Report No. 36764-TR

Turkey Public Expenditure Review December 18, 2006 Poverty Reduction and Economic Management Unit Europe and Central Asia Region

Document of the World Bank

CURRENCY AND EQUIVALENTS UNITS

Currency Unit = TL (Turkish Lira) - until December 31, 2004 = YTL (Yeni Turkish Lira) - from January 1, 2005

1 YTL = 1,000,000 TL 1 US Dollar ($) = 1.574 YTL, using exchange rate on June 30, 2006

ABBREVATIONS AND ACRONYMS BITT Banking and Insurance Transaction Tax BYEs Budget Management Information Systems (eBudget) CGG Consolidated General Government CIT Corporate Income Tax DFIF Support Price-Stabilization Fund DFIF Support Price Stabilization Fund DSI State Hydraulic Agency EBFs Extra Budgetary Funds GDBFC General Directorate of Budget and Fiscal Control GDRS General Directorate of Rural Services GFS Government Financial Statistics HPC High Planning Council IACB Internal Audit Coordination Board IPARD Instrument for Pre-Accession for Rural Development MOF Ministry of Finance MOH Ministry of Health MONE Ministry of National Education MTFS Medium Term Fiscal Strategy NEA National Education Accounts PAF Public Administration Framework PFMC Public Financial Management and Control PISA Program of International Student Assessment PIT Personal Income Tax PPA Public Procurement Authority PPL Public Procurement Law PPP Public Private Partnership RUSF Resource Utilization Support Fund SAI Supreme Audit Institution SCT Special Consumption Tax SDBs Standard Bidding Documents SDIF Savings Deposit Insurance Fund SPA Special Provincial Administration SPO State Planning Organization SSEs State Economic Enterprises SSI Social Security Institution TCA Turkish Court of Accounts UHI Universal Health Insurance VAT Value Added Tax VEDP Automated Tax Management System YOK Higher Education Council

FISCAL YEAR 2007 July 1 – June 30

Vice President: Shigeo Katsu Country Director: Ulrich Zachau Sector Director / Manager Cheryl Gray / Felipe Jaramillo Task Team Leader Aristomene Varoudakis

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 of the Treasury, is gratefully acknowledged, and especially from the services of State Planning Organization and the Ministry of Finance, which provided information on many parts of the study.

i

Table of Contents

EXECUTIVE SUMMARY ........................................................................ VI

OVERVIEW................................................................................................ IX

CHAPTER I SOURCES OF FISCAL CONSOLIDATION AND THE QUALITY OF FISCAL ADJUSTMENT................................................. 30 A. PATTERNS OF FISCAL ADJUSTMENT IN TURKEY................................................................. 30 B. THE SIZE OF GOVERNMENT AND COMPOSITION OF EXPENDITURES IN

INTERNATIONAL COMPARISON................................................................................................ 36 B.1 Public expenditure by economic category ......................................................................................... 37 B.2 Public Expenditure by Government function..................................................................................... 40 C. THE REVENUE SIDE ...................................................................................................................... 43 D. STRUCTURAL AND CYCLICAL COMPONENTS OF THE BUDGET BALANCE ................... 50

CHAPTER II OPTIONS FOR EFFICIENCY GAINS IN PUBLIC EXPENDITURES—HORIZONTAL ISSUES......................................... 60 A. PUBLIC SECTOR EMPLOYMENT AND THE WAGE BILL ....................................................... 60 B. OBJECTIVES AND REFORM OPTIONS IN THE 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 III OPTIONS FOR EFFICIENCY GAINS IN PUBLICEXPENDITURES—SECTOR ISSUES .................................... 94 A. PUBLIC EXPENDITURE IN THE EDUCATION SECTOR .......................................................... 94 A1. Educational Outcomes ....................................................................................................................... 95 A2. Structure of Education Expenditures in Turkey................................................................................. 98 A3. Key reform directions ...................................................................................................................... 115 B. THE PENSION SYSTEM AND ITS REFORM ............................................................................. 117 B1. Evolution of 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 C1. Health outcomes .............................................................................................................................. 127 C2. Main characteristics of the health sector.......................................................................................... 130 C3. Public expenditures on health .......................................................................................................... 135 C4. Short-and medium-term measures to improve efficiency and enhance equity of 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

ii

CHAPTER IV MANAGING MEDIUM TERM EXPENDITURE PRESSURES ............................................................................................. 153 A. THE GOVERNMET’S MEDIUM-TERM FISCAL STRATEGY ................................................. 154 B. EXPENDITURE PRESSURES IN KEY SECTORS AND MEDIUM-TERM CHALLENGES.... 161 B1. 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 EXPENDITURE PRESSURES...... 171 C.1. A sustainable medium-term expenditure envelope .......................................................................... 172 C.2. An assessment of required fiscal space............................................................................................ 174

CHAPTER V STRENGTHENING BUDGETARY INSTITUTIONS FOR EFFECTIVE GOVERNMENT ..................................................... 180 A. STATUS OF PUBLIC EXPENDITURE MANAGEMENT REFORMS ....................................... 182 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 of Expenditures Annex II Classification of Central Government Expenditures across........................................... 216

Functional and Economic Categories, 2004-2008 Annex III Estimating Structural and Cyclical Budget Balances..................................................... 222 Annex IV Tax performance: A Regression Model ......................................................................... 236 Annex V Draft Tax Benchmarks for Tax Expenditures ............................................................... 240 Annex VI Towards Estimates of Tax Expenditures in PIT and CIT ............................................. 243 Annex VII Classification of Budget Institutions.............................................................................. 251 Annex VIII Baseline Indicator System (BIS) for Public Procurement.............................................. 253

iii

Tables Table 1.1 Institutional Breakdown of 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 IECD Countries 46 Table 1.4 Outut Elasticity of Taxes 51 Table 1.5 Fiscal Stance and the Output Gap 55 Table 1.6 Output Elasticity of 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 of 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 of Public Investments 75 Table 2.5 Public Sector Investment, 1999-2006 76 Table 2.6 Annual Investment Spending (% of GNP) 76 Table 2.7 Effect of Economic returns from 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 on educational institutions as a percentage of GDP from Public and private

sources by source of fund 110 Table 3.5 Per Student Education Expenditure by Education Level 111 Table 3.6 Comparison of 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 from various sources 128 Table 3.9 In128nt Mortality and Under-five Mortality Rates by area of residence and region 128 Table 3.10 Co-payments for health care services and pharmaceuticals 131 Table 3.11 Number of insured individuals by type of health insurance coverage 2002 132 Table 3.12 No of hospitals and hospital beds by type of institution 133 Table 3.13 Number of 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 on 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 of Transfer Budget 150 Table 3.22 Consolidated Budget Expenditures Including Recurrent Costs 151 Table 4.1 Institutional breakdown of the primary surplus 154 Table 4.2 Central Government Budget Revenues in the MTFS 157 Table 4.3 Adjusted Primary expenditures of 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 of Additional Expenditures Needed for Financing New 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 on Health 169 Table 4.10 Estimated Revenue and Expenditure Savings from Co-payments 170 Table 4.11 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

iv

Table 4.15 Expenditure Pressures in Key Areas 175 Table 4.16 Medium term 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 of Recent Laws to Strengthen Public Financial Management 181 Table 5.2 Importance of 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 of the General Government 31 Figure 1.2 General Government Expenditure as % of GDP 37 Figure 1.3 Compensation of General Government employees 38 Figure 1.4 General Government Wage Bill in percent of 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 of General Government Expenditures 41 Figure 1.9 General Government Revenue in % of 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 21 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 of GDP 61 Figure 2.2 Composition of Wage Bill by Function 62 Figure 2.3 Employment in 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 of total personnel

expenditures of consolidated budget agencies 66 Figure 2.7 Gross Capital Formation by Function 74 Figure 2.8 Maintenance of Expenditures (% of NP) 80 Figure 3.1 Percent of 20-24 year olds with a secondary education diploma 96 Figure 3.2 Distribution of 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 of Education , 2002 106 Figure 3.6 Out of 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 of Spending on educational Inputs in Turkey in 2002 109 Figure 3.11 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 of the Population over Age 65 in Turkey and Selected OECD Countries 118 Figure 3.17 Pension System Deficits Before and After Passage of the 2006 Law 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

v

Figure 3.21 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 of recurrent public expenditures on health 138 Figure 4.1 Unadjusted and Unadjusted SSIs deficit 159 Figure 4.2 (a-d) Total Education Expenditure Framework Without and With New Education Policies 166 Figure 5.1 Schematic Diagram of the Internal Financial Control Framework for the general Budget

Institutions as Defined by the 2003 PFMC Law 196 Boxes Box 1.1 Conversion between SPO definition of CGG primary surplus to IMF’s program

definition of public sector primary surplus 32 Box 2.1 Guidelines to Identify Normative Tax Structure 86 Box 2.2 Tax expenditure by type of tax 87 Box 3.1 Principal sources of financing for the education sector 101 Box 3.2 Expenditure Procedures at the Provincial Level 102 Box 3.3 The system of financial protection for health expenditures 130 Box 3.4 The role of the Revolving Funds in the Turkish Health Care System 136 Box 3.5 Examples from Rural Agencies, Issue of High Personnel Expenditures 151 Box 4.1 2007 Budget and Medium-Term Fiscal Strategy 155 Box 4.2 The Methodological changes in the Reporting of Health Expenditures and Social Security

Transfers 159 Box 4.3 Data sources for the medium-term education expenditure framework 164 Box 4.4 Projecting the budgetary impact of new education policies and proposals 164 Box 5.1 Towards Full Implementation of PFMC 185 Box 5.2 Amendments to the PFMC Law 187

vi

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 on the revenue side so far—has been the cornerstone of 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 of recent global market volatility on Turkey. In the current environment, maintaining continuous fiscal discipline, with a high primary fiscal surplus, will be necessary in order to stem pressures on the current account, help keep disinflation on track, and achieve a reduction of 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 of medium-term growth. As strong fiscal discipline needs to be maintained in the years ahead, there is little room for Turkey to further increase total expenditure in order to meet pressing development challenges. Structural fiscal reforms, aimed at improving the quality of fiscal consolidation, are the only viable means of 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.

• Sector-specific reforms to improve the efficiency of expenditure programs would be required in areas where expenditure pressures are being felt;

• Trade-offs in expenditure allocations would need to be considered, with the aim of shifting resources to priority programs for growth and social development;

• Sector-specific expenditure reforms should be underpinned by horizontal reforms that improve the efficiency of expenditure programs across sectors;

• Initiatives to further 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

vii

account the expenditure pressures in key sectors such as health and education, the fiscal space needed to meet these expenditures is estimated at 1.1 percent of 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 of 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 of non personnel expenditures to levels similar to private schools and universities; improving the regional balance of 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 of 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 of poor health outcomes and relatively high expenditures suggests that important room exists for improving the efficiency of the health care system. The introduction of Universal Health Insurance (UHI) will improve access and equity but, to remain fiscally viable, it should go in tandem with cost savings and marked improvements in efficiency of public spending. The principal measure to control costs and dissuade indiscriminate use of health care services is the introduction of copayments for outpatient visits and for drugs. But cost containment measures should go in tandem with more comprehensive structural reform—such as, full expansion of family medicine; reforming the patient referral system; introduction of payment mechanisms for providers which will discourage unnecessary use of physician and diagnostic services; consolidation of 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 of a previously fragmented system, but future challenges should be kept in sight. With adequate implementation of the reform, the pension system deficit is expected to be absorbed by 2040, with a fiscal gain of 7 percent of GDP compared to a “no reform scenario”. However, long-run financial sustainability will be secured under a high level of payroll taxes by international standards—a key impediment to job creation in the formal sector. The results of the 2006 reform should be closely monitored, and if lowering contribution rates is 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 of expenditure programs across sectors. Horizontal reforms should include initiatives aimed at strengthening incentives

viii

in the civil service towards the achievement of results, while containing pressures on the wage bill. Adoption of efficient project evaluation methods, effective multi-year programming, and adequate provisioning for long-term operation and maintenance of public capital will be key to improving the quality of the investment program across functional expenditure areas. A framework conducive to a greater role of the private sector in the financing, development, and operation of infrastructure should be created—requiring a predictable policy and regulatory environment, together with careful design of Government commitments to private operators to minimize the risk of 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 of GDP in 2003, tax expenditures in Turkey were close to the high end of 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 of 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 of 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 modern public expenditure management and accountability. The main challenge ahead is the implementation of 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 of the internal audit structures; (ii) enactment of the Turkish Court of 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 of the overall Public Financial Management reform agenda.

ix

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 of the economic program, with the public sector primary balance moving to a surplus of 6.9 percent of GDP in 2005, from a deficit of 1.6 percent in 1999. Gross public debt was thus reduced from 106 percent of GDP in 2001 to 71 percent in 2005. Resolute fiscal consolidation relieved the conduct of monetary policy from fiscal dominance, thus enhancing the credibility of the commitment to low inflation and facilitating the implementation of 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 oil prices, have all resulted in a widening deficit of the external current account that reached 6.4 percent of GDP in 2005. As a result of 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 of 2005, will likely be further prejudiced in the short term by the pass-through of depreciation of 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 of 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 to 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 is little room for Turkey to further increase total expenditure in order to meet pressing development challenges. Structural fiscal reform, aimed at improving the quality of fiscal consolidation, is the only viable means of sustaining the adjustment in the future while making appropriate fiscal space for growth-enhancing expenditures and lower taxes. The focus of efforts should now shift to the expenditure side in critical sectors, with the aim of 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 of 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

x

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 of fiscal 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 of the Consolidated General Government increased by 5.7 percentage points of GDP. This improvement was largely driven by primary fiscal revenues, which increased by 6.1 percentage points of 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 of low-productivity expenditures and improve priority setting, which have supported a better control of 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 of 2.1 percent of GDP in 1999, which was turned into a surplus of 0.15 percent of GDP in 2005. These SOEs thus added the equivalent of almost 3 percentage points of GDP to fiscal consolidation of the public sector as a whole. This improvement was generated by a close to 2 percentage points of GDP reduction in their investment expenditures, mainly for infrastructure, and a cut in personnel compensation equivalent to 1 percent of GDP—as a result of a 45 percent downsizing of 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 of the social security institutions, which—despite Turkey’s favorable demographics—reached 4.8 percent of GDP in 2005, from 1.9 percent of GDP in 2000. In the absence of the social security reform enacted in June 2006, the projected deficit of the system would have further increased to reach 7 percent of GDP in the long term. 7. Most of 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 of the consolidated general government (excluding SOEs) increased by 2 percentage points of GDP. Over this period, buoyant fiscal revenues, due to robust growth, contributed an estimated 3.4 percentage points of GDP to fiscal consolidation—more than the improvement in the primary fiscal balance. Thus, since 2002 the structural primary fiscal surplus—a better measure of policy-induced changes in the fiscal stance—has actually declined (by an estimated 1.2 percent of potential GDP).

xi

8. The decline in the structural primary surplus is of concern both for the sustainability of fiscal adjustment and the management of 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 of 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 of 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 of 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 of GDP in 2005, primary expenditures of the general government were lower than in European Union (EU) member countries, but about 10 percentage points higher than in a comparison group of emerging economies. The difference mainly reflects a larger general government wage bill, which, at 9 percent of 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 of comparator countries by the equivalent of 0.8 and 1.2 percentage points of GDP respectively. By contrast, expenditure on environmental protection is comparatively low and would need to increase on the way to the EU. Expenditure on health care is relatively high in international comparison, while outcomes are below the average of comparator countries. Similarly, expenditure on social protection is high by international comparison in view of Turkey’s favorable demographics. Expenditure on education is in line with comparator countries but educational attainment remains weak. Horizontal public expenditure reforms 11. Sector-specific expenditure reforms will provide a better pay-off when combined with “horizontal” reforms that improve the efficiency of expenditure programs across sectors. Horizontal reforms should include initiatives aimed at strengthening incentives in the civil service towards the achievement of results, while containing pressures on the wage bill. A larger share of expenditures could thus be spent on materials and equipment necessary for the provision of high-quality public services. Adoption of efficient project

xii

evaluation methods, effective multi-year programming, and adequate provisioning for long-term operation and maintenance of public capital will also help improve the quality of 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 of GDP, before declining to 9 percent of GDP in 2005. However, the wage bill could increase again as a result of the additional pay rise for civil servants granted for 2006, the estimated cost of which is 0.35 percent of GDP. While the wage bill has remained rigid, employment in consolidated budget agencies and municipalities has declined by 3 percent since 1999, resulting in fast growth of average compensation. 13. The share of personnel expenditures on economic affairs and security is particularly high in international comparison. Personnel expenditures are concentrated on education, security, and health care, which absorb more than two-thirds of the wage bill. The share of personnel compensation in economic affairs—for the provision of infrastructure services, rural development, and services to agriculture—is twice as high as the average of 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 of average compensation to per capita GDP remained constant at its average level in 1999-2000, the wage bill of consolidated budget agencies would have been reduced by about 1 percentage point of GDP. Linking wage increases to inflation has not proven enough to contain growth in average compensation, probably because other forms of 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 of personnel compensation is composed of various allowances, compensations, and rewards. Many of these side benefits are exempt from the income tax, with estimated foregone tax revenue of around 0.5 percent of GDP. The system of 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:

• improving efficiency—by making it easier to adjust staffing composition when required; enhancing sources of performance motivation; facilitating opportunities for human capital improvement of public servants;

• 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;

xiii

• making it easier to ensure a fiscally sound wage bill—with the consolidation of legal frameworks expected to facilitate monitoring of employment and pay policy changes;

• 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 of civil 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 of compensation harmonization by setting a new structure of basic salary and wage scale index; and (vi) monitoring the reliability of recruitment and promotion procedures. 18. Public investment was cut significantly as part of the fiscal adjustment efforts during 2001-2004 at a time when rationalization of the public investment portfolio was overdue. Total public sector investment fell to 4.2 percent of GNP in 2004, from 6.8 percent in 2000, before recovering to 5.1 percent in 2005, and the brunt of adjustment was borne by investment in infrastructure. In the past, unclear criteria and processing rules had resulted in “over programming” of the public investment portfolio (PIP), with the average project completion time as high as 12 years in 2001. Rationalization of the PIP has been quite effective in reducing the number of projects and the average completion time to 6.7 years by 2005. However, as a result of the downsizing of the PIP, the number of projects put on hold has significantly increased, with “trace allocation” projects accounting for about 10 percent of the total value of the PIP in 2003-04, a significant rise from the 1999-2000 average of only 3 percent. 19. Meeting financing requirements of infrastructure is a key challenge for sustained growth. In order to emulate EU standards of infrastructure service provision, Turkey must systematically improve the quality of infrastructure investments in the medium term. The quantity of 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 of existing assets, (iii) ensuring adequate security and reliability of supply of infrastructure services, and (iv) ensuring that infrastructure investment addresses economic development objectives. More effective utilization of the recently introduced medium-term expenditure framework and enhancing the participation of 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 financing, development, and operation of infrastructure. In

xiv

view of the large financing requirements in infrastructure, achieving the Government’s objectives of 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 of Government commitments to private operators to minimize the risk of contingent liabilities. In the energy sector private investment now makes up about 50 percent of electricity generation, but this has been made possible only by significant contingent liabilities on the Government in the form of guarantees and off-take agreements. Letting the regulatory framework operate independently, and allowing prices to fully reflect costs would make it possible to attract private investment without significant guarantees. 21. The quality of 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 on a project-level evaluation and revising the rationalization policies and criteria, while establishing a proper prioritization of 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 of projects and improved incentives and pricing/subsidy policies are necessary for enhancing efficiency. 22. Appropriate allocations for operation and maintenance of public capital need to be ensured. This is a requirement not only for improving the quality of services, but also for restricting the cost of the projects as accelerated decay of capital goods may unduly burden the investment program in the future. Appropriations for maintenance declined from 0.4 percent of 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 of 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 of roughly 2.5 percent of 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 of investments that can be made is constrained not so much by the availability of financing but by the need to meet the primary balance targets in the face of 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 of infrastructure services will also help alleviate existing constraints.

xv

Structural expenditure reforms in key sectors 24. Public education expenditures have been on a downward trend since the 1999 crisis. At approximately 4 percent of GNP on average over the past decade, public education expenditures have reached an estimated 4.17 percent of GNP in the 2006 budget—less than the 4.33 percent of GNP spent in 2002. Including private financing, Turkey spends on education the equivalent of 7 percent of GDP. There is significant private financing of courses for high school and college entrance exams, and private financing also significantly contributes to the operation of public schools as a consequence of 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 of education. Turkey’s per student spending on pre-primary schools is 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 of its per capita income. 26. Even though per student expenditures in secondary education are comparatively high, student achievement is 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 of the relatively high per student expenditures in secondary education. Although student achievement, as measured by PISA indicators, is expected to increase with per capita spending, it would be counterproductive to further increase expenditure without, at the same time, taking steps to ensure that the impact of spending has improved. 27. The differences in educational expenditures at public and private schools are significant. Per-student spending in public primary schools is almost two-thirds lower than in private primary schools, while per-student expenditure in public general secondary schools is almost half the spending at private schools offering the same education. Public schools, regardless of the level of education provided, spend more on personnel relative to private schools, while use of 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:

xvi

• improving the quality of educational achievement in public schools would require increasing the share of non personnel expenditures to levels similar to private schools and universities;

• improving the regional balance of educational services across the country would require enhanced public education service provision in most backward provinces to compensate for possible lack of private sector initiative—while taking advantage of opportunities for public school retrenchment where private sector initiative is strong;

• increasing per student expenditure on pre primary education—an area where progress is critical to encourage greater participation of 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 of spending by reducing the number of teachers employed for the same level of enrollment. Second, schools need to be given autonomy over their resources and held accountable for their results. A system of per-student financing could be considered, with funding allocated in proportion to the number of students a school enrolls, up to its capacity. Schools would be responsible for using these funds 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 of the year. Within a financially autonomous system, the amount of funding allocated to each university should not be based on inputs, but on performance, in terms of 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 of higher education institutions. In a complementary way, the university entrance examination system should be modernized in such a way as to eliminate the advantage of the widespread private examination preparation courses. 31. The Turkish pension system has been a perennial drain on public finances as it continues to run large deficits. Turkey is a young country and should not be experiencing pension system deficits of the current magnitude. However, because of very low retirement ages, high benefit accrual rates, and other disincentives to remain in the formal labor force, pension deficits have persisted, with the exception of a temporary improvement in 1999, when parameters improved with the passage of a pension reform law. 32. The 2006 reform will bring fiscal balance of the pension system in the long term as well as the unification of a previously fragmented system. With adequate implementation of the reform, the pension system deficit is expected to decline by 1

xvii

percent of GDP below a “no reform” projection by 2020. The deficit will be absorbed by 2040, with a fiscal gain of 7 percent of GDP compared to a “no reform scenario”. However, the beneficial impact of 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 of payroll taxes by international standards. High contribution rates hamper job 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 of 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 of benefits per year of 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 of the labor force—a key requirement for a better fiscal performance of 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 of a partial pension with only 15 years of contributions. 34. The combination of poor health outcomes and relatively high expenditures suggests that important room exists for improving the efficiency of the health care system. Turkey spent approximately 6.6 percent of its GDP on health in 2004—a higher level than comparator countries. Public health care expenditures accounted for 68 percent of total, up from 61 percent in 1999. Although the health status of the population has considerably improved, health outcomes still compare unfavorably with countries that have similar or lower level of GDP per capita and approximately the same share of 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 of the country. High levels of infant and under-five mortality, as well as high maternal mortality, are related to the poor usage of preventive care. 35. Public expenditures on health care are on a strong upward drift. Public expenditures have exceeded 5 percent of GNP in 2004 and 2005, up from 3.3 percent in 1999. The increase in health expenditures in recent years is almost entirely explained by fast growing spending on hospitals and pharmaceuticals, while the share of expenditures on preventive services has actually decreased (from 11 percent in 1999 to 7 percent in 2005). The increase in the quantity of medical services partly reflects expanded coverage and access—even before the official enactment of Universal Health Insurance (UHI). The change of incentives facing medical providers in Ministry of Health hospitals and

xviii

outsourcing of services may have led to supplier-induced demand and an increase in production, provision and utilization of health services. 36. Efficiency and equity should be simultaneous objectives of health care reform. The introduction of UHI will improve access and equity but will undoubtedly generate greater demand for public expenditures on health. To remain fiscally viable, the introduction of UHI should go in tandem with marked improvements in efficiency of public spending. Turkey should also consider devoting a higher share of public expenditures to preventive activities, while targeting resources where they are most needed, to address the persistent variation of health indicators between urban and rural areas and among regions. 37. Consolidation of hospitals and greater management autonomy hold the promise of efficiency gains. Many MOH hospitals are too small in size to allow for efficient operation and provision of care, and have significantly lower utilization rates compared to University hospitals. The introduction of UHI provides a good opportunity to further strengthen the gains from the merger of MOH and SSK hospitals under MOH ownership and management. Further gains will come about by improving efficiency in the use of hospital resources and overall management and accountability, which would be facilitated by granting financial and administrative autonomy to public hospitals. 38. The challenge for cost--containing measures is to manage quantity—especially to contain pharmaceutical expenditures—without adversely affecting access, utilization and effectiveness of services. The principal measure suggested to control quantity and dissuade indiscriminate use is the introduction of copayments for outpatient visits and for drugs. Even nominal levels of copayments would help rationalize the consumption of health services and result in savings, not only in the production of health services but also in the consumption of 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 on the basis of capitation fee per enrollee, so as to encourage conservative use of 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 of 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 of rural public expenditures would be needed to improve their efficiency. At 1.2 per cent of GDP, the level of public rural investments is low compared to the size of the agricultural sector and its modernization needs—although rural investments have recently been on 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 on safety and phytosanitary standards—with particular focus on small farmers with weak links to markets—and away from irrigation projects. Agricultural

xix

transfers have been significantly reduced and restructured by switching the main focus of agricultural transfers to the Direct Income Support (DIS) program, with reduced artificial incentives for inputs and particular crops. As recognized by the Government, the DIS Program needs to be accompanied by transfer payments with the aim of promoting productivity and market development more directly. But with more than a third of rural expenditures on salaries of government employees, there is little room in the budget for investments as well as other operation and maintenance expenses. With greater emphasis on participatory approaches and private sector involvement in service provision, more effective and efficient use of 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 of 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 of 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 of 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 of a sector and thus facilitate the realization of 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 and progress in tax reform 42. The composition of 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 of GDP in 1999 to 16.7 percent of GDP in 2005. During the same period, personal income tax revenue fell from 6.5 percent of GDP in 2001 to only 4.9 percent GDP in 2005, while corporate income tax revenue remained broadly stable. As a result of the reliance on revenue-increasing measures, tax revenues in Turkey (including social security contributions) reached nearly 32 percent of GDP in 2005, with an estimated tax effort of about 30 percent higher than expected in international comparison. Primary fiscal revenues stand at 40 percent of 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 of ensuring some degree of compliance of those who would have otherwise escaped taxes altogether. In addition, indirect taxes uniformly

xx

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 of special regimes and exemptions. In particular, reduced VAT rates introduced recently for textile and clothing products are of questionable effectiveness as a means of assisting domestic producers to withstand competition and tackle the problem of fraudulent refunds. Multiple VAT rates and exemptions increase the complexity of the tax system, create more opportunities for misreporting and misclassification for tax purposes, and thus work against efforts of 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 of the Corporate and Personal Income Taxes have been initiated, with a reduction in the CIT rate and tax base broadening, and a simplification of the rate structure of the PIT, together with the harmonization of the taxation of wage and non-wage income.

• 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 on married workers; maintaining neutrality in the taxation of interest income across asset classes and avoiding discrimination among domestic and foreign investors.

• 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 of dividends.

45. Establishing an adequate tax expenditure framework would promote fiscal accountability and transparency. Aware of 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 of the major tax expenditures are granted with the aim of promoting investment, employment, regional development, R&D, and trade incentives (primarily through the CIT). A range of tax expenditures aims at assisting low-income tax payers, in support of social policy. These are delivered primarily through the PIT and CIT and, to a lesser extent, the VAT. Some major tax expenditures in this latter group include the deductibility (for PIT purposes) of 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 of tax expenditures is still in its early stages but preliminary estimations point to a significant fiscal cost. As part of the PFMC Law, the authorities have estimated that tax expenditures for 15 tax provisions will cost 1.6 percent of GDP per year over 2006-08—including, however, the investment tax allowance, since phased out. According to very approximate estimations, including part of omitted tax

xxi

expenditures would raise total tax expenditures to at least 5 percent of GDP per year, or about 18 percent of government revenues in 2003. This level of tax expenditures would place Turkey towards the high end of 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 of 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 amplifying expenditure needs.

• Several new policies have been initiated or proposed in education that will have a significant impact on the medium-term expenditure program. Simulated new or proposed policies would increasingly boost annual expenditures, by up to 0.5 percent of GDP in 2009.

• Health expenditures are projected to increase from 5.2 percent of 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 of GNP during 2007-2009.

48. The social security system deficit is set to decline, generating some modest fiscal space. With the enactment of the new social security laws, revenues of the system are projected to increase by 0.3 percent of GDP until 2009, while expenditures would decline by 0.2 percent of GDP. As a result, the social security system deficit is estimated to decrease by half a percentage point of GDP until 2009—with further savings expected in the long term. 49. A sustainable budget envelope for the financing of primary expenditures at the general government level is estimated at around 33 percent of GDP until 2009. A sustainable expenditure envelope is 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 of 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 of public debt, Turkey’s debt ratio is relatively high and its maturity short. The recent turmoil in international markets and its 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 is calculated as the difference between the projected primary expenditures and the sustainable expenditure envelope. In the baseline projection, the needed fiscal space is estimated at 1.1 percent of GDP in 2007, reaching 1.7 percent in

xxii

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 of 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 of GDP. The required fiscal space would thus increase by an equivalent amount. With a stronger effort in education and more ambitious reduction of the debt ratio the fiscal space needed would increase from 2.4 percent of 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. 51. 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 of fiscal consolidation, are the only viable means of 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 of 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 of shifting resources to priority programs for growth and social development. Sector-specific expenditure reforms should be underpinned by horizontal reforms that improve the efficiency of expenditure programs across sectors. Initiatives to further rationalize the tax system—including the tax expenditures—would also create fiscal space by broadening the tax bases. Strengthening budgetary institutions 52. Far-reaching public financial management reforms have improved budget coverage, formulation, execution, accounting, audit, and procurement. Enactment of 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 of 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 of EU accession and by the government’s desire to introduce greater performance orientation in budget management. 53. The coverage of 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 of 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

xxiii

PFMC Law required incorporation of extrabudgetary funds into the budget of the related administration, they continue to operate outside of budget discipline, with insufficient scrutiny of their operations. Similarly, revolving funds continue to operate outside the budget coverage at the risk of distorting budget discipline, further to a decision to restructure the revolving funds by the end of 2007 with an amendment made to the PFMC in December 2005. The Government’s preferred approach is to restructure remaining revolving funds, rather than including them in the budget. 54. Properly phasing-in and completing the implementation of the accrual accounting system requires addressing complex issues. Key issues include valuation of assets and inventories, depreciation rates, and gains and losses from foreign exchange. As shown by the experience of countries that have switched to accrual accounting, implementation generally works best when done in phases. Extensive training is 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 of all commitments will ensure budget discipline and avoid the possible accumulation of arrears. 55. The PFMC Law has introduced a modern internal audit framework— The PFMC Law requires each public administration to establish an internal audit unit. The law clearly articulates the accountability of ministers and other high-level civil servants. It also requires every line ministry to prepare an annual accountability report—also including performance data and details in the implementation of the strategic plan and budget. 56. —but implementation of the law itself poses 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 of the size or capacity of the agencies concerned. This raises an implementation challenge, as smaller municipalities may either ignore the provisions that they may consider too expensive, or they may implement the provisions by focusing on 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 of Account’s (TCA) draft law creates a sound basis for the development of government audit and should be complemented by additional initiatives. In particular, the audit side of TCA is not yet fully developed as the draft law is still insufficient in terms of structures and responsibilities. Even after enactment of the law, TCA will still 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 of audit policies for local government. TCA’s role might be redefined in several ways—for example, with different strategies for local administrations of differing sizes; creating a new local government audit service; using private sector auditors to audit local administrations.

xxiv

58. Since 2001 the public procurement system has made good progress, yet some significant issues require attention. A new Public Procurement Law (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 of certain services from the scope of the law. Internal consistency and integrity of 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 of the alignment with the EU acquis—in particular because: (i) the 15 percent national preference for public contracts above the threshold is incompatible with the acquis; (ii) there is lack of sector (utilities) legislation; (iii) a horizontal legal framework for concessions is lacking; and (iv) a platform for Electronic Procurement is still 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 of the reform agenda has been hesitant, and some regressive tendencies in the implementation of the PFMC law tend to perpetuate budgety fragmentation. Although some aspects of the new legal framework are being piloted intermittently, some of the secondary regulations related to practical implementation are yet to be issued. The government’s decision to undertake fundamental changes in the responsibilities of provincial and municipal governments also injects new complexity and considerable uncertainty into the PFMC implementation schedule. Enactment of the Omnibus Law has created concern that its provisions run counter to the intention of the PFMD and PFMC Law as approved by the Parliament. Without addressing the unfinished elements of the agenda and arrangements to ensure effective leadership and coordination among agencies, implementation of the overall PFM reform agenda will remain at risk.

xxv

Table A: Horizontal Public Expenditure Reforms—key policy options Short-term options (one year)

Medium-term options (2-3 years)

Public sector employment and the wage bill

• Refrain from ad hoc pay rises for civil servants and ensure that the extra (40+40) pay rise granted in 2006 is not built in the salary base for calculating the 2007 pay increase

• Formulate a MTEF over 2007-2010 with the aim of restraining Consolidated General Government personnel compensation in percent of GDP

• Establish a consistent set of criteria to distinguish between public servants and contracted personnel

• Harmonize salary scales for civil servants across various career streams

• Design harmonization of salary scales with a view to minimizing its fiscal cost

• Concentrate salary in elements most closely linked to human capital requirements of positions

• Consolidate the legal framework governing salary setting throughout the public sector

• Reform the allowance and side benefit system for civil servants to incorporate all elements of compensation in the personal income tax base

• Consider replacing uniform step salary increases with unit-specific salary increases based on annual performance ratings and a fixed unit-specific budget envelope

• Design an annual personnel performance appraisal process that ensures reliable and fair performance ratings

• Continuously monitor that the integrity of competitive selection procedures and procedures governing changes in grade is ensured

Managing the public investment program

• Formulate a plan for a detailed assessment of operation and maintenance (O&M) expenses of public capital

• Integrate responsibilities across agencies for planning of investment and O&M during the budget formulation process

• Conduct the assessment of O&M requirements and formulate a MTEF with the aim of ensuring sufficient appropriations to meet these requirements

• Design and implement policies to enhance the environment for private investment

• Assess the public investment portfolio to

xxvi

determine which projects are worth pursuing further, dropped, or redefined in their scopes, with a view to eliminating projects in “trace allocation”

• Ensure that the newly introduced MTEF is effectively and realistically utilized as a tool for formulating a sustainable investment program, including effective participation by the line agencies

• Formulate a MTEF with the aim of gradually increasing appropriations for public investment in infrastructure conditional on reforms to contain pressures in recurrent expenditures

Managing tax expenditures • Finalize benchmark tax structures for each tax type

• Complete the list of tax expenditures for all tax types

• Provide a functional classification of tax expenditures

• Identify gaps in existing and/or currently planned data bases and tax models to estimate and forecast tax expenditures and undertake initial steps to construct improved databases and models

• Plan required changes in tax laws, tax reporting requirements or modalities and/or database construction procedures including linking tax returns between years in databases.

• Draft discussion documents and hold working meetings with all stakeholders to discuss concepts, benefits and issues of tax expenditure accounting and budgeting and modalities for implementation

• Complete the construction of databases • Complete models (depends upon completion

of databases) • Refine the cost estimate of tax expenditures

by tax and by government function • Conduct a systematic review of tax

expenditures with the aim of eliminating or modifying those with most distorting impact on efficiency and tax payer equity, overlap with direct expenditures, and/or where objectives can be more efficiently achieved through direct expenditures

• All tax expenditures should have sunset dates; be subject to period review; and be repelled if the original economic and social situations on which they were enacted have changed

• Estimate and publish tax expenditure accounts for completed financial years.

