 29.1 Social Protection 0.3 0.9 0.3 ... 1.5 Total 79.6 3.9 9.4 7.0 100.0 29Steep increases in the wage bill in 1999 and 2001 reflect the sharp drop in GDP in these two crisis years. 30 Personnel expenditure in education and security i s incurred by the consolidated budget, but more than h a l f o f the wage bill for health care is borne by Revolving Funds and Social Security Institutions. 61 2.5 The wage bill in economic affairs and security is oversized by international comparison. The hnctional breakdown o f General Government personnel compensation in Turkey was compared to the average o f seven other new EU members, cohesion countries, and emerging economies for which similar information i s available: Czech Republic, Greece, Ireland, Mexico, Portugal, Spain, and South Africa. The share o f personnel compensation in security spending (Defense and Public Order and safety combined) is somewhat above the average o f comparators (Figure 2.2). In particular, the share o f personnel expenditures on economic affairs is twice as high. An example o f oversized wage bill with possible room for savings i s that o f the rural agencies budget (Chapter 111, section D2). I Fire 2.3 :Enploymentin the Fublii sector(1997-2004) Figure 2.2 Composition of Wale Bill by Function (in percent of Total Wage 1 , t- 1,920,000 ' 1,900,000 l 400000 1,880,000 I Consolidated 1 350000 / budget agenues 1,860,000 I 300000 ' \ 1,840,000 General publlc Defense 8 Economic Health Educatmn, S m 250000 i Munclpalibes 1,820,000 seMceS PuMic Order Mars, Culture, ptotechon 1 Enwonmental Relglon 8 200000 1,800,000 ~ 1 Protechon HouSlng 1997 1998 1999 2000 2001 2002 2003 2004 Source: World Bank staff calculations Source: Ministry o f Finance 2.6 Public sector employment has been significantly downsized, mostly in State Economic Enterprises. Total public sector employment (excluding Revolving Funds) was 2,460,740 in 2004-about 10 percent o f the economically active population. Three quarters o f public sector employees are in consolidated budget agencies. C i v i l servants comprise the bulk o f public employment, while public workers+oncentrated in SEES and State banks- represent about 22 percent, and contracted personnel only 5 percent o f total (Table 2.2).31 Overall, there has been a decreasing tendency in employment o f all categories o f public sector employees. Sector-wise, employment in consolidated budget agencies increased until 2000, and then shrank by 4.5 percent up until 2004 (Figure 2.3). The most significant employment downsizing has occurred in state economic enterprises, o f around 30 percent from 1997 to 2004. Employment in municipalities has remained stable. However, a significant change in the composition o f municipal employment has occurred, with the share o f temporary workers increasing to 45 percent o f total in 2004, from 25 percent in 1995. 3 1 Civil servants are defined as those public employees carrying out public services o f a primary and continuous nature. They enjoy constitutional protections which effectively provide a lifetime j o b guarantee. Contracted personnel in theory are employed t o fill temporary needs (typically unfilled c i v i l service positions), and are hired o n one-year, renewable contracts. In practice, Turkey's courts have ruled that they enjoy a l l benefits and protections o f c i v i l servants. Public workers have a contract but do not occupy a c i v i l service position. They do not enjoy the same j o b protections as the other categories but, unlike c i v i l servants, they are b o t h allowed collective bargaining and have a right t o strike. 62  Table 2.2: Total employment in Turkey’s public sector (l), September 2004 Civil servants Contracted Workers Total % of Total Personnel Public Sector Consolidated 1,620,341 17,182 184,621 1,822,144 74.0 budget Municipalities 92,487 2,178 106,700 201,365 8.2 Special provincial 3,924 87 762 4,773 0.2 administrations Social Security 74,391 349 3,728 78,468 3.2 institutions State economic 7,880 106,670 209,614 324,264 13.2 enterprises, State banks 58 29 29,639 29,726 1.2 Total 1,799,081 126,495 534,892 2,460,740 Percent of total 73.1 5.1 21.7 100 (1): Excluding Revolving Funds and Extra budgetary Funds Source: Ministry o f Finance 2.7 Civilian public sector employment is not exceptionally high by international standards. Using only central government employment data (consolidated budget agencies for Turkey), and excluding the functions o f defense and public order and safety, public sector employment in Turkey amounted to 2.2 percent o f population in 1999. This is somewhat above the scale o f public employment in other emerging economies such as Chile, Korea, and Mexico. However, i t is below the figures seen in other Eastern European and Central Asian countries-including new EU members (Figure 2.4a). Excessive employment in the public sector does not thus seem to be the main reason for the.relatively large wage bill by international comparison. 2.8 Average compensation o f Central Government employees is high by international comparison and has grown fast in recent years. High average compensation has been the main factor driving up the wage bill and largely accounts for differences with comparator countries. Indeed, the average wage in consolidated budget agencies in Turkey was in 1997- 2000 2.6 times higher than per capita GDP. This ratio was the second highest among the group o f comparators for which similar data are available (Figure 2.4b). At the same time, since 2000, average compensation in consolidated budget agencies has grown faster than various measures o f per capita income for the economy as a whole. Average compensation has thus increased to 2.9 times o f per capita GDP in 2001-2004 (Figure 2 . 4 ~ ) . ~ * Measured as a percentage o f GDP per worker in the non agricultural sector, average compensation has followed a very similar path, increasing from 54 percent in 1997 to 68 percent in 2004. Fast growth in average compensation has occurred despite the government’s policy o f restricting nominal compensation in the public sector so as not to exceed the annual change in the CPI. After offsetting fluctuations in 1999-2001, real average compensation surged by more than 6 percent in 2002 and by close to 10 percent in 2003 and 2004, as nominal increases by far surpassed CPI inflation and GDP growth (Figure 2.4d). 32 In this calculation, per capita GDP and compensation are measured in constant 1987 prices. A similar picture emerges f r o m nominal G D P and compensation figures, with the ratio o f average wage to per capita GDP increasing from 2.8 in 2000 to 3.4 in 2004. 63 Figure 2.4: Employment and average wage cost in the public sector F z u re 2.4a : C e n h l Government employment (excludmg Defeme and recuiity) in Figure 2.4b : Average prronnel compemation in c e n m l government as a ralio to - p r r e n t of population 1999 per capita GDP (1W-2000 for all countries and 2001-04 for Turkey) ~ 351 35 30 ~3 25 25 20 2 15 15 10 05 00 1 Russia Bulgaria Poland Romania Turkey Chile Korea Mexw Public sector employment i s not oversized.. . ...but average compensation is high by international standards.. . Figure 2 . 4 ~ : Trrnd~ in average personnel compmation m comolidated budget Figurr 2.4d: Average compmatmn m e o m d i a t e d budget agencies and CPI agencies (1997-2004; as rabos to p r c a p i t a i n c o r n measurrs) inflation (1998-2004; in percent) 31 07 n 90 15 j T T 3 IAverage / 1068 !! J , . ~ rnnagricuitual~ employee (ngM a m ) ~~ per 06 0 58 056 054 23 0 52 22 05 -15 1997 1998 1999 2OM) 2001 2002 2003 2004 1998 1999 2000 2001 2002 2003 2004 ...and has been growing faster than per capita income.. . . ..because nominal compensation increases have b y far exceeded inflation Source: World Bank staff calculations 2.9 Slower growth in average compensation would have created considerable room for fiscal savings. From 2000 to 2004 personnel compensation in consolidated budget agencies has increased from 8 to 8.5 percent o f GDP despite downsizing o f employment by about 90,000 employees (see Figures 2.1 and 2.3). The evidence summarized above suggests that the increase in the wage bill i s largely driven by the increase in average compensation. For example, had the ratio o f average compensation to per capita GDP remained constant at i t s average level in 1999-2000, the wage bill o f consolidated budget agencies would have fallen to around 7.6 percent o f GDP in 2004-that is, about 1 percentage point lower than i t s actual level (Figure 2.5). Better management o f wage increases in the public sector would thus create significant scope for savings in the budget. Linking wage increases to inflation has not proven enough to contain the growth in the wage bill, probably because other forms o f compensation have increased higher than the inflation rate. Moving to a simpler and more transparent compensation system would contribute to containing pressures in the wage bill. Compensation increases will also need to be more strongly anchored on expected inflation to contain pressures on the wage bill. 64  Figure 2.5 :Wage bill in consolidated budget agencies - actual and alternative scenario (1999-2004; in % ofGDP) . " 9- L \ Actual 8.5 \ * \ a- 0.9% 7.5 - Constant ratio of average wage to pc 7- GDP at I99912000 level 1 6.5 ' 1999 2000 2001 2002 2003 2004 Source: World Bank staff calculations 2.10 The wage bill could further increase as a result o f the additional pay rise for civil servants granted for 2006. In addition to the 5.1 percent salary increase agreed for 2006, the Government introduced in March 2006 additional payments o f YTL 40 for the first half and YTL 80 for the second h a l f o f the year to the public servants who do not receive additional payments. The estimated cost o f this action i s about YTL 1.7 billion, the equivalent o f 0.35 percent o f GDP. Part o f this extra cost, YTL 1.2 billion, had been budgeted and the difference will be financed from the personnel contingency in the 2006 budget. If this pay rise i s built in the salary base for pay increases in 2007 and future years, or if it triggered similar demands in the future, the fiscal cost could be sizable as i t would have a permanent impact on the Government's wage bill. 2.11 The compensation system is dominated by a multitude o f side benefits. In 2004, about 32 percent o f personnel compensation in consolidated budget agencies was composed o f various allowances, compensations, and rewards on top o f the basic salary o f employees. Moreover, overtime pay represented 6.5 percent o f total compensation. While wages are calculated on a transparent basis (driven by a combination o f grade and years in service), allowances and overtime pay seem complicated, seemingly offering substantial opportunities for discretion. Allowances and other rewards have been somewhat reduced in 2003 and 2004, by about 4 percentage points o f total pay (Figure 2.6). Overtime pay has been on an upward trend since 1997, but has been stabilized in the past two years. Overtime pay (0.5 percent o f GDP) i s a more expensive form o f compensation in the civil service and would need to be reduced more aggressively to contain pressures on the wage bill. This would call for an appropriate redefinition o f basic civil service obligations. 65  Figure 2.6 : Compensations, R e w a d , AUowances, and overtime pay in percent of total personnel expenditms ofcomolidated budget agencies (1997-2004) 40 l------ 1 8 38 -- 36 -- 34 -- 32 -- 30 -- /Overtime pay Compensations, J- 28 ~- rewards, 26 -- allowances 24 -- c-- t‘ I 1997 1998 1999 2000 2001 2002 2003 2004 Source: Ministry o f Finance 2.12 Many allowances are exempt from the income tax. Examples include “horizontal” allowances, such as those for representation; foreign service; university allowance; or allowance for knowledge o f foreign language. Several allowances for specific duties and overtime payments are also exempt from the income tax. Although a precise breakdown o f the amount o f non taxable allowances is not available, in view o f the substantial size o f allowances and overtime pay (3.3 percent o f GDP for consolidated budget agencies), their cost in terms o f foregone tax revenue i s likely to be significant. Assuming that 70 percent o f allowances and overtime pay are non taxable, and based on a 20 percent effective marginal tax rate, foregone tax revenue could be around 0.5 percent o f GDP (also see section D below). 2.13 Sizable side benefits reduce transparency and may be distorting incentives in the civil service compensation system. The share o f allowances i s disproportionately large for the highest grades. For the highest grade o f c i v i l servants, wages (including basic seniority and indicator salary plus side payment) account for only 15 percent o f total compensation. For the lowest grade, wages typically account for about 75 percent o f compensation. The exceedingly high role o f allowances among the highest grades i s cause for concern for transparency and incentives in the c i v i l service, and warrants further analysis. Moreover, as allowances represent a much larger share o f income for higher-grade c i v i l servants, exemptions from the income tax are detrimental to vertical taxpayer equity. One plausible hypothesis i s that these more discretionary forms o f pay are the means via which Turkey achieves de facto decompression o f what i s otherwise an excessively compressed set o f pay differentials between more and less skilled c i v i l servants.33 Achieving decompression in compensation scales will have to be an important part o f c i v i l service pay reform. Without 33Adequate data does not exist t o assess the compression o f wage scales in c i v i l service in international comparison. While the distribution o f wages can be assessed by grade for basic pay, similar information i s n o t available for allowances and other side benefits. 66 decompression, taxation o f allowances would reduce pay differentials and may have unintended effects o n incentives. B. OBJECTIVES AND REFORM OPTIONS IN THE MODERNIZATION OF . CIVIL SERVICE 2.14 The Government is considering plans for civil service reform to address some o f the deficiencies of the existing system. The proposed draft law o n public personnel seeks to address a number o f major objectives. These include: 0 improving efficiency in staffing and performance o f duties; eliminating discrepancies in remuneration across similar positions within the public sector; 0 making i t easier to manage a fiscally sound wage bill; 0 providing competitive opportunities for entry into the public administration The analysis below reviews some o f the reform options to meet these challenges but further detailed analysis o f opportunities for modernizing the rules governing c i v i l service pay-and any related opportunities for budgetary savings-is warranted. In addition, c i v i l service reform would need to address issues related to the deployment o f a number o f public employees to the special provincial administration as a result o f the recently enacted legislative framework. The majority o f those employees are still associated with the consolidated budget in terms o f their status and wages, even though centrally determined compensation is often not directly related to their performance, and does not reflect differences in regional living standards. (0 Main reform objectives 2.15 Improving efficiency: T o improve efficiency this draft law mandates changes aimed at three broad sub-objectives: (i) make i t easier to adjust staffing composition when required;34 (ii) enhance sources o f performance motivation for public personnel;35 (iii) facilitate opportunities for public servants to improve their human capital over the course o f their careers within the public service36 To address (i), the draft law would establish a legal distinction between “civil servants” and “contracted personnel”, so as to permit greater flexibility regarding the termination o f employment for “contracted personnel” than i s possible for “civil servants”, given Article 128 o f Turkish Republic C o n ~ t i t u t i o nIn ~ . ~ addition, the proposal would mandate dismissal o f 34 To increase allocative efficiency 35 To increase technical efficiency 36 To increase dynamic efficiency 37 M o r e specifically, the draft l a w would define “contracted personnel” as those undertaking responsibilities that were either not “primary” or not “permanent” . .. “services required for public services” (Article 3), and would allow such staff t o have their contracts terminated upon end o f contract within the first 15(-) years o f such a contractual relationship (Art. 17), although they would still enjoy the same retirement (pension) rights as “public servants”. I t would also allow contracted personnel to “annul their contracts according t o the methods and principles stated in their contracts” (Arts. 19, 63 and 64). 67 public servants receiving repeated “unsatisfactory” annual perfonnance evaluation^.^' To address (ii), the draft law would: (a) authorize a modest amount o f performance bonuses, subject to restrictions o n both the amount received by any individual employee as w e l l as the total expenditures on such bonuses by each public entity;39 and (b) mandate dismissal o f public servants receiving repeated “unsatisfactory” annual performance evaluations. T o address (iii),the draft law would provide generous opportunities for public servants to obtain either graduate level education or valuable work experience during periods o f secondment to international organizations or other overseas employers, during their public service careers, which can significantly enhance the attractiveness o f a career in Turkey’s public service. Moreover, heavier reliance would be placed o n Turkey’s higher education and private sectors for providing in-service training to its public service staff, thereby taking advantage o f competition in the provision o f in-service training. 2.16 Eliminating discrepancies in remuneration across similar positions within the public sector: To accomplish this, this draft proposal would consolidate the legal framework governing salary setting throughout the public sector, so as to: (i) establish seven uniform salary scales for public servants in each o f four separate career streams, and public contracted personnel in each o f three career streams;40 (ii) concentrate salary in (a) basic duty salary or wage,41 and (b) duty difference a l l ~ c a t i o n .Salary ~~ would thus be concentrated in those elements most clearly linked to human capital requirements and demands o f the position.43 38 Article 75(4) provides that “Public servants, whose personnel and success evaluation i s negative t w o times successively are appointed to the order o f another personnel and success evaluation supervisor. Public servants whose evaluation i s here again negative and those public servants who get negative evaluation a total o f four times within ten years are dismissed and these can never be employed again as public servants o r contracted personnel.” 39 Performance bonuses cannot exceed 5% o f an employee’s basic duty salary o r basic duty wage, while each public institution’s payments o f such bonuses cannot exceed 1% o f its total personnel expenditures, nor can it exceed the institution’s budget allocation for performance pay during the applicable budget year. 40 The public servant basic duty salary scales apply to: (i) general public servants, (ii) military public servants (i.e., officers, non-commissioned officers, and specialist gendarmeries), ( iiijudicial sector public servants (i.e., ) judges and prosecutors), and (iv) academic public servants (Le., university instructors); while the contracted personnel basic duty wage scales apply t o (v) general contracted personnel, (vi) military contracted personnel (i.e., specialist enlisted leaders), and (vii) academic contracted personnel (Le., university instructors). 4’ Le., the part o f salary o r wages fixed by the seven uniform basic duty salary and wage scales established by t h i s law. 42 The draft l a w constrains the awarding o f these “duty difference” pay amounts by specifying (i) grounds justifying “duty difference” payments, (ii) procedures for setting payment levels, and ( iii) limits o n their magnitudes. (i) Grounds: Such “duty difference” pay would b e based o n “the hardship o f the service, responsibility and risks, work intensity, work conditions, socio-economic development level o f the service location, living conditions and geographic characteristics, duty and other characteristics relating to the service location” (Art. 1lO(2)). ( ii) Procedures: The amounts and rules and procedures governing assignment o f such “duty difference” pay to individual personnel would be established through the following procedures: (a) proposals t o be prepared by the institution within which such “duty difference allocation” i s proposed to be established; (b) the State Personnel Administration t o provide i t s opinion o n each proposal; (c) the Ministry o f Finance t o submit each proposal to a commission consisting o f the Undersecretary o f the Ministry o f Finance, the Undersecretary o f the State Planning Organization, the Undersecretary o f Treasury and the Chairman o f the State Personnel Administration, under the chairmanship o f the Undersecretary o f the Prime Ministry, and (d) that commission to make the final decision (Article 110 (2) and (3)). ( iii) Limits: Such payments are limited to n o more than the lower o f (a) “twenty percent o f the highest basic duty salary” o r “forty percent o f the personnel’s own basic duty salary o r basic duty wage and fifty percent for” selected military positions (Article 110(1)). 43 Articles 108-1 11. Articles 112-1 16 provide other elements o f compensation. 68 In addition, in conjunction with related changes in the Social Security and Universal Health Insurance Law, the proposal would ensure that any elements o f remuneration that bring with them social security (pension) rights, also impose commensurate social security contribution (premium payment) obligations o n the recipient o f those elements o f remuneration. 2.17 Making it easier to ensure afiscally sound wage bill: T o address this objective, the proposal would consolidate the legal framework governing salary setting throughout the public sector as described above. This should make it easier for the Ministry o f Finance to monitor and analyze central government wage bill implications o f employment and pay policy proposals or changes. Moreover, this consolidation would ensure that any elements o f remuneration that bring with them social security (pension) rights, also impose commensurate social security contribution (premium payment) obligations o n the recipient o f those elements o f remuneration, thereby reducing perverse incentives to obtain eligibility for such elements o f salary, as well as eliminating a fiscally imprudent practice o f providing pension benefits not tied to pension contributions. 2.18 Providing competitive opportunities for entry into the public administration: T o accomplish this, the proposed draft law would continue Turkey’s current requirements for standardized examination-based competitive recruitment and selection procedures for both public servants and contracted personnel, as well as for promotions. Regulations governing those procedures would be established by the State Personnel Presidency. 2.19 All o f the proposed changes in the legal framework represent important steps toward achieving those objectives. With the aim o f supporting these efforts, it i s important to highlight implementation challenges posed by the proposed reforms in the public personnel management legal framework, as w e l l as a few additional areas in which complementary reforms could further the Government’s efforts to achieve the above four broad objectives. (ii) Implementation challenges and complementary reform areas 2.20 Some changes to be introduced by the proposed draft law are likely to pose implementation challenges: (i) establishing a clear legal distinction between c i v i l servants and contracted personnel; (ii)mandating dismissal o f c i v i l servants for repeated “unsatisfactory” performance rating; (iii) introducing performance bonuses; (iv) concentrating salary in basic duty and “duty difference” components; (v) setting basic salary and wage scale index (“indicators”) structure; and (vi) recruitment and selection procedures. Other risks to one or more o f the key reform objectives include: (vii) step salary increase provisions; and (viii) personnel performance evaluation procedures. 2.2 1 Establishing legal distinction between public servants and contracted personnel: The text o f this draft law which provides a legal basis for permitting non-renewal o f contracted personnel contracts has been very carefully drafted. I t i s unclear how the Turkish judicial system i s likely to r u l e o n challenges to these provisions. The Government has and continues to pay very close attention to the risk o f an adverse judicial finding once these provisions are challenged. This i s an unavoidable risk and challenge posed by the laudable 69 determination to create a more flexible labor market for an important subset o f Turkey’s public employees; a challenge that the Government fully recognizes. An issue that would merit attention i s that this draft law leaves entirely to other “special laws” the task o f defining which particular cadres (for public servants) and positions (for contracted personnel) shall exist. This leaves substantial scope for lack o f uniformity in the determination o f which jobs will be cadres (i.e., performed by public servants) and which will be positions (i.e., performed by contracted personnel). At a minimum, a set o f procedures should be established to provide adequate checks and balances and reasonable assurance that these determinations will follow a relatively consistent set o f criteria. 2.22 Dismissal for repeated “unsatisfactory” performance ratings: The current L a w o n allows State F u n ~ t i o n a r i e s ~~ termination o f employment for public servants only for events such as major disciplinary actions, voluntary departure, retirement, and death. Poor performance i s not grounds for dismissal. This new draft law would change that by making consistent poor performance grounds for dismissal. This i s a fundamental, important and much needed change. There i s a risk, however, that a poorly functioning annual personnel performance evaluation process will make this a change with little or no impact. This i s also an issue for promotions. 2.23 Performance bonuses: Performance bonuses are sensibly subject to important restrictions on their magnitudes in the draft legislation. Still, as the Government is aware, i t i s difficult to get a performance pay system to work as well as one might hope. In addition to limiting the magnitudes o f such performance bonuses, perhaps the most important challenge i s to get the performance evaluation process to work reasonably well. Some options for addressing this challenge are listed below. 2.24 Concentrating salary in basic duty and “duty difference” components: Although this i s a move in the right direction, the current draft law does not uniquely determine what fraction o f total remuneration will reflect these two components. The large number o f other elements o f remuneration and leave rights could compromise this objective. Careful analysis will be required to assess whether the implementation o f these provisions will actually concentrate salary in basic duty and “duty difference” elements o f remuneration. 2.25 Setting basic salary and wage scale index (%dicators’y structure: Because o f current significant discrepancies in remuneration for similar positions across institutions, implementation o f the more uniform salary structures mandated by this law will require changes in salary relativities. If implementation required n o change in the overall wage bill, some staff would see their remuneration rise while others would see i t fall. If, o n the other hand, any remuneration decreases are to be avoided, the wage bill would have to rise. This poses difficult trade-offs, o f which the Government i s well aware, especially as the wage bill is already high in international comparison and average remunerations increased significantly in recent years. 2.26 Managing the fiscal cost o f compensation harmonization. Multiple scenarios are being modeled so that the implications o f compensation harmonization (both for the overall 44 Law 657 o f 14 July 1965. 70 wage bill and for remuneration changes for individual staff) can be adequately understood before adopting a salary structure and implementation plan. In addition, phasing in o f the new salary structure is almost certain to be an important element o f the implementation plan. Appropriate design o f the reform would be needed so as to minimize its overall budgetary impact, while gradual phase in would be advisable so as to spread out the cost over time. 2.27 Recruitment and selection, and promotions procedures: The draft law would continue the examination-based, competitive recruitment and selection procedures as well as similar procedures governing changes in grade (ie., promotions), which were established within the last six years.45 The draft L a w leaves to the State Personnel Presidency the responsibility for regulating these processes. I t will be important to continuously monitor how well they are working and to make adjustments as necessary to ensure their integrity is not compromised. 2.28 Step salary increases: These are, essentially, basic salary or wage increases based o n seniority. Only personnel receiving unsatisfactory ratings in the annual performance review process at least once in any three-year period are excluded from these step pay increases. If the performance evaluation process does not reliably ensure that poor performers receive “unsatisfactory” ratings, these step salary increases will become guaranteed salary increases tied to seniority. Under these circumstances, these step increases will tend to be viewed as entitlements by staff, rather than as a source o f motivation. At best, they will motivate staff to avoid flagrant violation o f work rules. Moreover, they will pose the additional problem o f creating fiscal costs that cannot be easily controlled. 2.29 A number o f options exist for addressing both the low motivation potential o f such step increases as well as the dijj7culty o f controlling wage bill costs. These include: (i) requiring that the average magnitude o f step increases provided each year to be decided as part o f the annual budget process; (ii) eliminating uniform step increases, and replacing them with organizational unit-specific salary increases based on (a) annual performance ratings, and (b) a fixed organizational unit-specific wage bill budgeted for the next budget year;46 ( iii) placing a limit on the fraction o f staff to whom a supervisor can assign the highest performance ratings; (iv) including various design features in the annual personnel performance appraisal process designed to ensure that performance ratings reliably sort staff by their performance. 45 As noted in S I G M A ’ S July 2004 assessment “the current recruitment and promotion procedure was introduced in 1999, and Government Decree 24744 o f 3 M a y 2002 created a single recruitment centre and set out a recruitment procedure for the entire central administration”. Furthermore, “the Student Selection and Placement Centre ( O S Y M in the Turkish acronym), i s attached t o the Autonomous Council o f Higher Education and works in close co-operation with the State Personnel Presidency (Prime Minister’s Office).” Similar arrangements were mandated in 2000 for promotions. 46 Each organizational unit could, o f course, be assigned an identical percentage change in i t s wage bill budget f r o m year 0 to year 1; e.g., the average percentage change in the overall wage bill that w o u l d b e consistent with the Government’s overall fiscal policy for the upcoming budget year. A formula would define the relationship between performance ratings and salary changes. Managers o f organizational units, then, would have to assign performance ratings such that the total projected wage bill for their organizational unit for the upcoming budget year (year 1) would fit within the assigned wage bill envelope for their organizational unit. 71 2.30 While it is probably premature to consider a more radical approach to ensuring that regular salary increases reflect performance, the Government may want to consider in the future a reform o f this step increase system. Rather than guaranteeing step increases every third year, the annual salary adjustment mechanism could set an overall wage bill envelope for the succeeding year (perhaps for a given set o f staff; e.g., an organizational unit), and then employ the annual performance ratings (within that organizational unit) to apportion salary increases among the set o f staff covered by that wage bill envelope. A specific formula could be used to that effect, subject to the constraint that the sum o f individual salary increases cannot cause the resulting wage bill to exceed the fixed wage bill envelope. Such a mechanism would force managers evaluating their personnel to assign performance ratings that yield a fiscally feasible wage bill. 2.3 1 Personnel performance evaluation procedures: It i s notoriously difficult t o get performance evaluation procedures to reliably distinguish between actual performances across staff.47 Given the important role that annual personnel performance evaluations will play under this new law in promotions, bonus pay determinations, step increases, as w e l l as for providing a mechanism for terminating non-performing public servants, i t is especially important that c a r e h l attention be paid to ensuring that Turkey’s performance evaluation processes work well.48 The proposed draft law leaves to the State Personnel Presidency the task o f spelling out detailed personnel evaluation procedures and criteria. It also quite sensibly requires the Turkish Public Administration Institute (Turkiye ve Orta Dogu Amme Idaresi Enstitusu-TODAIE) to provide an independent assessment o f the system established by those detailed regulations. The one issue that would merit more attention here i s the process itself-and, in particular, at least four dimensions o f that process: 0 involving individual staff members in reaching agreements with those who will evaluate them before the beginning o f an evaluation period o n h o w their success in meeting the criteria specified in the regulations should be judged - i.e., what sorts o f evidence will be 0 requiring input from multiple persons in distinct positions to assess the performance o f any given staff member (e.g., supervisor, peers, clients); 0 subjecting the evaluator’s consolidation o f those inputs into a single evaluation to review by hisher superior before it is finali~ed;~’ and 47 OECD, Public Governance and Territorial Development Directorate, Public Governance Committee, H u m a n Resources Management Working Party, Performance-Related Pay Policies for Government Employees: Main Trends in OECD Member Countries draft, 8 October, 2004. Web address: h~://www.oecd.org/document/39/0,234O.en 2649 37405 33687079 1 1 1 37405,OO.html 48 These issues, are also examined in: http://web. worldbank. org/WB SITE/EXTERNAL/TOPICS/EXTPUBLICSECTORANDGOCE/EXT ADMINISTR4TIVEANDCIVILSERVICEREFORM/O,,contentMDK:20 13344 1-menuPK: 18289 10-pagePK: 1 48956-piPK:2 16618-theSitePK:286367,00.html 49 Art. 111(2) does require each public entity (“institutions and their units”) t o establish and announce performance criteria “within the first month of the financial year when the performance evaluation is made”. I t doesn’t, however, require any input from the staff who will be judged by those performance criteria; let alone, their agreement to those criteria. 50 The use o f “performance evaluation commissions”, as specified in Art. 111, i s one option for addressing this need for review by a third party, since those commissions are required to include five members, at least one o f which must represent the union representing the largest fraction o f staff within the work unit. 72  creating some rules or incentives designed to counter the natural tendency for supervisors to give most (if not all) staff the same (high) evaluations (e.g., setting fixed limits o n the fraction o f “very good” ratings - i.e., requiring supervisors to rate their staff o n a “curve”. 2.32 Striking a better balance between remuneration expenditures and expenditures on complementary inputs (training, capital, other recurrent costs): The effectiveness and efficiency with which organizations operate depends not solely o n their personnel and h o w motivated they are, but also on the other inputs with which those personnel work. Inadequate capital facilities, insufficient operating supplies, sub-optimal maintenance o f facilities, as well as failure to continuously upgrade the knowledge and skills o f personnel will all compromise operational effectiveness and efficiency. As such, i t i s important that reforms be complemented by steps to ensure that adequate resources are devoted to such complementary inputs. Given tight budgetary resources, this m a y require reallocation o f resources from wages and salaries to these other inputs. This poses another unavoidable and difficult trade- off, which all stakeholders must face. C. ISSUES IN MANAGING AND FINANCING THE PUBLIC INVESTMENT PROGRAM 51 2.33 Public investment in Turkey is geared more towards infrastructure, general public services-including public order and safety-and education, with respect to comparators, The composition o f C G G gross fixed capital formation by function reveals that public investment allocations in Turkey differ in some respects from the group o f comparators (Figure 2.7).52 Public investment for the provision o f infrastructure services (included in economic affairs) accounted for 60 percent o f total investment in 2003-04 and exceeded marginally the average o f comparators, although Turkey’s infrastructure gaps remain large.53 In stock terms, infrastructure projects represent more than 80 percent o f the investment portfolio (Table 2.3). The apparently large share o f infrastructure investment is very similar 51 This section focuses mostly o n the infrastructure sector recognizing the h i g h share o f the sector (more than 80 ercent) in the existing public investment stock. p2 Due to limited data availability o n functional breakdown o f investment spending, only the following countries have been included in the sample: Ireland, the Netherlands, Greece, Spain, Portugal, Sweden, Finland, Czech Republic, Belgium, Austria, and South Africa. 53 Provision o f infrastructure services involves three different types o f public entities at various tiers o f government. Central Government Agencies (CGAs) build major intercity express ways, v i l l a g e h a 1 roads, and most railways, seaports, and airports at national level. They also provide major urban water supply and treatment projects at local level. Independent central government-owned state economic enterprises (SEEs) provide electricity generation, transmission and distribution; railways, most airport and p o r t operation; o i l and gas distribution; postal services; and most telecommunications services. Local Authorities (LAs) - more precisely municipalities - are the m a i n deliverers o f infrastructure at the sub-national level. They are responsible for urban roads, urban public transportation, and solid waste disposal. 45 percent o f the total public investment in infrastructure i s spent through the budget. The remaining share is spent through SEEs (roughly 20 percent) and LAs (about a third). B o t h on-budget and off-budget public infrastructure providers are included in the estimation o f fiscal targets. 73 to what was observed for infrastructure public investment in Latin America countries particularly during the 198Os, before the introduction o f major private participation in infrastructure services but - most importantly - before major fiscal adjustments were .~~ ~ n d e r t a k e nAs a percentage o f GDP, investment in public goods (general public services and defense, public order and safety) and in education also exceeded that o f comparators. By contrast, investment for environmental protection remains negligible and may need to increase soon to levels comparable with benchmark countries. Figure 2.7: Gross Capital Formation by Function-Turkey (2004) and comparators (2003), in % of GDP . ,.".I...". . .. . . . ... I ...... . . . ... . .."- . . .. . .I " 1.4 - 1.39 I .Average of 11 comparators OTurkey, 1.2 - 1.0 ~ -IdllL I' 0 49 - 0.54 0.49 0.49 0.02 General public Economic affairs Housing and Education Other social Defense, S ~ N I ~ S , community sectors Public order and amenities, safety Environmental protection Source: SPO, World Bank Staff calculations Table 2.3: Infkastructure Investment Portfolio: Value and Number o f Projects, 2000-2005 Non- Infrastructure Total Value (Cost) Portfolio as a of Portfolio Share o f the Value o f Investment Stock Infrastructure I Infrastructure Stock as a Share Infrastructure Infrastructure Investment o f # Projects Total Portfolio I ( 000 YTL) (percent) (# Projects) (percent) 2000 71,925,876 14,293,392 83 5,321 I 1,100 21 2001 118,589,325 24,329,656 83 5,047 1,052 21 2002 138,143,540 28,653,287 83 4,414 20 2003 155,384,720 31,725,593 83 3,851 2004 161,373,229 34,739,639 82 3,555 I 785 22 2005(*) 169,961,9 18 36,722,401 82 2.627 I 742 28 2.34 Over the last fifty years, the urban population has grown twice as fast as the overall population and reached 65 percent o f the total population in 2000. H a l f the urban population resides in 16 metropolitan areas. This has placed severe pressures, particularly on municipal service operators to provide quality services despite the lack o f sufficient finances. Over the same period the number o f municipalities roughly doubled from over 1,700 to about 3,200. 54 That was the case o f Argentina, Brazil and Bolivia. F o r details, refer t o Easterly and Serven (2003). 74 The rural population (about 35 percent o f the total), resides in about 35,000 villages with an average size o f 500 persons. 2.35 Meeting financing requirements o f infrastructure is thus a key challenge for sustained growth. K e y issues in infrastructure financing in Turkey include (i) ensuring that infrastructure investment addresses economic development objectives, including those arising from the need to accelerate convergence and facilitate EU accession; ( ii) protecting high-priority investment projects from fiscal adjustments, and ensuring that key multi-year projects are carried on and completed on time, ( iii) long-term provisioning for operation and maintenance o f existing assets, (iv) ensuring adequate security and reliability o f supply o f infrastructure services; and (v) gradually increasing the role o f the private sector in the financing, development and operation o f infrastructure in Turkey, with careful design o f Government commitments to private operators to minimize the risk o f contingent liabilities. (4 Ensuring sufficient allocation and timely completion of high-priority projects 2.36 Public Investment was cut significantly as part o f the fiscal adjustment efforts during 2001-2004. Total public investment level (inclusive o f local administrations and SOEs investment), stood at 4.2 percent o f GNP in 2004, down from 6.8 percent in 2000 (Table 2.4). In 2001-2002, fiscal adjustment fell primarily on public investment by SEEs while investment by the consolidated budget increased. In 2003-2004 however, the cut was predominantly from the investment budgets o f both the consolidated budget agencies and SEEs. Investment spending in 2005 increased by close to 1 percent o f GNP, mainly due to increased investments by local administrations. Table 2.4. InstitutionalBreakdown of Public Investments ('YOof GNP) 1/ 1999 2000 2001 2002 2003 2004 2005 2006 Consolidated Budget 2.5 2.8 3.3 3.6 2.5 2.3 2.5 2.3 SEEs 1.7 1.8 1.3 1.3 0.8 0.7 0.7 0.8 Province Bank 0.2 0.2 0.2 0.1 0.1 0.1 0.1 0.1 Local Adm. 1.4 1.3 1.5 1.3 1.5 1.2 1.6 1.6 Revolving Funds 0.1 0.1 0.1 0.1 0.0 0.0 0.1 0.1 SOCSec Adm 0.0 0.0 0.1 0.0 0.0 0.0 0.0 0.0 Funds 0.7 0.6 0.2 0.0 0.0 0.0 0.0 0.0 TOTAL PUBLICSECTOR 6.6 6.8 6.4 6.3 4.9 4.2 5.1 5.0 I/Figures include investment worker payments. Source: SPO, World Bank Staff calculations 2.37 Rationalization o f the PIP in 2001 has been quite effective. I n2001, the government initiated a rationalization process, in order to allocate the reduced amounts o f appropriation and improve the average project completion time (Table 2.5). The SPO rationalized the investment portfolio by eliminating about 1000 projects in the 2001 program consisting mostly o f transport, energy and agriculture projects and by reducing the number o f multi-year projects. In addition, the program included new feasibility studies for some projects while criteria for inclusion o f new projects in the pipeline were tightened, with only those projects which have certainty o f funding, external or domestic, being considered for approval. As a result the total cost o f projects in the PIP was reduced and average completion time was cut by 32 percent. 75 2.38 Rationalization efforts continued in 2002-2005. The total number o f projects declined from over 5000 in 2001 to 2627 in 2005. W h i l e the average completion times (based on actual annual investment spending) are higher in 2003 and 2004, this reflected more the decline in the budgetary allocation, than the composition o f the portfolio. Energy and transportation investments were hit the hardest, since jointly they constitute about 40-50 percent o f total public sector investments. In its Medium-term (Economic) Program for 2007- 09, published in June 2006, the Government remained committed to continuing PIP rationalization by reducing the number o f new projects entering the portfolio and by increasing the share o f investments in the Budget. Table 2.5 : Public Sector Investment, 1999-2006 11 (current prices, million YTL) Number of l ~ o t aValue Initial Initial Actual Actual Investment Average Time o f Average Time of New of Projects Allocation of GNP)( "' Investment (in % o f GNP) Completion 31 Completion 41 Projects Total Multiyear 1999 5458 57,126 3,560 4.5% 3,649 4.7 10.2 9.9 1,186 305 2000 5321 86,219 5,905 4.7% 6,183 4.9 9.2 8.7 1,176 249 2001 21 5047 142,919 7,167 4.1% 7,570 4.3 12.5 11.8 1,234 286 2002 4414 166,797 10,590 3.Yh 12,223 4.4 8.5 7.2 1,066 128 2003 3851 187,110 12,464 3.5% 10,386 2.9 7.6 9.7 1,032 134 2004 3555 196,113 11,978 2.8% 10,835 2.5 8.1 8.6 1,079 149 2005 2627 206,684 15,875 3.3% 14,222 2.9 6.7 7.6 985 137 2006 2525 200,391 17,522 3.2% 5.5 1,013 155 I1 Excluder local admmistrauons. U includessupplementar). budget allocauon ofTL 280 million and T L 555 m l l m from Law 3418 ~\enues. 31 Average ume of sompleuon IS calculatedas Ihe amount of time required 1 0 finish up the remainingstock completely, assuming that no other proleem are laken inlo the invesment program in the following years and an appropriation and expenditure IS made ofthe magnitude of the current year Ieyel. 41 Bawd on achlal invesmenl Source: SPO 2.39 The brunt o f adjustment was borne by investment in infrastructure. Table 2.6 below gives the annual investment spending for public sector excluding local administrations for the period 1999-2006. The annual allocations for infrastructure investment came down from above 3 percent o f GNP in 1999-2000 to below 2 percent in 2003-05. This was the main source o f adjustment in public investment program, although education and health investments also declined. Table 2.6. Annual InvestmentSpending (% of GNP) I / 1999 2000 2001 2002 2003 2004 2005 200621 Economic Infrastructure 3.1 3.4 2.6 3.0 1.8 1.6 1.8 1.7 Water Resources-Imgation 0.3 0.4 0.3 0.4 0.2 0.2 0.2 0.2 Energy 0.9 0.9 0.8 1.2 0.7 0.5 0.5 0.4 Transport and Communication 1.7 1.8 1.2 1.1 0.6 0.8 0.9 0.9 Dnnkable Water,Sanitation 0.3 0.3 0.3 0.2 0.2 0.1 0.1 0.2 Social Infrastructure 0.9 0.8 0.9 0.8 0.7 0.5 0.6 0.6 Education 0.7 0.6 0.7 0.6 0.5 0.4 0.4 0.4 Health 0.2 0.2 0.3 0.2 0.2 0.1 0.2 0.2 Economic Sector 0.3 0.4 0.4 0.3 0.2 0.2 0.3 0.3 Others 0.3 0.4 0.3 0.3 0.2 0.2 0.2 0.3 TOTAL 4.1 4.9 4.3 4.4 2.9 2.5 2.9 2.9 1/ Excludes local adrmnistrations. Includes capital transfers that have public investment attributes and capital expenditure--netof expropnation. 2/ Budget allacatlon. Source: SPO 76 2.40 Some o f these infrastructure investment cuts were probably overdue, and occurred in the context o f the rationalization o f the public investment portfolio. In the past, the investment program seems to have been overloaded with low-priority projects. Unclear criteria and processing rules had resulted in “over programming’ o f the PIP. As a result, the stock o f approved but unfinished projects grew to an average o f more than 5,000 during the latter h a l f o f the 1 9 9 0 ’ ~ ~ the average completion time increased to about 10 years. Many and projects received “trace” allocations, i.e. amounts nowhere near enough to implement the project, but assigned merely to keep i t in the PIP. While in 2001, one-third o f the number o f projects with trace allocations were in the health sector, agriculture and education and other social sectors also had a sizeable share in the stock o f these “inactive” projects. In terms o f the amounts o f the projects in trace status, the agriculture, energy and transportation projects comprised a majority o f all the projects with trace allocation. (a combined share o f 85 percent in 2005). Furthermore, many projects had stayed in the program, and remained unfinished, for 30 years or longer. 2.41 As a result o f the significant downsizing o f the PIP the number o f projects put on hold has increased. In 2005, 254 infrastructure projects got allocations o f less than 10 percent o f their remaining investment, putting o n hold a portfolio o f a total cost o f YTL 54 billion (approximately US$40 billion). Trace allocation projects accounted on average for about 10 percent o f total value o f projects in 2003-04, a significant rise from the 1999-2000 average o f only 3 percent. I t i s not clear whether these projects in “trace” include projects which are: (i) non-economic projects that should not have entered the portfolio in the first place, (ii) projects once considered economically justified but which are no longer aligned with the new priorities o f the Government, or (iii) strategically important projects which are being constrained by fiscal adjustments and therefore have limited access to financing. In the absence o f a sustained and in depth review o f the portfolio, i t i s difficult to ascertain the quality o f the ongoing investments. O f the current portfolio, transport, irrigation and hydropower have the highest number o f stalled or stopped projects. Projects under the regular budget thus appear to have been more prone to be started and contracted upon without certainty o f requisite allocations in the following years. As an example, in transport 40 percent o f ongoing projects got a trace allocation o f less than 10 percent o f remaining investment. 55 2.42 I n addition, there have been significant delays in the implementation o f water supply and sanitation projects. The average completion period o f water supply projects implemented was 10.4 years in 2003. The average completion period o f water supply projects was 2.9 years and average completion period o f wastewater projects was 4.8 years implemented in 2003. Long implementation periods and slow project execution reduce the economic rates-of-return o f projects as the illustrative calculation in Table 2.7 below. 55 I t would be 25 percent o f on-going projects on trace if we take the threshold as 5 percent o f the requisite follow-up allocations. 77  Table 2.7. Effect on Economic Returns from Delayed Project Completion (US$ millions) Period of Annual Present Annual Benefits Present BenefWCost Economic Execution Investment Costs Value Value Ratio Rate-of- costs Benefits Return 5 years 20 each years 1-5 75 20 each years 6- 117 1.54 14.7% 35 10 years 10 each years 1- 61 20 each years 11- 73 1.18 11.3% 10 40 Source: W o r l d Bank staff (id I n vestment needs and f i n an cing requirements 2.43 Infrastructure gaps emerge as an obstaclefor doing business. A preliminary review o f the current status o f infrastructure service provision shows that in order to enable Turkey to emulate OECD standards o f service provision, Turkey must systematically improve the quality, and perhaps also the quantity, o f infrastructure investments in the medium term. The EBRD-World Bank Business Environment and Enterprise Performance Survey (BEEPS) indicates that the percentage o f Turkish firms stating telecommunications, electricity, and transportation as a problem for doing business i s significantly higher than the average for EU8 countries and E C A countries. And although the percentage o f these perceptions has slightly declined for telecommunications and electricity between 2002 and 2005, it has increased for transportation. In the same vein, the World Bank’s Investment Climate Survey for Turkey shows that nearly 82 percent o f the 1,300 f i r m s surveyed experienced an average o f 28 power outages in 2005, leading to a loss o f more than 4 percent o f annual revenues.56 In comparison, only 26 percent o f f i r m s in Poland reported having experienced outages, at an average o f 1.5 times per firm. The survey confirms that the quality o f transport services i s poor, compared to other countries at similar levels o f development. 2.44 A project-level evaluation o f the existing portfolio will be instrumental in improving the quality of the investments. I t has not been possible, in the context o f the present study, to estimate the infrastructure needs in Turkey because o f the varying requirements for (and often, the difficulty in agreeing on) quality o f infrastructure service delivery standards across the country. Bridging infrastructure gaps will most likely require additional funding in the future. Quantity o f services will need to improve, but quality i s also an issue as infrastructure gaps may also reflect inadequate maintenance and inefficient operation. Regulatory reform in infrastructure services can play a significant part in improving efficiency (World Bank 2006, CEM). An in depth study i s urgently needed to inform policy on the key gaps and policy options to address them. 2.45 The Ninth Development Plan o f the Government focuses on improving the quality of infrastructure, ensuring supply security, and increasing competition. Improving the quality o f infrastructure in terms o f improving reliability and reducing costs, will likely 56 The Bank conducts Investment Climate Surveys (ICs) in many countries o n a regular basis, with the objective o f assessing the m a i n factors w h i c h constrain growth by manufacturing firms. The Turkey I C s covered 1,323 f m s representing nearly every industry and ownership structure. 78 require additional funding, but may also require more adequate maintenance and better operation. Especially in energy, the country is likely to face growing pressures o n supply security in the short to medium term, given the high growth in demand. The Government i s therefore planning to invest in completing ongoing projects and in new investments which can mitigate this risk. Further, the Government is also trying to enhance the incentives for private investment. 2.46 Financing requirements for the existing stock o f infrastructure projects are sizeable in themselves. In order to complete all on-going projects, Turkey would require about YTL 82 billion over the next 5 years: an annual average investment o f roughly 2.5 percent o f GNP for 2006-10. This implies that the Government would have to substantially increase i t s funding for infrastructure from the current levels. N e w investment projects also enter the portfolio every year which places an additional burden o n project financing. The existing infrastructure portfolio imposes not only a significant financing challenge but also prevents the Government from pursuing new projects. Financial commitments associated with projects already in the portfolio take most o f the annual allocation for infrastructure investment. In the year 2005 about 90 percent o f the infrastructure investment allocation was related to ongoing investments, leaving 10 percent for new projects. 2.47 Expenditure pressures in investment will only rise further as alignment with the EU accelerates. Although the government recognizes these infrastructure needs and further states in the Preliminary National Development Plan that improving infrastructure services i s not only a strategic priority unto itself but also a key element for achieving sustainable growth and better quality o f life, much needs to be done towards faster convergence with EU. As part o f i t s EU accession process, the cost for Turkey to meet the environmental standards o f the European Union, where at least 11 EU Directives in the environmental area o f water supply and wastewater apply, is estimated to be high. The corresponding investment cost estimates over the next twenty two years range between US$15-20 billion, or US$70&900 million annually. The EU-driven estimated annual investment could be as much as 0.45 percent o f the country’s GNP.57When additional investment requirements are taken in to account, for example for alignment with the environmental acquis, effectiveness o f investment spending and better project selection and monitoring will be even more urgently needed. In the medium-tern program for 2007-09, the government has committed to further reducing the average completion time, which i s one o f the available indicators o f the quality o f the PIP, from an estimated 5.5 years in 2006 to 5 years in 2009. The reduction in completion times alone does not certainly ensure the rationalization o f the PIP. Structural measures are bound to complement the rationalization o f the investment stock, by enhancing management o f the public investments. ’’The Under-secretariat o f the State Planning Organization believes that EU financial assistance will be necessary t o complement the domestic investment funding in order for the EU acquis t o be met. The SPO ’ expects that the level o f available municipal funding will be reduced due to (i) a reluctance o n part o f the municipalities to borrow for investments where in the past they have been used t o receiving de facto grants; and (iithe intention o f the Debt L a w No. 4749 t o more firmly ensure that municipalities will be forced to use a ) portion o f internal sources for repayment o f foreign loans. 79 2.48 Appropriate allocations for operation and maintenance o f public capital need to be ensured. This is an important requirement not only for improving the quality o f services, but also for restricting the cost o f the projects. Whilst the investment budget i s under SPO oversight, maintenance expenditures below a certain threshold, as defined in annual budget laws, are recorded in the current budget, in the "purchases o f goods and services" line. Total maintenance expenditures including this line item, declined from 0.5 percent o f GNP in 1999-2000 to 0.3 percent in 2003-04.58 Appropriation for maintenance expenditures declined particularly for transportation sector, health sector and other public services sector. (see figure 2.8) Such reductions in O&M expenses should be avoided since they are not sustainable while they reduce the contribution o f public capital to efficiency and Figure 2.8. Maintenance Expenditures (% o f GNP) 0.09% - "1__.__ "1 _" " "__-I- " "-" "l-.-"l" __"_ - -"-- ..__._.II_._ 0.08% 0.07% 0.06% 0.05% 0.04% 0.03% 0.02% 0.01% 0.00% Source: Ministry o f Finance, SPO, World Bank Staff calculations 2.49 Inappropriate maintenance and accelerated decay o f capital goods may unduly burden the investmentprogram in thefuture. The existing structure o f budget preparation- with separate guidelines for current and investment expenditures-poses a risk o f misallocation o f resources. Agencies may not necessarily strategically allocate resources between investing in new assets and maintenance o f existing assets. The importance o f timely maintenance i s more prominent in the infrastructure sector and thus proper accounting o f the actual amounts spent for maintenance i s needed. For example, for a road with a 20- year life, GD o f State Highways (KGM) research shows that if maintenance i s not done at the end o f the 12th year, the road would start deteriorating at a rate 8 times faster than in the first '* The total maintenance expenditure is estimated as the sum o f the small amounts o f expenditure recorded in the recurrent budget and the expenditure that i s part o f the investment budget. There is a potential drawback for using the latter series. I t i s obtained by extracting f r o m the investment program database those projects having as part o f its title or characteristics, words like maintenance, repair, renovation. However, other projects that d o involve a maintenance component are not picked up by this method if the names or characteristics specified in the database d o not involve the relevant words. General Directorate o f State Highways i s reported t o have many projects involving maintenance work. As more reliable and refined data becomes available, the data could b e updated. '' Evidence o f lower than expected capacity utilization and shorter than planed economic life o f some projects points t o problematic maintenance, especially in the road network, where indicators o f physical condition point t o deteriorating quality.. However, the relative roles o f inadequate budget allocations and lack o f appropriate incentives/culture for maintenance seems difficult to establish. 80 years o f its lifetime. This, in practical terms, means that if the cost o f capital i s the same for maintenance as i t is for investing in new roads, undertaking investment for rebuilding an existing road is up to 2.5 times more expensive than providing timely maintenance. (iii) Policy priorities over the medium term 2.50 Apart from focusing on completing key infrastructure investments as discussed above, it i s important that: (a) a policy environment conducive to greater private investment be created, and (b) the budgetary and investment planning process be improved. In order to ensure supply security and reliability, key investments should be prioritized and completed as soon as feasible. These would include important dams and irrigation investments, as also transport, railroad, water supply, wastewater and solid waste projects. At the same time, it i s imperative to take steps towards improving the investment climate for private investment. Finally, in a longer-term perspective past deficiencies in the budgetary and planning process need to be addressed. 2.51 Policy environment for private investment. Given the very large financing requirements in infrastructure, it i s unlikely that the Government will be able to achieve i t s objectives o f secure and reliable infrastructure services-a key development challenge and requirement for EU convergence-in a sustainable way without sufficient support from private investment. International experience confirms that private investment will be directed to infrastructure if the policy and regulatory environment i s robust and predictable. In the absence o f such an environment, private investment will either not materialize in adequate quantity, or will require significant guarantees and assurances. This i s borne out in the energy sector where, over the last decade, private investment n o w makes up about 50 percent o f electricity generation, but this has been made possible only by significant contingent liabilities on the Government in the form o f guarantees and o f f take agreements. 2.52 One way o f increasing operating efficiency could be to use the private sector through Public-Private Partnerships (PPP). Although there i s n o single PPP prescribed model, the development o f PPP in countries brings several benefits. PPPs are seen as a tool for increasing efficiency and for substituting private financing for government financing. International experience from PPP is that efficiency and quality o f service could be expected to rapidly improve, provided that private operators are employed under contracts with incentives for improvements and sanctions for failure to reach explicit performance targets. PPP contracts bring the additional benefits o f forcing the government contractual party to carehlly set contractual performance targets and allocate risks efficiently. In general, risks should be borne by the contractual party that is best able to control them. This implies that most risks with the exception o f the political risk should be borne by the private partner. 2.53 PPPs can be implemented well i f political and regulatory risk perceptions have been taken in account. An essential condition for implementation is to ensure that sufficient attention has been paid to policy development, involvement o f stakeholders, information sharing at all levels and maximum focus o n institutional strengthening o f the respective entities requiring training o f c i v i l servants o n the appropriate frameworks and design o f PPPs 81 as well as adequate support t o procurement capacity. The existing commercial code and the code related to corporations are sufficient to implement and manage successful PPPs in Turkey. The problem usually occurs in the ownership structure o f the PPP. The issue i s who controls the majority shares, and thus controls the decision-making process (through the PPP Board). In most PPPs, the contract design i s obviously central t o whether private management will deliver benefits. Three aspects are crucial: the need to identify the extent o f control exercised by private managers, the need for third-party contract oversight, and (perhaps most importantly) the need to link remuneration o f private managers to performance. If these needs are met, successful PPPs can support commercialization. 2.54 One means for commercialization is to introduce the private sector through management contracts or leases, ofen used when full privatization is not feasible. Under a management contract, a public authority makes a private company responsible for managing and delivering a service, typically for a period o f three to five years. The public authority retains financial responsibility for the service, thus limiting the r i s k for the contractor. Remuneration for the contractor i s in the form o f a flat fee with the possibility o f additional performance-related payments. Using management contracts in utility commercialization is a relatively new phenomenon and some evidence i s emerging about the types o f benefits and problems that might ensue. Results have been varied with management contracts. Well- structured management contracts could be a useful interim solution, both to commercialize the utility and to wait for the investment climate to improve. There are also lease-type arrangements, to the extent they can be meaningfully differentiated from management contracts which can also be considered if a country i s interested. 2.55 Based on assessments o f infrastructure and the country's investments climate, there is significant scope for PPP for all o f the infrastructure sub sectors - energy, water and wastewater companies and rail freight companies in Turkey. I n k e y infrastructure service areas, Turkey has made substantial progress in designing the changes to legal and regulatory structures needed to meet EU acquis requirements. Even though much implementation work remains, Turkey i s fully aware o f i t s internal market liberalization challenges and fully understands the key role o f integrating with international markets. A s a result, Turkey should be attractive and able to successfully implement many PPP models within both infrastructure and other sectors as well. 2.56 Price signals and competitive markets are needed. In the electricity sector, the Government i s in the process o f establishing a competitive market which will enable suppliers and buyers o f electricity to participate efficiently and at economic prices. I n order for the market to function successfully, and in order for new investment to flow into the sector however, the Government will have to ensure that wholesale electricity prices reflect the true marginal costs o f building new generation capacity. At present, despite the existence o f an independent regulatory authority, in order to keep inflationary pressures down, the Government artificially keeps electricity prices l o w by reducing electricity generation using natural gas (the price o f which i s very high due to record high international o i l prices) and using its substantial hydroelectric resources instead (hydroelectricity i s not priced at the true opportunity cost o f the use o f water). This has the impact o f keeping the wholesale market price at a low level, which provides a disincentive to potential investors in electricity 82 generation. If the Government were to let the regulatory framework operate independently, and were to allow prices to reflect economics, i t would be possible to attract private investment without significant guarantees. 2.57 I n a longer-term perspective, the Government is conscious o f the need to ensure sustainability in medium-term infrastructure investment allocations by addressing the past deficiencies of the budgetary and investment planning process. I nthe recent past, volatility o f annual investment allocations has hampered implementation in line with strategic priorities, and has increased total costs and average completion times o f projects. A prerequisite to achieving stable and foreseeable annual investment allocations is macroeconomic stability and sustained fiscal discipline. But, in addition, Turkey further needs to effectively utilize the newly introduced medium-term expenditure framework. The replacement o f the annual process o f budgetary allocations by a 3-year process will help smooth out cash flows for multi-year projects. However, additional measures are likely to be required for ensuring that economically justified projects are completed in time. 2.58 The quality o f the infrastructure portfolio needs to be assessed in order to determine which projects are worth pursuing further, dropped, or redefined in their scopes. This would be important in view o f the need to re-prioritize the investment program in line with requirements for EU accession. A project-level evaluation should be conducted, together with a revision o f the rationalization policies and criteria, while establishing a proper prioritization o f investments aligned with the Government’s development objectives. Projects that are considered economically unviable due to cost escalation, changing government priorities, or even an initial deficient project appraisal, should be individually assessed and either redefined or dropped as the case may be. The Government should not rule out dropping such projects assuming that the financial penalties, if any, are justifiable. This option can be economically more beneficial in the long run than (i) keeping the project in an endless trace status and (ii) constraining the fiscal room o f infrastructure allocations available for new and high-return projects. I n summary, i t would be useful to develop more realistic financing plans before embarking on construction in order to reduce the historical long implementation periods. Proper economic and financial appraisals o f projects and improved incentives and pricinglsubsidy policies are necessary for enhancing efficiency. The training o f personnel is also fundamental to promoting efficiency. I n some instances, the focus should be o n utilizing existing assets optimally rather than adding new assets. 2.59 The absence o f well defined allocations for operation and maintenance expenditure in the investment budget is aproblem. It i s not clear if, and if so at which stage, there i s a strategic link between the funding and planning for investment and operation and maintenance (O&M). There is a clear split o f responsibilities between SPO and Ministry o f Finance during the Budget process for investment and O&M. Since there i s n o streamlined procedure, new investments might be taken without provisioning for the necessary maintenance streams and investment funds are increasingly sought for maintenance, therefore, likely misclassified. O&M requirements under the investment budget further reduce the headroom for new or ongoing investment. 83 2.60 Budget allocation decisions should involve operating agencies more than they do currently. Final decisions on investment allocations do not involve operating units. As such, the whole process encourages fragmentation between coordination (central units) and management (business-operating units). For instance, budget allocations are significantly smaller than investment proposals made by operating units, suggesting potential room for strengthening coordination communication among central agencies (Treasury, SPO and MoF) and line agencies. Unrealistic expectations should be curbed at the early stage o f the process when investment resource requests are invited from public agencies. Progress is noted with the issuance o f budget preparation notices in which priorities are mentioned more clearly than in the past. This has contributed to improving the realism o f proposals, but further efforts would be appropriate t o protect priority projects and help strengthen the overall credibility o f the budget process. 2.61 Overall, the major challenge for the Government is to create fiscal space for public investment while preserving fiscal discipline. As stated earlier, in order to complete all on- going projects in infrastructure over the next five years the Government would need an additional 2.5 percent o f GDP annually in infrastructure investments. Due to the overarching requirement to preserve fiscal sustainability, the level o f investments that can be made is constrained not so much by the availability o f financing but by the need to meet the primary balance targets in the face o f strong expenditure pressures in some areas-especially in pensions and health care. There are cases where ongoing projects have financing sources, but these cannot be used due to the l o w allocations provided to the projects. As explained in chapter 1, these l o w investment allocations have served so far to mainly offset a soaring deficit o f the social protection system that reflects slow or pending reform. The challenge for the Government i s thus not the availability o f this extent o f financing for infrastructure, but the ability to use the financing that i s already available. Creating the needed fiscal space, through the structural expenditure reforms reviewed in chapter 3, would thus be the k e y requirement for stepped up investment in infrastructure in the future. W e l l designed Private- Public Partnerships for the provision o f infrastructure services will also help alleviate existing constraints. 84 D. ESTABLISHING TAX EXPENDITURE ACCOUNTS-TOWARDS BETTER FISCAL ACCOUNTABILITY AND TRANSPARENCY 2.62 Establishing an adequate tax expenditure framework would promote fiscal accountability and transparency. Inadequate treatment o f tax expenditures is a concern for fiscal accountability and the transparency o f governance. A basic requirement o f a public sector budget is comprehensiveness, that is, the budget should cover all resources available to the government and expenditure allocations to all Government functions. Tax expenditures often represent the equivalent o f a major share o f the discretionary expenditures o f a government. Without the inclusion o f tax expenditures in the budget process, the tradeoffs in allocations across functions, sectors, regions and other target groups are limited to expenditures financed by the net revenues. Such net revenues are already reduced by the tax revenues forgone as a result o f a range o f incentives and preferences offered to taxpayers through the provisions o f the tax laws. Without their inclusion in the budget process, tax expenditures do not get the same level o f strict budget scrutiny as normal budget expenditures. Accordingly, to make the budget process more complete and subject all expenditures to the same level o f transparency and budget scrutiny, governments in OECD countries over recent decades have been estimating the revenue cost o f tax expenditures, publishing tax expenditure accounts and including estimated expenditures in the budget processes. (i) Tax Benchmarks and Tax Expenditures 2.63 Estimating tax expenditures requires establishing tax benchmarks. Tax expenditures, in broad terms, are tax provisions that deviate from a normative or benchmark tax system. Tax expenditures may take a number o f forms: exclusions, exemptions, allowance, deductions, credits, preferential tax rates, or tax deferrals. In order to identify tax expenditures, a normative, or a benchmark tax structure, has to be established. The normative or benchmark tax structure does not contain any tax provisions, which are used to implement government spending programs for favored activities and groups. The tax benchmark reflects tax decisions that take into account considerations o f the ability o f persons to pay the tax, and the economic, administrative and compliance costs o f the tax. 2.64 Establishing tax benchmarks in Turkey i s not as straightforward as in many OECD countries that have used the standard guidance (Box 2.1). Turkey’s tax bases and economic structures have led to tax structures specific to its circumstances that differ from the average country in the OECD. Turkey has a high share o f employment in the agricultural sector (about 29.5 percent as indicated in the Ninth Development Plan); and because o f extensive informality in agriculture and services, i t has a large share employed in the informal and unrecorded sectors (about 53 percent o f total employment). There i s also a large share o f employment at the minimum wage. Data on employment earnings from the social insurance system for 2003, for example, show that 59 percent o f employees in the private sector were reported to receive minimum wages while 4 percent o f public sector workers received minimum wages. Even if there i s significant underreporting o f employment earnings for social insurance purposes, it would s t i l l leave a significant share o f low-wage workers. 85  Box 2.1: Guidelines to Identify Normative T a x Structure Identification o f the normative tax structure can best proceed by determining whether a given provision responds to one o f the following questions: 1) I s the provision necessary t o determine the base o f the tax, normatively defined, in accordance with the fundamental nature o f the tax? 2) I s the p,rovision part o f the generally applicable rate structure? 3) I s the provision necessary t o define the taxable units liable for the tax? 4) I s the provision necessary t o assure that the tax i s determined within the time period selected for imposition o f the tax? 5) I s the provision necessary to implement the tax in international transactions? 6) I s the provision necessary t o adrmnister the tax? Seven developed countries (Canada, Germany, France, the Netherlands, Sweden, United K i n g d o m and United States) use these broad guidelines t o identify their normative tax structures and tax expenditures. Source: . McDaniel, P.R, and S. Surrey, “International Aspects o f Tax Expenditures: A comparative Study,” Kluwer, 1984, and Messere, K e n C. “Tax Policy i n OECD Countries, Choices and Conflicts,” IBFD Publications BV, Amsterdam, the Netherlands, 1993 2.65 The peculiar economic structure, together with the past economic instability, has led to specific choices o f tax structures in Turkey. These include: (i)reliance o n indirect taxation through the VAT (with limited rehnds o f input taxes) and Special Consumption Taxes (excise taxes) in contrast to most high-income OECD countries that receive a majority o f tax revenues from direct taxes o n income and payrolls; (ii) significant reliance o n withholding taxes in the income tax o n financial instrument income, , s e l f employment gross earnings, and wages and salaries, includingfinal withholding taxation o f interest deposits and employment income; (iii)heavy reliance o n employers as major tax agents for administering the taxes o f employees including PayAsYouEarn income taxes; (iv)special credits for expenditures o n health, education, basic food and rent, social security premiums, and pension contribution deductions; (v) exemption o f many self-employed, rural, household and agricultural workers based o n occupations rather than levels o f income; (vi) a complex set of inflation adjustments, investment incentives and offsetting taxes o n these adjustments and incentives that reduced the tax burden o n investors in real and financial assets. 2.66 This complex structure has supported revenue collections over recent years, but it has left the authorities in a weak position to analyze the tax system and its impacts on economicperformance. This lack o f information about the tax bases weakens the ability o f the authorities to formulate tax and economic policy, and, in the context o f tax expenditures, creates challenges for defining tax benchmarks and estimating the tax expenditures arising 86 from the various tax incentives and the tax assistance provided to various disadvantaged groups. 2.67 Aware o f the challenges, the Government has initiated work to analyze and estimate tax expenditures. The authorities, with support from the World Bank, have initiated a review o f the tax provisions o f the personal income tax (PIT), corporate income tax (CIT), value-added tax (VAT) and special consumption tax (SCT). Some 290 tax provisions (articles and items) were identified by the authorities in these legislations requiring consideration as to whether they should form part o f the tax benchmark or be regarded as tax expenditures. The major features o f the tax benchmarks for each o f the tax types and the draft tax benchmarks are presented in Annex V. 2.68 Based on the draft tax benchmarks, out o f the 290 taxprovisions, initially some 186 items were identified as potential tax expenditures. Currently, the authorities have identified a list o f 84 tax provisions as tax expenditures and included this l i s t in the 2006 Budget Law. The most significant o f those tax expenditures can be divided into seven different categories: (1) Free zones; (2) Investment incentives; ( 3 ) Financial instruments; (4) Regional employment incentives; (5) Agricultural sector support; (6) Social insurance and pension contributions; and (7) Assistance to low-income and disadvantaged persons. These categories are not exhaustive and sometimes overlap. 2.69 Tax expenditures are primarily delivered through the Personal and Corporate Income Taxes. By type o f tax, the l i s t o f tax expenditures is outlined in B o x 2.2. Some o f the major tax expenditures are granted with the aim o f promoting: investment, employment, regional development, R&D, and trade incentives primarily delivered through the CIT. A range o f tax expenditures aims to assist low-income individuals, in support o f social policy. These are delivered primarily through the PIT and C I T and to a lesser extent through the VAT (see B o x 2.2). Some major tax expenditures in this latter group include the deductibility o f contributions to the social security system for pensions, disability, unemployment and health benefits as w e l l as contributions to private pension and insurance company-based pension and health policies. Box 2.2: T a x expenditures by t w e of tax I. Personal and corporate income tax 0 Assistance to disadvantaged and low-income persons T a x credit for basic needs expenditures (VAT-invoiced expenditures o n education, health, basic foods, clothing and rent) by employees; Exemption o f income small self-employed; professional income o f writers, translators, sculptors, painters, computer programmers, composers and inventors; non-diplomatic staff o f embassies; wages o f workers in small villages; wages o f farm workers; wages paid to students, convicted and detainees and indigent people working at workshops o f school o f art and similar institutes, correctional institutions and reformatories, and alms’ houses; Exemption for income earned by new educational and training schools over first 5 years 0 Preferential treatment of certain employee benefts of civil servants and other workers Exemption for aid for heating purchases by c i v i l servants 87 Exemption for child allowances, marriage and nuptial supports, severance payments (up to 24 months pay) paid to civil servants and workers Exemption for allowances and expenses paid to civil servants and other workers undergoing education and training programs Exemption for wages o f apprentices 0 Social insurance and pension contributions, investment income and benefits Deductibility o f contributions to social insurance, pension and unemployment schemes Deductibility o f employer sponsored contributions to state social insurance and private insurance health schemes.60 For individually sponsored tax assisted pension savings, tax deductible contributions are limited to 10% of earnings up to a maximum contribution o f the annual minimum wage. In addition, tax-deductible premiums are permitted to individual purchases o f non-pension non-property insurance (life, disability, health, etc insurance) up to 5% o f earnings. Tax treatment o f private pension savings: exemption o f 25% o f pension income from pension savings investment for 10 or more years Exemption o f social security pensions, death, disability, illness and unemployment compensation and aid. Exempt investment income in pension plans, except that a 10% capital gains tax i s charged o n gains in employer-sponsored pension funds. For individually and state sponsored tax assisted pension savings, all investment income and gains are tax exempt. Exemptions for charitable donations and sponsorship o f sportsmen and donations to Izmir University Games Preparation and Regulation Board Exemption o f enterprises o f sub-national government in agricultural, water, electricity, transportation, cold storage and slaughterhouse businesses Exemption o f organizations involved in development o f industrial zones. 0 Preferential treatment o f income on selectedfinancial instruments Exemptions for government bond income up to TL 50 billion and certain mutual fund income through end of 2004. Exemption (except for withholding taxes) up to global limit o f YTL 12,000 for total declared investment income from: (i)Interest on deposits and repos (ii) Inflation-adjusted TL bond income ( iii)FX bond income (iv) TL bond income, including government bonds (v) Mutual fund income (vi) Capital gains on shares o f listed companies held for less than 3 months and amounting to more than YTL 10,000 (vii) Capital gains o n unlisted shares held for less than one year. (viii) Inflation-adjusted capital gains on immovable property. Real estate held for more than 4 years i s exempt. 0 Investment incentives 40% tax allowance on depreciable assets in addition to regular depreciation allowance; indexed and indefinite carry forward o f unused allowances, but as o f 2006 carry-forwards o f allowances that cannot be deducted w i l l be limited to 3 years. 40% tax allowance o n R&D expenditures i n new technology and know-how 6oState-sponsored social security pensions require contributions o f 11% from the employer and 9% from the employee aside further contributions to health, maternity and disability insurance. These insurance contributions are subject to a maximum pensionable or insured earnings limit based on the salary o f a senior civil servant. Employees o f banks, police, military etc have access to employer-sponsored tax assisted pension savings managed by some 21 foundations. The tax-deductible contributions are subject to the same limits as social security contributions. Employer contributions to group life, health and pension plans are tax deductible up to the social security limits. 88 Regional employment and investment incentives in regions below $1,500 per capita 5 year tax holidays for companies increasing employment by at least 10 employees till 2007 Tax exemption o f up t o 100% f r o m PIT and Social Security premiums o n minimum wage for additional workers 0 Technology Development Regions Tax exemption f r o m income f r o m R&D activities o f businesses in region for 5 (extendable t o 10) years Employment tax exemption for R&D employees 0 Free Trade Zones (wage andprofits taxes) Prior t o 2004, tax holiday for corporate and wage income taxes Since 2004, phasing out o f o l d licenses by 2008 and only granting CIT exemption t o income f r o m export production. Earnings received f r o m construction, repair, assembly works camed out and technical services rendered abroad and transferred t o general accounts in Turkey 11. VAT Full exceptions o f specified mining and petroleum exploration3 Partial exemptions o f deliveries, operation, construction and servicing o f various craft; supplies and services in the construction o f ports; servicing o f craft at ports3 Partial exemptions o f imports for military purposes; presidential use; for donations for health, education and social purposes3 1% VAT rate and 8% rate o n items other than those eligible for reduced rates according t o EU VAT directives(EU Directive o n VAT 77/388/EEC, Annex H), including textiles, clothing and leather articles rebated to 8% Exempt r a w minerals Exempt services provided in a Free Trade Zone. Exempt sale o f assets in case o f winding up o f financial institutions 111. Special Consumption Tax Exemptions for fuel and weapon supplies to security and intelligence agency Exemption for fuel supplies for petroleum exploration Exemption for vehicles for disabled Exemption for airplanes and helicopters for Turkish Air Foundation Exemption o f imports for military purposes; presidential use; for donations for health, education and social purposes Note: 1. In addition to tax expenditure estimates, it i s also useful to provide tax revenue change estimates as memorandum items for various structural items even where considered part o f tax benchmark. Examples o f memorandum items that are useful t o present for information purposes would be the tax revenue losses f r o m a l l reduced VAT rates, and dividend deductions. 2. Given significant changes in tax policy starting in 2001, the tax expenditures noted above focus o n the current tax expenditures, and omit tax expenditures removed o r phased out in recent years. 3. “Full exceptions” are those for which input VAT i s creditable or refundable (also referred t o as “zero VAT rating” in some VAT structures); “Partial exemptions” are those for which input VAT is not creditable (the equivalent o f “VAT exempt items” in other VAT structures). Source: W o r l d Bank Staff 89 (io The Size of Tax Expenditures 2.70 The estimation o f tax expenditures is still in its early stage in Turkey. The estimation o f tax expenditures i s highly demanding, more so than revenue forecasting. Estimating the incidence o f tax expenditures on tax revenues requires detailed analysis o f the individual tax measures. This requires detailed information on the tax base, including that related to exempt persons and activities. Existing tax databases are s t i l l very incomplete because o f the limited tax return information. The current status o f tax databases and models i s described in Annex VI. The starting point for tax expenditure accounts are the estimates for the actual tax expenditures in past years which then form the basis o f forecasting future tax expenditures to be incorporated into the medium-term budget process. Databases and tax simulation models for 2003 are used as the starting point o f assessing the size o f tax expenditures. 2.71 Authorities estimated tax expenditures for 15 out o f 84 tax expenditure provisions for 2006-08 as part o f the Budget Law largely based on 2003 databases and models. The estimates for these 15 tax expenditure provisions are presented in Panel A o f Table 2.8. The combined effect o f these tax expenditures in 2003 is a revenue loss o f 1.5 percent o f GDP. The largest tax expenditure, at 0.66 percent o f GDP, arises from the investment tax allowance, and the second largest, at 0.33 percent o f GDP, i s from the PIT tax credit for personal expenditures on basic needs items. The PIT and C I T tax expenditures are based on micro-simulation models which capture the distribution o f marginal tax rates across individuals and the partial value o f deductions and exemptions arising from a lack o f taxable income due to losses in the year or carry forwards o f losses from prior years as discussed in detail in Annex VIS6* 2.72 Additional tax expenditure estimates can be approximated for other items in the summary CIT and PIT tax returns for 2003 and for deductions in social security contributions. Panel B contains estimates o f the “other deductions” taken under the C I T and the revenue cost o f deductions o f social security contributions from taxable income in 2003.62 In addition, i t is recognized that loss carry forwards arise both from adverse market circumstances and from unused tax deductions or exemptions in prior years. The elimination o f tax expenditures, therefore, would result in lower loss carry forwards into 2003, and hence, higher revenue collections. Here i t is assumed that 25 percent o f the loss carry forwards are caused by unused preferences in prior years that would result tax increases in 2003 to gain a rough approximation o f the tax cost. The need for further work to estimate the impact o f prior year tax preferences on current year tax expenditures i s discussed in Annex VI. Panel B tax expenditure estimates add a further 1.66 percent o f GDP to the total tax expenditures in 2003. In fact, comparison o f these estimates, which reflect partial capture o f the tax value o f deductions and exemptions, with those that w o u l d arise using the data from the summary company tax returns for 2003 show the C I T tax expenditures are some 28 percent smaller. 62 “Other deductions” include a l l the deductions and exemptions allowed that are not specifically itemized o n the tax return. These include the deductions allowed under Article 14 for donations o r aid provided t o educational institutions for health facilities, or t o support sports, culture and national heritage. 90  Table 2.8: Selected tax expenditures for Turkey in 2003 YTL millions % GDP C.T.L. Art. 814 Exemptions for certain investment income earned by corporations 935.03 0.26% from mutual funds. Art. 815 Exemption for sale of preference nghts certificates and issuance 20.83 0.01% premiums Art. 817 Exemption for gains earned in construction, repair, installation 71.04 0.02% works and technical services in other counmes and transferred to general accounts in Turkey. Art. 8/12 Exception for gains arising from the sale of participation shares 259.73 0.07% and real estates that companies have had under their assets and decide to add to their capital (revised form o f C.T.L. Provisional Article 28) Art. 19 of I T Law Investment tax allowance 2,367.56 0.66% 3218 S.K. Art. 3 Exemptions provided under the Free Trade Zones Law no.3218 159.43 0.04% 4691 S.K. Art. 8 Exemption for income in the Technology Development Regions 54.00 0.02% I.T.L. Art. 8911 Deduction for individual insurance premiums and contributions to 8.48 0.00% individual pension system I.T.L. Art. 8914 Deduchon for chantable donations and grants. 2.77 0.00% I.T.L. Art. 121-bis Tax credit for expenditures on education, health, food, rent, and 1201.oo 0.33% clothing of the taxpayer, spouse and children provided against taxes on wage income 4325 S.K. Art. 3 Exemptions provided for income for employers that employ ten or 30.27 0.01% more workers in the provinces indicated in the Law Value Added Tax L a w No. 3065 VAT Art. 13 Exception for vehicles, search of precious mines and petroleum 166.13 0.05% and national security expenditures and investments (subparagraph (d) covered under the basic taxing system) VAT Provisional Art. 15 Exception for construction contracts made w t h housing 123.72 0.03% construction cooperatives and social secunty organizations established by law exclusive to housing units not exceeding 150m2 and construction contracts made to municipalities regarding the constructions for which building construction licenses were obtained pnor to 29/7/1998 I Total A 5,400 1.51% Other CIT deductions 1,773.86 0.49% Deduction of social secunty contributions for pensions, 3,300.00 0.92% unemployment insurance and health benefits Preferences in loss carry forwards absorbed in year in CIT and PIT 915.29 0.25% Total B 5,989.15 1.66% 91  Exemption for aid for fuel purchases by civil servants and certain 0.50% self employed workers; exemption for child allowances, marnage and nuptial supports, severance payments (up to 24 months pay) paid to civil servants and workers; exemption for allowances and expenses paid to civil servants and other workers undergoing education and training programs; exemption for wages o f I apprentices I Total A + B + C 3.67% Notes: In Panel A, PIT and CIT are estimates by authonties, and VAT estimates are based on estimates by authonhes for 2006 adjusted for growth in nominal GDP. In Panel B, bank staff estimates are based on summary tax return data for 2003 assuming 82% absorption of exemptions and deductions, and 25% o f loss cany forwards being tax preference carry forwards. Social secunty data base used to o b h n effective marginal tax rate for social security deductions. Panel C based on IMF and Bank staff estimates 2.73 Including part o f other omitted tax expenditures would raise estimated tax expenditures to at least 5% of GDP, or about 18% of government revenues in 2003. Panel C in Table 2.8 estimates the tax expenditure from a number o f preferences for employee benefits for civil servants and other workers. The combined estimate o f tax expenditures in Table 2.8 for 2003 amounts to 3.67 percent o f GDP. W h i l e most o f the major tax expenditures are included, no estimates are included o n tax expenditures for exemptions o f income from a wide range o f trade, service and farm activities, as well as most o f the VAT and SCT tax expenditures. Tax expenditures in OECD countries have fallen in the range o f 4 percent to 35 percent o f gross expenditure^.^^ I t appears that a comprehensive assessment o f tax expenditures would place Turkey towards the high end o f the OECD country range. (iii) Refining the framework for the analysis of tax expenditures-key steps ahead 2.74 A major effort i s required to upgrade the tax databases and related tax simulation models to obtain more accurate and reliable tax expenditure estimates. Annex V I details the current status o f the tax database and models by tax type and the required work to upgrade them for reliable tax expenditure estimates. 2.75 The following are the key steps to develop and implement a full framework for publishing tax expenditure accounts and incorporating tax expenditure estimates into the annual budget appropriations and the medium-term budgetingframework to achieve fiscal accountability. Work on steps a through d i s largely complete. a. Development o f benchmark taxes for each tax type and identification o f tax expenditures in each tax type b. Sectorhnctional classification o f tax expenditures 63 Zicheng Li Swift, “Managing the Effects o f Tax Expenditures o n National Budgets,” Tax Notes International, 41(10), M a r c h 13, 2006 reports tax expenditures as shares o f gross expenditures for Canada, Netherlands, US, and UK in Table 1. 92  c. Estimation o f some major tax expenditures as illustrations o f size and importance o f tax expenditures. d. Development o f tax expenditure discussion documents e. Development o f tax expenditure estimation methodologies for all tax types and tax expenditures f. Choice o f year to start publication o f tax expenditure accounts g. Development o f tax expenditure projection procedures for multi-year forward budgets h. Discussion and decisions about how tax expenditures will enter budgeting system i.Integration o f tax expenditures into comprehensive budget allocation process 2.76 Tax expenditures should become an integral part o f governmental spending programs. To achieve this goal, a major and sustained work effort is required over the medium term. The above analysis provides a road map for establishing the tax benchmarks and identifying tax expenditures, for establishing “reasonable” databases and “good” databases and the related tax simulation models for estimating tax expenditures. The analysis Wher outlines the steps in development o f tax expenditure accounts and integration o f accounts into budget system in order to achieve fiscal accountability in such spending. In addition, tax expenditures have to be evaluated and audited for performance and procedural compliance. Tax expenditure accounts and their evaluations and audits should be published to achieve fiscal transparency. 2.77 A tax expenditure framework, similar to what i s described above, is widely used in P the EU forfiscal accountability and trans arency. In particular, the EU has adopted “State Aid Rules”64 and a “Code o f Conduct”6 for i t s member states. The “State Aid Rules” restrict or prohibit state assistance to industry, the scope o f which is broad enough to cover many types o f tax expenditures. The “Code o f Conduct” requires member states to eliminate 66 identified special tax incentives in order to refrain from certain types o f tax competition. Elaborating a hll-fledged tax expenditure framework would help Turkey conform to the EU’s “State Aid Rules” and “Code o f Conduct” as part o f the EU accession process. 64 EC Treaty, A r t s . 87-89. See Schon, “Taxation and State Aid Law in the European Union”, 36 Common Market L. Rev. 91 1 (1999) 65 Communication from the Commission to the Council: Towards Tax Co-ordination in the European Union. A Package to Tackle HarmfUl Tax Competition, Doc. C O M (97) 495 final. 93  CHAPTER I11 OPTIONS FOR EFFICIENCY GAINS IN PUBLIC EXPENDITURES-SECTOR ISSUES 3.1 Structural public expenditure reforms in key sectors would be the main option to create theJiscal space needed to meet Turkey’s development challenges. As public expenditures are already high compared to other emerging economies and some new EU members, there i s little room for Turkey to hrther increase overall expenditure in order to meet pressing development challenges. Policy would thus need to focus on reforms that improve the efficiency and productivity o f expenditure programs in areas where expenditure pressures are being felt (such as in health care and social protection), or are likely to be felt in the future as a result o f the need to fill Turkey’s gaps in key areas for growth (education). Greater emphasis should be placed on areas in which Turkey is spending large amounts and getting poor results in international comparison. Structural expenditure reforms in key sectors (such as health care and the pension system) would free up resources over the medium term, which would be reallocated from low-value to high-value and growth-promoting programs. Structural reforms in education would improve efficiency, and thus help Turkey get better results for given resources. To maximize fiscal space, sector-specific reforms should be combined with the horizontal reforms discussed in the previous chapter-focused on the modernization o f civil service pay and employment system, the rationalization o f the investment program, and the integration o f tax expenditures in the budget process. At the same time, possible trade- offs in expenditure allocations would also merit particular attention, by reducing spending in functional areas where i t appears to be oversized in international comparison (such as general public services and defense, public order and safety), so as to create fiscal space for growth-enhancing programs. 3.2 The focus o f this chapter i s on structural expenditure reforms in education, social protection (pensions), and health care. These three sectors accounted together for 58.5 percent o f primary expenditures o f the Consolidated General Government (CGG) in 2004, the equivalent o f 19.2 percent o f GDP. The chapter also reviews rural public expenditures, which i s the most important component o f public expenditures on economic affairs, accounting for almost 2 percent o f GDP. These four sectors together accounted for about 65 percent o f primary C G G expenditures in 2004. A. PUBLIC EXPENDITURE IN THE EDUCATION SECTOR 3.3 Turkey has remarkably improved its education system in recent years, particularly at the primary school level. In just eight years, between 1997 and 2005, Turkey increased the enrollment o f 7-14 year-olds from 80 percent to 90 percent. The 94 Government has been making tremendous efforts to provide good-quality schools for the poor, especially by means o f additional investments to establish more primary schools in rural and marginalized urban areas. The World Bank’s 2005 Education Sector Study reports that the equity o f public spending on primary education has also improved significantly since 1997. A new curriculum under development by the Ministry o f National Education’s Board o f Education represents a great leap forward for comprehensively updating teaching standards. Government initiatives have succeeded in bringing significant amounts o f private sector funds into the sector, especially for the expensive physical infrastructure investments. Also, Turkey has increasingly been engaging in international education policy dialogue forums. 3.4 Despite this progress, Turkey will have to make carefully targeted investments over the next few years to foster the competitiveness of its labor force. Employers both in Turkey and in the global economy are demanding that high school graduates are not only able to demonstrate advanced technical competences but also that they are able to think critically, reason and ask questions, solve problems, and show proficiency in a foreign language. The European Union has recognized these changing demands and in 2005 adopted the “Guidelines on Employment Policy” aimed at upgrading the education systems o f EU members. EU members increasingly understand that economic growth depends o n having higher skills in the workforce and that improving education levels i s the best way to acquire these skills. The EU has thus called on i t s member countries to ensure that their school systems set and meet high-quality learning standards, increase education opportunities at all levels, allow youth to pursue flexible learning paths, and make sure the education system is meeting the needs o f the labor market. A1.Educational Outcomes66 3.5 Despite improvements in the area o f primary education, there is an important gap between Turkey and the EU in the proportion o f young people that have a complete upper secondary school education. W h i l e there has been important progress during the last five years on this indicator, marked by a 12 percentage point gain, only 39 percent o f i t s 20-24 year old population in 2005 had a secondary school diploma, compared with 78 percent o f this population in the European Union (Figure 3.1). The EU’s “Lisbon Agenda” target o f 85 percent o f 22-year-olds with a complete secondary education i s a long way off. Based on rates o f historical progress, i t will take Turkey several decades to catch up with today’s EU levels, which are o f course themselves climbing upwards. In addition, in several regions o f the country, Turkey’s outcomes are considerably worse. For example, in Diyarbakir, a Southeastern province, 28 percent o f 20-24 year olds have a secondary education diploma, and only 17 percent o f girls in this age group do. Results in Diyarbakir are similar to those in some o f the other Eastern and Southeastern Provinces. 66 The W o r l d Bank report “Education Sector Study: Sustainable Pathways to an Effective, Equitable, and Efficient Education System,” published in 2005, provides in depth information o n the education system, its performance and outcomes, and recommendations for improvements. 95  - Figure 3.1: Percent o f 20-24 year-olds with a secondary education diploma 90% N 15 Total 80% - N15 Female 70% -Ankara Total "" ," - -* -Ankara Female 50% -Turkey Total L. Turkey - 40% +Turkey Female (total) - - X- . Kars Total 30% 20% no/, I - Kars Females Ckyarbakir Total I i --C- Ckyarbakir Female I"'" 1995 2000 2005 Source: Ministry of National Education 3.6 Improving student learning achievements is an additional challenge for Turkey. Recent international assessments have ranked Turkey well behind most o f i t s counterparts on student learning achievement as well.67 The PISA study published in December 2004 shows that 52.3 percent o f 15 year-olds in Turkey cannot achieve beyond level 1 on a six point scale o f learning proficiency (Figure 3.2). This i s compared to only 16.6 percent o f students in EU countries who perform at this proficiency standard. Student achievement results based on other international comparative assessments are similar. Although Turkey's results are consistent with the performance o f other countries at similar per capita income, it nonetheless means that Turkey's labor force is not competitive with Europe's. In other words, Turkey's schools are failing to adequately prepare the majority o f i t s students at a time when the labor markets in Europe and Turkey are demanding increasingly better qualified graduates. '' For example, Trends in International Mathematics and Science (TIMSS), Progress in International Readin'g Literacy Study (PIRLS) and, more recently, the International Student Assessment Program (PISA). One o f the most reliable international measures used by many advanced and middle-income countries to assess the competencies o f the emerging labor force is the OECD's P I S A (Program o f International Student Assessment). The P I S A instrument measures what 15-year-olds (students near the end o f their compulsory schooling) k n o w and are able t o do. 96  Figure 3.2. Distribution o f Students by P I S A Proficiency Level in T u r k e y a n d the EU, 2003 30 20 c S a, 10 Turkey u 3 5 u- New Members 0 c S a, =Members e $ 0 OECD ,00 1,OO 2,OO 3,OO 4,OO 5,OO 6,OO Proficiency Source: PISA, 2003. 3.7 There are a number o f reasons for the still (relative to OECD) poor performance o f such a large share o f Turkey's students. Weak incentives and organizational deficiencies: Compared with the education systems in most o f the countries in the world, Turkey's education system i s highly centralized and i t s schools are given very little autonomy over their resources, staff deployment, textbook selection, allocation o f instructional time, and selection o f programs offered. Without autonomy, schools cannot be held accountable for their results and they do not have any incentives to improve their quality. At the same time, the central authority, which controls and determines the allocation o f nearly a l l o f the financial and human resources in the sector, has not succeeded in ensuring equity across schools, a proper function o f the center. Social and geographical educational disparities: Despite Turkey's educational progress since 1997, significant disparities in educational quality between social and economic classes and geographic locations persist. For example, in 2004, the average expenditure (from the Ministry o f National Education's allocated budget) per primary student nation-wide was approximately YTL 1,250 (approximately US$925). In some provinces, however, per-student expenditure was only about half that amount (see also below). The Education Sector Study (WB, 2005) reports that children from poor families are enrolled in schools that have the lowest level o f resources and the least experienced teachers. Nonetheless, there i s 97  no explicit strategy at the central level to improve poorly performing schools or help poorly performing students. 0 Poor teacher training and support: The education system in Turkey does not adequately prepare or systematically support teachers to assure that they are able to educate their students at the levels needed to develop the kinds o f learning outcomes needed by today's graduates. The Education Sector Study states that teacher preparation and professional support programs are out o f date, disjointed, poorly evaluated, discontinuous, and are not based on best global practice and evaluated experience. 0 Inefficient exam screening: A substantial portion o f students' time and effort is absorbed by preparation for the university entrance examination (and an earlier secondary school selection examination) rather than in school learning, and that this tendency is reinforced by social, economic, and family incentives. Because o f the structure o f the examination-a relatively short test composed only o f multiple choice items, and its limited curriculum coverage-it tends to discourage the learning o f the advanced competencies described above. This creates a misalignment between what the new curriculum and labor market expect students to learn, and what students are motivated to learn. Given that this examination has such an overriding influence on university acceptance and placement, it has generated a sizeable exam preparation industry that absorbs vast amounts o f private resources and provides little value in terms o f raising human capital and the educational qualifications o f the future work force. A2.Structure o f Education Expenditures in Turkey (a) Overall Trends in Public Expenditure on Education 3.8 Public education expenditures have been on a downward trend since the 1999 crisis. Public education expenditures have been, on average, at approximately four percent o f GNP over the past decade (Table 3.1), although they peaked in 1999 at 4.48 percent, and have been decreasing somewhat since then due to fiscal pressures.68 The Government estimates that 4.17 percent o f GNP will be spent on education in the 2006 budget-still less than 4.33 percent o f GNP spent in 2002. The reduction in public General budget allocations t o Ministry o f National Education (MONE) have been adjusted t o address misclassification o f expenditures due t o problems in the coding system. For example, adjustments were made t o eliminate health expenditures o f MONE and related institutions which are recorded as education expenditures. Adjustments have also been made t o eliminate transfer items such as transfers t o General Directory and Dormitory that have been transferred t o MONE budget f r o m the Ministry o f Finance budget since 2004. Using unadjusted expenditure allocations f r o m the general budget to MONE leads to a n overestimation o f education expenditures o f about 8 percent, with average expenditure in 2002-2006 at 4.5 percent o f GDP, compared t o 4.18 percent estimated in Table 3.1. I t should be noted that expenditure estimations by SPO in the 2007 Annual Program are consistent with the estimations presented above (in trends and levels), with the average level o f expenditures in 2002-2006 estimated at 4.16 percent o f GDP. 98 expenditures over 1999-2006 i s evenly distributed between expenditures o n personnel and public investment, while expenditures and transfers have increased (Table 3.2). By contrast, expenditures o n education grew significantly between 1997 and 1999. This was the period when public spending was driven by the policy priority o f the then-enacted 8 year compulsory schooling. The increase in expenditures over this period i s entirely attributable to greater spending o n personnel (Table 3.2). Table 3.1. Overall Public Education Expenditures by Institution (1997-2006) Source: World Bank staff calculations based on MONE and Emil, Social Expenditure Study, November 2005 (background study) 3.9 Fluctuations in public expenditures may also reflect the weak linkages between policies, planning, and the budget. This was the main deficiency o f the Turkish sectoral budget process until recently. The new budget system, with foundations laid by the new financial management and control law (Law 5018), corrects this problem by introducing strategic planning and performance based management into the budgeting process. Still, the disconnect between objectives and budget will remain as long as the this linkage problem continues as the line units at the central level are not equipped with sufficient knowledge and understanding o f strategic planning and performance budgeting concepts for effective implementation. The capacity issues in this regard are especially a problem in education sector due to heavy policy agenda for coming years, as discussed in chapter 4. In the absence o f the effective mechanisms leading to greater linkages between policies and financial resources, there could be a massive mismatch between what is committed f i o m the government and what i s affordable regarding budget constraints. 