• Forecast tax expenditures for use in budgeting • Review budgeting process to provide for

selection of and trade-offs between tax expenditures and direct expenditures

xxvii

Table B: Public expenditures in key sectors—summary of reform options Short-term options (one year)

Medium-term options (2-3 years)

Education

• Design an incentive system and action plan to ensure that secondary schools fill their classrooms up to capacity levels

• Introduce a legally viable mechanism for primary and secondary schools to manage certain, limited financial resources

• Design mechanisms that would give higher education institutions autonomy over financial resources in an accountability framework whereby they would receive their funding in blocks rather than as line items; would be allowed to carry over unspent funds between years, and receive their funding on the basis of output performance for the three main responsibilities of tertiary education: teaching (using student numbers), research and service provision

• Pilot test mechanisms for higher education institutions autonomy

• Develop a policy to measure and report on student learning performance at every school, coherent with the new national curriculum

• Introduce detailed strategic planning for educational institutions, linking outcome objectives to education expenditures at all levels

• Evaluate primary and secondary schools' financial controls and other outcomes with respect to their limited autonomy over certain financial resources

• Introduce a system of per student financing for schools to cover a significantly larger share of the acquisition of educational materials, equipment, professional development services and other services normally acquired by MONE or provincial offices

• Modernize the university entrance examination system to promote improved learning outcomes and in such a way as to eliminate the advantage of private examination preparation courses

• Roll out the financial autonomy mechanisms to all higher education institutions based on the lessons learned in the pilot implementation

• Change the composition of spending in public schools to increase the share of non personnel expenditures on educational materials to levels similar to private schools and universities

• Enhance public education service provision in most backward provinces to compensate for lack of private sector initiative

• Consider increasing per student expenditure on pre

xxviii

primary education and introduce financing instruments that leverage private initiative

Pensions

• Implementing the Social Security and Universal Health Insurance Law as enacted in 2006 to ensure the sustainability of the pension system

• Initiate a nation-wide consultation on a medium-term action plan to realize savings in the pension system, with the aim of financing a future reduction in payroll taxes that would promote employment creation and formalization

• Consider phasing in earlier than currently planed the requirement of 15 minimum years of contribution to gain entitlement to partial pension

• Consider further reducing accrual rates of pension benefits to bring them to the average level in OECD countries

• To strengthen incentives to contribute to the pension system consider amending the Law so that the pensionable base be calculated on the basis of 100 percent of average wage growth

Health care

• Reduce range of reimbursable drugs (i.e., difference between original and closest generic) to 10% of price of generic drug

• Reassess and possibly reduce the number of drugs on the positive list

• Adjust and possibly increase the co-payment rates in drugs by category

• Introduce a flat fee (co-payment) per prescription

• Devote a larger share of well-targeted resources on preventive health care activities

• Rationalize price of generics and original drugs • Introduce medical audit for pharmaceuticals • Expand bioequivalence to avoid switch to more

expensive drugs • Training of physicians on appropriate use of drugs

(mostly antibiotics) and development of standard treatment protocols

• Switch to dispensing drugs by the actual # of pills rather than bottles/vials

• Grant financial and administrative autonomy to public hospitals with a view to creating appropriate incentives for innovative management and stronger accountability

• Consider consolidating small MOH hospitals to improve capacity utilization and allow more efficient operation

• Consider introducing prospective payment mechanisms at

xxix

the hospital level to strengthen incentives to contain costs

Rural expenditures

• Step up financing in the rural investment budget to accelerate project completion

• Redeploy expenditures from the transfer budget into investment-oriented programs with the aim of improving agricultural productivity and off-farm rural employment growth

• Shift emphasis in rural investment from irrigation to high-return activities such as land consolidation, and storage and pre-marketing facilities

• Address over programming in the rural investment budget and maintain adequate level of financing to accelerate project completion

• Maintain focus on rural investments to improve drinking water facilities

• Emphasize participatory approaches, cooperatives, and private sector involvement in service provision to reduce over reliance on government staff for agricultural and rural development activities

Note: Measures highlighted in grey denote options that may require additional expenditures

CHAPTER I SOURCES OF FISCAL CONSOLIDATION AND THE QUALITY OF FISCAL ADJUSTMENT

1.1 Improving the quality of 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. As high public sector borrowing requirements became unsustainable and structural problems mounted, it became inevitable for Turkey to embark on an ambitious reform program aiming at macroeconomic stability and fundamental restructuring of the economy following the 2001 financial crisis. The analysis in this chapter focuses on the quality and sustainability of the strong fiscal adjustment achieved so far, since maintaining hard-won macroeconomic stability is a pre-requisite for promoting faster growth on a sustained basis. At the same time, the analysis benchmarks Turkey’s size and composition of public expenditures and fiscal revenues in international comparison, as achieving an efficient allocation of expenditures and an efficient distribution of the tax burden are key preconditions for fast growth. 1.2 The first section reviews the patterns of fiscal consolidation since 1999 and examines shifts in the focus of 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 of 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 of frequent changes in taxes—this analysis provides a more accurate view of 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 of 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 of the Consolidated General Government (CGG), from a negligible primary surplus of 0.5

31

percent of GDP in 1999 to a surplus of 6.1 percent in 2005 (Figure 1.1).1 (See Box 1.1 for the conversion from the SPO-based consolidated general government primary surplus to IMF program definition of public sector primary surplus). The CGG borrowing requirement excluding privatization proceeds decreased by 11.8 percentage points of GDP between 1999 and 2005, driven by the increase in the primary surplus and the decline of interest payments on public debt. Turkey’s public sector—comprising the consolidated general government plus the State Owned Enterprises (SOEs)—achieved an even larger fiscal consolidation in 1999-2005, from a primary deficit of 1.6 percent of GDP in 1999 to a primary surplus of 6.9 percent in 2005.2 1.4 —which was largely revenue-driven. Fiscal adjustment at the CGG 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. 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 of the Government to restore macroeconomic stability—have led 1 While Annex I presents detailed data sources and methodology, it should be noted that the definition of general government used in this report includes the (i) consolidated budget, (ii) social security institutions (SSI), (iii) a subset of 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 of the Central Bank and other public depository institutions, 36 non-financial state economic enterprises (SOEs), nine regulatory and supervisory agencies, and 45 out of a total of 50 special budget institutions listed in the Public Financial Management and Control Law (PFMC). Special budget institutions refer to 50 public entities established as affiliated or related to a ministry to provide certain public services. These special budget institutions receive revenues and are authorized to spend them. The analysis includes the net subsidies and transfers between these institutions and the consolidated budget. 2 The IMF’s Stand-By Arrangement (SBA) program target of 6.5 percent of GNP primary surplus for 2005-2007, includes primary surplus of a subset of SOEs together with the general government. The primary surplus data for the SOEs provided in Table 1.1 are from IMF staff reports.

Figure 1.1 Fiscal balances and primary expenditures and revenues of the General Government (1999-2005, in % of GDP)

-20.0

-15.0

-10.0

-5.0

0.0

5.0

10.0

1999 2000 2001 2002 2003 2004 200529.0

31.0

33.0

35.0

37.0

39.0

41.0

Fiscal Balance Primary SurplusPrimary Revenues (right axis) Primary Expenditures (right axis)

32

to sharp declines in interest payments on public debt: Interest payments, which had peaked at 25.4 percent of GDP in 2001, up from 14.6 percent in 1999, declined to 9.6 percent of GDP in 2005. However, interest payments still represented the equivalent of 28 percent of total primary expenditures of 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 of 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 is 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 from the SPO to the IMF Program definition

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, its legislation of public procurement, financial management and financial control with international standards. A public procurement agency and a risk management unit for public debt (under the

(% of GDP) 1999 2000 2001 2002 2003 2004 2005Primary Surplus (SPO definition of 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 of 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 (+) Support Price Stabilization Fund repayment 0.0 0.0 0.0 0.0 (+) Interest revenues of SSIs 0.0 0.0 0.1 0.0 0.0 0.0 0.0Adjustments 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(+) Privatization Fund 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(+) Mass 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 5.8SOEs primary surplus (program definition) (2.1) (1.5) 0.1 1.1 0.7 1.1 0.8Program adjusted public sector primary surplus (2.5) 2.6 5.1 4.2 6.3 7.1 6.6

33

General Directorate of Public Finance in the Treasury) have been established; most of the extra budgetary funds have been integrated into the budget; the so called “duty losses” of state banks (obligations to state banks and SOEs due to quasi-fiscal policies) have been accounted in a more transparent manner while generation of 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 of December 2003. The PFMC Law aims at establishing a public financial management and control system compatible with international standards and EU norms. It extends the scope of the budget; provides budgetary unity, increase effectiveness, fiscal transparency, and accountability during the process of preparation and implementations of 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 of the fiscal stance is 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 of CGG expenditures in 2005) by itself created a fiscal adjustment of more than 6 percentage point of GDP (Table 1.1). With the primary surplus of 9.9 percent of GDP, the CB more than doubled its primary surplus compared to 1999. Controlling the extensive off-budget activities of the past by eliminating the earmarked revenue system through the closure of numerous Extra-budgetary Funds (EBFs) contributed to fiscal consolidation. Abolishment of all budgetary funds, with the exception of the Support Price Stabilization Fund (DFIF), and all but five extra budgetary Funds (i.e., the Social Solidarity Fund, the Defense Fund, the Promotion Fund, SDIF and the Privatization Fund) improved fiscal discipline and brought the deficit of the EBFs to a small surplus in 2005.3 Table 1.1: Institutional breakdown of Consolidated General Government Primary Balance

(% 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 Revolving funds 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

3 The number of EBFs has been reduced in 2000 and 2001, and the related earmarked revenue system was abolished in 2004.

34

1.8 Further fiscal adjustment has been impeded by a growing social security deficit, which reached 4.1 percent of GDP in 2005.4 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 of the Social Security Institutions (SSIs) increased from 1.9 percent of GDP in 2000 to 4.1 percent in 2005. Given the favorable demographic profile of Turkey, the currently high deficits pose a bigger challenge in the long run considering that the number of elderly will start to increase as the demographic shift starts to kick in. As further explained in chapter 3, in the absence of the social security reform enacted in June 2006, the projected deficit of the system could have reached 6.7 percent of GDP over the long term. 1.9 Recently enacted social security reform will help contain the deficits over the long term. With the help of the recent reform the pension system deficit is expected to decline by 1 percent below the baseline projection of “no reform” by 2020 (chapter 3). However, the beneficial impact of these measures is expected to be rather limited over the medium run, thus creating little 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 future in order to facilitate formalization and job creation. 1.10 The primary balance of State-Owned Enterprises has significantly improved since 1999. The SOEs covered under the IMF program generated a deficit of 2.1 percent of GDP in 1999 but their fiscal situation improved substantially, resulting in a surplus of 0.15 percent of GDP in 2005 (Table 1.1).5 This improvement was mainly an outcome of reductions of close to 2 percentage points of GDP in their investment expenditures, and 1 percentage point cut in their personnel compensation. The personnel retrenchment of the SOEs has been supported by the PFPSAL program. 1.11 With the privatization of profitable SOEs, like Turk Telekom and Tupras, the contribution of 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 of the consolidated budget through their net transfer to the budget. In 2005, the net transfer from the SOEs to the budget was YTL 0.8 billion or 0.17 percent of GDP. 6 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 of growing business activity in liberalized sectors. In addition, with

4 The deficit of the social security institutions does not include payments made on behalf of the consolidated budget since those expenditures are treated as part of the consolidated budget expenditures. Therefore the deficit of 4.1 percent of GDP is consistent with the more commonly used definition, (including payments made on behalf of the consolidated budget) which corresponds to 4.8 percent of GNP in 2005. 5 The number of SOEs covered in the program was eight in 1999, and this number increased to 26 in 2005. 6 The transfers from SOEs to the budget include: Dividend payments of YTL 2.13 billion and Treasury levy of YTL 0.17 billion. Transfers from the budget to SOEs include capital transfers of YTL 0.9 billion and “duty loss” payments of YTL 0.57 billion. Net transfers from SOEs to the budget amount thus to YTL 0.8 billion.

35

privatization revenues used to pay down public debt, there should be room for expenditure savings as a result of 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 of fiscal consolidation until 2001 reflects predominantly an increase in general government revenues, excluding SSI, and an improvement in the primary balance of the SOEs—the latter up until 2002.7 During 1999-2001 primary revenues increased by 5.9 percentage points of GDP, driven by an increase in indirect taxes of 3.2 percentage points of GDP (Table 1.2). By contrast, primary expenditure of the general government remained almost constant in percent of GDP.

Table 1.2: Fiscal Consolidation in Turkey, 1999-2005 and sub-periods1 (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 1.0 0.5 Total Revenues 7.2 6.2 1.0 Primary Revenues 6.1 5.9 0.2 Current Expenditures -0.3 -0.2 0.0 of 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

1 Changes, between the years indicated, in percent of GDP 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

7 Although tax revenues in percent of 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 of GDP.

36

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 on revenue-increasing measures rather than expenditure rationalization. From 1999 to 2005, almost 100 percent of the 6.1 percentage points of GDP increase in the CGG primary surplus came from revenue-increasing measures—especially from higher indirect taxes. Indirect tax revenues increased from 11.7 percent of GDP in 1999 to 16.7 percent of 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 of investment and reduced expenditures for personnel in 2005. It should be noted, however, that initiatives have been taken to prevent low 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 of expenditure growth than in the past.

B. THE SIZE OF GOVERNMENT AND COMPOSITION OF EXPENDITURES IN INTERNATIONAL COMPARISON8

1.15 Turkey’s primary expenditures are still at the low end of EU countries— Turkey’s public expenditure allocations have been benchmarked against a set of countries comprising cohesion EU countries, new EU members, and emerging market economies. Reflecting still high interest payments on public debt, total expenditures of the Consolidated General Government (CGG) are among the highest of the comparison group (Figure 1.2). However, at 32.7 percent of GDP in 2004, primary expenditure is much lower, and compares favorably with the average size of government seen in cohesion countries of the EU and new EU members—although Turkey’s primary expenditure in percent of GDP is 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 of emerging economies (excluding Mexico where available data do 8 In this section expenditure data for Turkey refer to 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.

37

not capture properly spending by sub national governments) represented 23 percent of GDP in 2003, about ten percentage points of GDP lower than in Turkey. As primary expenditure was only marginally reduced in percent of GDP between 1999 and 2004, Turkey’s relative position remained stable within the benchmark group.

Figure 1.2 : General Government Expenditure as % of GDP

0

10

20

30

40

50

60

Czech

Slovak

Hunga

ry

Poland

Portug

al

Greece

Spain

Bulgari

a

Turkey

99

Irelan

d

Turkey

04

Roman

iaBraz

il

Malays

iaKore

a

S. Afric

a

Argenti

naChil

e

Thaila

nd

Mexico

Primary Expenditure Interest

Source: World Bank staff calculations

B.1. Public expenditure by economic category9

1.17 The Government wage bill is high compared to other emerging economies and some new EU members. At 10 per cent of GDP in 2004, Turkey’s expenditure on wages and salaries at the level of the General Government is relatively high. It is still lower than in Southern EU countries (Greece, Portugal, Spain) or in some new EU members with a tradition of 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 of total General Government primary expenditure in 2004, up from 28.6 percent in 2000, the wage bill represents a large fraction of 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) of primary CGG expenditures in 2004, or the equivalent of 24.6 percent of GDP. Remaining primary expenditures (the equivalent of 8 percent of GDP) are allocated for purchases of goods and services (7.3 percent of GDP) and capital transfers. The bulk of purchases of goods and services are allocated to health (3.8 percent of GDP) and Defense, Public order and safety (1.8 percent of GDP combined). A detailed cross-classification (i.e., economic and functional) of CGG expenditures is provided in Annex I.

38

Figure 1.4: General Government Wage Bill in percent of Primary Public Expendidture (2004)

05

1015202530354045

S. A

frica

Bra

zil

Thai

land

Arg

entin

aP

ortu

gal

Turk

eyH

unga

ryG

reec

eS

pain

Irela

ndK

orea

Pol

and

Mal

aysi

aC

hile

Slo

vak

Rom

ania

Cze

chB

ulga

riaM

exic

o

Figure 1.3: Compensation of General Government employees (2003-04; in % of GDP)

0

2

4

6

8

10

12

14

16

Por

tuga

lH

unga

ryP

olan

dG

reec

eB

razi

lS

pain

Turk

eyS

. Afri

caIre

land

Slo

vak

Arg

entin

aC

zech

Kor

eaTh

aila

ndM

alay

sia

Rom

ania

Bul

garia

Chi

leM

exic

o

comparison (Figure 1.4). This imparts rigidity in the budget at a time when more fiscal space is needed. Source: World Bank staff calculations 1.18 Expenditures for current transfers, mainly comprising social benefits, are oversized by international standards, in view of Turkey’s favorable demographics. At 11.7 percent of GDP in 2004, expenditure on current transfers is lower than in most EU comparator countries (15 percent and up, with the exception of Ireland), but higher than emerging market economies such as Chile, Korea and Malaysia (Figure 1.5). Current transfers in their majority (8.5 percent of GDP) are for social protection, so that spending in this category is better analyzed in the context of the age distribution of population. Given that Turkey has a very young population and the lowest old age dependency ratio among a selected group of comparators, its social protection expenditures are oversized by international comparison (Figure 1.6). In addition, the proportion of 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.10 Therefore, despite the current favorable demographic situation, in the future, Turkey is likely to face an upward pressure on social spending in order to expand the social safety net.

10 OECD Fact book 2005 – ISBN 92-64-01869-7 – © OECD 2005. See Section “Population and migration - demographic trends - ageing societies”.

39

Figure 1.5: Expenditure by Economic Category: Current transfers--Turkey (2004) and comparators (2003), in % of GDP

0

5

10

15

20

25

Pol

and

Cze

ch

Gre

ece

Por

tuga

l

Hun

gary

Spa

in

Slo

vak

Turk

ey

Irela

nd

Chi

le

Kor

ea

Mal

aysi

a

Figure 1.6: General Government Expenditure on Social Benefits and Social Transfers (as % of GDP)

vs. Old Age Dependency Ratio

0.0

5.0

10.0

15.0

20.0

25.0

0.00 0.05 0.10 0.15 0.20 0.25 0.30

Old Age Dependency Ratio

GG

Exp

endi

ture

on

Soci

al B

enef

its a

nd S

ocia

l Tra

nsfe

rs

TUR

MAL

KOR

CHL

IRL

SVK

POL

CZEHUN

POR

ESP

GRE

Source: World Bank staff calculations

Source: World Bank staff calculations 1.19 Turkey’s domestic investment level is close to the average of 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 of 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, careful consideration should be given to increasing the size of the PIP based on sound cost/benefit assessment of investment projects. The large portfolio of frozen projects suggests that the PIP has been oversized in the past, including with projects of questionable priority. Issues in the management of the public investment program are reviewed in chapter 2.

40

Figure 1.7: Consolidated General Government Gross Fixed Capital Formation-- Turkey (2004) and comparators (2003), in % of GDP

0

1

2

3

4

5

6

7

Mal

aysi

a

Kor

ea

Cze

ch

Irela

nd

Bulg

aria

Gre

ece

Chi

le

Rom

ania

Turk

ey

Spa

in

Pol

and

Hun

gary

Por

tuga

l

Thai

land

Slo

vak

Mex

ico

Arge

ntin

a

Bra

zil

S. A

frica

Source: World Bank staff calculations

B.2. Public Expenditure by Government function 1.20 Implementation of functional classification of expenditures is very recent in Turkey. Turkey initiated the GFS 2001 consistent functional classification for the consolidated budget institutions in 2004, and rest of 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 of social sector expenditures, an official functional breakdown of the general government expenditures is not available for 1999-2005 .11 The cross functional and economic classification for 2003 and 2004 used in this study is a first attempt to estimate the general government functional expenditures. Completing a functional classification of expenditures for past years will help analyze more accurately the sectoral impact of the fiscal adjustment. Turkey’s functional distribution of consolidated general government expenditure has been benchmarked against a set of 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:

11 General Government social expenditures first initiated under the PEIR 2001 study of the World Bank. The coverage has been expanded under the PFPSAL program. 12 Comparators include: Russia, Portugal, Greece, Czech Republic, Spain, Ireland, S. Africa, Romania, Malaysia, Mexico, Thailand, Brazil, and Chile. Turkey’s size of government, as measured by primary expenditure in percent of GDP (32.6 percent in 2004), was broadly comparable to the average of these countries (30 percent of GDP in 2003).

41

Figure 2.9: Functional allocation of general government expenditures -- Turkey (2004) and comparators (2003); in % of GDP

0

2

4

6

8

10

12

Gen. Pub.Services

Defense andPublic orderand safety

Economicaffairs

Environmentprotection

Housing andcommunityamenities

Health Recreation,culture and

religion

Education Socialprotection

Turkey Comparators

Source: World Bank staff calculations 1.21 Expenditure for the provision of 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 of comparator countries by 0.8 and 1.2 percentage points of 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 of the investment budget and continuing reform of subsidies to agriculture since 1999. 1.23 Expenditure on health care is relatively high in international comparison. At 5.9 percent of GDP in 2004 (including Hospital Revolving Funds), CGG expenditure on health care is 2.2 percentage points of GDP higher than in the average of comparator countries. Public expenditure on health care depends of course on many factors, such as the structure of the health care system, demographics, and per capita GDP. Further analysis in this area is 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 of GDP is similar to other EU countries and comparable to emerging economies, but educational attainment in Turkey is still 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 of these gaps. This would be a pre-requisite for faster long-term growth (World Bank, 2006 CEM). However, in view of current levels of public spending, existing educational gaps also

Figure 1.8: Functional Allocation of General Government Expenditures – Turkey (2004) and Comparators (2003); in % of GDP

42

suggest potential for efficiency improvements that should go in tandem with increases in the level of 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. As noted earlier, although Turkey’s expenditure in percent of GDP is on par with the average of 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 of air pollution, quality of water supply, wastewater and waste management. Although considerable uncertainty surrounds the cost estimates of necessary investments, it should be noted that costs of meeting the Acquis requirements will also depend on policy choices and reforms in infrastructure services with environmental externalities, while there is considerable room for participation of the private sector.13 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 of fiscal consolidation since 1999 but improving the quality of fiscal adjustment will be a key challenge on the way to the EU. This is important for a number of reasons: (i) international experience suggests that the quality of fiscal adjustment is a key factor for sustainability and is closely associated with stronger growth performance;14 (ii) expenditure pressures are likely to be felt over the medium term, either as a result of 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 of regional disparities, while maintenance needs to be scaled up (see chapter 2); (iv) Government will have to make room in later stages of EU accession for expenditures in areas such as environmental protection where harmonization with the EU Acquis is 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 is little room for Turkey to further increase expenditure in order to meet pressing development challenges. 13 Markandya, A. 2005. "Turkey on the Path of to the Accession: The Environmental Acquis" p. 295 - 309 in "Turkey. Economic Reforms & Accession to the European Union", co publication of the World Bank and the Centre for Economic Policy Research 14 A. Alesina and R. Peroti, “Fiscal Adjustments in OECD Countries: Composition and Macroeconomic Effects”, IMF Staff Papers, 1997; A. Alesina and S. Ardagna, “Tales of Fiscal Adjustment”, Economic Policy, 1998; R. Peroti, “Fiscal Consolidation in Europe: Composition Matters”, American Economic Review, 1996.

43

Figure 1.9: General Government Revenue in % of GDP vs per capita GDP in PPP

0

10

20

30

40

50

60

0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000

GDP per capita, PPP

Tota

l Rev

enue

, % o

f GD

P

20031998

TUR 99BUL

POL SVK

HUN CZE

BRAROM

ARG

MALS. A.

THAMEX

TUR 04GRE

IREKOR

POR

SP

CHL

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 on 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 of improving the efficiency of 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 of civil service pay and employment system, and the rationalization of the investment program would also help contain expenditure pressures on the wage bill and investment spending and thus better control public expenditure across functional areas. Incorporation of 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 of revenue-driven fiscal adjustment, general government revenues have increased to levels comparable to EU countries and much higher than in emerging market economies. In most EU cohesion countries and new EU members the fiscal burden has remained constant, or has even decreased in percentage of GDP, as a result of initiatives to contain the increase in the size of Government (Figure 1.9). But contrary to this trend, and as a result of the pattern of fiscal adjustment since 1999, general government revenues in Turkey stood at 40 percent of GDP in 2005, a level that is now similar to EU countries.15 The increase in the fiscal burden in 1999-2005 was the largest among the comparator countries—similar only to Brazil.

Source: World Bank staff calculations

15 General government revenues include tax revenues, social security contributions, factor income (mainly revenues from SOE’s), and interest income.

44

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

0.4 0.6 0.8 1 1.2 1.4 1.6 1.8

Tax effort ratio in 1995

Tax

effo

rt r

atio

in 2

004

Figure 1.10: Tax effort indicators (1995 and 2004)

Turkey MoldovaCroatia

Denmark

Slovak R

RussiaGreece

Czech R

Romania UkrainePoland

Bulgaria

Hungary

Netherlands

Ireland

AustraliaJapan

Thailand

UK

ArgentinaKorea,Mexico

1.30 The tax effort in Turkey is largely above the average of comparator countries. Tax revenues (including social security contributions) have increased to 31.7 percent of 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 of tax performance across countries in the period 1995-2004 confirms that differences in par capita income, the share of agriculture in GDP, and openness to trade, are consistent factors accounting for international variations in tax ratios (see Annex IV). Accounting for such determinants of tax performance, the estimated “tax effort ratio” in Turkey was 1.3 in 2004, indicating an actual tax ratio of about 30 percent higher than expected based on international experience.16 Turkey’s tax effort was one of the highest among comparators, mainly OECD countries and emerging economies, equal to that of Moldova and surpassed only by Croatia (Figure 1.10). 1.31 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 of 0.9 (Figure 1.10). It 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 1.11). The overall tax effort has stabilized since 2002 at around 1.3, on the back of a slightly lower direct tax effort while, at the same time, the indirect tax effort has kept increasing reflecting the introduction of special consumption taxes and a heavy reliance on excises.

Source: World Bank staff calculations 16 Tax effort is measured by comparing the actual tax ratio of a country with that predicted by using a panel regression. An index of one means the country’s tax effort is at the expected level, given the structural factors of the country. In that case the country is using its taxable capacity at level consistent with the average of the other countries in the sample. See annex IV for the calculations. A tax effort ratio above one potentially indicates excessive taxation, which may have a negative incidence on efficiency and growth for many reasons (capital is mobile; labor and sales could go to the informal market, etc).

45

Turkey

00.20.40.60.8

11.21.41.61.8

2

1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

overall dir indir

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 of fiscal policy in Turkey, especially in view of the extensive informal economy. However, in most comparator countries the tax effort ratio is close to one—on 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 is overstretched, most likely reflecting high and multiple tax rates, given Turkey’s level of development and economic structure. There is 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 of 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 of GNP in 1999 to 16.7 percent of GNP in 2005. During the same period, personal income tax (PIT) revenue fell from 6.5 percent of GNP in 2001 to only 4.9 percent GNP in 2005, while corporate income tax (CIT) revenue remained broadly stable. As a result of these trends, Turkey is one of the countries with the heaviest reliance on indirect taxes among OECD countries, with indirect taxes representing 52 percent of tax revenues, against 36.5 percent on average in the OECD (Table 1.3).

Figure 1.11: Tax effort ratios in Turkey (1995-2004)

46

Table 1.3. Revenue Breakdown for selected OECD Countries 2003

General Government (in percent of GDP)

Taxes & SS Taxes Income

Tax PIT CIT VAT Excises Imp Duty SS

Austria 43.0 28.3 12.9 10.4 2.3 9.5 2.7 0.0 14.7Belgium 45.2 30.6 16.5 13.4 3.0 12.2 2.2 0.3 14.6Canada 33.6 28.6 15.6 ... ... 11.8 0.3 0.2 4.9Czech Republic 36.5 21.3 9.6 4.9 4.7 8.4 2.5 1.3 15.2Denmark 48.7 47.1 28.8 26.0 2.8 17.8 4.0 ... 1.7Finland 44.8 32.6 17.5 14.0 3.5 14.0 4.3 0.0 12.2France 43.4 26.8 10.3 8.1 2.2 14.2 2.5 0.0 16.6Germany 40.3 22.8 10.3 ... ... 7.7 2.6 0.7 17.5Greece 36.1 23.2 8.3 4.9 3.3 11.9 3.2 0.0 12.9Hungary 39.1 26.4 9.3 7.1 2.2 13.2 3.3 1.1 12.7Ireland 29.6 25.0 11.9 7.0 3.8 9.1 2.0 1.4 4.6Italy 42.6 29.8 13.2 10.8 2.3 11.7 2.5 0.0 12.9Netherlands 38.7 24.2 10.1 6.9 3.2 10.7 1.3 1.3 14.5Norway 43.4 33.4 19.3 11.2 4.3 14.4 1.7 0.1 9.9Poland 36.6 22.5 6.3 4.1 2.2 12.8 3.9 1.0 14.1Portugal 36.9 25.1 9.1 5.9 3.2 14.8 3.3 0.2 11.8Slovak Republic 31.2 18.8 7.0 3.4 2.8 9.3 ... 1.4 12.4Spain 35.8 23.0 10.2 6.9 3.3 10.3 2.6 0.0 12.8Sweden 50.2 35.8 18.1 16.0 2.1 17.1 3.3 0.0 14.4United Kingdom 36.1 28.9 13.2 10.4 2.8 10.5 3.5 0.0 7.3United States 25.5 18.5 10.8 8.9 2.0 10.0 1.6 0.2 7.0Turkey (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 of tax ratios need to be interpreted with care, three features of the tax structure in Turkey are especially noteworthy:

• A very heavy reliance on excises, which account for nearly one-third of all tax receipts and raise at least twice as much, relative to GDP, as in any other country in the sample.17 Heavy reliance on excises is of concern as it may encourage activity in the informal sector, which would further undermine the tax bases and impede efficiency for the economy as a whole. Reducing this high level of 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.18

17 This appears to mainly reflect high rates on ‘traditional’ excisable goods: around half is from petroleum products alone—indeed, at about $8/gallon, pump prices for gasoline are about the highest in Europe. In addition, the taxes on “luxury goods” raise revenue of about 0.1 percent of GDP. 18 Of course excises need to be raised to ensure that negative externalities are fully offset. While a systematic international comparison of specific excises would be appropriate, cursory evidence suggests that excises in Turkey are comparatively high. For 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

47

• Low receipts from, and reliance on, income taxation. Although the PIT yield in percent of 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 may be the existence of a large informal sector. Similarly, Corporate Income Tax (CIT) revenues are somewhat lower than the OECD average, despite the fact that the CIT rate in 2004 was a substantial 33 percent (though reduced in 2005 to 30 percent).19 This indicates that the base of 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 on wages are combined, total payments to the government out of labor income can approach 43 percent of 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 on 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 of ensuring some degree of compliance of those who would have otherwise escaped taxes altogether. Moreover, indirect taxes offer a more symmetrical treatment of labor, transfer, and capital income, thus meeting some criteria for horizontal taxpayer equity better than income taxes that discriminate against some types of income or are subject to considerable evasion. These factors tend to alleviate the distributional consequences of indirect taxes. In addition, indirect taxes uniformly levied on consumption offer advantages on 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 mix relying more on indirect taxes may be more conducive to growth. Caveats apply, however, to these considerations, if greater reliance on indirect taxes is accompanied by proliferation of 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 of indirect taxes would call for careful study and long-term design. Although the balance between direct and indirect taxation in Turkey is considerably out of line 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 to increase by 11 percent, the increase in SCT on Petroleum and Natural Gas is lower, estimated at 3 percent. 19 Further reduction in the CIT rate to 20% has been initiated starting from January 2006.

48

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 of direct for indirect taxes would not be advisable in view of the above-mentioned considerations regarding the efficiency benefits of indirect taxes and their equity-improving incidence when significant informality is present, as in the case of 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 of the debt ratio and create room for a lower primary surplus in percent of GDP; (ii) better control of primary expenditures in low-value programs; (iii) tax base broadening and reduction in the size of the informal sector. The analysis that follows presents options for income tax reform with the aim of improving efficiency and broadening the tax base. Eventually, these initiatives will also facilitate a shift in the balance of the tax burden between direct and indirect taxation. 1.37 The tax landscape is changing fast— The authorities are currently in the process of reviewing and rewriting each of their major tax laws, also following commitments taken in the last Stand-By Arrangement with the IMF. The authorities eliminated stamp duties on 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 on previously exempt government securities, to final withholding at 15 percent. This would move Turkey closer to an explicit “dual income tax” (DIT) model of 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 of the futility of trying to tax a highly mobile base like 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 of 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 on 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 of asset-specific distortions that the withholding tax regime was supposed to achieve. As for offering different tax treatment to non-residents, despite the temporary boost to the attractiveness of 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 on building a broader array of tax treaties that would carve out particular exceptions for non-residents. 1.39 The 2006 budget and draft legislation to amend the existing PIT and CIT laws initiated significant reforms of the Corporate Income Tax and the Personal Income Tax. They include: (i) a reduction of the CIT rate from 30 to 20 percent, with a parallel

49

elimination of the Investment Tax Allowance (ITA); and, (ii) a reduction of the top personal income tax rate from 40 to 35 percent for non-wage earners, accompanied by a reduction in the number of tax brackets from 5 to 4 and a unification of the tax regimes for wage and non-wage income.20 As a result of tax base broadening, due to the elimination of the ITA, ongoing tax administration reforms, and broader strengthening of the CIT code (regarding, for example, transfer pricing and thin capitalization),21 the Government does not expect a significant impact of these measures on CIT collections, and stands ready to take offsetting expenditure measures if necessary to keep the primary surplus on target. The CIT tax reforms have the potential to promote investment on a sound basis and stimulate job creation in the formal sector. The PIT reform will improve the efficiency of the tax system, as the dual tax system for wages and non 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 CIT legislation has been put into force, while only the PIT legislation concerning the tax brackets and tax rates has been implemented. 1.40 Reduced VAT rates introduced recently are of questionable effectiveness and not supportive of base broadening efforts. As part of its tax initiatives the Government announced in March 2006 reduced VAT rates on 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 is not appropriate for the goal pursued, as reduced VAT 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 of the tax system, create more opportunities for misreporting and misclassification for tax purposes, and thus work against efforts of strengthening tax compliance. 1.41 Initiatives in the years ahead should focus on tax-base broadening. Two goals should be pursued with appropriate prioritization: (i) In the short term, promoting tax rationalization, with the aim of reducing the distortions associated with a given level of the tax burden and broadening the tax bases; (ii) Over the medium term, making fiscal room for lower taxes, by taking advantage of the reduction in the debt ratio and pursuing public expenditure reforms aimed at containing expenditure pressures.

20 By the new reform, it is also planned to reduce the lowest tax rate for income components other than wages from 20 percent to 15. The highest and the lowest tax rates on wages will be maintained at 35 percent and 15 percent respectively, thus effectively unifying the two tax schedules. 21 “Thin capitalization” occurs when multinational firms are able to shift profit across jurisdictions for tax purposes by increasing the amount of debt financing in jurisdictions with high tax rates in order to benefit from the deductibility of interest from the corporate tax base. 22 Authorities argue that the rate reduction will also tame incentives to use forged documents to get fictitious rebates in VAT refund system, thus helping improve compliance in the short term.

50

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 of fiscal 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 of fiscal adjustment has to distinguish between structural and cyclical parts of the budget. 1.43 Changes in the structural deficit can contribute to the assessment of the sustainability of fiscal adjustment in the medium term by identifying the impact of the growth cycle on the budget. Short-term improvements in the budget, due to robust economic growth, may 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 of 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. It would also depend on whether the policy-induced changes are sustainable themselves over the medium term. The lack of major restraint in non-interest expenditures can raise questions as to the sustainability of the adjustment as discussed earlier. This section aims to address only the first issue above related to the cyclicality of the observed fiscal adjustment. 1.45 The basic methodology for 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 III. Because long time series data is not available at general government level, estimations relied on 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 level.23 On the expenditure side, no adjustments were made for the impact of cycles in economic activity. The main budget expenditure item sensitive to the cycle is unemployment

23 The consolidated budget accounts for more than 85 percent of 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.

51

insurance. It should be noted that unemployment insurance is a recent development in Turkey with very little insurance payments to unemployed so far and is also out of the coverage of 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 of indirect taxes is 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 CIT (see Annex III). The estimated elasticities are close to the OECD averages. The fact that the elasticities are greater than one reflects the progressivity of taxes in Turkey, as in most OECD countries. Table 1.4: Output Elasticity of Taxes (1)

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 1.1 United Kingdom 1.2 1.7 Canada 1.1 1.5 Australia 1.0 1.4 Austria 1.3 1.7 Belgium 1.1 1.6 Denmark 1.0 1.6 Finland 0.9 1.6 Greece 1.7 1.1 Ireland 1.4 1.3 Netherlands 1.7 1.5 New Zealand 0.9 1.4 Norway 1.0 1.4 Portugal 1.5 1.2 Spain 1.9 1.2 Sweden 0.9 1.8 Average 1.3 1.5 Turkey (2) 1.5 1.6

(1): Elasticities for all countries except Turkey are taken from the study “Girouard, N. and C. André (2005), Measuring Cyclically-Adjusted Budget Balances for OECD Countries, Economics Department Working Papers, No. 434.

(2): World Bank staff estimations

52

1.47 Estimated structural primary budget balances indicate a strong consolidated budget performance from 2000 onwards.24 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 of 1-2 percent of 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 of 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 further 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 of changes in the structural overall balance of the consolidated budget (Figure 1.13). Figure: 1. 12. Actual and Structural Primary Budget Balances for Consolidated Budget

-5.0

-2.5

0.0

2.5

5.0

7.5

10.0

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

Per c

ent

0.85

0.90

0.95

1.00

1.05

1.10

1.15

1.20

1.25

1.30

y / y

p

Structural PBB Actual PBB y/yp

Note: Structural budget balances are percent of potential GNP, while actual budget balances are percent of 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 not steady revenue sources.

53

Figure: 1. 13. Actual and Structural Budget Balances for Consolidated Budget

-18.0

-9.0

0.0

9.0

18.0

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

Per c

ent

0.82

0.91

1.00

1.09

1.18

y / y

p

Structural BB Actual BB y/yp

Note: Structural budget balances are percent of potential GNP, while actual budget balances are percent of 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 of 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 of the cycle to the performance of the primary budget balance has reached substantial levels as of 2005.25 This indicates that the current fiscal performance can be significantly undermined in case of 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 of agriculture (which does not pay taxes) and the informal economy, and also the large share of indirect taxes which are less cyclical compared to direct taxes.

54

Figure: 1.14 Cyclical Budget Balance for Consolidated Budget

-5.0

-4.0

-3.0

-2.0

-1.0

0.0

1.0

2.0

3.0

4.0

5.019

90

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

Per c

ent

0.85

0.90

0.95

1.00

1.05

1.10

1.15

y / y

p

Cyclical BB y/yp

Source: World Bank staff calculations

1.49 Changes in the structural budget balance are a useful indicator of whether fiscal policy 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 of the changes in the structural fiscal balances with the changes in the output gap implies a counter-cyclical fiscal stance: It shows that fiscal adjustments are taking off 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 of the consolidated budget, the fiscal stance has often turned pro-cyclical. The analysis of 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 half of 1990s. In this context, it 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 of the recovery, the weakening in the structural primary balances in 2004-2005 indicates that the fiscal stance has become pro-cyclical again.

55

Table 1.5: Fiscal Stance and the Output Gap(1) (Consolidated Budget)

Change from the Previous Year (Percentage points) Output gap

(actual/potential) Structural Primary

Budget Balance (% of Potential GNP)

Structural Budget Balance

(% of 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 -1.1 -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 1.51 The estimated structural primary budget balance confirms that fiscal performance was also strong at the general government level.26 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 of general government, social security contributions are also adjusted for the cycle. The estimated elasticity of 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 of 0.82 is close to the average of OECD countries (Table 1.6). The estimated structural primary balance of the general government was, on average, about 6 percent for the period of 2000-2005 (Figure 1.15). The overall structural deficit of the general government has declined substantially after 2001 with the help of high primary surpluses and declining real interest rates. The structural deficit declined from a peak of 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.