99  MoNE 0.07 0,lO 0,09 0,lO 0,ll 0,08 0,07 0,06 0,07 0,09 Universities 0,lO 0,lO 0,12 0,04 0,04 0,12 0,ll 0,09 0,lO 0.10 Other Ins. 0.07 0,07 0,09 0,lO 0,ll 0,ll 0,16 0,15 0,19 0,19 Total 3,77 4,11 4,48 4,26 4,31 4,33 4,22 4,03 4,11 4,19 (3) Financing of Education Expenditures 3.10 The fragmentation o f firzancing sources has been reduced. The analysis o f public expenditures on education in Turkey presented in the World Bank's 2001 Public Expenditure and Institutional Review (PER) highlighted the fragmentation in sources o f finance.as a key feature o f the sector. These sources o f education finance included general budget appropriations to the Ministry o f National Education (MONE) and to other central government agencies providing educational services, various budgetary funds (Law 4306 for Basic Education, L a w 4318, Law 3308), extra budgetary funds (Social Aid and Solidarity Fund), revolving funds, aid from the community, and aid from associations and foundations. After the extra-budgetary funds were closed in 2001 and the earmarked expenditures attached to the closed extra budgetary funds were abolished in 2003, the main source o f finance for the education sector became the consolidated budget. Until 2006, educational services had been provided by the consolidated budget agency's schools. After the adoption o f the new Public Financial and Management and Control Law (PFMC), educational institutions are now classified under the central budget. In the light o f new definition o f the general government sector in L a w 5018, the percentage distribution o f public education expenditures i s given in Table 3.3 and further explained in Boxes 3.1 and 3.2.69 69L a w 5018 defines general government as the total o f central government, social security institutions and local authorities. Central government includes general budget agencies (such as MONE), special budget agencies (such as universities) and regulatory and supervisory authorities. Our definition o f central government in this study also covers extra budgetary funds and revolving finds. This suggests more comprehensive coverage o f general government sector than l a w 5018. However, t h i s definition does not cover foundations and P T A s (parent teacher association) and other communal aid. 100  Table 3.3: Public Education Expenditures in Turkey by Institution (percentage shares in total) Total General (million YTL) MONE Universities Others Government 2004 11.184 3.855 2.235 17.275 2005 12.870 4.050 3.048 19.969 2006 14.393 4.749 3.455 22.598 (Percentage Share) 2004 64,7 22,3 12,9 100,o 2005 64,5 20,3 15,3 100,o 2006 63,7 21,o 15,3 100,o Source: World Bank Staff calculations based on MONE Box 3.1: Principal sources of financing for the education sector Central Budget Agencies These are comprised o f direct allocations to MONE under the general budget agencies, University budgets under the special budget agencies, and other central budget institutions which use public sources for education and are covered by the central budget (e.g., the Ministry o f Health, the General Directorate o f the Police, etc.). The total share o f the general government revenues in total education expenditures i s approximately 98 percent, including the revolving funds (discussed below) and extra budgetary funds. Extra Budgetaw Funds -the Social Solidaritv and Sumort Fund Under the current framework, only the Social Solidarity and Support Fund provides extra budgetary funds to the education sector. This hnd provides free school books and financial assistance to the children from poor families. The education related expenditures o f Social Solidarity and Support Fund are categorized as “not classified elsewhere” in the Budget since these expenditures are not considered for education, per se, but have more to do with scholarships and pro-poor support. Revolving Funds At the pre-university level, secondary vocational schools generate most o f the revolving fund revenues through the sales o f products produced by their students. These contributions are usually allocated to pay for non-personnel current expenditures (such as materials). At the postsecondary level, university hospitals generate a significant amount o f revolving funds revenues, but as they are spent on health services they are not considered in this analysis. Local Governments There are mainly two types o f local authorities in Turkey providing funding to the education sector, namely municipalities and Special Provincial Administrations (SPAS). L a w 222, which regulates local governments, establishes that Provinces must spend at least 20 percent o f their budget o n education, especially at the primary education level. 101 Box 3.2 Expenditure Procedures at the Provincial Level (SPA Involvement) SPA involvement in carrying out investment on education at the provincial level stems from the practice that SPAs contribute 20 percent their revenues to primary education at the provincial level. This enables SPAs to merge their own funds with the central budget appropriations to MONE to be spent on education that province. SPAs have been developing capacity in tendering, contracting and managing procurement o f civil works, goods, and services, as well as in making payments to the contractors. Criteria for which procurement activities are done at the center and which by SPAs are not subject to any specific legislation and left to the discretion o f MONE. A flow chart o f the expenditure procedure under SPA involvement case: I Budget Office SPA (Center) Procurement Unit* 0 issues call for tenders, 0 forms a tender commission (with the participation o f the MONE Releases appropriations sends representatives), payment orders and cash awards contract Office o f the Chief Financial Officer o f the Provincial Unit o f Civil Works 0 Controls the work of the contractors Province 0 Checks the work completion reports for payments (hakedisler) calculated by the contractor Informs the appropriationto 0 Sends completion report and payment order to the SPA Provincial Office o f MONE Payments order signed by governor, cash transferred to SPA bank account by the accountingofficer (muhasebe muduru) 3.1 1 Comprehensive information on education expenditure has been collected. Until 2002, no comprehensive public expenditure information was available for the education sector. Even for public expenditures on education, because o f budget fragmentation due to o f f budget activities and a lack o f functional budget classification systems, it was very difficult to collect data from the institutions budgets and then consolidate these data. With the new budget coding system introduced in 2004 for the consolidated budget institutions and subsequently expanded to the entire central government budget agencies, data for the public sector have been collected in a more coherent way and are n o w accessible for analysis. In addition to this improvement, in 2002-2003 the Turkish Statistical Institute (TURKSTAT - formerly entitled the State Institution o f Statistics) conducted institutional 102 surveys aimed at collecting detailed educational expenditures and revenues from educational institutions and other provider^.^' All o f these surveys, other than the school survey and the crhche survey, collected information from the complete universe o f such educational institutions in Turkey, and therefore the data collected represented total spending on education in Turkey by each o f these institution^.^^ 3.12 Including private financing, Turkey spends on education the equivalent o f 7 percent o f GDP. The 2005 World Bank Education Sector Study included a detailed analysis o f the flow o f funds in Turkey’s education system based in part on the above survey. This analysis, presented in a background study for the education system, was called the National Education Accounts (NEA) report.72 The analysis found that Turkey spent on education in 2002 the equivalent o f 6.97 percent o f GDP.73 O f this, about 62.3 percent was financed from public sources and 35.3 percent by private sources (Figure 3.3). International resources accounted for 0.1 percent o f total education expenditures, while 2.3 percent o f expenditures could not be unequivocally accounted as public or private. As noted earlier, public expenditures on education equaled 4.34 percent o f GDP in 2002, with private and other expenditures on education accounting for 2.63 percent o f GDP. The vast majority o f household spending (90.4 percent) came from own funds, 3.5 percent came from central government revenues in the form o f scholarships, 1.4 percent came from private foundations, and 0.2 percent from municipalities in the form o f scholarships. The remaining amount came from “other” sources that could not be identified by type. 70 The TURKSTAT surveys were sent to: (a) a nationally representative sample o f public and private schools providing pre-primary, primary, secondary general, vocational or special education; (b) all public and private universities providing 2 year, 4 year, 5 year, masters and doctorate programs; (c) a nationally representative sample o f private crbches; (d) the Ministry o f National Education; (e) the Higher Education Board, the Council o f Higher Education (YOK); and (e) all other ministries and public agencies such as the Ministry o f Finance, Ministry o f Health, Ministry o f Labor, etc. that have any expenditures in education. 71 A separate nationally representative household education expenditure survey was also conducted by TURKSTAT, which captured total household spending on education. The survey was designed to collect detailed information on household educational expenses, including expenditures related to family members in public schools, private schools, foundation universities, non-formal education, and tutoring or exam preparation programs (dersanes). The survey also collected information on educational items received by the household, in-kind, credit, and financial assistance (scholarships etc) given to the household members. 72 The NEA employed a sources-to-uses framework and the accounts are disaggregated according to financing agent (public and private), provider (levels and types o f schools and universities), and inputs (personnel, school supplies, etc.). The National Education Accounts analysis focused on recurrent expenditures, since policy and strategy are principally influenced by recurrent expenditures, as opposed to investment allocations. 73 The mid-year 2002 exchange rate was US$1=1,519,668 TL; GDP in 2002 was 277,449 trillion TL. 103 I Figure 3.3. Education Expenditures in Turkey in 2002, by Source (YTL 19.35 billion) International Resources Other Sources Household Funds 33.4% i (6.46) Central Foundations, Government Associations, Revenue I Firms etc. 61.4% I 1.9% (11.89) (0.37) 0.9% (0.17) Source: World Bank Staff calculations (c) Providers of Education Services 3.13 Primary and higher education absorb the largest shares o f expenditures. O f the total spending on education, 89.5 percent went to the providers o f educational services, that is, the public schools and universities (including public providers o f informal education), and the balance o f 10.5 percent went to private schools and (foundation) universities (Figure 3.4). 0 The largest share o f education spending went to primary education schools, which received 40.3 percent o f the total spending on education. O f the amount spent on primary schools, 95 percent went to public primary schools, and the rest went to private primary schools. 0 General secondary schools and vocational/technical schools accounted for approximately 10 percent and 8 percent o f spending on education, respectively. 0 Public universities accounted for 30 percent o f the total spending, while foundation universities and private course houses providing higher education received only 2.3 percent o f total education spending. 3.14 Private examination preparation programs (dersanes) at various levels o f education lfrom primary to university level) accountedfor 4.2 percent o f total spending on education. This includes dersanes that are providing instructional services to students preparing for university exams, as well as courses providing examination preparation instruction to younger students preparing for the secondary school selection examination and those preparing for the private school selection examination (now combined into a single selection examination called the OKS). Expenditures on private examination preparation programs were the equivalent o f one-third o f total expenditures on public and private secondary schools. 104  Figure 3.4. Total Education Spending in Turkey in 2002, by Provider Private Course Not known by kind ruuiic pre-primary Private pre-primary Private University 0% Public Vocational and Public general Technical 1 Private 'general 8% secondary secondary 11% 1% Source: World Bank Staff calculations 3.15 Private financing significantly contributes to primary public schools and universities. Private financing agents (i.e., families) contributed to about one-fourth the spending at public primary schools, but a much greater share o f spending at both pre- primary and general secondary public schools (45.2 percent o f total spending at pre- primary public schools and 46.2 percent at general secondary public schools). Public financing agents contributed to almost all o f the spending at vocational/technical schools and special needs schools (more than 90 percent o f the spending at each level). Private financing agents accounted for 36 percent o f spending at public universities (YTL 2.12 billion) and all o f the spending at private universities (YTL 447 million-Figures 3.5 and 3.6). Private out-of-pocket financing o f public schools is a consequence o f the continuing financial constraints faced by the public schools. Principals request that the parents contribute to the maintenance, costs o f paper, and other needs o f their schools. Funds are collected from time to time or before the registration o f the new students. 105  Figure 3.5. Financing of Education, 2002 100% u - n 80% 60% 40% 20% 0% Pre-primary Primary General Vocational University Special Secondary and Technical education , % . financed by public financing agents 0 %financed by private financing agents~ 0 % financed by other financing agents Source: World Bank Staff calculations Figure 3.6. Out o f Pocket Spending in Turkey in 2002, by Providers (YTL 6.93 billion) Private Exam rNot known by kind Preparation Courses 1.O% 11.7% Pre-primary Schools Special Education Schools 0.2% Primary Schools Universities 1 I Vocational General High 35.3% Tech./High Schools Schools 3.3% 15.2% Source: World Bank Staff calculations 106  (d) Per-student expenditure 3.16 Education expenditures per student vary substantially by the level o f education. The average annual expenditure per student at primary school was YTL 810 (US$ 533), while it was four times higher at Y T L 3,260 (US$ 2,145) for a University student. The average expenditure on students attending general secondary schools and vocational/ technical schools was U S $ 902 and U S $ 1144, respectively (Figure 3.7). The government spends U S $ 395 per student at the primary level and U S $ 1,088 at the university level. f8 The Government spends the hi est amount per student, equivalent to U S $ 1,343, at vocational and technical schools. Figure 3.7. Total Education Spending per Student (US%), 2002 (US$l=YTL1.52) 1 1 2500 Pre-primary Primary General Vocational and University ' , Secondary Technical ' Source: World Bank Staff calculations 3.17 The differences in educational expenditures at public and private schools are significant. The per-student spending at public primary schools (US$ 5 16) i s almost one- third o f the per-student spending at private primary schools (US$ 1,524). Similarly, the per-student expenditure at public general secondary schools (US$ 876) is almost h a l f the spending at private schools (US$ 1,587) offering the same education (Figure 3.8). 74 T h i s issue i s covered in more detail in the 2005 Turkey Education Sector Study, Chapter 3: Developing the Basic Competencies o f All Students. 107  Figure 3.8. Education Spending per Student at Public and Private Schools (US$), 2002, (US$l=YTL 1.52) Source: World Bank Staff calculations 3.18 Not surprisingly, educational outcomes vary by school type. Primary school students studying in private primary schools perform about twice as well as those in public primary schools, whether they are enrolled in the well-supported boarding schools (YIBOS) or in the normal "Basic Education" primary schools (Figure 3.9). F i g u r e 3.9: Achievement across school types ,8 I ,7 - $6 - ,5 - ,4 - TURKISH MATH ,3 - -SCIENCE SOCIAL Bask Boarding Private School type Source: World Bank Staff calculations based on MONE 108 3.19 Public schools, regardless o f the level of education provided, spend a higher share o f their expenditure on personnel expenses relative to private schools. This i s most likely a factor accounting for the differences in outcomes between public and private schools. The difference in the structure o f spending reflects a more intensive use o f books, supporting educational materials, and school support services in private schools. Research in much o f the world indicates that there is a positive, causal relationship between the level o f non-salary inputs and educational outcomes. 3.20 Public schools also allocated a larger share o f their expenditures on investment spending relative to corresponding private schools (Figure 3.1 0). Again, the only exception is general secondary level o f education where both public and private schools spent an equal proportion o f their money (4 percent) on capital items. Public universities spent a higher share o f their total expenditures on personnel expenses (40 percent) compared to foundation universities that spent only 30 percent on personnel expenses. Though, both providers had similar shares o f investment expenses, at about 12 percent. Salaries and other personnel expenses accounted for 64 percent o f total expenditures at public primary schools, but only for 39 percent o f the total expenditures at university Figure 3.10. Distribution of Spending on Educational Inputs in Turkey in 2002, (YTL 19.35 billion) 100 80 UlTransfer 60 0 Capital 40 Other current 20 Personnel Source: W o r l d Bank Staff calculations (e) International Comparison of Education Expenditures 3.21 Total education expenditures, as a percent o f GDP, are higher in Turkey than in most OECD countries (Table 3.4). Exceptionally high private expenditures as a share o f GDP (2.2 percent) i s the main reason for Turkey's outlier status. Compared to OECD partners, Turkey also allocates a smaller share o f government spending to education (4.5 percent). 109  Table 3.4. Expenditure on educational institutions as a percentage of GDP frc 1 public and pri :e o f fund Public ' Total Malaysia3 72 Korea 48 82 United States 591 7,3 Turkey' 495 697 France 5,6 6,O Poland3 56 Mexico 5,1 5,9 Finland ' 5,7 53 United Kingdom 4,7 5,5 Germany 4,3 583 Spain 4,3 4,9 Czech Republic 42 4,6 Brazil3' 41 m India5 4,o 42 Slovak Repub l i ~ ~ , ~ 4,o 4,l 1. Including public subsidies to households attributable for educational institutions. Including direct expenditure on educational institutions from international sources. 2. Net of public subsidies attributable for educational institutions. 3. Public subsidies to households not included in public expenditure, but in private expenditure. 4. Direct expenditure on educational institutions from international sources exceeds 1.5% of all public expenditure. 5. Year of reference 2000. 6. Year of reference 2002 Source: OECD Education at a Glance, 2003 and Turkey National Education Accounts, ESS Background Study 3.22 Turkey spends the least per student than any OECD member country, in line with its lower per capita income level, and with large discrepancies across levels o f education (Table 3.5). Total per student education expenditures in Turkey i s about one third the per student expenditure in OECD, on a~erage.'~Per student expenditure on pre primary is particularly low, as Turkey spends almost 16 times less than the OECD average. O n primary education, Turkey spends four times less than the OECD average, and for secondary education, Turkey spends less than half o f the OECD average. By contrast, Turkey's per-student spending on tertiary education is much closer to OECD averages. An analysis o f the distributional incidence o f such discrepancies o n per student spending by level o f education would be warranted. 75In order to compare education expenditure among the countries in more analytical perspective, a relevant indicator i s per student expenditures. The indicator "Education Expenditures as a Share o f GDP" i s not entirely relevant if there i s a significant per capita income gap across the comparison countries. 110  Table 3.5. Per Student Education Expenditure by Education Level (2004 US$ Source: Education at a glance 2004; W o r l d Bank Staff calculations Note: "P": only Public 3.23 I n view o f Turkey's level of development, per student spending is very low in pre-primary education but relatively high in secondary and tertiary education. Although there are large differences in public spending per student comparing Turkey with European and other OECD countries, i t i s important to measure Turkey's spending levels on education with appropriate comparator countries. A s shown in the scatter diagrams below (Figures 3.1 1-3.14), Turkey's per student spending o n pre-primary schools is very low, while spending in primary school i s consistent with expectations given per capita income. For secondary and tertiary education, Turkey spends somewhat more per student than what would be predicted o n the basis o f i t s per capita income. 111  Figure 3.1 1. Expenditureson Pre-Primary Education (PPP Adjusted) 9,uw 8,000 7,000 6,000 m 5,000 : 4,000 a r 0 5 : 3,000 : W 2,000 1,ow 0 -1 ,ow Per Capita Income (PCI), PPP wee: World Bank Staff calculations Figure 3.12. Expenditures on Primary Education (PPP Adjusted) 8000 7000 0 v) .- C 6000 4 C 5000 a 4000 n 5 3000 .- C '0 f 2000 n 1000 0 0 5000 10000 15000 20000 25000 30000 35000 40000 Per-Capita Income (PCI) Source: World Bank Staff calculations 112  Figure 3.13. Expenditureson Secondary Education 10000 9000 5 CI 7000 C 0) u 6000 E a 5000 P f 4000 e 3000 5 8 2000 1000 0 0 5000 10000 15000 20000 25000 30000 35000 40000 Per-Capita Income (PCI) , ~ I Source: World Bank Staff calculations Figure 3.14. Expenditureson Tertiary Education 25000 g 20000 2. .cI E P) U a 15000 iz z n 3 10000 .- a CI U c % X 5000 W 0 0 5000 10000 15000 20000 25000 30000 35000 40000 Per-Capita Income (PCI) Source: World Bank Staff calculations 113  Figure 3.15. PISA Results by Secondary Expenditure per Student (PPP) 500 B 8 f 450 e g 400 1 B 350 0 1,000 2,000 3.000 4,000 5.000 6,000 7,CW 8,000 9,000 10,000 Secondary Expenditure Per Student (PPP) - ~ u r c e :World Bank Staff calculations 3.24 Even though per student expenditures in secondary education are comparatively high, student achievement is lagging behind comparators. As already mentioned, Turkey’s educational achievements as measured by the learning scores in the PISA study remain relatively poor. When these scores are combined with levels o f per student spending, they provide an indication o f the existing room for improving expenditure efficiency in education. A relatively comprehensive assessment o f student achievement is provided by the average PISA scores o f 15-year old students in three fields: Literacy; Mathematics; and Science. The average score in 2003 i s compared to average per student expenditure in secondary education in constant purchasing power parity (Figure 3.15). Turkey’s score remains below the average that could have been expected in view o f the important level o f per student expenditures. For example, compared to Thailand-a country with similar per capita income-Turkey spends approximately three times more per student o n secondary education (Figure 3.13). However, Turkish students’ average PISA score i s only marginally superior to that o f Thai students, indicating that there i s ample scope for efficiency gains in secondary education. Although student achievement, as measured by PISA indicators, i s expected to increase with per capita spending (Figure 3.15), i t would be counterproductive to further increase expenditure without, at the same time, taking steps to ensure that the efficiency o f spending has improved. 3.25 There is a considerable difference in MONE per-student education expenditures by province. There i s also a noticeable difference in per student expenditure on primary and secondary education across the provinces: In primary education, per student expenditures in Istanbul, the region with the lowest expenditure per student, is less than one-third that o f Tunceli-the region with the highest spending 114 per student. There is also a three-fold difference between the lowest (Istanbul) and highest provinces (Ardahan in 2004 and Gumushane in 2005) in terms o f per student expenditures in secondary education. 3.26 These differences potentially signal an inter-region or inter-province equity problem. Most o f the provinces which spent the least per student at the primary level are from the Southern and Southeastern regions. However, as the example o f Istanbul shows, some provinces with a high per capita income and developed areas are also in this group: Per student expenditures in Adana, Kocaeli, Bursa, Antalya, and Izmir are also very low. Indeed, both for primary and secondary education, there is no statistical relation between per capita income level and per student expenditure. This result, in particular for primary education, indicates that per capita income disparities may need to be taken in greater consideration in setting polices for the education sector. A3.Key reform directions 3.27 Improving education outcomes will require increased institutional efficiency as a prerequisite for committing additional financial resources. As further examined in chapter 4, new education policies to be introduced by the Government will have a significant medium-term impact o n the budget. Moreover, various educational gaps identified in this section suggest the need for additional fiscal space for educational expenditures (not explicitly evaluated in chapter 4): 0 improving the quality o f educational achievement in public schools may require increasing the share o f n o n personnel expenditures to levels similar to private schools and universities, as more intensive use o f supporting educational material and ICT in private schools is conducive to better quality and outcomes; 0 improving the regional balance o f educational services across the country may require enhanced public education service provision in most backward provinces to compensate for possible lack o f private sector initiative; 0 increasing per student expenditure o n pre primary education, an area where Turkey lags far behind comparators even accounting for differences in level o f development, yet where progress i s critical to improve participation o f women in the labor force. The tight fiscal framework in the years ahead indicates that the necessary condition for the sector to continue to expand is to improve the efficiency o f public spending o n education. Key policy directions for improved efficiency are presented in the World Bank’s 2005 Education Study and further summarized below. 3.28 A first priority is to implement policies that encourage or force schools to fill their classrooms up to acceptable capacity levels. This i s especially true for vocational school classrooms that are under capacity by an average o f 12 students per classroom. By improving school capacity utilization, it would be possible to increase the efficiency o f spending by reducing the number o f teachers employed for the same level o f enrollment. 115 3.29 Second, schools need to be given autonomy over their resources and held accountable for their results. As mentioned in chapter 3 , compared with Europe and most o f the world, Turkey’s public schools have the least autonomy over resources, staff deployment (at the school), textbook selection, allocation o f instructional time, and selection o f programs offered. M o r e school autonomy and accountability will help improve the allocation o f resources. 3.30 The first and second strategies could be combined in an innovative way by introducing a system ofper-studentfinancing. This system could be similar to the one implemented in the United Kingdom, where funding i s allocated in proportion to the number o f students a school enrolls up to its capacity. The school is responsible for using these f i n d s to cover their personnel, material, and professional development costs, but would be held accountable by Government for results. 3.3 1 Third, higher education institutions should be given greater autonomy over financial resources with incentives to manage these resources efficiently, and they need to be held accountable for reporting and outcomes. T o be able to become more efficient and provide better results, universities need to move funds between items and activities during the course o f the year. At present, higher education institutions have very little flexibility as their public funds are allocated in a large number o f line items and it is difficult for a university to reallocate between them. Further, the amount allocated to a university under each line item is determined centrally, and mainly o n the basis o f historical precedent. Financial autonomy can be increased by reducing the number o f ‘blocks’ - o r line items - within which a university receives its general public funds, excluding funds allocated for specific purposes (such as for research from TUBITAK). Universities should also be able to use accrual accounting with the provision reintroduced to allow unspent finds to be ‘carried over’ from one year to the next (which would enable universities to build up ‘reserves’). Within a financially autonomous system, the amount o f funding allocated to each university should not be based o n inputs (such as staff), but on ‘performance’ in terms o f outputs for the three main activities: teaching (using student numbers), research and service development. Importantly, an increase in financial autonomy is unlikely to lead to any increase in public funding. Increases in financial autonomy should be coupled with a commensurate increase in accountability for internal accounting, reporting to independent boards, and producing annual reports. I t should be recognized, however, that a comprehensive approach should be designed to achieve needed administrative and financial autonomy o f higher education institutions. 3.32 Fourth, the university entrance examination system should be modernized in such a way as to eliminate the advantage o f private examination preparation courses (dersanes and tutors). YOK is considering basing such a revision o f the current examination with a system based to some degree o n models such as the French or International Baccalaureate, the German Arbitur, or the British ‘A’ levels. The Turkish model would be a comprehensive set o f exams that test both the knowledge and skills that students learn in secondary school. I t could have a core o f subjects with blocks o f subject options which pupils can choose. Such a system would be designed to assess what students learned throughout their secondary school programs, what they are able t o 116 do with what they have learned, h o w they reason, and h o w they apply their skills. Such assessments would create standards for graduating from secondary school, provide a way to measure how schools are performing, provide an incentive for students to learn what i s taught in the classroom rather than at the dersanes, and serve as a measuring rod for universities and employers to assess knowledge and reasoning competencies, not test taking skills, which have absolutely n o value in the real world - and certainly should have n o value in institutions o f higher learning. If h a l f o f the private f i n d s used to finance examination preparation today could be redirected to subsidize public secondary education, i t would represent a boost o f that spending by nearly 40 percent. B. THE PENSION SYSTEM AND ITS REFORM 3.33 The Turkish pension system has been a perennial headache for fiscal policy as the system continues to run large deficits. Fiscal deficits have persisted, with the exception o f a temporary improvement in 1999, as parameters improved from passage o f a pension reform law. Policymakers are hoping to have solved the major problems with the passage o f the newest pension and health insurance law in M a y o f 2006. The law should bring the system to fiscal sustainability, but only in the long run as all the parametric changes embedded in the new law unfold. B1.Evolution of the Pension System 3.34 The Turkish pension system historically consisted o f three pension institutions. Sosyal Sigortalar Kurumu (SSK) covered private sector workers and state employees not considered career c i v i l servants. Emekli Sandigi (ES) covered state employees with c i v i l servant status. Bag-Kur (BK) covered the self-employed and farmers. BK ran separate schemes for the self-employed and the farmers, with completely different parameters for each, thus resulting in 4 public sub-systems. 3.35 Turkey is otherwise a young country and should not be experiencing pension system deficits of the magnitude that it has experienced. Turkey’s population i s twice as young as Ireland’s; the next OECD country with the youngest population, which itself has a twice as young population as Italy, the OECD country facing the most severe trend o f population aging (Figure 3.16). The challenge o f ensuring sustainability o f the pension system faced by Turkey i s also encountered in Latin America countries where the population i s also young, and i t i s mainly a result o f excessive benefits in relation to revenues. Solutions in other countries have ranged from privatization o f the system, to drastic cutbacks in benefits and/or increases in contributions and/or increases in retirement ages. 117 I Figure 3.16: Population over Age 65 in Turkey and Selected OECD Countries (in %) 25 20 15 10 5 0 Source: OECD 3.36 The first critical feature which led to the downfall o f the Turkish pension system was the removal o f the minimum retirement age in 1991. At this point, individuals could retire once they had completed the minimum years o f service required for retirement, which were 25 years for men and 20 for women. But years o f service did not imply years o f contribution. In the case o f SSK, an individual needed to fulfill only 5000 days o f contributions (14 years) before becoming eligible for a pension which led to the phenomena o f substantial numbers o f individuals retiring before the age o f 40. 3.37 The 1999pension law tried to remedy this by reimposing a minimum retirement age uniformly for all three pension institutions. The retirement age for new entrants to the labor force after the passage o f the 1999 law was 60 for men and 58 for women. This is lower than the retirement age in other OECD countries, but certainly better than what had existed before. However, given the previous structure o f the pension system, the new retirement ages were only to be gradually phased in, with the earliest retirees after passage o f the 1999 law allowed to retire at 43 for men and 38 for women. The phase-in for this retirement age i s painfully slow with average retirement ages well below 50 even today. 3.38 The second feature o f the Turkish pension system which raised costs was the Government’s decision, beginning in the mid-1980, to supplement pensions with uniform flat amounts awarded to all pensioners. These payments were intended to supplement the very l o w pensions for which many individuals qualify-as a result o f the l o w ages and the l o w contribution periods required to be eligible for pensions. Since the amount o f the flat pension was decided by Parliament each year, without regard to the finances o f the pension system, these flat payments eventually became larger than the pensions themselves. Aside from the fiscal consequences, these flat payments had substantially negative impacts o n incentives. If the majority o f the pension is unrelated to the number o f years o f contributions made or the amount on which one contributes, individuals have an incentive to declare the fewest number o f years and the lowest salary 118 possible. These negative incentives substantially reduce revenues which only exacerbates the fiscal situation caused by the increased expenditure. In the late 1990’s the Government froze these flat payments in nominal terms, and with the subsequent era o f high inflation, the value o f these flat payments i s currently quite small. 3.39 However, social support payments, to all pensioners, were made again in 2002, followed by additional payments in 2003. Following the 1999 law, which attempted to link contributions and benefits much more closely for S S K and BK self-employed and indexed the pensions post-retirement to inflation, again there were complaints about the l o w level o f pensions. For S S K and BK, social support payments were taken as prepaid inflation increments. Inflation increments are paid in Turkey o n a monthly basis. The flat payments were added to the pension payments at the beginning o f the year and no inflation increments were given until the sum o f the unpaid inflation increments equaled the value o f the flat social support payment. By midyear the value o f the pension in S S K was essentially where it would have been without the social support payments. However, in BK, where the pension levels are lower and the inflation increments are much lower, and the flat payment was higher, there was s t i l l a residual increase at the end o f the year over where the pension would have been with inflation increments. Rather than reduce the nominal amount o f the pension payments the following year, the Government chose to simply add the residual amount o f the flat payment to the base pension and then index normally from that point forward. T o avoid inequities for new pensioners, an equivalent flat amount was added to the base pensions o f new pensioners as well. B2. Comparison o f 2006 Pension Law with 1999 Law 3.40 While the 1999 pension law introduced the minimum retirement age for all three pension institutions, it changed benefit parameters for only SSK and the se& employed insured under BK. While the minimum retirement age and the new benefit parameters were a vast improvement over the previous law, they s t i l l were generous by international standards and provided individuals an incentive t o contribute as few years as possible. Previously, the minimum and maximum wages o n which contributions were paid were adjusted at the discretion o f Parliament and were not linked to any other parameters in the economy, including minimum wage. Thus, there was a period o f time when the maximum insurable earnings was below minimum wage. Since the insurable earnings determined the pensions, pensions were understandably quite low. The 1999 law tried to link the contribution limits and other pension parameters to observable economic factors. In the absence o f a recognized wage series, many o f the parameters were linked to GDP growth or inflation. Thefour achievements o f the 1999 law were: 0 the introduction o f the minimum retirement age for all three pension institutions and all sub-systems, although at an extremely slow phase-in rate, 0 an increase in the length o f contribution required in SSK, although again at a slow rate, 119  0 some decrease in the disincentives for longer contribution periods in SSK and BK self-employed, and 0 a linking o f important pension system parameters in S S K and BK self-employed to economic factors. The law very importantly did not achieve a unification o f the pension parameters in a l l four sub-systems. 3.41 The 2006 law focuses on the unification o f the four sub-systems, bringing them institutionally into a single institution with uniform parameters for all categories o f workers. This i s a significant achievement for improving the flexibility o f the Turkish labor market, for improving compliance, and for improving the fiscal sustainability o f the overall pension system, given the rapidly increasing costs o f the c i v i l service system prior to the reform. Table 3.6 provides a comparison o f the pension system parameters before and after the 2006 reform. The most significant changes apply to the c i v i l servant system and the BK farmers’ system bringing them in line with the other sub-systems. 120  ru 0 I .^ B K n N E L 0 h W L 9 3 3 N W f3 -= L W ed U 3 2 x 0 I ; v) L W Y 8 E cp a E W Y v) h VY .I m i!i 8 .e 4 c) W 9 .e r, 0 s a sf .I 3 L cp 3 W 8 P .. a : I c, Y E 2 a W % L YI . b 2  B3.Impact o f 2006 Law and remaining challenges 3.42 From a fiscal sustainability perspective, the 2006 law brings the pension system in the long run to fiscal sustainability. However, as with all pension systems which do not adjust benefits to l i f e expectancy changes, as life expectancy continues to rise in the outer years o f the simulation period, pension deficits will begin to re-emerge, but these will not be anywhere close to the magnitude o f the current deficits and will remain below 1 percent for the remainder o f the simulation period. The pension projections below show the fiscal impact o f the reform (Figure 3.17). Figure 3.17: Pension SystemDeficits Before andAfter Passage ofthe 2006 Law I 1% 0% ~ -1% -2% -3% -4% -5% -6% -7% -8% 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060 2065 2070 2075 Source: PROST projections produced by Bank staff Year '- +B - a se Case -ReformCase 3.43 The positive achievements o f the reform include the substantial fiscal gains as well as the unification o f a previously fragmented system. Some o f the fiscal gains come from the uniform move to inflation indexation o f pensions post-retirement, which is a substantial achievement. Previously, only SSK and BK self-employed pensioners received pension adjustments indexed to inflation by law, although the Government had granted higher increases in recent history. Furthermore, prior to reform, there were strong disincentives for continuing work, with the benefits given for each additional work year much lower than those given for the normal working career. While not necessarily providing incentives for continued work, this new law removes the disincentives, with all work years being given equal credit whether they are below what i s considered a normal working career or not. However, there remain some areas o f concern for the longer t e r m future. 3.44 The contribution rates in Turkey are relatively high by OECD standards despite Turkey's young population (Figure 3.18). High contribution rates hamper activity and job creation in the formal sector, and eventually deprive the budget from significant 122 revenues. The Turkish government has recently expressed interest in lowering contribution rates in order to encourage growth o f the formal sector labor force. Currently, there i s no fiscal space to lower contribution rates and Figure 3.17 suggests that there would be no space to lower contribution rates in the foreseeable future. Even at i t s peak in 2045, the system will just be fiscally balanced, but given the projected return to deficits shortly thereafter, lowering contribution rates at that juncture would soon lead to substantial fiscal deficits. If lowering contribution rates i s a desired objective, then more changes will be required to achieve that goal. Options for such changes are further outlined below. Figure 3.18: Contribution rates in selected OECD countries % - 40 g) 35 30 25 g 20 u) 0 15 g 10 s 5 0 Source: OECD 3.45 Retirement ages in Turkey remain substantially below those in other OECD countries. Only in 2048 will Turkey achieve the retirement ages o f most other O E C D countries today (Figure 3.19). Since most OECD countries do not currently have sustainable pensions, more changes will be required in these countries, suggesting that even in 2048, Turkey’s retirement age may lag those o f other countries. One might argue that Turkey’s life expectancy is below that o f the other OECD countries, but OECD countries like Mexico are quite comparable to Turkey and s t i l l have retirement ages o f 65 for both men and women even today. The l o w retirement ages in Turkey are a residual from the 1991 decision to remove the minimum retirement age entirely and the 1999 decision to reinstate a minimum retirement age very, very slowly. In the 2006 law, the decision was made to keep the 1999 retirement age schedule intact and only impact those joining the covered labor force after the new law enters into force. But this results in very l o w retirement ages for a very long time. Since these young retirees draw pensions for such a long time, once they have retired, it becomes very difficult to reduce expenditures. 123  F i g u r e 3.19: Retirement ages in OECD countries 80 70 60 - - _. 50 Men 40 30 Women 20 ~ 10 0 I Source: OECD and World Bank Staff calculations 3.46 Accrual rates, the benefit paid per year o f contribution, are also approaching OECD levels, but have not reached OECD averages even with the new pension law. Figure 3.20 shows the accrual rates in OECD countries. The leftmost Turkey bar represents where SSK and BK self-employed are today. ES and BK farmers are even more generous. The second Turkey bar from the left shows where Turkey will be under the new law beginning in 2007, just even with the most generous country in the OECD. The third Turkey bar shows where Turkey will be as the law unfolds in 2016, now even with the second most generous country in the OECD. Clearly, the rest o f the OECD has gone much farther in reducing accrual rates. 1. Figure 3.20: Pension Accrual Rates in OECD Countries Source: OECD and World Bank Staff calculations 3.47 A third source of potential savings to finance a reduction in contribution rates would come from an increase in the formalization o f the labor force, growth in the wage base, and better work incentives. Currently, almost 50 percent o f the contributors in SSK contribute on the basis o f minimum wages; most workers in all systems 124 contribute for as few years as possible in addition to those who work entirely in the informal labor force. This behavior i s linked with the pension parameters in the Turkish pension system. They are all predicated on the belief that the typical working or contribution l i f e i s relatively short and will remain so and should remain so. In most countries, the typical working life is 40-45 years, and individuals who participate in the labor force contribute for the full working life. In Turkey, the new law stipulates that the full working career will be 25 years, but this is being phased in extremely slowly such that only those who begin work in 2026 will face the 25 year constraint. As a result o f the few expected years o f contribution, Turkish policymakers are forced to choose between raising the accrual rate per year o f service or facing the prospect o f many people receiving very l o w pensions. But the high accrual rates undermine work incentives. If individuals can obtain decent pensions with only 20 years o f work and contribution, why bother to work and contribute for 30 years, much less 40 or 45 years as people in other countries do? The high accrual rates are thus self-reinforcing, as they encourage the policymakers to opt for high accrual rates from the beginning. The new law improves the situation in that before, private sector workers earned 3.5% accrual rates for their first 10 years o f work, which then dropped to 2% per year for the next 15 years o f work and then to 1.5% per year thereafter. N o w there is a flat rate for all years o f work. Actuarially fair insurance would reward those who work beyond retirement age by approximately 6% additional pension per year o f delay. 3.48 Furthermore, even in the new system, there is the possibility o f a partial pension. While the new law stipulates that the new entrant as o f 2026 must contribute 25 years before receiving a full pension, partial pensions can be received with only 15 years o f contributions and at an age three years older than the retirement age. But again, the 15 year requirement is being slowly phased in from the current 12.5 and the previous 10 year req~irement.~~ 3.49 While the 2006 law generally moves away from the previous structure and toward a rule-based system which encourages longer careers, one o f the pension system parameters chosen is pushing in the opposite direction. The 2006 law stipulates that when calculating the pensionable base, past wages will be revalued on the basis o f 50 percent o f past average wage growth and 50 percent o f past inflation. As a result, an individual who contributes on the basis o f average wage throughout his career will receive a pension based on substantially less than average wage.77 Furthermore, the wages furthest from the point o f retirement, the early wages, will be valued less than 76 As a result, there is a relatively large number o f pensioners, especially in BK where individuals contribute as little as possible, who are receiving very l o w pensions. As already noted, the typical response has been t o complain through the political process that pensions are too l o w and t o receive some kind o f supplement which i s equally distributed t o those w h o contributed 10 years as w e l l as to those w h o contributed 30. These periodic supplements h r t h e r reinforce the incentives to contribute as little and for as few years as possible. The recent supplements in 2003 more than doubled the pensions for BK farmers, causing not only fiscal stress, but also undermining compliance. 77 Typically, countries revalue o n the basis o f 100 percent o f average wage growth. Thus, an individual who paid contributions o n the basis o f average wage gets a pension calculated as a percentage o f the current average wage. 125 wages earned closer to retirement, giving individuals an incentive not to contribute for long durations o f time. Figure 3.21: Average Benefits Relative to Average Wage in SSK, post-reform 40.0% 35.0% 30.0% 25.0% 20.0% 15.0% 10.0% 5.0% .O% I Source: World Bank Staff calculations 3.50 The 2006 reform will reduce the level ofpension benefits relative to the average wage (Figure 3.21). The fall in pension benefits can be attributed to a number o f factors. The 1999 law lengthened the wage period on which pensions were based from the previous average o f the last 5 years to lifetime wages. However, SSK did not have a historical data base to implement the shift to lifetime wages immediately and i s increasing the number o f years entered into the average by one year per calendar year since 1999. The longer time period tends to reduce the pensionable base and thus the pension. Furthermore, the benefits being provided in 2006 are prorated benefits since individuals had contribution periods prior to the 1999 law. The pension for the pre- 1999 contribution period i s credited under pre-1999 rules, which were more generous. Thus, the 2006 value includes some grandfathering o f the pre-1999 rights. The same grandfathering principle applies to the 1999 law so that not until 2040 are individuals retiring who are fully under the new law. But a significant portion o f the change in pension benefits can be attributed to this change in the revaluation o f past wages. The 1999 law had set the revaluation parameter to nominal GDP growth which was overly generous, given that nominal GDP growth exceeded the typical average wage growth used by other countries, but the change to a 50-50 mix is much less generous. As a result, the pension benefits are falling. Note that this fall is occurring despite the fact that retirees in 2006 contributed for about 17 years on average, while retirees in 2040 will be contributing for about 30 years on average. While the example o f S S K is provided here, the impact o f the change will affect all the other sub-systems as well, particularly ES and BK farmers where the pension was historically based on last wage rather than an average career wage. 126 3.51 Summing up-important steps have been taken forward but, to address the challenge o f formalization, the reform agenda would need to be completed in thefuture based on the results o f the 2006 reform. The 2006 Turkish pension l a w has made substantial changes and will move Turkey to a fiscally sustainable pension system in the future with a unified labor market which improves labor flexibility and compliance. The remaining agenda involves increasing formalization o f the labor force and reducing contribution rates. In addition, the law removes many o f the disincentives to longer working careers embedded in previous pension laws, but may not provide sufficient incentives to increase the length o f the contribution period and the formalization o f the labor force. Further reform steps in the future should be envisaged in light o f the results o f the 2006 reform. C. PUBLIC EXPENDITURE IN THE HEALTH SECTOR C1.Health outcomes 3.52 Turkey has achieved considerable progress in terms o f health status of its population but the health sector is still under-performing. Health outcomes in Turkey compare unfavorably with health outcomes in countries that have similar or lower level o f GDP per capita and approximately the same share o f public spending o n health (Table 3.7).78 In general, these indicators are worse than average indicators in the Europe and Central Asia region or the group o f Upper Middle Income countries (with the exception o f maternal mortality ratio). However, certain health outcomes have been improving steadily through time (Table 3.8). Under-5 mortality rates have declined and a similar decline i s observed for infant mortality. These trends are consistent among different sources and seem to indicate that Turkey will probably reach the MDG related target for under-5 mortality o f 23 per 1,000 live births by 2015. Trends regarding maternal mortality, o n the other hand, are more difficult to establish, as data are scarce and not very reliable and differ greatly among various sources.79 In general, although a declining trend i s observed, this i s not very strong and maternal mortality remains high as compared to other countries in the region or countries with similar income level as Turkey.8o 78 Countries with similar shares of public spending on health may not have similar levels of health outcomes, as there i s no clear pattern between public spending and health outcomes when controlling for the country’s income level (S. Devarajan, presentation “Development Policy Lending and Health”, World Bank, Washington DC, March 1,2006). 79 See also discussion in World Bank (2003), Chapter 1 “Status and Trends in Health Indicators”. Information on mortality and morbidity in Turkey i s based on hospital data (discharges and deaths) and for that reason tends to underestimate the actual mortality and morbidity levels. 127  Tab1 3.7: Health indicators for selected countries (2003 or latest available year) GDP per THE PHE LE Infant Under-5 Maternal Vaccination capita (% (% Mortality Mortality MortalitJ rates (US$ GDP)' GDP)' DPT Measles 2000 con.) Spain 14,691 8.6 5.4 80 4 4 4 98 97 Greece 11,449 9.5 5.0 78 4 5 9 88 88 Portugal 10,284 9.3 6.6 76 4 5 5 99 96 Brazil 6,957 7.9 3.6 69 33 35 260 96 99 Czech 5,899 7.0 6.4 75 4 5' 9 97 99 Rep. Mexico 5,803 6.1 2.7 74 23 28 83 91 96 Hungary 5,105 7.8 5.5 73 8 7 16 99 99 Poland 4,634 6.1 4.4 75 6 7 13 99 97 Malaysia 4,011 3.8 2.0 73 7 7 41 96 92 Slovak 4,254 5.9 5.3 73 7 8 3 99 99 Rep. Turkey 2,977 6.5 4.3 69 33 39 70 68 75 Thailand 2,276 4.4 3.0 69 23 26 44 90 96 Russia 2,138 6.2 3.5 66 16 21 67 98 96 Romania 1,963 6.3 4.2 70 18 20 49 97 97 Bulgaria 1,838 7.3 4.4 72 12 17' 32 96 96 E&CA 2,283 6.3 4.2 68 29 36 59 90 92 UMI 4,075 6.4 3.5 69 24 30 92 88 90 Notes: 1. Figures for total health expenditures (THE) and public expenditures on health (PEH) are for 2002. Figures for Source WDI 64.0 .. ..50.0 .. .. 38.0 I .. .. 33.0 DHS 93 .. 52.6 DHS 98 .. 43.0 .. DHS 03 .. 29.0 Under 5 mortality (per 1,000 live births) WDI 78.0 .. ..60.0 .. ..45.0 .. .. 39.0 DHS 93 .. 60.9 DHS 98 .. 52.1 .. DHS 03 .. 37.0 WDI ..70.0 .. 128 Note: 1. WDI: World Development Indicators, World Bank database; DHS: Demographic Health Survey; WHO/HFA: World Health Organization, Health for All database. Source: World Bank (2004). 3.53 Improvements in health indicators during the last decade were not spread evenly across urban and rural Turkey and across regions. Infant and under-five mortality rates in rural areas are about 70 percent higher than in urban areas (Table 3.9).81 Lnfant and under-five mortality rates are higher in the North and East regions. 3.54 The prevalence o f HIWAIDS seems to be relatively low in Turkey but is increasing. The first AIDS/HIV case was registered in Turkey in 1985. By the end o f 2004 there were 1,992 AIDS/HIV cases registered at the MOHe8* A comparison with 2,001 suggests that the number o f AIDS cases in 2004 in the 25-34 age group has increased by 29 percent, while for the 15-24 age group by 40 percent .83 Although these increases are very small in terms o f number o f cases as compared to the total population, effective prevention programs are considered necessary for keeping infection rates low in the future. Table 3.9: Infant Mortality and 1 nder-five Mortality Rates by area of residence and region, 2003 Infant Mortality Under-Jive Mortality Residence Urban 23 30 Rural 39 50 Region West 22 30 South 29 30 Central 21 33 North 34 48 East 41 49 Source: Turkey Demographic Health S rvey, 2003. 3.55 Better use o f preventive health services among high risk groups (mothers and children) is key to improving core health outcomes. Turkey's high levels o f infant and under-5 mortality, as well as high maternal mortality, are related to the poor usage o f preventive care.84 According to the 2003 Demographic and Health Survey, one-fifth o f mothers do not receive any antenatal care during their pregnancies. The share o f mothers who do not receive any antenatal care increases to one-third in rural areas. With respect to the place o f delivery, although 78 percent o f all births are delivered in a health facility, nearly half o f births in the East region take place at home. Similarly, the proportion o f births delivered with the assistance o f a doctor or trained health personnel decreases from " Maternal mortality ratios are not reported in the 2003 TDHS. Draft tables prepared for the MOH Statistical Yearbook 2004 (forthcoming). 83 MOH data. 84 Preventive services include the programs o f TB, Malaria and Maternal and Child Health, Health Education and Cancer Control, as w e l l as activities o f the Institute o f Hygiene R e f i k Saydam. Ths definition o f preventive activities does not coincide with the one employed by MOH, which includes primary care activities as well. However, a l l primary care activities cannot be considered as public goods o r goods with large public externalities (e.g., personal services). 129  83 percent for the total population to only 60 percent for deliveries in the East region. The same survey also provides information with respect to vaccination rates where only .~~ 54 percent o f children 12-23 months are fully ~ a c c i n a t e d Vaccination rates also vary among regions, with the percentage o f children fully vaccinated in the East being as l o w as 35 percent. However, according to MOH administrative data vaccination rates have increased in the last two years with 77 percent o f children 12-23 months fully vaccinated at the end o f 2005 nationwide and 62.5 percent in the South East Anatolia region. C2.Main characteristics o f the health sector86 3.56 Estimates o f the number o f individuals enrolled with the different insurance entities in Turkey differ. Various public entities offer financial protection from out-of- pocket health expenditures in Turkey (see Box 3.3). Data from the State Planning Organization show that around 84.5 percent o f the total population (59 million) enjoys health insurance coverage (Table 3.11). S K K has the highest number o f enrollees with around 32 million (46 percent o f total population), while Bag-Kur and Emekli Sandigi cover around 15.5 million individuals (22 percent) and 10.5 million individuals (15 percent), respectively. Privately insured individuals represent less than 1 percent o f the population. Nevertheless, issues such as multiple enrollment, non up-dated records and the use o f estimates for the number o f dependents have been put forward as problems affecting the reliability o f these data. Estimates from the 2002-2003 Turkey National Household Health Expenditure Survey, undertaken in the framework o f National Health Accounts, report a national insurance coverage significantly below that reported in national statistics, with only 67 percent o f the population being insured, including the Green Card program.87 I Box 3.3: The system of financial protection for health expenditures I The most important entity is the Social Security Organization (Sosyal Sigortalar Kurumu, SSK). SSK i s a social security institution for private sector employees and blue-collar public workers. SSK is financed by health insurance premiums paid by employees (5 percent o f earnings) and employers (6 percent o f earnings), while deficits are covered by budget transfers. SSK’s benefit package offers family coverage and includes emergency, outpatient and inpatient services, while in general SSK does n o t pay for preventive services. (Table 3.10).” Self-employed individuals, agricultural laborers, as w e l l as their dependents are insured by the Social Security Institution o f Craftsmen, Tradesman and other Self-Employed (Bag-Kur). Bag-Kur is financed by 85 Children who are fully vaccinated have received BCG, measles and three doses o f DPT and polio. 86 This section provides a b r i e f description o f the health sector in Turkey with respect t o financial protection mechanisms and the provision and utilization o f health care services. F o r a detailed description o f financial protection mechanisms and provision and utilization o f health care services in Turkey, see W o r l d Bank (2003) ” Analysis o f household surveys for 2001 and 2002 provide similar results regarding health insurance coverage; data indicate that only 62 percent t o 63 percent o f the total population in Turkey has access t o some f o r m o f health insurance, including Green Card holders (World Bank, 2005). 88 This group also includes invalid individuals, widowsiers, and orphans. 130 health premiums (equal to 20 percent o f each contributor’s earnings). Civil servants (active and retired) and their dependents acquire health insurance coverage by a third social security institution, the Pension Fund (Emegli Sandigi). Health expenditure for active (working) individuals and their dependents are financed by the budgets o f their public institutions, while expenditure for pensioners and their dependents are financed by Emekli Sandigi. Emekli Sandigi does not collect premiums, but it receives a direct budget transfer. The benefit package offered by Bag-Kur and Emekli Sandigi i s similar t o the SSK benefit package. The Green Card program was introduced in Turkey in 1992 and aims at providing financial protection against out-of-pocket health expenditures t o poor individuals. The Green Card program i s managed by the MOH and it provides free health services (with the exception o f pharmaceuticals prescribed during outpatient visits since M a y 2005). This program was envisaged as a transitory measure until the introduction o f the universal health insurance coverage takes place. I I I I I Notes: 1. The distinction between active workers and retired applies only to SSK, ES and Bag-Kur. 2. In the case o f drugs prescribed by a physicians’ committee, no co-payments apply. 3.57 The percentage o f insured individuals in rural areas is considerably lower than the national average. According to the 2002-2003 household data, only 57 percent o f individuals are insured in rural areas, while the corresponding percentage in urban areas 131 i s 72.5 percent. Across regions, the lowest insurance coverage i s registered in East Turkey at 54 percent. Type State Planning Organization Household Health Expenditure Survey 2002-2003 2002 % o f total population % o f total population Emekli Sandigi 15.4. 12.5 SSK 46.3 33.5 Bag-Kur 22.3 11.7 Private Funds 0.5 0.4 Green Card 8.6 Others 0.5 Total insured 84.5 67.2 3.58 Universal health insurance will be introduced starting January 2007. The introduction o f universal health insurance (UHI) aims at merging the various social security and other schemes under a common fund that will pool all public resources for health and be responsible for contracting health care providers. The newly created fund will have the capacity to contract with public and private providers that offer high quality services. Universal insurance coverage will also offer the same level o f financial protection against out-of-pocket health expenditures to all individuals in Turkey. This applies to differences in financial protection currently observed between insured and non- insured individuals, but also to differences within the group o f insured individuals. For example, in the case o f the Green Card program i t is well known that resources fall short o f the actual needs o f the program’s beneficiaries. Among the SSI beneficiaries, the lowest out-of-pocket expenditures are reported for SSK beneficiaries, followed by Emegli Sandigi and Bag-Kur beneficiaries. 3.59 The system o f health care service providers is evolving. Until 2004, health care services in Turkey were mainly provided by the MOH facilities, the S K K facilities and public university hospitals, with only 6 percent o f hospital beds in the private sector (Table 3.12). Starting 2005, all SSK facilities were transferred to the MOH, as a transitional measure until the transformation o f health care providers to autonomous entities. This transformation forms part o f the Government’s health sector reform agenda and aims at separating the financing from the provision o f health care services. 3.60 The recent changes have caused the M O H to reinforce its position as the main provider o f health care services. In particular, the MOH i s the sole provider o f preventive and primary health care services, provided through a network o f health centers, health posts, mother and child health and family planning centers, tuberculosis dispensaries, and malaria and cancer control centers. The MOH also owns and operates 770 hospitals, comprising 68 percent o f all hospitals beds in the country. University 132 hospitals complement the provision o f secondary and tertiary level care through 56 tertiary hospitals (15.2 percent o f hospital beds), while the private sector i s the third largest player with 287 hospitals, but only 6.7 percent o f total bed capacity. Another 75 hospitals (around 10 percent o f hospitals beds) are owned by other ministries, municipalities, associations, foreigners and minorities. Type of institution No. of hospitals % (hospitals) No. of beds % (beds) Ministry o f Health 770 67.7 132,062 50.4 Ministry o f Defense 42 3.5 15,900 8.5 Faculty o f Medicine (University) 56 4.7 29,700 14.5 Municipalities 7 0.6 1,221 0.7 Associations & Foundations 18 1.5 2,224 0.8 Foreigners 3 0.3 268 0.1 Minorities 5 0.4 618 0.5 Private 287 24.2 12,990 6.3 Total 1,188 100.0 194,983 100.0 3.61 The total number o f hospitals and hospital beds has increased considerably, with a stronger role of the private sector. From 1994 to 2006, the number o f hospitals and hospital beds has grown by 17 and 32 percent respectively.” The highest increase has been registered in the number o f hospital beds in the private sector (63 percent), followed by increases in the number o f university hospital beds (50 percent). 3.62 A large number o f hospitals were - and continue to be - substantially underutilized. Moreover, important variations s t i l l exist in hospital occupancy rates among provinces, ranging in 2005 from 25 percent in Tunceli State Hospital to 96 percent in Gaziantep State Hospital. Overall, bed occupancy rate in Turkey i s 62 percent, which i s much below other European countries (Table 3.13). 3.63 The number o f physicians and nurses has also increased but regional disparities remain. For the periods 1994-2004 the number o f physicians and nurses grew by 5 1 percent and 74 percent respe~tively.’~ Their numbers have thus approached the corresponding averages in the EU-15 countries.” Nevertheless, health care providers are unequally distributed, with fewer providers operating in the East and Southeast regions as compared to the rest o f the country. As a result, some areas have an over-supply o f health care providers. 3.64 The availability o f diagnostic technologies has considerably improved. In 2003, the number of magnetic resonance imaging (MU) units in Turkey was 3 per million population, less than h a l f o f the OECD average o f 7.6 MRI units, but higher than France (2.8), Hungary (2.6), Czech Republic (2.4), Greece (2.3), Slovakia (2.0), Poland (1 .O) and Mexico (0.2). Similarly, the number o f computed tomography scanners in Turkey stood 89 Turkish Statistical Service. 90 MOH data. 91 Health for A l l Database, WHO/Europe, 2006. 133 at 7.3 per million population in 2003, less than half the OECD average o f 17.9, but higher than Hungary (6.9), Poland (6.3), the UK (5.8) and Mexico (1 S), Notes: 1. * 2002; ** 2001; 2. Only for acute beds; 3. EU-IO: the countries who joined the EU after May 2004. Source: Health for All Database, WHO/Europe 2005; 3.65 Health care service utilization is growing rapidly but still remains comparatively low. There has been a 75 percent increase in the number o f outpatient contacts per person per year between 1993 and 2003 (Table 3.14). A similar pattern i s observed for inpatient admissions per 100 individuals, albeit with a smaller percentage increase (38 percent). However, utilization rates in Turkey continue to be much lower than the utilization rates in other European countries (Figures 3.22 and 3.23).92 On average, in the EU-25 countries, outpatient contacts per person per year are more than three times higher than in Turkey, while inpatient admission per 100 individuals are twice higher than in Turkey. 3.66 The implementation o f the Health Transformation Programme (HTP) which was initiated in 2003 i s expected to influence further the number o f health personnel, the availability o f diagnostic technologies and health care utilization in Turkey, thus close monitoring o f indicators in these areas i s recommended in the following years. 92 The Turkish population has a much lower median age and lower percentage o f the population above 65 years and these figures are not standardized or adjusted for age. 134  Out-patient contacts Inpatient admission Year per person per year p e r 100 individuals 1993 1.so 5.88 1994 1.60 6.17 1995 1.70 6.28 1996 1.so 6.50 1997 2.00 6.88 1998 2.10 7.12 1999 2.10 7.34 2000 2.40 7.53 2001 2.60 7.72 2002 7.91 2003 8.11 % change 75.33l 37.90 Figure 3.22: Inpatient admissions for selected Figure 3.23: Outpatient contacts for selected countries, 2003 countries, 2002 EUZ5 EU15(200Z) R o w h l y Rlal" CzcrhRlQubk SlOVlh POlUd B ~ U U spUn(1002) Poflvgrl Tuky 11 5 m 5 I 10 25 10 ooo zoo 100 600 800 moo uoo 11.00 1600 Source: Health for All Database, WHO/Europe, 2006. C3.Public expenditures on health 3.67 Turkey spent approximately 6.6 percent o f its GDP on health in 2004 (World Bank, 2006). Health expenditures relative to GDP in Turkey are higher than Romania, Russia, Mexico and Poland (all around 6 percent o f GDP) but lower than Spain, Portugal, Greece and Brazil (all between 8 and 10 percent o f GDP). In terms o f public-private mix, public expenditures on health in Turkey accounted for 68 percent o f total health expenditure in 2004, up from 61 percent in 1999. The allocation o f public health care expenditures among the various entities who incur them poses certain challenges mainly because o f the existence o f revolving funds in M O H and university facilities (Box 3.4). The combination o f poor health outcomes and relatively high expenditures suggests that 135 some important inefficiencies exist in the health care system. Measures to address these inefficiencies are proposed in section C4 below. Box 3.4: The role of the Revolving Funds in the Turkish Health Care System Public expenditures o n health consist o f expenditures incurred by the MOH, General Directorate o f Coastal Services, Universities, Social Solidarity Fund, other Ministries and agencies, c i v i l servants, Social Security Institutions (i.e. SSK, Bag-Kur and Emegli Sandigi), state economic enterprises and local authorities. Revolving funds exist in MOH and university facilities. MOH and university facilities offer their services t o individuals belonging t o the three SSI and payments for the services provided are channeled t o the revolving funds o f each facility. T h i s means that part o f the resources that appear as expenditures in the financial statements o f SSK, Bag-Kur and Emegli Sandigi, also appear as revenues in the financial statements o f the MOH and university facilities. In order not t o double-count these resources, one could (a) assign a l l expenditures t o the SSI, leaving the revolving fund accounts only with out-of-pocket expenditure^^^ or (b) discount the funds p a i d t o the MOH and university facilities f r o m the SSI accounts and present them under the revolving fund accounts. Table 7 presents public expenditures o n health using the second method. T h i s method n o t only guarantees that public expenditures are single-counted, but also clearly presents the size and importance o f revolving funds in the financing o f MOH and university facilities. Data show that the financing o f MOH and university facilities increasingly relies on revolving funds. Revolving funds represented 17.6 percent o f total health expenditure in 1999, increasing t o 28.7 percent in 2005, with the share o f the revolving funds o f the MOH increasing at a higher pace than the share o f the revolving funds o f university hospitals. Resources f r o m revolving funds for MOH and university facilities mainly finance personnel salaries, capital investment, maintenance and recurrent costs.94 The fact that providers are allowed t o allocate these resources with greater flexibility than any budget related resources introduces a variety o f perverse incentives into the system. First, it raises equity considerations since individuals with no ability t o pay for user charges cannot be treated in the facility o f their choice. Second, i t might lead to an increase o f utilization o f health care services as a result o f induced demand f r o m the side o f the health care provider^.'^ Third, having t o choose among patients whose institutions pay into revolving funds and individuals whose provision o f services i s financed by the budget, health care providers might show a preference for the first group, crowding out the poorest f r o m the system. This is especially the case after the introduction o f a performance-based payment mechanism for the distribution o f revolving funds among individual professionals working in hospitals. The importance o f revolving funds has increased even more f r o m 2005, with the transfer o f SSK facilities t o the MOH and the set up o f revolving funds for these facilities as well. Data o f January 2006 o n the number o f revolving funds o f the MOH facilities suggest that a l l hospitals under the MOH have operational revolving funds (Table 3.15).96 The distribution o f revolving funds across the country suggests that the poorest regions (East Anatolia and South East Anatolia) have the lowest number o f revolving funds. 93 These are co-payments o f the SSI enrollees o r out-of-pocket expenditure for individuals who are n o t qualified t o receive free care. 94 After deducting the shares o f the Treasury (1 1 percent), Social Assistance Fund for Children (around 1.5 percent), and MOH (2 percent), the rest o f revolving funds are distributed for salaries (up t o 50 percent o f the remaining balance), capital investment, maintenance and recurrent costs. 95 F o r these implications o f revolving funds, see W o r l d Bank (2003). 96 The total number o f revolving funds in MOH hospitals in the B o x Table is lower then the sum o f MOH and SSK hospitals in Table 5 due to the merging o f some o f the MOH and SSK facilities after the transfer o f the latter. 136 3.68 Total public expenditures on health have increasedfrom 3.3percent o f GNP in 1999 to above Spercent in 2004 and 2005 (Table 3.16). All available statistics confirm that public expenditures on health have risen significantly in 2005 as w e l l as in the first quarter o f 2006. Public expenditures on health o f social security institutions, Green card and C i v i l servants increased from YTL 13.9 billion in 2003 to YTL 17.9 b i l l i o n in 2005 (representing an increase o f 29 percent). At the same time, MOH and Universities budget expenditures o n health increased f r o m YTL 3.4 b i l l i o n in 2003 to YTL 6.2 billion in 2005 (representing an increase o f 82 percent). The share o f personnel expenditures o n health has increased during this period f r o m 1.1 percent to 1.6 percent o f GDP, w h i l e the share o f investments has fallen. Table 3.16: Public expenditure on health, 1999-2005 1999 2000 2001 2002 2003 2004 2005 SERVANTS + Ba -Kur 137 TPHE (Yo I GDP) I 3.28 3.65 4.19 4.78 4.82 5.25 5.06 Source: World Bank calculationsbased on data from MOF, SPO, MOH, SSK, BK, ED. 3.69 The bulk o f health care expenditures are on curative care and medical goods. Accurate data on the functional classification o f recurrent public expenditures on health and the distribution o f recurrent public expenditures by type o f provider are available through the Turkey National Health Accounts for the years 1999-2000 only (Figure 3.24). In 2000, around 57 percent o f all recurrent expenditures were spent on curative care and 34 percent on medical goods dispensed to outpatients. Only 5 percent was spent for prevention and public health services. Regarding the type o f provider, 54 percent was spent on services provided at hospitals and only 11.2 percent at primary care facilities. Available data for 2005 shows that expenditures on pharmaceuticals has increased to almost 40 percent o f total public spending on health. Figure 3.24: Functional classification o f recurrent public expenditures o n health, 1999-2000 Services o f Rehabilitative Care Ancillary Services t o B a l t h Care M e d i c a l G o o d s Dispensed t o Outpatients Prevention and Public B a l t h Services B a l t h Administration and Insurance 0 .o 10 .o 20.0 30.0 40.0 50 .O 60.0 70 .O Source: Turkey National Health Accounts 1999-2000, MOH, 2004. 3.70 Public expenditures on prevention are very limited. MOH is the major provider o f public health and prevention services in Turkey. MOH runs special programs for tuberculosis, malaria and cancer control, as well as maternal and child health and health education. The Institute o f Hygiene o f Refik Saydam i s responsible for the provision o f laboratory based preventive services. The share o f expenditures on preventive services has decreased from 11 percent in 1999 to 7 percent in 2005 (Table 3.17). In the area o f maternal and child health, the MOH spent only 1.6 percent o f i t s budget in 2005. This definition o f expenditure on preventive activities does not coincide with the one 138 employed by MOH, which includes primary care activities as well. However, all primary care activities cannot be considered as public goods or goods with large public externalities (e.g., personal services). Malaria program 1.48 1.85 1.90 1.75 1.39 1.05 0.93 Maternal and Child Health Program 2.14 2.21 1.95 2.22 1.89 1.81 1.61 Institute o f Hygiene Rafiq Saydam 1.07 1.04 0.94 0.94 0.77 0.71 0.68 Health Education 4.26 4.16 3.45 3.94 3.00 2.83 2.55 Cancer Control 0.01 0.01 0.02 0.10 0.06 0.05 0.10 Total preventive services 10.72 10.99 9.82 10.60 8.40 7.79 7.12 Memo item Outpatient Care 33.66 32.61 29.80 34.34 28.83 31.65 30.11 (General Dir. o f Primarv Health Care) Source: Health Statistics, MOH. 3.71 The increase in health expenditures in recent years is almost entirely explained by increased spending on hospitals and pharmaceuticals. Within these categories, increases in the quantity o f medical services and number o f prescriptions for drugs have contributed most to expenditure growth. For instance, the number o f outpatient visits in M O H hospitals increased from 110 million in 2003 to 130 million in 2004 to 170 million in 2005. 3.72 Across the different categories o f beneficiaries, the bulk o f the increase in expenditures is explained by increase in SSK enrollees and Green Card holders. The number o f SSK enrollees holding a health card increased from about 22 million in 2003 to about 31 million in 2005, while the number o f Green Card holders increased from about 6 million to 10 million during this period. This was accompanied by an increase in public reimbursement for drugs, which resulted in a shift from out-of-pocket expenditures to public expenditures. 3.73 A series o f events starting from February 2004 help explain the observed trends in expenditure on pharmaceuticals. The more important o f these are: Feb 2004 - Reference pricing introduced March 2004 - VAT on prescribed drugs reduced by 8 percentage points December 2004 - Protocol signed with suppliers o f drugs results in a discount between 7.5% and 14% in the price o f drugs Jan 2005(a) - Discount o f 8% negotiated on non-prescription drugs Jan 2005(b) - Prescription drugs extended to Green Card holders as part o f their benefit package Feb 2005(a) - S S K enrollees allowed to obtain drugs from private pharmacies Feb 2005(b) - Positive drug l i s t introduced Feb 2005(c) - SSK hospitals transferred and pharmacies closed 139  0 May 2005 - Green Card holders required to make copayments, like others June 2005 - A hrther reduction o f 8.83% obtained in drug pricing following the fall in the Euro. 3.74 These measures have had mixed results, as was expected. Some (those o f March 2004, December 2004, January 2005a, and June 2005) resulted in a fall in price o f drugs, some (May 2005) resulted in a fall in consumption o f drugs, while others (January 2005b, and February 2005a) resulted in increase in consumption o f drugs. The net result i s that the price o f drugs has, on average, fallen, but the quantity procured has increased significantly, particularly following the inclusion o f drugs in the benefit package o f Green Card holders and by easing the restrictions on SSK enrollees by allowing them to obtain drugs from private pharmacies. In effect, therefore, changes in pharmaceutical policies in the last year or so have resulted in new access for Green Card holders and improved access for SSK enrollees. This has significantly outweighed the gains realized through reductions negotiated in prices. 3.75 The increase in the quantity o f medical services partly reflects the increase in coverage and access, effectively covering the entire population o f the country- The increase in coverage, tantamount to effective introduction o f universal health insurance even before the official enactment o f the legislation, manifests itself in the increased utilization o f services by Green Card holders and SSK beneficiaries. B o t h groups o f beneficiaries have seen an impressive addition in the consumption o f health goods and services in the course o f the last 15 months. 3.76 - a n d partly the change o f incentives facing medical providers in MOH hospitals and outsourcing o f services. In parallel with the take-over o f S S K hospitals by MOH, the incentives structure for physicians working in MOH hospitals has also been changed and their reimbursement from the net collections o f Revolving Funds i s linked to their performance, measured principally by the revenue generated for the Revolving ~ the same Funds, which is a direct hnction o f the number o f services p e r f ~ r m e d . ’ At time, the limits to which MOH physicians can augment their salaries from the Revolving Fund have also been raised, to maximum o f 5 times the official salary for medical residents, 7 times for mid-level doctors and 8 times for senior doctors. There are concerns that this has led to supplier-induced demand and an increase in production, provision and utilization o f health services. In addition, almost all medical tests and diagnostics in MOH hospitals have been outsourced to private parties. W h i l e this has undoubtedly resulted in increased efficiency and lower unit costs o f tests, it has also led to an increase in the number and range o f available tests, and thus boosted expenditures. 3.77 While these changes have significantly improved access, increased utilization from the earlier low levels, and practically eliminated waiting lists, their impact on costs is of concern. These changes - by increasing the incentives for suppliers o f health services to induce demand - have led to an increase in production and delivery o f health services without necessarily improving the clinical quality and effectiveness o f health ’’The net figure i s determined by subtracting the 15% ‘tax’ o n Revolving Funds that i s collected by the Treasury and by subtraction o f the costs o f medical tests and diagnostics. 140 services. While some o f the pent-up demand i s almost certainly being met, new and unnecessary health expenditures are being generated in the system due to an increase in the total volume and m i x o f services as well as per episode o f care. 3.78 I n addition to the real increase in quantity o f services being produced and delivered, there are concerns that the billing system further exaggerates the number o f services provided. There is n o working and effective claim processing system at present, and it i s almost impossible to verify the nature, extent and authenticity o f the bills produced and presented by MOH hospitals. The extent o f this overstatement and inflation is difficult to establish and ascertain, but there i s little doubt that this practice is not uncommon. C4.Short-and medium-term measures to improve efficiency and enhance equity of public spending 3.79 Efficiency and equity should be simultaneous drivers o f health care reform. Health sector interventions include pure public goods, (e.g., communicable disease control, preventive activities, etc.), pure private goods (e.g., medical treatment for non- communicable illness), and a host o f health services in-between. Reflecting this mixed nature o f interventions, measures should be designed with the aim o f simultaneously improving efficiency and enhancing equity o f public spending. K e y directions and underlining rationale include: 0 strengthening o f public health and preventive services (public good rationale); 0 universal health insurance (equity rationale); 0 consolidating public hospitals, limiting hospital-based outpatient care, managing claims for hospital services, controlling the quantity o f medical services, containing pharmaceutical expenditures, and reforming provider payment mechanisms (all efficiency enhancing measures). Strengthening preventive health services 3.80 Low health outcomes, especially concerning child and maternal mortality, in combination with meager public resources suggest that Turkey should consider devoting a higher share o f public expenditures on preventive activities in the future in i t s effort to improve the health status o f the population. Furthermore, resources need to be targeted where they are most needed, if it is for the persistent variation o f health indicators between urban and rural areas and among regions to decrease. 3.81 Efforts to strengthen the delivery o f primary health care services, in general, and preventive activities, in particular, have been initiated with the piloting o f the family medicine system. Evidence from OECD countries indicates that a family- medicine centered health system has great potential to improve the cost-effectiveness o f care while strengthening equity. Successful implementation o f family medicine, in the 141 medium-term, will also contribute to reducing the problem o f excessive outpatient consultations at the hospital level and a fragmented referral system. The family medicine system will bring the physician and family members into closer and more personal contact, enabling the physician to play an important role in the family’s health and prevention o f illness. Under the family medicine system, simple and routine diagnostic services and consultations could be provided under a single-window and common illnesses could be treated across a broad spectrum o f medicine domains, including internal medicine, gynecology and pediatrics. Family medicine places special emphasis on continuity o f care and on quality o f health services, and integrates preventive health services with basic health services and provides the full package under one window. The family medicine system has the potential for the strengthening o f the patient referral system as well.’’ A family medicine pre-pilot has been completed in Duzce and a recent public opinion p o l l indicates high user satisfaction. An evaluation o f the Duzce pilot comparing with non-pilot sites i s also underway. The Duzce pilot is currently being rolled out in three provinces. Another seven provinces will be covered by the end o f 2007. The piloting and roll-out o f family medicine i s an important achievement given that these reforms have been planned in Turkey for over a decade and never implemented. Universal Health Insurance 3.82 The introduction of universal health insurance in Turkey will trigger a number o f changes not only in health financing but also in the production and delivery o f health services, and will generate new demands on the management and regulation o f the health care system. An estimation o f the effects o f moving from a regime o f social health insurance, administered by multiple agencies offering different sets o f benefits but covering less than the total population, to a regime o f social health insurance administered by a single agency offering uniform benefits requires an estimation o f the effects o f the proposed change on not only the flow o f funds but also on the level o f spending itself. The flow o f funds will change as a result o f the creation o f a single payer agency replacing the multiple agencies and the addition o f the hitherto uninsured to the pool o f insured. The level o f spending will change on account o f the “induction impact,” i.e., on account o f changes in utilization o f health services by the already insured as they adjust to new boundaries o f coverage, and the utilization patterns o f the newly insured. The level o f spending will also change on account o f the “efficiency impact,” i.e., on account o f changes in expenditures following efficiency improvement measures that may be instituted with Universal Health Insurance. 3.83 I n the new environment that will inevitably follow the introduction o f universal health insurance, the role o f public expenditures and institutions will become even more significant. The expansion o f UHI will undoubtedly place greater demand on public expenditures on health; at the same time, the introduction o f UHI offers a unique opportunity for the introduction o f efficiency and equity enhancing measures as well as 98 At the time o f writing, the family medicine pilot in Duzce has already made significant headway, with training o f over 100 family doctors completed and over two-thirds o f the 300,000 population o f Duzce registered with family physicians. 142 strengthening the overall effectiveness o f the health system. Indeed, the introduction of UHI can be a fiscally viable proposition only i f there are marked improvements in efficiency of public spending in addition to improvements on the revenue side o f the social security system. Consolidation of hospitals and greater manaEement autonomv 3.84 The transfer o f all SSK facilities to the M O H was the first step towards a more unified network o f health care providers. However, as documented earlier, there i s ample room for improving hospital efficiency. Prior to the merger o f MOH and SSK hospitals, public hospitals in Turkey were owned and run by MOH, S S K and Universities, and there was a lot o f variation in their performance and efficiency levels. 3.85 Consolidation o f hospitals and greater management autonomy could be the next step towards a more efficient provider network. Many MOH hospitals are too small in size to allow for efficient operation and provision o f care, and have significantly lower utilization rates compared to SSK and University hospitals. The fact that after the transfer o f SSK facilities, the MOH has found itself with more than one hospital in a great number o f places, especially in urban areas, offers the opportunity to carry out a re- organization o f i t s provider network. In addition, hospital managers enjoy very limited administrative and financial autonomy, and have very few incentives to adopt efficiency- enhancing measures. The proposed introduction o f UHI provides a good opportunity to further strengthen the gains from the merger o f MOH and SSK hospitals under MOH ownership and management. 3.86 While some gains in efficiency can be brought about simply by consolidating and reducing the number o f hospital beds in many provinces, further gains will come about only by improving efficiency in the use o f hospital resources and overall management and accountability. The separation o f provision and financing provides an opportunity to introduce innovative methods in management o f health facilities, which can be achieved by granting financial and administrative autonomy to public hospitals. The introduction o f hospital autonomy will require appropriate legislation that will allow for public assets to be managed outside the direct purview o f the government, and related laws and regulation would need to be amended in order to facilitate the transformation o f MOH and S S K facilities to autonomous bodies. This process has started, and the MOH is planning to implement a hospital reorganization pilot that would allow the testing o f different hospital reform models in Turkey, but it is essential that momentum is maintained during the transition phase. Building the capacity o f hospital managers to operate within an environment o f greater autonomy i s an important pre-condition for the success o f hospital reorganization and provider payment reforms. Another pre-condition for success i s the development and implementation o f an appropriate financial and performance accountability framework. For example, under a model o f hospital reorganization and autonomy, the hospital revolving funds will no longer operate as extra-budgetary funds but will be integrated into a comprehensive hospital resource management framework with appropriate rules and regulations agreed between the Social Security Institute and providers. As an interim short-term measure to prepare for these 143 reforms, capacity building and financial and performance reporting by the hospitals, including on the revolving funds could be encouraged. 3.87 Since the increase in health expenditures is almost entirely driven by increase in quantity, it is imperative that cost-containing measures focus on quantity. The challenge would be to reduce quantity without adversely affecting access, utilization and effectiveness o f services. The principal measure suggested to control quantity and dissuade frivolous use i s the introduction o f copayments for outpatient visits and for drugs (see chapter 4, section B 3 for the estimated impact o f various options). Even nominal levels o f copayments will reduce utilization o f health services and result in savings not only in the production o f health services but also in the consumption o f drugs (since almost all visits to a doctor result in a prescription). This i s perhaps the only demand-side measure that will have the effect o f rationalizing consumption. M a n aaina Pharmaceutical Expenditures 3.88 Drugs deserve particular attention as they are perhaps the single largest cost driver in almost all healthcare systems, and have been the most dynamically growing element in overall costs o f healthcare services in recent years. Growth in drug spending has outpaced total health expenditure in most OECD countries, and Turkey is no exception. Expenditure on pharmaceutical products constitutes a significant proportion o f total expenditures on health in Turkey, accounting for almost half o f all SSK, Emekli Sandigi and Bag-Kur spending on health. While pharmaceutical prices have increased broadly in line with general inflation, there has been a much larger change in consumption levels, including subtle changes in consumption in favor o f newer and more expensive drugs. One o f the reasons why Turkey spends a huge amount o n drugs and pharmaceutical products i s that most o f the insured population i s insensitive to pharmaceutical prices, with out-of-pocket payments for medicines constituting between 10 and 20 percent o f total medicine bill o f the insured.99 3 3 9 Managing consumption o f pharmaceuticals is critical in order to contain expenditures on drugs. Many countries have successfully adopted demand-side measures o f controlling consumption, and cost-sharing has proved to be the most effective such measure. loo The consumption o f pharmaceutical products among the insured i s actually not l o w by international standards, and there i s a strong scope for cost 99 Indeed, the h i g h percentage o f pharmaceutical expenditures in terms o f overall health expenditures is as m u c h a reflection o f l o w overall expenditures o n health as i t i s o f high expenditures o n drugs. In addition, drug prices are to a great extent at international levels, while other costs (salaries, etc.) are at a national level, thus increasing the relative weight/proportion o f drug expenditures in overall expenditures. loo In the Netherlands, for example, the introduction o f co-payments o n prescribed pharmaceuticals (a fixed amount per prescription) l e d t o substantial decrease in the total number o f prescriptions. In Germany, drug cost-containment measures take the f o r m o f cost-sharing, prescription limitations, reference prices and the pharmaceutical spending cap that makes physicians’ associations liable for any overspending with n o upper limit. These measures l e d t o substantive decreases in pharmaceutical expenditures for social health insurance, mainly attributable t o price reductions, changes in physicians’ prescribing behavior resulting in a reduced number o f prescriptions by 11.2 percent and increased prescriptions for generics. The French government imposes a fine o n pharmaceutical companies if pharmaceutical expenditures surpass budget ceilings either due to price o r quantity increases. 144 containment if indiscriminate consumption can be curbed. The policy table at the end o f this chapter provides a l i s t o f possible measures to curtail pharmaceutical expenditures in the short and medium term, including co-payments. The proposed measures are both demand and supply driven, including interventions that aim to reduce the utilization o f drugs through training for physicians on appropriate use o f drugs and the development o f standard treatment protocols. Reform ina urovider uavment mechanisms 3.90 Changes in provider payment systems should introduce incentives for physicians to provide quality care at lowest costs. Some o f these changes are already being planned, and a system o f paying family physicians o n the basis o f capitation i s being worked out. Physicians paid on the basis o f a capitation fee per enrollee receive a fixed amount per enrollee regardless o f the type and extent o f treatment sought. Physicians participating in this scheme bear most risks o f treating a patient, and therefore are likely to be conservative in the amount o f health care they provide. Introducing a system o f performance-incentives, especially for reaching high risk populations such as mothers and children with preventive health care services could be considered to complement the capitation system. Such a system would need to be extended to cover all outpatient care as family medicine is scaled up from the Duzce pilot. 3.9 1 Likewise, prospective payment mechanisms"' introduced at the hospital level would provide incentives to hospitals to contain costs. A project to identify diagnostic related groups (DRGs) is already under implementation and the medium-term objective o f hospital payment reforms i s to implement case-based payments. The implementation o f case-based payment systems has shown great promise in OECD countries. Under case- based payments, "money follows the patient" and hospitals that are unable to attract enough patients are forced to implement efficiency measures and in some cases merge with other hospitals or close down. Implementation o f DRGs must be combined with management reforms to give autonomy to health providers to allocate inputs as needed to implement efficiency gains. Without the implementation o f budget management reforms, DRGs will not work in Turkey. An interim measure i s to move to global budgets that cap health expenditures in hospitals, provide a disciplined environment for providers within lo* Prospective payment mechanisms rely o n the fact that services associated with a particular treatment are reasonably predictable and can be bundled into a group to which a monetary value can be attached. The hospital then gets reimbursed according t o a pre-fixed rate per bundle. Such payment mechanisms d o n o t encourage excessive use, since the hospital can conceivably make a profit (or a surplus) by being carefbl about inputs and hospital lengths o f stay. One o f the most widely-known prospective payment systems i s the Diagnosis Related Group, o r DRG. Developed to classify treatments according t o the resource costs o f its treatment, D R G s employ a complete and consistent coding system o f patient-level information obtained f r o m medical records t o establish and cost bundles o f appropriate inputs for one o r more diagnosis-based treatment. Typically requiring about 10-20 data items per discharge over a few years, the DRG system provides a valuable tool not only for reimbursing hospitals but also for overall planning and resource management.'" To be certain, a DRG-based system by itself will n o t necessarily promote efficient use o f resources. Hospital care providers and managers need the flexibility and tools t o actively manage their resources and redirect their use, which will ensure that cost-savings in treatment o f one case are passed through the entire system. 145 which to behave combined with capacity building o f providers to manage global budgets. While global budgets can lock in previous (inefficient levels o f spending) and do not generate incentives for efficiency, when combined with the implementation o f performance monitoring, can be a powerful too to control health expenditures in the short term and also build provider capacity to facilitate more sophisticated payments reforms. Contain ina outpatient care provided bv hospitals 3.92 A large number o f outpatient services in Turkey are provided in hospitals, accounting for almost 43 percent o f total costs o f outpatient services.'02 Outpatient services provided in hospitals cost significantly more than outpatient services provided in outpatient clinics, and i t is imperative that the introduction o f universal health insurance and family medicine be accompanied by a significant reduction in number o f outpatient visits in hospitals that are paid out o f the health insurance fund. This can be managed by establishing clear and transparent rules restricting reimbursement by the health fund o f outpatient treatment carried out in hospitals. Managing claims from providers to financiers 3.93 The creation o f the health insurance fund and accompanying internal control systems is expected to ease the problem of verifiing the validity of bills sent by heath care providers that some social security institutions, namely SSK, face at present. Important progress has been made in strengthening internal controls by developing a claims management system in the SSI as well as a pharmaceutical expenditure tracking systems. Furthermore, under the new system, the package o f health care services provided to the population should be clearly defined and its cost evaluated, while resources for the financing o f this package should be allocated to the health insurance fund. The fund, in turn, should be responsible for the management o f these resources and should not be bailed out in case its expenditures exceed its revenues, as it i s currently the case with most social security institutions. D. RURAL PUBLIC EXPENDITURES D1.Rural Investments 3.94 The level o f rural public investments is low in per capita terms owing to the large share o f employment in agriculture. The overall level o f rural public investments in Turkey is stable (at roughly YTL 350 million, in 1999 prices), amounting to 0.4 percent o f GDP. The rural public investments figure i s not limited to the agricultural sector; i t also includes environment, forestry as well as rural infrastructure including roads and electricity. Rural public investment represents around 12 percent o f total general government investment, which i s equivalent to the share o f agriculture in total GDP (Table 3.18). However, the level o f rural public investment i s still considered low, MOH (2004). lo* 146 due to the fact that rural employment i s as much as 30 percent o f total employment. At roughly U S $ 30 per rural inhabitant, rural public investments appear to be l o w by international comparison, but rural employment i s also declining as agriculture i s being modernized. Key to improving the contribution o f rural expenditures to the modernization o f agriculture and rural development will be: (i) the redeployment o f hnds in the transfer budget into investment oriented programs; and, (ii) a shift o f funds within the current non-transfer budget out o f salaries o f government employees into investments. Table 3.18- Total Public Investment and Rural Investment, (Million YTL, in 1999 real terms) 1999 2000 2001 2002 2003 2004 Total Public Investment 2,992 2,264 2,803 3,210 2,974 2,551 %of GDP 3.8% 4.1% 3.7% 4.0% 3.5% ‘Rural’ 327 384 318 365 351 328 Share of Total 10.9% 17.0% 11.3% 11.4% 11.8% 12.8% Of which, Agriculture 242 263 215 244 272 249 Share of Total 8.1% 11.6% 7.7% 7.6% 9.2% 9.7% Source: State Planning Organization. 3.95 The distribution o f investments within the agricultural sector is skewed to irrigation. The State Hydraulic Agency (DSI) has accounted for an average o f 45 percent o f rural investments in 1999-2004 (Table 3.19). Greater emphasis i s needed in areas such as land consolidation activities and storage and pre-marketing facilities (with particular focus on small farmers with weak l i n k s to markets). 3.96 Despite the high level o f irrigation investment, the execution o f these investments is very slow. D S I focuses on completion o f main canals and head works in certain areas but leaves on-farm irrigation activities (mainly executed by the former General Directorate o f Rural Services, GDRS) waiting in line for long period^."^ Over- programming and lack o f funds to complete the planned projects in time causes delays and uncertainties. Currently, D S I would need more than 20 years to complete i t s 183 projects with the existing annual investment allocations. Though D S I investments expenditures have been increasingly focused away from the portfolio o f dormant/ l o w return projects over the past two to three years, this trend needs to be accelerated in a way which gives greater emphasis to completing schemes down to the farm level so that the benefits to a wider group o f beneficiaries can be more tangible in the medium term. 3.97 At less than a quarter o f the total, rural non-agricultural investments are quite low. More than half o f this investment i s focused on rural roads and there is relatively little investment for drinking water and sanitation. At only US$4 and $US1.5 per rural inhabitant per year, respectively, these investments in roads and non-irrigation water supply are low by most standards. With approximately 317,000 km o f rural roads network and 478.7 k m / l O O O k m 2 rural road density Turkey can be considered quite 103 GDRS was closed in 2004, with i t s functions allocated to the Ministry o f Agriculture and Rural Affairs and Special Provincial Administrations. 147 developed (compared to Poland 343.9 k m / l O O O k m 2 and Thailand 124 k m / l O O O km2). Nevertheless, only approximately half o f these roads are asphalt roads, the remaining being stabilized, graveled or earth roads, which require considerable upgrading. Regarding drinking water, 25 percent o f villages and sub-villages and 15 percent o f rural population is without drinking water facility. Therefore, investments particularly in these areas are o f great importance for Turkey. Closer coordination between line agencies as well as SPO in determining priorities for rural investments i s necessary if this imbalance i s to be adjusted. 3.98 Participatory approaches in investments are needed. In order to improve focus and get better leverage o f public investments in the rural sector, local communities, farmers' organizations and/or NGOs should be involved in the design, implementation and management o f investments. When investments become more demand-driven they usually gain greater relevance and can be held to higher accountability in terms o f efficiency and timely execution. Moreover, when beneficiaries are given greater voice in deciding on the local focus o f public investments they can be expected to be more willing to mobilize greater co-financing for these infrastructure investments. Table 3.19 Investment by Rural Agencies (Constant 1999 Trillion TL) 1999 2000 2001 2002 2003 2004 Agriculture 242 263 215 244 272 249 Total Agricultural Share, of which: 74% 68% 68% 67% 77% 76% DSI-Irrigation 47% 44% 39% 38% 48% 48% GDRS-Irrigationand Other Agriculture 9% 8% 7% 7% 7% 8% Ministryof Agriculture and Rural Affairs 4% 4% 7% 6% 7% 7% Ministry of Forestry and Environment 3% 3yo 3yo 4% 5% 5yo GDF 6% 5% 5% 4% 4% 4yo Other 5% 5Yo 6% 7% 7% 4% Rural Non-Agricultural 85 121 103 120 79 79 Total Share, o f which: 26% 32% 32% 33% 23% 24% Rural Roads-GDRS&GDR 14% 13% 20% 18% 14% 13% Rural Drinking Water-GDRS 8% 7% 7% 6% 5% 5% Rural Electrification-TEDAS 4% 11% 5% 9% 4% 6% Total 100% 100% 100% 100% 100% 100% Source: State Planning Organization. D2.Non Investment Spending for Rural Areas 3.99 Agricultural transfers have been restructured- The Government started in 2000 to change the system o f agricultural support to promote fiscal stabilization and allocative efficiency. By the end o f 2002, the implementation o f the reform o f agricultural transfers had significantly reduced artificial incentives for inputs and particular crops and switched the main focus o f agricultural transfers to the Direct 148 Income Support (DIS) Program.‘04 By 2004, the annual fiscal cost had been reduced from US$6.1 billion (3.1 percent o f GDP) in 1999 to US$2.4 billion (0.8 percent o f GDP). This has been achieved largely through elimination o f credit subsidies and reduction o f debt write-offs, and substantial reduction in state financed crop purchases and fiscal transfers to cover the losses o f state agricultural enterprises. (These all fall under the “other” category in Table 3.20.) At the same time, the DIS Program expanded to become the main instrument o f rural income support. In 2003-4, the D I S Program covered over 75 percent o f farmers and accounted for three quarters o f the total agricultural transfer budget (Table 3.20). 3.100 -but some reversals have been noted recently. In September 2006 the Agriculture Product Office (TMO) announced the intention to borrow up to $400m from international markets to purchase between 100,000-125,000 tones o f hazelnuts at prices higher than market prices so as to support distressed producers. Such interventions are o f concern because they: (i) represent a material deviation from agricultural sector policies o f de-linking subsidies from production decisions; and, (ii) reflect a structural weakness in the budgeting process whereby duty losses could be created by TMO (and possibly other similar entities) through state guarantees. This type o f allocation o f public resources should be capped and subject to the scrutiny o f the parliamentary process, while amendments to the public debt law may also be appropriate so as to limit the Treasury’s ability to finance duty losses through state guarantees to be paid in the hture. - Table 3.20 Total Agricultural Transfer Budget, 1999-2005 Amounts (million US$) 1999 2000 2001 2002 2003 2004 2005 Preliminary Direct Income Support 0.0 2.7 68.3 1,246.7 1,556.7 1,788.0 1687.5 Premium Payments 234.6 312.0 296.6 149.8 189.9 212.1 218.8 Alternative Crops and A S C U 0.0 0.0 0.0 0.0 34.2 69.0 9.4 Restructuring A n i m a l Husbandry 2.4 17.6 35.9 55.3 84.5 120.0 123.8 Agricultural Set-Aside 0.0 0.0 0.0 0.0 0.0 0.0 3.0 Rural Development Grants 0.0 0.0 0.0 0.0 0.0 0.0 15.0 Other 5,838.2 2,804.9 1,470.7 235.2 266.8 206.0 105.8 Total 6,075.2 3,137.2 1,871.5 1,686.8 2,132.2 2,395.1 2163.1 Shares of Total (in YO) Direct Income Support 0.0 0.1 3.6 73.9 73.0 74.7 78.0 Premium Payments 3.9 9.9 15.8 8.9 8.9 8.9 10.1 Alternative Crops and A S C U 0.0 0.0 0.0 0.0 1.6 2.9 0.4 Restructuring A n i m a l Husbandry 0.0 0.6 1.9 3.3 4.0 5.0 5.7 Agricultural Set-Aside 0.0 0.0 0.0 0.0 0.0 0.0 0.1 Rural Development Grants 0.0 0.0 0.0 0.0 0.0 0.0 0.7 Other 96.1 89.4 78.6 13.9 12.5 8.6 4.9 Source: Undersecretariat o f Treasuly and World Bank Stafcalculations. I O 4 F o r details see, Turkey: Review o f the Impact o f the Reform o f Agricultural Sector Subsidization” W o r l d Bank, 2003. 149 3.101 I n 2004, the Government took the initial steps of reshaping the agricultural transfer budget with the aim of initiating programs more broadly targeted to rural development. There are two key factors motivating this approach. First, there i s the widely held view that the D I S Program, though serving the useful purpose o f income support for farmers (in the transition after subsidy reduction), should be better targeted to exclude wealthier farmers. Second, there is the view that the DIS Program needs to be accompanied by transfer policies which aim to promote productivity and market development more directly. 