56

Table 1.6: Output Elasticity of Social Security Contributions (1)

Social Security Contributions United States 0.6Japan 0.5Germany 0.5France 0.8Italy 0.9United Kingdom 0.9Canada 0.6Australia -Austria 0.6Belgium 0.8Denmark 0.7Finland 0.6Greece 0.8Ireland 0.9Netherlands 0.6New Zealand -Norway 0.8Portugal 0.9Spain 0.7Sweden 0.7 Average 0.7 Turkey (2) 0.8

(1): Elasticities for all countries except Turkey are taken from the study “Girouard, N. and C. André (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 of potential GDP since then. Similar to the case of 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 of the changes in the structural fiscal balances with changes in the output gap.

57

Figure: 1.15. Actual and Structural Primary Budget Balances for General Government

-5.0

-2.5

0.0

2.5

5.0

7.5

10.0

1999

2000

2001

2002

2003

2004

2005

Per c

ent

0.85

0.90

0.951.00

1.05

1.10

1.151.20

1.25

1.30

y / y

p

Structural PBB Actual PBB y/yp

Note: Structural budget balances are percent of potential GNP, while actual budget balances are percent of actual GNP.

Source: World Bank staff calculations Figure: 1.16. Actual and Structural Budget Balances for General Government

-20.0

-10.0

0.0

10.0

20.0

1999

2000

2001

2002

2003

2004

2005

Per c

ent

0.82

0.91

1.00

1.09

1.18

y / y

p

Structural BB Actual BB y/yp

Note: Structural budget balances are percent of potential GNP, while actual budget balances are percent of actual GNP.

Source: World Bank staff calculations 1.53 The pro-cyclicality of 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 of the one-off, revenue raising measures vanished in 2002, structural budget balances deteriorated although growth was recovering. This pattern of fiscal adjustment contributed to the pro-cyclical stance of fiscal policy in 2002.

58

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 of the consolidated budget, social security institutions and local administrations. The structural primary balances of the consolidated budget declined by 0.7 percent of the potential GNP in 2004-2005.27 The structural primary balances of the social security institutions weakened by 1.2 percent of potential GNP in the period 2004-2005 and by 0.7 percent of potential GNP in 2005 alone. The long-awaited amnesty for premium collections must have contributed to this outcome. It is expected that the recently enacted social security reform will slowly contribute to the reversal of this trend over the coming years (see chapter 3). Similarly, the structural primary balances of the local administrations fell by 0.6 percent of potential GNP in 2005 but the overall decline over 2004-2005 was only 0.4 percent of potential GNP due to a small improvement in 2004. The pro-cyclicality in the last period is likely to have had a magnifying effect on the growth cycle. Table 1.7: Fiscal Stance and the Output Gap(1) (General Government)

Change from the Previous Year (Percentage points) Output gap

(actual/potential) Structural Primary

Budget Balance (% of Potential GNP)

Structural Budget Balance

(% of 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): 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 1.55 The findings are robust to alternative modeling assumptions. The impact of 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 of potential output. The cyclical part of the general government budget was recalculated by setting all tax elasticities to unity in order to see if higher tax elasticities overstate the impact of the cycle and by estimating potential output with a different methodology in order to test robustness of the results with respect to the estimation of 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). 27 The consolidated budget figures here are defined within a context of general government, which requires some adjustments in the data. Therefore the figures presented here may differ from the ones given in the previous section. 28 To estimate potential output, the actual output time series was “filtered” to remove cyclical fluctuations using a Hodrick-Presscott filter.

59

Table 1.8: General Government Cyclical Budget in % of GDP 1999 2000 2001 2002 2003 2004 2005 Potential output is estimated by production function Cyclical Budget αCIT=1.57 αPIT=1.50 αSSC=0.82 -0.8 0.2 -4.0 -1.8 -0.8 1.1 2.2 Cyclical Budget αCIT=1.0 αPIT=1.0 αSSC=1.0 -0.7 0.2 -3.5 -1.6 -0.7 1.0 2.1 Potential output is estimated by alternative method

(Hodrick-Presscott filter) Cyclical Budget αCIT=1.57 αPIT=1.50 αSSC=0.82 -0.5 0.6 -3.8 -2.0 -1.3 0.8 2.2 Cyclical Budget αCIT=1.0 αPIT=1.0 αSSC=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 of higher than expected inflation and exchange rate depreciation, can trigger a slowdown in growth. The operation of automatic fiscal stabilizers would limit the size and/or duration of any growth slowdown, so that their impact on the primary fiscal surplus would normally be welcome. However, even under a soft landing scenario, a declining primary surplus may have secondary unwanted effects on the economy, ranging from a deterioration of debt dynamics, to even higher real interest rates, and a non-negligible effect on market sentiment. 1.57 To avoid a pro-cyclical fiscal stance, the structural primary surplus should be maintained unchanged during years of 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 of the debt ratio would continue even in case of 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 of 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 still have to be preserved in cases of slower growth, thus making the functioning of fiscal stabilizers “asymmetric”. In view of the positive output gap, the decline in the structural primary surplus in 2004 and 2005, the widening deficit of the external current account, and the need to ensure that the debt ratio continues to fall despite higher interest rates, some withdrawal of fiscal stimulus would seem appropriate in the 2007 budget.

60

CHAPTER II OPTIONS FOR EFFICIENCY GAINS IN PUBLIC EXPENDITURES—HORIZONTAL ISSUES

2.1 To maximize the efficiency of expenditure programs, reforms in key functional areas should be underpinned by horizontal public expenditure reforms. The natural focus of public expenditure reforms is to improve the efficiency—for a given cost—of 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 of expenditure programs across sectors. Horizontal reforms include initiatives that improve, for example, incentives in the civil service towards the achievement of results, or help contain pressures on the wage bill, so that a larger share of expenditures can be spent on materials and equipment necessary for the provision of high-quality public services. These reforms can obviously help achieve superior outcomes in key sectors that absorb large amounts of public financial resources, such as education or health care. Adoption of efficient project evaluation methods, multi-year programming, and adequate provisioning for long-term operation and maintenance needs are examples of initiatives that improve the quality of 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 civil service reform. The third section reviews trends in the public investment program across sectors and highlights some priority directions to ensure the sustainability of the medium-term investment program in infrastructure. The fourth section provides a preliminary analysis of tax expenditures in Turkey—an important area for the improvement of fiscal accountability, the transparency of governance, and the safeguarding of a strong fiscal framework. Transparency of tax expenditures would support their rationalization; help avoid duplication with direct expenditure programs in key 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 of

61

5

6

7

8

9

10

11

1999 2000 2001 2002 2003 2004

Consolidated budget SSIs Local administrations Revolving Funds EBFs

Figure 2.1: Personnel expenditures in percent of GDP (1999-2004)

GDP, growing from 9.5 percent to 10 percent between 2000 and 2004 (Figure 2.1).29 The lion’s share of personnel expenditures, about 80 percent, is incurred by consolidated budget agencies. Local administrations and revolving funds account for 9 and 7 percent of total personnel expenditure. These shares have remained stable over time.

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 of the wage bill in 2004 (Table 2.1).30 The fourth major category of personnel expenditure is in economic affairs, accounting for the provision of infrastructure services, rural development, and services to agriculture. The share of this functional category in total personnel expenditure is equal to that of public order and safety and higher than defense. Provision of general public services is the fifth largest category of personnel spending for the General Government as a whole. It is relatively oversized at the level of local administrations, representing the largest component of their wage bill. Table 2.1 : Personnel expenditures by government function and level of General Government

(2004; in percent). Consolidated

budget Social Security Institutions

Local administrations

Revolving Funds

Total

General Public Services 7.9 0.1 2.4 0.1 10.4

Defense 10.1 … 10.2 Public Order and Safety 12.7 0.8 … 13.5 Economic Affairs 10.5 … 1.6 1.1 13.2 Environmental Protection 0.1 0.5 … 0.7 Housing and Community Amenities … 1.7 … 1.8 Health 7.7 3.0 0.6 5.3 16.4 Recreation, Culture and Religion 2.5 … 0.6 … 3.1 Education 27.6 1.0 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 Source: World Bank staff

29 Steep 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 is incurred by the consolidated budget, but more than half of the wage bill for health care is borne by Revolving Funds and Social Security Institutions.

62

Figure 2.2 Composition of Wage Bill by Function (in percent of Total Wage Bill; 2004)

20.7

7.4

19.1

6.410.4

23.7

13.916.4

34.0

1.5

35.7

10.7

05

10152025303540

General publicservices

Defense &Public Order

EconomicAffairs,

EnvironmentalProtection

Health Education,Culture,

Religion &Housing

Socialprotection

Average of Comparator Countries Turkey

200000

250000

300000

350000

400000

450000

500000

1997 1998 1999 2000 2001 2002 2003 20041,800,000

1,820,000

1,840,000

1,860,000

1,880,000

1,900,000

1,920,000

Figure 2.3 : Employment in the Public sector (1997-2004)

Consolidated budget agencies

SEEs

Municipalities

2.5 The wage bill in economic affairs and security is oversized by international comparison. The functional breakdown of General Government personnel compensation in Turkey was compared to the average of seven other new EU members, cohesion countries, and emerging economies for which similar information is available: Czech Republic, Greece, Ireland, Mexico, Portugal, Spain, and South Africa. The share of personnel compensation in security spending (Defense and Public Order and safety combined) is somewhat above the average of comparators (Figure 2.2). In particular, the share of personnel expenditures on economic affairs is twice as high. An example of oversized wage bill with possible room for savings is that of the rural agencies budget (Chapter III, section D2).

Source: World Bank staff calculations Source: Ministry of 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 of the economically active population. Three quarters of public sector employees are in consolidated budget agencies. Civil servants comprise the bulk of public employment, while public workers—concentrated in SEEs and State banks—represent about 22 percent, and contracted personnel only 5 percent of total (Table 2.2).31 Overall, there has been a decreasing tendency in employment of all categories of 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, of around 30 percent from 1997 to 2004. Employment in municipalities has remained stable. However, a significant change in the composition of municipal employment has occurred, with the share of temporary workers increasing to 45 percent of total in 2004, from 25 percent in 1995. 31 Civil servants are defined as those public employees carrying out public services of a primary and continuous nature. They enjoy constitutional protections which effectively provide a lifetime job guarantee. Contracted personnel in theory are employed to fill temporary needs (typically unfilled civil service positions), and are hired on one-year, renewable contracts. In practice, Turkey’s courts have ruled that they enjoy all benefits and protections of civil servants. Public workers have a contract but do not occupy a civil service position. They do not enjoy the same job protections as the other categories but, unlike civil servants, they are both allowed collective bargaining and have a right to strike.

63

Table 2.2: Total employment in Turkey’s public sector (1), September 2004 Civil servants Contracted

Personnel Workers Total % of Total

Public Sector Consolidated budget

1,620,341 17,182 184,621 1,822,144 74.0

Municipalities 92,487 2,178 106,700 201,365 8.2 Special provincial administrations

3,924 87 762 4,773 0.2

Social Security institutions

74,391 349 3,728 78,468 3.2

State economic enterprises,

7,880 106,670 209,614 324,264 13.2

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 of 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 of defense and public order and safety, public sector employment in Turkey amounted to 2.2 percent of population in 1999. This is somewhat above the scale of public employment in other emerging economies such as Chile, Korea, and Mexico. However, it 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 of 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 of 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 of per capita income for the economy as a whole. Average compensation has thus increased to 2.9 times of per capita GDP in 2001-2004 (Figure 2.4c).32 Measured as a percentage of 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 of 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 from nominal GDP and compensation figures, with the ratio of average wage to per capita GDP increasing from 2.8 in 2000 to 3.4 in 2004.

64

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

Turkey 01-04

Malaysia

Turkey 97-00

Chile

ArgentinaSpain

KoreaPoland

Mexico

Slovak Republic

Bulgaria

Russian Federation

Czech Republic

Hungary

Romania

Figure 2.4b : Average personnel compensation in central government as a ratio to per capita GDP (1997-2000 for all countries and 2001-04 for Turkey)

2.22.32.4

2.52.62.72.8

2.93

3.1

1997 1998 1999 2000 2001 2002 2003 20040.50.520.540.560.580.60.620.640.660.680.7

Figure 2.4c : Trends in average personnel compensation in consolidated budget agencies (1997-2004; as ratios to per capita income measures)

Average compensation / non agricultural GDP per employee (right axis)

Average compensation / per capita GDP (left axis)

0

0.5

1

1.5

2

2.5

3

3.5

4

Russia Bulgaria Poland Romania Turkey Chile Korea Mexico

Figure 2.4a : Central Government employment (excluding Defense and security) in percent of population -- 1999

0

10

20

30

40

50

60

70

80

90

1998 1999 2000 2001 2002 2003 2004-15

-10

-5

0

5

10

15

Figure 2.4d : Average compensation in consolidated budget agencies and CPI inflation (1998-2004; in percent)

Average compensation

CPI

Average compensation in real terms

Figure 2. 4: Employment and average wage cost in the public sector

Public sector employment is not oversized… …but average compensation is high by international standards…

…and has been growing faster than per capita income… …because nominal compensation increases have by

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 of GDP despite downsizing of 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 is largely driven by the increase in average compensation. For example, had the ratio of average compensation to per capita GDP remained constant at its average level in 1999-2000, the wage bill of consolidated budget agencies would have fallen to around 7.6 percent of GDP in 2004—that is, about 1 percentage point lower than its actual level (Figure 2.5). Better management of 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 of 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.

65

6.5

7

7.5

8

8.5

9

9.5

1999 2000 2001 2002 2003 2004

Figure 2.5 : Wage bill in consolidated budget agencies -- actual and alternative scenario (1999-2004; in % of GDP)

Actual

Constant ratio of average wage to pc GDP at 1999/2000 level

0.9%

Source: World Bank staff calculations 2.10 The wage bill could further increase as a result of 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 of YTL 40 for the first half and YTL 80 for the second half of the year to the public servants who do not receive additional payments. The estimated cost of this action is about YTL 1.7 billion, the equivalent of 0.35 percent of GDP. Part of 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 is 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 it would have a permanent impact on the Government's wage bill. 2.11 The compensation system is dominated by a multitude of side benefits. In 2004, about 32 percent of personnel compensation in consolidated budget agencies was composed of various allowances, compensations, and rewards on top of the basic salary of employees. Moreover, overtime pay represented 6.5 percent of total compensation. While wages are calculated on a transparent basis (driven by a combination of 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 of 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 of GDP) is a more expensive form of 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 of basic civil service obligations.

66

2022242628303234363840

1997 1998 1999 2000 2001 2002 2003 20040

1

2

3

4

5

6

7

8

Figure 2.6 : Compensations, Rewards, Allowances, and overtime pay in percent of total personnel expenditures of consolidated budget agencies (1997-2004)

Overtime pay Compensations, rewards, allowances

Source: Ministry of 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 of foreign language. Several allowances for specific duties and overtime payments are also exempt from the income tax. Although a precise breakdown of the amount of non taxable allowances is not available, in view of the substantial size of allowances and overtime pay (3.3 percent of GDP for consolidated budget agencies), their cost in terms of foregone tax revenue is likely to be significant. Assuming that 70 percent of 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 of 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 of allowances is disproportionately large for the highest grades. For the highest grade of civil servants, wages (including basic seniority and indicator salary plus side payment) account for only 15 percent of total compensation. For the lowest grade, wages typically account for about 75 percent of compensation. The exceedingly high role of allowances among the highest grades is cause for concern for transparency and incentives in the civil service, and warrants further analysis. Moreover, as allowances represent a much larger share of income for higher-grade civil servants, exemptions from the income tax are detrimental to vertical taxpayer equity. One plausible hypothesis is that these more discretionary forms of pay are the means via which Turkey achieves de facto decompression of what is otherwise an excessively compressed set of pay differentials between more and less skilled civil servants.33 Achieving decompression in compensation scales will have to be an important part of civil service pay reform. Without

33 Adequate data does not exist to assess the compression of wage scales in civil service in international comparison. While the distribution of wages can be assessed by grade for basic pay, similar information is not available for allowances and other side benefits.

67

decompression, taxation of allowances would reduce pay differentials and may have unintended effects on 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 of the deficiencies of the existing system. The proposed draft law on public personnel seeks to address a number of major objectives. These include:

• improving efficiency in staffing and performance of duties; • eliminating discrepancies in remuneration across similar positions within the

public sector; • making it easier to manage a fiscally sound wage bill; • providing competitive opportunities for entry into the public administration

The analysis below reviews some of the reform options to meet these challenges but further detailed analysis of opportunities for modernizing the rules governing civil service pay—and any related opportunities for budgetary savings—is warranted. In addition, civil service reform would need to address issues related to the deployment of a number of public employees to the special provincial administration as a result of the recently enacted legislative framework. The majority of those employees are still associated with the consolidated budget in terms of 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. (i) Main reform objectives 2.15 Improving efficiency: To improve efficiency this draft law mandates changes aimed at three broad sub-objectives:

(i) make it easier to adjust staffing composition when required;34 (ii) enhance sources of performance motivation for public personnel;35 (iii) facilitate opportunities for public servants to improve their human capital over the

course of 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 of employment for “contracted personnel” than is possible for “civil servants”, given Article 128 of Turkish Republic Constitution.37 In addition, the proposal would mandate dismissal of

34 To increase allocative efficiency 35 To increase technical efficiency 36 To increase dynamic efficiency 37 More specifically, the draft law 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 to have their contracts terminated upon end of contract within the first 15(-) years of such a contractual relationship (Art. 17), although they would still enjoy the same retirement (pension) rights as “public servants”. It would also allow contracted personnel to “annul their contracts according to the methods and principles stated in their contracts” (Arts. 19, 63 and 64).

68

public servants receiving repeated “unsatisfactory” annual performance evaluations.38 To address (ii), the draft law would: (a) authorize a modest amount of performance bonuses, subject to restrictions on both the amount received by any individual employee as well as the total expenditures on such bonuses by each public entity;39 and (b) mandate dismissal of public servants receiving repeated “unsatisfactory” annual performance evaluations. To address (iii), the draft law would provide generous opportunities for public servants to obtain either graduate level education or valuable work experience during periods of secondment to international organizations or other overseas employers, during their public service careers, which can significantly enhance the attractiveness of a career in Turkey’s public service. Moreover, heavier reliance would be placed on Turkey’s higher education and private sectors for providing in-service training to its public service staff, thereby taking advantage of competition in the provision of 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 of four separate career streams, and public contracted personnel in each of three career streams;40 (ii) concentrate salary in (a) basic duty salary or wage,41 and (b) duty difference allocation.42 Salary would thus be concentrated in those elements most clearly linked to human capital requirements and demands of the position.43 38 Article 75(4) provides that “Public servants, whose personnel and success evaluation is negative two times successively are appointed to the order of another personnel and success evaluation supervisor. Public servants whose evaluation is here again negative and those public servants who get negative evaluation a total of four times within ten years are dismissed and these can never be employed again as public servants or contracted personnel.” 39 Performance bonuses cannot exceed 5% of an employee’s basic duty salary or basic duty wage, while each public institution’s payments of such bonuses cannot exceed 1% of 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), (iii) judicial sector public servants (i.e., judges and prosecutors), and (iv) academic public servants (i.e., university instructors); while the contracted personnel basic duty wage scales apply to (v) general contracted personnel, (vi) military contracted personnel (i.e., specialist enlisted leaders), and (vii) academic contracted personnel (i.e., university instructors). 41 I.e., the part of salary or wages fixed by the seven uniform basic duty salary and wage scales established by this law. 42 The draft law constrains the awarding of these “duty difference” pay amounts by specifying (i) grounds justifying “duty difference” payments, (ii) procedures for setting payment levels, and (iii) limits on their magnitudes. (i) Grounds: Such “duty difference” pay would be based on “the hardship of the service, responsibility and risks, work intensity, work conditions, socio-economic development level of the service location, living conditions and geographic characteristics, duty and other characteristics relating to the service location” (Art. 110(2)). (ii) Procedures: The amounts and rules and procedures governing assignment of such “duty difference” pay to individual personnel would be established through the following procedures: (a) proposals to be prepared by the institution within which such “duty difference allocation” is proposed to be established; (b) the State Personnel Administration to provide its opinion on each proposal; (c) the Ministry of Finance to submit each proposal to a commission consisting of the Undersecretary of the Ministry of Finance, the Undersecretary of the State Planning Organization, the Undersecretary of Treasury and the Chairman of the State Personnel Administration, under the chairmanship of the Undersecretary of the Prime Ministry, and (d) that commission to make the final decision (Article 110 (2) and (3)). (iii) Limits: Such payments are limited to no more than the lower of (a) “twenty percent of the highest basic duty salary” or “forty percent of the personnel’s own basic duty salary or basic duty wage and fifty percent for” selected military positions (Article 110(1)). 43 Articles 108-111. Articles 112-116 provide other elements of compensation.

69

In addition, in conjunction with related changes in the Social Security and Universal Health Insurance Law, the proposal would ensure that any elements of remuneration that bring with them social security (pension) rights, also impose commensurate social security contribution (premium payment) obligations on the recipient of those elements of remuneration. 2.17 Making it easier to ensure a fiscally sound wage bill: To 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 of Finance to monitor and analyze central government wage bill implications of employment and pay policy proposals or changes. Moreover, this consolidation would ensure that any elements of remuneration that bring with them social security (pension) rights, also impose commensurate social security contribution (premium payment) obligations on the recipient of those elements of remuneration, thereby reducing perverse incentives to obtain eligibility for such elements of salary, as well as eliminating a fiscally imprudent practice of providing pension benefits not tied to pension contributions. 2.18 Providing competitive opportunities for entry into the public administration: To 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 of the proposed changes in the legal framework represent important steps toward achieving those objectives. With the aim of supporting these efforts, it is important to highlight implementation challenges posed by the proposed reforms in the public personnel management legal framework, as well 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 civil servants and contracted personnel; (ii) mandating dismissal of civil 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 of the key reform objectives include: (vii) step salary increase provisions; and (viii) personnel performance evaluation procedures. 2.21 Establishing legal distinction between public servants and contracted personnel: The text of this draft law which provides a legal basis for permitting non-renewal of contracted personnel contracts has been very carefully drafted. It is unclear how the Turkish judicial system is likely to rule on challenges to these provisions. The Government has and continues to pay very close attention to the risk of an adverse judicial finding once these provisions are challenged. This is an unavoidable risk and challenge posed by the laudable

70

determination to create a more flexible labor market for an important subset of Turkey’s public employees; a challenge that the Government fully recognizes. An issue that would merit attention is that this draft law leaves entirely to other “special laws” the task of defining which particular cadres (for public servants) and positions (for contracted personnel) shall exist. This leaves substantial scope for lack of uniformity in the determination of 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 of procedures should be established to provide adequate checks and balances and reasonable assurance that these determinations will follow a relatively consistent set of criteria. 2.22 Dismissal for repeated “unsatisfactory” performance ratings: The current Law on State Functionaries44 allows termination of employment for public servants only for events such as major disciplinary actions, voluntary departure, retirement, and death. Poor performance is not grounds for dismissal. This new draft law would change that by making consistent poor performance grounds for dismissal. This is a fundamental, important and much needed change. There is a risk, however, that a poorly functioning annual personnel performance evaluation process will make this a change with little or no impact. This is 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, it is difficult to get a performance pay system to work as well as one might hope. In addition to limiting the magnitudes of such performance bonuses, perhaps the most important challenge is 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 is a move in the right direction, the current draft law does not uniquely determine what fraction of total remuneration will reflect these two components. The large number of other elements of remuneration and leave rights could compromise this objective. Careful analysis will be required to assess whether the implementation of these provisions will actually concentrate salary in basic duty and “duty difference” elements of remuneration. 2.25 Setting basic salary and wage scale index (“indicators”) structure: Because of current significant discrepancies in remuneration for similar positions across institutions, implementation of the more uniform salary structures mandated by this law will require changes in salary relativities. If implementation required no change in the overall wage bill, some staff would see their remuneration rise while others would see it fall. If, on the other hand, any remuneration decreases are to be avoided, the wage bill would have to rise. This poses difficult trade-offs, of which the Government is 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 of compensation harmonization. Multiple scenarios are being modeled so that the implications of compensation harmonization (both for the overall 44 Law 657 of 14 July 1965.

71

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 of the new salary structure is almost certain to be an important element of the implementation plan. Appropriate design of 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 (i.e., promotions), which were established within the last six years.45 The draft Law leaves to the State Personnel Presidency the responsibility for regulating these processes. It 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 on 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 of motivation. At best, they will motivate staff to avoid flagrant violation of work rules. Moreover, they will pose the additional problem of creating fiscal costs that cannot be easily controlled. 2.29 A number of options exist for addressing both the low motivation potential of such step increases as well as the difficulty of controlling wage bill costs. These include: (i) requiring that the average magnitude of step increases provided each year to be decided as part of 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 of 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 SIGMA’s July 2004 assessment “the current recruitment and promotion procedure was introduced in 1999, and Government Decree 24744 of 3 May 2002 created a single recruitment centre and set out a recruitment procedure for the entire central administration”. Furthermore, “the Student Selection and Placement Centre (ÖSYM in the Turkish acronym), is attached to the Autonomous Council of 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, of course, be assigned an identical percentage change in its wage bill budget from year 0 to year 1; e.g., the average percentage change in the overall wage bill that would be 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 of 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.

72

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 of 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 of staff; e.g., an organizational unit), and then employ the annual performance ratings (within that organizational unit) to apportion salary increases among the set of staff covered by that wage bill envelope. A specific formula could be used to that effect, subject to the constraint that the sum of 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. 2.31 Personnel performance evaluation procedures: It is notoriously difficult to 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 well as for providing a mechanism for terminating non-performing public servants, it is especially important that careful 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 of spelling out detailed personnel evaluation procedures and criteria. It also quite sensibly requires the Turkish Public Administration Institute (Turkiye ve Orta Dogu Amme İdaresi Enstitusu—TODAIE) to provide an independent assessment of the system established by those detailed regulations. The one issue that would merit more attention here is the process itself—and, in particular, at least four dimensions of that process:

• involving individual staff members in reaching agreements with those who will evaluate them before the beginning of an evaluation period on how their success in meeting the criteria specified in the regulations should be judged – i.e., what sorts of evidence will be relevant,49

• requiring input from multiple persons in distinct positions to assess the performance of any given staff member (e.g., supervisor, peers, clients);

• subjecting the evaluator’s consolidation of those inputs into a single evaluation to review by his/her superior before it is finalized;50 and

47 OECD, Public Governance and Territorial Development Directorate, Public Governance Committee, Human Resources Management Working Party, Performance-Related Pay Policies for Government Employees: Main Trends in OECD Member Countries draft, 8 October, 2004. Web address: http://www.oecd.org/document/39/0,2340,en_2649_37405_33687079_1_1_1_37405,00.html 48 These issues, are also examined in: http://web.worldbank.org/WBSITE/EXTERNAL/TOPICS/EXTPUBLICSECTORANDGOVERNANCE/EXTADMINISTRATIVEANDCIVILSERVICEREFORM/0,,contentMDK:20133441~menuPK:1828910~pagePK:148956~piPK:216618~theSitePK:286367,00.html 49 Art. 111(2) does require each public entity (“institutions and their units”) to establish and announce performance criteria “within the first month of the financial year when the performance evaluation is made”. It 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 of “performance evaluation commissions”, as specified in Art. 111, is one option for addressing this need for review by a third party, since those commissions are required to include five members, at least one of which must represent the union representing the largest fraction of staff within the work unit.

73

• 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 on the fraction of “very good” ratings – i.e., requiring supervisors to rate their staff on 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 on their personnel and how motivated they are, but also on the other inputs with which those personnel work. Inadequate capital facilities, insufficient operating supplies, sub-optimal maintenance of facilities, as well as failure to continuously upgrade the knowledge and skills of personnel will all compromise operational effectiveness and efficiency. As such, it is important that reforms be complemented by steps to ensure that adequate resources are devoted to such complementary inputs. Given tight budgetary resources, this may require reallocation of 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 of CGG gross fixed capital formation by function reveals that public investment allocations in Turkey differ in some respects from the group of comparators (Figure 2.7).52 Public investment for the provision of infrastructure services (included in economic affairs) accounted for 60 percent of total investment in 2003-04 and exceeded marginally the average of comparators, although Turkey’s infrastructure gaps remain large.53 In stock terms, infrastructure projects represent more than 80 percent of the investment portfolio (Table 2.3). The apparently large share of infrastructure investment is very similar 51 This section focuses mostly on the infrastructure sector recognizing the high share of the sector (more than 80 percent) in the existing public investment stock. 52 Due to limited data availability on functional breakdown of 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 of infrastructure services involves three different types of public entities at various tiers of government. Central Government Agencies (CGAs) build major intercity express ways, village/rural 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 port operation; oil and gas distribution; postal services; and most telecommunications services. Local Authorities (LAs) – more precisely municipalities — are the main deliverers of infrastructure at the sub-national level. They are responsible for urban roads, urban public transportation, and solid waste disposal. 45 percent of the total public investment in infrastructure is spent through the budget. The remaining share is spent through SEEs (roughly 20 percent) and LAs (about a third). Both on-budget and off-budget public infrastructure providers are included in the estimation of fiscal targets.

74

to what was observed for infrastructure public investment in Latin America countries particularly during the 1980s, before the introduction of major private participation in infrastructure services but – most importantly – before major fiscal adjustments were undertaken.54 As a percentage of GDP, investment in public goods (general public services and defense, public order and safety) and in education also exceeded that of 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

0.46

1.10

0.490.38

0.49

0.7

1.39

0.02

0.54 0.49

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

General publicservices, Defense,Public order and

safety

Economic affairs Housing andcommunityamenities,

Environmentalprotection

Education Other socialsectors

Average of 11 comparators Turkey

Source: SPO, World Bank Staff calculations

Table 2.3: Infrastructure Investment Portfolio: Value and Number of Projects, 2000-2005

Total Value (Cost) of Portfolio Investment Stock

Infrastructure Non-Infrastructure

Infrastructure Portfolio as a

Share of the Value of Total Portfolio

Public Investment Infrastructure

Infrastructure Stock as a

Share of # Projects

( 000 YTL) (percent) (# Projects) (percent) 2000 71,925,876 14,293,392 83 5,321 1,100 21 2001 118,589,325 24,329,656 83 5,047 1,052 21 2002 138,143,540 28,653,287 83 4,414 899 20 2003 155,384,720 31,725,593 83 3,851 829 22 2004 161,373,229 34,739,639 82 3,555 785 22

2005(*) 169,961,918 36,722,401 82 2,627 742 28 Source: SPO (*) Budget Allocation

2.34 Over the last fifty years, the urban population has grown twice as fast as the overall population and reached 65 percent of the total population in 2000. Half 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 of sufficient finances. Over the same period the number of municipalities roughly doubled from over 1,700 to about 3,200.

54 That was the case of Argentina, Brazil and Bolivia. For details, refer to Easterly and Serven (2003).

75

The rural population (about 35 percent of the total), resides in about 35,000 villages with an average size of 500 persons. 2.35 Meeting financing requirements of infrastructure is thus a key challenge for sustained growth. Key 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 of existing assets, (iv) ensuring adequate security and reliability of supply of infrastructure services; and (v) gradually increasing the role of the private sector in the financing, development and operation of infrastructure in Turkey, with careful design of Government commitments to private operators to minimize the risk of contingent liabilities. (i) Ensuring sufficient allocation and timely completion of high-priority projects 2.36 Public Investment was cut significantly as part of the fiscal adjustment efforts during 2001-2004. Total public investment level (inclusive of local administrations and SOEs investment), stood at 4.2 percent of 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 of both the consolidated budget agencies and SEEs. Investment spending in 2005 increased by close to 1 percent of GNP, mainly due to increased investments by local administrations. Source: SPO, World Bank Staff calculations 2.37 Rationalization of the PIP in 2001 has been quite effective. In 2001, the government initiated a rationalization process, in order to allocate the reduced amounts of 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 of transport, energy and agriculture projects and by reducing the number of multi-year projects. In addition, the program included new feasibility studies for some projects while criteria for inclusion of new projects in the pipeline were tightened, with only those projects which have certainty of funding, external or domestic, being considered for approval. As a result the total cost of projects in the PIP was reduced and average completion time was cut by 32 percent.

Table 2.4. Institutional Breakdown of Public Investments (% of 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 Soc Sec 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 PUBLIC SECTOR 6.6 6.8 6.4 6.3 4.9 4.2 5.1 5.0 1/ Figures include investment worker payments.

76

2.38 Rationalization efforts continued in 2002-2005. The total number of projects declined from over 5000 in 2001 to 2627 in 2005. While 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 of the portfolio. Energy and transportation investments were hit the hardest, since jointly they constitute about 40-50 percent of 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 of new projects entering the portfolio and by increasing the share of investments in the Budget.

Total Multiyear1999 5458 57,126 3,560 4.5% 3,649 4.7 10.2 9.9 1,186 3052000 5321 86,219 5,905 4.7% 6,183 4.9 9.2 8.7 1,176 2492001 2/ 5047 142,919 7,167 4.1% 7,570 4.3 12.5 11.8 1,234 2862002 4414 166,797 10,590 3.9% 12,223 4.4 8.5 7.2 1,066 1282003 3851 187,110 12,464 3.5% 10,386 2.9 7.6 9.7 1,032 1342004 3555 196,113 11,978 2.8% 10,835 2.5 8.1 8.6 1,079 1492005 2627 206,684 15,875 3.3% 14,222 2.9 6.7 7.6 985 1372006 2525 200,391 17,522 3.2% 5.5 1,013 155

1/ Excludes local administrations.2/ Includes supplementary budget allocation of TL 280 trillion and TL 555 trillion from Law 3418 revenues.3/ Average time of completion is calculated as the amount of time required to finish up the remaining stock completely, assuming that no other projects are taken into the investment program in the following years and an appropriation and expenditure is made of the magnitude of the current year level.4/ Based on actual investment.

Source: SPO

Average Time of Completion 4/

New ProjectsInitial Allocation (%

of GNP)

Table 2.5 : Public Sector Investment, 1999-2006 1/(current prices, million YTL)

Number of

Projects

Total Value of Projects

Initial Allocation

Actual Investment

Actual Investment (in % of GNP)

Average Time of Completion 3/

2.39 The brunt of 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 of GNP in 1999-2000 to below 2 percent in 2003-05. This was the main source of adjustment in public investment program, although education and health investments also declined.

1999 2000 2001 2002 2003 2004 2005 2006 2/

Economic Infrastructure 3.1 3.4 2.6 3.0 1.8 1.6 1.8 1.7 Water Resources-Irrigation 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 Drinkable Water,Sanitation 0.3 0.3 0.3 0.2 0.2 0.1 0.1 0.2Social 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 0.2Economic Sector 0.3 0.4 0.4 0.3 0.2 0.2 0.3 0.3Others 0.3 0.4 0.3 0.3 0.2 0.2 0.2 0.3TOTAL 4.7 4.9 4.3 4.4 2.9 2.5 2.9 2.9

2/ Budget allacation.Source: SPO

Table 2.6. Annual Investment Spending (% of GNP) 1/

1/ Excludes local administrations. Includes capital transfers that have public investment attributes and capital expenditure--net of expropriation.

77

2.40 Some of these infrastructure investment cuts were probably overdue, and occurred in the context of the rationalization of 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’ of the PIP. As a result, the stock of approved but unfinished projects grew to an average of more than 5,000 during the latter half of the 1990’s, and the average completion time increased to about 10 years. Many projects received “trace” allocations, i.e. amounts nowhere near enough to implement the project, but assigned merely to keep it in the PIP. While in 2001, one-third of the number of projects with trace allocations were in the health sector, agriculture and education and other social sectors also had a sizeable share in the stock of these “inactive” projects. In terms of the amounts of the projects in trace status, the agriculture, energy and transportation projects comprised a majority of all the projects with trace allocation. (a combined share of 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 of the significant downsizing of the PIP the number of projects put on hold has increased. In 2005, 254 infrastructure projects got allocations of less than 10 percent of their remaining investment, putting on hold a portfolio of a total cost of YTL 54 billion (approximately US$40 billion). Trace allocation projects accounted on average for about 10 percent of total value of projects in 2003-04, a significant rise from the 1999-2000 average of only 3 percent. It is 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 of the Government, or (iii) strategically important projects which are being constrained by fiscal adjustments and therefore have limited access to financing. In the absence of a sustained and in depth review of the portfolio, it is difficult to ascertain the quality of the ongoing investments. Of the current portfolio, transport, irrigation and hydropower have the highest number of stalled or stopped projects. Projects under the regular budget thus appear to have been more prone to be started and contracted upon without certainty of requisite allocations in the following years. As an example, in transport 40 percent of ongoing projects got a trace allocation of less than 10 percent of remaining investment. 55 2.42 In addition, there have been significant delays in the implementation of water supply and sanitation projects. The average completion period of water supply projects implemented was 10.4 years in 2003. The average completion period of water supply projects was 2.9 years and average completion period of wastewater projects was 4.8 years implemented in 2003. Long implementation periods and slow project execution reduce the economic rates-of-return of projects as the illustrative calculation in Table 2.7 below.

55 It would be 25 percent of on-going projects on trace if we take the threshold as 5 percent of the requisite follow-up allocations.