3.102 The other encouraging aspect o f planned adjustments is that 30 percent of the transfer budget is to be allocated to productivity enhancing measures. These measures intend to include: i)grants to farmers for stimulating environmentally sustainable agricultural practices and alternative crops to tobacco and hazelnut; i i)investment grants awarded on a competitive basis for agribusiness and other rural SMEs, as well as for public providers o f infrastructure and services; and, i ii)agricultural sector risk mitigation through partial coverage o f crop insurance premia for farmers. These areas are the main beneficiaries o f the planned reduction in the share o f the D I S program. - Table 3.21 Agricultural Support Instruments’ Shares of Transfer Budget (in%) Current Share, 2004 Shares Targeted by Government over 2006-2010 DIS Payments 77 45 Premium Payments 11 13 Animal Husbandry Support 7 12 Alternative Crop Programs 2 5 Environmental Set Aside Programs 0 5 Crop Insurance Premia 0 5 Rural Development Grants 0 10 Other instruments 3 5 Total 100 100 Source: Undersecretariat of Treasury and Draft Agricultural Support Strategy 3.1 03 The broader strategic agricultural development policy challenge for the Government is to once again contain and reduce premium payments and animal husbandry support. The agricultural transfer budget intends to give roughly equal weight to social aims (such as DIS) and productivity enhancing measures (such as alternative crop programs and rural development grants-Table 3.21). However, as has been demonstrated in the recent Country Economic Memorandum (World Bank, 2006), the trend in OECD countries i s away from these largely output-focused budget subsidies (and output-based transfers from consumers through import tariff policy), towards even greater emphasis on investments to more clearly and efficiently target programs that can improve agricultural productivity and off-farm rural employment growth. 3.1 04 The rural consolidated budget (including all transfers, investments and recurrent costs) i s characterized by roughly equal shares for transfers and recurrent expenditures (Table 3.22). The real level o f the current expenditure budget has been fairly constant over the past few years, but the level o f transfers was cut by roughly 60 150 percent in real terms. At the same time, the ratio o f current to investment expenditure has been stable at roughly 2 : l over the past five years. Another striking fact i s that, as much as 90 percent o f recurrent spending o f rural agencies is allocated for salaries alone. Therefore, there is limited funding (only 2-3 percent o f recurrent budget) for logistical and operational spending, which is extremely vital for effective service delivery. Table 3.22: Consolidated Budget Expenditures - Including Recurrent Costs (Constant 1999 prices, Million YTL) 1999 2000 2001 2002 2003 2004 Total Rural Expenditures 3,389 2,301 1,869 1,646 1,538 1,649 a) Investment - Share o f Rural Budget 9% 16% 18% 27% 17% 17% b) Transfer - Share o f Rural Budget 71% 50% 45% 31% 37% 43% c) Recurrent- Share o f Rural Budget,: 20% 33% 37% 42% 45% 41% - O f which, Personnel as Share o f Rural Budget 18% 28% 34% 37% 41% 35% Source: State Planning Organization, State Institute o f Statistics. Notes: Totals exclude spending by local government, spending o n SOEs, and spending o n ‘special’ funds, but includes the agricultural subsidy (transfer) budget. Rural comprises expenditures by MARA, MEF, DSI, GDF, and GDRS. 3.105 Over reliance on government staff for the planning and execution o f agricultural and rural development activities should be addressed (Box 3 S). When more than a third o f the rural expenditures are on salaries o f government employees, this leaves significantly small room in the budget for investments as w e l l as other operation and maintenance expenses. However, it should be noted that rural investments are expected to be o n an upward trend further to the implementation o f the Rural Support (“Koy-Des”) project for clean water and rural roads. With greater emphasis given to participatory approaches and private sector involvement in service provision and expansion o f the role o f cooperatives, producers’ unions and NGOs, more effective and efficient use o f funds could be achieved. Box 3.5: Examples from Rural Agencies, Issue of High Personnel Expenditures One example is State Hydraulic Works (DSI), which turned approximately 90 percent o f the management o f irrigation schemes to Waters Users Associations, but cut only 10 percent o f its work force and s t i l l continues t o allocate 25-30 percent o f its budget for personnel. Another example is Ministry o f Agriculture and Rural Affairs (MARA) whose budget i s soaked up by personnel expenditures (70-75 percent o f total budget) should consider contracting out mainly extension activities while reducing such personnel. Salary costs used t o account for approximately 75 percent o f the budget o f General Directorate o f Rural Services (GDRS). After i t s closure in M a r c h 2005, GDRS’s head quarters staff were transferred to MARA (roughly 5,000 employees) and the sizable regional staff (roughly 48,000 employees), to Special Provincial Administrations (SPAs). However, the issue o f GDRS’s overstaffing hasn’t been solved but merely passed o n t o MAR4 and SPAs. Moreover, the ambiguity regarding the transfer o f the 2005 budget o f GDRS t o the related agencies resulted in leaving ex-personnel o f GDRS predominantly idle. 151 E. SUMMARY OF EXPENDITURE REFORM OPTIONS IN KEY SECTORS 3.106 Structural expenditure reforms to contain costs should be combined with expenditure reallocations to maximize the expected efficiency gains. Table B in the Overview section o f the study summarizes some k e y reform options in expenditure programs discussed in this chapter with a view to promoting expenditure efficiency and realizing fiscal costs over the medium term. Some o f the proposed measures aim to directly contain short-term expenditure pressures, especially in health care, and the Government would be advised to consider implementation in 2007. Other reforms should be planed over the medium term, during 2008-09, but preparatory steps would need to be taken earlier-such as, for example, with the design o f new incentive systems and evaluation mechanisms for educational institutions, or a possible deepening o f the recent pension system reform. The table provides an illustrative implementation time frame o f the proposed reform options. Some proposed measures would entail expenditure reallocations or an increase in expenditure levels over the medium term-such as, for example, measures to strengthen pre-primary education or preventive health care (these options are highlighted in grey in table B). However, in some cases, increased expenditures could improve the fbnctioning o f a sector and thus facilitate the realization o f cost savings-such as, for example, the increase in preventive health care expenditures that would help contain pharmaceutical and hospital costs over the long run. Increased spending in specific areas would also facilitate better overall expenditure efficiency and thus achievement o f superior outcomes for a given amount o f resources-such as in the - case, for example, o f higher pre-primary education spending, increased n o n personnel spending in public schools, or strengthening o f overall rural investments and redeployment towards productivity-enhancing programs. 152  CHAPTER I V MANAGINGMEDIUM TERM EXPENDITURE PRESSURES 4.1 Despite the strong fiscal adjustment since 1999, expenditure pressures in key sectors as well as the need to enhance growth-promoting expenditures so as to meet development challenges are creating stress on the budget. As documented in the previous chapters, public expenditures in Turkey are relatively oversized compared to other fast-growing emerging market economies and the high public debt to GNP ratio continues to be a constraint+alling for large primary surpluses for a few more years.Io5 The development challenges on the way to the EU and the expenditure pressures in some key sectors are creating additional stress o n public spending. Freeing up resources over the medium t e r m in order to meet development challenges requires significant expenditure efficiency improvements and reallocations. This chapter reviews medium- term expenditure projections in key functional areas and elaborates a framework for the assessment o f expenditure pressures in view o f the budgetary envelope that would be consistent with a strong enough framework for debt sustainability. 4.2 The first section reviews the Medium-Term Fiscal Strategy (MTFS) presented by the Government with the 2006 budget. The second section reviews projections o f medium-term expenditure trends in four k e y functional areas analyzed in the previous chapter: education, pensions, health care, and rural expenditures. Section three elaborates the framework for the assessment o f expenditure pressures (or required fiscal space) by comparing the projected expenditures over the medium term with a sustainable expenditure envelope that i s consistent with continuous progress in reducing the debt ratio-a precondition for preserving a sound macroeconomic framework. Different scenarios o f expenditure pressures are examined depending o n the presumed ambition o f expenditure policy and the targeted reduction o f public debt. lo5 n a comfort The level o f public debt in proportion t o G D P that will need to be reached t o place Turkey i zone in the face o f surrounding vulnerabilities will depend inter alia o n the maturity and currency composition o f the stock o f debt. The still short average maturity o f domestic debt and the large share linked to foreign exchange increase Turkey's vulnerability t o liquidity and exchange risks. These risks need to be mitigated by a l o w level o f the debt ratio in proportion to GDP. With increasing average maturity and less foreign exchange exposure, the debt ratio consistent with a comfortable level o f risk would, all else equal, be higher. 153 A. THE GOVERNMENT'S MEDIUM-TERM FISCAL, STRATEGY lo6 4.3 The government plans, programs and budgets are informative sources for understanding foreseen future policy actions and available budgetary revenue sources. The Turkish government publicizes periodic documents which provide information on current and expected future policies, economic targets and budgetary resources. Such documents include development plans, annual programs, government budgets and other similar documents. This section first looks at the Government's plans, programs and budgets in order to identify foreseen policy changes, available budgetary resources and possible bottlenecks. Within this perspective, the following section evaluates the government programs both at the central government level and the general government level. I t also discusses how the authorities see fiture challenges and how these are addressed in policy documents. (0 Central Government 4.4 The first Medium Term Fiscal Strategy (MTFS) projects strong fiscal performance in 2006. The authorities have introduced their first medium t e r m budgeting framework at the central government at the end o f 2005. The medium term budgeting process has been initiated with the issuance o f the Medium Term Macro Framework (MTMF) as the Council Ministers' Decision. The MTMF sets the medium-term polices and determines sectoral priorities which create the basis for the MTFS. The MTFS covers the period o f 2006-2008 and it included revenue projections and indicative expenditure ceilings for 2007-2008 alongside the appropriations for 2006. lo7 Considering first the 2006 budget, the targeted primary surplus o f about 5 percent o f GNP i s consistent with the IMF program target o f 6.5 percent o f the GNP primary surplus for the whole public sector."' Table 4.1 below provides the institutional breakdown o f the public sector primary balance according to the IMF program definitions while Annex I1 provides the detailed economic and functional classification o f Central Government expenditures over 2004-2008. As can be seen, the central government budget i s expected to produce 5.1 percent o f GNP primary surplus in 2006, while the remaining part i s expected to come from the rest o f the public sector. The projected primary surplus target o f the central lo6 At the time o f drafting, the new M e d i u m T e r m Fiscal Strategy (MTFS) covering the period 2007-2009 was not released by the government and therefore the analysis in this section i s based o n the previous M T F S which covered 2006-2008. However, this section will be updated in line with the n e w M T F S u p o n receiving the related data f r o m the government authorities. lo' The indicative ceilings were not provided for the Regulatory and Supervisory institutions as these institutions are not subject t o article 16 o f the P F M C Law. 'Os I t is s t i l l not clear at this stage h o w the IMF will follow the public sector primary balance for 2006 and beyond since the coverage o f the central government budget includes n e w institutions such as Regulatory and Supervisory Agencies (RSAs), which were not part o f the public sector definition o f the IMF. Until 2005, they were only covered within the context o f transfer f r o m the budget, not with their total revenues and expenditures. Although the impact o f the n e w l y included institutions o n the primary surplus i s negligible, i t s t i l l needs t o be clarified whether there will be any changes in the definition o f the consolidated public sector primary balance definition to cover the whole central government budget institutions. 154 government i s consistent with the past trends. This can also be said for all other institutions covered under the public sector except the SEEs. The declining trend in the primary surplus o f SEEs is mainly the reflection o f the privatization o f profitable SEEs such as TUPRAS, TELEKOM and POAS. The temporary decline in EBFs’ primary balance in 2005 came from a capital injection to the SEES under the privatization portfolio. The EBFs are projected to yield a primary surplus o f 0.2 percent o f GNP in 2006. Table 4.1: Institutionalbreakdowno f the primary surplus, 2004-2006 (“ho f GNP) 2004 2005 2006 Public Sector primary balance 7.1 6.2 6.5 A- Central Government 5.1 5.0 5.1 B- Rest o f public sector 2.0 1.2 1.5 EBFs 0.2 0.2 0.2 Revolving Funds 0.2 0.2 0.2 Social Security Institutions -0.1 -0.0 0.0 Unemployment Insurance Fund 0.4 0.4 0.4 Local administrations 0.2 0.1 0.1 SEES 1.1 0.4 0.6 Source: IMF 4.5 The indicative primary balances for the central government in 2007 and 2008 presented in the MTFS are lower than in the past years. MTFS presents indicative primary surplus targets o f 4.6 percent in 2007 and 4.5 percent in 2008 for the central g~vernment.’~’ This will require that the rest o f the public sector produce more than 1.5 percent o f the GNP, to reach the public sector primary surplus target o f 6.5 percent o f GNP. The fall in primary surplus results from a decline in central government revenues. The authorities are projecting declines in both primary revenues and expenditures. However, the reduction in primary revenues between 2006 and 2008 more than offsets the reduction in primary expenditures, giving rise to deterioration in the projected primary fiscal balances. I t should be noted, however, that these projections are surrounded by uncertainty as the Government has revised the M T F S in less than three months, when submitting the budget to the Parliament in October 2006. Box 4.1 provides a preliminary update for 2006 realizations and 2007 budget projections. Box 4.1: 2007 Budget and Medium Term Fiscal Strategy The second medium-term fiscal plan covering 2007-2009 was prepared by the Government with the issuance o f M e d i u m T e r m Macro Program in the Official Gazette dated June 13, 2006. The second macro plan sets out the macro policies and targets for the next three years, defines sectoral policies and priorities. The related M e d i u m T e r m Fiscal Strategy which aims to map macro and sectoral policies and priorities o f the M T P into the budget was issued o n July 15,2006. Although the new M T F S has a three years perspective, the m a i n focus o f the strategy is still o n the current (first) year. In this regard, while there i s only 1 percent divergence between the 2006 central government non-interest expenditures realization estimate and the 2006-2008 M T F S target, this difference i s slightly IO9 The analysis in this section is based o n the budget and G N P figures presented in the M T F S . 155 above 13 percent for 2007 and 2008 data presented in the last M T F S . Some o f this divergence can be explained through higher than expected salary increase for the c i v i l servants in 2006 and increased estimate o f the central government expenditures for universal health insurance premium for the poor due to increased number o f green card holders. These differences highlight the need for more attention in the preparation o f the M T F S so as t o enhance i t s credibility. The Government revised its M T F S in less than three months after i t was announced in July 2006, when submitting the budget t o the Parliament in October 2006, which m a y also affect the credibility o f the MTFS. Since the details o f 2008 and 2009 budget are not available, the following analysis focuses only o n 2007 central government budget. The 2007 central government budget provides a consistent framework with the 6.5 percent target o f public sector primary surplus under the IMF program definition, with a central government primary surplus projected at 5 percent o f GNP. Although the 5 percent o f GNP primary surplus target for 2007 is quite lower than the estimate for 2006, the better than expected performance o f the 2006 budget i s a reflection o f one-off revenues such as social security arrears restructuring collection and collection f r o m the former owners o f SDIF banks (total o f 1.4 percent o f GNP according t o IMF calculations). This l e d t o a reduction in the deficit o f the social security institutions’ deficit and therefore current transfer o f the central government. While the revenue side o f 2007 budget does not foresee any policy change, f r o m the expenditures side, the m a i n change in the assumptions o f 2006 and 2007 i s the introduction o f Universal Health Insurance system and o f a n e w social security reform law. The increase in personnel expenditures partly reflects the introduction o f the n e w social security legislation, as the premium base o f the c i v i l servants will b e broadened t o cover almost their entire salaries so that the Government will pay an increased amount o f social security contributions. A s i t is envisaged in the enacted social security reform legislation, the Government i s required to pay premium increase o f the c i v i l servants t o eliminate any possible reduction in their wages. Personnel expenditures will also increase owing to ad hoc salary hikes and due to indexation to higher than expected inflation. A s a result, the quality o f 2007 fiscal stance requires special attention since the fiscal space gained f r o m the social security transfers will be allocated t o personnel expenditures. 4.6 The reduction in corporate income tax rates as well as declines in other tax and non tax revenues are expected to weaken budget revenues in the short term. Central government revenues are expected to decline by 2.5 percentage points, from 32.3 percent 156 o f GNP in 2005 to 29.8 percent in 2008. Although the breakdown o f this reduction is not possible in IMF program definition as it would require allocating tax rebates according to tax types, the total decline in IMF program definition would be 2.8 percent o f GNP. At unadjusted basis, the main declining revenue items are:"' From the tax revenue side (1.4percent o f the GDP) > The already announced reduction in the corporate income tax rates from 30 percent to 20 percent (0.6percent o f the GDP) 9 The negative impact o f the falling interest revenues o n personal income tax due to the decline in inflation and reduction o f number o f brackets as well as upper rate for personal income tax (0.3percent o f the GDP) 9 Fall in special consumption tax (STC) revenue due to not increasing the rates (0.7percent o f the GDP) From the non-tax revenues ( 1.1 percent o f GDP) 9 No dividend contribution from Turkish Telecom starting in 2007 9 State Bank dividend contribution expected to decline I t should be noted that to the extent lower income tax rates tame incentives for tax evasion, the impact o n tax collections could be mitigated over the medium term. However, for effectively reducing tax evasion, reduced C I T and PIT rates should form part o f a broader reform drive to improve formalization, tackling other obstacles to formality-such as, for example, high labor taxes and labor market rigidities-and strengthening tax administration.' '' Table 4.2: Central Government Budget Revenues in the MTFS 2005 11 2006 2007 2008 Tax revenues 27.1 27.0 26.3 25.7 Taxes o n Income 8.3 8.1 7.8 7.4 Personal Income Tax 5.5 5.4 5.3 5.3 Corporate Income Tax 2.8 2.7 2.5 2.2 Other Taxes o n Income Profits and Capital Gains 0.0 Taxes o n Property 0.4 0.9 0.8 0.8 O f which inheritance and gift tax 0.0 0.0 0.0 0.0 Domestic Taxes o n Good and services 12.6 12.3 12.1 11.8 Domestic VAT 4.5 4.7 4.8 4.9 SCT 6.9 6.9 6.5 6.2 Other 1.2 0.7 0.7 0.7 O f which M o t o r Vehicle Tax 0.5 0.6 0.6 0.6 Taxes o n international trade 4.5 4.5 4.4 4.4 Other Taxes 0.5 0.5 0.5 0.5 Fees and Fines 0.7 0.7 0.7 0.7 Non-Tax revenues 4.2 4.1 3.3 3.1 Capital Revenues 0.4 0.4 0.4 0.3 'loThe figures may not sum up t o the decline in budget total revenues due to changes in other items not stated here. 'I1OECD, "Economic Survey: Turkey 2006" provides an overview o f options t o improve formalization. 157  Grants and Aid 0.1 0.1 0.1 0.1 Annexed Budget Income 0.4 0.5 0.5 0.5 RSAs 0.1 0.1 0.1 Total Revenues 32.3 32.2 30.6 29.8 Direct 32.7 32.1 31.9 31.4 Indirect Taxes 67.3 67.9 68.1 68.6 4.7 The government predicts 1 percent decline in primary expenditures between 2005 and 2008 in addition to 3.4 percent fall in interest expenditures. The expenditure adjustments are projected to come mostly from a decline in the social security transfers recorded under the current transfers. A slight decline in personnel expenditures i s also projected. Table 4.3: Adjusted Primary expenditures of the central government (Yo o f GDP) 2005 2006 2007 2008 01 - Personnel 6.5 6.6 6.5 6.2 02- Social Security contribution 11 1.4 1.4 1.2 1.1 03 - Goods and Services 21 2.1 2.5 2.4 2.3 05 - Current Transfers 3/ 10.3 9.3 9.1 9.0 06 - Capital Expenditures 2.0 2.3 2.2 2.1 07 - Capital transfers 0.3 0.3 0.3 0.3 08 - Lending 0.3 0.7 0.5 0.5 09 - Contingencies 0.0 0.3 0.3 0.3 Total primary expenditures 22.9 23.4 22.4 21.8 Memo item Interest paym ents 9.4 8.5 7.4 6.0 Note: 2005 data are for the consolidated budget. 11Public personnel health expenditures under goods and services has been moved to social security premium in 2005 and 2006 for the purpose o f comparability. 21 Health expenditures for public personnel and green card holders has been deducted f r o m the goods and services data in 2005 and 2006. 31 2005 data was adjusted with the revenue share o f local governments and funds, and health expenditures for the public personnel and green card holders moved t o social security transfer under the current transfer. Source: MOF, and W o r l d Bank Staff calculation 4.8 Social Security Transfers from the consolidated budget are targeted to decline by 0.7percent o f GNP between 2005 and 2008. As a result o f the social security reform adopted in 2006 (chapter 3, section B) the social security deficit i s expected to cease to grow and decline by approximately 1 percent o f GNP until 2020 compared to pre-reform case. Over the medium term, the government has committed to containing the social security system deficit at the percent level o f 4.5 o f GNP. With the new social security reform legislation Turkey will move to a UHI system starting in 2007. This will bring 158 some changes with respect to the reporting o f social security transfers. Therefore, the social security system deficit announced by the M T F S for 2007 Unadjusted and Unadjusted SSIs deficit and 2008 requires some adjustment 5.0% in order to have comparable data. Without any adjustment, the total e 4.8% 4.6% - social security transfers o f the 4.4% M T F S do not show any decline. 5- - 4.2% 0 s 4.0% However, after the adjustments (explained in Box 4.2 below), the 3.8% overall deficit o f social security 3.6% transfers i s expected to decrease to 2005 2006 2007 2008 4.1 percent by the end o f 2008 Unadjusted social security deficit 0 adjusted social security deficit from 4.8 percent in 2005. Box 4.2: T h e Methodological Changes in the Reporting of Health Expenditures and Social Security Transfers Turkey will start implementing a universal health insurance system starting in 2007 as already declared in the medium term fiscal framework and annual program. With the implementation o f this system, a l l citizens will be covered by the health insurance system and the health insurance premium for poor people will be paid by the Government. Health expenditures for c i v i l servants will be converted into a health premium and will be paid by the related public administrations. This new implementation will bring about two changes in the reporting. -- First, as o f 2007 health expenditures related to green card holders (the poor) will be converted into premium payments and reported as transfers t o social security institutions under current transfers rather than as goods and services expenditures. -- Second, health expenditures for c i v i l servants previously reported under good and services will b e reported as social security premiums o f public personnel. Therefore, in order t o have comparable data for the social security transfers consistent with the 4.5 percent deficit target, YTL 2,400 and 2,500 m i l l i o n expected health premium payments to be made by the Government for poor people in 2007 and 2008 respectively have been deducted f r o m the transfers t o SSIS. Table: Adjusted Social Security Transfers for 2005-2008 159 (ii) General Government 4.9 General government primary revenues are projected to decline only marginally towards the end of the period. Although the authorities foresee a decline in central government revenues, the increase in primary revenues o f the social security system is expected to compensate for this decline and leave the total revenues close to the level in 2005. The introduction o f the UHI in 2007 is estimated to increase premium collections in the social security system. Table 4.4: General Government Revenues and Expenditures ( % o f GDP) 1999 2004 2005 2006 2007 2008 Total Revenues 34.23 41.09 42.31 42.83 41.76 40.95 Primary Revenues 33.42 39.05 39.65 40.22 40.14 39.39 Current Expenditures 17.25 17.53 17.09 16.77 17.79 17.30 o f which personnel 10.41 10.13 8.78 8.48 8.15 7.82 Investment Expenditures 4.42 3.17 3.80 3.73 3.69 3.68 Transfer Expenditures 25.72 25.65 22.22 22.13 21.19 19.97 Current Transfers 24.45 24.90 21.56 21.24 20.51 19.30 of which interest payments 14.43 13.48 9.55 8.67 6.99 5.96 Capital Transfers 1.27 0.75 0.66 0.88 0.68 0.67 Total Expenditures 47.39 46.35 43.11 42.62 42.67 40.96 Primary Expenditures 32.97 32.86 33.57 33.96 35.68 35.00 Borrowing Requirement (inc priv) 13.16 5.26 0.81 -0.21 0.91 0.01 Borrowing Requirement (excl. priv) 13.24 5.65 1.62 1.46 1.63 0.68 Primary Surplus 0.5 6.2 6.1 6.3 4.5 4.4 1/ Excluding Regulatory and Supervisory Institutions Source: SPO and the WB staff calculation 4.10 Primary expenditure o f the general government is projected to increase by almost 1.5 percent o f GNP. Current transfers, excluding interest payments, are expected to grow faster than GDP, reflecting increased premium payments under the new UHI system (see chapter 3, section C). Investment expenditures are not projected to increase in proportion to GDP, despite the need to bridge significant infrastructure gaps discussed earlier (chapter 2, section C). 4.11 Similar to the case o f the central government, the primary surplus o f the general government is expected to decline in 2007 and 2008. The predicted rise in general government expenditures will lead to a significant decline in primary surplus as this rise i s not expected to be compensated by revenue increases. However, i t i s not clear at this stage whether the projected decline in primary surplus is a policy choice or a trend to be addressed by measures beyond the current budget. At the current stage, neither the government MTFS at central government level, nor the projections at general government level provide a coherent estimation o f fiture expenditure pressures or measures for 160 addressing emerging challenges. The government’s initial economic policy documents are also silent about the need to create fiscal room for increasing growth enhancing spending. Table 4.5: General Government Primary Surplus (% o f GDP) 1999 2004 2005 2006 2007 2008 ConsolidatedBudget 3.8 9.2 9.8 9.3 7.2 7.2 Primary Revenues 20.9 24.6 25.0 25.3 23.7 23.2 Primary Expenditures 17.1 15.3 15.2 16.0 16.5 16.0 SSI -3.0 -3.8 -4.1 -3.5 -3.0 -3.2 Primary Revenues 5.6 7.1 7.1 7.4 8.9 8.7 Primary Expenditures 8.6 10.9 11.2 10.9 12.0 11.9 Local Administrations 0.3 0.1 -0.3 -0.3 -0.3 -0.3 Primary Revenues 4.0 3.7 4.0 4.0 4.0 3.9 Primary Expenditures 3.7 3.5 4.3 4.4 4.3 4.3 Revolvingfimds 0.0 0.4 0.3 0.3 0.3 0.3 Primary Revenues 1.6 2.6 2.5 2.4 2.5 2.5 Primary Expenditures 1.6 2.2 2.2 2.1 2.1 2.1 EBFs -0.6 -0.2 0.0 0.1 0.0 0.0 Primary Revenues 1.7 0.6 0.5 0.5 0.5 0.5 Primary Expenditures 2.3 0.7 0.5 0.5 0.6 0.5 UI 0.0 0.4 0.3 0.4 0.4 0.4 Primary Revenues 0.0 0.4 0.4 0.5 0.4 0.4 Primary Expenditures 0.0 0.1 0.1 0.1 0.1 0.1 Total Primary Surplus 0.5 6.2 6.1 6.2 4.5 4.4 Memo Item: SOEs -2.1 0.9 0.15 0.6 0.4 0.3 Total public Sector -1.6 7.2 6.8 6.8 4.9 4.7 Source: SPO and the WB staff calculation B. EXPENDITURE PRESSURES IN KEY SECTORS AND MEDIUM-TERM CHALLENGES 4.12 Expenditure pressures are growing in key sectors and the development challenges on the way to the EU are ampliJjting the needs. Although, the government’s economic policy documents lack any explicit consideration o f the growing expenditure pressures, those documents mention many policy changes for which budgetary implications need to be carefully analyzed. This section aims at identifying expenditure pressures in key economic sectors-education, health, pensions and rural development- and their budgetary implications. 161  B1. New and emerging education sector policies in Turkey 4.13 The Government has proposed several new policies that, i f and when implemented, will have a significant impact on the medium term expenditure program. These new policies were not accounted for in the 2006 budget document. Proposed policies examined below include: 0 Fifteen New Public Universities 0 Grants and Loans to Parents Sending their Children to Private Schools. This policy change was part o f a draft legislation submitted to the Parliament. Although, the government finally decided not to include this policy change in the Special Education Institutions law no: 5545 during the Parliamentary discussions, the fiscal impact o f this policy has been analyzed below for a illustrative purposes. 0 Extending school years to one more year In addition to these three policies, the Government i s also seeking to increase the secondary school enrollment rate from its current 61 percent to the EU target o f 85 percent. 4.14 Early in 2006, Parliament approved the opening o f 15 new public universities. This decision has increased the number o f public universities which are classified as special budget agencies from 53 to 68.lI2 Most o f these new universities were established by separating parts o f existing universities into new ones with new names. For this reason, many already have the required physical infrastructure. Nonetheless, with the enacted law, these 15 institutions will need new administrative and educational staff to run as a university. The total number o f required staff i s 20,157 academics and 6,024 administrative staff. The Ministry o f Finance plans to hire new staff gradually. 4.15 Better strategic planning and links with financing priorities is required. The high excess demand for university education i s well known and documented in Turkey. Nonetheless, the fact that the decision to establish these new universities was not taken within any explicit strategy formulated by the Higher Education Council (YOK) or the State Planning Organization demonstrates that more consistency is required with planning or financing priorities. Opening up new universities without explicit criteria for whom they should serve, where they should be located, which degrees they should grant, and how they should recruit creates an additional fiscal burden that may not be compensated with an equivalent level o f benefits to the public. 4.16 Grants and loans have been considered for parents sending their children to private schools. A draft Law submitted to the Parliament in March 2006 envisaged providing YTL 1,000 as a lump sum amount to subsidize the cost o f tuition fees for parents sending their children to private schools. According to this draft legislation, additional subsidy would also be provided to cover half o f the interest costs o f loans that 112 The new Universities are: Ahi Evran University; Kastamonu University; Duzce University; M.Akif Ersoy University; Usak University; Rize Universitesi University; Namik Kemal University; Erzincan University; Aksaray University; Giresun University; Hitit University; Bozok University; Adiyaman University; Ordu University; Amasya University 162 parents take on to pay for private education. Part o f the costs o f electricity, water and natural gas for private schools would also be financed-by subsidizing the lower tariffs for these services to make them equivalent to the tariffs paid by public schools. Pre primary, primary and secondary school education provided by the private sector were covered by the proposal. 4.17 The efficiency o f subsidies as a means o f improving enrolment and leveraging private initiative in education is uncertain. Whether these subsidies and credits are targeted to lower income families (such as voucher programs in the U.S. or Chile) or whether other criteria are to be used would make a difference for their impact. If i t is targeted to the poor, it may provide positive benefits for children from lower income families to benefit from private school education, which provide better quality education as reported above. For many parents already sending their children to private schools, the subsidy would simply reduce their o w n tuition payments. Moreover, the amount may be sufficient to motivate some parents to pull their children from public to private schools. Subsidies may also serve as an incentive for business people to open new private schools. However, they would not motivate entrepreneurs to invest in small towns where the average household income i s already low. The increased demand from small towns and villages, or from l o w income families in big cities, i s likely to be partly met by foundations that are able to open and run private schools at lower expenses than private schools. However, public service provision in most backward provinces may need to further increase in the future to compensate for possible lack o f private sector initiative. 4.18 MONE has announced that the school year o f the secondary education will be extended from three to four years, consistent with the Education Council's decision. Unlike the 1997 basic education reform law, which extended compulsory primary schooling from 5 to 8 years, this policy will be implemented by means o f secondary regulation rather than legislation. This new regulation requires that students who enrolled in general or vocational secondary school in the 2005-2006 school year may earn a diploma after four years o f education. B2.The Medium Term Education Sector Fiscal Framework for 2007-2009 4.19 The medium term expenditure framework 2004-2009 is reviewed both with and without these new policies. Expenditures on education are classified by separating education-only related expenditures, from administrative and headquarter costs which are not directly related to education functions (see B o x 4.3 for data sources and methodology). In the light o f this institutional structure and main assumptions, the medium-term expenditure framework mentioned above, covering the period 2006-2008, shows that Turkey plans to spend approximately 4.53 percent o f GDP by the education institutions o f which 4.09 percent goes to directly to education functions, with the remainder going to other functions, such as general administration, health etc.' l3Since This amount includes the health expenditure o f c i v i l servant which are recorded as an education expenditures by the MONE. 163 there is n o announced expenditure framework o f education institutions, the year 2009 projection i s calculated simply by applying inflation and growth adjustment factors. Box 4.3: Data sources for the medium-term education expenditure framework The data sources for this study were taken directly f r o m the final accounts o f MONE, the M e d i u m Term Fiscal Plan o f the 2006 budget, the 2006 budget accounts and related background tables prepared by the Ministry o f Finance (MOF), State Planning Organization (SPO) documents about extra budgetary funds (EBFs) and revolving h n d s (RFs). Under the heading "Central Government Budget," the analysis examined funding classified as MONE, R F s o f MONE, Universities, and R F s o f Universities and other budgetary institutions (such as education institutions o f M o H , General Directorate o f Police, Credit and Dormitories ). The data for R F s came f r o m SPO and f r o m M o F for 2004-2006. F o r the following years, the analytic tables included interpolated data to adjust for inflation and growth. The Treasury subsidies t o Credit and Dormitories were kept under MONE expenditures o f tertiary education. The education-related expenditures o f the Social Solidarity and Support Fund are classified as "not classified elsewhere" as these expenditures are not directly related t o education but more t o do with scholarships and p r o poor support, as mentioned above. Any budgetary institutions other than MONE and universities were kept under this item such as education and training related expenditures o f Police Force (such as police academies), and the Ministry o f Health (including training schools for nurses) . SSK, Bagkur, Emekli Sandigi and the Employment Agency come under this heading and data suggest very minimal education (probably mostly training) expenditures o f such institutions. 4.20 The medium term expenditure framework o f the government, updated in December 2005, does not include the new policies discussed above. The present analysis i s therefore constructed by first preparing the total education sector medium term expenditure framework without the policies included. Then, the estimated total cost o f new policies i s projected and reflected in the expenditure framework. In the absence o f new policies, the baseline projections presented below suggest that the expenditure level o f the education sector in the coming years would have been slightly reduced, as well as the total expenditure level o f central government institutions. Thus, the recent trends in decreasing public sector expenditures o n education would continue in the next three year^."^ The likely impact o f the new policies has been projected based o n the assumptions described in B o x 4.4. ~ Box 4.4: Projecting the budgetary impact of new education policies and proposals Opening new 15 Public Universities b > By using the approved and released academic and administrative staff cadres, the assumption i s that such cadres f r o m 2006 through 2009 will be employed gradually. 'I4The difference between total expenditure and total education expenditure at the public education sector are due t o (a) headquarter administrative expenses o f M o N E and Universities, (b) health expenditures o f the educational institutions, and (c) environmental expenditures (especially at the universities). 164 P > Current education system data for average salaries o f academicians and administrative staff are used to make forward estimates by adjusting for inflation. P > N o assumption i s made with respect to cost o f additional construction as most o f the universities w i l l likely be using the buildings o f the institutions that were part o f the existing universities. P > Te average cost o f new office equipment and stationary w i l l be 10 percent o f the total salaries for 2006 and 5 percent for the succeeding years. Supporting private education (proposal not implemented so far) > > By using the existing numbers o f private school students by the levels o f education, i t was possible to increment such numbers under different scenarios. T h e total cost o f state aid o f YTL 1000 per student was estimated under such scenario^."^ By using different assumptions about the ratios that the parents use loans for school fees, the average school fees, and the current interest rates offered by banks for education finance, i t was possible to estimate the total cost o f the interest subsidy. P > Savings o f operational expenditures have also been estimated due to a reduction i n the number o f students in public schools who would transfer to the private schools. Extending the duration of secondary schoolingfiom three to four years > > Both capital and current expenditures including personnel have been estimated. Costs o f constructing new classrooms needed to accommodate students for the additional school year are based on the assumption that 50 percent o f the classroom construction costs will be incurred i n 2008 and 50 n 2009. The cost reflects the fact that the new policy w i l l not take effect until 2009, when percent i students entering ninth grade in 2005 w i l l be required to spend an additional year (for the 2009-2010 school year) to earn a diploma. Besides the construction costs, we have the costs o f added additional teachers and operational needs. These calculations have been done separately for both general and vocational schools. For operational and personnel cost estimations, the assumption i s that that they would be effective in 2009. Expanding the secondary education enrollment rate towards EU convergence P Finally, it was assumed that Turkey plans to increase enrolment rate o f secondary school from 61 percent in 2005-2006 to 70 percent in 2008-2009 related with the EU targets o f 85 percent secondary completion rate by 2015. '15 Assumptions: no increase in number o f students assumed for 2006; 20 % incerase i s assumed with respect to 2006 in number o f students for 2007; 40 % incerase i s assumed with respect to 2006 in number o f students for 2008; 50 % incerase i s assumed with respect to 2006 in number o f students for 2009. 165 Figure 4.2 (a - d). Total Education Expenditure Framework Without and With N e w Education Policies and proposals(2004-2009) I 16a.3 -- m m , ,~,. . . . . . . .. .. .. .. . .. . i -~amw -c*hmu~ 4.2 1 All of the new policies and proposals will increase the financing needs o f the education sector. Based on these estimations, the gap between the currently programmed financial envelope and the envelope based on the new policies implemented in the timeframe discussed above is as follows: from an incremental 0.06 percent o f GDP in 2006, the additional expenditures required to continue the implementation o f the new policies in subsequent years would reach 0.44 percent o f GDP in 2009. The economic breakdown o f additional expenditures i s shown in Table 4.6. Table 4.6. Breakdown o f Additional Expenditures Needed for Financing New Policies Personnel Operational Capital Transfer Total In percent of GNP 2007 0.02 0.00 0.08 0.1 1 2008 0.04 0.00 0.21 0.09 0.34 2009 0.13 0.01 0.21 0.09 0.44 Source: W o r l d Bank Staff calculations 4.22 More than 85 percent o f the additional fiscal pressure will be incurred by MONE’s budget (Table 4.7). Each year, MONE would need to find additional sources o f financing. For some years, namely 2008 and 2009, this incremental amount will exceed 166 10 percent o f its total resources. For universities, the additional amount o f financing will be around 5 percent. More than 60 percent o f the additional cost i s for capital expenditures by M O M , mainly for new construction and equipment in 2008 and 2009. After capital expenditures, the biggest demand will be to cover the costs o f transfer expenditures and personnel expenditures. In the case o f higher education, the additional expenditure pressure will be for personnel. Source: World Bank Staff calculations 4.23 The above projections indicate that MONE and the higher education sector will need to substantially increase institutional efficiency, both on the side o f allocation and execution. The tight fiscal framework indicates that Turkey has a limited room to increase the expenditure level in real terms, especially due to the need to maintain fiscal discipline to hedge against surrounding risks and reduce the debt ratio. The only way for the sector to continue to expand i s to improve the efficiency o f public spending within the education sphere. Main policy directions along these lines were summarized in the World Bank’s 2005 Education Sector Study and in chapter 3. B3.Projecting social protection expenditures (0 The deficit of the pension system 4.24 Social security system deficit is set to decline. With the enactment o f the new social security laws (see chapter 3, section C), revenues o f the system are projected to increase by 0.3 percent o f GDP until 2009. The unification o f the pension institutions and data bases as well as introduction o f the UHI will contribute to better revenue collection. Furthermore, the reforms will also bring about a reduction o f 0.2 percent o f GDP in expenditures. A s a result, the social security system deficit i s estimated to decrease by half a percentage point until 2009. Table 4.8: Social Security System Deficit (Percent o f GDP) 2004 2005 2006 2007 2008 2009 Total Revenue 3.5 3.7 3.8 4.0 4.0 4.0 Total Expenditure 7.2 7.2 7.3 7.2 7.1 7.0 Balance -3.6 -3.5 -3.5 -3.2 -3.1 -2.9 Source: SSI and World Bank staff estimates 167 (ii) Health care expenditures-a baseline simulation 4.25 Health insurance protection has effectively been extended to the entire population even before the formal introduction o f Universal Health Insurance in January 2007. Health care costs o f Green Card holders were expected to increase following the introduction o f UHI, but a bulk o f this increase has already taken place even before the introduction o f UHI with the introduction o f a large number o f Green Cards effectively extending health insurance protection to almost all the poor. Bulk o f the increase in health expenditures in the future i s likely to come when the presently uninsured non-poor take up UHI upon i t s introduction and as utilization o f health services in Turkey increases to levels similar to other European countries. The projections assume no co-payments for outpatient visits and drugs and extrapolate the current rate o f increase o f outpatient expenses. Under these assumptions, and taking into account the administrative measures already introduced in June 2006 to partly contain the rise in health expenditures, it i s expected that utilization o f health services will increase at a rate o f 10 percent till 2008 and at a rate o f 7.5 percent in subsequent years. 4.26 The authorities worked on a series o f cost saving measures. The measures became effective on July 1,2006, which are collectively expected to result in a net saving o f Y T L 1.4 billion in the second half o f 2006, include: (i) Global budget protocol, underpinned by payment-per-case system for M O H hospitals effectively reducing the average price per hospital-based outpatient visit by 20 percent, which i s expected to yield a net saving o f YTL 600 million by end o f 2006. (ii) Introduction o f per-case payment for private hospitals at rates equal to MOH hospitals, which i s expected to yield a net saving o f YTL 400 million by end o f 2006. (iii)Introduction o f per-case payment system for university hospitals, with rates set at 50% above the new M O H tariff for outpatient care and equal to the MOH tariff for inpatient services, which i s expected to yield a net saving o f Y T L 300 million. (iv)Reduction in drugs from the positive l i s t that will be reimbursed by public funds, which i s expected to yield a net saving o f Y T L 100 million. 4.27 Taking into account the effect o f these measures, health expenditures in 2006 are projected to be Y T L 29,711 million. Future expenditure projections are based o n the following assumptions: (i) most o f the increase in costs due to the poor joining UHI has already taken place; (ii) inflation rate in the health sector is 11 percent year-on-year for 2006, 10 percent for 2007, 9 percent for 2008, 8 percent for 2009 and 7 percent 2010 and onward;'16 (iii) 10 percent o f the country's population will be covered by family medicine in 2007, increasing to 15 percent in 2008, 20 percent in 2009, 30 percent in 2010, 40 percent in 2011 and so on; (iv) hospital occupancy rates will increase to 67 percent in 2006, 70 percent in 2008 and 72 percent in 2011; and (vi) hospital based Inflation rate in the health sector in Turkey, driven by technology and pharmaceutical prices, has historically been 3 to 5 percentage points higher than CPI inflation. 168 outpatient visits will continue to increase at the current rate o f 10 percent until 2008 and 7.5 percent annually thereafter. 4.28 On these assumptions, health expenditures are projected to increase above ten percent per year over the medium term. Health expenditures o f social security institutions, civil servants and Green Card holders will increase by 19.4 percent in 2007, 18.3 percent in 2008, 15 percent in 2009.'17 Health expenditures by M O H and Universities as directed through their annual budgets are assumed to increase at a rate o f 5 percent every year. As a proportion o f GNP, health expenditures are projected to increase from 5.3 percent in 2006 to 6.0 percent in 2009, and escalate to around 6.6 percent by 201 1. As indicated in Table 4.9, total health care expenditures in nominal terms are projected to increase by 92 percent from 2006 to 201 1. A major fiscal risk not incorporated in these projections is that the presently uninsured non-poor take up UHI upon its introduction without a commensurate increase in contributions. Table 4.9: Projected Public Expenditures on Health, 2006-2011 2006 2007 2008 2009 2010 2011 SSI, Civil Servants, Green Card 23,001 27,468 32,499 37,382 42,608 48,576 MOH & Univ Budget Exp 6,710 7,046 7,398 7,768 8,156 8,564 TOTAL 29,711 34,514 39,897 45,150 50,764 57,140 Percent o f GNP 5.3 5.5 5.8 6.0 6.34 6.56 Source: W o r l d Bank Staff calculations (iii) Containing health care expenditures-illustrative scenarios 4.29 The introduction o f co-payments is expected to bring about saving in health expenditures. As mentioned in chapter 3, the challenge in the coming years will be to control the quantity without adversely affecting access, utilization and effectiveness o f health care services. The principal measure suggested to control quantity o f medical services and dissuade frivolous use is the introduction o f co-payments for outpatient visits and for drugs. Even nominal levels o f co-payments will reduce utilization o f health services and result in savings not only in the production o f health services but also in the consumption o f drugs (since almost all visits to a doctor result in a prescription). This i s perhaps the only demand-side measure that will have the effect o f rationalizing consumption. Table 4.10 presents the expected revenue generation from outpatient co- payments o f YTL 2 that the Government proposed to introduce in 2007, increasing to YTL 5 by 2012. Introduction o f co-payments would also reduce number o f outpatient visits as well as pharmaceutical consumption. The table presents the expected savings in utilization o f health services and drugs resulting from the introduction o f co-payments for outpatient services.' '' 'I7At the same time, revenues are also expected to increase, particularly because o f the presently non-poor joining UHI and making insurance contributions. Revenue projections have been not been carried out at this time. 'I8Even though price elasticity for demand health services in Turkey i s estimated t o be low, it would n o t be unreasonable t o expect the introduction o f a 2 YTL co-payment for outpatient health services t o result in a 5% reduction in utilization o f health services. 169 Table 4.10: Estimated Revenue and expenditure savings from Co-payments (million YTL) Revenue from Co-payment 473 I 497 [ 509 1 522 I 535 Reduced Expenditure on Drugs 483.5 [ 501.6 I 511.0 I 520..6 I 530.3 I Revenue + Savings I 1,206.0 I 1,260.3 1 1,288.7 I 1,317.7 I 1,347.4 I Source: World Bank Staff calculations 4.30 Even more ambitious measures will be needed to achieve a significant and sustainable reduction of costs. Most important o f these measures i s the introduction o f co-payments, as discussed above. Other measures proposed to be taken by MOH, including rationalizing drug use, widening the base for reference pricing o f drugs by including several other countries, re-negotiating and lowering prices and charges payable to private providers, setting a flat payment for internal medicine interventions in MOH hospitals, etc. Indeed, the key to lasting success would be the combination o f such measures with structural reforms. 4.31 Cost containment measures should go in tandem with more comprehensive structural reform. As explained in chapter 3, in the medium and longer term, however, cost containment will have to rely on building the system hierarchy by full expansion o f family medicine as gate-keeper, reforming the patient referral system to streamline hospital-based outpatient care, and finally introduce payment mechanisms for providers which will replace the current performance-based payment o f hospital physicians with one which will discourage unnecessary use o f physician and diagnostic services and introduce incentives to remain within budget. B4. Projections for Rural Expenditures 4.32 On current policies, agricultural transfers are gradually projected to increase. The Agricultural Policy Framework Law lays out the structure o f agricultural transfer programs financed from budget resources, and sets a target level o f not less one percent o f GDP. The projections below simulate a gradual increase to the 1.0 percent o f GDP target. The other assumptions are annual 5 percent real GDP growth, and constant shares o f GDP for expenditures on personnel, other recurrent expenditures, and for investment expenditures. 4.33 Reducing personnel expenditures would create more fiscal space for rural investments. Based on these assumptions total rural expenditures would grow by approximately 7 percent annually in real terms, and 32 percent in total between 2004 and 2008. Transfers would grow proportionally more than investments. The share o f personnel expenditures slightly decreases to 33 percent, without any reduction in personnel numbers. If the government tries to tackle the problem o f overstaffing through downsizing efforts, the share o f personnel costs can be further reduced in the medium 170 term. As explained in chapter 3, this could facilitate allocation o f more resources to investments as well as other recurrent expenditures. Table 4.11: Rural Expenditure Allocations and Projections (2004-2008) 1999 Constant Prices, YTL, Billion 2004 2005 2006 2007 2008 Non-Transfers Total 943,470 990,644 1,040,176 1,092,185 1,146,794 Personnel 583,340 612,507 643,133 675,289 709,054 Other Current 56,566 59,394 62,364 65,482 68,756 Investment 273,227 286,888 301,232 3 16,294 332,109 Transfer - inter-agency 30,337 31,854 33,447 35,119 36,875 Transfers Total 705,281 755,510 839,949 930,944 1,028,938 Ratio to Real GDP 0.83% 0.85% 0.90% 0.95% 1.OO% Real GDP (1999 Terms) 84,650,977 88,883,526 93,327,702 97,994,087 102,893,791 Grand Total 1,648,751 1,746,154 1,880,125 2,023,129 2,175,732 Transfer Share 43% 43% 45% 46% 47% Investment 17% 16% 16% 16% 15% Current 41% 40% 39% 38% 37% O f which, Personnel: 35% 35% 34% 33% 33% Source: Undersecretariat of Treasury, SPO and World Bank StaflCalculations 4.34 Fiscal space could also be created by a partial switch out o f transfers to investments and by use o f EU Pre-accession funds. Overall, the cumulative increment in investments outlays (above the current level) would be roughly US$300 m i l l i o n spread over 3 years (2006-2008). The challenge is that even achieving this real 22 percent (relative to 2004) increase in investment outlays (albeit o f improved quality and efficiency), would leave rural investment levels at a stable share o f GDP. The EU Instrument for Pre-Accession for Rural Development (PARD) would thus need to quickly be activated to be able to make a significant contribution to rural investments. In addition, this should be combined with a partial shift o f the transfer budget into investment-focused rural development programs, which could also be partially financed by the EU PARD. Important investments in main roads, energy, water, and sanitation could also be taken up by the private sector with appropriately designed pricing and user fee schemes. C. REQUIRED FISCAL SPACE TO MEET MEDIUM-TERM EXPENDITURE PRESSURES 4.35 An appropriate framework, integrating projected budgetary resources and conditions for debt sustainability, is required for assessing the fiscal space needed to meet medium-term challenges. The analysis requires a joint assessment o f available budgetary resources; changes in the composition o f public spending; k e y features o f the medium-term macroeconomic environment; and the underlying debt dynamics. This 171 section brings together all these considerations and provides an analysis o f the additional medium-termfiscal space needed in the light o f the available budgetary means. C.l. A sustainable medium-termexpenditure envelope 4.36 Current and expectedpolicies, as well as the government’s MTFS, are all taken into account in calculating the total budget spending envelope. Based on such information, general government budget revenues are projected over the next three years, and are used as a benchmark for estimating the total spending envelope that would be available for financing budgetary expenditures while allowing for a continuous reduction o f the public debt ratio towards a desired level. In a baseline scenario, the total envelope for the financing o f primary expenditures i s calculated by subtracting a predetermined primary surplus from the estimated total primary revenues. Such a scenario allows a reduction o f the public debt ratio towards a level that may not necessarily correspond to the desired level, as debt dynamics are also affected by risks imbedded into the medium- term macroeconomic framework. Alternative scenarios are then constructed where the primary surplus is determined in such a way as to meet a specific target for the public debt ratio over the medium term. General government data produced by SPO set the basis for such calculations. 