78

Table 2.7. Effect on Economic Returns from Delayed Project Completion

(US$ millions)

Source: World Bank staff (ii) Investment needs and financing requirements 2.43 Infrastructure gaps emerge as an obstacle for doing business. A preliminary review of the current status of infrastructure service provision shows that in order to enable Turkey to emulate OECD standards of service provision, Turkey must systematically improve the quality, and perhaps also the quantity, of infrastructure investments in the medium term. The EBRD-World Bank Business Environment and Enterprise Performance Survey (BEEPS) indicates that the percentage of Turkish firms stating telecommunications, electricity, and transportation as a problem for doing business is significantly higher than the average for EU8 countries and ECA countries. And although the percentage of 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 of the 1,300 firms surveyed experienced an average of 28 power outages in 2005, leading to a loss of more than 4 percent of annual revenues.56 In comparison, only 26 percent of firms in Poland reported having experienced outages, at an average of 1.5 times per firm. The survey confirms that the quality of transport services is poor, compared to other countries at similar levels of development. 2.44 A project-level evaluation of the existing portfolio will be instrumental in improving the quality of the investments. It has not been possible, in the context of the present study, to estimate the infrastructure needs in Turkey because of the varying requirements for (and often, the difficulty in agreeing on) quality of infrastructure service delivery standards across the country. Bridging infrastructure gaps will most likely require additional funding in the future. Quantity of services will need to improve, but quality is 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 is urgently needed to inform policy on the key gaps and policy options to address them. 2.45 The Ninth Development Plan of the Government focuses on improving the quality of infrastructure, ensuring supply security, and increasing competition. Improving the quality of infrastructure in terms of improving reliability and reducing costs, will likely 56 The Bank conducts Investment Climate Surveys (ICS) in many countries on a regular basis, with the objective of assessing the main factors which constrain growth by manufacturing firms. The Turkey ICS covered 1,323 firms representing nearly every industry and ownership structure.

Period of Execution

Annual Investment Costs

Present Value Costs

Annual Benefits Present Value Benefits

Benefit/Cost Ratio

Economic Rate-of-Return

5 years 20 each years 1-5 75 20 each years 6-35

117 1.54 14.7%

10 years 10 each years 1-10

61 20 each years 11-40

73 1.18 11.3%

79

require additional funding, but may also require more adequate maintenance and better operation. Especially in energy, the country is likely to face growing pressures on supply security in the short to medium term, given the high growth in demand. The Government is 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 of 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 of roughly 2.5 percent of GNP for 2006-10. This implies that the Government would have to substantially increase its funding for infrastructure from the current levels. New investment projects also enter the portfolio every year which places an additional burden on 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 of the annual allocation for infrastructure investment. In the year 2005 about 90 percent of 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 is not only a strategic priority unto itself but also a key element for achieving sustainable growth and better quality of life, much needs to be done towards faster convergence with EU. As part of its EU accession process, the cost for Turkey to meet the environmental standards of the European Union, where at least 11 EU Directives in the environmental area of 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$700–900 million annually. The EU-driven estimated annual investment could be as much as 0.45 percent of the country’s GNP.57 When additional investment requirements are taken in to account, for example for alignment with the environmental acquis, effectiveness of investment spending and better project selection and monitoring will be even more urgently needed. In the medium-term program for 2007-09, the government has committed to further reducing the average completion time, which is one of the available indicators of the quality of 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 of the PIP. Structural measures are bound to complement the rationalization of the investment stock, by enhancing management of the public investments.

57 The Under-secretariat of the State Planning Organization believes that EU financial assistance will be necessary to complement the domestic investment funding in order for the EU acquis to be met. The SPO expects that the level of available municipal funding will be reduced due to (i) a reluctance on part of the municipalities to borrow for investments where in the past they have been used to receiving de facto grants; and (ii) the intention of the Debt Law No. 4749 to more firmly ensure that municipalities will be forced to use a portion of internal sources for repayment of foreign loans.

80

2.48 Appropriate allocations for operation and maintenance of public capital need to be ensured. This is an important requirement not only for improving the quality of services, but also for restricting the cost of the projects. Whilst the investment budget is under SPO oversight, maintenance expenditures below a certain threshold, as defined in annual budget laws, are recorded in the current budget, in the “purchases of goods and services” line. Total maintenance expenditures including this line item, declined from 0.5 percent of 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 of public capital to efficiency and growth.59

Figure 2.8. Maintenance Expenditures (% of GNP)

0.00%0.01%0.02%0.03%0.04%0.05%0.06%0.07%0.08%0.09%

Agr

MiningManuf

Energy

Trans

Touris

mCoo

mm

Housing

Educat

ionHeal

thOPS

1999-20002003-2004

Source: Ministry of Finance, SPO, World Bank Staff calculations 2.49 Inappropriate maintenance and accelerated decay of capital goods may unduly burden the investment program in the future. The existing structure of budget preparation—with separate guidelines for current and investment expenditures—poses a risk of misallocation of resources. Agencies may not necessarily strategically allocate resources between investing in new assets and maintenance of existing assets. The importance of timely maintenance is more prominent in the infrastructure sector and thus proper accounting of the actual amounts spent for maintenance is needed. For example, for a road with a 20-year life, GD of State Highways (KGM) research shows that if maintenance is not done at the end of the 12th year, the road would start deteriorating at a rate 8 times faster than in the first 58 The total maintenance expenditure is estimated as the sum of the small amounts of expenditure recorded in the recurrent budget and the expenditure that is part of the investment budget. There is a potential drawback for using the latter series. It is obtained by extracting from the investment program database those projects having as part of its title or characteristics, words like maintenance, repair, renovation. However, other projects that do involve a maintenance component are not picked up by this method if the names or characteristics specified in the database do not involve the relevant words. General Directorate of State Highways is reported to have many projects involving maintenance work. As more reliable and refined data becomes available, the data could be updated. 59 Evidence of lower than expected capacity utilization and shorter than planed economic life of some projects points to problematic maintenance, especially in the road network, where indicators of physical condition point to deteriorating quality.. However, the relative roles of inadequate budget allocations and lack of appropriate incentives/culture for maintenance seems difficult to establish.

81

years of its lifetime. This, in practical terms, means that if the cost of capital is the same for maintenance as it 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 is 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 is 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 is unlikely that the Government will be able to achieve its objectives of 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 is robust and predictable. In the absence of such an environment, private investment will either not materialize in adequate quantity, or will require significant guarantees and assurances. This is borne out in the energy sector where, over the last decade, private investment now makes up about 50 percent of electricity generation, but this has been made possible only by significant contingent liabilities on the Government in the form of guarantees and off take agreements. 2.52 One way of increasing operating efficiency could be to use the private sector through Public-Private Partnerships (PPP). Although there is no single PPP prescribed model, the development of 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 of 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 of forcing the government contractual party to carefully 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 of the political risk should be borne by the private partner. 2.53 PPPs can be implemented well if 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 of stakeholders, information sharing at all levels and maximum focus on institutional strengthening of the respective entities requiring training of civil servants on the appropriate frameworks and design of PPPs

82

as well as adequate support to 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 of the PPP. The issue is who controls the majority shares, and thus controls the decision-making process (through the PPP Board). In most PPPs, the contract design is obviously central to whether private management will deliver benefits. Three aspects are crucial: the need to identify the extent of control exercised by private managers, the need for third-party contract oversight, and (perhaps most importantly) the need to link remuneration of 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, often 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 of three to five years. The public authority retains financial responsibility for the service, thus limiting the risk for the contractor. Remuneration for the contractor is in the form of a flat fee with the possibility of additional performance-related payments. Using management contracts in utility commercialization is a relatively new phenomenon and some evidence is emerging about the types of 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 is interested. 2.55 Based on assessments of infrastructure and the country's investments climate, there is significant scope for PPP for all of the infrastructure sub sectors - energy, water and wastewater companies and rail freight companies in Turkey. In key 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 is fully aware of its internal market liberalization challenges and fully understands the key role of integrating with international markets. As 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 is in the process of establishing a competitive market which will enable suppliers and buyers of electricity to participate efficiently and at economic prices. In 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 of building new generation capacity. At present, despite the existence of an independent regulatory authority, in order to keep inflationary pressures down, the Government artificially keeps electricity prices low by reducing electricity generation using natural gas (the price of which is very high due to record high international oil prices) and using its substantial hydroelectric resources instead (hydroelectricity is not priced at the true opportunity cost of the use of water). This has the impact of keeping the wholesale market price at a low level, which provides a disincentive to potential investors in electricity

83

generation. If the Government were to let the regulatory framework operate independently, and were to allow prices to reflect economics, it would be possible to attract private investment without significant guarantees. 2.57 In a longer-term perspective, the Government is conscious of the need to ensure sustainability in medium-term infrastructure investment allocations by addressing the past deficiencies of the budgetary and investment planning process. In the recent past, volatility of annual investment allocations has hampered implementation in line with strategic priorities, and has increased total costs and average completion times of 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 of the annual process of 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 of 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 of 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 of the rationalization policies and criteria, while establishing a proper prioritization of 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 of infrastructure allocations available for new and high-return projects. In summary, it 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 of projects and improved incentives and pricing/subsidy policies are necessary for enhancing efficiency. The training of personnel is also fundamental to promoting efficiency. In some instances, the focus should be on utilizing existing assets optimally rather than adding new assets.

2.59 The absence of well defined allocations for operation and maintenance expenditure in the investment budget is a problem. It is not clear if, and if so at which stage, there is a strategic link between the funding and planning for investment and operation and maintenance (O&M). There is a clear split of responsibilities between SPO and Ministry of Finance during the Budget process for investment and O&M. Since there is no 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.

84

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 of the process when investment resource requests are invited from public agencies. Progress is noted with the issuance of budget preparation notices in which priorities are mentioned more clearly than in the past. This has contributed to improving the realism of proposals, but further efforts would be appropriate to protect priority projects and help strengthen the overall credibility of 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 of GDP annually in infrastructure investments. Due to the overarching requirement to preserve fiscal sustainability, the level of investments that can be made is constrained not so much by the availability of financing but by the need to meet the primary balance targets in the face of 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 low allocations provided to the projects. As explained in chapter 1, these low investment allocations have served so far to mainly offset a soaring deficit of the social protection system that reflects slow or pending reform. The challenge for the Government is thus not the availability of this extent of financing for infrastructure, but the ability to use the financing that is already available. Creating the needed fiscal space, through the structural expenditure reforms reviewed in chapter 3, would thus be the key requirement for stepped up investment in infrastructure in the future. Well designed Private-Public Partnerships for the provision of infrastructure services will also help alleviate existing constraints.

85

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 of tax expenditures is a concern for fiscal accountability and the transparency of governance. A basic requirement of 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 of a major share of the discretionary expenditures of a government. Without the inclusion of 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 of a range of incentives and preferences offered to taxpayers through the provisions of the tax laws. Without their inclusion in the budget process, tax expenditures do not get the same level of strict budget scrutiny as normal budget expenditures. Accordingly, to make the budget process more complete and subject all expenditures to the same level of transparency and budget scrutiny, governments in OECD countries over recent decades have been estimating the revenue cost of 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 of 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 of the ability of persons to pay the tax, and the economic, administrative and compliance costs of the tax. 2.64 Establishing tax benchmarks in Turkey is 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 of employment in the agricultural sector (about 29.5 percent as indicated in the Ninth Development Plan); and because of extensive informality in agriculture and services, it has a large share employed in the informal and unrecorded sectors (about 53 percent of total employment). There is also a large share of employment at the minimum wage. Data on employment earnings from the social insurance system for 2003, for example, show that 59 percent of employees in the private sector were reported to receive minimum wages while 4 percent of public sector workers received minimum wages. Even if there is significant underreporting of employment earnings for social insurance purposes, it would still leave a significant share of low-wage workers.

86

2.65 The peculiar economic structure, together with the past economic instability, has led to specific choices of tax structures in Turkey. These include:

(i) reliance on indirect taxation through the VAT (with limited refunds of input taxes) and Special Consumption Taxes (excise taxes) in contrast to most high-income OECD countries that receive a majority of tax revenues from direct taxes on income and payrolls;

(ii) significant reliance on withholding taxes in the income tax on financial instrument income, , self employment gross earnings, and wages and salaries, including final withholding taxation of interest deposits and employment income;

(iii)heavy reliance on employers as major tax agents for administering the taxes of employees including PayAsYouEarn income taxes;

(iv) special credits for expenditures on health, education, basic food and rent, social security premiums, and pension contribution deductions;

(v) exemption of many self-employed, rural, household and agricultural workers based on occupations rather than levels of income;

(vi) a complex set of inflation adjustments, investment incentives and offsetting taxes on these adjustments and incentives that reduced the tax burden on 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 economic performance. This lack of information about the tax bases weakens the ability of the authorities to formulate tax and economic policy, and, in the context of tax expenditures, creates challenges for defining tax benchmarks and estimating the tax expenditures arising

Box 2.1: Guidelines to Identify Normative Tax Structure Identification of the normative tax structure can best proceed by determining whether a given provision responds to one of the following questions: 1) Is the provision necessary to determine the base of the tax, normatively defined, in

accordance with the fundamental nature of the tax? 2) Is the provision part of the generally applicable rate structure? 3) Is the provision necessary to define the taxable units liable for the tax? 4) Is the provision necessary to assure that the tax is determined within the time period

selected for imposition of the tax? 5) Is the provision necessary to implement the tax in international transactions? 6) Is the provision necessary to administer the tax? Seven developed countries (Canada, Germany, France, the Netherlands, Sweden, United Kingdom and United States) use these broad guidelines to identify their normative tax structures and tax expenditures. Source: . McDaniel, P.R, and S. Surrey, “International Aspects of Tax Expenditures: A comparative Study,” Kluwer, 1984, and Messere, Ken C. “Tax Policy in OECD Countries, Choices and Conflicts,” IBFD Publications BV, Amsterdam, the Netherlands, 1993

87

from the various tax incentives and the tax assistance provided to various disadvantaged groups. 2.67 Aware of 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 of the tax provisions of 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 of the tax benchmark or be regarded as tax expenditures. The major features of the tax benchmarks for each of the tax types and the draft tax benchmarks are presented in Annex V. 2.68 Based on the draft tax benchmarks, out of the 290 tax provisions, initially some 186 items were identified as potential tax expenditures. Currently, the authorities have identified a list of 84 tax provisions as tax expenditures and included this list in the 2006 Budget Law. The most significant of 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 of tax, the list of tax expenditures is outlined in Box 2.2. Some of the major tax expenditures are granted with the aim of promoting: investment, employment, regional development, R&D, and trade incentives primarily delivered through the CIT. A range of tax expenditures aims to assist low-income individuals, in support of social policy. These are delivered primarily through the PIT and CIT and to a lesser extent through the VAT (see Box 2.2). Some major tax expenditures in this latter group include the deductibility of 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. Box 2.2: Tax expenditures by type of tax I. Personal and corporate income tax Assistance to disadvantaged and low-income persons

Tax credit for basic needs expenditures (VAT-invoiced expenditures on education, health, basic foods, clothing and rent) by employees; Exemption of income small self-employed; professional income of writers, translators, sculptors, painters, computer programmers, composers and inventors; non-diplomatic staff of embassies; wages of workers in small villages; wages of farm workers; wages paid to students, convicted and detainees and indigent people working at workshops of school of 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 Preferential treatment of certain employee benefits of civil servants and other workers

Exemption for aid for heating purchases by civil servants

88

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 of apprentices Social insurance and pension contributions, investment income and benefits

Deductibility of contributions to social insurance, pension and unemployment schemes Deductibility of 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 of the annual minimum wage. In addition, tax-deductible premiums are permitted to individual purchases of non-pension non-property insurance (life, disability, health, etc insurance) up to 5% of earnings. Tax treatment of private pension savings: exemption of 25% of pension income from pension savings investment for 10 or more years Exemption of social security pensions, death, disability, illness and unemployment compensation and aid. Exempt investment income in pension plans, except that a 10% capital gains tax is charged on 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 of sportsmen and donations to Izmir University Games Preparation and Regulation Board Exemption of enterprises of sub-national government in agricultural, water, electricity, transportation, cold storage and slaughterhouse businesses Exemption of organizations involved in development of industrial zones. Preferential treatment of income on selected financial 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 of 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 of listed companies held for less than 3 months and amounting to more than YTL 10,000 (vii) Capital gains on unlisted shares held for less than one year. (viii) Inflation-adjusted capital gains on immovable property. Real estate held for more than 4 years is exempt. Investment incentives

40% tax allowance on depreciable assets in addition to regular depreciation allowance; indexed and indefinite carry forward of unused allowances, but as of 2006 carry-forwards of allowances that cannot be deducted will be limited to 3 years. 40% tax allowance on R&D expenditures in new technology and know-how 60 State-sponsored social security pensions require contributions of 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 of a senior civil servant. Employees of 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.

89

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 of up to 100% from PIT and Social Security premiums on minimum wage for additional workers Technology Development Regions

Tax exemption from income from R&D activities of businesses in region for 5 (extendable to 10) years Employment tax exemption for R&D employees Free Trade Zones (wage and profits taxes)

Prior to 2004, tax holiday for corporate and wage income taxes Since 2004, phasing out of old licenses by 2008 and only granting CIT exemption to income from export production. Earnings received from construction, repair, assembly works carried out and technical services rendered abroad and transferred to general accounts in Turkey II. VAT Full exceptions of specified mining and petroleum exploration3 Partial exemptions of deliveries, operation, construction and servicing of various craft; supplies and services in the construction of ports; servicing of craft at ports3 Partial exemptions of imports for military purposes; presidential use; for donations for health, education and social purposes3 1% VAT rate and 8% rate on items other than those eligible for reduced rates according to EU VAT directives(EU Directive on VAT 77/388/EEC, Annex H), including textiles, clothing and leather articles rebated to 8% Exempt raw minerals Exempt services provided in a Free Trade Zone. Exempt sale of assets in case of winding up of financial institutions III. 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 of imports for military purposes; presidential use; for donations for health, education and social purposes Note: 1. In addition to tax expenditure estimates, it is also useful to provide tax revenue change estimates as memorandum items for various structural items even where considered part of tax benchmark. Examples of memorandum items that are useful to present for information purposes would be the tax revenue losses from all reduced VAT rates, and dividend deductions. 2. Given significant changes in tax policy starting in 2001, the tax expenditures noted above focus on the current tax expenditures, and omit tax expenditures removed or phased out in recent years. 3. “Full exceptions” are those for which input VAT is creditable or refundable (also referred to as “zero VAT rating” in some VAT structures); “Partial exemptions” are those for which input VAT is not creditable (the equivalent of “VAT exempt items” in other VAT structures). Source: World Bank Staff

90

(ii) The Size of Tax Expenditures 2.70 The estimation of tax expenditures is still in its early stage in Turkey. The estimation of tax expenditures is highly demanding, more so than revenue forecasting. Estimating the incidence of tax expenditures on tax revenues requires detailed analysis of the individual tax measures. This requires detailed information on the tax base, including that related to exempt persons and activities. Existing tax databases are still very incomplete because of the limited tax return information. The current status of tax databases and models is 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 of 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 of assessing the size of tax expenditures. 2.71 Authorities estimated tax expenditures for 15 out of 84 tax expenditure provisions for 2006-08 as part of the Budget Law largely based on 2003 databases and models. The estimates for these 15 tax expenditure provisions are presented in Panel A of Table 2.8. The combined effect of these tax expenditures in 2003 is a revenue loss of 1.5 percent of GDP. The largest tax expenditure, at 0.66 percent of GDP, arises from the investment tax allowance, and the second largest, at 0.33 percent of GDP, is from the PIT tax credit for personal expenditures on basic needs items. The PIT and CIT tax expenditures are based on micro-simulation models which capture the distribution of marginal tax rates across individuals and the partial value of deductions and exemptions arising from a lack of taxable income due to losses in the year or carry forwards of losses from prior years as discussed in detail in Annex VI.61 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 of the “other deductions” taken under the CIT and the revenue cost of deductions of social security contributions from taxable income in 2003.62 In addition, it is recognized that loss carry forwards arise both from adverse market circumstances and from unused tax deductions or exemptions in prior years. The elimination of tax expenditures, therefore, would result in lower loss carry forwards into 2003, and hence, higher revenue collections. Here it is assumed that 25 percent of the loss carry forwards are caused by unused preferences in prior years that would result tax increases in 2003 to gain a rough approximation of the tax cost. The need for further work to estimate the impact of prior year tax preferences on current year tax expenditures is discussed in Annex VI. Panel B tax expenditure estimates add a further 1.66 percent of GDP to the total tax expenditures in 2003.

61 In fact, comparison of these estimates, which reflect partial capture of the tax value of deductions and exemptions, with those that would arise using the data from the summary company tax returns for 2003 show the CIT tax expenditures are some 28 percent smaller. 62 “Other deductions” include all the deductions and exemptions allowed that are not specifically itemized on the tax return. These include the deductions allowed under Article 14 for donations or aid provided to educational institutions for health facilities, or to support sports, culture and national heritage.

91

Table 2.8: Selected tax expenditures for Turkey in 2003

Panel A: Authorities estimates YTL millions % GDP

Corporate Tax Law No.5422 C.T.L. Art. 8/4 Exemptions for certain investment income earned by corporations

from mutual funds. 935.03 0.26%

Art. 8/5 Exemption for sale of preference rights certificates and issuance premiums

20.83 0.01%

Art. 8/7 Exemption for gains earned in construction, repair, installation works and technical services in other countries and transferred to general accounts in Turkey.

71.04 0.02%

Art. 8/12 Exception for gains arising from the sale of participation shares and real estates that companies have had under their assets and decide to add to their capital (revised form of C.T.L. Provisional Article 28)

259.73 0.07%

Art. 19 of IT 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%

Income Tax Law No.193

I.T.L. Art. 89/1 Deduction for individual insurance premiums and contributions to individual pension system

8.48 0.00%

I.T.L. Art. 89/4 Deduction for charitable donations and grants. 2.77 0.00%

I.T.L. Art. 121-bis Tax credit for expenditures on education, health, food, rent, and clothing of the taxpayer, spouse and children provided against taxes on wage income

1,201.00 0.33%

4325 S.K. Art. 3 Exemptions provided for income for employers that employ ten or more workers in the provinces indicated in the Law

30.27 0.01%

Value Added Tax Law No. 3065

VAT Art. 13 Exception for vehicles, search of precious mines and petroleum and national security expenditures and investments (subparagraph (d) covered under the basic taxing system)

166.13 0.05%

VAT Provisional Art. 15 Exception for construction contracts made with housing construction cooperatives and social security 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 prior to 29/7/1998

123.72 0.03%

Total A 5,400

1.51%

Panel B: Additional estimates based on 2003 returns

Other CIT deductions 1,773.86 0.49%

Deduction of social security contributions for pensions, unemployment insurance and health benefits

3,300.00 0.92%

Preferences in loss carry forwards absorbed in year in CIT and PIT 915.29 0.25%

Total B 5,989.15

1.66%

92

Panel C: Approximate estimates for selected other tax expenditure provisions

Exemption for aid for fuel purchases by civil servants and certain self employed workers; 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 of apprentices

0.50%

Total A + B + C 3.67%

Notes: In Panel A, PIT and CIT are estimates by authorities, and VAT estimates are based on estimates by authorities 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% of loss carry forwards being tax preference carry forwards. Social security data base used to obtain effective marginal tax rate for social security deductions.

Panel C based on IMF and Bank staff estimates

2.73 Including part of 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 of preferences for employee benefits for civil servants and other workers. The combined estimate of tax expenditures in Table 2.8 for 2003 amounts to 3.67 percent of GDP. While most of the major tax expenditures are included, no estimates are included on tax expenditures for exemptions of income from a wide range of trade, service and farm activities, as well as most of the VAT and SCT tax expenditures. Tax expenditures in OECD countries have fallen in the range of 4 percent to 35 percent of gross expenditures.63 It appears that a comprehensive assessment of tax expenditures would place Turkey towards the high end of the OECD country range. (iii) Refining the framework for the analysis of tax expenditures—key steps ahead 2.74 A major effort is required to upgrade the tax databases and related tax simulation models to obtain more accurate and reliable tax expenditure estimates. Annex VI details the current status of 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 budgeting framework to achieve fiscal accountability. Work on steps a through d is largely complete.

a. Development of benchmark taxes for each tax type and identification of tax expenditures in each tax type

b. Sector/functional classification of tax expenditures

63 Zicheng Li Swift, “Managing the Effects of Tax Expenditures on National Budgets,” Tax Notes International, 41(10), March 13, 2006 reports tax expenditures as shares of gross expenditures for Canada, Netherlands, US, and UK in Table 1.

93

c. Estimation of some major tax expenditures as illustrations of size and importance of tax expenditures.

d. Development of tax expenditure discussion documents e. Development of tax expenditure estimation methodologies for all tax types and

tax expenditures f. Choice of year to start publication of tax expenditure accounts g. Development of tax expenditure projection procedures for multi-year forward

budgets h. Discussion and decisions about how tax expenditures will enter budgeting system i. Integration of tax expenditures into comprehensive budget allocation process

2.76 Tax expenditures should become an integral part of 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 further outlines the steps in development of tax expenditure accounts and integration of 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 is described above, is widely used in the EU for fiscal accountability and transparency. In particular, the EU has adopted “State Aid Rules”64 and a “Code of Conduct”65 for its member states. The “State Aid Rules” restrict or prohibit state assistance to industry, the scope of which is broad enough to cover many types of tax expenditures. The “Code of Conduct” requires member states to eliminate 66 identified special tax incentives in order to refrain from certain types of tax competition. Elaborating a full-fledged tax expenditure framework would help Turkey conform to the EU’s “State Aid Rules” and “Code of Conduct” as part of the EU accession process.

64 EC Treaty, Arts. 87-89. See Schon, “Taxation and State Aid Law in the European Union”, 36 Common Market L. Rev. 911 (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. COM (97) 495 final.

94

CHAPTER III 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 the fiscal 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 is little room for Turkey to further increase overall expenditure in order to meet pressing development challenges. Policy would thus need to focus on reforms that improve the efficiency and productivity of 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 of 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 of civil service pay and employment system, the rationalization of the investment program, and the integration of 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 it 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 of this chapter is on structural expenditure reforms in education, social protection (pensions), and health care. These three sectors accounted together for 58.5 percent of primary expenditures of the Consolidated General Government (CGG) in 2004, the equivalent of 19.2 percent of GDP. The chapter also reviews rural public expenditures, which is the most important component of public expenditures on economic affairs, accounting for almost 2 percent of GDP. These four sectors together accounted for about 65 percent of primary CGG 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 of 7-14 year-olds from 80 percent to 90 percent. The

95

Government has been making tremendous efforts to provide good-quality schools for the poor, especially by means of 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 of public spending on primary education has also improved significantly since 1997. A new curriculum under development by the Ministry of National Education's Board of Education represents a great leap forward for comprehensively updating teaching standards. Government initiatives have succeeded in bringing significant amounts of 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 of EU members. EU members increasingly understand that economic growth depends on having higher skills in the workforce and that improving education levels is the best way to acquire these skills. The EU has thus called on its 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 of the labor market.

A1. Educational Outcomes66 3.5 Despite improvements in the area of primary education, there is an important gap between Turkey and the EU in the proportion of young people that have a complete upper secondary school education. While there has been important progress during the last five years on this indicator, marked by a 12 percentage point gain, only 39 percent of its 20-24 year old population in 2005 had a secondary school diploma, compared with 78 percent of this population in the European Union (Figure 3.1). The EU's "Lisbon Agenda" target of 85 percent of 22-year-olds with a complete secondary education is a long way off. Based on rates of historical progress, it will take Turkey several decades to catch up with today’s EU levels, which are of course themselves climbing upwards. In addition, in several regions of the country, Turkey’s outcomes are considerably worse. For example, in Diyarbakır, a Southeastern province, 28 percent of 20-24 year olds have a secondary education diploma, and only 17 percent of girls in this age group do. Results in Diyarbakır are similar to those in some of the other Eastern and Southeastern Provinces. 66 The World Bank report “Education Sector Study: Sustainable Pathways to an Effective, Equitable, and Efficient Education System,” published in 2005, provides in depth information on the education system, its performance and outcomes, and recommendations for improvements.

96

Figure 3.1: Percent of 20-24 year-olds with a secondary education diploma

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 of its counterparts on student learning achievement as well.67 The PISA study published in December 2004 shows that 52.3 percent of 15 year-olds in Turkey cannot achieve beyond level 1 on a six point scale of learning proficiency (Figure 3.2). This is compared to only 16.6 percent of 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 of 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 of its students at a time when the labor markets in Europe and Turkey are demanding increasingly better qualified graduates.

67 For example, Trends in International Mathematics and Science (TIMSS), Progress in International Reading Literacy Study (PIRLS) and, more recently, the International Student Assessment Program (PISA). One of the most reliable international measures used by many advanced and middle-income countries to assess the competencies of the emerging labor force is the OECD's PISA (Program of International Student Assessment). The PISA instrument measures what 15-year-olds (students near the end of their compulsory schooling) know and are able to do.

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

1995 2000 2005

EU 15 Total

EU 15 Female

Ankara Total

Ankara Female

Turkey Total

Turkey Female

Kars Total

Kars Females

Diyarbakir Total

Diyarbakir Female

Turkey (total)

EU 15(total)

97

Figure 3.2. Distribution of Students by PISA Proficiency Level in Turkey and the EU, 2003

Proficiency

6,005,004,003,002,001,00,00

Per

cent

of S

tude

nt

30

20

10

0

Turkey

New Members

Members

OECD

Source: PISA, 2003. 3.7 There are a number of reasons for the still (relative to OECD) poor performance of such a large share of Turkey's students.

• Weak incentives and organizational deficiencies: Compared with the education systems in most of the countries in the world, Turkey’s education system is highly centralized and its schools are given very little autonomy over their resources, staff deployment, textbook selection, allocation of instructional time, and selection of 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 of nearly all of the financial and human resources in the sector, has not succeeded in ensuring equity across schools, a proper function of 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 of 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 of resources and the least experienced teachers. Nonetheless, there is

98

no explicit strategy at the central level to improve poorly performing schools or help poorly performing students.

• 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 of learning outcomes needed by today's graduates. The Education Sector Study states that teacher preparation and professional support programs are out of date, disjointed, poorly evaluated, discontinuous, and are not based on best global practice and evaluated experience.

• Inefficient exam screening: A substantial portion of 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 of the structure of the examination—a relatively short test composed only of multiple choice items, and its limited curriculum coverage—it tends to discourage the learning of 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 of private resources and provides little value in terms of raising human capital and the educational qualifications of the future work force.

A2. Structure of 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 of 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 of GNP will be spent on education in the 2006 budget—still less than 4.33 percent of GNP spent in 2002. The reduction in public

68 General budget allocations to Ministry of National Education (MONE) have been adjusted to address misclassification of expenditures due to problems in the coding system. For example, adjustments were made to eliminate health expenditures of MONE and related institutions which are recorded as education expenditures. Adjustments have also been made to eliminate transfer items such as transfers to General Directory and Dormitory that have been transferred to MONE budget from the Ministry of Finance budget since 2004. Using unadjusted expenditure allocations from the general budget to MONE leads to an overestimation of education expenditures of about 8 percent, with average expenditure in 2002-2006 at 4.5 percent of GDP, compared to 4.18 percent estimated in Table 3.1. It 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 of expenditures in 2002-2006 estimated at 4.16 percent of GDP.

99

expenditures over 1999-2006 is evenly distributed between expenditures on personnel and public investment, while expenditures and transfers have increased (Table 3.2). By contrast, expenditures on education grew significantly between 1997 and 1999. This was the period when public spending was driven by the policy priority of the then-enacted 8 year compulsory schooling. The increase in expenditures over this period is entirely attributable to greater spending on personnel (Table 3.2).

Table 3.1. Overall Public Education Expenditures by Institution (1997-2006)

(million YTL) 1997 1998 1999 2000 2001 2002 2003 2004 2005 (E) 2006 (P)Ministry of National Education 698 1.476 2.330 3.561 5.118 7.614 10.009 11.184 12.870 14.393

General Budget 620 1.391 2.252 3.302 4.845 6.927 9.847 10.954 12.610 14.105Revolving Funds 9 19 29 62 85 131 162 230 260 289Extra Budgetary Funds (3418 and 3308) 67 61 39 177 155 509In Kind Foreign Credit 3 5 10 20 32 47

Universities 270 487 808 1.159 1.585 2.872 3.438 3.855 4.050 4.749Annexed Budget 233 406 679 931 1.240 2.331 2.778 3.171 3.690 4.365Revolving Funds 11 34 52 93 160 240 363 338 361 384Student Social Service Account 22 38 61 107 140 235 296 345In Kind Foreign Credit 5 8 16 27 45 66

Other Central Government Agencies 116 188 290 502 736 1.160 1.333 1.887 2.350 2.738Local Authorities 24 47 77 132 167 258 276 348 698 717General Government Total 1.108 2.198 3.506 5.354 7.606 11.903 15.056 17.275 19.969 22.598

Ministry of National Education 2,38 2,76 2,98 2,84 2,90 2,77 2,81 2,61 2,65 2,67General Budget 2,11 2,60 2,88 2,63 2,75 2,52 2,76 2,55 2,59 2,61Revolving Funds 0,03 0,03 0,04 0,05 0,05 0,05 0,05 0,05 0,05 0,05Extra Budgetary Funds (3418 and 3308) 0,23 0,11 0,05 0,14 0,09 0,18In Kind Foreign Creit 0,01 0,01 0,01 0,02 0,02 0,02

Universities 0,92 0,91 1,03 0,92 0,90 1,04 0,96 0,90 0,83 0,88Annexed Budget 0,79 0,76 0,87 0,74 0,70 0,85 0,78 0,74 0,76 0,81Revolving Funds 0,04 0,06 0,07 0,07 0,09 0,09 0,10 0,08 0,07 0,07Student Social Service Account 0,07 0,07 0,08 0,09 0,08 0,09 0,08 0,08In Kind Foreign Creit 0,02 0,02 0,02 0,02 0,03 0,02

Other Central Government Agencies 0,39 0,35 0,37 0,40 0,42 0,42 0,37 0,44 0,48 0,51Local Authorities 0,08 0,09 0,10 0,10 0,09 0,09 0,08 0,08 0,14 0,13General Government Total 3,77 4,11 4,48 4,26 4,31 4,33 4,22 4,03 4,11 4,19

As a Share of GNP (%)

Source: World Bank staff calculations based on MONE and Emil, Social Expenditure Study, November 2005 (background study), Yilmaz (2006)

3.9 Fluctuations in public expenditures may also reflect the weak linkages between policies, planning, and the budget. This was the main deficiency of 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 of 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 of the effective mechanisms leading to greater linkages between policies and financial resources, there could be a massive mismatch between what is committed from the government and what is affordable regarding budget constraints.

100

Table 3.2. Overall Public Education Expenditures by Economic Classification (1997-2006)

1997 1998 1999 2000 2001 2002 2003 2004 2005 (E) 2006 (P)Personnel 2,42 2,68 3,09 2,79 2,89 2,86 2,87 2,73 2,73 2,82

MoNE 1,80 2,06 2,45 2,15 2,22 2,19 2,24 2,16 2,15 2,18Universities 0,53 0,53 0,57 0,54 0,55 0,54 0,55 0,49 0,50 0,53Other Ins. 0,09 0,08 0,08 0,10 0,12 0,13 0,08 0,07 0,08 0,10

Other Current 0,36 0,47 0,38 0,38 0,43 0,47 0,47 0,53 0,48 0,46MoNE 0,13 0,14 0,14 0,14 0,13 0,14 0,17 0,19 0,19 0,17Universities 0,09 0,10 0,11 0,11 0,11 0,17 0,17 0,16 0,10 0,11Other Ins. 0,14 0,23 0,13 0,13 0,19 0,16 0,12 0,18 0,18 0,19

Investment (Capital) 0,75 0,68 0,70 0,86 0,73 0,69 0,54 0,46 0,54 0,52MoNE 0,44 0,47 0,42 0,51 0,43 0,33 0,31 0,22 0,24 0,23Universities 0,14 0,13 0,12 0,15 0,14 0,17 0,10 0,13 0,13 0,14Other Ins. 0,17 0,08 0,17 0,20 0,16 0,19 0,13 0,12 0,17 0,16

Transfer 0,25 0,27 0,30 0,23 0,25 0,31 0,34 0,31 0,36 0,38MoNE 0,07 0,10 0,09 0,10 0,11 0,08 0,07 0,06 0,07 0,09Universities 0,10 0,10 0,12 0,04 0,04 0,12 0,11 0,09 0,10 0,10Other Ins. 0,07 0,07 0,09 0,10 0,11 0,11 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 Source: World Bank Staff calculations based on MONE, Yilmaz (2006)

(b) Financing of Education Expenditures

3.10 The fragmentation of financing sources has been reduced. The analysis of public expenditures on education in Turkey presented in the World Bank's 2001 Public Expenditure and Institutional Review (PEIR) highlighted the fragmentation in sources of finance as a key feature of the sector. These sources of education finance included general budget appropriations to the Ministry of National Education (MONE) and to other central government agencies providing educational services, various budgetary funds (Law 4306 for Basic Education, Law 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 of 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 of the new Public Financial and Management and Control Law (PFMC), educational institutions are now classified under the central budget. In the light of new definition of the general government sector in Law 5018, the percentage distribution of public education expenditures is given in Table 3.3 and further explained in Boxes 3.1 and 3.2.69

69 Law 5018 defines general government as the total of 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 of central government in this study also covers extra budgetary funds and revolving funds. This suggests more comprehensive coverage of general government sector than law 5018. However, this definition does not cover foundations and PTAs (parent teacher association) and other communal aid.

101

Table 3.3: Public Education Expenditures in Turkey by Institution (percentage shares in total)

(million YTL) MONE Universities Others Total General 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 12,9 100,0 2005 64,5 20,3 15,3 100,0

2006 63,7 21,0 15,3 100,0 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 of 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 of Health, the General Directorate of the Police, etc.). The total share of the general government revenues in total education expenditures is approximately 98 percent, including the revolving funds (discussed below) and extra budgetary funds. Extra Budgetary Funds – the Social Solidarity and Support Fund Under the current framework, only the Social Solidarity and Support Fund provides extra budgetary funds to the education sector. This fund provides free school books and financial assistance to the children from poor families. The education related expenditures of 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 of the revolving fund revenues through the sales of 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 of 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 of local authorities in Turkey providing funding to the education sector, namely municipalities and Special Provincial Administrations (SPAs). Law 222, which regulates local governments, establishes that Provinces must spend at least 20 percent of their budget on education, especially at the primary education level.