4.37 On current policies, general government revenues are projected to stay around the level of 2005 over the medium term. While general government total revenues reached 42.6 percent o f GNP in 2005, primary revenues were recorded at 40 percent o f GNP. General government primary revenues are projected to remain at around this same level over the next three years. Tax revenues are projected to decline slightly in 2006 compared to the level in 2005 but stabilize at this level thereafter. Non-tax revenues are also expected to decline by 0.5 percentage points. Factor incomes are estimated to decline until 2007 and stay constant thereafter. Finally social funds are projected to increase mainly due to expected rise in premium collection under the new social security and UHI systems. Table 4.12: General Government Revenue Proiections (Yo of GNP) 2004 2005 2006 2007 2008 2009 Tax Revenues 23.7 24.9 24.4 24.5 24.5 24.5 Non-Tax Revenues 3.1 3.1 2.9 2.5 2.4 2.4 Factor Incomes 6.9 7.0 6.3 5.4 5.3 5.3 Interest revenues 1.6 1.9 0.9 0.9 0.8 0.8 Social Funds 7.0 6.9 7.1 8.2 8.4 8.4 Total Revenues 40.7 41.8 40.6 40.7 40.6 40.6 -Privatization 0.4 0.8 1.6 0.7 0.6 0.6 Total Revenues 41.1 42.6 42.3 41.4 41.2 41.2 Primary Revenues 39.1 40.0 39.7 39.8 39.8 39.8 Source: Ministry o f Finance and World Bank Staff calculations 172 4.38 I n a baseline scenario, the general government primary surplus is assumed to stay at the IMFprogram target level until 2009. The high level o f public debt to GNP, the s t i l l risky composition o f the debt stock and its short maturity still require large primary surpluses in order to secure a declining debt to GNP ratio needed to mitigate existing vulnerabilities. Under the current policy framework, a primary surplus consistent with IMF program targets (6.5 percent o f GNP for the whole public sector) would deliver a falling debt to GNP ratio. Taking into account the most recent developments and under certain plausible macroeconomic assumptions, this level o f primary surplus would deliver a debt to GNP ratio o f 55.3 percent by 2009 (Table 4.13). Table 4.13: Public Debt Forecasts Actual Forecast (in % of GNP) 2005 2006 2007 2008 2009 Gross Debt Stock 71.6 68.9 64.0 59.9 55.3 Macroeconomic Indicators Nominal Interest Rate (%) 16.2 17.6 17.2 15.0 13.7 CPI Inflation (%, Dec/Dec) 7.7 9.7 6.5 5.0 4.0 Depreciation (Dec/Dec) 0.1 20.0 1.9 1.o 1.o GNP growth rate (%) 7.6 4.5 4.0 5.0 5.0 Primary Balance (% o f GNP) 6.2 6.5 6.5 6.5 6.5 Source: World Bank Staff calculations 4.39 Fundamentals became more dominant driving forces behind the decline in debt ratios. Although high inflation and real appreciation o f YTL were the driving forces behind the sharp fall in debt to GNP ratios during the initial recovery period after the crisis in 2001, improved fundamentals, basically growth and the fiscal stance, became dominant factors in 2004 and 2005. Fiscal stance and high growth will gain further importance in reducing the debt to GNP ratios over the next three years since the effect o f inflation and exchange rate movements i s expected to be relatively smaller. Therefore, a primary surplus consistent with the targets under the IMF program would be needed for securing a significant decline in the public debt ratio. 4.40 Debt dynamics are also analyzed under alternative primary surplus levels. However, the amount o f decline in debt to GNP ratio that can be achieved until 2009 is highly sensitive to the level o f primary surplus for the reason explained above. For example, all else equal, a one percentage point decline in the primary surplus in proportion to GNP would cause the debt to GNP ratio to be 4.5 percentage points higher in 2009 compared to the base case. Moreover, Turkey has to opt for a faster reduction in its debt ratio in order to increase i t s resilience to shocks and preserve high growth rates for a sustained period o f time. In this context, a higher primary surplus would serve more to the purpose. 173 4.41 Turkey has to target a significant decline in its public debt to GNP ratio. Although the recent improvements in debt to GNP ratio, its composition and maturity are strong achievements, Turkey’s debt to GNP ratio i s relatively high and its maturity i s short. Turkey needs to aim at much lower debt to GNP ratio in order to ensure i t s sustainability and reduce vulnerabilities in i t s macroeconomic environment. The sensitivity o f domestic markets to the recent turmoil in international markets justifies the need for a lower debt ratio in order to ensure lasting macroeconomic stability. Although domestic market volatility was triggered by international events, i t was magnified by domestic conditions and therefore Turkey was the most heavily hit country among emerging economies as demonstrated by the slide in the stock market and the depreciation o f the currency. 4.42 A base line scenario is constructed for analyzing medium-term expenditure pressures based on the continuation o f the current fiscal policy stance. A budget envelope that would be available for financing primary expenditures over the next three years i s calculated with the assumption that the primary surplus would be kept at 6.5 percent o f GNP for the whole public sector. Given that the contribution o f the SOEs to the primary surplus is declining as the most profitable SOEs are being privatized, the assumption o f constant primary surplus ratio to GNP would mean that the general government has to compensate for the falling fiscal contribution o f SOEs. W h i l e calculating the primary surplus for the general government, the contribution o f SOEs i s deducted from 6.5 percent primary surplus and then the remaining primary surplus which needs to be generated by the general government i s converted into SPO definition (see Box 1 o f Chapter 1 for the conversion methodology). Within this context, the total amount o f resources that would be available for financing primary expenditures i s estimated to be around 33 percent o f GNP until 2009. Table 4.14: Total Spending Envelope for Primary Expenditures. Actual Forecast (Yo of GNP) 2004 2005 2006 2007 2008 2009 Total spending envelope for 32.9 33.8 33.4 33.4 33.3 33.2 primary expenditures Source: W o r l d Bank Staff calculations C.2. An assessment of required fiscal space 4.43 Additional spending needs in key areas are substantial. Under the baseline scenario, i t i s expected that the current policies-including the recently introduced social security and health reforms-would continue. Additionally, i t i s assumed that the government starts implementing the announced policy changes in the area o f education (see Table 4.7 in section B 1 o f this chapter). Moreover, spending in key areas would increase to alleviate needs resulting from either depressed spending (such as investment in infrastructure) or development challenges (such as need to increase environmental 174 spending for EU alignment). Considering first the investment needs in infrastructure, as it i s documented in chapters 1 and 2, fiscal adjustment following the 2001 crisis brought a significant cut in public investment, especially in infrastructure. To bolster long-term growth and achieve income convergence with the EU countries, Turkey needs to increase its spending in infrastructure investment. Under a conservative assumption, i t i s assumed that general government investment in infrastructure would increase by 0.5 percent o f GNP during 2007-2009. Similarly i t is assumed that environmental spending would increase by 0.1 percentage points o f GNP in 2007 and 0.2 percentage points o f GNP in 2008-2009. These are rather small numbers to facilitate any alignment with the EU Acquis but a gradual approach i s thought to be more appropriate and realistic. Table 4.15: Expenditure Pressures in Key Areas Actual Additional Spending Need (Yo of GNP) 2004 2005 2006 2007 2008 2009 Economic Affairs 4.7 4.4 -infrastructure investment 0.0 0.5 0.5 0.5 Environmental Protection 0.2 0.2 0.0 0.1 0.2 0.2 Education 4.5 4.6 0.1 0.1 0.3 0.4 Total 0.1 0.7 1.0 1.1 Source: World Bank Staff calculations 4.44 Expenditure pressures are estimated to be significant during 2006-2009 even under conservative assumptions. Based on Table 4.15, general government total expenditures as well as expenditure pressures are projected. Table 4.16 presents these projections in a full-fledged functional classification. The calculations show that expenditure pressures are significant at around 1 percent o f GNP in 2006, rising to 2 percent by 2009. 4.45 I n a second scenario, with a stronger reform initiative in education, the expenditure pressures can be even larger. In this scenario, it is assumed that the government would target more ambitious outcomes in education starting from 2007 onwards. Such targets would include, for example, reducing student per classroom to 30. The additional cost o f such policies would be around 0.3 percentage points o f GNP, which would raise the estimated expenditure pressures in Table 4.16 to 1.7, 2.2 and 2.5 percentage points o f GNP in 2007.2008 and 2009 respectively. 175 Table 4.16: Medium term Expenditure Pressures, Baseline Scenario Actual Forecast (Yo o f GNP) 2004 2005 2006 2007 2008 2009 General Public Services 16.6 12.9 11.7 11.0 9.9 9.0 -Interest expenditures 13.5 9.6 8.5 7.7 6.7 5.7 Defense 2.3 2.2 2.2 2.2 2.2 2.2 Public Order and Safety 2.0 2.1 2.1 2.1 2.1 2.1 Economic Affairs 4.6 4.4 4.4 4.9 4.9 4.9 Environmental Protection 0.2 0.3 0.3 0.3 0.4 0.4 Housing and Community Amenities 0.7 0.8 0.8 0.8 0.8 0.8 Health * 5.9 6.2 6.3 6.5 6.9 7.1 Recreation culture and religion 0.6 0.7 0.7 0.7 0.7 0.7 Education 4.5 4.4 4.8 4.7 4.7 4.8 Social Protection 8.8 9.3 9.2 9.1 9.0 8.9 -Social Security 7.2 7.2 7.3 7.2 7.1 7.0 A. Total Expenditures 46.3 43.5 42.5 42.3 41.6 40.9 A.l Primary Expenditures 32.9 33.8 34.0 34.6 34.9 35.2 B. Available Spending Envelope 32.9 33.8 33.4 33.4 33.3 33.2 C. Expenditure Pressures (Al-B) / Needed Fiscal Space 0.0 0.0 0.6 1.1 1.7 2.0 Source: SPO, World Bank Staff calculations * Figuresfor the health expenditures o f the general government differ by about 1 percentage points o f GNP from the ones presented in Table 3.16 because o f some methodological reasons. 4.46 I n a third scenario, the requiredfiscal effort for reducing public debt to GNP ratio to the levels projected before the recent market turmoil is estimated. The public debt to GNP ratio was estimated to be around 51 percent by 2009 before the recent market turmoil. Although the volatility in markets was triggered by international events-including interest rate rises by U S Fed and ECB and a “flight out o f risk” due to perceived mounting inflation pressures worldwide-domestic developments that occurred prior to this episode o f global volatility have magnified the incidence into local markets. The currency depreciated by 20 percent, intensifying inflationary pressures, while the Central Bank o f Turkey had to increase i t s policy rates by 4 percentage points within less than a month in response to these developments. In the face o f such abrupt changes in interest rates and the value o f the exchange rate that have a strong repercussion on the fiscal deficit and debt dynamics, the high debt to GNP ratio and i t s structure are s t i l l creating risks for macro stability. Therefore Turkey has to aim at faster reductions in its debt to GNP ratio in the coming years. 4.47 A policy of reducing the public debt ratio to 51 percent by 2009 would put more pressures on budgetary resources. A s discussed before, primary budget surpluses and economic growth are now the major driving forces behind the fall in debt ratios. Therefore, hrther reductions in the total available spending envelope would be required 176 in order to reduce the debt to GNP ratio, under current conditions, to the levels projected before the recent turmoil in markets and the resulting increases in interest rates and the depreciation o f the YTL. The calculations show that the total spending envelope would need to be reduced by an additional 1 percentage point o f GDP (Table 4.17). Table 4.17: Primary Surplus Needed for Achieving Pre-turmoil Debt Ratios Actual Forecast (in YOof GNP) 2005 2006 2007 2008 2009 Gross Debt Stock 71.6 67.9 62.1 56.8 51.1 Macroeconomic Indicators Nominal Interest Rate (%) 16.2 17.6 17.2 15.0 13.7 CPI Inflation (%, DecDec) 7.7 9.7 6.5 5.0 4.0 Depreciation (DecDec) 0.1 20.0 1.9 1.o 1.o GNP growth rate (%) 7.6 4.5 4.0 5.0 5.0 Primary Balance (YO o f GNP) 6.2 7.5 7.5 7.5 7.5 Source: W o r l d Bank Staff calculations 4.48 Expenditure pressures can reach as much as 3 percent o f GNP i f the debt to GNP rate is targeted as 51 percent in 2009. Under the assumption o f 7.5 percent primary surplus, the estimated expenditure pressures would reach 3 percent o f GNP if no additional spending i s made in education for reducing student per classroom to 30. The expenditure pressures would reach as much as 3.3 percent o f GNP if the government targets reducing student per classroom to 30 at the same time. Table 4.18: Required fiscal space to meet medium-term expenditure pressures Forecast (Yo o f GNP) 2006 2007 2008 2009 Expenditure Pressures with higher primary surplus to further reduce the debt ratio and without additional spending in education for increasing preschool -1.6 -2.1 -2.7 -3.0 enrolment ratio and reducing student per classroom to 30. As above with additional spending in education for increasing preschool enrolment ratio and reducing -1.6 -2.4 -3.0 -3.3 student per classroom to 30. Source: W o r l d Bank Staff calculations 4.49 Creating the required fiscal space to meet medium-term expenditure pressures would call for combined initiatives on several fronts. The needed fiscal space over the next three years can be created only by structural reforms that affect the composition and efficiency o f public expenditures on the one hand and improve revenue performance on the other. By far the strongest efforts should be put on the former, because the tax burden i s already high and more tax revenues should only be a marginal source o f fiscal space. If anything, fiscal space should be created in due time in order to reduce the tax burden. 177 Structural reforms that improve the efficiency o f expenditures in key functional areas (chapter 3) should be underpinned by horizontal reforms aimed at improving performance across sectors (chapter 2). Changes in the composition o f expenditures should also be sought, with the aim o f redirecting resources from functional areas where expenditures are oversized to priority areas for growth and social development. At the same time, i t is important that the ambitious reform agenda o f public financial management initiated since 2001 (see chapter 5) be properly implemented, to prevent wasting o f scarce financial resources and to ensure that policy priorities are adequately reflected in the budget. 4.50 Reforms that improve expenditure efficiency and help contain costs should be prioritized. Some o f these reforms have been enacted or planed but implementation will be a k e y challenge. For example, the 2006 pension reform will be instrumental in capping the deficit o f the pension system and gradually creating fiscal space for more productive expenditures. Strict implementation o f the parametric changes and strengthening o f the contribution base will be needed to ensure that the projected fiscal savings indeed materialize over the coming years. Similarly, the implementation o f the health care reform agenda with introduction o f co-payments for outpatient hospital care and drugs will be a precondition to containing booming health care expenditures. Other efficiency- improving reforms should be phased in at a faster pace. For example, creating incentives for secondary schools to fill up classrooms to capacity levels will reduce the number o f personnel needed to support an increasing student population. Similarly, hospital consolidation will improve capacity utilization and allow more efficient operation while introduction o f prospective payments systems at the hospital level will improve incentives to contain costs. 4.5 1 Changes in expenditure composition could reflect deliberate policy trade offs and should rely on greater private sector participation in the provision o f public services. Fiscal space could be created by downsizing certain expenditure programs in functional areas where expenditures appear oversized in international comparison and in accordance with government policy priorities over time. Leveraging private sector initiative, through appropriately designed Public-Private Partnerships--or other market- type mechanisms such as contracting out-the right price signals, and competitive markets, would be key in bridging gaps in the provision o f infrastructure services. This option should be exploited dynamically because there are very limited options for a significant increase in the financing o f infrastructure in the coming years. At the same time, education sector reforms should redirect private resources from l o w productivity uses-such as the widespread examination preparation classes-to more productive uses, such as for example pre-primary school education or vocational training. Horizontal reforms, with a view, for example, o f establishing a distinction between “core public service”, to be performed by c i v i l servants, and public services that could be delegated to contracted personnel, would also facilitate greater flexibility in the use o f resources and partnerships with the private sector. 4.52 On the revenue side, fiscal space should be created through tax base broadening and judicious use of privatization revenues. With the tax effort high in 178 international comparison, tax rate increases should be ruled out as they could further discourage activity in the formal sector and impede growth. Tax initiatives should aim at establishing tax neutrality by avoiding tax exemptions and special regimes, as has been the case recently with some initiatives in the VAT and the taxation o f interest income. As previously analyzed, although more refined estimates would be necessary, tax expenditures-such as tax exemptions o f allowances and side benefits to c i v i l servants, or the deductibility o f social security contributions from taxable i n c o m e - c a r r y a significant cost in terms o f foregone tax revenues. Rationalization o f tax expenditures could thus go a long way towards broadening the tax bases and creating needed fiscal space. Finally, it would be important that the expected privatization revenues be used to repay public debt. The impact o f the recent market turmoil on the debt ratio could thus be offset and the required fiscal space to bring the debt ratio back o n its previous track would be significantly reduced. 179  CHAPTER V STRENGTHENING BUDGETARY INSTITUTIONS FOR EFFECTIVE GOVERNMENT 5.1 Longstanding institutional weaknesses in budget and financial management had ofen contributed to fiscal crises in the past. Public expenditure management was hampered by budget fragmentation-in particular, widespread off-budget arrangements; lack o f a comprehensive resource framework; and divided responsibilities for formulation, execution, and oversight o f the budget. These weaknesses contributed to high degree o f non-transparency, which also impeded effective policy guidance to the budget in both aggregate fiscal management and in evaluation and management o f sector policy. Turkey’s excessive use o f quasi-fiscal policies and the poor coverage and control o f contingent commitments had exposed the government to fiscal risks. A plethora o f control and inspection agencies focused on “catching violators” rather than on fixing the internal control system. The Supreme Audit Institution (SAI) was limited by its restricted mandate. 5.2 The government has implemented a number o f priority actions to improve budget coverage, formulation, execution, accounting, and audit. A far-reaching overhaul o f the laws governing public finance management has been undertaken over the past few years. Ensuring the comprehensiveness and transparency o f public expenditure was identified as a fundamental priority for reform. Moreover, the new laws go beyond the initial reform strategy in a number o f respects-inspired in part by the prospect o f EU accession and by the government’s desire to introduce greater performance orientation in budget management. However, in some instances implementation o f the reform agenda has been hesitant, and some regressive tendencies in the implementation o f the PFMC law tend to perpetuate old problems o f budgety fragmentation. I t is important that the unfinished agenda outlined below be addressed in a focused and determined way. 5.3 Although the initial public sector reform program, was initiated to improve the budget and financial management reform, the coverage o f the public sector reform has been expanded to cover the administrative aspects o f the public sector. Therefore several other laws have been or are being enacted (Table 5.1). Expansion o f the public sector reform will add additional challenges for the implementation o f the reform. Sequencing and complementarity o f these reforms needs to be carefully designed for a successhl implementation o f public sector reforms. 180  Table 5.1. Public Sector Reform Laws and Draft Laws Law Status The Constitution The Constitution was amended in October 2005 to adopt the terminology used in the PFMC law and to expand the scope o f TCA’s mandate to cover all central government institutions, social security institutions, and local administrations. Special Provincial Administration Law Reenacted in March 2005; however, several articles have been referred to the Constitutional Court by the President. Municipalities Law Enacted in December 2004. The Constitutional Court cancelled the law, and i t was reenacted in July 2005. However, one article was referred to the Constitutional Court by the President. Secondary legislation i s being prepared. Metropolitan Municipalities Law n July 2004. Secondary legislation i s being prepared. Passed i SPASand Municipalities’ Revenues A draft i s submitted to Parliament in September 2006. Law Civil Service Law A draft i s prepared but not submitted to Parliament as o f October 2006. Regional Development Agency L a w Enacted in January 2006. Turkish Court o f Accounts Law A draft law was submitted to the Parliament in February 2005. Source: World Bank Staff 5.4 Both international experience and Turkey’s own experience indicate that implementation o f a major budget and public financial management reform by itself needs strong coordination and monitoring. Because o f their impact o n the whole o f government, public sector reforms tend to be cross-cutting, and they take time to implement. They require strong high-level leadership to articulate a clear vision o f the objectives. Pragmatic mid-stream adjustments must adapt to changing circumstances, and coordination must be effective across government. Weak interagency coordination would pose the inherent risk o f provisions being enacted that are inconsistent with or contradict the previous reforms. I t i s crucial that Turkey maintain the momentum and credibility o f these major reforms; so the need for strong, effective coordination to reconcile conflicts cannot be overstated. 5.5 This chapter assesses the status o f public financial management reform implementation and identifies further actions that would need to be initiated in order to complete the Government’s reform agenda. The first section provides a short overview o f the status o f public financial management reforms further to the presentation o f the 2006 budget. The following sections break down the challenge to public financial management in Turkey along the following lines: budget formulation, accounting and 181 reporting, budget execution and cash management, revenue, internal control and audit, external audits, legislative oversight, and public procurement. A. STATUS OF PUBLIC EXPENDITURE MANAGEMENT REFORMS 5.6 Enactment o f the PFMC Law in 2003 marked a defining moment for public financial management. The Public Financial Management and Control (PFMC) L a w (No. 5018) enacted by Parliament in December 2003 replaced the 1927 Public Accounting Law (No. 1050), providing a new legal framework for modern public expenditure management and accountability. The law articulates a modem performance- oriented public sector management. I t provides the first comprehensive framework for public finance management using a modern Government Finance Statistics (GFS) concept for the coverage o f revenue and expenditure. It clarifies to the public the nature o f ministerial and official accountability, and i t strengthens public expenditure and financial management processes in line with EU practice. The Public Financial Management and Control (PFMC) L a w was amended in December 2005 with the main changes indicated in B o x 5.1. A.1 Budget Formulation: 5.7 The 2006 budget was prepared and is being implemented according to the PFMC law. The 2006 budget and 2007-2008 indicative proposals have been prepared according to the functional, economic and institutional classifications defined under the PFMC law. With this year’s budget, the Government Finance Statistics (GFS) analytical budget classification and the accrual based accounting systems have been expanded to all general government institutions, including local governments. The implementation o f this year’s budget will also be carried out in line with the PFMC structure. This will enable the Ministry o f Finance to consolidate the budget realization for all general government institutions. Details on the implementation o f the PFMC law are provided in B o x 5.2. 5.8 Although full implementation o f the PFMC law is expected to be completed by the end o f 2007, with the 2006 budget the PFMC law has been satisfactorily implemented in key areas. These areas include: (i) improving budget formulation processes to enable a medium-term fiscal strategy to guide fiscal aggregates and enable sector budgets to be linked to policy priorities; (ii)expanding budget coverage to all central government budget entities; (iii) expanding the GFS-consistent economic and functional budget classification to the entire general government institutions; (iv) delegating more responsibility (with related accountability) to spending agencies through reforming the internal control regime and abolishing central-level ex-ante control; (v) improving accounting through implementation o f accrual-based accounting for the entire general government. (4 Improved Credibility of the Budget Preparation Process 5.9 Relative to the pre-2001 situation, there have been steady improvements in budget formulation and preparation processes, including better central guidance to the 182 line ministries. The framework for these actions was established by a High Planning Council (HPC) decision issued in June 2001 to accompany the Prime Minister’s Budget Call. The HPC decision provided a macro-fiscal framework for the preparation o f the 2002 budget. I t established indicative ceilings for recurrent and investment budgets for ministries and line agencies, based on an indexation formula applied to the actual budget allocations that each ministry and line agency received in 2001. As a further step, the Prime Minister’s Budget Call for the 2003, 2004 and 2005 budgets included individually- specified indicative ceilings for each ministry and line agency, based on the 2002 budget preparation experience. 5.10 The policy-planning-budgeting linkage is being strengthened further through top-down budget formulation announced by the Medium Term Program ( MTP) and the Medium-Term Fiscal Stratem (MTFS) in the 2006 budget. The preparation and political endorsement o f an MTFS is expected to reinforce fiscal discipline and management, as well as providing the basis for improved strategic allocation o f budgetary resources to priorities. The MTP which was issued in the Official Gazette dated May 3 1, 2005 as a Council o f Ministers’ Decision set out a three year (2006-2008) macro-economic framework.’ l9 The MTP outlined policies and priorities in five development areas and nine main sectors.12’ In addition, the MTFS was issued in July 2, 2005 as a HPC decision and laid out medium term fiscal aggregates and institutional ceilings by economic classification consistent with the MTP. ’*’ (ii) Reforms to improve capacity at the institutional level 5.1 1 The PFMC law underlines the need for institutional capacity strengthening by requiring that the fiscal implications o f new policies be estimated and policy decisions be made consistent with the M T F S . The 2001 PER noted that the institutional processes for policy formulation and management, both in ministries and at the level o f the Council o f Ministers, were weak. Thus in addition to the strengthening o f “top-down” processes for setting medium term fiscal strategy, reforms would also have to strengthen “bottom-up” processes to improve policy formulation planning and performance at agency level. I t i s important to ensure that the capacity o f ministries and departments to propose policies in line with strategic objectives is strengthened. 119 Macro economic targets for 2006-2008 set out in the M T P were consistent with April 2005 S B A with the IMF. Another implicit underlying factor for determining the macro economic framework for has been convergence with the Maastricht criteria o f the EU. The 2005 Pre-accession Economic Program document states that securing price stability, reducing the public sector deficit and debt to G D P ratios must be riorities if the macro-economic policies are to converge with the Maastricht criteria. ’’ Five development areas covered in the MTP are (a) improving human resources and employment generation, (b) social inclusion and the fight against poverty, (c) enhancing the competitiveness o f enterprises, (d) regional development and reduction in regional development disparities and (e) good governance. And the nine m a i n sectors included in the MTP are (a) education, (b) health, (c) environment and urban infrastructure, (d) science and technology, (e) agriculture, (f) manufacturing industry and mining. energy, (h) transportation and communication, (ix) tourism. “‘The M T F S provided aggregate fiscal balances for the entire central government institutions. Whereas, institutional ceilings were provided only for general and special budget institutions excluding the regulatory and supervisory institutions in line with the P F M C law. 183 5.12 The PFMC law requires ministries to undertake a strategic planning exercise and performance budgeting in order to clarifj, policy goals as well as objectives and to prepare budget requests consistent with performance goals. The SPO has issued guidelines for strategic planning in key line ministries and departments. In July 2003, the HPC issued a decision announcing the launch o f the strategic planning initiative on a pilot basis in eight agencies.'22 The Government launched these pilot cases in 2004 to test the guidelines developed by SPO and MOF for strategic planning and performance related budgeting respectively. With support from SPO, these eight agencies are expected to finalize their strategic plans during 2006 with a view to reflecting these plans in the 2007 Budget. Based on the experiences o n these pilots, a phased program to expand strategic planning to the rest o f the government has been announced by the SPO. A three year phased approach was decided to expand implementation o f strategic plannin to the % entire central government and social security institutions through the end o f 2009. 1 5.13 The Government has initiated an accelerated implementation o f strategic planning for local administrations starting in 2006. This initiative was part o f the Government's decentralization reform efforts, which is expected to lead to an efficient allocation o f resources at the local level and improved public service delivery. The Metropolitan Municipalities Law (no: 5216, enacted in July 2004), the Municipality L a w (no: 5393 enacted in July 2005), and SPA law (no: 5302, enacted in February 2005) require all the local administrations with more than 50,000 populations to prepare their strategic plans within a year after the effectiveness o f these laws. A total o f 206 local administrations and 8 1 Special Provincial Administrations are required to complete their strategic plans by July 2006. 5.14 I n order to ensure a smooth transition to the new system, new institutional structures have been established. The MOF and SPO established new departments to prepare guidelines and the secondary legislation, undertake the iloting activities, provide training, and better respond to the newly defined structure.'2a The performance based budgeting department in the M O F and the strategic planning department in the SPO were established in 2004 and 2005, respectively. The experience o f the pilots demonstrates both the difficult challenge posed by these agency level reforms and the need for coordination and carefbl design in the implementation o f performance-based budgeting. I t i s particularly difficult to undertake major performance reforms in a context where the financial control and management reforms are being implemented with resulting environmental changes and uncertainties. Thus the main challenge for the government i s to draw lessons from the pilot cases to design further stages o f the reform, including consideration o f the appropriate pre-conditions for such reform. '22 Ministry o f Agriculture and Rural Affairs; Turkish Statistical Institute (TURKSTAT); General Directorate o f Health for Borders and Coasts; General Directorate o f Highways; Hacettepe University; Denizli SPA; Iller Bank; Kayseri Metropolitan Municipality. 123 The time table for the regulatory and supervisory institutions which are part o f the central government has not been announced yet. 124 In addition to Pilot agencies, the SPO has adopted a new regulation o n strategic planning. The regulation requests public administrations t o prepare their strategic plans. Thus, piloting activities have become less necessary now. All the administrative u n i t s are requested t o prepare their strategic plans within a timeframe set forth in the regulation. However, these administrative u n i t s are in need o f substantial technical support. 184 Box 5.1: Amendments to the PFMC Law T h e Government amended the PFMC law in December 22, 2005 with law no. 5436. The main motivation behind this amendment was to clarify some o f the terminology used in the law, thus elimnating some of the foreseen implementationproblems. The most important features of the amendment can be explained as follows. Change in the Coverage: As a result of this amendment, there have been some changes in the list of the chart I,I1 and 111 . Table below provides a summary of these changes. Court of Jurisdictional Disputes [ ChartI I Ministry o f Justice (Chart I) High Council for Elections I ChartI I Ministry o f Justice (Chart I) Near and Middle-East Labor Training Center Chart 11 Ministry o f Labor (Chart I) Center o f Research on Ataturk Chart I1 Presidency of Ataturk Culture, language and Historv ...- ,,...__.., -. IPACLH) . (Chart In ~ Ataturk Cultural Center Chart I 1 PACLH budget (Chart 11) Turkish Lanmage Arrencv I - I I Chart I 1 - . PACLH budget (Chart 11) Turkish History Agency [ Chart11 PACLH budget (Chart U) Turkish Industry Management and Administration Institute I Chart I1 TUBITAK (Chart 11) I Ministry of Health (ChGI) ~~ Presidency of Refik Saydam Health Center r C h a n 11 InstitutionsRemoved from the Coverage of the PFMC Presidency o f EuropeanUnion Educ and Youth Programs 1 Chart I GD o f the Turkish Radio and Television Agency Chart I1 Presidency o f Ministry of Defense Carburant Supply and Chart 11 NATO POL Plant Operattons Presidency of the Mass Housing Adnunistration Chart I1 Natural Ihsaster Insurance Agency Chart 11 General Directorate o f the Nattonal Lottery Chart I 1 General Directorate of Spor-Toto Chart I1 hivate Employee’s Social Secunty Fund Chart I1 Turkish Accountine Standards Board 1 Chart II Higher Specialization and Research Hospital Chart I 1 General Directorate for Rural Services Chart I abolished Presidency o f Revenue Administrations 1 Newly established Chart 1 DG o f Social Assistance and Solidarity I Newly established Chart I Structural Changes: The financial control officer positions have been eliminated and the authority for ex-ante control has been moved to the Financial Services Units in general. As i s the case for regulatory and supervisory institutions, the Parliament and the TCA will submit their budgets directly to the Parliament, with a copy given to the Ministry of Finance. The Parliament’s external audit will be undertaken by an independent audit commission as i s the case for the TCA. Revenue surplus of the regulatory and supervisory institutions will be transferred to the general budget on a quarterly, rather than annual, basis. T h e revolving funds, which were scheduled to be closed by the end of 2007, will now be restructured within the same timetable. 185 (iii) Comprehensiveness of the Budget 5.15 To support the Government’s ongoing efforts towards improved budgetary formulation, a number o f steps have been taken to achieve comprehensive budget framework for supporting sound fiscal management. 5.16 Better Budget Definition: Adoption o f new budgetary definitions helped Turkey converge with the international practice. A new general government budget definition has been introduced with the implementation o f the PFMC law in 2006. The general government definition covers: (i) central government budget institutions (total o f 143); (ii) social security institutions’ budgets including the unemployment insurance agency (4); and (iii)local administrations (around 3,305 including 81 SPAS). 5.17 Improved Coverage o f the Budget: Moreover, the coverage o f the central government budget approved by the Parliament has been considerably expanded. The 2006 budget approved by the Parliament covers all the general budget (chart I), special budget (chart 1 1) and regulatory and supervisory institutions (chart 111). The previous coverage o f the budget (i.e. the consolidated budget) comprised only 61 percent o f the total general government expenditures. With this structure, the Parliament i s enabled to review the total expenditures under the central government. 5.18 Improved Budget Classification: Until recently, Turkey was not implementing the functional classification o f government expenditure that is recommended by the IMF’s Manual on GFS and which i s essential for policy analysis o f expenditure. This situation has been addressed by the government’s adoption o f h c t i o n a l budget classification in line with the GFS 2001 and by parallel amendments o f the chart o f accounts for all entities forming part o f the general government. The GFS classification, including the ten sector functional classification has been applied in the 2004 Budget for consolidated budget agencies. The expansion o f the GFS to the entire general government started to be implementedwith the 2006 budget. 5.19 Reduction in off-budget activities: There has been a drastic reduction in the number and size o f off-budget institutions and activities since 2001 but the agenda remains unfinished. With the exception o f the Support Price Stabilization Fund (DFIF), all budgetary funds and all but five extrabudgetary funds (EBFs) have been eliminated. The remaining EBFs are: Promotion and the Publicity Fund, the Defense Industry Support Fund, the Privatization Fund, the Social Aid and Solidarity Incentive Fund, and the Savings Deposit Insurance Fund (SDIF). However, EBFs s t i l l remain outside the scope o f the budget discipline (see below). Rationalization and reform o f revolving funds, which represent another form o f off-budget activity, i s under way, but their closure has been delayed firther to a 2005 amendment o f the PFMC L a w (see below). 5.20 Formalization o f providing information outside the central government budget: The Parliament has been informed about the expenditures and revenues o f institutions 186 outside the central government. The revenues and expenditures o f local governments (2003-2008), social security institutions (2003-2006), revolving finds (2006-2006) and extra-budgetary finds (2003-2006) have been submitted to the Parliament as an attachment to the central government budget law. The Parliament, however, discussed and approved the central government budget for 2006 only. 5.21 First effort towards reporting tax expenditures: The MOF has included a l i s t o f tax expenditures and their estimated amount for 2006-2008 as an attachment to the budget for the first time. The coverage o f this study i s based on the four main tax legislations: personnel income tax (PIT), corporate income tax (CIT), value added tax (VAT) and special consumption tax (SCT). Issues in the estimation o f tax expenditures have been addressed in chapter 2. Box 5.2: Towards Full Implementation of PFMC The Public Financial Management and Control L a w (PFMC), which was enacted in 2003, defines the new principals, rules and structures for the Turkish public financial management systems including budget formulation and execution, financial and internal control systems, and internal and external audit structures. The original effectiveness date o f the P F M C was postponed for one year to 2006. The main reasoning behind this postponement was that since the 2005 budget was prepared according to Public Accounting L a w no. 1050, its implementation cannot be carried out with a n e w structure. Although the coverage o f the 2005 budget was n o t in line with the expanded definition o f the central government budget, major improvements had been undertaken in terms o f the quality o f budget reporting. 1/ With the 2006 budget, coverage was expanded t o a l l central budget institutions, the sum o f the general budget, special budget, and regulatory and supervisory institutions. With t h i s expansion, a total o f 45 n e w institutions were included in the budget coverage. However, although the P F M C requires the revolving funds to be attached to the general and special budget institutions, these funds were not covered within the 2006 budget. The original plan for the revolving funds was t o report them under the relevant institution’s budget until their abolishment by the end o f 2007. However, with an amendment t o the P F M C enacted in December 2005, the government decided not t o abolish them but “restructure” them by 2007. Until the restructuring i s completed, the rules and principles for their budget preparation, execution, accounting and auditing will be determined by a secondary legislation issued by the Ministry o f Finance. As a transitional measure, the revenues and expenditures o f the revolving funds have been submitted t o the Parliament as an annex to the Central Government budget since 2003 and were submitted along with their projected revenues and expenditures for 2006. The Parliament was also informed about the revenues and expenditures o f the social security institutions and local administration for the 2006-2008 period. This information was provided as an annex to the budget proposal. None o f these institutions’ expenditures has been discussed in the Parliament. In the case o f the social security institutions, the data presented to the Parliament were the estimates o f the institutions themselves. As for the local governments, the data produced through consolidation o f the individual local adrmnistrations’ budgets was not used since the GFS budget coding structure was not in effect for the a l l o f the local administrations. Therefore, the projection o f the State Planning Organization was used. Financial services units for all o f the general government budget institutions were established in January o f 2006, and with the establishment o f these units the authority o f the M o F for ex-ante spending control was delegated t o the line agencies. The accounting responsibility and m a i n budget responsibilities other than the general budget institutions were transferred t o the special budget institutions. 2/ 187 The internal audit structure foreseen by the PFMC w i l l be completed before end o f 2007. The Ministry o f Finance i s planning to set up the internal audit structure o f the 15 pilot institutions together with their personnel during 2006. I t should be noted that two areas that require more time before full implementation are strategic planning and performance budgeting. The Government has been undertaking pilot activities in these areas since April o f 2004. The PFMC authorized the SPO and the M o F to determine the expansion timetable for the institutional strategic planning and performance based budgeting. As expected, the 2006 budgets o f the institutions were not based o n institutional plans and performance criteria. According to the Municipality law, strategic planning i s necessary for local administrations with populations o f more than 50,000 i n 2006 Implementation o f the external audit part o f the PFMC depends entirely o n the enactment o f the new Turkish Court o f Account (TCA) law which was submitted to the Parliament in February o f 2005. The new draft expands the audit mandate o f the T C A to all public sectors and provides a legal basis for expanding the scope o f the TCA to include the performance and financial audits along with the compliance audit. 1/ Details o f the new features o f the 2005 Budget are explained in “Turkey, Discussion Note on the 2005 Budget” prepared by Mediha Agar and Rodrigo Chaves dated January 2005. 2/ In the former structure MoF, staff was in charge of the accounting oflice and the personnel o f some of the annex budget institutions now classified as special budget institutions. As a transitional measure, MoF accounting staff was appointed to the relevant line agencies with temporary duty status (iv) Unfinished Agenda: 5.22 While the linkage between the policy-planning and budgeting has greatly improved through the medium-term budgeting approach ,in 2006, more efforts are required for further improvements. Although 2006 was the first year o f implementation, the achievements in linking the policies/priorities with the budget have been quite satisfactory. However, further improvements can be achieved through; e Ensuring that the MTP ’s sectoral classification is appropriately translated into the budget functional classification: The sectoral breakdown provided in the M T P i s not consistent with the functional classification o f the budget structure. Due to inconsistent classifications o f the M T P and the budget, i t is difficult to follow the link between the policies envisaged in the MTP and the related budget allocations. Mapping specific sectoral categories o f the MTP priorities into the functional categories o f the budget would allow the government to easily translate MTP decisions into budget allocations. e Supporting policy objectives with required measures: Policies and priorities defined in the MTP are too generic and do not provide sufficient guidance for line agencies to translate these policies into their budget proposals. A detailed policy matrix defining priorities/policies accompanied by a set o f measures together with their cost estimates and implementation timetables could provide better guidance for the line agencies. e Making program budgeting an integral part o f the PFMC system. I n addition to mainstreaming MTP, what i s equally important i s adopting program budgeting approach to annual budgeting. New budget coding will not be enough 188  for program budgeting unless the latter i s integrated in the PFMC system. In the absence o f program budgeting, strategic planning and performance evaluation will be impeded. 0 f policy formulation documents: Currently, Rationalizing the number o there are too many documents that can be treated as the basis o f policy f ~ r m u l a t i o n . ' Although ~~ these documents have different stated purposes and perspectives, i t would be worth considering options for merging/consolidating some o f these documents for a simplified policy guidance structure. This would help eliminate potential sources o f confusion for the line agencies; reduce the workload o f the key central agencies; and diminish the risk o f inconsistencies among these documents. The Government i s aware o f these challenges and efforts to improve institutional capacity for medium-term budget planning are under way. 5.23 The implementation o f strategic planning and performance budgeting will require more efforts of the key central agencies. Several issues need to be addressed for a sound implementation o f the strategic planning and performance budgeting initiatives. Timing and sequencing i s a critical issue. Technical and human resource capacity will be another major challenge for the implementation o f strategic planning at the local level. Moreover, coming up with a provincial level strategic plan which was not foreseen by the legislation will be an additional complication for the government given that all municipalities, SPAS and metropolitan municipalities will prepare their own strategic plans without appropriate linkage and sequencing. 5.24 Revolving funds continue to operate outside the budget coverage at the risk o f distorting budget discipline. Many spending units maintain revolving finds financed by their own sources o f revenues-more than 1,000 such funds as o f December 31, 2005. The original PFMC law required the closure o f the revolving funds before the end o f 2007. However, as mentioned before, the government decided to restructure the revolving funds attached to the general and special budget institutions by the end o f 2007 with an amendment made to the PFMC in December o f 2005. More than six percent o f the central government expenditures will thus s t i l l not be covered by the overall financial management and control structure defined in the PFMC. The M O F started to issue the revenues and expenditures and balance sheets o f the revolving funds on a quarterly basis as o f June 2005 and provided information to the Parliament on the revolving funds budget as an annex o f the 2006 budget proposal. 5.25 The share o f revolving funds is quite important in health and education sectors. As can be seen in Table 5.2, the off-budget revolving fund expenditures o f the Ministry Policy formulation in 2006 was based on the following l i s t of documents prepared by the central agencies: i) 2006 annual program (SPO), ii)general economic targets and investments 2006 (SPO), iii) 2005 Pre-Accession Economic Program (SPO), iv) MTP (SPO), v) MTFS (MOF). Additionally, the ninth Development Plan, the Rural Development Plan, and the Regional Development Plan will be included in the l i s t starting next year. 189 o f Health (MOH) reflect almost 90 percent o f the total appropriation allocated to the MOH budget. Therefore, the MOH i s spending an amount almost equal to its budget with almost no appropriate financial control mechanism. Table 5.2: Importance of the Revolving Funds (million YTL) 2003 2004 2005e 2006P Higher Education Institutions and universities 1,783 2,237 2,360 2,532 Other Central Budget Institutions 3,429 5,732 7,924 8,6 15 Total 5,212 7,969 10,284 11,146 Memo Items Central Government Budget 1/ 140,109 151,702 168,549 173,807 Revolving Funds expenditures (% o f Central Government Budget) 3.72 5.25 6.10 6.41 Ministry o f Health RF expenditures (% o f MOH budget) 86.5 1/ 2003-2005 data i s for the consolidated budget. F o r the purpose o f comparability, revenue transfers to the local administrations and funds were included in the consolidated budget expenditures. The number for 2006 does not include the regulatory institutions’ budgets. Source: Ministry o f Finance and W o r l d Bank Staff calculation 5.26 Extrabudgetary funds still operate outside o f budget discipline. The PFMC L a w intends to further strengthen the control framework for the EBFs by the following measures: (i) incorporation o f extra budgetary funding into the budgets o f the related administration submitted for parliamentary approval; (ii) requiring external audit by the Turkish Court o f Accounts; and, (iii) monthly reporting by EBF in the consolidated report o f the central government. However, implementation o f these measures has been delayed. Even though the PFMC L a w required incorporation o f extra budgetary funds into the budget o f the related administration, the 2006 budget fails to do so. Instead, revenues and expenditures (without details) o f the EBFs were submitted as an annex to the budget document. Their budgets were neither discussed nor approved by the Parliament. Unlike previous years, however, transfers to EBFs from the general budget are shown as separate line items in the budget. A s it i s a legal requirement, EBFs should be subject to the internal control and internal audit regime that i s defined in the PFMC Law. Total expenditures channeled through the EBFs as a share o f the central government budget will be around 2.24 percent for the 2003-2006 period (Table 5.3). 190  Table 5.3: Extra Budgetary Funds, 2003-2006 (Million YTL) 2003 2004 2005e 2006P Privatization Fund ( Privatization Administration, Chart 11) 1,120 1,126 1,897 811 Defense Industry Support Fund (Undersecretariat o f Defense Industry, Chart 1 1) 1,251 1,233 809 1,053 Promotion Fund (Prime Ministry, Chart I ) 41 111 121 128 Social Solidarity Fund (GD o f Social Solidarity, 658 1,284 1,262 1,258 Total 3,070 3,754 4,08'9 3,250 EBFs' expenditureskentral government budget (%) 2.2 2.5 2.4 1.9 Source: SPO and MoF B. BUDGET EXECUTION, ACCOUNTING, AND REPORTING (0 Developments and Reforms Since 2001 5.27 The PFMC Law has balanced the accountabilities and authorities o f spending agencies. The PFMC Law has revoked the authority o f the Court o f Accounts and Ministry o f Finance to grant visas. By setting up Strategy Development Directorates with responsibilities for financial services and internal control in each spending agency, the law has delegated certain authorities and responsibilities to the spending agencies previously in the domain o f the Ministry o f Finance. For example, the spending agency now has the authority to authorize payments, and the accounting offices o f GDPA no longer perform regularity audit and detailed verification o f payments. Previously budget officers who worked in spending agencies were employees o f Ministry o f Finance; now they will be part o f the financial service units within each spending agency. 5.28 The law clearly defines the role o f the Treasury as provider o f cash and the Ministry o f Finance as the payer o f budget expenses. The 1994 L a w on the Structures and Duties o f the Treasury Undersecretariat and the Foreign Trade Undersecretariat (No. 4059) explicitly requires that the treasury find the cash required for state expenditures (Article 2). The PFMC Law (and the earlier Law on Public Accounting) delegated the task o f making payments to the GDPA o f the Ministry o f Finance (Article 61). 5.29 There is predictability in terms o f allocation of budget funds to line ministries. The division o f responsibility is clear between Treasury and the Ministry o f Finance. The explicit requirement that the treasury find the required cash for budget execution has provided predictability to spending agencies executing their budget programs. Other than under exceptional circumstance, rationing o f cash has thus been avoided. 5.30 An automated system has been implemented to facilitate budget management. The General Directorate o f Budget and Fiscal Control (GDBFC) has developed an automated Budget Management Information System (BYES - also called eBudget) to keep track o f budget appropriations, allocations, and utilization. The interface between 191  Say2OOOi and BYES ensures the transfer o f data on budget allocations to the Say2OOOi system as well as actual expenditures from Say2OOOi back to BYES. 5.3 1 Legislation to restructure and strengthen Turkey’s tax administration has been enacted and is under implementation. Revenue Administration has been established as a semiautonomous entity under the Ministry o f Finance. It i s structured functionally, with local tax offices directly under its control. The Ministry o f Finance retains responsibility for tax policy, allowing the new entity to focus on tax administration. Additionally, it is recommended that a large taxpayer unit be set up by the Revenue Administration, with the aim also o f assuming increasing responsibility for collection o f social security contributions. 5.32 Implementation o f an automated tax management system (VEDOP). VEDOP integrates all tax offices with the Turkish communications network and allows taxpayers to transfer their tax returns and check balances. I t includes both the assessment and the collection o f taxes. Automation has had several benefits: quicker collection o f tax returns and declaration forms; better utilization o f back-office staff; valuable taxpayer information directly transferred to a database, where it is used for macroeconomic forecasting and monitoring; and better integration with the Tax Intelligence Department in detecting unreported income. 5.33 The PMFC Law replaced the outdated 1927 Law on Public Accounting. The new law introduced important features o f a modern accounting framework, including an accrual-based accounting system and consolidated reporting requirements for the general government; and i t established an official body for setting government accounting standards. 5.34 Turkey has a single accounting system, which is maintained by the General Directorate o f Public Accounting (GDPA). Chart I institutions (comprising line ministries) do not maintain separate accounting systems. The detailed books and records for accounting are maintained in more than 1,500 GDPA offices in cities and towns throughout the country. This classic division o f function thus enables strict control o f . accounting transactions. 5.35 Periodic financial statements are prepared regularly and on time. GDPA prepares aggregated financial reports for the central government, which are published as a monthly bulletin. Since late 1998, the monthly data have also been publicly available on the GDPA website (http://www.muhasebat.gov.tr/mbulten/indexE.asp). 5.36 Turkey has adopted a “dual” system for budgeting and accounting using a cash-based approach for the budget and an accrual-based approach for accounting. Turkey has consciously switched from cash-based to an accrual basis o f accounting. Since the annual budgets are cash-based, the GDPA in the Ministry o f Finance has devised an intricate method for accounting entries that keeps track o f both the cash-based and accrual entries. Cash-based budget outturns are therefore easily distinguished from accrual-based expenditures. As a practical matter, the immediate replacement o f a cash- based budget by an accrual-based budget would demand enormous resources from both 192 the government and Parliament. These resources are not readily available. The solution- maintaining the current cash-based system for the budget while changing government accounting-is a progressive, realistic approach that i s consistent with orderly development and implementation. Switchover to accrual-based budgeting must be viewed as a longer-term objective. 5.37 Commitment accounting has been introduced. Say2OOOi can capture and set aside spending agency commitments against the budget appropriation and budget allocation, thus preventing budget commitments in excess o f appropriations. 5.38 A uniform chart o f accounts that is harmonized with budget classijication has been implemented. The GDPA has issued a new framework for the accrual-based chart o f accounts. I t i s harmonized with the economic classification o f the newly adopted GFS budget classification system. A new chart o f accounts enabled the GDPA to compile financial statements consistent with the 2001-GFS budget classification.'26 5.39 Uniform accounting standards have been introduced and a single authority to set standards has been established. Article 49 o f the PFMC law mandates that accounting be harmonized with the international standards and standards be issued by a Government Accounting Standards Board (GASB), which was established as the general government's sole standard-setting authority, and has been operating since June 2006. The GASB i s under the Ministry o f Finance, comprised o f representatives from the TCA, State Planning Organization (SPO), Ministry o f Finance, Treasury, and other agencies. 5.40 An automated online accounting system has been implemented. Using a state- of-the-art Oracle database, the Ministry o f Finance developed in-house an automated online accounting system, Say2000i.'27 Up and running since 2002, the system networks and ca tures receipts and payments from more than 1,500 national nodes as they are made.lP8 Because the database i s linked through a central server, a l l transactions are immediately available. The system can produce periodic financial statements without the typical delays o f decentralized accounting systems. The system covers all general budgetary institutions except the Office o f the President and accounting office for the State Debt within the Treasury. I t includes the majority o f s ecial budget institutions (Chart 1 1) and the regulatory supervisory agencies (Chart III).p29 The Say2000i system does not extend to social security institutions as well as local governments such as special provincial administrations and municipalities. 126 The GDPA will only be compiling the financial statements for the general government. The Department o f Statistics i s responsible for compiling GFS-compliant financial statistics. 12' The total hardware and software investment was about US$8 million. See the General Directorate o f Public Accounting (GDPA) web site, http://www.muhasebat.~ov,trlsav2000/indexE.asp for a presentation o f the technological architecture o f the system. 128 Serving more than 6,000 users, the system i s comprised o f 7,700 computers, 3,300 printers, 400 routers, 400 modems, two database servers, and seven application servers. 129 There are only 7 institutions not covered by the Suy2000i because o f technical difficulties : The Turkish Standards Institute, Turkish Patent Institute, Small and M e d i u m Sized Industry Development Organization (KOSGEB), Telecommunication Agency, Banking Regulatory and Supervision Agency (BRSA), Tobacco Agency, and Punishment and Prison Institutions. 193 (ii) The Unfinished Agenda and Remaining Challenges 5.41 The scope o f tax audit activities needs to be improved. Over the medium term, the government has committed itself to increasing the number o f auditors f i o m 5 percent o f the Revenue Administration staff to the international standard o f 20 percent or more. 5.42 Implementing a Say2000i-compatible system in the social security institutions and local administrations. The government i s in the process o f implementing Say2000i- compatible accounting systems in the social security institutions and local administrations. However, the large number o f local administrations (more than 3,200) poses an implementation challenge in the short term. 5.43 Properly phasing-in and completing the implementation o f the accrual accounting system. The government intends to move to a full accrual accounting. Since 2004, the Say2000i system has been able to capture the acquisition o f new assets. However, valuation o f existing assets (including heritage assets) and depreciation policies are still to be worked out. Full transition to accrual accounting represents a major undertaking for any government. Like any large-scale project, the shift requires careful planning and management. Governments have adopted accrual accounting over a wide range o f timefiames, often in stages. For example, the United Kingdom produced its first accrual accounts for individual departments in 1999-2000. I t consolidated central government accounts using generally accepted accounting principles in 2003-04, and i t will prepare whole-of-government accounts for 2005-06. 5.44 The implementation o f full accrual accounting requires taking on complex issues. K e y issues include valuation o f assets and inventories, depreciation rates, and gains and losses from foreign exchange. As shown by the experience o f countries that have switched to accrual accounting, implementation generally works best when done in phases. 30 Even where countries have successfully produced accrual-based financial statements, auditors have issued disclaimers o f opinion on the accrual-based financial statements for a number o f years. For example, the U.S. Government Accountability Office (GAO) has issued disclaimers on the consolidated financial statements o f the United States government since 1997, the first year in which accrual-based consolidated financial statements were prepared. 5.45 Training users of accrual-based financial statements. The Ministry o f Finance has made significant efforts in training a large body o f public accountants across the country. Accrual-based financial reporting i s not an end in itself but provides information u s e h l for multiple purposes. Accrual-based accounting statements are difficult to understand, especially for non-accountants. For accrual-based financial statements to be fully useful as management decision-making tools, it i s therefore important that users, I 3 O For more information refer to I F A C publication “Transition to the Accrual Basis o f Accounting” dated December 2003 ( htta://www.ifac.org/Store/Details.tm~1?S1D=102026702640546 ) 194 including heads o f departments and program managers, be extensively trained in how to use and interpret accrual accounting statements. 5.46 Implementing Say2OOOi interfaces with other applications. The Say2000i system needs an interface with the treasury system for accessing data on public debt, with the budget management and information system (BYES) for obtaining budget appropriation and cash allocations and transmitting actual expenditure data), and VEDOP revenue system (for obtaining accrual-based tax data). Interface to the VEDOP revenue system i s fully operational. The interface to B Y E S and the treasury system i s currently operational as well. Once all interfaces are fully operational, the need for manual intervention will be eliminated; and automatic transfer and the integrity o f data would be ensured. 5.47 Issuing and implementing secondary legislation as required under the PFMC Law. The PFMC Law envisages secondary legislation to facilitate implementation o f the law and provide detailed guidance to officials. For example, secondary legislation includes procedures for preliminary payment, transfer and set-off transactions, working principles for the GASB, and accounting and reporting standards. 5.48 Establishing GASB and issuing accounting standards. Public accountability o f government is demonstrated in part by accounting standards that require fair presentation and full disclosure. The accounting regulation issued by GDPA in November 2003 was revised in June 2005. However, GASB- not GDPA-is the authority to issue accounting standards for the general government. 5.49 Providing access to the Say2OOOi system for line ministries. Line ministries do not maintain separate accounting records although the PFMC L a w provides them with an enhanced role in financial management; so i t is particularly important that line ministries have unhindered read-only access to the Say2000i system. This will enable them to monitor the financial performance o f their respective units. Currently, GDPA policy provides line ministry staff with access upon request, but access needs to be expanded and made more widely available. 5.50 Implementing full commitment accounting. Liabilities are fully captured in the Say2OOOi system.13’ However, until recently only commitments for large multiyear capital expenditure projects were fully captured. The revised accounting regulation issued in June 2005 requires recording o f all commitments to ensure budget discipline and avoid accumulation o f arrears. 5.5 1 Establishing and implementing formal procedures for system controls audit. An online system with a database o f the magnitude o f Say2000i system requires formal 131 The terms commitments, expenditures, and payments are often confused. Simply stated, commitments are the value o f the purchase orders and contracts that bin an agency t o buy certain goods or services t o an amount specified. Expenditures are the actual amounts o f goods and services for which full payment m a y o r may not have been made. Payment i s the amount transferred f r o m the treasury bank account to the credit o f the supplier’s account. 195 policies and procedures to audit system controls on a regular basis, including mechanisms for monitoring the implementation o f audit recommendations. C. INTERNAL CONTROL AND INTERNAL AUDIT (0 Developments and Reforms since 2001 5.52 The PFMC Law has defined a modern financial management control framework for the public sector in Turkey.132 It has introduced major policy changes in internal financial controls and internal audit arrangements. The spending agencies will have increased authority and concomitant accountability for executing the budget. As shown in Figure 5.1, the law envisages a financial control framework. Each spending agency will have an authorizing officer and a financial services expert who belongs to the spending unit’s administration. Figure 5.1. Schematic Diagram o f the Internal Financial Control Framework for the General Budget Institutions as Defined by the 2003 P F M C Law I Head o f Public Administration Centralized Authorizing Financial Services Expert Internal Accounting Officer (Strategy Development Auditor Officer (Spending Unit) Directorate) (decentralized) Payment processing controls: * Strategy Development Directorate. Financial services unit responsible for ex-ante controls. * Authorizing Officer. Approves and i s held accountable for payments. * Accounting Officer. Verifies, makes payment, and maintains accounts. 5.53 There are positive aspects o f the internal control system. The most important positive feature i s the separation o f duties and responsibilities between those who incur expenditures and those who make payments. The GDPA accounting offices exercise controls with respect to payments by verifying the documentation supporting the payment and authority to make the expenditure. Under the framework newly specified by the 13* The original provision in the PFMC Law required establishment o f a new position o f financial control officer with fill responsibility for ex-ante controls within each budget spending unit. This provision was deleted by the 2005 amendment and responsibility for ex-ante controls has now been assigned to the financial services unit within in each budget spending unit, thus eliminating new positions but at the same time clearly articulating the role o f the financial services unit. 196 PFMC Law, these controls and compliance with rules and regulations would be exercised by the Strategy Development Directorate before payment. The GDPA accounting office would restrict i t s e l f to verifying the completeness o f documentation and to ensuring that no material errors exist. Another positive feature o f the internal controls i s that Say2OOOi system does not allow deletion o f transactions, so any corrections must be entered as new transactions. 5.54 The PFMC Law has introduced a modern internal audit framework. With devolution o f authority and responsibility to spending agencies, the need has increased for a modem internal audit organization that can assure the head o f the spending agency on the soundness o f the internal control system. The PFMC L a w requires each public administration to establish an internal audit unit within i t s administration. The law envisions a decentralized internal audit model in which internal auditors report to the head o f the administration. This is similar to the current model o f inspection boards in each line ministry, except that inspection boards report directly to the minister. To provide a centralized oversight mechanism over the dispersed internal audit function, the law requires a central Internal Audit Coordination Board (IACB) to be attached to the Ministry o f Finance. This board would set internal audit standards, organize training for internal auditors, and provide quality assurance for internal audit work carried out by the line ministry internal audit units. 5.55 The law clearly articulates the accountability o f ministers and other high-level civil servants. In the old accountability model, accountants and accrual officers were responsible to T C A for financial misdemeanors. This has been replaced with broad accountability for government officials. Articles 10 and 11 o f the PFMC L a w deal exclusively with ministers and heads o f public administrations. The new provisions contrast sharply with the previous law, which carefully defined the accountability o f the accountant but generally glossed over the accountability o f the higher-level officials. 5.56 The “visa” power o f the Turkish Court o f Accounts and the Ministry o f Finance has been revoked. Under the previous system, the Turkish Court o f Accounts and the Ministry o f Finance were required to issue “visas”-that is, authority to incur expenditures and provide other approvals for expenditures above predetermined thresholds, which were in the form o f controls normally exercised by an executive authority. T C A thus played an indirect role in the budget execution process. This created apparent conflict o f interests, because T C A also audited budget execution reports. The PFMC L a w has transferred this power to the Strategy Development Directorates to be established within each budget spending unit, which would exercise ex-ante control over budget expenditures. 5.57 Introduction o f accountability reports. The PFMC L a w (Article 41) requires every line ministry to prepare an annual accountability report. In addition to providing financial data, the report would include performance data and details o f activities in implementing the strategic plan and budget. These reports would be audited by the T C A and submitted to Parliament. The introduction o f accountability reports breaks with a long tradition o f providing Parliament with financial reports only. It thus strengthens the 197 accountability o f line ministries. It should, however, be noted that recent studies have questioned the assumption that performance information necessarily serves the legislature. While this measure is laudable and at least sends a signal in the right direction, i t i s too early to evaluate i t s actual effectiveness in Turkey. 5.58 The Ministry o f Finance has created three new departments to prepare secondary legislation and to implement the PFMC Law. These are the Financial Control Department, which prepares secondary legislation in the area o f financial controls; the Internal Audit Department, which develops secondary legislation in the area o f internal audit and also serves as a secretariat for the IACB; and the Financial Management Department, which i s responsible for work in the area o f performance budgeting. 5.59 The accrual accounting system would enable tracking o f both movable and immovable assets. The Say2000i accounting system has captured purchase and construction o f new assets since January 1, 2004. This will enable regular reconciliation between the fixed asset register and accounting data, and it will strengthen control over fixed assets. (io Unfinished Agenda and Remaining Challenges 5.60 Implementation o f the PFMC Law itself poses a major challenge. Beyond serving as a more transparent and accountable approach to public resource management, the PFMC Law envisions a substantial change in the public internal financial control framework; and i t devolves financial management authority and responsibility to line agencies. However, with the continuing uncertainty regarding the enactment o f the Public Administration Framework Law-which envisages abolition o f inspection boards in line agencies-the mapping o f inspectors to internal audit positions has become less certain. 5.61 Phased implementation o f the PFMC Law across the general government. The PFMC Law is applicable to the central government as well as to social security institutions and local governments, consisting o f 8 1 Special Provincial Administrations (SPAs) and more than 3,200 municipalities. Some o f the SPAs and several smaller municipalities lack adequate resources for creating new institutional structures for internal audits. The law envisages a uniform implementation schedule across the general government, irrespective o f the size or capacity o f the agencies to implement the law. This poses an enormous implementation risk: smaller municipalities may either ignore the provisions that they consider too expensive, or they may implement the provisions by focusing on form rather than substance. 5.62 Involve spending agencies in the implementation o f the PFMC Law. The implementation o f the PFMC Law had been orchestrated by the Ministry o f Finance. Spending agencies had generally been aloof distant observers. If PFMC implementation i s to proceed smoothly and if the spending agencies are now to support the reforms wholeheartedly, they must become actively involved in the process. This would require 198 further efforts in dissemination and training. The government has taken the first step by creating a high-level Change Management Committee. Working Change Management Teams have been created to guide the implementation process and to enlist active involvement among spending agencies. These are steps in the right direction. Efforts must now turn toward making them functional. 5.63 The Public Administration Framework (PAF) Law adopted by the parliament and vetoed by the president has generated uncertainty regarding the role o f inspection boards and internal audit units. The PAF L a w envisioned abolition o f inspection boards in line ministries that carried out inspections, investigations, and minimal internal audit in their respective .ministries. Since the veto, however, uncertainty has reigned with regard to respective roles o f inspection boards and internal audit units. Previously, the inspection boards focused o n identifying violators o f laws and regulations. In general, they paid little attention to internal control systems. The creation o f internal audit units to focus o n internal control systems would be a step forward. Nevertheless, every environment needs investigative capacity to uncover abuse o f power and misuse o f public resources. I t will therefore be important to retain inspection and investigation functions, albeit with fewer personnel. Also, the scenario needs to be considered o f inspection boards being abolished before internal audit units become fully operational. This could weaken the control framework and increase the risk o f misuse o f funds. A detailed transition strategy urgently needs to be devised, outlining interim measures while the new institutional structures become fully operational. 5.64 Implement a payroll module-that is, a centralized payroll database-within the Say2OOOi system. Payroll for the approximately 1.8 m i l l i o n general government staff i s not yet centralized, exposing payroll to the inherent risks o f misuse and malpractice. A personnel module is ready in the Say2OOOi system. When implemented, all employees will be tracked in the central database. Payments could be made electronically to an employee’s bank account, thereby improving the control framework. D. EXTERNAL AUDIT (0 Key Features of the External Audit Arrangements 5.65 There are three external audit agencies in the public sector in Turkey. These are: Turkish Court o f Accounts (TCA); the High Audit Board under the Prime Minister’s Department (YDK); and the State Audit Board under the Presidency (DDK). YDK was established in 1938 to audit state-owned enterprises on behalf o f the parliament. DDK has a broad mandate, covering all public sector bodies except the military and the courts. Because i t undertakes research studies, investigations, and audits only at the request o f the president, i t therefore does not function as a regular external audit agency. 5.66 TCA is independent o f the legislature and the executive. TCA i s a constitutional establishment and is entrusted with the task o f auditing o n behalf o f the Turkish Grand National Assembly. T C A carries out its work on behalf o f the parliament but remains 199 independent o f both the legislative and executive branches o f government. The parliament elects i t s president by secret ballot from two candidates, who are put forth by an ad hoc committee o f 15 members drawn from the parliament’s Plan and Budget Commission. The president o f the T C A is elected for a seven-year term and i s eligible for reelection. T C A has complete discretion in deciding what accounts and subjects to look at, including how and when to examine them. 5.67 TCA and YDK are financially independent. The T C A enjoys financial independence by preparing its own budget, which it submits to parliament without intervention. The president o f T C A i s the sole authority who can authorize expenditures from the budget, leaving no room for either the legislative or the executive bodies to interfere with its budgetary spending. Similarly, YDK enjoys financial independence as i t does not receive any budgetary support. The entire budget o f YDK i s funded out o f audit fees collected from SOEs. 5.68 As an institution, TCA also exercises judicial power. The T C A i s a collegiate body which includes an elected president and elected members as well as appointed audit staff. While the auditors carry out audit function, the colleges-namely chambers and board o f appeals-perform the judicial function. 5.69 TCA and YDK report to the parliament. T C A reports o n the general and annexed budgets. YDK reports on state-owned enterprises and on several funds. By contrast, DDK reports to the president. Under the new PFMC Law, T C A will be required to submit an opinion to the parliament on the annual accountability reports prepared by public administrations; an annual general conformity statement on the final accounts; and a report on evaluation o f financial statistics in terms o f accuracy, reliability, and conformity with set standards. In addition to these mandatory reports, T C A would submit ad hoc reports on performance audits and other special audits. An audit report for each SOE audited by YDK is submitted to the parliament. In addition, YDK submits a consolidated report on all SOEs audited during the year. (io Developments and Reforms since 2001 5.70 The Constitution was amended in 2005 explicitly mandating TCA to audit social security institutions and local administrations. Ambiguity regarding the mandate o f the T C A to audit social security institutions and local administrations has been removed by the 2005 amendments to the Constitution. 5.71 The new draft TCA Law (January 2004) along with the PFMC Law (2003) addresses several key issues. These include: 0 The PFMC Law extended TCA’s audit mandate to cover a l l administrations forming part o f general government. The earlier anomalies have been addressed by granting TCA the authority to audit the presidency. However, the December 2005 amendment to the PFMC Law has once again restricted TCA’s mandate to audit the parliament. 200  a Audit of TCA by an independent commission comprising professional audit staff i s included in the PFMC Law (Article 69). a The draft law defines audit as a planned activity to be carried out in accordance with international standards. a The draft law specifically provides for financial audit by TCA, thereby putting financial audit to the forefront along with the regularity audit. a The PFMC Law has effectively removed the role for TCA in the visa approval process. The draft law similarly contains no provisions dealing with erstwhile powers o f TCA regarding visa approval. a The draft law contains an article explicitly requiring TCA to make i t s reports public within 15 days o f their submission to Parliament. a The draft law recognizes the need to carry out interim audits at any time during the financial year. a The draft law permits the audit o f accounts and transactions of public administrations by independent private auditors. However, such audits must be carried out under supervision o f TCA auditors in compliance with i t s audit principles. (iii) Unfinished Agenda and Remaining Challenges 5.72 TCA 's draft law creates a sound basis for the development o f government audit in Turkey that is in line with modern government audit concepts and accepted international auditing standards. I t i s important that the draft law be enacted quickly. This will enable TCA to concretely begin the implementation o f reforms. The following additional measures would facilitate the transformation o f TCA into a modem supreme audit institution. 5.73 Strengthening the audit function within TCA. The audit side o f TCA is not yet fully developed. The draft law i s s t i l l insufficient in terms o f structures and responsibilities. Even after enactment o f the law, TCA will still be dominated by i t s judicial side. Its audit reports and the audit planning and coordination activities will largely remain in the hands o f those primarily employed o n non-audit issues. TCA will s t i l l lack an authoritative internal voice representing audit interests. And those aiming for senior positions will s t i l l find only one way forward: leaving audit to j o i n the chambers as members o f the court. 5.74 Developing capacity for financial statement audits. I n the future, TCA intends to carry out financial audits. This type o f audit includes technical opinion o n the reliability and accuracy o f the financial statements. TCA is therefore committed to issuing short-form audit opinion for the more than 5,000 entities currently subject to its audit. This fundamental change o f approach would require sustained training for TCA auditors to undertake financial statement audits and to issue audit opinions. Since the government has adopted accrual basis o f accounting and would be producing accrual- based financial statements in addition to cash-based budget execution reports, this development requires capacity building within TCA o n the skills to audit accrual-based financial statements. 201 5.75 Adopting a strategy for risk-based audit. Until 2006, audits carried out by the T C A were based on the 100 percent transaction examination instead o f risk assessment. T T C A has been maintaining the fiction o f 100 percent audit based on the related articles o f abolished Public Accounting Law no: 1050. With the effectiveness o f the PFMC, the Parliament plans to adopt a new T C A Law which will make the external audit consistent with the application o f internationally recognized audit standards. 5.76 Designing the role and structure o f the YDK: The Public Administration Framework L a w enacted by the parliament but vetoed by the president envisioned merger o f YDK with TCA. This welcome proposal was consolidating the mandate o f the country’s Supreme Audit Institution and reduces the number o f external audit agencies. However, with the presidential veto to the PAFL, the legal basis for the merger was abolished. Therefore a relevant article needs to be integrated to the T C A draft legislation. Moreover, issues posed by the merger must be carefblly considered, including questions o f personnel, grades, and levels o f staff. 5.77 Devising a strategy for audit o f local governments. Local governments are part o f TCA’s audit mandate. There are more than 3,200 local administrations in Turkey, about 15 percent o f which are audited by TCA. Since those selected are generally the largest, T C A estimates that about 80 percent o f local government expenditures are thus subject to audit. However, this s t i l l implies a widespread lack o f audit in this sector. Moreover, with the increased role for local administration envisioned under the decentralization agenda, the complexity and size o f activities carried out by local administrations will doubtlessly increase significantly. This prompts the need for reconsideration o f audit policies for this sector. Depending on the choice o f a strategy, TCA’s role might be redefined in several ways-for example, dividing local administrations into categories based on population size, with different strategies for local administrations o f differing sizes; creating a new local government audit service; using private sector auditors to audit local administrations. 5.78 Developing secondary legislation, manuals, and guidelines. The draft law provides a fiamework o f principles. Secondary legislation, manuals, guidelines, and extensive training will be needed to achieve the aims. In particular, T C A must change from an institution dominated by its judicial fbnctions to one that i s adapted to carrying out modem audits and reporting its audit findings to the parliament. 5.79 Adopting international audit standards. In the future, T C A intends to carry out audits in line with accepted international audit standards. Article 35 o f the draft law defines audit as a planned activity carried out in accordance with these standards. I t is important that T C A adopts (without any changes) the audit standards o f either INTOSAI . or the International Federation o f Accountants (IFAC). In addition, T C A should develop and issue ethical guideline for auditors. 202 E. LEGISLATIVE OVERSIGHT 5.80 The Plan and Budget (P&B) Commission o f the parliament oversees the budgetary process. The commission is chaired by a member o f the ruling party and i s comprised o f 40 parliamentarians. Parliamentarians are entitled to recruit advisers paid from the parliamentary budget. They provide expert advice on various issues. Advisors can come from the civil service or the private sector. Advisors from the civil service retain their previous job. 5.81 However, the P&B Commission holds little influence over the overall budget, because budgets must be prepared annually within the broader economic program agreed upon with the IMF. The parliament also exerts little influence over intersectoral allocations. Although the parliament and the P&B Commission have limited influence over overall budget size, it is not unusual for informal discussion to take place earlier in the budget process among members o f the parliament, the commission, and the government. Considerable discussion occurs annually while the budget i s being drafted, though the parliament has limited capacity to make actual changes. The three-year timeframe envisaged in the PFMC may facilitate better interactions and greater dialogue o n budget between the parliament and the government. 5.82 Neither the parliament nor the commission reviews interim budget execution reports. The final accounts to parliament include an enormous quantity o f numeric data but scant detail or discussion on program implementation, outcomes, or outputs. The reports are not conducive to discussion. The PFMC L a w o f 2003 envisioned annual accountability reports with details on costs and program achievements, thereby enabling meaninghl debate in the parliament. 5.83 Little discussion takes place on in the parliament on TCA audit reports because the reports contain little information on financial management or accountability issues. Like the annual final account, the T C A audit report i s primarily quantitative and contains little on financial management or accountability issues that emerged during the audit. To date, as described, the T C A has primarily discharged judicial functions and carried out compliance checks rather than financial audits. The new draft T C A law would change this by requiring T C A to carry out financial audits with opinions on findings, thereby strengthening parliamentary oversight over the executive branch. 5.84 Effective implementation o f the PFMC Law would facilitate better parliamentary oversight over the executive branch. The law improves the framework for parliamentary oversight by requiring three-year budget estimates, a medium-tern fiscal strategy, and annual accountability reports to the parliament. Furthermore, the draft law requires the T C A to render opinions on financial accounts, including any issues that emerged during the course o f the audit process. 203 F. PUBLIC PROCUREMENT MANAGEMENT 5.85 Between 2001 and February 2006, the public procurement system made significant progress, yet significant issues and challenges remain. Inefficient procurement procedures in the past contributed to waste and opportunities for corruption in the award o f public sector contracts. The 2001 P E R and the Country Procurement Assessment Report (June 200 1) highlighted procurement practices that encourage unrealistic bids and nontransparent decision making. This has led to many public investments remaining incomplete, with huge economic costs attached to these inefficiencies. Reforms implemented since 2001 have taken on a new dimension with the deepening dialogue related to EU Accession, specifically the “Procurement Screening Process”, Baseline Indicator System (BIS) to assess public procurement system, the Consultants Assessment Study to assess the current status and problems o f the consultants’ sector in cooperation with the World Bank, and the e-GP initiatives. (i) Developments and Reforms since 2001 5.86 A satisfactory new Public Procurement Law and secondary legislation was prepared, including procedural guidelines and instructions to provide guidance to procuring entities. Three national and one international consultant were hired to assist the Public Procurement Authority (PPA) in preparing the corresponding documentation, and all documents were finalized in 2002 before the Public Procurement Law came into effect in January 2003. All documents were published in the Official Gazette and are available through the PPA website. 5.87 Standard bidding documents (SBDs) for use by procuring entities were prepared. The consultants also assisted PPA to prepare SBDs for procurement o f goods, works, technical services, and consultant services. Here too all documents were completed before the Public Procurement L a w came into effect in January 2003, and all documents were published in the Official Gazette and are available through the PPA website. 5.88 Documents to standardize public procurement operations were printed and disseminated for all procuring entities. They are all made available through the PPA website, providing much easier and wider accessibility among public and private entities. Although reproduction and dissemination o f standard documents was foreseen within the scope o f grant-financed activities, the PPA published all documents on its website, an activity h l l y financed by its own resources rather than through grant funds. 5.89 Trainers were trained on the new Public Procurement Law. PPA’s o w n staff provided advance training on the PPL upon request o f procuring agencies, which also was financed from its o w n resources. As o f M a y 2005, 9,634 staff from 278 procuring entities were trained by PPA trainers. In addition to three days o f classroom training, a more comprehensive and systematic training module was developed using a web-based distance learning module. A consulting firm was hired to develop an interactive learning 204 program for procuring entities and the private sector. This activity was completed in January 2004, and the learning tool was made available through PPA’s website. Five thousand copies o f interactive learning CDs were reproduced and distributed to public and private entities as w e l l as subscribers o f the Public Procurement Bulletin using PPA financial resources. 5.90 A Public Procurement Law for Utilities was prepared. An international and a national consultant were hired to assist the PPA in preparing a new public procurement law for state owned economic enterprises for water, transport, energy, and the telecommunication sectors. This was done in line with the EU Utilities Directive. This activity was not explicitly defined under the scope o f this project. However, in response to difficulties and urgent needs faced by the state owned enterprises, the Bank and PPA agreed to use grant funds for technical assistance. T w o consultants were hired in January 2004. They submitted a draft law to the PPA in M a y 2004. The draft law was submitted by the PPA to the government, circulated amongst stakeholders for comments and inputs, and returned to the Ministry o f Finance for h r t h e r action. Currently the draft law i s considered within the “Screening Process o f Procurement with EU” - whether to incorporate it within the PPL or issue i t as a single bill. (ii) Unfinished Agenda and Remaining Challenges 5.91 Use o f countryprocurement systems. To some extent, all projects in Turkey rely o n the existing country procurement systems with national competitive bidding (NCB). Thresholds are being raised, enabling use o f N C B or other procurement methods for a larger portion o f project procurement. Subject to certain conditions, exceptions, and waivers, the World Bank permits the use o f country public procurement system for N C B contracts. 5.92 Procurement legislation. There have been two amendments to the PPL in June 2002 and July 2003. However, other laws have also made many other amendments to the PPL, mainly to provide exemptions to various agencies and to exclude procurement o f certain services from the scope o f the law. Later, a decision was made to hold all amendments o n the PPL until after the Procurement Screening Process with EU i s complete. It i s recommended that the government should ensure internal consistency and integrity o f the original law, considering a systematic and well thought out approach before making amendments. Since the PPA was established as a regulatory agency to prepare required legislation in public procurement, i t i s suggested that all draft laws amending the PPL be prepared and finalized in close consultation with the PPA. 5.93 e-Government Procurement Strategy. There is presently n o e-Government Procurement Strategy. However, the PPA has initiated a process o n its o w n initiative to identify actions to be taken and the necessary legislation to be prepared by the government. Currently, procurement notices are published electronically; and e-bulletin and tender documents can be obtained in C D format. Most public procuring agencies maintain their own websites. Notices are published o n the web in addition to obligatory issuance in the Public Procurement Bulletin, which i s available in hard copy and electronic format. There i s strategic integration within the context o f e-government 205 planning. PPA has agreed to establish a specific portal to serve public and private agencies such as the Social Security Agency and the Chamber o f Commerce and Industry. (Such portals will simplify the collection o f authenticating documents and certificates required in the bidding documents, for example, document authenticating membership in the chamber o f commerce.) Within the context o f the EU accession process some funds are allocated to PPA for electronic infrastructure and staff development. 5.94 Implementation o f baseline indicators to assess the public procurement system. As part o f the ongoing dialogue supported by an IDF grant for institutional strengthening o f Public Procurement Authority, i t was agreed to conduct a joint assessment o f public procurement system through the Baseline Indicators System (BIS). The B I S provides a structured approach to analyze the health o f the existing system. A rating system was finalized in February 2006, with the participation o f Public Procurement Agency (see Annex VI11 for details). The report provides a comprehensive picture o f the current public procurement system, which would become a valuable tool for identifying, tracking and prioritizing key areas for further strengthening. This study may require further refinements based on the latest OECD-DAC-World Bank rating B I S system and as necessitated by expected changes in laws and regulations. However major issues and concerns are summarized below: ~ Indicator Summary of Self World Bank’s commentslconcerns assessment by PPA Indicator No. 1- Public All the 26 sub- (i)New utility law to be enacted;(ii)PPA to maitain its procurement legislative indicators rated as oversight on all amendments and new law which affects and regulatory framework “fully achieved” public procurement;(iii) Bank remain concerned about achieves the agreed exemptions provided to certain types of entities or services standards and complies through other laws;(iv) open competitive bidding to be a with applicable preferred method in the law ;(v)all procurement above or obligations below the threshold should be open to international competitors;(vi) data on direct procurement to be kept; (vii)threshold for mandatory publication of contract award to be increased and monitored by PPA; (viii) time limits for submission of tenders to be aligned with EU Directives; (ix) data on debarment process to be monitored; (x) unfair competitive advantage to state-owned enterprises to be removed and equal treatment to all bidders considered; (xi) PPA to prepare and issue a Good practice Guidance Note on how to prepare neutral technical specification; and (xii)electronic system to be put in place to monitor current status of complaint review process Indicator No. 2 Existence “fully achieved” on all (i) PPA to ensure consistency within all implementing of Implementing the 6 sub-indicators regulation; (ii) implementing regulation to be updated Regulations and regularly; and (iii)User’s Guide manual for contracting Documentation entities to be updated regularly Indicator No.3 Out o f 6 sub-indicators Further studies required to provide data on budget MainstreamingProcedures PPA has rated 1 as formulation. into Public Financial fullyachieved, 4 a s Management substantiallyachieved and 1 as not achieved Indicator No.4 Functional Out o f 4 sub-indicators, Administrative capacity o f PPA to be strenghthened and Management Normative 3 “fully achieved’ and I responsibilities within the organization clearly defined Body substantiallyachieved 206 Indicator No.5 Existence of Out o f 4 sub-indicators (i)PPA to develop a more proactive strategy for capacity Institutional Development 2 “fully achieved”, 1 development through twinning arrangement( such as Capacity “substantially mandatory training,minimum qualification standards to achieved” and 1 “not professionalize procurement staff, on-line cerification o f staff, achieved” awareness compaign etc); and (ii) collection and monitoring o f national procurement statistics t o be enforced in line with Article 53(b) 9 Indicator N o 6: Efficient Out o f 6 sub-indicators (i) Need to improve procurement performance through Procurement Operations 1 i s fully achieved and availabilty, analysis and application o f data Capacity and Practices 5 are substantially achieved Indicator N o 7 Out o f three applicable There are certain systemic constraint in access to credit Functionality o f Public sub-indicators, 2 “fully market. This needs to be addressed. Procurement Market achieved” and 1 “substantially achieved” Indicator N o 8 Existence of Out o f 2 sub-indicators Contract administration 1 “fully achieved” and There are implementation issues. More study is required to and Dispute Resolution 1 “substantially determine coverage o f existing procedure to facilitate efficient Provisions achieved” dispute resolution process Indicator N o 9 Effective Out o f 5 sub-indicators, Implemenatation o f PFMC law to be monitored Control and Audit System 2 are “fully achieved”, 1 “substantially achieved” and 1 “not achieved” Indicator N o 10 Efficiency All 6 sub-indicators Complaints are not resolved at the initial stage( at the o f Appeal Mechanism “fully achieved” contractingentities) and generally referred to PPA, which delays the contract award for one month and increases the workload o f PPA. PPA should develop and provide guidance to the contracting entities to address the complaints from participants Indicator N o 1 1 Degree o f All 4 sub-indicators N o Issue identified by the B I S team. Access to Information “fully achieved” Indicator N o 12 Out o f 7 sub-indicators [mplementation issues remain in this area. Signature o f Ethics and Anti-cormption all “fully Achieved” :thical contract is required for all public employees. measures Disclosure o f financial assests by the public staff i s required :very 5 years. But this issue requires more study. 5.95 Status o f Dialog between EC and Turkey on Procurement. As part o f Turkey’s accession to the EU, process screening meetings between EU and Turkey are being held. Public Procurement i s one these topics and discussions o f this topic took place in November 2005. Draft screening report prepared by E C notes improvements required in the following areas (i) public contracts( 15% national preference above the threshold i s incompatible with the acquis); (ii)Lack o f sector (utilities) legislation; (iii)Concessions - lack o f horizontal legal framework; (iv)Strengthening the Administrative capacity o f the PPA; and (v) .Electronic Procurement - such as need for electronic auction o n a pilot basis 5.96 Need for an Action Plan. The action plan on CPAR 2001 recommendations summarizing responsibilities and achievements to date are given in Table 5.4. The table also benchmarks progress made on the action points identified in the 2001 CPAR. 207  Table 5.4. Action Plan as Provided in CPAR 2001 Actions Responsibility Achievement Establish an interministerial committee to draft a new public GOT Done procurement law. Consult interested stakeholders. Hire international legal experts to assist the Drafting Committee. Present a new public procurement law to the TGNA by October 2001. Apply .. . .public procurement law to all agencies at central and local GOT Partially done levels, using budgetary or extrabudgetary public funds, and to SEES. Draft and enact implementing regulations. GOT Done Prepare and enact a medium-termplan for progressive reform o f the GOT Pending EU public procurement legislation to achieve full alignment with EU procurement Procurement Directives. screening process Undertake a review of public investment programming, budget GOT Partially done planning, and procurement planning. (Coordinate with the PEIR.) MOF Introduce a prequalificationprocedure for large or complex GOT Done contracts. (Include pre-qualification provisions in new public procurement law.) Reform the Contractor Certificate system. GOT, MPWS DONE Improve market access by graduated reform and relaxation o f legal GOT Partially done and administrative obstacles to participation by foreign bidders. Revise the range of procurement methods: reinforce open bidding as GOT Done the main procurement method; introduce methods for small-value purchases, direct contracting and consultants’ services. (Include methods and their conditions for use in new public procurement law.) Increase transparency and fairness: lengthen bidding periods, reform GOT Done bid opening procedures, use objective bid evaluation criteria; introduce more rigorous requirements for advertising tenders and announcing contract awards; abolish verbal bidding, bracketing and expression of bids as percentage discounts off unit prices. (Include in new public procurement law.) Introduce standard prequalification and bidding documents for civil GOT Done works, goods, and services. Improve the bid protest mechanism; provide a legal basis and GOT Done institutional arrangements for administrative review. (Link to the establishment o f a central public procurement office.) Establish a new central public procurement office. (Establish and GOT Done define its function in new public procurement law.) Draw up a national training strategy for public procurement. GOT Not yet Develop a national training program for public procurement and GOT Training Not yet deliver it through identified training institutions. institutions Establishprocurement as a profession within the civil service. GOT Not yet Establish a professional accreditation scheme for government GOT Not yet procurement officers. Develop a more constructiverelationship with the business sector GOT Partially done through outreach programs Increase resources and investigative powers o f TCA. TGNA Draft bill submitted to TGNA Train TCA auditors in the application o f the public procurement law GOT TCA Publish results of procurement audits. TCA Not yet Strengthen internal audit functions of procuring entities to audit GOT Partially done procurement. Procuring entities Strengthen provisions in the criminal law on procurement-related GOT Done corruption and enforce them more rigorously. (Include reference to criminal law provisions in new public procurement law.) Insert specific prohibitions against fraudulent and corrupt activities GOT Done by bidders in bidding and contract documents. (Coordinate with Procuring entities development of standard bidding documents.) Put in place facilities for reporting allegations of fraud and GOT Done corruption. Introduce clear conflict of interest rules. (Include in new public GOT Done procurement law and standard bidding documents.) Source: World Bank Staff G. COMMON REFORM DIRECTIONS 5.97 Implementation o f the newly enacted laws represents a major challenge. Never in Turkey’s history have so many new laws been enacted altogether affecting public financial management. If successfully implemented, they could dramatically alter longstanding attitudes and practice. A bureaucratic culture that has traditionally been closed, conservative, centralized, control-oriented, and hierarchical could be transformed toward transparency and public service, opening itself to participation by c i v i l society and enhancing accountability to citizens. Yet to succeed, the reform agenda must be fully planned and effectively executed. Although some aspects o f the new legal framework are being piloted intermittently, most o f the secondary regulations (that is, those related to practical implementation) are yet to be issued. None o f the reforms have yet been fully implemented. The government’s decision to undertake fundamental changes in the responsibilities o f provincial and municipal governments also injects new complexity and considerable uncertainty into the PFMC implementation schedule. This, in part, has contributed to the delay in PFMC implementation. 5.98 Implementation o f major public reforms needs strong coordination and monitoring. By the nature o f the issues that they address, public sector reforms tend to be cross-cutting, and they require time for implementation. Strong high-level leadership must articulate a clear vision o f the objectives. Pragmatic midstream adjustments must adapt to changing circumstances, and coordination must be effective across units o f government. In part, the PFMC and the Public Finance and Debt Management (PFMD) laws spearheaded the public sector reforms to address concerns about recurring fiscal crises. Weak interagency coordination would pose the inherent risk o f provisions being enacted that are inconsistent with or contradict the previous reforms. Enactment o f the Omnibus L a w has created concern that i t s provisions run counter to the intention o f the P F M D and P F M C L a w as approved by the Parliament-a measure to improve fiscal 209 transparency and to ensure that policy decisions are subject to consideration in the context o f a medium-tern fiscal strategy that restrains expenditure commitments. It i s crucial that Turkey maintains the credibility o f its major public financial management reforms; so the need for strong, effective coordination to reconcile such conflicts cannot be overstated. Without arrangements to ensure effective leadership, implementation o f the reform agenda will remain at high-risk. 210 ANNEXES 21 1  ANNEX I Data and Methodology for Consolidated General Government Classification o f Expenditures 1. Economic Classification: In this analysis, the State Planning Organization (SPO) data produced for the Pre- Accession Economic Program (PEP) has been used for the consolidated budget, social security institutions, local administrations, revolving funds, extra-budgetary funds and unemployment insurance fund. With the implementation o f the PFMC law in January 2006, the institutional coverage has changed to be more in line with the international definitions and to increase the scope o f the general government institutions, however, since the last year covered in Chapter Ii s 2005, the terminology used here and the coverage are not consistent with the PFMC law. The SPO is the only public institution which estimates the size o f the general government for Turkey. In generating consolidated general government data the SPO is following the GFS manual. The IMF i s another source for the general government data in Turkey. However, the purpose o f their data i s not to measure the overall size o f the revenues and expenditures o f the general government, but to monitor the performance criteria set in terms o f the primary surplus o f the public sector. The definition o f general government used in this report includes: consolidated budget (CB), local administrations, revolving funds, social security institutions (SSI), Revolving Funds and a common subset o f budgetary and extra budgetary funds (EBFs) Consolidated Budget Institutions: All o f the general and annex budget institutions. As o f 2005 the total number o f the institutions was 98. Local Administrations: 3225 municipalities, 8 1 special provincial administrations, Iller Bank, 16 water and sewerage companies o f metropolitan municipalities and 10 natural gas and public transportation companies. Revolving funds: 1,083 (in 2005) enterprises established under the consolidated budget institutions and TRT (Turkish Radio and Television), DG o f Dormitory and Student Credits, National Lottery and A O C (Ataturk Forestry Farm). Social Security Institutions: SSK, BagKur, Emekli Sandigi and Unemployment Insurance Funds. The UI was included into the balance since year 2000. Extra Budgetary Funds: The total number o f the funds included in the general government balance decreased from 12 in 1999 to 4 in 2005. 212 I t should also be noted that the analysis does not include the expenditures o f the Central Bank and other public depository institutions, 36 non-financial state economic enterprises (SOEs), eight regulatory and supervisory agencies, and 38 institutions covered under the general and special budget institutions listed in the Public Financial Management and Control L a w (PFMC).'33 The analysis includes the net subsidies and transfers between these institutions and the consolidated budget. Although the SPO i s in general in line with the GFS methodology three adjustments were made to their general government data for the following reasons; 0 To eliminate the double counting: Invoiced payments which are social assistance type o f payments made by the Emekli Sandigi through transfers from the consolidated budget were deducted from the social security institutions balance. Spending on common retirement from BagKur has been netted o f in the S S K balance because the same amount o f spending i s already reported in the BagKur's expenditures. To transfer cash -based accounting into accrual accounting; Interest payments realized in 2001 but reported in 2002 were deducted from 2002 expenditures and added to 2001. 