102

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 of 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 of MONE. A flow chart of the expenditure procedure under SPA involvement case:

3.11 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 of budget fragmentation due to off budget activities and a lack of 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 now accessible for analysis. In addition to this improvement, in 2002-2003 the Turkish Statistical Institute (TURKSTAT – formerly entitled the State Institution of Statistics) conducted institutional

Office of the Chief Financial Officer of the

Province

Releases appropriations sends payment orders and cash

Provincial Office of MONE

Informs the appropriation to

SPA Procurement Unit*

• issues call for tenders, • forms a tender commission (with the

participation of the MONE representatives),

• awards contract

Payments order signed by governor, cash transferred to SPA bank account by the accounting officer (muhasebe muduru)

Provincial Unit of Civil Works • Controls the work of the contractors • Checks the work completion reports

for payments (hakedisler) calculated by the contractor

• Sends completion report and payment order to the SPA

MONE Budget Office

(Center)

103

surveys aimed at collecting detailed educational expenditures and revenues from educational institutions and other providers.70 All of these surveys, other than the school survey and the crèche survey, collected information from the complete universe of such educational institutions in Turkey, and therefore the data collected represented total spending on education in Turkey by each of these institutions.71

3.12 Including private financing, Turkey spends on education the equivalent of 7 percent of GDP. The 2005 World Bank Education Sector Study included a detailed analysis of the flow of 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 of 6.97 percent of GDP.73 Of 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 of total education expenditures, while 2.3 percent of expenditures could not be unequivocally accounted as public or private. As noted earlier, public expenditures on education equaled 4.34 percent of GDP in 2002, with private and other expenditures on education accounting for 2.63 percent of GDP. The vast majority of household spending (90.4 percent) came from own funds, 3.5 percent came from central government revenues in the form of scholarships, 1.4 percent came from private foundations, and 0.2 percent from municipalities in the form of 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 of 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 of private crèches; (d) the Ministry of National Education; (e) the Higher Education Board, the Council of Higher Education (YOK); and (e) all other ministries and public agencies such as the Ministry of Finance, Ministry of Health, Ministry of 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 of 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.

104

Source: World Bank Staff calculations

(c) Providers of Education Services

3.13 Primary and higher education absorb the largest shares of expenditures. Of the total spending on education, 89.5 percent went to the providers of educational services, that is, the public schools and universities (including public providers of informal education), and the balance of 10.5 percent went to private schools and (foundation) universities (Figure 3.4).

• The largest share of education spending went to primary education schools, which received 40.3 percent of the total spending on education. Of the amount spent on primary schools, 95 percent went to public primary schools, and the rest went to private primary schools.

• General secondary schools and vocational/technical schools accounted for approximately 10 percent and 8 percent of spending on education, respectively.

• Public universities accounted for 30 percent of the total spending, while foundation universities and private course houses providing higher education received only 2.3 percent of total education spending.

3.14 Private examination preparation programs (dersanes) at various levels of education (from primary to university level) accounted for 4.2 percent of 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 of one-third of total expenditures on public and private secondary schools.

Figure 3.3. Education Expenditures in Turkey in 2002, by Source (YTL 19.35 billion)

Other Sources2.3%(0.45)

International Resources

0.1% (0.01) Household

Funds 33.4% (6.46)

Local Government

Revenue0.9%(0.17)

Foundations, Associations,

Firms etc.1.9% (0.37)

Central Government

Revenue61.4%(11.89)

105

Figure 3.4. Total Education Spending in Turkey in 2002, by Provider

Public special education

0%

Public general secondary

11%

Public primary41%

Private primary2%

Private University2%

Public pre-primary0%

Not known by kind1%

Private Course Houses

4%

Private pre-primary0%

Public Vocational and Technical

8%Private general

secondary1%

Public University30%

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 of spending at both pre-primary and general secondary public schools (45.2 percent of total spending at pre-primary public schools and 46.2 percent at general secondary public schools). Public financing agents contributed to almost all of the spending at vocational/technical schools and special needs schools (more than 90 percent of the spending at each level). Private financing agents accounted for 36 percent of spending at public universities (YTL 2.12 billion) and all of the spending at private universities (YTL 447 million—Figures 3.5 and 3.6). Private out-of-pocket financing of public schools is a consequence of the continuing financial constraints faced by the public schools. Principals request that the parents contribute to the maintenance, costs of paper, and other needs of their schools. Funds are collected from time to time or before the registration of the new students.

106

Figure 3.5. Financing of Education, 2002

0%

20%

40%

60%

80%

100%

Pre-primary Primary GeneralSecondary

Vocationaland Technical

University Specialeducation

% financed by public financing agents % financed by private financing agents% financed by other financing agents

Source: World Bank Staff calculations

Figure 3.6. Out of Pocket Spending in Turkey in 2002, by Providers

(YTL 6.93 billion)

Universities35.3%

General High Schools15.2%

Vocational Tech./High Schools

3.3%

Not known by kind1.0%

Pre-primary Schools0.7%

Special Education Schools

0.2%

Private Exam Preparation Courses

11.7%

Primary Schools32.5%

Source: World Bank Staff calculations

107

(d) Per-student expenditure 3.16 Education expenditures per student vary substantially by the level of education. The average annual expenditure per student at primary school was YTL 810 (US$ 533), while it was four times higher at YTL 3,260 (US$ 2,145) for a University student. The average expenditure on students attending general secondary schools and vocational/ technical schools was US$ 902 and US$ 1144, respectively (Figure 3.7). The government spends US$ 395 per student at the primary level and US$ 1,088 at the university level. The Government spends the highest amount per student, equivalent to US$ 1,343, at vocational and technical schools.74

Figure 3.7. Total Education Spending per Student (US$), 2002(US$1=YTL1.52)

0

500

1000

1500

2000

2500

Pre-primary Primary GeneralSecondary

Vocational andTechnical

University

Tota

l spe

ndin

g pe

r chi

ld in

dol

lars

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$ 516) is almost one-third of 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 half the spending at private schools (US$ 1,587) offering the same education (Figure 3.8).

74 This issue is covered in more detail in the 2005 Turkey Education Sector Study, Chapter 3: Developing the Basic Competencies of All Students.

108

Figure 3.8. Education Spending per Student at Public and Private Schools (US$), 2002, (US$1=YTL 1.52)

0

500

1000

1500

2000

2500

3000

Pre-primary

Primary

General Secondary

Vocational and Technical

University(public& private)

Public schools Private schools

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).

Source: World Bank Staff calculations based on MONE

Graph 1. Achievement Across School Types

School type

PrivateBoardingBasic

Mea

n

,8

,7

,6

,5

,4

,3

,2

TURKISH

MATH

SCIENCE

SOCIAL

Figure 3.9: Achievement across school types

109

3.19 Public schools, regardless of the level of education provided, spend a higher share of their expenditure on personnel expenses relative to private schools. This is most likely a factor accounting for the differences in outcomes between public and private schools. The difference in the structure of spending reflects a more intensive use of books, supporting educational materials, and school support services in private schools. Research in much of the world indicates that there is a positive, causal relationship between the level of non-salary inputs and educational outcomes.

3.20 Public schools also allocated a larger share of their expenditures on investment spending relative to corresponding private schools (Figure 3.10). Again, the only exception is general secondary level of education where both public and private schools spent an equal proportion of their money (4 percent) on capital items. Public universities spent a higher share of 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 of investment expenses, at about 12 percent. Salaries and other personnel expenses accounted for 64 percent of total expenditures at public primary schools, but only for 39 percent of the total expenditures at university

0

20

40

60

80

100

Public primary

Private primary

Public General Secondary

Private General Secondary

Public University

Private University

Figure 3.10. Distribution of Spending on Educational Inputs in Turkey in 2002, (YTL 19.35 billion)

Other

Transfer

Capital

Other current

Personnel

Source: World Bank Staff calculations

(e) International Comparison of Education Expenditures

3.21 Total education expenditures, as a percent of GDP, are higher in Turkey than in most OECD countries (Table 3.4). Exceptionally high private expenditures as a share of GDP (2.2 percent) is the main reason for Turkey's outlier status. Compared to OECD partners, Turkey also allocates a smaller share of government spending to education (4.5 percent).

110

Table 3.4. Expenditure on educational institutions as a percentage of GDP

from public and private sources, by source of fund

Public 1 Private 2 Total Malaysia3 7,2 - - Korea 4,8 3,4 8,2 United States 5,1 2,3 7,3 Turkey6 4,5 2,2 6,7 France 5,6 0,4 6,0 Poland3 5,6 - - Mexico 5,1 0,8 5,9 Finland 5,7 0,1 5,8 United Kingdom 4,7 0,8 5,5 Germany 4,3 1,0 5,3 Spain 4,3 0,6 4,9 Czech Republic 4,2 0,4 4,6 Brazil3, 5 4,1 M m India5 4,0 0,2 4,2 Slovak Republic3,4 4,0 0,1 4,1 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 of education (Table 3.5). Total per student education expenditures in Turkey is about one third the per student expenditure in OECD, on average.75 Per student expenditure on pre primary is particularly low, as Turkey spends almost 16 times less than the OECD average. On primary education, Turkey spends four times less than the OECD average, and for secondary education, Turkey spends less than half of the OECD average. By contrast, Turkey's per-student spending on tertiary education is much closer to OECD averages. An analysis of the distributional incidence of such discrepancies on per student spending by level of education would be warranted.

75 In order to compare education expenditure among the countries in more analytical perspective, a relevant indicator is per student expenditures. The indicator "Education Expenditures as a Share of GDP" is not entirely relevant if there is a significant per capita income gap across the comparison countries.

111

Table 3.5. Per Student Education Expenditure by Education Level (2004 US$

adjusted with PPP)

Pre

primary Primary Secondary Tertiary USA 8,522 7,560 8,779 22,234 Denmark 4,542 7,572 8,113 14,280 Australia 5,713 6,571 8,562 11,274 Italy (P) 5,972 6,783 8,258 8,347 Sweden 3,504 6,295 6,482 15,188

Belgium 4,062 5,321 7,912 11,589 France 4,323 4,777 8,107 8,837 Netherlands 4,228 4,862 6,403 12,974 Germany 4,956 4,237 6,620 10,504 England 7,595 4,415 5,933 10,753 Spain 3,608 4,168 5,442 7,455 Ireland 4,026 3,743 5,245 10,003 Greece - 3,299 3,768 4,280 Hungary (P) 2,882 2,592 2,633 7,122 Mexico 1,410 1,357 1,915 4,341 Turkey (P)* 99 899 1,957 5,105 Turkey 276 1,214 3,046 7,847

Sele

cted

OEC

D C

ount

ries

OECD Total 4,490 4,819 6,688 12,319 Argentina 1,745 1,655 2,306 3,775

Brazil (P) 1,044 832 864 -

Chile 1,766 2,110 2,085 6,901

Uruguay (P) 1,200 1,202 1,046 2,201

Indonesia 73 108 322 1,414

India 57 405 650 2,522

Israel 3,428 4,650 5,617 11,494

Jordan (P) 342 811 840 -

Malaysia (P) 611 1,562 2,600 11,303

Philippines (P) 75 492 465 1,648

Sele

cted

Oth

er C

ount

ries

Thailand 764 1,045 1,081 1,851 Source: Education at a glance 2004; Yilmaz (2006), staff calculations Note: "P": only Public

3.23 In view of 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, it is important to measure Turkey’s spending levels on education with appropriate comparator countries. As shown in the scatter diagrams below (Figures 3.11-3.14), Turkey’s per student spending on pre-primary schools is very low, while spending in primary school is consistent with expectations given per capita income. For secondary and tertiary education, Turkey spends somewhat more per student than what would be predicted on the basis of its per capita income.

112

Expenditures on Pre-Primary Education (PPP Adjusted)

-1,000

0

1,000

2,000

3,000

4,000

5,000

6,000

7,000

8,000

9,000

0 5000 10000 15000 20000 25000 30000 35000 40000 45000

Per Capita Income (PCI), PPP

Expe

nditu

re p

er S

tude

nt

India

Indonesia

SwedenIsrael

SpainHungary

ArgentinaChile

Mexico

UruguayBrazil

Thailand Malaysia

PolandTurkeyJordan

Belgium

Netherlands

Denmark

France

Germany

Australia

United States

UK

Italy

Ireland

Source: World Bank Staff calculations

Figure 12. Expenditures on Primary Education (PPP Adjusted)

Thailand

IsrailIndiaIndonesia Chile

Argent inaTurkeyM exico

GreeceIreland

SpainEnglandGermany

NetherlandsFrance

Belgium

SwedenAustralia

Denmark USA

0

1000

2000

3000

4000

5000

6000

7000

8000

0 5000 10000 15000 20000 25000 30000 35000 40000

Per-Capita Income (PCI)

Expe

nditu

re p

er S

tude

nt in

US$

Source: World Bank Staff calculations

Figure 3.11. Expenditures on Pre-Primary Education (PPP Adjusted)

Figure 3.12. Expenditures on Primary Education (PPP Adjusted)

113

Figure 13. Expenditures on Secondary Education

Thailand

Israil

IndiaIndonesia

Chile Argentina

Turkey

M exico

Greece

IrelandSpainEngland

GermanyNetherlands

FranceBelgium

Sweden

AustraliaDenmark

USA

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

0 5000 10000 15000 20000 25000 30000 35000 40000

Per-Capita Income (PCI)

Expe

nditu

re p

er S

tude

nt (U

S$)

Source: World Bank Staff calculations

Figure 14. Expenditures on Tertiary Education

USA

Denmark

Australia

Sweden

Belgium

France

Netherlands

GermanyEngland

Spain

Ireland

GreeceM exico

Turkey

Argent ina

Chile

IndonesiaIndia

Israil

Thailand

0

5000

10000

15000

20000

25000

0 5000 10000 15000 20000 25000 30000 35000 40000

Per-Capita Income (PCI)

Expe

nditu

res

per S

tude

nt (U

S$)

Source: World Bank Staff calculations

Figure 3.13. Expenditures on Secondary Education

Figure 3.14. Expenditures on Tertiary Education

114

PISA Results by Secondary Expenditure per Student (PPP)

300

350

400

450

500

550

0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000 10,000

Secondary Expenditure Per Student (PPP)

PISA

Ave

rage

Per

form

ance

(200

3)

Uruguay

Thailand

Brazil

Mexico

Turkey

Sweden

Netherlands

Spain

Ireland

HungaryGreece

DenmarkFrance

BelgiumAustralia

Germany

ItalyUS

Source: 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 of per student spending, they provide an indication of the existing room for improving expenditure efficiency in education. A relatively comprehensive assessment of student achievement is provided by the average PISA scores of 15-year old students in three fields: Literacy; Mathematics; and Science. The average score in 2003 is 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 of the important level of per student expenditures. For example, compared to Thailand—a country with similar per capita income—Turkey spends approximately three times more per student on secondary education (Figure 3.13). However, Turkish students’ average PISA score is only marginally superior to that of Thai students, indicating that there is ample scope for efficiency gains in secondary education. Although student achievement, as measured by PISA indicators, is expected to increase with per capita spending (Figure 3.15), it would be counterproductive to further increase expenditure without, at the same time, taking steps to ensure that the efficiency of spending has improved.

3.25 There is a considerable difference in MONE per-student education expenditures by province. There is 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 of Tunceli—the region with the highest spending

Figure 3.15. PISA Results by Secondary Expenditure per Student (PPP)

115

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 of per student expenditures in secondary education.

3.26 These differences potentially signal an inter-region or inter-province equity problem. Most of the provinces which spent the least per student at the primary level are from the Southern and Southeastern regions. However, as the example of 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 on 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):

• improving the quality of educational achievement in public schools may require increasing the share of non personnel expenditures to levels similar to private schools and universities, as more intensive use of supporting educational material and ICT in private schools is conducive to better quality and outcomes;

• improving the regional balance of educational services across the country may require enhanced public education service provision in most backward provinces to compensate for possible lack of private sector initiative;

• increasing per student expenditure on pre primary education, an area where Turkey lags far behind comparators even accounting for differences in level of development, yet where progress is critical to improve participation of 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 of public spending on 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 is especially true for vocational school classrooms that are under capacity by an average of 12 students per classroom. By improving school capacity utilization, it would be possible to increase the efficiency of spending by reducing the number of teachers employed for the same level of enrollment.

116

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 of the world, Turkey’s public schools have the least autonomy over resources, staff deployment (at the school), textbook selection, allocation of instructional time, and selection of programs offered. More school autonomy and accountability will help improve the allocation of resources.

3.30 The first and second strategies could be combined in an innovative way by introducing a system of per-student financing. This system could be similar to the one implemented in the United Kingdom, where funding is allocated in proportion to the number of students a school enrolls up to its capacity. The school is responsible for using these funds to cover their personnel, material, and professional development costs, but would be held accountable by Government for results.

3.31 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. To be able to become more efficient and provide better results, universities need to move funds between items and activities during the course of the year. At present, higher education institutions have very little flexibility as their public funds are allocated in a large number of 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 on the basis of historical precedent. Financial autonomy can be increased by reducing the number of ‘blocks’ – or 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 funds 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 of funding allocated to each university should not be based on inputs (such as staff), but on ‘performance’ in terms of 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. It should be recognized, however, that a comprehensive approach should be designed to achieve needed administrative and financial autonomy of higher education institutions.

3.32 Fourth, the university entrance examination system should be modernized in such a way as to eliminate the advantage of private examination preparation courses (dersanes and tutors). YOK is considering basing such a revision of the current examination with a system based to some degree on models such as the French or International Baccalaureate, the German Arbitur, or the British ‘A’ levels. The Turkish model would be a comprehensive set of exams that test both the knowledge and skills that students learn in secondary school. It could have a core of subjects with blocks of 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 to

117

do with what they have learned, how they reason, and how 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 is 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 no value in the real world – and certainly should have no value in institutions of higher learning. If half of the private funds used to finance examination preparation today could be redirected to subsidize public secondary education, it would represent a boost of 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 of a temporary improvement in 1999, as parameters improved from passage of a pension reform law. Policymakers are hoping to have solved the major problems with the passage of the newest pension and health insurance law in May of 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 of three pension institutions. Sosyal Sigortalar Kurumu (SSK) covered private sector workers and state employees not considered career civil servants. Emekli Sandigi (ES) covered state employees with civil 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 is 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 of population aging (Figure 3.16). The challenge of ensuring sustainability of the pension system faced by Turkey is also encountered in Latin America countries where the population is also young, and it is mainly a result of excessive benefits in relation to revenues. Solutions in other countries have ranged from privatization of the system, to drastic cutbacks in benefits and/or increases in contributions and/or increases in retirement ages.

118

Percentage of Population > 65

0

5

10

15

20

25Tu

rkey

Irelan

d

Icelan

d US

Cana

da

Denm

ark

Norw

ay

Finlan

d

Switz

erlan

d

UK

Spain

Fran

ce

Austr

ia

Portu

gal

Swed

en

Belgi

um

Gree

ce

Germ

any

Italy

Source: OECD

3.36 The first critical feature which led to the downfall of the Turkish pension system was the removal of the minimum retirement age in 1991. At this point, individuals could retire once they had completed the minimum years of service required for retirement, which were 25 years for men and 20 for women. But years of service did not imply years of contribution. In the case of SSK, an individual needed to fulfill only 5000 days of contributions (14 years) before becoming eligible for a pension which led to the phenomena of substantial numbers of individuals retiring before the age of 40. 3.37 The 1999 pension 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 of 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 of the pension system, the new retirement ages were only to be gradually phased in, with the earliest retirees after passage of the 1999 law allowed to retire at 43 for men and 38 for women. The phase-in for this retirement age is painfully slow with average retirement ages well below 50 even today. 3.38 The second feature of 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 low pensions for which many individuals qualify—as a result of the low ages and the low contribution periods required to be eligible for pensions. Since the amount of the flat pension was decided by Parliament each year, without regard to the finances of 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 on incentives. If the majority of the pension is unrelated to the number of years of contributions made or the amount on which one contributes, individuals have an incentive to declare the fewest number of years and the lowest salary

Figure 3.16: Population over Age 65 in Turkey and Selected OECD Countries (in %)

119

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 of high inflation, the value of these flat payments is 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 SSK and BK self-employed and indexed the pensions post-retirement to inflation, again there were complaints about the low level of pensions. For SSK and BK, social support payments were taken as prepaid inflation increments. Inflation increments are paid in Turkey on a monthly basis. The flat payments were added to the pension payments at the beginning of the year and no inflation increments were given until the sum of the unpaid inflation increments equaled the value of the flat social support payment. By midyear the value of the pension in SSK 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 still a residual increase at the end of the year over where the pension would have been with inflation increments. Rather than reduce the nominal amount of the pension payments the following year, the Government chose to simply add the residual amount of the flat payment to the base pension and then index normally from that point forward. To avoid inequities for new pensioners, an equivalent flat amount was added to the base pensions of new pensioners as well.

B2. Comparison of 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 self-employed insured under BK. While the minimum retirement age and the new benefit parameters were a vast improvement over the previous law, they still were generous by international standards and provided individuals an incentive to contribute as few years as possible. Previously, the minimum and maximum wages on which contributions were paid were adjusted at the discretion of Parliament and were not linked to any other parameters in the economy, including minimum wage. Thus, there was a period of 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 of a recognized wage series, many of the parameters were linked to GDP growth or inflation. The four achievements of the 1999 law were:

• the introduction of the minimum retirement age for all three pension institutions and all sub-systems, although at an extremely slow phase-in rate,

• an increase in the length of contribution required in SSK, although again at a slow rate,

120

• some decrease in the disincentives for longer contribution periods in SSK and BK self-employed, and

• a linking of important pension system parameters in SSK and BK self-employed to economic factors.

The law very importantly did not achieve a unification of the pension parameters in all four sub-systems. 3.41 The 2006 law focuses on the unification of the four sub-systems, bringing them institutionally into a single institution with uniform parameters for all categories of workers. This is a significant achievement for improving the flexibility of the Turkish labor market, for improving compliance, and for improving the fiscal sustainability of the overall pension system, given the rapidly increasing costs of the civil service system prior to the reform. Table 3.6 provides a comparison of the pension system parameters before and after the 2006 reform. The most significant changes apply to the civil servant system and the BK farmers’ system bringing them in line with the other sub-systems.

121

Table 3.6: Comparison of the pension system parameters before and after the 2006 reform New Unified

System SSK BK self-employed ES BK farmers

Retirement Age Eventually 65/65 Rising from 43/38 to 60/58

Rising from 43/38 to 60/58 Rising from 43/38 to 60/58 Rising from 43/38 to 60/58

Years of Contribution

Eventually 25 years

Rising from 5000 days to 7000 days

Rising from 5000 days to 7000 days

25M/20 F 25M/20 F, but with partial pension option

Accrual rate 2.5% per year until 2015; 2% per year thereafter

3.5% for first 10 years; 2% for next 15; 1.5% thereafter

3.5% for first 10 years; 2% for next 15; 1.5% thereafter

3% per year for 25 years; 1% thereafter

70% total benefit for 25 years; 1% increment and decrement for more or fewer years

Pensionable base Gradually increased to lifetime salary

Gradual increase to lifetime salary from 5 years in 1999

Gradual increase to lifetime salary from 3 years in 1999

Last salary Last salary

Revaluation of pension base

50% to inflation; 50% to average wage growth

100% to CPI inflation + real GDP growth

100% to CPI inflation + real GDP growth

NA since based on last salary

NA since based on last salary

Indexation of pension post-retirement

100% to inflation 100% to inflation 100% to inflation 100% wage indexed to wage growth in position from which retired

100% wage growth of civil servants

Contribution Rate

11% for employers; 9% for employees; or 20% for self-employed

11% for employers; 9% for employees

20% for self-employed 20% for employers; 16% for employees, but covers health expenditures in retirement

20% for farmers

Minimum Insurable Earnings

Minimum wage Minimum wage as of June 2004; previously linked to nominal GDP growth

Wages based on self-determined steps with mandatory increase to next step for first 12 years

Minimum base wage Wages based on self-determined steps

Maximum Insurable Earnings

6.5 times minimum wage

6.5 times minimum wage

24th step amount Full base wage 24th step amount

122

B3. Impact of 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 life expectancy changes, as life expectancy continues to rise in the outer years of the simulation period, pension deficits will begin to re-emerge, but these will not be anywhere close to the magnitude of the current deficits and will remain below 1 percent for the remainder of the simulation period. The pension projections below show the fiscal impact of the reform (Figure 3.17).

Figure 3.17: Pension System Deficits Before and After Passage of the 2006 Law

-8%

-7%

-6%

-5%

-4%

-3%

-2%

-1%

0%

1%

2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060 2065 2070 2075

Year

% o

f GD

P

Base Case Reform CaseSource: PROST projections produced by Bank staff

3.43 The positive achievements of the reform include the substantial fiscal gains as well as the unification of a previously fragmented system. Some of the fiscal gains come from the uniform move to inflation indexation of 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 is considered a normal working career or not. However, there remain some areas of concern for the longer term 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

123

revenues. The Turkish government has recently expressed interest in lowering contribution rates in order to encourage growth of the formal sector labor force. Currently, there is 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 its 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 is a desired objective, then more changes will be required to achieve that goal. Options for such changes are further outlined below.

Pension Contribution Rates in OECD Countries

05

10152025303540

Portu

gal

Italy

Spain

Nethe

rland

s

Austr

ia

Finlan

d

Gree

ce

Turke

y

Germ

any

Swed

en

Japa

n

Belgi

um

Fran

ce

Luxe

mbou

rg US

Icelan

d

Switz

erlan

d

Cana

da

% o

f Gro

ss W

age

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 of most other OECD 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 of other countries. One might argue that Turkey’s life expectancy is below that of the other OECD countries, but OECD countries like Mexico are quite comparable to Turkey and still have retirement ages of 65 for both men and women even today. The low 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 low 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.

Figure 3.18: Contribution rates in selected OECD countries

124

Retirement Ages for Men and Women in OECD

01020304050607080

Austr

alia

Austr

iaBe

lgium

Cana

daDe

nmar

kFin

land

Fran

ceGr

eece

Germ

any

Icelan

dIre

land

Italy

Japa

nLu

xemb

ourg

Mexic

oNe

therla

nds

New

Zeala

ndNo

rway

Portu

gal

Spain

Swed

enSw

itzer

land

Turke

y 200

6Tu

rkey 2

048 UK US

MenWomen

Source: OECD and World Bank Staff calculations

3.46 Accrual rates, the benefit paid per year of 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 of the OECD has gone much farther in reducing accrual rates.

Accrual Rates in OECD pension systems

00.5

11.5

22.5

3

Turke

yTu

rkey

Spain

Turke

yPo

rtuga

lIta

lyLu

xFr

ance

Austr

iaNe

therla

ndSlo

vakia

Finlan

dKo

rea

Czec

hBe

lgium

Swed

enSw

itzer

land

Icelan

d USHu

ngar

yGe

rman

yNo

rway

Japa

nPo

land

Cana

da UK

% o

f wag

e per

year

of s

ervic

e

20062007

2016

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 of the labor force, growth in the wage base, and better work incentives. Currently, almost 50 percent of the contributors in SSK contribute on the basis of minimum wages; most workers in all systems

Figure 3.20: Pension Accrual Rates in OECD Countries

Figure 3.19: Retirement ages in OECD countries

125

contribute for as few years as possible in addition to those who work entirely in the informal labor force. This behavior is linked with the pension parameters in the Turkish pension system. They are all predicated on the belief that the typical working or contribution life is 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 of the few expected years of contribution, Turkish policymakers are forced to choose between raising the accrual rate per year of service or facing the prospect of many people receiving very low pensions. But the high accrual rates undermine work incentives. If individuals can obtain decent pensions with only 20 years of 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 of work, which then dropped to 2% per year for the next 15 years of work and then to 1.5% per year thereafter. Now there is a flat rate for all years of work. Actuarially fair insurance would reward those who work beyond retirement age by approximately 6% additional pension per year of delay.

3.48 Furthermore, even in the new system, there is the possibility of a partial pension. While the new law stipulates that the new entrant as of 2026 must contribute 25 years before receiving a full pension, partial pensions can be received with only 15 years of 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 requirement.76

3.49 While the 2006 law generally moves away from the previous structure and toward a rule-based system which encourages longer careers, one of 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 of 50 percent of past average wage growth and 50 percent of past inflation. As a result, an individual who contributes on the basis of average wage throughout his career will receive a pension based on substantially less than average wage.77 Furthermore, the wages furthest from the point of retirement, the early wages, will be valued less than

76 As a result, there is a relatively large number of pensioners, especially in BK where individuals contribute as little as possible, who are receiving very low pensions. As already noted, the typical response has been to complain through the political process that pensions are too low and to receive some kind of supplement which is equally distributed to those who contributed 10 years as well as to those who contributed 30. These periodic supplements further 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 on the basis of 100 percent of average wage growth. Thus, an individual who paid contributions on the basis of average wage gets a pension calculated as a percentage of the current average wage.

126

wages earned closer to retirement, giving individuals an incentive not to contribute for long durations of time.

Average Pension Benefits as % of Average Wage in SSK

.0%5.0%

10.0%15.0%20.0%25.0%30.0%35.0%40.0%

2006

2010

2014

2018

2022

2026

2030

2034

2038

2042

2046

2050

2054

2058

2062

2066

2070

2074

Source: World Bank Staff calculations

3.50 The 2006 reform will reduce the level of pension benefits relative to the average wage (Figure 3.21). The fall in pension benefits can be attributed to a number of factors. The 1999 law lengthened the wage period on which pensions were based from the previous average of 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 is increasing the number of 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 is credited under pre-1999 rules, which were more generous. Thus, the 2006 value includes some grandfathering of 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 of the change in pension benefits can be attributed to this change in the revaluation of 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 of SSK is provided here, the impact of 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.

Figure 3.21: Average Benefits Relative to Average Wage in SSK, post-reform

127

3.51 Summing up—important steps have been taken forward but, to address the challenge of formalization, the reform agenda would need to be completed in the future based on the results of the 2006 reform. The 2006 Turkish pension law 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 of the labor force and reducing contribution rates. In addition, the law removes many of the disincentives to longer working careers embedded in previous pension laws, but may not provide sufficient incentives to increase the length of the contribution period and the formalization of the labor force. Further reform steps in the future should be envisaged in light of the results of the 2006 reform.

C. PUBLIC EXPENDITURE IN THE HEALTH SECTOR

C1. Health outcomes 3.52 Turkey has achieved considerable progress in terms of 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 of GDP per capita and approximately the same share of public spending on health (Table 3.7).78 In general, these indicators are worse than average indicators in the Europe and Central Asia region or the group of Upper Middle Income countries (with the exception of 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 is 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 of 23 per 1,000 live births by 2015. Trends regarding maternal mortality, on 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 is observed, this is not very strong and maternal mortality remains high as compared to other countries in the region or countries with similar income level as Turkey.80

78 Countries with similar shares of public spending on health may not have similar levels of health outcomes, as there is 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”. 80 Information on mortality and morbidity in Turkey is based on hospital data (discharges and deaths) and for that reason tends to underestimate the actual mortality and morbidity levels.

128

Table 3.7: Health indicators for selected countries (2003 or latest available year) Vaccination

rates GDP per

capita (US$ 2000 con.)

THE (%

GDP)1

PHE (%

GDP)1

LE Infant Mortality

Under-5 Mortality

Maternal Mortality2

DPT Measles

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 Rep.

5,899 7.0 6.4 75 4 51 9 97 99

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 Rep.

4,254 5.9 5.3 73 7 8 3 99 99

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 12 171 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 THE should be treated with caution as they might underestimate the level of private spending on health; 2. Figures for Maternal Mortality Ratios are for 2000; 3. Infant Mortality Rate: the probability of dying in the first year of life per 1,000 live births, Under-5 Mortality Rate: the probability of dying before the fifth birthday per 1,000 live births, Maternal Mortality Ratio: modeled estimated, the number of deaths among women aged 15-49 per 100,000 live births, Vaccination rates: percentage of children aged 12-23 months; 4. DPT: Diphtheria, Pertussis and Tetanus; E & CA: Europe and Central Asia region; UMI: Upper Middle Income countries. Source: World Development Indicators, 2006.

Table 3.8: Health-related MDG indicators for Turkey from various sources, 1990-2000 Source

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

Infant mortality (per 1,000 live births) WDI 64.0 .. .. .. .. 50.0 .. .. .. .. 38.0 .. .. 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

Maternal mortality (per 100,000 live births) WDI .. .. .. .. .. .. .. .. .. .. 70.0 .. .. ..

WHO/HFA 180.0 .. .. .. .. 55.0 .. .. 130.0 .. .. .. .. ..

129

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 Infant and under-five mortality rates are higher in the North and East regions. 3.54 The prevalence of HIV/AIDS seems to be relatively low in Turkey but is increasing. The first AIDS/HIV case was registered in Turkey in 1985. By the end of 2004 there were 1,992 AIDS/HIV cases registered at the MOH.82 A comparison with 2001 suggests that the number of 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 of number of 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 Under-five Mortality Rates by area of residence and region, 2003 Infant Mortality Under–five 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 Survey, 2003. 3.55 Better use of preventive health services among high risk groups (mothers and children) is key to improving core health outcomes. Turkey’s high levels of infant and under-5 mortality, as well as high maternal mortality, are related to the poor usage of preventive care.84 According to the 2003 Demographic and Health Survey, one-fifth of mothers do not receive any antenatal care during their pregnancies. The share of mothers who do not receive any antenatal care increases to one-third in rural areas. With respect to the place of delivery, although 78 percent of all births are delivered in a health facility, nearly half of births in the East region take place at home. Similarly, the proportion of births delivered with the assistance of a doctor or trained health personnel decreases from 81 Maternal mortality ratios are not reported in the 2003 TDHS. 82 Draft tables prepared for the MOH Statistical Yearbook 2004 (forthcoming). 83 MOH data. 84 Preventive services include the programs of TB, Malaria and Maternal and Child Health, Health Education and Cancer Control, as well as activities of the Institute of Hygiene Refik Saydam. This definition of preventive activities does not coincide with the one 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).

130

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 of children 12-23 months are fully vaccinated.85 Vaccination rates also vary among regions, with the percentage of children fully vaccinated in the East being as low as 35 percent. However, according to MOH administrative data vaccination rates have increased in the last two years with 77 percent of children 12-23 months fully vaccinated at the end of 2005 nationwide and 62.5 percent in the South East Anatolia region.

C2. Main characteristics of the health sector86 3.56 Estimates of the number of 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 of the total population (59 million) enjoys health insurance coverage (Table 3.11). SKK has the highest number of enrollees with around 32 million (46 percent of 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 of the population. Nevertheless, issues such as multiple enrollment, non up-dated records and the use of estimates for the number of dependents have been put forward as problems affecting the reliability of these data. Estimates from the 2002-2003 Turkey National Household Health Expenditure Survey, undertaken in the framework of National Health Accounts, report a national insurance coverage significantly below that reported in national statistics, with only 67 percent of the population being insured, including the Green Card program.87 Box 3.3: The system of financial protection for health expenditures The most important entity is the Social Security Organization (Sosyal Sigortalar Kurumu, SSK). SSK is 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 of earnings) and employers (6 percent of 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 not pay for preventive services. (Table 3.10).88 Self-employed individuals, agricultural laborers, as well as their dependents are insured by the Social Security Institution of 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 of DPT and polio. 86 This section provides a brief description of the health sector in Turkey with respect to financial protection mechanisms and the provision and utilization of health care services. For a detailed description of financial protection mechanisms and provision and utilization of health care services in Turkey, see World Bank (2003) 87 Analysis of household surveys for 2001 and 2002 provide similar results regarding health insurance coverage; data indicate that only 62 percent to 63 percent of the total population in Turkey has access to some form of health insurance, including Green Card holders (World Bank, 2005). 88 This group also includes invalid individuals, widows/ers, and orphans.

131

health premiums (equal to 20 percent of 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 of 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 is similar to 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 to poor individuals. The Green Card program is managed by the MOH and it provides free health services (with the exception of pharmaceuticals prescribed during outpatient visits since May 2005). This program was envisaged as a transitory measure until the introduction of the universal health insurance coverage takes place.

Table 3.10: Co-payments for health care services and pharmaceuticals SSK ES Bag-Kur Green

Card Uninsured

Emergency services No co-payment

No co-payment

No co-payment

No co-payment

No co-payment

Outpatient services Active workers and dependants No co-

payment; YTL 0.8 for dependents

No co-payment

20% per visit

Pensioners and dependants YTL 0.8 per visit

No co-payment

10% per visit

No co-payment

Full price

Inpatient services Active workers and dependants No co-

payment No co-payment

No co-payment

No co-payment

Full price

Pensioners and dependants No co-payment

No co-payment

No co-payment

No co-payment

Full price

Pharmaceuticals1 Active workers and dependants 20%

outpatient care / Free inpatient care

20% outpatient care / Free inpatient care

20% outpatient care / Free inpatient care

Pensioners and dependants 10% outpatient care / Free inpatient care

10% outpatient care / Free inpatient care

10% outpatient care / Free inpatient care

20% outpatient care / Free inpatient care

Full price

Notes: 1. The distinction between active workers and retired applies only to SSK, ES and Bag-Kur. 2. In the case of drugs prescribed by a physicians’ committee, no co-payments apply. 3.57 The percentage of insured individuals in rural areas is considerably lower than the national average. According to the 2002-2003 household data, only 57 percent of individuals are insured in rural areas, while the corresponding percentage in urban areas

132

is 72.5 percent. Across regions, the lowest insurance coverage is registered in East Turkey at 54 percent.

Table 3.11: Number of insured individuals by type of health insurance coverage, 2002 Type State Planning Organization

2002 Household Health Expenditure Survey 2002-2003

% of total population % of total population Emekli Sandigi 15.4* 12.5 Civil

servants - 7.4

Retired - 5.1 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 Note: * This figure includes both active and retired civil servants. Source: State Planning Organization, www.dpt.gov.tr, Economic and Social Indicators and Turkey National Household Health Expenditure Survey 2002-2003. 3.58 Universal health insurance will be introduced starting January 2007. The introduction of 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 of 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 of insured individuals. For example, in the case of the Green Card program it is well known that resources fall short of the actual needs of 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 of health care service providers is evolving. Until 2004, health care services in Turkey were mainly provided by the MOH facilities, the SKK facilities and public university hospitals, with only 6 percent of 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 of health care providers to autonomous entities. This transformation forms part of the Government’s health sector reform agenda and aims at separating the financing from the provision of health care services. 3.60 The recent changes have caused the MOH to reinforce its position as the main provider of health care services. In particular, the MOH is the sole provider of preventive and primary health care services, provided through a network of 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 of all hospitals beds in the country. University

133

hospitals complement the provision of secondary and tertiary level care through 56 tertiary hospitals (15.2 percent of hospital beds), while the private sector is the third largest player with 287 hospitals, but only 6.7 percent of total bed capacity. Another 75 hospitals (around 10 percent of hospitals beds) are owned by other ministries, municipalities, associations, foreigners and minorities.