0 Moreover, in order to be in line with the primary surplus definition o f the GFS 2001 manual, interest revenues o f the institutions were deducted. In their current definition o f primary surplus, the SPO has been deducting only the interest expenditures but not interest revenues. 2. Functional Classification: 7. In this study a cross classification o f general government expenditures -- functional vs economic -- for 2003 and 2004 by combining the expenditures o f the institutional coverage defined in the economic classification was used. Regulatory and Supervisory Institutions and some o f the special budget institutions which were recently brought under the definition o f general government by law 5018 were not included because o f the data problem. 8. M a i n data sources for the study are as follows. 0 SPO economic classification o f the general government data Ministry o f Finance, consolidated budget and revolving funds 133 Special budget institutions refer to 83 public entities including 55 higher education institutions and universities established as affiliated or related to a ministry t o provide certain public services. These special budget institutions receive revenues and are authorized t o spend them. The complete list o f these institutions i s presented in the Public Financial Management and Control Law. 213  0 SSK, Bagkur, Emekli Sandigi 0 High Audit Board (YDK) on the accounts o f SSK, Bagkur and TRT (Turkish Radio and Television Board ) ,AOC ( Atatiirk’s Farm) , General Directorate o f Credit and Dormitories. 0 National Lottery Directorate 0 IMF fiscal tables o f June 2006 0 Ministry o f Internal Affairs, General Directorate o f Local Authorities 9. As a general principle the GFS consolidation methodology has been used while producing the functional classification o f the general government. A detailed analysis on the methodology and assumptions for producing the cross classification i s available in Ferhat Emil, 2005, Turkey-Economic and Functional Classification o f General Government Expenditures, background analysis note for the CEM: promoting Sustained Growth and Convergence with the European Union. Table A1 below shows the results o f this study. 214 a .I a 9 w .3 cd h d W r, 0 .. r10 cec  1 3 F W 0 0 3 e 2 0 3 0 '51 c.l W C .I g . U I 1 P .I L m U .I U 2 I 0 .I U B 3 L .I d 8 0 0 d c.l 4 g 1 : P 5 P 5 c: e c U 5 ; i a C z .C U E (: I A m I E (: I - I 8 i ; d A c 0 z E C z .C U h m r: G 8 a E U E A 8 *s U i 8 C C 9 8 h P Y v m \o 0 0 t-4 1 e a4 t 3 c I c) 8 u d . 0 I c) 9 E m P 5 m u E m I B .I c) 2 s k e 9 i B w s k a4  7 7 1 U *$ .$ I$ Y  I 1 U Y od 0 0 2 -f 0 0 N C .I k .I c) n I B L3 B L m L3 a c) r, r, 0 0 s s I A U C v 0I d . c) IA 0 0 u 0 0 e C N N I 1 r- u- L .I 8 3 e 8 C 0 e 8 s s t t U t A I E E 3 c, .;3 B c) 8 c) e C u u 5 . u I U 9 5 .I 9 c) 3 9 E 0 E m - . c)I m A A L P B m B 0 t c) B .3 C .I m e m E u c E u - z z c I 2 I r 0 C 0 0 x I x c )I . . c)I 8 I 3 cr I 0 r U 9 2 A U C cd -2 4 i 0 C E 2 C : w 8 E P 8 w a 8 L 8 3 E d 8 h 8 n z m c) a 3 4 3  1I 8 I c1 - co S 0 X I C A X CI 0 g \o 0 0 r4 *" E e E il I d L W u $ .I W La 0 5 v) 1 I B e v) u I d e 0 s U .I 1 k x 0 -2 s 8 W e. % 8 J 7 X 1  ANNEX111 Estimating Structural and Cyclical Budget Balances 1. The structural budget balance reflects a fiscal policy adjusted for the effects o f cyclical movements in economic activity on budget. In other words, the structural budget balance reflects what government revenues and expenditures would be if the economy was operating at potential level o f output and therefore does not include the effect o f cyclical developments in economic activity. In contrast, the actual budget balance does include the effect o f the cyclical component o f economic activity and therefore fluctuates around the structural budget balance. The methodology for estimating structural budget balance consists o f two steps; measuring the potential output and linking the related budget items, revenues and expenditures, to the cycle. However, the common difficulties in application are the estimation o f robust revenue and expenditure elasticities with respect to economic activity and the estimation o f potential output which i s itself an unobservable variable. There are different approaches used by many countries or international institutions and the difference mainly comes from the method to identify the cycle and determine the sensitivity o f budget to the cycle. 2. There are two widely used methodologies to identify the economic cycle. Smoothing methods such as Hodrick-Prescott filters are commonly used to estimate trend output and the cyclical component o f the economic activity i s calculated as the difference between trend and actual output. Although, this approach i s simple and transparent, it suffers from end point bias. That i s the difference between actual and trend series gets smaller for recent observations with a meaning that economic slowdown or overheating may not be truly cached. A second approach aimed at eliminating this drawback i s to estimate potential output based on a production function. The production function approach is relatively more complex since i t requires judgments on the rate o f technological change and the rate o f structural unemployment. However, i t i s less susceptible to the end point bias for recent observations. The production function approach i s used by many international institutions such as OECD and the EU as well as by many countries. I t i s also used by the Turkish government since 2003, in i t s annual Pre-Accession Economic Programmes (PEP) submitted to the EU. Considering the methods to determine the sensitivity o f budget to the cycle, there are different approaches for estimating revenue and expenditure elasticities, ranging from the ones using single equation regression analysis to the ones using a macro-econometric model 3. The structural budget balances are in practice calculated from revenues and expenditures adjusted for the deviation o f actual output from potential using estimated elasticities linking revenues and expenditures to economic activity. Therefore, the structural budget balance i s defined as follows: sb* : Structural budget balance (as a ratio o f national income), 222  Taxi* : The structural value o f tax revenues in category i, GEi* : The structural value o f government expenditures in category i, OER : Other expenditures and revenues which are not affected by the growth cycles, Y' : Potential output. The relationship between structural tax revenues and expenditures and their actual values can be shown as follows: Tax, : Actual tax revenues in category i , GEi : Actual government expenditures in category i , Y : Actual national income, ai : The output elasticity o f category itaxes, pi : The output elasticity o f category igovernment expenditures. 4. n this study, the potential output is calculated by using the Production function I Method. Tax revenues are divided into 3 categories showing cyclical movements. These categories are taxes levied o n income, corporate taxes and indirect taxes. Following the standard practice in the literature and in international applications, the elasticity o f indirect taxes is taken as unit elasticity. O n the other hand, the elasticity o f the other two taxes was estimated from regression analysis. In the case o f general government, social security contributions were also adjusted for the cycle using an elasticity estimated by a regression analysis. On the expenditure side theoretically only unemployment benefits and social security benefits are assumed to be affected by the growth cycles. However, introduction o f unemployment benefits in Turkey is a very recent issue and the eligibility conditions to apply for a benefit are quite restrictive. Therefore, unemployment benefits in Turkey s t i l l do not show a cyclical pattern. In this context, while estimating structural balances for the general government, unemployment benefits are not adjusted for the cycle. This leads to the following equation for the calculation o f the structural budget balance; i : Income tax, corporate tax and indirect tax while making estimations for consolidated budget and it also includes social security contributions as a separate tax item while making estimations for the general government. 5. Once the structural budget balance calculated, the difference between the actual budget balance and structural budget balance can be defined as the cyclical budget balance. cb** = b - sb* 223  cb** : Cyclical budget balance (as a ratio o f national income) b : Actual budget balance (as a ratio o f national income) 6. In the calculation o f the potential output with a production function, it was assumed that output i s determined as a function o f labor, capital stock and TFP. A Cobb- Douglas function with constant returns to scale i s used as production function and TFP is calculated as Solow residual. r, = L,"K,'-"TFP, Y: GNP, L: Labour demand (employment) and TFP: Total Factor Productivity 7. In the transition from GNP to potential GNP, TFP has to be de-trended and the potential labor series has to be estimated. TFP was de-trended using Hodrick-Prescott filter. In estimating the potential labor-force series, supply o f labor-force was estimated by taking into consideration the historical trend o f the participation rate in the labor-force and then the estimated labor-force supply series was corrected by nairu (non-accelerating inflation rate o f unemployment). Concerning the estimation o f nairu, under the variable nairu assumption, the unemployment rate was put through the Hodrick-Prescott filter and the nairu series was calculated*34. =[POP15, *PART, *(l-nah,)]" Y p t = LPt"K,'-"TFPPt *K,!-" *TFP,' K: Capital stock TFPP: Potential total factor productivity Yp: Potential GNP Lp:Potential labour force POP15: Working age population PART: Participationrate nairu: Non-accelerating inflation rate of unemployment 8. Regression methodology i s used for the estimation o f tax elasticities. First o f all, following the standard practice in the literature and in international applications, the elasticity o f indirect taxes i s taken as unit elasticity. Therefore, tax elasticities are estimated only for Personal Income Tax (PIT), Corporate Income Tax (CIT) and Social Security Contributions (SSC). A co-integration between income and C I T would normally be expected and therefore, using Engel-Granger methodology, a co-integrating relation i s tested between C I T and GNP, between PIT and GNP and between SSC and GNP. To this end, all variables are tested for stationarity and the results presented in Table A3.1 below show that all variables are not level stationary but their first differences are stationary. That i s to say all variables are integrated o f order one and therefore a con-integration relation can be tested between income and tax variables. 9. Considering first CIT, the time path o f CIT as ratio to GNP is presented in the Figure A3.1 below, which shows three distinct periods. An increasing path through 1975- 1987, a declining path through 1988-1993 and an increasing path once again through 1993- 2005. A co-integration i s tested by regressing log o f CIT on a constant, log o f income and two time dummies in order to account for the broken time path and then residuals are tested i 3 4 For a detailed discussion o f the topic see Denis C., K.Mc Morrow, W. Roger, "Production Function Approach to - Calculating Potential Growth and Output Gaps Estimates for the EU Member States and the US", European Commission Directorate-General for Economic and Financial Affairs, Economic Papers 176, September 2002. 224 for stationarity. The results show a strong and statistically significant relation between income and C I T with an estimated elasticity o f 1.57. The stationarity o f the residuals indicate a co-integrating relation between the two variables. 10. Similarly, the time path o f PIT as ratio to GNP is presented in the Figure A3.3 below. The figure shows that the ratio o f PIT to GNP fluctuated a lot throughout the period. Following an initially increasing path, the ratio declines through 1980-1984 and shows fluctuations around an increasing trend thereafter. Especially the reversed u-shaped pattern at the beginning o f the sample i s hard to be explained by the developments in GNP. Therefore, we chose the more stable second part o f the sample (1984-2005) to search for a reliable relation between PIT and the GNP. However, even for the second part, there is stable decline (PIT declines also in real terms) at the end o f the sample, particularly in 2004 and 2005. This could be due to fact that the government adopted a medium term tax strategy covering the period o f 2002-2004 and within this context many changes took place in taxation regime starting from 2003. The government also introduced tax allowances for new employment created in some lagging regions in 2003. Moreover, there has been a shift towards indirect taxes and all these changes may have brought a decline in the PIT to GNP ratio at the end o f sample. The weak employment generation in the post-2001 era could be another factor supporting this decline in the ratio. Therefore a co-integration relation i s tested for two different samples, 1984-2003 and 1984-2005. 11. First, a co-integration i s tested by regressing log o f PIT on a constant, log o f income and a time dummy for the end-sample in order to account for the decline in PIT in real terms and then residuals are tested for stationarity. The results show a strong and statistically significant relation between income and PIT with an estimated elasticity o f 1.50. The stationarity o f the residuals indicate a co-integrating relation between the two variables. A second co-integration relation is tested by regressing log o f PIT o n a constant and log o f income for the sample 1984-2003 in order to exclude the end period decline. The results again show a strong and statistically significant relation between income and PIT with an estimated elasticity o f 1.46. The stationarity o f the residuals indicate a co-integrating relation between the two variables. Both equations support existence o f a co-integration relation between PIT and GNP and the estimated elasticities are very close to each other. For that reason, we use the elasticity o f 1.5 estimated for the full sample in the calculation o f structural budgets. 12. Finally the time path o f SSC as a ratio to GNP is presented in the Figure A3.6 below, which shows two distinct periods. A declining path through 1975-1986, and then an increasing path through 1987-2005. A co-integration is tested by regressing log o f SSC on a constant, log o f income and a time dummy in order to account for the broken time path and then residuals are tested for stationarity. The results show a strong and statistically significant relation between income and SSC with an estimated elasticity o f 0.82. The stationarity o f the residuals indicate a co-integrating relation between the two variables. 225 Table A3.1: Unit Root tests NullHypothesis: Variable has a unit root (Sample: 1975-2005, Lag selection: automatic based on SIC, MAXLAG+) Level with constant 1' M e r e n c e with constant t-statistic Rob.* t-Statistic Rob.* CIT 0.3575 18 0.9775 -5.672049 0.0001 PIT -1.425974 0.5563 -5.5415 18 0.0001 ssc 2.134058 0.9998 -3.093936 0.0382 GNP 0.950752 0.9948 -5.674495 0.0001 *MacKinnon (1996) one-sided gvalues Test critical values: 1% level -3.808546 5% level -3.020686 10% level -2.650413 Level with trend 1' difference with trend CIT -1.472806 0.8166 -5.807175 0.0003 PIT -2.394180 0.3749 -5.429284 0.0007 ssc -0.611427 0.9702 4.904890 0.0026 GNP -2.109044 0.5202 -6.056936 0.0001 *MacKinnon (1996) one-sided pvalues Test critical values: 1% level 4.498307 5% level -3.658446 10% level -3.268973 Source: World Bank Staff calculations 226  CorDorate Income T a x Figure A3.1: CIT as a Ratio to GNP 0.51, 1 , , , , , , , , , , , , , 1 , , , , 1 , , ,~ I , , , , I , , 1975 1980 1985 1990 1995 2000 2005 I -CIT/GNP*100 I Source: World Bank Staff calculations Table A3.2: Regression o f CIT on Income Dependent Variable: LOG(C1T) Method: Least Squares Date: 05/18/06 Time: 10:13 Sample (adjusted): 1975 2005 Includedobservations: 3 1 after adjustments Variable Coefficient Std. Error t-Statistic Prob. C -11.08901 1.947756 -5.693223 0.0000 LOG(GNP) 1.570558 0.174202 9.015715 0.0000 T 1-CIT 0.047 138 0.008424 5.595815 0.0000 T3-CIT 0.053354 0.015482 3.446170 0.0019 R-squared 0.9 18644 Mean dependent var 6.966399 Adjusted R-squared 0.909605 S.D. dependent var 0.654503 S.E. o f regression 0.19678 1 Akaike info criterion -0.293532 Sum squared resid 1.045519 Schwarz criterion -0.108501 Log likelihood 8.549744 F-statistic 101.6256 227 D u r b i n - W a t s o n stat 1.558320 Prob(F-statistic) 0.000000 Source: World Bank Staff calculations Figure A3.2: Actual, Fitted and Residuals 8 .O 7.5 7.0 6.5 6 .O 5.5 \ 1975 1980 1985 1990 1995 2000 2005 I -Residual -Actual -Fitted I ource: World Bank Staff calculations 228 Table A3.3: Residual Stationarity Null Hypothesis: RESID01 has a unit root Exogenous: Constant Lag Length: 0 (Automatic based on SIC, MAXLAG=7) t-Statistic Prob." Augmented Dickey-Fuller test statistic -4.174036 0.0029 Test critical values: 1% level -3.670170 5% level -2.963972 10% level -2.621007 *MacKinnon (1996) one-sided p-values. Source: World Bank Staff calculations 229  Personal Income Tax Figure A3.3: PIT as a Ratio to GNP 1 -PIT/GNP*100 I ource: World Bank Staff calculations Table A3.4: Regression o f PIT on Income Dependent Variable: LOG(P1T) Method: Least Squares Date: 05/24/06 Time: 17:49 Sample: 1984 2005 Included observations: 22 Variable Coefficient Std. Error t-Statistic Prob. C -8.766768 1.557270 -5.629577 0.0000 LOG(GNP) 1.502484 0.136301 11.02326 0.0000 T 1-PIT -0.142184 0.043240 -3.288240 0.0039 R-squared 0.875356 Mean dependent var 8.430307 Adjusted R-squared 0.862236 S.D. dependent var 0.339875 S.E. o f regression 0.1261 50 Akaike info criterion -1.176568 Sum squared resid 0.302362 Schwarz criterion -1.027790 L o g likelihood 15.94225 F-statistic 66.71730 Durbin-Watson stat 1.730146 Prob(F-statistic) 0.000000 Source: World Bank Staff calculations 230 Figure A3.4: Actual. Fittedand Residuals -Residual -Actual ~ Fitted Source: World Bank Staff calculations Table A3.5: Residual Stationarity Null Hypothesis: RESID01 has a unit root Exogenous: Constant Lag Length: 0 (Automatic based on SIC, MAXLAG=4) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -4.101883 0.0057 Test critical values: 1% level -3.83 15 11 5% level -3.029970 10% level -2.655194 *MacKinnon (1996) one-sided p-values. Source: World Bank Staff calculations 23 1 Table A3.6: Regression o f PIT on Income Dependent Variable: LOG(P1T) Method: Least Squares Date: 05/24/06 Time: 17:50 Sample: 1984 2003 Included observations: 20 Variable Coefficient Std. Error t-Statistic Prob. C -8.290114 1.585252 -5.229526 0,0001 LOG(GNP) 1.460238 0.138625 10.53372 0.0000 R-squared 0.860421 M e a n dependent var 8.405536 Adjusted R-squared 0.852667 S.D. dependent var 0.346960 S.E. of regression 0.133 177 Akaike info criterion -1.099630 Sum squared resid 0.319252 Schwarz criterion -1.000057 Log likelihood 12.99630 F-statistic 110.9593 Durbin-Watson stat 1.570086 Prob(F-statistic) 0.000000 Source: World Bank Staff calculations 232 'igure A3.5: Actual, Fitted and Residuals - 9.2 - 8.8 - 8.4 - 8.0 - 7.6 I -Residual -Actual -Fitted I ource: World Bank Staff calculations Table A3.7: Residual Stationarity Null Hypothesis: RESID01 has a unit root Exogenous: Constant Lag Length: 3 (Automatic based on SIC, M A X L A G 4 ) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -4.434567 0.0037 Test critical values: 1% level -3.920350 5% level -3.065585 10% level -2.673459 *MacKinnon (1996) one-sided p-values. Source: World Bank Staff calculations 233  Social Security Contributions Figure A3.6: SSC as a Ratio to GNP .07 .06 - .05 - .04 - .03 - .02 , , , , , 1 , , 1 , , , , 1 , , , , 1 , , , , 1 , , , , 1975 Id80 1985 1990 1995 2000 2005 1- ource: World Bank Staff calculations Table A3.8: Regression o f SSC on Income Dependent Variable: LOG(SSC) Method: Least Squares Date: 06/02/06 Time: 12:28 Sample: 1978 2005 Included observations: 28 Variable Coefficient Std. Error t-Statistic Prob. C -1.649974 2.795110 -0.590307 0.5603 LOG(GNP) 0.815429 0.253353 3.218549 0.0036 SSC-TI 0.062682 0.012469 5.027038 0.0000 R-squared 0.960094 Mean dependent var 8.027606 Adjusted R-squared 0.956902 S.D. dependent var 0.684470 S.E. of regression 0.142097 Akaike info criterion -0.963658 Sum squared resid 0.504788 Schwarz criterion -0.820922 Log likelihood 16.49122 F-statistic 300.7369 Durbin-Watson stat 0.674616 Prob(F-statistic) 0.000000 Source: World Bank Staff calculations 234 Figure A3.7: Actual, Fitted and Residuals II -Residual -Actual __ Fitted Source: World Bank Staff calculations Table A3.9: Residual Stationarity Null Hypothesis: RESID01 has a unit root Exogenous: Constant Lag Length: 1 (Automatic based on SIC, MAXLAG=6) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -3.250778 0.0282 Test critical values: 1% level -3.71 1457 5% level -2.981038 10% level -2.629906 *MacKinnon (1996) one-sided p-values. Source: World Bank Staff calculations 235  ANNEX I V T a x performance: A Regression There has been only limited effort to develop comprehensive tools for assessing tax performance across countries so far. Typically, there are two main approaches used to measure a country’s tax effort. In its simplest form comparisons can be based on differences between the effective tax rates and the standard tax yield following the methodology developed in Tanzi (1981), Schaffer and Turley (2000). An alternative is to calculate a tax effort index as the ratio o f actual tax share to the predicted (or potential) tax share (regression approach). The predicted tax ratio is determined from regression relating tax shares to various explanatory variables that serve as proxies for tax bases or other factors that might affect country’s ability to tax. Following recent tax effort literature (e.g., Stotsky and WoldeMariam (1997), Piancastelli (2001), Eltony and Nagy (2002) Bird, Martinez-Vazquez and Torgler (2004) and Hudson and Teere (2004)) a stochastic model i s used for estimating tax revenue, where T N is the tax ratio and Xi (i= l....n) represent various independent variables expected to influence the tax ratio, while U i s the error term: T N = f(Xi ... Xn, U) The independent variables employed in the basic model are similar to those used in the most recent literature: gross national product per capita, the ratio o f trade to GDP (imports plus exports over GDP), the share o f the mining sector and the agricultural sector in GDP, and population growth. An overview o f the variables applied in previous empirical studies i s provided in the table below. Other variables, such as external debt, CPI, rural population etc, are expected to check robustness of the base results. A time trend is intended to capture any overall trend in taxation. The analysis uses panel data for 57 developed and developing countries, including 26 E C A and 6 non-ECA comparators countries over the period 1995-2004. The choice o f sample is motivated by the need to obtain a data set composed o f countries with similar characteristics to E C A and comparator countries, as well as data availability o f information. Data were obtained from World Development Indicators, IMF Regional Fiscal Data Set, IMF Country Profile Chapter IV, Schneider and Klinglmair 2003 and the Government Finance Statistics from M O F s in respective countries. A set o f 57 countries was used, comprising three groups: 10 in the lower middle income 16 in the upper middle income group and 31 in the higher income group137,as defined by the World Development Indicators 2004. 135 This annex draws on empirical work currently conducted for a World Bank study on Public Finance Policies and Growth in Europe and Central Asia. 136 $825-$3,255 GNI per capita 13’ above$l0,065 GNI per capita (31 countries in our sample) 236 Table 1 Overview of empirical findings in tax effort studies Variable GDP per capita + +/- + + +/- + Population density + Population growth Urban Pomlation + Agriculture, value added as % o f GDP + - Manufacturing, value added as YOo f GDP + + +/- Mining, value added as % o f GDP Services, value added as YOo f GDP GDP + Import, as % o f GDP +I- + Export, as % o f GDP + Trade (Export + Import as % o f GDP) + + External Debt, as % o f GDP - Consumer Price Index Inequality Aid, as percentage to GDP + Share o f Fuel in total export + - I +/- Index governance I + I /+/ Regulation to entry Composite Institutional quality Tax morale I +/- Method o f estimation The panel data model was estimated with both “fixed effects” (using the least squares dummy variable (LSDV) approach) and “random effects” (applying generalized least squares (GLS) approach). The Hausman test consistently rejects the random effects model in favor o f the fixed effects model. 237 However, in the next step the normal distribution o f the error term was rejected and diagnostic tests revealed problems o f cross-sectional correlation. To deal with the problem o f cross- sectional correlation, the Prais-Winsten estimator^'^' were employed. (5.690) GDPpeaCZ#aT& 2.740* (4.150) ChEt 324..56 20.96* 22.26* 25.72* 29.59* 17.38* 16.51* (198.52) (367) (367) (1.754) (2.271) (1.386) (1.337) N 485 509 426 426 282 485 485 R-sq 0.59 0.59 0.59 0.60 0.44 0.60 0.63 13* The method i s an alternative t o feasible generalized last squares for fitting the linear cross-sectional time -series models when disturbances are not assumed t o be independent and identically distributed, and it is preferable t o the feasible GLS when the number o f observations and time span are limited. 238 The model performs generally well with estimated coefficients for the explanatory variables in line with the previous findings in the literature (see Table 2). Higher GDP per capita i s associated with a higher tax ratio. The structure o f the economy seems also t o matter. The tax ratio is negatively related to the share o f agriculture in GDP and positively related to the share o f manufacturing sector in GDP, but the latter variable proved to be statistically insignificant (equation 1). The insignificance o f manufacturing i s somewhat surprising, although i t may potentially be explained by a negative correlation with agricultural share or the fact that the manufacturing share varies across countries dependent o n the stage o f development. Thus the regression was reestimated with interactive terms between manufacturing and GDP per capita (equation 6) included. Accordingly, manufacturing has become significant. Moreover, the significantly negative coefficient on the manufacturing-GDP per capita interactive t e r m m a y indicate that as countries develop the importance o f manufacturing as a source o f tax revenue declines. Moreover, the agriculture sector i s much more difficult to tax for less developed countries, as indicated by the significantly positive coefficient o n the agriculture-GDP per capita interactive term. A faster rate o f population growth leads to a lower tax ratio, while openness is associated with a higher tax ratio. Inclusion o f dummy variables controls for differences in stage o f development in the sample and reveals that the lower middle income countries have statistically significant lower tax ratio than other countries in the sample. On the contrary, i t is observed that high income countries have a statistically significant higher tax ratio than other countries in the sample. Moreover, in order to get a more realistic picture o f a country’s taxable capacity vis-a-vis its natural resource base a dummy variable13’ for important o i l producer countries was included. The easiness o f taxing natural resource extraction is likely to generate more tax revenue than non-fuel activities. The coefficient for o i l dummy has the predicted negative signs and i s statistically significant in all equations. Finally, the trend variable i s generally negative, indicating that, all else equal, tax ratios are o n a downward trend as a result o f global tax competition. However, i t is not statistically significant (equation 1). A similar exercise was carried out with respect to the indirect and direct taxes collection across countries. 13’ OIL dummy takes value o f 1 if the share o f fuel (and related products) in total merchandize exports exceeds 40% i s negative and insignificant 239  ANNEX V DRAFT TAX BENCHMARKS FOR TAX EXPENDITURES A tax benchmark identifies the normal tax arrangements that apply to similar classes o f taxpayers or types o f activity. Tax expenditures are defined as deviations from the benchmark. The principal criterion o f benchmark design is that the benchmark represents a consistent tax treatment o f similar activities or classes o f taxpayers and neither favors nor disadvantages similarly placed activities or classes o f taxpayers. includes certain tax provisions (such as exemptions, deductions, tax credits, and other tax preferences) to adjust taxable income in order to: o comply with the ability-to-pay principle o enhance the economic and collection efficiency o f taxation o simplify or make feasible tax administration with respect to a class o f taxpayers or type o f activity ensures that tax expenditure report provides sufficient information for policy formulation. In addition, each country will have i t s own purposes or demands for a tax expenditure report so that the benchmark should reflect such purposes or demands. The following benchmarks summarize some o f the basic features that could be consider forming the benchmarks o f the personal income tax (PIT), corporation income tax (CIT), Value-added tax (VAT) and special consumption tax (SCT) in Turkey. Personal Income Tax Benchmark Tax period, calendar year (as given by the personal income tax law) Tax unit: individual (as given by PIT law) Residents taxed o n world wide income, non-residents on income sourced from Turkey Taxable net income is gross income reduced by the allowable cost s o f earning income (allowable deductions) Gains and losses realized on movable and immovable property Losses carried forward for five years (given by the law) Inflation. Tax base i s fully adjusted for inflation (as given by the law, since 2004) Tax rate schedule and income brackets for individual income earners as applicable to a tax year (as given in the law) Final withholding taxes (no declaration o f income) (as given in the law) o Final withholding on specified types o f investment income o Final PAYE taxes paid on employment income (including the special tax credit for basic expenditures) Provisions to reduce or eliminate double taxation (50% dividend deduction) Foreign tax credit: Credit is available to residents for income tax paid or incurred to any foreign country (as given by the law) 240  Exemptions provided for persons subject to diplomatic privileges on a reciprocal basis Exemption provisions for central and local governments and non-government organizations o f a public nature in the context o f providing services o f a public nature o n a non-commercial basis Corporate Income Tax Benchmark Tax period: calendar year /end o f the financial period o f the company (given by the law) Tax unit: legal entity (given by the law) Resident entities taxed on world wide income, non-residents on income sourced from Turkey Taxable income as calculated on the basis o f balance sheet methods Gains and losses realized on movable and immovable property Tax rate: General standard rates applicable to year (given by the law) Valuation: inflation adjustment policies as applicable (given by the law) Depreciation: depreciation based on the economic life, using the straight-line method or declining balance method. Capital revaluation exemption (given by the law) Losses carried forward for five years (given by the law) Intercorporate dividends are exempted (given by the law) Deductible distributions o f surplus o f cooperatives as part o f tax integration Up to 50% deduction o f repatriated profits as part o f tax integration Foreign tax credit: Credit is available to domestic corporations for income tax paid or incurred to any foreign country (as given by the law) Tax deferred on capital gains arising from mergers and acquisitions or (detail see the law) Tax exemption for income transfer due to corporate restructure or merger (given by the law) Exemption provisions for central and local governments and non-government organizations o f a public nature in the context o f providing services o f a public nature on a non-commercial basis VAT Tax Benchmark Tax period: month for domestic deliveries (supplies) and quarter for imports Valuation o f domestic deliveries and imports (given in the law) Timing o f recognition o f domestic deliveries and imports (given in the law) Tax rate: a nominal rate at 18% and a reduced rate (8%) for items on Annex H o f the EU Directives (see below) that permits VAT at a reduced rate, and super reduced rate (1%). Destination-based tax on domestic deliveries and imports (exports and related international transport services are zero-rated) Multi-stage tax with deduction o f input VAT and refund o f excess VAT only in case o f exports, and supplies at rates below standard rate; excess input VAT carried forward. Exemptions related to normal customs treatment o f transit trade, free zones, bonded warehousing, temporary importation, change o f residence, etc Deferment mechanisms (or equivalent structures) for VAT on imported and domestically supplied machinery and equipment 24 1  0 Exemption provisions for government and i t s agencies and non-government organization providing services o f a pubic character and o n a non-commercial basis (as specified in the law) 0 Exemption for self-employed businessmen, tradesmen and farmers (as specified in the law) 0 Exemption o f mergers and acquisitions, delivery and rental o f non-commercial immovable property; financial sector; delivery o f financial instruments. 0 Deliveries to persons with diplomatic privileges are fully exempted (zero rated) Special Consumption Tax Benchmark 0 Destination based single stage tax 0 Tax period: monthly 0 Valuation o f imports and domestic deliveries at factory gate, dealer or retail level as specified in law for the different specified commodities 0 Timing o f payments (as specified in the law) 0 Tax rates as specified for lists o f goods o List I :tax rates specified for petroleum products . o List 1 1, tax rates specified for vehicles . o List I11 tax rates specified for tobacco products tax rates specified for alcoholic products o L i s t N. Tax rate at 6.7% o n specified luxury items 0 Exemptions for exports and for persons with diplomatic privileges 0 Exemptions related to normal customs treatment o f transit trade, free zones, bonded warehousing, temporary importation, change o f residence, etc 0 Exemption for donations to government and its agencies 242  ANNEX V I Towards Estimates o f Tax Expenditures in PIT and CIT The following estimates are based on the summary PIT and C I T tax return for 2003. Based on analysis o f tax expenditure tables and some added data, there are a number o f questions, comments and observations towards estimates o f tax expenditures in PIT and CIT. PIT 1. Little can be gathered from the 2003 PIT return. Out o f the YTL 6.393 billion in income, some 29.2% i s payable in tax or YTL 1.868 billion. The only tax expenditures that can be identified are in item 26 which gives YTL 16.5 million in deductions for individual insurance and pension premiums (possibly under ITL Articles 89/1-3) and YTL 16.99 million in deductions for donations and grants to health and education organizations (possibly under I T L Articles 89/58) These represent small tax expenditures o f about YTL 5 million for item 26 and YTL 5.1 million for item 27 respectively as the bulk o f these deductions would occur as a result o f premiums paid by employees and would be captured under the PAYE. In addition, and importantly, most o f the PIT tax expenditures (some 100 provisions o f the law identified as tax expenditures) under the income tax law are structured as “exemptions” or “tax holidays” and hence no returns are filed and no data is directly available for tax expenditure estimates. CIT 2. The CIT summary return data i s considerably richer in tax expenditure information, but without (i)use o f a tax model and (ii) consideration o f the inter-temporal impacts on loss and exemptioddeductiodallowance carryovers, it still has limitations. For example, a deduction taken in one period can result in a loss o f taxable income in a later period through loss carryovers. These issues are expanded upon below. 3. A general problem with tax expenditures in the income tax arises from the changes effective tax rate depending upon the (a) the income level, (b) the size o f the exemption or deduction, and (c) the difference in timing o f accrual and realization o f the exemption or deduction: 243 a. Where there are different tax rates in different tax brackets, the income level o f the beneficiary o f the exemption or deduction affects the marginal tax rate (MTR) applied to the exemption or deduction. Ideally, a weighted average MTR i s required to apply to the amount o f the exemption or deduction. This is clearly an issue in estimating tax expenditures in the PIT, including any flow through o f dividends and other investment income from the company to personal levels. This i s typically solved by calculating tax expenditures using a micro simulation tax model based on a representative sample o f taxpayers so that the appropriate weights and M T R s are applied depending o n the distribution o f taxable income. b. The size o f exemptions and deductions affects tax expenditure estimates in two ways. i. The MTR i s affected by how many tax brackets are covered by the exemption or deduction. This i s most important if tax expenditure items are considered collectively rather than individually or one-at-a time. The best-practice tax expenditure convention i s to consider each one-at-a-time as a marginal decision. Even then a deduction may straddle one or more tax brackets and an average MTR i s applicable. Again the use o f micro simulation tax models with a sample o f representative taxpayers in terms o f taxable incomes and deductions allows the accurate estimation o f each tax expenditure item. ii. An exemption or deduction may make a taxpayer move from being taxable to non-taxable such that only part o f the exemption or deduction has current cash flow consequences. The tax value o f the remainder may be (a) lost forever, (b) captured immediately as a refund or loss carry back against prior year taxable income, (c) carried forward as an identifiable exemptiorddeduction with or without indexation (for example, in Turkey, unusable investment allowances are carried forward indefinitely with indexation), or (d) carried forward as a general loss o f taxable income to be deducted against future taxable income that may arise within a limited time period (such as the next five years, in the case o f Turkey). From a tax flow perspective, the potential tax value o f an exemption or deduction in the current year i s reduced in all cases except (b). In cases (c) and (d) additional tax value may be captured in some future year. Alternatively, the tax flow consequences in the current year arise from a combination o f the new deductions in the year plus the carry forwards o f deductions taken but not used in prior years. This issue o f the timing o f when tax expenditures are realized i s a major and difficult problem. I t i s discussed at some length below. c. As noted above in b.ii, exemptions and deductions may not be usable in the current year for lack o f taxable income. This problem i s treated in two different ways as discussed in cases (c) and (d) above. These two types are illustrated in Table 1 as “exemptions to be deducted even when these i s a loss,” which generate either traceable carry forwards or general (untraceable) loss carry forwards, and as “exemptions deducted when there i s a profit,” which have to be carried forward separately when there i s not a profit. The actual cash flow tax expenditure cost o f the former i s difficult to determine in any current year. In addition to some o f the exemptions or deductions not being claimable in the current year, unused amounts from prior years may be carried forward into the current year, and furthermore, the removal o f the exemption or deduction in an earlier year 244 affects the loss carry forwards into the current year and hence affects the current tax value o f current exemptions or deductions. The tax expenditure treatment o f exemptions that do not cause a loss carry forward (but unused amounts are carried forward separately) are treated subsequently below. 245 Table 1: Towards some preliminary CIT tax expenditure estimates and identification of issues 30% of exemption Or allowance Questions, issues, concerns YTL millions Exemptions t o be deducted even when there is loss Income from Investment Funds (CTL Article 1,671 Correct tax rate? Ideally should be taxed at 814) rate of person holding investment units or shares. Income from Pre-emptive Rights & Issue 30 Not all current exemption generates a Premiums (CTL Art. 8/51) current tax loss, but prior years exemptions may be currently deductible Income From Constructing Abroad (CTL 126 Not all current exemption generates a Art.8i7) current tax loss, but prior years exemptions may be currently deductible. Also issue of whether subject to 10% withholding such that tax loss is reduced C.T.L. Temporary Article 28la 498 Not all current exemption generates a current tax loss, but prior years exemptions may be currently deductible. Also issue of whether subject to 10% withholding such that tax loss is reduced Income from Free Trade Areas (Free Zones 89 Not all current exemption generates a Law 3218) current tax loss, but prior years exemptions may be currently deductible. Possibly missing tax returns. Does not include PIT exemptions on labor income Exempt Income under the Law 4325 for 33 Not all current exemption generates a Less Developed Regions current tax loss, but prior years exemptions may be currently deductible Income From Technology Develop. Regions 8.8 Not all current exemption generates a current tax loss, but prior years exemptions may be currently deductible Other Deductions and Exemptions 2,283 Given large size of this exemption, need to identify source of these other exemptions, deductions and allowances. Carryforwards Other Previous Year Loss 4,815 Previous Year Loss Resulting From 236 Exemptions Exemptions t o be deducted if there is profit Realized Invest. Allowance 2,934 Is this all 40% allowances or does it include some higher rate allowances subject to 19.8% withholding tax on the exemption? Does this amount include unused allowances carried forward from earlier years? What are the stocks of unused investment allowances and how much was added to these stocks in 2003? Provisioned Invest. Allowance 86 Other Deductions 57 Do be identified Source: Tax Revenue Administration and World Bank Staff 246 6. The cash f l o w tax cost can be broken out into two main components. i. The actual amount o f the exemption or deduction that can be absorbed in the year against current taxable income. This typically requires a tax micro-simulation model to calculate the amount o f the exemption or deduction used in the year. Alternatively, the amount o f the current deduction i s limited by the amount o f total loss carry forwards from the current year to the next year (LCF,,,). This LCF, i s the total o f traced and untraced LCFc sna the Ilwsed amounts o f the taxable income limited deductions (such as the investment allowance) 1,671 i e current deduction exceeds the L CF, its removal w i l l result in no current tax, but merely a decrease in LCF . ,, ii. The removal o f the exemption or deduction o f the tax expenditure type in previous years could have resulted in lower loss carry forwards into the year, and hence the amount o f the current exemption or deduction that would have been claimed in (i) above increases. The amount o f carry-forwards o f unused deductions or exemptions o f the same type used in the year. This can be precisely determined if the amount o f the unused exemptions or deductions are traceable (or are in type-specific pools carried forward). Otherwise some share o f the loss or exemption carry forwards has to be attributed to the type o f tax expenditure. Table 1 above shows a tax cost o f some YTL 222 millions in exemptions carried forward and claimed in 2003. In addition, some share o f the YTL 3.577 billon i n general losses carried forward into 2003 may be arising from untraced exemptions and deductions in earlier years rather than economic losses. Where use o f carry forwards o f traced deductions i s known, the impacts o f changes in these deductions on the carry forwards o f the unused income limited deductions (such as the unused investment allowances) also has to be accounted for. Any tax expenditure estimate has to bring into the picture the total losses carried into the current year as well as the total losses carried forward to the next year, These effects are jointly analyzed in Table 2. Here the tax expenditure is hown to depend whether he tax payer i s tax payable in the year (T>O) or not and (ii) critically o n (i) whether any losses are carried forward into the year. A taxpaying person will not be generating loss carry forwards to the next year and will get the full value out o f current deductions and will have used up all LCF into the year. The total LCF into the year determines the maximum carry forward o f unused deductions into the year and put an upper limit o n h o w much o f prior year unused losses can affect the tax flows in the current year. Where n o tax is payable in the year and losses are carried forward into the following year, then any increase in current year taxable income because o f the removal o f a specific deduction, may not will not result in the full value o f this deduction being captured as a tax increase to the extent the loss carry forward to the next year first has to be covered before any tax comes payable. Table 2 provides a useful way o f cross tabulating the tax returns in order to analyze and estimate the tax expenditures. Tax Payable, T = 0 Tax Payable, T > 0 Loss carry A. Partial TE for cases where (D B. Full TE = Dt forward, LCF - LCF,,t) t>0 =O Loss carry Dprior > LCF C1. Partial TE for cases where (D D1. Partial TE = (D -LCF) t forward, LCF +LCF - LCF,,,t)t >0 247 >O Dph r or LCF C2. Partial TE for cases where (D D2. Full TE = (D + Dpnor )t ’ + Dprior - LCFnexJ t 0 7. What are the implications o f the above analysis o f estimating tax expenditures o n a cash flow basis? First, that an annual micro simulation model i s necessary, but not sufficient to estimate tax expenditures where there are loss carry forwards or losses created by exemptions or deductions. Second, additional information i s required to estimate “TE”, namely, total LCF, LCFneXt and Dpnor . L C F and LCFnext should be knowable for each tax return and should be added to any tax return report. D p i o r requires the further work o f linking tax returns over the years. Finally, i t is useful to cross-tabulate estimates o f tax expenditures by the taxpayers status in terms o f being taxpaying or not (taxes payable in the current year > 0 or not) and whether the taxpayer has total losses carried forward. 8. Some special comments are necessary about the carry forward o f unused income-limited deductions and special deductions where any losses generated and carried forward are traced as a special pool o f losses: a. In the case o f income-limited losses, the actual amount o f losses used in the current year from current deductions and unused carry forwards gives the correct TE estimate. It i s automatically the sum o f D+ L C F - LCFnext . b. In the case o f traced deductions, the sum o f the current deduction taken (D) plus the amount o f any L C F o f that type used in the year has to corrected for any changes in the stock o f unused income-limited deductions over the current year to be an accurate estimate. Alternatively, the methodology in table 2 based o n the current deduction and changes in the stock o f total L C F overhe current year can be used. 9. A r e these concerns about loss carry forward quantitatively important or not in tax expenditure estimates. Table 3 provides some clues. I t i s based on the summary tax return data readily at hand for 1998, 1999 and 2003 for companies. I t shows that tax losses are extremely high, even excluding the impacts o f investment allowances, which are only taken when there is available taxable income and do not generate explicit loss carry forwards. The following i s evident: a. Loses in a year vary between 2 1% and 54% o f taxable profits b. Combined current year losses and losses carried forward vary between 83% and 99% o f losses c. Loss carry forwards are absorbed relatively slowly. This can be found by comparing the new losses generated in 1998 and 1999 with the losses carried into year 2003 from these years. These indicate that these stocks o f loss carry forwards have declined at the rate o f 23% per year through use or companies folding. This means that at least 26% o f the 248  original losses are lost because o f the five-year carry forward restriction. Overall there is a relatively low probability (less than 23%) that a loss generated in one year will be used in the next. d. Losses absorbed in a year reduced between 10% and 33% o f taxable income. e. In addition, exemptions that are taken only when there is a profit (essential investment allowances) have eliminated another 27% to 35% o f taxable income in the year. Unfortunately, no information i s available on the investment allowances created by vintage year and the outstanding stocks o f these vintages in order to find out h o w rapidly they are absorbed. The large size o f the investment allowances claimed each year (a similar magnitude to the new losses) would suggest that investment allowance carry forwards are more rapidly absorbed. This coupled with their indexation and indefinite carry forward makes the analysis o f the stocks o f unused investment allowances a source o f serious concerns to both tax revenue and tax expenditure analysis. 10. Next which exemptions or deductions that generate losses are important. From Table 1, the major item o f deductions i s “other exemptions and deductions” (line 38) generating a direct tax expenditure o f about YTL 2.1 billion in 2003. Clearly i t would be useful to identify what are the major sources o f this large amount o f exemptions and deductions?? Aside from this item, a number o f other items amongst the exemptions generating losses have other issues as noted in Table 1. The treatment o f investment income requires tax integration considerations. Two other items may be subject to withholding taxes that affects the TE estimates. Companies in free trade zones have perpetual tax holidays so that loss carry forwards are o f no value until and unless their licenses expire. That leaves the exemptions for companies in less developed regions, technology regions and the large “other” group. To get an upper bound on the potential values, following table 2, the sample o f companies in 2003, if divided into those with and without these exemptions, and then into the four cells in table 2 assuming the case o f Dpfior > LCF, . An accurate estimate would require knowing the exemption claims (Dpnor ) o f these companies currently with L C F >O in 2003 over the previous 5 years. The L C F i t s e l f would be the upper bound on the amount o f the exemption carried forward into 2003 . Studies o f time series f tax returns i s required to find Dpfior and estimate usefl probabilities such as the probability o f also claiming the same type o f deduction in prior years and also being not-taxpaying and generating LCF. 11. Finally, based on 8 above, the tax expenditure for investment allowances i s the dominant item and clearly very important. Table 1 shows a tax expenditure o f YTL 3 billion in 2003 The amount o f unused investment allowances, which are carried forward indefinitely with indexation is unknown. If the amounts o f deductions, losses and loss carry forwards declines in the future these investment allowances could grow in the hture as the available taxable income grows, even if the investment allowance were limited or used less. It i s important to get estimates o f the amount o f unused investment allowances by companies broken out into the cross tabulations o f table 2. 249 a  ANNEX VI1 Classificationo f Budget Institutions Chart I :General Budget Institutions 1. Presidency o f Republic 2. Turkish Grand National Assemblv 3. Constitutional Court 4. _. Court o f Atmeals 5. Court o f State 6. Turkish Court o f Accounts 7. PrimeMinistrv 8. State Planning Organization 9. Treasury Undersecretariat 10. Foreign Trade Undersecretariat Y 11. Undersecretariat o f Customs 12. State Institute o f Statistics 13. Ministrv o f Religious Affairs 14. Ministry o f Justice 15. Ministry o f National Defense 16. Ministrv o f Interior 17. General Directorate o f Security 18. General Commandership o f Gendarmerie 19. Commandership o f Coastal Security 20. Ministry o f Foreign Affairs 2 1. Ministrv o f Finance 22. Ministry o f National Education 23. Ministry o f Public Works and Settlement 24. General Directorate o f L a n d Registrv and Cadastre 27. Undersecretariat o f Marine Affairs t 30. 3 1. Ministry o f Industry and Trade Ministry o f Energy and Natural Resources 32. Ministrv o f Culture and Tourism 33. Ministry o f Environment and Forestry 34. General Directorate o f State Meteorological Affairs 25 1 252  ANNEX VI11 BASELINE INDICATOR SYSTEM (BIS) FOR PUBLIC PROCUREMENT JOINT ASSESSMENT REPORT FOR TURKEY I n January 2005, the World Bank proposed to Public Procurement Authority (PPA) of Turkey to undertake a joint study on Baseline Indicators System (BIS) for the assessment of the structure of national public procurement system in Turkey. Then, the Ministry of Finance and PPA agreed to implement this system in Turkey to analyze the health of existing public procurement system. The Bank mission visited the PPA in April 2005 to discuss the applicable indicators for Turkey and methodology to be used in this assessment. All parties accepted that a completed rating system would provide a clear and comprehensive picture of the current status of the public procurement system and would also provide a very valuable tool for identifying and prioritizing key areas for further development. The indicators and the format of the assessment report were discussed and agreed by the Bank and the PPA in April 2005. The assessment was conducted on the basis of four (4) key areas called “pillar” which are composed of twelve (12) indicators (baselines). These twelve (12) indicators are composed of 80 sub-indicators that are considered as desirable standards against which Turkey’s public procurementsystem i s assessed. I t was agreed that all indicatorshave the same weight and within each indicator, all sub- indicators have the same weight. F o r indicators with multiple sub-indicators; a “yes” o r “no” answer will be provided for each sub-indicator. I f over 90% o f the sub-indicators are met, then the assessment for that indicator will be “fully achieved”. The agreed standards for the assessment would be as follows: Over 90% = baseline i s fully achieved (FA) Between 70-90% = baseline i s substantially achieved (SA) Under 70% = baseline i s not achieved (NA) The International Relations and Coordination with EU Department o f PPA i s assigned to work with the Bank team in this assessment. The team established within the PPA with the help of other related agencies working in the field of public procurement conducted the assessment and submitted the preliminary assessment report to the Bank by the end of May 2005. Each sub-indicator was rated against the scale of “fully achieved”, “substantially achieved” or “not achieved” by the PPA. The summary o f ratings o f the self-assessment by the PPA i s as follows: Indicator 1 : Public Procurement legislative and regulatory framework achieves the agreed standards and complies with applicable obligations = FA Indicator 2: Existence of Implementing Regulations and Documentation = FA Indicator 3: MainstreamingProcedures into Public Financial Management = SA Indicator 4: Functional ManagementmormativeBody = SA Indicator 5: Existence of Institutional Development Capacity = SA Indicator 6: Efficient Procurement Operations Capacity and Practice = SA Indicator 7 : Functionality of the Public Procurement Market = SA 253 Indicator 8: Existence o f Contract Administration and Dispute ResolutionProvisions = SA Indicator 9: Effective Control and Audit System = SA Indicator 10: Efficiency o f Appeal Mechanisms = FA Indicator 11: Degree o f Access to Information = FA Indicator 12: Ethics and Anti-corruption Measures = FA The Bank mission visited PPA to discuss the initial report in June 2005. During the mission, indicator by indicator review and discussions were held down to detail o f each sub-indicatorwith the PPA and the Bank. Following detailed discussion o f ratings; the actions that might be taken by the PPA (or other relevant organizations) t o improve the rating that was indicated for substantially achieved or not achieved sub-indicators are identified. Even o n indicators rated as fully achieved by the PPA; the discussions enabled the Bank and the PPA t o identify implementation issues that are affecting in achieving the full benefit o f law and regulations and to look at possible actions to improve performance. The sub-indicators, where the Bank has comments, need to be addressed further by the PPA in order to be more compatible with international standards and good practices. Therefore, PPA will take actions by giving priority to the “mandatory sub-indicators”. I n February 2006, the draft final report was updated together with the PPA and the Bank based on the progress in the last 8 months and recent developments are incorporated into the assessment table. 254 - c e 5 0 10 10 N  \o m N I ,  v, N 1 m m N  0 3 \o 2 N b  N \o N m c d 0 z 4 4 8 $ 8 Z0 lZ0 3 B t d if 3