Table 3.12. No. of hospitals and hospital beds by type of institution, 2004 Type of institution No. of hospitals % (hospitals) No. of beds % (beds) Ministry of Health 770 67.7 132,062 50.4 Ministry of Defense 42 3.5 15,900 8.5 Faculty of 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 Source: MOH. 3.61 The total number of hospitals and hospital beds has increased considerably, with a stronger role of the private sector. From 1994 to 2006, the number of hospitals and hospital beds has grown by 17 and 32 percent respectively.89 The highest increase has been registered in the number of hospital beds in the private sector (63 percent), followed by increases in the number of university hospital beds (50 percent). 3.62 A large number of hospitals were – and continue to be – substantially underutilized. Moreover, important variations still 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 is 62 percent, which is much below other European countries (Table 3.13). 3.63 The number of physicians and nurses has also increased but regional disparities remain. For the periods 1994-2004 the number of physicians and nurses grew by 51 percent and 74 percent respectively.90 Their numbers have thus approached the corresponding averages in the EU-15 countries.91 Nevertheless, health care providers are unequally distributed, with fewer providers operating in the East and Southeast regions as compared to the rest of the country. As a result, some areas have an over-supply of health care providers. 3.64 The availability of diagnostic technologies has considerably improved. In 2003, the number of magnetic resonance imaging (MRI) units in Turkey was 3 per million population, less than half of the OECD average of 7.6 MRI units, but higher than France (2.8), Hungary (2.6), Czech Republic (2.4), Greece (2.3), Slovakia (2.0), Poland (1.0) and Mexico (0.2). Similarly, the number of computed tomography scanners in Turkey stood

89 Turkish Statistical Service. 90 MOH data. 91 Health for All Database, WHO/Europe, 2006.

134

at 7.3 per million population in 2003, less than half the OECD average of 17.9, but higher than Hungary (6.9), Poland (6.3), the UK (5.8) and Mexico (1.5).

Table 3.13: Number of Hospital Beds and Utilization (2003 or latest available year) Beds per 100,000 Average Length of Stay Bed occupancy rate (%)2

Austria 834.7 7.9 76.2 Denmark 398.7 5.4 84.0**

France 760.3 11.7 84.0 Germany 874.6 10.6 77.6 Italy 411.8 7.6* 76.9*

Netherlands 457.7* 12.5** 58.4**

Portugal 363.9 8.7 85.2 Sweden … 6.5 ... Czech 855.5 11.2 74.1 Hungary 783.5 8.4 77.2 Latvia 781.3 10.8 ... Lithuania 868.2 10.3 73.4 Poland 547.0 7.7 ... Slovakia 732.3 9.1 64.8 Slovenia 495.5 9.4 68.1 Romania 732.3 8.0 ... Turkey 257.4 5.8 61.9 EU-25 595.7 9.5 77.5 EU-103 655.5 8.6 72.7 Notes: 1. * 2002; ** 2001; 2. Only for acute beds; 3. EU-10: 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 of outpatient contacts per person per year between 1993 and 2003 (Table 3.14). A similar pattern is 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 of the Health Transformation Programme (HTP) which was initiated in 2003 is expected to influence further the number of health personnel, the availability of diagnostic technologies and health care utilization in Turkey, thus close monitoring of indicators in these areas is recommended in the following years.

92 The Turkish population has a much lower median age and lower percentage of the population above 65 years and these figures are not standardized or adjusted for age.

135

8.02

11.52

11.85

17.54

17.32

18 .54

21.76

22 .82

24 .98

24 .92

18 .32

19 .41

0 5 10 15 20 25 30

Turkey

Portug al

Spain (2002)

Bulgaria

Po land

Slovakia

Czech Repub lic

Russ ia

Hungary

Romania

EU 15 (2002)

EU 25

2 .6 3

3 .6 7

5.90

9 .40

9 .50

12 .22

12 .96

15.00

8 .54

0 .0 0 2 .00 4 .00 6 .0 0 8 .0 0 10 .00 12 .00 14 .00 16 .0 0

Turkey (2001)

Po rtugal

Po land

Russia

Spain

Hungary

Slovakia

Czech Repub lic

EU 25

Table 3.14: Health care service utilization, 1993-2003

Year Out-patient contacts per person per year

Inpatient admission per 100 individuals

1993 1.50 5.88 1994 1.60 6.17 1995 1.70 6.28 1996 1.80 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 75.331 37.90

Note: 1. 1993-2001; 2. 1993-2003 Source: Health for All Database, WHO/Europe, 2006. Figure 3.22: Inpatient admissions for selected Figure 3.23: Outpatient contacts for selected countries, 2003 countries, 2002 Source: Health for All Database, WHO/Europe, 2006.

C3. Public expenditures on health 3.67 Turkey spent approximately 6.6 percent of 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 of GDP) but lower than Spain, Portugal, Greece and Brazil (all between 8 and 10 percent of GDP). In terms of public-private mix, public expenditures on health in Turkey accounted for 68 percent of total health expenditure in 2004, up from 61 percent in 1999. The allocation of public health care expenditures among the various entities who incur them poses certain challenges mainly because of the existence of revolving funds in MOH and university facilities (Box 3.4). The combination of poor health outcomes and relatively high expenditures suggests that

136

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 on health consist of expenditures incurred by the MOH, General Directorate of Coastal Services, Universities, Social Solidarity Fund, other Ministries and agencies, civil 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 to individuals belonging to the three SSI and payments for the services provided are channeled to the revolving funds of each facility. This means that part of the resources that appear as expenditures in the financial statements of SSK, Bag-Kur and Emegli Sandigi, also appear as revenues in the financial statements of the MOH and university facilities. In order not to double-count these resources, one could (a) assign all expenditures to the SSI, leaving the revolving fund accounts only with out-of-pocket expenditures93 or (b) discount the funds paid to the MOH and university facilities from the SSI accounts and present them under the revolving fund accounts. Table 7 presents public expenditures on health using the second method. This method not only guarantees that public expenditures are single-counted, but also clearly presents the size and importance of revolving funds in the financing of MOH and university facilities. Data show that the financing of MOH and university facilities increasingly relies on revolving funds. Revolving funds represented 17.6 percent of total health expenditure in 1999, increasing to 28.7 percent in 2005, with the share of the revolving funds of the MOH increasing at a higher pace than the share of the revolving funds of university hospitals. Resources from revolving funds for MOH and university facilities mainly finance personnel salaries, capital investment, maintenance and recurrent costs.94 The fact that providers are allowed to allocate these resources with greater flexibility than any budget related resources introduces a variety of perverse incentives into the system. First, it raises equity considerations since individuals with no ability to pay for user charges cannot be treated in the facility of their choice. Second, it might lead to an increase of utilization of health care services as a result of induced demand from the side of the health care providers.95 Third, having to choose among patients whose institutions pay into revolving funds and individuals whose provision of services is financed by the budget, health care providers might show a preference for the first group, crowding out the poorest from the system. This is especially the case after the introduction of a performance-based payment mechanism for the distribution of revolving funds among individual professionals working in hospitals. The importance of revolving funds has increased even more from 2005, with the transfer of SSK facilities to the MOH and the set up of revolving funds for these facilities as well. Data of January 2006 on the number of revolving funds of the MOH facilities suggest that all hospitals under the MOH have operational revolving funds (Table 3.15).96 The distribution of revolving funds across the country suggests that the poorest regions (East Anatolia and South East Anatolia) have the lowest number of revolving funds.

93 These are co-payments of the SSI enrollees or out-of-pocket expenditure for individuals who are not qualified to receive free care. 94 After deducting the shares of the Treasury (11 percent), Social Assistance Fund for Children (around 1.5 percent), and MOH (2 percent), the rest of revolving funds are distributed for salaries (up to 50 percent of the remaining balance), capital investment, maintenance and recurrent costs. 95 For these implications of revolving funds, see World Bank (2003). 96 The total number of revolving funds in MOH hospitals in the Box Table is lower then the sum of MOH and SSK hospitals in Table 5 due to the merging of some of the MOH and SSK facilities after the transfer of the latter.

137

Table 3.15: MOH Hospital Revolving Funds, January 2006. Regions Number of revolving funds Percent Mediterranean 83 10.16 East Anatolia 94 11.51 Aegean 113 13.83 South East Anatolia 56 6.85 Central Anatolia 151 18.48 Black Sea 163 19.95 Marmara 157 19.22 Total 817 100.00

Source: MOH 3.68 Total public expenditures on health have increased from 3.3 percent of GNP in 1999 to above 5 percent in 2004 and 2005 (Table 3.16). All available statistics confirm that public expenditures on health have risen significantly in 2005 as well as in the first quarter of 2006. Public expenditures on health of social security institutions, Green card and Civil servants increased from YTL 13.9 billion in 2003 to YTL 17.9 billion in 2005 (representing an increase of 29 percent). At the same time, MOH and Universities budget expenditures on health increased from YTL 3.4 billion in 2003 to YTL 6.2 billion in 2005 (representing an increase of 82 percent). The share of personnel expenditures on health has increased during this period from 1.1 percent to 1.6 percent of GDP, while the share of investments has fallen.

Table 3.16: Public expenditure on health, 1999-2005 Million YTL 1999 2000 2001 2002 2003 2004 2005

CIVIL SERVANTS HEALTH 190,191 285,429 479,749 836,776 1,063,078 1,242,000 1,096,000

PHARMA 198,272 308,001 543,393 817,224 972,410 1,220,000 1,100,000

TOTAL 388,463 593,430 1,023,142 1,654,000 2,035,488 2,462,000 2,196,000

SSK HEALTH 389,238 622,780 1,123,942 1,493,315 2,559,000 3,287,500 3,370,000

PHARMA 304,017 572,409 992,616 1,878,558 2,450,000 2,471,528 3,476,000

OTHER 55,445 85,000 141,400 222,477 352,079 446,500 458,000

TOTAL 748,700 1,280,189 2,257,958 3,594,350 5,361,079 6,205,528 7,304,000

ES HEALTH 135,635 222,152 345,171 645,117 867,046 1,135,000 1,077,080

PHARMA 198,076 358,185 660,135 1,099,664 1,516,601 1,524,000 1,589,860

OTHER 25,737 42,736 84,089 95,440 114,543 136,000 223,060

TOTAL 359,448 623,073 1,089,395 1,840,221 2,498,190 2,795,000 2,890,000

Bag-Kur HEALTH 140,327 215,605 330,043 692,900 996,700 1,111,000 1,396,000

PHARMA 231,419 458,336 780,446 1,321,531 1,892,100 2,302,000 2,056,000

OTHER 41,672 56,355 118,360 180,900 200,700 247,000 163,000

TOTAL 413,418 730,296 1,228,849 2,195,331 3,089,500 3,660,000 3,615,000 GREEN CARD HEALTH 108,161 166,580 300,817 536,937 665,000 612,000 700,000

PHARMA 19,273 31,939 91,438 112,878 252,000 450,000 1,109,000

138

TOTAL 127,434 198,519 392,255 649,815 917,000 1,062,000 1,809,000 MOH+UNIV +LOC GOV+OTHER HEALTH

529,907

933,638

1,615,585

3,180,988

3,414,363

6,453,948 6,239,539

TOTAL HEALTH 1,493,459 2,446,184 4,195,307 7,386,033 9,565,087 13,841,448 13,878,619

PHARMA 951,057 1,728,870 3,068,028 5,229,855 7,083,111 7,967,528 9,330,860

OTHER 122,854 184,091 343,849 498,817 667,322 829,500 844,060

TOTAL 2,567,370 4,359,145 7,607,184 13,114,705 17,315,520 22,638,476 24,053,539 TPHE (% GDP) 3.28 3.65 4.19 4.78 4.82 5.25 5.06 Source: World Bank calculations based on data from MOF, SPO, MOH, SSK, BK, ED. 3.69 The bulk of health care expenditures are on curative care and medical goods. Accurate data on the functional classification of recurrent public expenditures on health and the distribution of recurrent public expenditures by type of provider are available through the Turkey National Health Accounts for the years 1999-2000 only (Figure 3.24). In 2000, around 57 percent of 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 of 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 of total public spending on health. Figure 3.24: Functional classification of recurrent public expenditures on health, 1999-2000

1.6

5.2

30 .2

1.5

1.5

60 .0

2 .1

4 .3

33 .8

1.3

1.4

57.1

0 .0 10 .0 2 0 .0 3 0 .0 4 0 .0 50 .0 6 0 .0 70 .0

Health Adminis trat io n andInsurance

Prevent ion and Pub lic HealthServices

Med ical Go od s Dispensedto Outp at ients

Ancillary Services to HealthCare

Services o f Rehab ilitat iveCare

Services o f Curat ive Care

20 001999

Source: Turkey National Health Accounts 1999-2000, MOH, 2004. 3.70 Public expenditures on prevention are very limited. MOH is the major provider of 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 of Hygiene of Refik Saydam is responsible for the provision of laboratory based preventive services. The share of expenditures on preventive services has decreased from 11 percent in 1999 to 7 percent in 2005 (Table 3.17). In the area of maternal and child health, the MOH spent only 1.6 percent of its budget in 2005. This definition of expenditure on preventive activities does not coincide with the one

139

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).

Table 3.17: Preventive activities (as a share of MOH budget), 1999-2005 Percentage of MOH Budget 1999 2000 2001 2002 2003 2004 2005 TB program 1.76 1.72 1.56 1.65 1.29 1.34 1.25 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 of 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 (General Dir. of Primary Health Care)

33.66 32.61 29.80 34.34 28.83 31.65 30.11

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 of medical services and number of prescriptions for drugs have contributed most to expenditure growth. For instance, the number of outpatient visits in MOH hospitals increased from 110 million in 2003 to 130 million in 2004 to 170 million in 2005. 3.72 Across the different categories of beneficiaries, the bulk of the increase in expenditures is explained by increase in SSK enrollees and Green Card holders. The number of SSK enrollees holding a health card increased from about 22 million in 2003 to about 31 million in 2005, while the number of 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 of events starting from February 2004 help explain the observed trends in expenditure on pharmaceuticals. The more important of these are:

• Feb 2004 – Reference pricing introduced • March 2004 – VAT on prescribed drugs reduced by 8 percentage points • December 2004 – Protocol signed with suppliers of drugs results in a discount

between 7.5% and 14% in the price of drugs • Jan 2005(a) – Discount of 8% negotiated on non-prescription drugs • Jan 2005(b) – Prescription drugs extended to Green Card holders as part of

their benefit package • Feb 2005(a) – SSK enrollees allowed to obtain drugs from private pharmacies • Feb 2005(b) – Positive drug list introduced • Feb 2005(c) – SSK hospitals transferred and pharmacies closed

140

• May 2005 – Green Card holders required to make copayments, like others • June 2005 – A further reduction of 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 of March 2004, December 2004, January 2005a, and June 2005) resulted in a fall in price of drugs, some (May 2005) resulted in a fall in consumption of drugs, while others (January 2005b, and February 2005a) resulted in increase in consumption of drugs. The net result is that the price of drugs has, on average, fallen, but the quantity procured has increased significantly, particularly following the inclusion of drugs in the benefit package of 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 of medical services partly reflects the increase in coverage and access, effectively covering the entire population of the country— The increase in coverage, tantamount to effective introduction of universal health insurance even before the official enactment of the legislation, manifests itself in the increased utilization of services by Green Card holders and SSK beneficiaries. Both groups of beneficiaries have seen an impressive addition in the consumption of health goods and services in the course of the last 15 months. 3.76 —and partly the change of incentives facing medical providers in MOH hospitals and outsourcing of services. In parallel with the take-over of SSK hospitals by MOH, the incentives structure for physicians working in MOH hospitals has also been changed and their reimbursement from the net collections of Revolving Funds is linked to their performance, measured principally by the revenue generated for the Revolving Funds, which is a direct function of the number of services performed.97 At the same time, the limits to which MOH physicians can augment their salaries from the Revolving Fund have also been raised, to maximum of 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 of health services. In addition, almost all medical tests and diagnostics in MOH hospitals have been outsourced to private parties. While this has undoubtedly resulted in increased efficiency and lower unit costs of tests, it has also led to an increase in the number and range of 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 of health services to induce demand – have led to an increase in production and delivery of health services without necessarily improving the clinical quality and effectiveness of health 97 The net figure is determined by subtracting the 15% ‘tax’ on Revolving Funds that is collected by the Treasury and by subtraction of the costs of medical tests and diagnostics.

141

services. While some of the pent-up demand is almost certainly being met, new and unnecessary health expenditures are being generated in the system due to an increase in the total volume and mix of services as well as per episode of care. 3.78 In addition to the real increase in quantity of services being produced and delivered, there are concerns that the billing system further exaggerates the number of services provided. There is no working and effective claim processing system at present, and it is almost impossible to verify the nature, extent and authenticity of the bills produced and presented by MOH hospitals. The extent of this overstatement and inflation is difficult to establish and ascertain, but there is 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 of 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 of health services in-between. Reflecting this mixed nature of interventions, measures should be designed with the aim of simultaneously improving efficiency and enhancing equity of public spending. Key directions and underlining rationale include:

• strengthening of public health and preventive services (public good rationale);

• universal health insurance (equity rationale); • consolidating public hospitals, limiting hospital-based outpatient care,

managing claims for hospital services, controlling the quantity of 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 of public expenditures on preventive activities in the future in its effort to improve the health status of the population. Furthermore, resources need to be targeted where they are most needed, if it is for the persistent variation of health indicators between urban and rural areas and among regions to decrease. 3.81 Efforts to strengthen the delivery of primary health care services, in general, and preventive activities, in particular, have been initiated with the piloting of the family medicine system. Evidence from OECD countries indicates that a family-medicine centered health system has great potential to improve the cost-effectiveness of care while strengthening equity. Successful implementation of family medicine, in the

142

medium-term, will also contribute to reducing the problem of 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 of 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 of medicine domains, including internal medicine, gynecology and pediatrics. Family medicine places special emphasis on continuity of care and on quality of 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 of the patient referral system as well.98 A family medicine pre-pilot has been completed in Duzce and a recent public opinion poll indicates high user satisfaction. An evaluation of the Duzce pilot comparing with non-pilot sites is also underway. The Duzce pilot is currently being rolled out in three provinces. Another seven provinces will be covered by the end of 2007. The piloting and roll-out of family medicine is 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 of changes not only in health financing but also in the production and delivery of health services, and will generate new demands on the management and regulation of the health care system. An estimation of the effects of moving from a regime of social health insurance, administered by multiple agencies offering different sets of benefits but covering less than the total population, to a regime of social health insurance administered by a single agency offering uniform benefits requires an estimation of the effects of the proposed change on not only the flow of funds but also on the level of spending itself. The flow of funds will change as a result of the creation of a single payer agency replacing the multiple agencies and the addition of the hitherto uninsured to the pool of insured. The level of spending will change on account of the “induction impact,” i.e., on account of changes in utilization of health services by the already insured as they adjust to new boundaries of coverage, and the utilization patterns of the newly insured. The level of spending will also change on account of the “efficiency impact,” i.e., on account of changes in expenditures following efficiency improvement measures that may be instituted with Universal Health Insurance. 3.83 In the new environment that will inevitably follow the introduction of universal health insurance, the role of public expenditures and institutions will become even more significant. The expansion of UHI will undoubtedly place greater demand on public expenditures on health; at the same time, the introduction of UHI offers a unique opportunity for the introduction of efficiency and equity enhancing measures as well as

98 At the time of writing, the family medicine pilot in Duzce has already made significant headway, with training of over 100 family doctors completed and over two-thirds of the 300,000 population of Duzce registered with family physicians.

143

strengthening the overall effectiveness of the health system. Indeed, the introduction of UHI can be a fiscally viable proposition only if there are marked improvements in efficiency of public spending in addition to improvements on the revenue side of the social security system. Consolidation of hospitals and greater management autonomy 3.84 The transfer of all SSK facilities to the MOH was the first step towards a more unified network of health care providers. However, as documented earlier, there is ample room for improving hospital efficiency. Prior to the merger of MOH and SSK hospitals, public hospitals in Turkey were owned and run by MOH, SSK and Universities, and there was a lot of variation in their performance and efficiency levels. 3.85 Consolidation of 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 of care, and have significantly lower utilization rates compared to SSK and University hospitals. The fact that after the transfer of SSK facilities, the MOH has found itself with more than one hospital in a great number of places, especially in urban areas, offers the opportunity to carry out a re-organization of its 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 of UHI provides a good opportunity to further strengthen the gains from the merger of 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 of hospital beds in many provinces, further gains will come about only by improving efficiency in the use of hospital resources and overall management and accountability. The separation of provision and financing provides an opportunity to introduce innovative methods in management of health facilities, which can be achieved by granting financial and administrative autonomy to public hospitals. The introduction of hospital autonomy will require appropriate legislation that will allow for public assets to be managed outside the direct purview of the government, and related laws and regulation would need to be amended in order to facilitate the transformation of MOH and SSK facilities to autonomous bodies. This process has started, and the MOH is planning to implement a hospital reorganization pilot that would allow the testing of different hospital reform models in Turkey, but it is essential that momentum is maintained during the transition phase. Building the capacity of hospital managers to operate within an environment of greater autonomy is an important pre-condition for the success of hospital reorganization and provider payment reforms. Another pre-condition for success is the development and implementation of an appropriate financial and performance accountability framework. For example, under a model of 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

144

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 of services. The principal measure suggested to control quantity and dissuade frivolous use is the introduction of copayments for outpatient visits and for drugs (see chapter 4, section B3 for the estimated impact of various options). Even nominal levels of copayments will reduce utilization of health services and result in savings not only in the production of health services but also in the consumption of drugs (since almost all visits to a doctor result in a prescription). This is perhaps the only demand-side measure that will have the effect of rationalizing consumption. Managing 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 of 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 of total expenditures on health in Turkey, accounting for almost half of 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 of newer and more expensive drugs. One of the reasons why Turkey spends a huge amount on drugs and pharmaceutical products is that most of the insured population is insensitive to pharmaceutical prices, with out-of-pocket payments for medicines constituting between 10 and 20 percent of total medicine bill of the insured.99 3.89 Managing consumption of pharmaceuticals is critical in order to contain expenditures on drugs. Many countries have successfully adopted demand-side measures of controlling consumption, and cost-sharing has proved to be the most effective such measure.100 The consumption of pharmaceutical products among the insured is actually not low by international standards, and there is a strong scope for cost 99 Indeed, the high percentage of pharmaceutical expenditures in terms of overall health expenditures is as much a reflection of low overall expenditures on health as it is of high expenditures on 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 of drug expenditures in overall expenditures. 100 In the Netherlands, for example, the introduction of co-payments on prescribed pharmaceuticals (a fixed amount per prescription) led to substantial decrease in the total number of prescriptions. In Germany, drug cost-containment measures take the form of cost-sharing, prescription limitations, reference prices and the pharmaceutical spending cap that makes physicians’ associations liable for any overspending with no upper limit. These measures led to substantive decreases in pharmaceutical expenditures for social health insurance, mainly attributable to price reductions, changes in physicians’ prescribing behavior resulting in a reduced number of prescriptions by 11.2 percent and increased prescriptions for generics. The French government imposes a fine on pharmaceutical companies if pharmaceutical expenditures surpass budget ceilings either due to price or quantity increases.

145

containment if indiscriminate consumption can be curbed. The policy table at the end of this chapter provides a list of 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 of drugs through training for physicians on appropriate use of drugs and the development of standard treatment protocols. Reforming provider payment mechanisms 3.90 Changes in provider payment systems should introduce incentives for physicians to provide quality care at lowest costs. Some of these changes are already being planned, and a system of paying family physicians on the basis of capitation is being worked out. Physicians paid on the basis of a capitation fee per enrollee receive a fixed amount per enrollee regardless of the type and extent of treatment sought. Physicians participating in this scheme bear most risks of treating a patient, and therefore are likely to be conservative in the amount of health care they provide. Introducing a system of 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. 3.91 Likewise, prospective payment mechanisms101 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 of hospital payment reforms is to implement case-based payments. The implementation of 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 of 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 of budget management reforms, DRGs will not work in Turkey. An interim measure is to move to global budgets that cap health expenditures in hospitals, provide a disciplined environment for providers within 101 Prospective payment mechanisms rely on 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 to a pre-fixed rate per bundle. Such payment mechanisms do not encourage excessive use, since the hospital can conceivably make a profit (or a surplus) by being careful about inputs and hospital lengths of stay. One of the most widely-known prospective payment systems is the Diagnosis Related Group, or DRG. Developed to classify treatments according to the resource costs of its treatment, DRGs employ a complete and consistent coding system of patient-level information obtained from medical records to establish and cost bundles of appropriate inputs for one or 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.101 To be certain, a DRG-based system by itself will not necessarily promote efficient use of resources. Hospital care providers and managers need the flexibility and tools to actively manage their resources and redirect their use, which will ensure that cost-savings in treatment of one case are passed through the entire system.

146

which to behave combined with capacity building of providers to manage global budgets. While global budgets can lock in previous (inefficient levels of spending) and do not generate incentives for efficiency, when combined with the implementation of 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. Containing outpatient care provided by hospitals 3.92 A large number of outpatient services in Turkey are provided in hospitals, accounting for almost 43 percent of total costs of outpatient services.102 Outpatient services provided in hospitals cost significantly more than outpatient services provided in outpatient clinics, and it is imperative that the introduction of universal health insurance and family medicine be accompanied by a significant reduction in number of outpatient visits in hospitals that are paid out of the health insurance fund. This can be managed by establishing clear and transparent rules restricting reimbursement by the health fund of outpatient treatment carried out in hospitals. Managing claims from providers to financiers 3.93 The creation of the health insurance fund and accompanying internal control systems is expected to ease the problem of verifying 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 of health care services provided to the population should be clearly defined and its cost evaluated, while resources for the financing of this package should be allocated to the health insurance fund. The fund, in turn, should be responsible for the management of these resources and should not be bailed out in case its expenditures exceed its revenues, as it is currently the case with most social security institutions.

D. RURAL PUBLIC EXPENDITURES

D1. Rural Investments 3.94 The level of rural public investments is low in per capita terms owing to the large share of employment in agriculture. The overall level of rural public investments in Turkey is stable (at roughly YTL 350 million, in 1999 prices), amounting to 0.4 percent of GDP. The rural public investments figure is not limited to the agricultural sector; it also includes environment, forestry as well as rural infrastructure including roads and electricity. Rural public investment represents around 12 percent of total general government investment, which is equivalent to the share of agriculture in total GDP (Table 3.18). However, the level of rural public investment is still considered low,

102 MOH (2004).

147

due to the fact that rural employment is as much as 30 percent of total employment. At roughly US$ 30 per rural inhabitant, rural public investments appear to be low by international comparison, but rural employment is also declining as agriculture is being modernized. Key to improving the contribution of rural expenditures to the modernization of agriculture and rural development will be: (i) the redeployment of funds in the transfer budget into investment oriented programs; and, (ii) a shift of funds within the current non-transfer budget out of salaries of 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 of investments within the agricultural sector is skewed to irrigation. The State Hydraulic Agency (DSI) has accounted for an average of 45 percent of rural investments in 1999-2004 (Table 3.19). Greater emphasis is needed in areas such as land consolidation activities and storage and pre-marketing facilities (with particular focus on small farmers with weak links to markets). 3.96 Despite the high level of irrigation investment, the execution of these investments is very slow. DSI focuses on completion of main canals and head works in certain areas but leaves on-farm irrigation activities (mainly executed by the former General Directorate of Rural Services, GDRS) waiting in line for long periods.103 Over-programming and lack of funds to complete the planned projects in time causes delays and uncertainties. Currently, DSI would need more than 20 years to complete its 183 projects with the existing annual investment allocations. Though DSI investments expenditures have been increasingly focused away from the portfolio of dormant/ low 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 of beneficiaries can be more tangible in the medium term. 3.97 At less than a quarter of the total, rural non-agricultural investments are quite low. More than half of this investment is 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 of rural roads network and 478.7 km/1000 km2 rural road density Turkey can be considered quite 103GDRS was closed in 2004, with its functions allocated to the Ministry of Agriculture and Rural Affairs and Special Provincial Administrations.

148

developed (compared to Poland 343.9 km/1000 km2 and Thailand 124 km/1000 km2). Nevertheless, only approximately half of these roads are asphalt roads, the remaining being stabilized, graveled or earth roads, which require considerable upgrading. Regarding drinking water, 25 percent of villages and sub-villages and 15 percent of rural population is without drinking water facility. Therefore, investments particularly in these areas are of great importance for Turkey. Closer coordination between line agencies as well as SPO in determining priorities for rural investments is necessary if this imbalance is to be adjusted. 3.98 Participatory approaches in investments are needed. In order to improve focus and get better leverage of public investments in the rural sector, local communities, farmers’ organizations and/or NGOs should be involved in the design, implementation and management of investments. When investments become more demand-driven they usually gain greater relevance and can be held to higher accountability in terms of efficiency and timely execution. Moreover, when beneficiaries are given greater voice in deciding on the local focus of 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-Irrigation and Other Agriculture 9% 8% 7% 7% 7% 8% Ministry of Agriculture and Rural Affairs 4% 4% 7% 6% 7% 7% Ministry of Forestry and Environment 3% 3% 3% 4% 5% 5% GDF 6% 5% 5% 4% 4% 4% Other 5% 5% 6% 7% 7% 4% Rural Non-Agricultural 85 121 103 120 79 79 Total Share, of 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 of agricultural support to promote fiscal stabilization and allocative efficiency. By the end of 2002, the implementation of the reform of agricultural transfers had significantly reduced artificial incentives for inputs and particular crops and switched the main focus of agricultural transfers to the Direct

149

Income Support (DIS) Program.104 By 2004, the annual fiscal cost had been reduced from US$6.1 billion (3.1 percent of GDP) in 1999 to US$2.4 billion (0.8 percent of GDP). This has been achieved largely through elimination of credit subsidies and reduction of debt write-offs, and substantial reduction in state financed crop purchases and fiscal transfers to cover the losses of 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 of rural income support. In 2003-4, the DIS Program covered over 75 percent of farmers and accounted for three quarters of 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 $400m from international markets to purchase between 100,000-125,000 tones of hazelnuts at prices higher than market prices so as to support distressed producers. Such interventions are of concern because they: (i) represent a material deviation from agricultural sector policies of de-linking subsidies from production decisions; and, (ii) reflect a structural weakness in the budgeting process whereby duty losses could be created by TMO (and possibly other similar entities) through state guarantees. This type of allocation of public resources should be capped and subject to the scrutiny of the parliamentary process, while amendments to the public debt law may also be appropriate so as to limit the Treasury's ability to finance duty losses through state guarantees to be paid in the future. Table 3.20 - Total Agricultural Transfer Budget, 1999-2005 Amounts (million US$) 1999 2000 2001 2002 2003 2004 2005

PreliminaryDirect 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 ASCU Restructuring

0.0 0.0 0.0 0.0 34.2 69.0 9.4

Animal 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 %) 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 ASCU Restructuring

0.0 0.0 0.0 0.0 1.6 2.9 0.4

Animal 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 of Treasury and World Bank Staff calculations.

104 For details see, Turkey: Review of the Impact of the Reform of Agricultural Sector Subsidization” World Bank, 2003.

150

3.101 In 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 is the widely held view that the DIS Program, though serving the useful purpose of 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 of 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; ii) investment grants awarded on a competitive basis for agribusiness and other rural SMEs, as well as for public providers of infrastructure and services; and, iii) agricultural sector risk mitigation through partial coverage of crop insurance premia for farmers. These areas are the main beneficiaries of the planned reduction in the share of the DIS 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 3.103 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 is 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. 3.104 The rural consolidated budget (including all transfers, investments and recurrent costs) is characterized by roughly equal shares for transfers and recurrent expenditures (Table 3.22). The real level of the current expenditure budget has been fairly constant over the past few years, but the level of transfers was cut by roughly 60

151

percent in real terms. At the same time, the ratio of current to investment expenditure has been stable at roughly 2:1 over the past five years. Another striking fact is that, as much as 90 percent of recurrent spending of rural agencies is allocated for salaries alone. Therefore, there is limited funding (only 2-3 percent of 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 of Rural Budget 9% 16% 18% 27% 17% 17% b) Transfer - Share of Rural Budget 71% 50% 45% 31% 37% 43% c) Recurrent- Share of Rural Budget,: 20% 33% 37% 42% 45% 41%

- Of which, Personnel as Share of Rural Budget 18% 28% 34% 37% 41% 35%

Source: State Planning Organization, State Institute of Statistics. Notes: Totals exclude spending by local government, spending on SOEs, and spending on ‘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 of agricultural and rural development activities should be addressed (Box 3.5). When more than a third of the rural expenditures are on salaries of government employees, this leaves significantly small room in the budget for investments as well as other operation and maintenance expenses. However, it should be noted that rural investments are expected to be on an upward trend further to the implementation of the Rural Support (“Köy-Des”) project for clean water and rural roads. With greater emphasis given to participatory approaches and private sector involvement in service provision and expansion of the role of cooperatives, producers’ unions and NGOs, more effective and efficient use of 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 of the management of irrigation schemes to Waters Users Associations, but cut only 10 percent of its work force and still continues to allocate 25-30 percent of its budget for personnel. Another example is Ministry of Agriculture and Rural Affairs (MARA) whose budget is soaked up by personnel expenditures (70-75 percent of total budget) should consider contracting out mainly extension activities while reducing such personnel. Salary costs used to account for approximately 75 percent of the budget of General Directorate of Rural Services (GDRS). After its closure in March 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 of GDRS’s overstaffing hasn’t been solved but merely passed on to MARA and SPAs. Moreover, the ambiguity regarding the transfer of the 2005 budget of GDRS to the related agencies resulted in leaving ex-personnel of GDRS predominantly idle.

152

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 of the study summarizes some key 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 of 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 of new incentive systems and evaluation mechanisms for educational institutions, or a possible deepening of the recent pension system reform. The table provides an illustrative implementation time frame of 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 functioning of a sector and thus facilitate the realization of 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 of superior outcomes for a given amount of resources—such as in the case, for example, of higher pre-primary education spending, increased non personnel spending in public schools, or strengthening of overall rural investments and redeployment towards productivity-enhancing programs.

153

CHAPTER IV MANAGING MEDIUM 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—calling for large primary surpluses for a few more years.105 The development challenges on the way to the EU and the expenditure pressures in some key sectors are creating additional stress on public spending. Freeing up resources over the medium term 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 of expenditure pressures in view of 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 of medium-term expenditure trends in four key functional areas analyzed in the previous chapter: education, pensions, health care, and rural expenditures. Section three elaborates the framework for the assessment of expenditure pressures (or required fiscal space) by comparing the projected expenditures over the medium term with a sustainable expenditure envelope that is consistent with continuous progress in reducing the debt ratio—a precondition for preserving a sound macroeconomic framework. Different scenarios of expenditure pressures are examined depending on the presumed ambition of expenditure policy and the targeted reduction of public debt.

105 The level of public debt in proportion to GDP that will need to be reached to place Turkey in a comfort zone in the face of surrounding vulnerabilities will depend inter alia on the maturity and currency composition of the stock of debt. The still short average maturity of domestic debt and the large share linked to foreign exchange increase Turkey’s vulnerability to liquidity and exchange risks. These risks need to be mitigated by a low level of the debt ratio in proportion to GDP. With increasing average maturity and less foreign exchange exposure, the debt ratio consistent with a comfortable level of risk would, all else equal, be higher.

154

A. THE GOVERNMENT’S MEDIUM-TERM FISCAL STRATEGY 106

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. It also discusses how the authorities see future challenges and how these are addressed in policy documents. (i) Central Government 4.4 The first Medium Term Fiscal Strategy (MTFS) projects strong fiscal performance in 2006. The authorities have introduced their first medium term budgeting framework at the central government at the end of 2005. The medium term budgeting process has been initiated with the issuance of 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 of 2006-2008 and it included revenue projections and indicative expenditure ceilings for 2007-2008 alongside the appropriations for 2006.107 Considering first the 2006 budget, the targeted primary surplus of about 5 percent of GNP is consistent with the IMF program target of 6.5 percent of the GNP primary surplus for the whole public sector.108 Table 4.1 below provides the institutional breakdown of the public sector primary balance according to the IMF program definitions while Annex II provides the detailed economic and functional classification of Central Government expenditures over 2004-2008. As can be seen, the central government budget is expected to produce 5.1 percent of GNP primary surplus in 2006, while the remaining part is expected to come from the rest of the public sector. The projected primary surplus target of the central

106 At the time of drafting, the new Medium Term Fiscal Strategy (MTFS) covering the period 2007-2009 was not released by the government and therefore the analysis in this section is based on the previous MTFS which covered 2006-2008. However, this section will be updated in line with the new MTFS upon receiving the related data from the government authorities. 107 The indicative ceilings were not provided for the Regulatory and Supervisory institutions as these institutions are not subject to article 16 of the PFMC Law. 108 It is still not clear at this stage how the IMF will follow the public sector primary balance for 2006 and beyond since the coverage of the central government budget includes new institutions such as Regulatory and Supervisory Agencies (RSAs), which were not part of the public sector definition of the IMF. Until 2005, they were only covered within the context of transfer from the budget, not with their total revenues and expenditures. Although the impact of the newly included institutions on the primary surplus is negligible, it still needs to be clarified whether there will be any changes in the definition of the consolidated public sector primary balance definition to cover the whole central government budget institutions.

155

government is 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 of SEEs is mainly the reflection of the privatization of 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 of 0.2 percent of GNP in 2006.

Table 4.1: Institutional breakdown of the primary surplus, 2004-2006 (% of 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 of 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 of 4.6 percent in 2007 and 4.5 percent in 2008 for the central government.109 This will require that the rest of the public sector produce more than 1.5 percent of the GNP, to reach the public sector primary surplus target of 6.5 percent of 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. It should be noted, however, that these projections are surrounded by uncertainty as the Government has revised the MTFS 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 of Medium Term 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 Medium Term Fiscal Strategy which aims to map macro and sectoral policies and priorities of the MTP into the budget was issued on July 15, 2006. Although the new MTFS has a three years perspective, the main focus of the strategy is still on the current (first) year. In this regard, while there is only 1 percent divergence between the 2006 central government non-interest expenditures realization estimate and the 2006-2008 MTFS target, this difference is slightly

109 The analysis in this section is based on the budget and GNP figures presented in the MTFS.

156

Table : 2007 Central Government Budget (% of GNP) 2006 2007 1/ Total central government revenues 30.6 29.8 I. General Budget Revenues 30.0 29.2 Tax revenues 24.5 25.0 Non-tax revenues 5.5 4.1 II. List II and III Institutions net revenues 0.6 0.6 Program Definition Primary Revenues 29.2 28.9 Total Expenditures 2/ 31.2 32.5 Personnel 7.6 8.5 Goods and Services 2.3 2.5 Interest 8.2 8.4 Current Transfers 10.0 9.6 Capital Expenditures 1.9 1.9 Capital transfers 0.5 0.6 Lending 0.7 0.6 Contingencies - 0.4 Program definition primary expenditures 22.9 23.9 Program definition primary surplus 6.3 5.0 1/ 2007 data are from the proposed budget 2/ The 2006 goods and services and current transfer data has been adjusted in order to have a comparable data with 2007. Health expenditures for civil servants and green card holders has been deducted from the goods and services category and added to capital transfers. Source: MOF, and staff calculation

above 13 percent for 2007 and 2008 data presented in the last MTFS. Some of this divergence can be explained through higher than expected salary increase for the civil servants in 2006 and increased estimate of the central government expenditures for universal health insurance premium for the poor due to increased number of green card holders. These differences highlight the need for more attention in the preparation of the MTFS so as to enhance its credibility. The Government revised its MTFS in less than three months after it was announced in July 2006, when submitting the budget to the Parliament in October 2006, which may also affect the credibility of the MTFS. Since the details of 2008 and 2009 budget are not available, the following analysis focuses only on 2007 central government budget. The 2007 central government budget provides a consistent framework with the 6.5 percent target of public sector primary surplus under the IMF program definition, with a central government primary surplus projected at 5 percent of GNP. Although the 5 percent of GNP primary surplus target for 2007 is quite lower than the estimate for 2006, the better than expected performance of the 2006 budget is a reflection of one-off revenues such as social security arrears restructuring collection and collection from the former owners of SDIF banks (total of 1.4 percent of GNP according to IMF calculations). This led to a reduction in the deficit of the social security institutions’ deficit and therefore current transfer of the central government. While the revenue side of 2007 budget does not foresee any policy change, from the expenditures side, the main change in the assumptions of 2006 and 2007 is the introduction of Universal Health Insurance system and of a new social security reform law. The increase in personnel expenditures partly reflects the introduction of the new social security legislation, as the premium base of the civil servants will be broadened to cover almost their entire salaries so that the Government will pay an increased amount of social security contributions. As it is envisaged in the enacted social security reform legislation, the Government is required to pay premium increase of the civil servants to 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. As a result, the quality of 2007 fiscal stance requires special attention since the fiscal space gained from the social security transfers will be allocated to 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

157

of GNP in 2005 to 29.8 percent in 2008. Although the breakdown of 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 of GNP. At unadjusted basis, the main declining revenue items are:110

• From the tax revenue side (1.4 percent of the GDP) The already announced reduction in the corporate income tax rates from

30 percent to 20 percent (0.6 percent of the GDP) The negative impact of the falling interest revenues on personal income

tax due to the decline in inflation and reduction of number of brackets as well as upper rate for personal income tax (0.3 percent of the GDP)

Fall in special consumption tax (STC) revenue due to not increasing the rates (0.7 percent of the GDP)

• From the non-tax revenues ( 1.1 percent of GDP) No dividend contribution from Turkish Telecom starting in 2007 State Bank dividend contribution expected to decline

It should be noted that to the extent lower income tax rates tame incentives for tax evasion, the impact on tax collections could be mitigated over the medium term. However, for effectively reducing tax evasion, reduced CIT and PIT rates should form part of 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.111

Table 4.2: Central Government Budget Revenues in the MTFS

2005 1/ 2006 2007 2008

Tax revenues 27.1 27.0 26.3 25.7 Taxes on 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 on Income Profits and

Capital Gains 0.0 - - - Taxes on Property 0.4 0.9 0.8 0.8 Of which inheritance and gift tax 0.0 0.0 0.0 0.0 Domestic Taxes on 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

Of which Motor Vehicle Tax 0.5 0.6 0.6 0.6 Taxes on 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

110 The figures may not sum up to the decline in budget total revenues due to changes in other items not stated here. 111 OECD, “Economic Survey: Turkey 2006” provides an overview of options to improve formalization.

158

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

1/ 2005 data are adjusted with the revenue share of the local governments and funds, and YTL 3,050 is deducted from the interest revenues under the non-tax category Source: MOF, and World Bank Staff calculation.

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 is also projected.

Table 4.3: Adjusted Primary expenditures of the central government (% of GDP) 2005 2006 2007 2008 01 – Personnel 6.5 6.6 6.5 6.2 02- Social Security contribution 1/ 1.4 1.4 1.2 1.1 03 - Goods and Services 2/ 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 payments 9.4 8.5 7.4 6.0 Note: 2005 data are for the consolidated budget. 1/ Public personnel health expenditures under goods and services has been moved to social security premium in 2005 and 2006 for the purpose of comparability. 2/ Health expenditures for public personnel and green card holders has been deducted from the goods and services data in 2005 and 2006. 3/ 2005 data was adjusted with the revenue share of local governments and funds, and health expenditures for the public personnel and green card holders moved to social security transfer under the current transfer. Source: MOF, and World Bank Staff calculation

4.8 Social Security Transfers from the consolidated budget are targeted to decline by 0.7 percent of GNP between 2005 and 2008. As a result of the social security reform adopted in 2006 (chapter 3, section B) the social security deficit is expected to cease to grow and decline by approximately 1 percent of 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 of 4.5 of GNP. With the new social security reform legislation Turkey will move to a UHI system starting in 2007. This will bring

159

some changes with respect to the reporting of social security transfers. Therefore, the social security system deficit announced by the MTFS for 2007 and 2008 requires some adjustment in order to have comparable data. Without any adjustment, the total social security transfers of the MTFS do not show any decline. However, after the adjustments (explained in Box 4.2 below), the overall deficit of social security transfers is expected to decrease to 4.1 percent by the end of 2008 from 4.8 percent in 2005.

Box 4.2: The 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 of this system, all 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 civil 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 of 2007 health expenditures related to green card holders (the poor) will be converted into premium payments and reported as transfers to social security institutions under current transfers rather than as goods and services expenditures.

-- Second, health expenditures for civil servants previously reported under good and services will be reported as social security premiums of public personnel.

Therefore, in order to have comparable data for the social security transfers consistent with the 4.5 percent deficit target, YTL 2,400 and 2,500 million expected health premium payments to be made by the Government for poor people in 2007 and 2008 respectively have been deducted from the transfers to SSIs.

Table: Adjusted Social Security Transfers for 2005-2008

(YTL million) 2005 2006 2007 2008 SSK 7,507 5,494 Bag-Kur 6,863 6,750 Emekli Sandigi (including duty losses) 8,946 11,103 Unified Social Security Agency 23,778 26,411 Total SSIs transfers 23,316 23,347 23,778 26,411 (% of the GDP) 4.8% 4.3% 4.0% 4.1% Memo item Emekli Sandigi (Duty Loses) 3,397 4,403 4,562 4,403

Unadjusted and Unadjusted SSIs deficit

3.6%

3.8%4.0%

4.2%

4.4%

4.6%4.8%

5.0%

2005 2006 2007 2008

(% o

f GD

P)

Unadjusted social security deficit adjusted social security deficit

160

(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 of the social security system is expected to compensate for this decline and leave the total revenues close to the level in 2005. The introduction of the UHI in 2007 is estimated to increase premium collections in the social security system. Table 4.4: General Government Revenues and Expenditures

(% of 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 of 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 of the general government is projected to increase by almost 1.5 percent of 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 of the central government, the primary surplus of 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 is not expected to be compensated by revenue increases. However, it is 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 of future expenditure pressures or measures for

161

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

(% of GDP) 1999 2004 2005 2006 2007 2008 Consolidated Budget 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 Revolving funds 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 amplifying the needs. Although, the government’s economic policy documents lack any explicit consideration of 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.

162

B1. New and emerging education sector policies in Turkey 4.13 The Government has proposed several new policies that, if 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:

• Fifteen New Public Universities • Grants and Loans to Parents Sending their Children to Private Schools. This

policy change was part of 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 of this policy has been analyzed below for a illustrative purposes.

• Extending school years to one more year In addition to these three policies, the Government is also seeking to increase the secondary school enrollment rate from its current 61 percent to the EU target of 85 percent. 4.14 Early in 2006, Parliament approved the opening of 15 new public universities. This decision has increased the number of public universities which are classified as special budget agencies from 53 to 68.112 Most of these new universities were established by separating parts of 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 of required staff is 20,157 academics and 6,024 administrative staff. The Ministry of 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 is 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 of 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 of tuition fees for parents sending their children to private schools. According to this draft legislation, additional subsidy would also be provided to cover half of the interest costs of loans that 112 The new Universities are: Ahi Evran University; Kastamonu University; Duzce University; M.Akif Ersoy University; Usak University; Rize Universitesi University; Namık Kemal University; Erzincan University; Aksaray University; Giresun University; Hitit University; Bozok University; Adıyaman University; Ordu University; Amasya University

163

parents take on to pay for private education. Part of the costs of 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 of subsidies as a means of 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 it 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 own 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 is already low. The increased demand from small towns and villages, or from low income families in big cities, is 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 of private sector initiative. 4.18 MONE has announced that the school year of 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 of 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 of 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 Box 4.3 for data sources and methodology). In the light of 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 of GDP by the education institutions of which 4.09 percent goes to directly to education functions, with the remainder going to other functions, such as general administration, health etc.113 Since 113 This amount includes the health expenditure of civil servant which are recorded as an education expenditures by the MONE.

164

there is no announced expenditure framework of education institutions, the year 2009 projection is 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 from the final accounts of MONE, the Medium

Term Fiscal Plan of the 2006 budget, the 2006 budget accounts and related background tables prepared by the Ministry of Finance (MOF), State Planning Organization (SPO) documents about extra budgetary funds (EBFs) and revolving funds (RFs). Under the heading "Central Government Budget," the analysis examined funding classified as MONE, RFs of MONE, Universities, and RFs of Universities and other budgetary institutions (such as education institutions of MoH, General Directorate of Police, Credit and Dormitories ). The data for RFs came from SPO and from MoF for 2004-2006. For the following years, the analytic tables included interpolated data to adjust for inflation and growth. The Treasury subsidies to Credit and Dormitories were kept under MONE expenditures of tertiary education.

The education-related expenditures of the Social Solidarity and Support Fund are classified as “not classified elsewhere” as these expenditures are not directly related to education but more to do with scholarships and pro 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 of Police Force (such as police academies), and the Ministry of Health (including training schools for nurses) .

SSK, Bagkur, Emekli Sandıgı and the Employment Agency come under this heading and data suggest very minimal education (probably mostly training) expenditures of such institutions.

4.20 The medium term expenditure framework of the government, updated in December 2005, does not include the new policies discussed above. The present analysis is therefore constructed by first preparing the total education sector medium term expenditure framework without the policies included. Then, the estimated total cost of new policies is projected and reflected in the expenditure framework. In the absence of new policies, the baseline projections presented below suggest that the expenditure level of the education sector in the coming years would have been slightly reduced, as well as the total expenditure level of central government institutions. Thus, the recent trends in decreasing public sector expenditures on education would continue in the next three years.114 The likely impact of the new policies has been projected based on the assumptions described in Box 4.4. Box 4.4: Projecting the budgetary impact of new education policies and proposals Opening new 15 Public Universities

> By using the approved and released academic and administrative staff cadres, the assumption is that such cadres from 2006 through 2009 will be employed gradually.

114 The difference between total expenditure and total education expenditure at the public education sector are due to (a) headquarter administrative expenses of MoNE and Universities, (b) health expenditures of the educational institutions, and (c) environmental expenditures (especially at the universities).

165

> Current education system data for average salaries of academicians and administrative staff are used to make forward estimates by adjusting for inflation.

> No assumption is made with respect to cost of additional construction as most of the universities will likely be using the buildings of the institutions that were part of the existing universities.

> Te average cost of new office equipment and stationary will be 10 percent of 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 of private school students by the levels of education, it was possible to increment such numbers under different scenarios. The total cost of state aid of YTL 1000 per student was estimated under such scenarios.115 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, it was possible to estimate the total cost of the interest subsidy.

> Savings of operational expenditures have also been estimated due to a reduction in the number of students in public schools who would transfer to the private schools.

Extending the duration of secondary schooling from three to four years

> Both capital and current expenditures including personnel have been estimated. Costs of constructing new classrooms needed to accommodate students for the additional school year are based on the assumption that 50 percent of the classroom construction costs will be incurred in 2008 and 50 percent in 2009. The cost reflects the fact that the new policy will not take effect until 2009, when students entering ninth grade in 2005 will 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 of 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 is that that they would be effective in 2009.

Expanding the secondary education enrollment rate towards EU convergence

Finally, it was assumed that Turkey plans to increase enrolment rate of secondary school from 61 percent in 2005-2006 to 70 percent in 2008-2009 related with the EU targets of 85 percent secondary completion rate by 2015.

115 Assumptions: no increase in number of students assumed for 2006; 20 % incerase is assumed with respect to 2006 in number of students for 2007; 40 % incerase is assumed with respect to 2006 in number of students for 2008; 50 % incerase is assumed with respect to 2006 in number of students for 2009.

166

Figure 4.2 (a – d). Total Education Expenditure Framework Without and With New Education Policies and proposals(2004-2009)

TOTAL EXPENDITURE LEVEL at current Prices EDUCATIOAN EXPENDITURE LEVEL at Current Prices

TOTAL EXPENDITURE LEVEL as a share of GDP EDUCATION EXPENDITURE LEVEL as a share of GDP

16.000

18.000

20.000

22.000

24.000

26.000

28.000

30.000

32.000

34.000

36.000

2004 2005 2006 2007 2008 2009

No Policy with Policies

3,5

4,0

4,5

5,0

2004 2005 2006 2007 2008 2009

No Policy with Policies

3,50

4,00

4,50

5,00

2004 2005 2006 2007 2008 2009

No Policy with Policies

16.000

18.000

20.000

22.000

24.000

26.000

28.000

30.000

32.000

2004 2005 2006 2007 2008 2009

No Policy with Policies

Source: World Bank Staff calculations 4.21 All of the new policies and proposals will increase the financing needs of 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 of GDP in 2006, the additional expenditures required to continue the implementation of the new policies in subsequent years would reach 0.44 percent of GDP in 2009. The economic breakdown of additional expenditures is shown in Table 4.6.

Table 4.6. Breakdown of Additional Expenditures Needed for Financing New Policies

Personnel Operational Capital Transfer Total

In percent of GNP 2007 0.02 0.00 0.08 0.11 2008 0.04 0.00 0.21 0.09 0.34 2009 0.13 0.01 0.21 0.09 0.44

Source: World Bank Staff calculations 4.22 More than 85 percent of the additional fiscal pressure will be incurred by MONE's budget (Table 4.7). Each year, MONE would need to find additional sources of financing. For some years, namely 2008 and 2009, this incremental amount will exceed

167

10 percent of its total resources. For universities, the additional amount of financing will be around 5 percent. More than 60 percent of the additional cost is for capital expenditures by MONE, mainly for new construction and equipment in 2008 and 2009. After capital expenditures, the biggest demand will be to cover the costs of transfer expenditures and personnel expenditures. In the case of higher education, the additional expenditure pressure will be for personnel. Table 4.7. Total Education Expenditures by Institutions (% of GDP)

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 of 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 is to improve the efficiency of 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 (i) The deficit of the pension system 4.24 Social security system deficit is set to decline. With the enactment of the new social security laws (see chapter 3, section C), revenues of the system are projected to increase by 0.3 percent of GDP until 2009. The unification of the pension institutions and data bases as well as introduction of the UHI will contribute to better revenue collection. Furthermore, the reforms will also bring about a reduction of 0.2 percent of GDP in expenditures. As a result, the social security system deficit is estimated to decrease by half a percentage point until 2009. Table 4.8: Social Security System Deficit

(Percent of 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

OtherNo Policy with Policies No Policy with Policies No Policy with Policies

2004 3.16 3.16 1.13 1.13 0.33 4.62 4.622005 3.09 3.09 1.10 1.10 0.40 4.58 4.582006 3.11 3.17 1.15 1.15 0.44 4.70 4.762007 3.04 3.12 1.09 1.12 0.41 4.54 4.652008 2.93 3.22 1.05 1.09 0.38 4.36 4.702009 2.93 3.32 1.05 1.11 0.39 4.37 4.81

MoNE Universities Total

168

(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 of Universal Health Insurance in January 2007. Health care costs of Green Card holders were expected to increase following the introduction of UHI, but a bulk of this increase has already taken place even before the introduction of UHI with the introduction of a large number of Green Cards effectively extending health insurance protection to almost all the poor. Bulk of the increase in health expenditures in the future is likely to come when the presently uninsured non-poor take up UHI upon its introduction and as utilization of 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 of increase of 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 is expected that utilization of health services will increase at a rate of 10 percent till 2008 and at a rate of 7.5 percent in subsequent years. 4.26 The authorities worked on a series of cost saving measures. The measures became effective on July 1, 2006, which are collectively expected to result in a net saving of YTL 1.4 billion in the second half of 2006, include:

(i) Global budget protocol, underpinned by payment-per-case system for MOH hospitals effectively reducing the average price per hospital-based outpatient visit by 20 percent, which is expected to yield a net saving of YTL 600 million by end of 2006.

(ii) Introduction of per-case payment for private hospitals at rates equal to MOH hospitals, which is expected to yield a net saving of YTL 400 million by end of 2006.

(iii)Introduction of per-case payment system for university hospitals, with rates set at 50% above the new MOH tariff for outpatient care and equal to the MOH tariff for inpatient services, which is expected to yield a net saving of YTL 300 million.

(iv) Reduction in drugs from the positive list that will be reimbursed by public funds, which is expected to yield a net saving of YTL 100 million.

4.27 Taking into account the effect of these measures, health expenditures in 2006 are projected to be YTL 29,711 million. Future expenditure projections are based on the following assumptions: (i) most of 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;116 (iii) 10 percent of 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

116 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.

169

outpatient visits will continue to increase at the current rate of 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 of 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 2009.117 Health expenditures by MOH and Universities as directed through their annual budgets are assumed to increase at a rate of 5 percent every year. As a proportion of 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 2011. As indicated in Table 4.9, total health care expenditures in nominal terms are projected to increase by 92 percent from 2006 to 2011. 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 of GNP 5.3 5.5 5.8 6.0 6.34 6.56

Source: World Bank Staff calculations (iii) Containing health care expenditures—illustrative scenarios 4.29 The introduction of 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 of health care services. The principal measure suggested to control quantity of medical services and dissuade frivolous use is the introduction of co-payments for outpatient visits and for drugs. Even nominal levels of co-payments will reduce utilization of health services and result in savings not only in the production of health services but also in the consumption of drugs (since almost all visits to a doctor result in a prescription). This is perhaps the only demand-side measure that will have the effect of rationalizing consumption. Table 4.10 presents the expected revenue generation from outpatient co-payments of YTL 2 that the Government proposed to introduce in 2007, increasing to YTL 5 by 2012. Introduction of co-payments would also reduce number of outpatient visits as well as pharmaceutical consumption. The table presents the expected savings in utilization of health services and drugs resulting from the introduction of co-payments for outpatient services.118

117 At the same time, revenues are also expected to increase, particularly because of the presently non-poor joining UHI and making insurance contributions. Revenue projections have been not been carried out at this time. 118 Even though price elasticity for demand health services in Turkey is estimated to be low, it would not be unreasonable to expect the introduction of a 2 YTL co-payment for outpatient health services to result in a 5% reduction in utilization of health services.

170

Table 4.10: Estimated Revenue and expenditure savings from Co-payments (million YTL)

2007 2008 2009 2010 2011 Revenue from Co-payment 473 497 509 522 535 Reduced Expenditure on OP Services 249.2 261.6 268.2 274.9 281.7 Reduced Expenditure on Drugs 483.5 501.6 511.0 520.6 530.3 Revenue + Savings 1,206.0 1,260.3 1,288.7 1,317.7 1,347.4

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 of these measures is the introduction of co-payments, as discussed above. Other measures proposed to be taken by MOH, including rationalizing drug use, widening the base for reference pricing of 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 of 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 of 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 of hospital physicians with one which will discourage unnecessary use of 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 of agricultural transfer programs financed from budget resources, and sets a target level of not less one percent of GDP. The projections below simulate a gradual increase to the 1.0 percent of GDP target. The other assumptions are annual 5 percent real GDP growth, and constant shares of 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 of personnel expenditures slightly decreases to 33 percent, without any reduction in personnel numbers. If the government tries to tackle the problem of overstaffing through downsizing efforts, the share of personnel costs can be further reduced in the medium

171

term. As explained in chapter 3, this could facilitate allocation of 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 273,227 286,888 301,232 316,294 332,109 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% 1.00% 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% Of which, Personnel: 35% 35% 34% 33% 33%

Source: Undersecretariat of Treasury, SPO and World Bank Staff Calculations 4.34 Fiscal space could also be created by a partial switch out of transfers to investments and by use of EU Pre-accession funds. Overall, the cumulative increment in investments outlays (above the current level) would be roughly US$300 million spread over 3 years (2006-2008). The challenge is that even achieving this real 22 percent (relative to 2004) increase in investment outlays (albeit of improved quality and efficiency), would leave rural investment levels at a stable share of GDP. The EU Instrument for Pre-Accession for Rural Development (IPARD) 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 of the transfer budget into investment-focused rural development programs, which could also be partially financed by the EU IPARD. 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 of available budgetary resources; changes in the composition of public spending; key features of the medium-term macroeconomic environment; and the underlying debt dynamics. This

172

section brings together all these considerations and provides an analysis of the additional medium-term fiscal space needed in the light of the available budgetary means.

C.1. A sustainable medium-term expenditure envelope

4.36 Current and expected policies, 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 of the public debt ratio towards a desired level. In a baseline scenario, the total envelope for the financing of primary expenditures is calculated by subtracting a predetermined primary surplus from the estimated total primary revenues. Such a scenario allows a reduction of 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 of GNP in 2005, primary revenues were recorded at 40 percent of 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 Projections (% 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 of Finance and World Bank Staff calculations

173

4.38 In a baseline scenario, the general government primary surplus is assumed to stay at the IMF program target level until 2009. The high level of public debt to GNP, the still risky composition of 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 of 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 of primary surplus would deliver a debt to GNP ratio of 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 1.0 1.0 GNP growth rate (%) 7.6 4.5 4.0 5.0 5.0 Primary Balance (% of 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 of 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 of inflation and exchange rate movements is 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 of decline in debt to GNP ratio that can be achieved until 2009 is highly sensitive to the level of 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 its resilience to shocks and preserve high growth rates for a sustained period of time. In this context, a higher primary surplus would serve more to the purpose.

174

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 is relatively high and its maturity is short. Turkey needs to aim at much lower debt to GNP ratio in order to ensure its sustainability and reduce vulnerabilities in its macroeconomic environment. The sensitivity of 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, it 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 of the currency. 4.42 A base line scenario is constructed for analyzing medium-term expenditure pressures based on the continuation of the current fiscal policy stance. A budget envelope that would be available for financing primary expenditures over the next three years is calculated with the assumption that the primary surplus would be kept at 6.5 percent of GNP for the whole public sector. Given that the contribution of the SOEs to the primary surplus is declining as the most profitable SOEs are being privatized, the assumption of constant primary surplus ratio to GNP would mean that the general government has to compensate for the falling fiscal contribution of SOEs. While calculating the primary surplus for the general government, the contribution of SOEs is deducted from 6.5 percent primary surplus and then the remaining primary surplus which needs to be generated by the general government is converted into SPO definition (see Box 1 of Chapter 1 for the conversion methodology). Within this context, the total amount of resources that would be available for financing primary expenditures is estimated to be around 33 percent of GNP until 2009. Table 4.14: Total Spending Envelope for Primary Expenditures. Actual Forecast (% of GNP) 2004 2005 2006 2007 2008 2009 Total spending envelope for primary expenditures

32.9 33.8 33.4 33.4 33.3 33.2

Source: World 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, it is expected that the current policies—including the recently introduced social security and health reforms—would continue. Additionally, it is assumed that the government starts implementing the announced policy changes in the area of education (see Table 4.7 in section B1 of 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

175

spending for EU alignment). Considering first the investment needs in infrastructure, as it is 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, it is assumed that general government investment in infrastructure would increase by 0.5 percent of GNP during 2007-2009. Similarly it is assumed that environmental spending would increase by 0.1 percentage points of GNP in 2007 and 0.2 percentage points of GNP in 2008-2009. These are rather small numbers to facilitate any alignment with the EU Acquis but a gradual approach is thought to be more appropriate and realistic. Table 4.15: Expenditure Pressures in Key Areas Actual Additional Spending Need (% of GNP) 2004 2005 2006 2007 2008 2009

Economic Affairs 4.7 4.4 -infrastructure investment 0.0 0.5 0.5 0.5Environmental Protection 0.2 0.2 0.0 0.1 0.2 0.2Education 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 of GNP in 2006, rising to 2 percent by 2009. 4.45 In 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 of such policies would be around 0.3 percentage points of GNP, which would raise the estimated expenditure pressures in Table 4.16 to 1.7, 2.2 and 2.5 percentage points of GNP in 2007.2008 and 2009 respectively.

176

Table 4.16: Medium term Expenditure Pressures, Baseline Scenario Actual Forecast (% of 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.1 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 (A1-B) / Needed Fiscal Space 0.0 0.0 0.6 1.1 1.7 2.0

Source: SPO, World Bank Staff calculations * Figures for the health expenditures of the general government differ by about 1 percentage points of GNP from the ones presented in Table 3.16 because of some methodological reasons. 4.46 In a third scenario, the required fiscal 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 US Fed and ECB and a “flight out of risk” due to perceived mounting inflation pressures worldwide—domestic developments that occurred prior to this episode of global volatility have magnified the incidence into local markets. The currency depreciated by 20 percent, intensifying inflationary pressures, while the Central Bank of Turkey had to increase its policy rates by 4 percentage points within less than a month in response to these developments. In the face of such abrupt changes in interest rates and the value of the exchange rate that have a strong repercussion on the fiscal deficit and debt dynamics, the high debt to GNP ratio and its structure are still 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. As discussed before, primary budget surpluses and economic growth are now the major driving forces behind the fall in debt ratios. Therefore, further reductions in the total available spending envelope would be required

177

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 of the YTL. The calculations show that the total spending envelope would need to be reduced by an additional 1 percentage point of GDP (Table 4.17). Table 4.17: Primary Surplus Needed for Achieving Pre-turmoil Debt Ratios Actual Forecast

(in % of 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 (%, Dec/Dec) 7.7 9.7 6.5 5.0 4.0 Depreciation (Dec/Dec) 0.1 20.0 1.9 1.0 1.0 GNP growth rate (%) 7.6 4.5 4.0 5.0 5.0 Primary Balance (% of GNP) 6.2 7.5 7.5 7.5 7.5

Source: World Bank Staff calculations 4.48 Expenditure pressures can reach as much as 3 percent of GNP if the debt to GNP rate is targeted as 51 percent in 2009. Under the assumption of 7.5 percent primary surplus, the estimated expenditure pressures would reach 3 percent of GNP if no additional spending is made in education for reducing student per classroom to 30. The expenditure pressures would reach as much as 3.3 percent of 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 (% of 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 enrolment ratio and reducing student per classroom to 30.

-1.6 -2.1 -2.7 -3.0

As above with additional spending in education for increasing preschool enrolment ratio and reducing student per classroom to 30.

-1.6 -2.4 -3.0 -3.3

Source: World 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 of 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 is already high and more tax revenues should only be a marginal source of fiscal space. If anything, fiscal space should be created in due time in order to reduce the tax burden.

178

Structural reforms that improve the efficiency of 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 of expenditures should also be sought, with the aim of redirecting resources from functional areas where expenditures are oversized to priority areas for growth and social development. At the same time, it is important that the ambitious reform agenda of public financial management initiated since 2001 (see chapter 5) be properly implemented, to prevent wasting of 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 of these reforms have been enacted or planed but implementation will be a key challenge. For example, the 2006 pension reform will be instrumental in capping the deficit of the pension system and gradually creating fiscal space for more productive expenditures. Strict implementation of the parametric changes and strengthening of the contribution base will be needed to ensure that the projected fiscal savings indeed materialize over the coming years. Similarly, the implementation of the health care reform agenda with introduction of 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 of personnel needed to support an increasing student population. Similarly, hospital consolidation will improve capacity utilization and allow more efficient operation while introduction of prospective payments systems at the hospital level will improve incentives to contain costs. 4.51 Changes in expenditure composition could reflect deliberate policy trade offs and should rely on greater private sector participation in the provision of 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 of infrastructure services. This option should be exploited dynamically because there are very limited options for a significant increase in the financing of infrastructure in the coming years. At the same time, education sector reforms should redirect private resources from low 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, of establishing a distinction between “core public service”, to be performed by civil servants, and public services that could be delegated to contracted personnel, would also facilitate greater flexibility in the use of 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

179

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 of interest income. As previously analyzed, although more refined estimates would be necessary, tax expenditures—such as tax exemptions of allowances and side benefits to civil servants, or the deductibility of social security contributions from taxable income—carry a significant cost in terms of foregone tax revenues. Rationalization of 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 of the recent market turmoil on the debt ratio could thus be offset and the required fiscal space to bring the debt ratio back on its previous track would be significantly reduced.

180

CHAPTER V STRENGTHENING BUDGETARY INSTITUTIONS FOR EFFECTIVE GOVERNMENT

5.1 Longstanding institutional weaknesses in budget and financial management had often contributed to fiscal crises in the past. Public expenditure management was hampered by budget fragmentation—in particular, widespread off-budget arrangements; lack of a comprehensive resource framework; and divided responsibilities for formulation, execution, and oversight of the budget. These weaknesses contributed to high degree of non-transparency, which also impeded effective policy guidance to the budget in both aggregate fiscal management and in evaluation and management of sector policy. Turkey’s excessive use of quasi-fiscal policies and the poor coverage and control of contingent commitments had exposed the government to fiscal risks. A plethora of 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 of priority actions to improve budget coverage, formulation, execution, accounting, and audit. A far-reaching overhaul of the laws governing public finance management has been undertaken over the past few years. Ensuring the comprehensiveness and transparency of public expenditure was identified as a fundamental priority for reform. Moreover, the new laws go beyond the initial reform strategy in a number of respects—inspired in part by the prospect of EU accession and by the government’s desire to introduce greater performance orientation in budget management. However, in some instances implementation of the reform agenda has been hesitant, and some regressive tendencies in the implementation of the PFMC law tend to perpetuate old problems of budgety fragmentation. It 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 of the public sector reform has been expanded to cover the administrative aspects of the public sector. Therefore several other laws have been or are being enacted (Table 5.1). Expansion of the public sector reform will add additional challenges for the implementation of the reform. Sequencing and complementarity of these reforms needs to be carefully designed for a successful implementation of public sector reforms.

181

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 of 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 it was reenacted in July 2005. However, one article was referred to the Constitutional Court by the President. Secondary legislation is being prepared.

Metropolitan Municipalities Law Passed in July 2004. Secondary legislation is being prepared.

SPAs and Municipalities’ Revenues Law

A draft is submitted to Parliament in September 2006.

Civil Service Law A draft is prepared but not submitted to Parliament as of October 2006.

Regional Development Agency Law Enacted in January 2006.

Turkish Court of 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 of a major budget and public financial management reform by itself needs strong coordination and monitoring. Because of their impact on the whole of 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 of 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 of provisions being enacted that are inconsistent with or contradict the previous reforms. It is crucial that Turkey maintain the momentum and credibility of these major reforms; so the need for strong, effective coordination to reconcile conflicts cannot be overstated. 5.5 This chapter assesses the status of 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 of the status of public financial management reforms further to the presentation of the 2006 budget. The following sections break down the challenge to public financial management in Turkey along the following lines: budget formulation, accounting and

182

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 of the PFMC Law in 2003 marked a defining moment for public financial management. The Public Financial Management and Control (PFMC) Law (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 modern performance-oriented public sector management. It provides the first comprehensive framework for public finance management using a modern Government Finance Statistics (GFS) concept for the coverage of revenue and expenditure. It clarifies to the public the nature of ministerial and official accountability, and it strengthens public expenditure and financial management processes in line with EU practice. The Public Financial Management and Control (PFMC) Law was amended in December 2005 with the main changes indicated in Box 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 of this year’s budget will also be carried out in line with the PFMC structure. This will enable the Ministry of Finance to consolidate the budget realization for all general government institutions. Details on the implementation of the PFMC law are provided in Box 5.2. 5.8 Although full implementation of the PFMC law is expected to be completed by the end of 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 of accrual-based accounting for the entire general government. (i) 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

183

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 of the 2002 budget. It 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 Strategy (MTFS) in the 2006 budget. The preparation and political endorsement of an MTFS is expected to reinforce fiscal discipline and management, as well as providing the basis for improved strategic allocation of budgetary resources to priorities. The MTP which was issued in the Official Gazette dated May 31, 2005 as a Council of Ministers’ Decision set out a three year (2006-2008) macro-economic framework.119 The MTP outlined policies and priorities in five development areas and nine main sectors.120 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.121 (ii) Reforms to improve capacity at the institutional level 5.11 The PFMC law underlines the need for institutional capacity strengthening by requiring that the fiscal implications of new policies be estimated and policy decisions be made consistent with the MTFS. The 2001 PEIR noted that the institutional processes for policy formulation and management, both in ministries and at the level of the Council of Ministers, were weak. Thus in addition to the strengthening of “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. It is important to ensure that the capacity of ministries and departments to propose policies in line with strategic objectives is strengthened. 119 Macro economic targets for 2006-2008 set out in the MTP were consistent with April 2005 SBA with the IMF. Another implicit underlying factor for determining the macro economic framework for has been convergence with the Maastricht criteria of the EU. The 2005 Pre-accession Economic Program document states that securing price stability, reducing the public sector deficit and debt to GDP ratios must be priorities if the macro-economic policies are to converge with the Maastricht criteria. 120 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 of enterprises, (d) regional development and reduction in regional development disparities and (e) good governance. And the nine main 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. (g) energy, (h) transportation and communication, (ix) tourism. 121 The MTFS 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 PFMC law.

184

5.12 The PFMC law requires ministries to undertake a strategic planning exercise and performance budgeting in order to clarify 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 of the strategic planning initiative on a pilot basis in eight agencies.122 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 on these pilots, a phased program to expand strategic planning to the rest of the government has been announced by the SPO. A three year phased approach was decided to expand implementation of strategic planning to the entire central government and social security institutions through the end of 2009.123 5.13 The Government has initiated an accelerated implementation of strategic planning for local administrations starting in 2006. This initiative was part of the Government’s decentralization reform efforts, which is expected to lead to an efficient allocation of resources at the local level and improved public service delivery. The Metropolitan Municipalities Law (no: 5216, enacted in July 2004), the Municipality Law (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 of these laws. A total of 206 local administrations and 81 Special Provincial Administrations are required to complete their strategic plans by July 2006. 5.14 In 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 piloting activities, provide training, and better respond to the newly defined structure.124 The performance based budgeting department in the MOF and the strategic planning department in the SPO were established in 2004 and 2005, respectively. The experience of the pilots demonstrates both the difficult challenge posed by these agency level reforms and the need for coordination and careful design in the implementation of performance-based budgeting. It is 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 is to draw lessons from the pilot cases to design further stages of the reform, including consideration of the appropriate pre-conditions for such reform.

122 Ministry of Agriculture and Rural Affairs; Turkish Statistical Institute (TURKSTAT); General Directorate of Health for Borders and Coasts; General Directorate of Highways; Hacettepe University; Denizli SPA; Iller Bank; Kayseri Metropolitan Municipality. 123 The time table for the regulatory and supervisory institutions which are part of the central government has not been announced yet. 124 In addition to Pilot agencies, the SPO has adopted a new regulation on strategic planning. The regulation requests public administrations to prepare their strategic plans. Thus, piloting activities have become less necessary now. All the administrative units are requested to prepare their strategic plans within a timeframe set forth in the regulation. However, these administrative units are in need of substantial technical support.

185

Box 5.1: Amendments to the PFMC Law The Government amended the PFMC law in December 22, 2005 with law no. 5436. The main motivation behind this amendment was to clarify some of the terminology used in the law, thus eliminating some of the foreseen implementation problems. 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, II and III. Table below provides a summary of these changes. Original After Amendments

Institutions that have been included under another institution’s budget Court of Jurisdictional Disputes Chart I Ministry of Justice (Chart I) High Council for Elections Chart I Ministry of Justice (Chart I) General Directorate of Mint and Stamp Printing House Chart I Treasury Near and Middle-East Labor Training Center Chart II Ministry of Labor (Chart I) Center of Research on Ataturk Chart II Presidency of Ataturk Culture, language and

History (PACLH) (Chart II) Ataturk Cultural Center Chart II PACLH budget (Chart II) Turkish Language Agency Chart II PACLH budget (Chart II) Turkish History Agency Chart II PACLH budget (Chart II) Turkish Industry Management and Administration Institute Chart II TUBITAK (Chart II) Presidency of Legal Medicine Chart II Ministry of Justice (Chart I) Presidency of Refik Saydam Health Center Chart II Ministry of Health (Chart I)

Institutions Removed from the Coverage of the PFMC Presidency of European Union Educ. and Youth Programs Chart II - GD of the Turkish Radio and Television Agency Chart II - Presidency of Ministry of Defense Carburant Supply and NATO POL Plant Operations

Chart II -

Presidency of the Mass Housing Administration Chart II - Natural Disaster Insurance Agency Chart II - General Directorate of the National Lottery Chart II - General Directorate of Spor-Toto Chart II - Private Employee’s Social Security Fund Chart II - Turkish Accounting Standards Board Chart II - Higher Specialization and Research Hospital Chart II - General Directorate for Rural Services Chart I abolished General Directorate of Building Land Office Chart II abolished Sugar Agency Chart III abolished Savings Deposit Insurance Fund Chart III Eregli Coal Region Workers Savings and Aid Fund Chart IV

Recently Included Presidency of Revenue Administrations Newly established Chart I DG of Social Assistance and Solidarity 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 is 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 is 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. The revolving funds, which were scheduled to be closed by the end of 2007, will now be restructured within the same timetable.

186

(iii) Comprehensiveness of the Budget 5.15 To support the Government’s ongoing efforts towards improved budgetary formulation, a number of steps have been taken to achieve comprehensive budget framework for supporting sound fiscal management. 5.16 Better Budget Definition: Adoption of new budgetary definitions helped Turkey converge with the international practice. A new general government budget definition has been introduced with the implementation of the PFMC law in 2006. The general government definition covers: (i) central government budget institutions (total of 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 of the Budget: Moreover, the coverage of 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 II) and regulatory and supervisory institutions (chart III). The previous coverage of the budget (i.e. the consolidated budget) comprised only 61 percent of the total general government expenditures. With this structure, the Parliament is enabled to review the total expenditures under the central government. 5.18 Improved Budget Classification: Until recently, Turkey was not implementing the functional classification of government expenditure that is recommended by the IMF’s Manual on GFS and which is essential for policy analysis of expenditure. This situation has been addressed by the government’s adoption of functional budget classification in line with the GFS 2001 and by parallel amendments of the chart of accounts for all entities forming part of 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 of the GFS to the entire general government started to be implemented with the 2006 budget. 5.19 Reduction in off-budget activities: There has been a drastic reduction in the number and size of off-budget institutions and activities since 2001 but the agenda remains unfinished. With the exception of 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 still remain outside the scope of the budget discipline (see below). Rationalization and reform of revolving funds, which represent another form of off-budget activity, is under way, but their closure has been delayed further to a 2005 amendment of the PFMC Law (see below). 5.20 Formalization of providing information outside the central government budget: The Parliament has been informed about the expenditures and revenues of institutions

187

outside the central government. The revenues and expenditures of local governments (2003-2008), social security institutions (2003-2006), revolving funds (2006-2006) and extra-budgetary funds (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 list of tax expenditures and their estimated amount for 2006-2008 as an attachment to the budget for the first time. The coverage of this study is 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 of tax expenditures have been addressed in chapter 2. Box 5.2: Towards Full Implementation of PFMC The Public Financial Management and Control Law (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 of the PFMC 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 Law no. 1050, its implementation cannot be carried out with a new structure. Although the coverage of the 2005 budget was not in line with the expanded definition of the central government budget, major improvements had been undertaken in terms of the quality of budget reporting. 1/ With the 2006 budget, coverage was expanded to all central budget institutions, the sum of the general budget, special budget, and regulatory and supervisory institutions. With this expansion, a total of 45 new institutions were included in the budget coverage. However, although the PFMC 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 to report them under the relevant institution’s budget until their abolishment by the end of 2007. However, with an amendment to the PFMC enacted in December 2005, the government decided not to abolish them but “restructure” them by 2007. Until the restructuring is completed, the rules and principles for their budget preparation, execution, accounting and auditing will be determined by a secondary legislation issued by the Ministry of Finance. As a transitional measure, the revenues and expenditures of the revolving funds have been submitted to 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 of the social security institutions and local administration for the 2006-2008 period. This information was provided as an annex to the budget proposal. None of these institutions’ expenditures has been discussed in the Parliament. In the case of the social security institutions, the data presented to the Parliament were the estimates of the institutions themselves. As for the local governments, the data produced through consolidation of the individual local administrations’ budgets was not used since the GFS budget coding structure was not in effect for the all of the local administrations. Therefore, the projection of the State Planning Organization was used. Financial services units for all of the general government budget institutions were established in January of 2006, and with the establishment of these units the authority of the MoF for ex-ante spending control was delegated to the line agencies. The accounting responsibility and main budget responsibilities other than the general budget institutions were transferred to the special budget institutions. 2/

188

The internal audit structure foreseen by the PFMC will be completed before end of 2007. The Ministry of Finance is planning to set up the internal audit structure of the 15 pilot institutions together with their personnel during 2006. It 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 of 2004. The PFMC authorized the SPO and the MoF to determine the expansion timetable for the institutional strategic planning and performance based budgeting. As expected, the 2006 budgets of the institutions were not based on institutional plans and performance criteria. According to the Municipality law, strategic planning is necessary for local administrations with populations of more than 50,000 in 2006 Implementation of the external audit part of the PFMC depends entirely on the enactment of the new Turkish Court of Account (TCA) law which was submitted to the Parliament in February of 2005. The new draft expands the audit mandate of the TCA to all public sectors and provides a legal basis for expanding the scope of the TCA to include the performance and financial audits along with the compliance audit. 1/ Details of the new features of 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 office and the personnel of 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 of implementation, the achievements in linking the policies/priorities with the budget have been quite satisfactory. However, further improvements can be achieved through;

• Ensuring that the MTP’s sectoral classification is appropriately translated into the budget functional classification: The sectoral breakdown provided in the MTP is not consistent with the functional classification of the budget structure. Due to inconsistent classifications of the MTP and the budget, it is difficult to follow the link between the policies envisaged in the MTP and the related budget allocations. Mapping specific sectoral categories of the MTP priorities into the functional categories of the budget would allow the government to easily translate MTP decisions into budget allocations.

• 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 of measures together with their cost estimates and implementation timetables could provide better guidance for the line agencies.

• Making program budgeting an integral part of the PFMC system. In addition to mainstreaming MTP, what is equally important is adopting program budgeting approach to annual budgeting. New budget coding will not be enough

189

for program budgeting unless the latter is integrated in the PFMC system. In the absence of program budgeting, strategic planning and performance evaluation will be impeded.

• Rationalizing the number of policy formulation documents: Currently, there are too many documents that can be treated as the basis of policy formulation.125 Although these documents have different stated purposes and perspectives, it would be worth considering options for merging/consolidating some of these documents for a simplified policy guidance structure. This would help eliminate potential sources of confusion for the line agencies; reduce the workload of the key central agencies; and diminish the risk of inconsistencies among these documents.

The Government is aware of these challenges and efforts to improve institutional capacity for medium-term budget planning are under way. 5.23 The implementation of strategic planning and performance budgeting will require more efforts of the key central agencies. Several issues need to be addressed for a sound implementation of the strategic planning and performance budgeting initiatives. Timing and sequencing is a critical issue. Technical and human resource capacity will be another major challenge for the implementation of 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 of distorting budget discipline. Many spending units maintain revolving funds financed by their own sources of revenues—more than 1,000 such funds as of December 31, 2005. The original PFMC law required the closure of the revolving funds before the end of 2007. However, as mentioned before, the government decided to restructure the revolving funds attached to the general and special budget institutions by the end of 2007 with an amendment made to the PFMC in December of 2005. More than six percent of the central government expenditures will thus still not be covered by the overall financial management and control structure defined in the PFMC. The MOF started to issue the revenues and expenditures and balance sheets of the revolving funds on a quarterly basis as of June 2005 and provided information to the Parliament on the revolving funds budget as an annex of the 2006 budget proposal. 5.25 The share of revolving funds is quite important in health and education sectors. As can be seen in Table 5.2, the off-budget revolving fund expenditures of the Ministry

125 Policy formulation in 2006 was based on the following list 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 list starting next year.

190

of Health (MOH) reflect almost 90 percent of the total appropriation allocated to the MOH budget. Therefore, the MOH is 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 2005 e 2006 P Higher Education Institutions and universities 1,783 2,237 2,360 2,532 Other Central Budget Institutions 3,429 5,732 7,924 8,615 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 (% of Central Government Budget) 3.72 5.25 6.10 6.41 Ministry of Health RF expenditures (% of MOH budget) 86.5 1/ 2003-2005 data is for the consolidated budget. For the purpose of 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 of Finance and World Bank Staff calculation

5.26 Extrabudgetary funds still operate outside of budget discipline. The PFMC Law intends to further strengthen the control framework for the EBFs by the following measures: (i) incorporation of extra budgetary funding into the budgets of the related administration submitted for parliamentary approval; (ii) requiring external audit by the Turkish Court of Accounts; and, (iii) monthly reporting by EBF in the consolidated report of the central government. However, implementation of these measures has been delayed. Even though the PFMC Law required incorporation of extra budgetary funds into the budget of the related administration, the 2006 budget fails to do so. Instead, revenues and expenditures (without details) of 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. As it is a legal requirement, EBFs should be subject to the internal control and internal audit regime that is defined in the PFMC Law. Total expenditures channeled through the EBFs as a share of the central government budget will be around 2.24 percent for the 2003-2006 period (Table 5.3).

191

Table 5.3: Extra Budgetary Funds, 2003-2006

(Million YTL) 2003 2004 2005 e 2006 P Privatization Fund ( Privatization Administration, Chart II) 1,120 1,126 1,897 811 Defense Industry Support Fund (Undersecretariat of Defense Industry, Chart II) 1,251 1,233 809 1,053 Promotion Fund (Prime Ministry, Chart I) 41 111 121 128 Social Solidarity Fund (GD of Social Solidarity, 658 1,284 1,262 1,258 Total 3,070 3,754 4,089 3,250 EBFs’ expenditures/central government budget (%) 2.2 2.5 2.4 1.9

Source: SPO and MoF

B. BUDGET EXECUTION, ACCOUNTING, AND REPORTING

(i) Developments and Reforms Since 2001 5.27 The PFMC Law has balanced the accountabilities and authorities of spending agencies. The PFMC Law has revoked the authority of the Court of Accounts and Ministry of 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 of the Ministry of Finance. For example, the spending agency now has the authority to authorize payments, and the accounting offices of GDPA no longer perform regularity audit and detailed verification of payments. Previously budget officers who worked in spending agencies were employees of Ministry of Finance; now they will be part of the financial service units within each spending agency. 5.28 The law clearly defines the role of the Treasury as provider of cash and the Ministry of Finance as the payer of budget expenses. The 1994 Law on the Structures and Duties of 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 of making payments to the GDPA of the Ministry of Finance (Article 61). 5.29 There is predictability in terms of allocation of budget funds to line ministries. The division of responsibility is clear between Treasury and the Ministry of 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 of cash has thus been avoided. 5.30 An automated system has been implemented to facilitate budget management. The General Directorate of Budget and Fiscal Control (GDBFC) has developed an automated Budget Management Information System (BYES – also called eBudget) to keep track of budget appropriations, allocations, and utilization. The interface between

192

Say2000i and BYES ensures the transfer of data on budget allocations to the Say2000i system as well as actual expenditures from Say2000i back to BYES. 5.31 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 of Finance. It is structured functionally, with local tax offices directly under its control. The Ministry of 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 of assuming increasing responsibility for collection of social security contributions. 5.32 Implementation of 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. It includes both the assessment and the collection of taxes. Automation has had several benefits: quicker collection of tax returns and declaration forms; better utilization of 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 of a modern accounting framework, including an accrual-based accounting system and consolidated reporting requirements for the general government; and it established an official body for setting government accounting standards. 5.34 Turkey has a single accounting system, which is maintained by the General Directorate of 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 of function thus enables strict control of 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 of accounting. Since the annual budgets are cash-based, the GDPA in the Ministry of Finance has devised an intricate method for accounting entries that keeps track of 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 of a cash-based budget by an accrual-based budget would demand enormous resources from both

193

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 is 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. Say2000i can capture and set aside spending agency commitments against the budget appropriation and budget allocation, thus preventing budget commitments in excess of appropriations. 5.38 A uniform chart of accounts that is harmonized with budget classification has been implemented. The GDPA has issued a new framework for the accrual-based chart of accounts. It is harmonized with the economic classification of the newly adopted GFS budget classification system. A new chart of accounts enabled the GDPA to compile financial statements consistent with the 2001-GFS budget classification.126 5.39 Uniform accounting standards have been introduced and a single authority to set standards has been established. Article 49 of 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 is under the Ministry of Finance, comprised of representatives from the TCA, State Planning Organization (SPO), Ministry of 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 of Finance developed in-house an automated online accounting system, Say2000i.127 Up and running since 2002, the system networks and captures receipts and payments from more than 1,500 national nodes as they are made.128 Because the database is linked through a central server, all transactions are immediately available. The system can produce periodic financial statements without the typical delays of decentralized accounting systems. The system covers all general budgetary institutions except the Office of the President and accounting office for the State Debt within the Treasury. It includes the majority of special budget institutions (Chart II) and the regulatory supervisory agencies (Chart III).129 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 of Statistics is responsible for compiling GFS-compliant financial statistics. 127 The total hardware and software investment was about US$8 million. See the General Directorate of Public Accounting (GDPA) web site, http://www.muhasebat.gov.tr/say2000/indexE.asp for a presentation of the technological architecture of the system. 128 Serving more than 6,000 users, the system is comprised of 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 Say2000i because of technical difficulties : The Turkish Standards Institute, Turkish Patent Institute, Small and Medium Sized Industry Development Organization (KOSGEB), Telecommunication Agency, Banking Regulatory and Supervision Agency (BRSA), Tobacco Agency, and Punishment and Prison Institutions.

194

(ii) The Unfinished Agenda and Remaining Challenges 5.41 The scope of tax audit activities needs to be improved. Over the medium term, the government has committed itself to increasing the number of auditors from 5 percent of the Revenue Administration staff to the international standard of 20 percent or more. 5.42 Implementing a Say2000i-compatible system in the social security institutions and local administrations. The government is in the process of implementing Say2000i-compatible accounting systems in the social security institutions and local administrations. However, the large number of local administrations (more than 3,200) poses an implementation challenge in the short term. 5.43 Properly phasing-in and completing the implementation of 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 of new assets. However, valuation of 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 of timeframes, often in stages. For example, the United Kingdom produced its first accrual accounts for individual departments in 1999–2000. It consolidated central government accounts using generally accepted accounting principles in 2003–04, and it will prepare whole-of-government accounts for 2005–06. 5.44 The implementation of full accrual accounting requires taking on complex issues. Key issues include valuation of assets and inventories, depreciation rates, and gains and losses from foreign exchange. As shown by the experience of countries that have switched to accrual accounting, implementation generally works best when done in phases.130 Even where countries have successfully produced accrual-based financial statements, auditors have issued disclaimers of opinion on the accrual-based financial statements for a number of years. For example, the U.S. Government Accountability Office (GAO) has issued disclaimers on the consolidated financial statements of 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 of Finance has made significant efforts in training a large body of public accountants across the country. Accrual-based financial reporting is not an end in itself but provides information useful 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 is therefore important that users,

130 For more information refer to IFAC publication “Transition to the Accrual Basis of Accounting” dated December 2003 ( http://www.ifac.org/Store/Details.tmpl?SID=102026702640546 )

195

including heads of departments and program managers, be extensively trained in how to use and interpret accrual accounting statements. 5.46 Implementing Say2000i 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 is fully operational. The interface to BYES and the treasury system is currently operational as well. Once all interfaces are fully operational, the need for manual intervention will be eliminated; and automatic transfer and the integrity of 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 of 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 of 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 Say2000i system for line ministries. Line ministries do not maintain separate accounting records although the PFMC Law provides them with an enhanced role in financial management; so it is particularly important that line ministries have unhindered read-only access to the Say2000i system. This will enable them to monitor the financial performance of 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 Say2000i system.131 However, until recently only commitments for large multiyear capital expenditure projects were fully captured. The revised accounting regulation issued in June 2005 requires recording of all commitments to ensure budget discipline and avoid accumulation of arrears. 5.51 Establishing and implementing formal procedures for system controls audit. An online system with a database of the magnitude of Say2000i system requires formal

131 The terms commitments, expenditures, and payments are often confused. Simply stated, commitments are the value of the purchase orders and contracts that bin an agency to buy certain goods or services to an amount specified. Expenditures are the actual amounts of goods and services for which full payment may or may not have been made. Payment is the amount transferred from the treasury bank account to the credit of the supplier’s account.

196

policies and procedures to audit system controls on a regular basis, including mechanisms for monitoring the implementation of audit recommendations.

C. INTERNAL CONTROL AND INTERNAL AUDIT

(i) 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 of the Internal Financial Control Framework for the General Budget Institutions as Defined by the 2003 PFMC Law

Payment processing controls:

* Strategy Development Directorate. Financial services unit responsible for ex-ante controls. * Authorizing Officer. Approves and is held accountable for payments. * Accounting Officer. Verifies, makes payment, and maintains accounts.

5.53 There are positive aspects of the internal control system. The most important positive feature is the separation of 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 132 The original provision in the PFMC Law required establishment of a new position of financial control officer with full 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 of the financial services unit.

Head of Public Administration

Authorizing Officer

(Spending Unit)

Internal Auditor

GDPA Centralized Accounting

Accounting Officer

(decentralized)

Financial Services Expert (Strategy Development

Directorate)

197

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 itself to verifying the completeness of documentation and to ensuring that no material errors exist. Another positive feature of the internal controls is that Say2000i system does not allow deletion of transactions, so any corrections must be entered as new transactions. 5.54 The PFMC Law has introduced a modern internal audit framework. With devolution of authority and responsibility to spending agencies, the need has increased for a modern internal audit organization that can assure the head of the spending agency on the soundness of the internal control system. The PFMC Law requires each public administration to establish an internal audit unit within its administration. The law envisions a decentralized internal audit model in which internal auditors report to the head of the administration. This is similar to the current model of 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 of 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 of ministers and other high-level civil servants. In the old accountability model, accountants and accrual officers were responsible to TCA for financial misdemeanors. This has been replaced with broad accountability for government officials. Articles 10 and 11 of the PFMC Law deal exclusively with ministers and heads of public administrations. The new provisions contrast sharply with the previous law, which carefully defined the accountability of the accountant but generally glossed over the accountability of the higher-level officials. 5.56 The “visa” power of the Turkish Court of Accounts and the Ministry of Finance has been revoked. Under the previous system, the Turkish Court of Accounts and the Ministry of 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 of controls normally exercised by an executive authority. TCA thus played an indirect role in the budget execution process. This created apparent conflict of interests, because TCA also audited budget execution reports. The PFMC Law 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 of accountability reports. The PFMC Law (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 of activities in implementing the strategic plan and budget. These reports would be audited by the TCA and submitted to Parliament. The introduction of accountability reports breaks with a long tradition of providing Parliament with financial reports only. It thus strengthens the

198

accountability of 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, it is too early to evaluate its actual effectiveness in Turkey. 5.58 The Ministry of 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 of financial controls; the Internal Audit Department, which develops secondary legislation in the area of internal audit and also serves as a secretariat for the IACB; and the Financial Management Department, which is responsible for work in the area of performance budgeting. 5.59 The accrual accounting system would enable tracking of both movable and immovable assets. The Say2000i accounting system has captured purchase and construction of 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. (ii) Unfinished Agenda and Remaining Challenges 5.60 Implementation of 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 it devolves financial management authority and responsibility to line agencies. However, with the continuing uncertainty regarding the enactment of the Public Administration Framework Law—which envisages abolition of inspection boards in line agencies—the mapping of inspectors to internal audit positions has become less certain. 5.61 Phased implementation of 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 of 81 Special Provincial Administrations (SPAs) and more than 3,200 municipalities. Some of 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 of the size or capacity of 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 of the PFMC Law. The implementation of the PFMC Law had been orchestrated by the Ministry of Finance. Spending agencies had generally been aloof distant observers. If PFMC implementation is 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

199

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 of inspection boards and internal audit units. The PAF Law envisioned abolition of 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 of inspection boards and internal audit units. Previously, the inspection boards focused on identifying violators of laws and regulations. In general, they paid little attention to internal control systems. The creation of internal audit units to focus on internal control systems would be a step forward. Nevertheless, every environment needs investigative capacity to uncover abuse of power and misuse of public resources. It will therefore be important to retain inspection and investigation functions, albeit with fewer personnel. Also, the scenario needs to be considered of inspection boards being abolished before internal audit units become fully operational. This could weaken the control framework and increase the risk of misuse of 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 Say2000i system. Payroll for the approximately 1.8 million general government staff is not yet centralized, exposing payroll to the inherent risks of misuse and malpractice. A personnel module is ready in the Say2000i 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

(i) 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 of 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 of the parliament. DDK has a broad mandate, covering all public sector bodies except the military and the courts. Because it undertakes research studies, investigations, and audits only at the request of the president, it therefore does not function as a regular external audit agency. 5.66 TCA is independent of the legislature and the executive. TCA is a constitutional establishment and is entrusted with the task of auditing on behalf of the Turkish Grand National Assembly. TCA carries out its work on behalf of the parliament but remains

200

independent of both the legislative and executive branches of government. The parliament elects its president by secret ballot from two candidates, who are put forth by an ad hoc committee of 15 members drawn from the parliament’s Plan and Budget Commission. The president of the TCA is elected for a seven-year term and is eligible for reelection. TCA 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 TCA enjoys financial independence by preparing its own budget, which it submits to parliament without intervention. The president of TCA is 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 it does not receive any budgetary support. The entire budget of YDK is funded out of audit fees collected from SOEs. 5.68 As an institution, TCA also exercises judicial power. The TCA is 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 of appeals—perform the judicial function. 5.69 TCA and YDK report to the parliament. TCA reports on 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, TCA 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 of financial statistics in terms of accuracy, reliability, and conformity with set standards. In addition to these mandatory reports, TCA 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. (ii) 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 of the TCA 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:

• The PFMC Law extended TCA’s audit mandate to cover all administrations forming part of 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.

201

• Audit of TCA by an independent commission comprising professional audit staff is included in the PFMC Law (Article 69).

• The draft law defines audit as a planned activity to be carried out in accordance with international standards.

• The draft law specifically provides for financial audit by TCA, thereby putting financial audit to the forefront along with the regularity audit.

• 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 of TCA regarding visa approval.

• The draft law contains an article explicitly requiring TCA to make its reports public within 15 days of their submission to Parliament.

• The draft law recognizes the need to carry out interim audits at any time during the financial year.

• The draft law permits the audit of accounts and transactions of public administrations by independent private auditors. However, such audits must be carried out under supervision of TCA auditors in compliance with its audit principles.

(iii) Unfinished Agenda and Remaining Challenges 5.72 TCA’s draft law creates a sound basis for the development of government audit in Turkey that is in line with modern government audit concepts and accepted international auditing standards. It is important that the draft law be enacted quickly. This will enable TCA to concretely begin the implementation of reforms. The following additional measures would facilitate the transformation of TCA into a modern supreme audit institution. 5.73 Strengthening the audit function within TCA. The audit side of TCA is not yet fully developed. The draft law is still insufficient in terms of structures and responsibilities. Even after enactment of the law, TCA will still be dominated by its judicial side. Its audit reports and the audit planning and coordination activities will largely remain in the hands of those primarily employed on non-audit issues. TCA will still lack an authoritative internal voice representing audit interests. And those aiming for senior positions will still find only one way forward: leaving audit to join the chambers as members of the court. 5.74 Developing capacity for financial statement audits. In the future, TCA intends to carry out financial audits. This type of audit includes technical opinion on the reliability and accuracy of 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 of 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 of accounting and would be producing accrual-based financial statements in addition to cash-based budget execution reports, this development requires capacity building within TCA on the skills to audit accrual-based financial statements.

202

5.75 Adopting a strategy for risk-based audit. Until 2006, audits carried out by the TCA were based on the 100 percent transaction examination instead of risk assessment. T TCA has been maintaining the fiction of 100 percent audit based on the related articles of abolished Public Accounting Law no: 1050. With the effectiveness of the PFMC, the Parliament plans to adopt a new TCA Law which will make the external audit consistent with the application of internationally recognized audit standards. 5.76 Designing the role and structure of the YDK: The Public Administration Framework Law enacted by the parliament but vetoed by the president envisioned merger of YDK with TCA. This welcome proposal was consolidating the mandate of the country’s Supreme Audit Institution and reduces the number of 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 TCA draft legislation. Moreover, issues posed by the merger must be carefully considered, including questions of personnel, grades, and levels of staff. 5.77 Devising a strategy for audit of local governments. Local governments are part of TCA’s audit mandate. There are more than 3,200 local administrations in Turkey, about 15 percent of which are audited by TCA. Since those selected are generally the largest, TCA estimates that about 80 percent of local government expenditures are thus subject to audit. However, this still implies a widespread lack of audit in this sector. Moreover, with the increased role for local administration envisioned under the decentralization agenda, the complexity and size of activities carried out by local administrations will doubtlessly increase significantly. This prompts the need for reconsideration of audit policies for this sector. Depending on the choice of 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 of 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 framework of principles. Secondary legislation, manuals, guidelines, and extensive training will be needed to achieve the aims. In particular, TCA must change from an institution dominated by its judicial functions to one that is adapted to carrying out modern audits and reporting its audit findings to the parliament. 5.79 Adopting international audit standards. In the future, TCA intends to carry out audits in line with accepted international audit standards. Article 35 of the draft law defines audit as a planned activity carried out in accordance with these standards. It is important that TCA adopts (without any changes) the audit standards of either INTOSAI or the International Federation of Accountants (IFAC). In addition, TCA should develop and issue ethical guideline for auditors.

203

E. LEGISLATIVE OVERSIGHT

5.80 The Plan and Budget (P&B) Commission of the parliament oversees the budgetary process. The commission is chaired by a member of the ruling party and is comprised of 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 of the parliament, the commission, and the government. Considerable discussion occurs annually while the budget is 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 on 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 of numeric data but scant detail or discussion on program implementation, outcomes, or outputs. The reports are not conducive to discussion. The PFMC Law of 2003 envisioned annual accountability reports with details on costs and program achievements, thereby enabling meaningful 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 TCA audit report is primarily quantitative and contains little on financial management or accountability issues that emerged during the audit. To date, as described, the TCA has primarily discharged judicial functions and carried out compliance checks rather than financial audits. The new draft TCA law would change this by requiring TCA to carry out financial audits with opinions on findings, thereby strengthening parliamentary oversight over the executive branch. 5.84 Effective implementation of 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-term fiscal strategy, and annual accountability reports to the parliament. Furthermore, the draft law requires the TCA to render opinions on financial accounts, including any issues that emerged during the course of the audit process.

204

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 of public sector contracts. The 2001 PEIR and the Country Procurement Assessment Report (June 2001) 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 of 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 of goods, works, technical services, and consultant services. Here too all documents were completed before the Public Procurement Law 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 of standard documents was foreseen within the scope of grant-financed activities, the PPA published all documents on its website, an activity fully financed by its own resources rather than through grant funds. 5.89 Trainers were trained on the new Public Procurement Law. PPA’s own staff provided advance training on the PPL upon request of procuring agencies, which also was financed from its own resources. As of May 2005, 9,634 staff from 278 procuring entities were trained by PPA trainers. In addition to three days of 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

205

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 of interactive learning CDs were reproduced and distributed to public and private entities as well as subscribers of 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 of 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. Two consultants were hired in January 2004. They submitted a draft law to the PPA in May 2004. The draft law was submitted by the PPA to the government, circulated amongst stakeholders for comments and inputs, and returned to the Ministry of Finance for further action. Currently the draft law is considered within the “Screening Process of Procurement with EU” - whether to incorporate it within the PPL or issue it as a single bill. (ii) Unfinished Agenda and Remaining Challenges 5.91 Use of country procurement systems. To some extent, all projects in Turkey rely on the existing country procurement systems with national competitive bidding (NCB). Thresholds are being raised, enabling use of NCB or other procurement methods for a larger portion of project procurement. Subject to certain conditions, exceptions, and waivers, the World Bank permits the use of country public procurement system for NCB 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 of certain services from the scope of the law. Later, a decision was made to hold all amendments on the PPL until after the Procurement Screening Process with EU is complete. It is recommended that the government should ensure internal consistency and integrity of 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, it is 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 no e-Government Procurement Strategy. However, the PPA has initiated a process on its own 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 CD format. Most public procuring agencies maintain their own websites. Notices are published on the web in addition to obligatory issuance in the Public Procurement Bulletin, which is available in hard copy and electronic format. There is strategic integration within the context of e-government

206

planning. PPA has agreed to establish a specific portal to serve public and private agencies such as the Social Security Agency and the Chamber of Commerce and Industry. (Such portals will simplify the collection of authenticating documents and certificates required in the bidding documents, for example, document authenticating membership in the chamber of commerce.) Within the context of the EU accession process some funds are allocated to PPA for electronic infrastructure and staff development. 5.94 Implementation of baseline indicators to assess the public procurement system. As part of the ongoing dialogue supported by an IDF grant for institutional strengthening of Public Procurement Authority, it was agreed to conduct a joint assessment of public procurement system through the Baseline Indicators System (BIS). The BIS provides a structured approach to analyze the health of the existing system. A rating system was finalized in February 2006, with the participation of Public Procurement Agency (see Annex VIII for details). The report provides a comprehensive picture of 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 BIS system and as necessitated by expected changes in laws and regulations. However major issues and concerns are summarized below: Indicator Summary of Self

assessment by PPA

World Bank’s comments/concerns

Indicator No. 1- Public procurement legislative and regulatory framework achieves the agreed standards and complies with applicable obligations

All the 26 sub-indicators rated as “fully achieved”

(i)New utility law to be enacted;(ii)PPA to maitain its oversight on all amendments and new law which affects public procurement;(iii) Bank remain concerned about exemptions provided to certain types of entities or services through other laws;(iv) open competitive bidding to be a preferred method in the law ;(v)all procurement above or 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 of Implementing Regulations and Documentation

“fully achieved” on all the 6 sub-indicators

(i) PPA to ensure consistency within all implementing regulation; (ii) implementing regulation to be updated regularly; and (iii) User’s Guide manual for contracting entities to be updated regularly

Indicator No.3 Mainstreaming Procedures into Public Financial Management

Out of 6 sub-indicators PPA has rated 1 as fully achieved, 4 as substantially achieved and 1 as not achieved

Further studies required to provide data on budget formulation.

Indicator No.4 Functional Management /Normative Body

Out of 4 sub-indicators, 3 “fully achieved’ and I substantially achieved

Administrative capacity of PPA to be strenghthened and responsibilities within the organization clearly defined

207

Indicator No.5 Existence of Institutional Development Capacity

Out of 4 sub-indicators 2 “fully achieved”, 1 “substantially achieved” and 1 “not achieved”

(i)PPA to develop a more proactive strategy for capacity development through twinning arrangement( such as mandatory training,minimum qualification standards to professionalize procurement staff, on-line cerification of staff, awareness compaign etc); and (ii) collection and monitoring of national procurement statistics to be enforced in line with Article 53(b) 9

Indicator No 6: Efficient Procurement Operations Capacity and Practices

Out of 6 sub-indicators 1 is fully achieved and 5 are substantially achieved

(i) Need to improve procurement performance through availabilty, analysis and application of data

Indicator No 7 Functionality of Public Procurement Market

Out of three applicable sub-indicators, 2 “fully achieved” and 1 “substantially achieved”

There are certain systemic constraint in access to credit market. This needs to be addressed.

Indicator No 8 Existence of Contract administration and Dispute Resolution Provisions

Out of 2 sub-indicators 1 “fully achieved” and 1 “substantially achieved”

There are implementation issues. More study is required to determine coverage of existing procedure to facilitate efficient dispute resolution process

Indicator No 9 Effective Control and Audit System

Out of 5 sub-indicators, 2 are “fully achieved”, 1 “substantially achieved” and 1 “not achieved”

Implemenatation of PFMC law to be monitored

Indicator No 10 Efficiency of Appeal Mechanism

All 6 sub-indicators “fully achieved”

Complaints are not resolved at the initial stage( at the contractingentities) and generally referred to PPA, which delays the contract award for one month and increases the workload of PPA. PPA should develop and provide guidance to the contracting entities to address the complaints from participants

Indicator No 11 Degree of Access to Information

All 4 sub-indicators “fully achieved”

No Issue identified by the BIS team.

Indicator No 12 Ethics and Anti-corruption measures

Out of 7 sub-indicators all “fully Achieved”

Implementation issues remain in this area. Signature of ethical contract is required for all public employees. Disclosure of financial assests by the public staff is required every 5 years. But this issue requires more study.

5.95 Status of Dialog between EC and Turkey on Procurement. As part of Turkey's accession to the EU, process screening meetings between EU and Turkey are being held. Public Procurement is one these topics and discussions of this topic took place in November 2005. Draft screening report prepared by EC notes improvements required in the following areas (i) public contracts(15% national preference above the threshold is incompatible with the acquis); (ii)Lack of sector (utilities) legislation; (iii)Concessions - lack of horizontal legal framework; (iv)Strengthening the Administrative capacity of the PPA; and (v) .Electronic Procurement - such as need for electronic auction on 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.

208

Table 5.4. Action Plan as Provided in CPAR 2001

Actions Responsibility Achievement

Establish an interministerial committee to draft a new public 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.

GOT Done

Apply public procurement law to all agencies at central and local levels, using budgetary or extrabudgetary public funds, and to SEEs.

GOT Partially done

Draft and enact implementing regulations. GOT Done

Prepare and enact a medium-term plan for progressive reform of the public procurement legislation to achieve full alignment with EU Procurement Directives.

GOT Pending EU procurement screening process

Undertake a review of public investment programming, budget planning, and procurement planning. (Coordinate with the PEIR.)

GOT MOF

Partially done

Introduce a prequalification procedure for large or complex contracts. (Include pre-qualification provisions in new public procurement law.)

GOT Done

Reform the Contractor Certificate system. GOT, MPWS DONE

Improve market access by graduated reform and relaxation of legal and administrative obstacles to participation by foreign bidders.

GOT Partially done

Revise the range of procurement methods: reinforce open bidding as 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.)

GOT Done

Increase transparency and fairness: lengthen bidding periods, reform 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.)

GOT Done

Introduce standard prequalification and bidding documents for civil works, goods, and services.

GOT Done

Improve the bid protest mechanism; provide a legal basis and institutional arrangements for administrative review. (Link to the establishment of a central public procurement office.)

GOT Done

Establish a new central public procurement office. (Establish and define its function in new public procurement law.)

GOT Done

Draw up a national training strategy for public procurement. GOT Not yet

Develop a national training program for public procurement and deliver it through identified training institutions.

GOT Training institutions

Not yet

Establish procurement as a profession within the civil service. GOT Not yet

Establish a professional accreditation scheme for government procurement officers.

GOT Not yet

Develop a more constructive relationship with the business sector through outreach programs

GOT Partially done

209

Increase resources and investigative powers of TCA. TGNA Draft bill submitted to TGNA

Train TCA auditors in the application of the public procurement law GOT TCA

Publish results of procurement audits. TCA Not yet

Strengthen internal audit functions of procuring entities to audit procurement.

GOT Procuring entities

Partially done

Strengthen provisions in the criminal law on procurement-related corruption and enforce them more rigorously. (Include reference to criminal law provisions in new public procurement law.)

GOT Done

Insert specific prohibitions against fraudulent and corrupt activities by bidders in bidding and contract documents. (Coordinate with development of standard bidding documents.)

GOT Procuring entities

Done

Put in place facilities for reporting allegations of fraud and corruption.

GOT Done

Introduce clear conflict of interest rules. (Include in new public procurement law and standard bidding documents.)

GOT Done

Source: World Bank Staff

G. COMMON REFORM DIRECTIONS

5.97 Implementation of 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 civil society and enhancing accountability to citizens. Yet to succeed, the reform agenda must be fully planned and effectively executed. Although some aspects of the new legal framework are being piloted intermittently, most of the secondary regulations (that is, those related to practical implementation) are yet to be issued. None of the reforms have yet been fully implemented. The government’s decision to undertake fundamental changes in the responsibilities of 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 of major public reforms needs strong coordination and monitoring. By the nature of 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 of the objectives. Pragmatic midstream adjustments must adapt to changing circumstances, and coordination must be effective across units of 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 of provisions being enacted that are inconsistent with or contradict the previous reforms. Enactment of the Omnibus Law has created concern that its provisions run counter to the intention of the PFMD and PFMC Law as approved by the Parliament—a measure to improve fiscal

210

transparency and to ensure that policy decisions are subject to consideration in the context of a medium-term fiscal strategy that restrains expenditure commitments. It is crucial that Turkey maintains the credibility of 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 of the reform agenda will remain at high-risk.

211

ANNEXES

212

ANNEX I

Data and Methodology for Consolidated General Government Classification of 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 of 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 of the general government institutions, however, since the last year covered in Chapter I is 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 of the general government for Turkey. In generating consolidated general government data the SPO is following the GFS manual. The IMF is another source for the general government data in Turkey. However, the purpose of their data is not to measure the overall size of the revenues and expenditures of the general government, but to monitor the performance criteria set in terms of the primary surplus of the public sector.

The definition of general government used in this report includes: consolidated budget (CB), local administrations, revolving funds, social security institutions (SSI), Revolving Funds and a common subset of budgetary and extra budgetary funds (EBFs)

Consolidated Budget Institutions: All of the general and annex budget institutions. As of 2005 the total number of the institutions was 98.

Local Administrations: 3225 municipalities, 81 special provincial administrations, Iller Bank, 16 water and sewerage companies of 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 of Dormitory and Student Credits, National Lottery and AOC (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 of the funds included in the general government balance decreased from 12 in 1999 to 4 in 2005.

213

It should also be noted that the analysis does not include the expenditures of 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 Law (PFMC).133 The analysis includes the net subsidies and transfers between these institutions and the consolidated budget.

Although the SPO is in general in line with the GFS methodology three adjustments were made to their general government data for the following reasons;

• To eliminate the double counting:

Invoiced payments which are social assistance type of 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 of in the SSK balance because the same amount of spending is 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.

• Moreover, in order to be in line with the primary surplus definition of the GFS 2001 manual, interest revenues of the institutions were deducted. In their current definition of 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 of general government expenditures -- functional vs economic -- for 2003 and 2004 by combining the expenditures of the institutional coverage defined in the economic classification was used. Regulatory and Supervisory Institutions and some of the special budget institutions which were recently brought under the definition of general government by law 5018 were not included because of the data problem. 8. Main data sources for the study are as follows.

• SPO economic classification of the general government data • Ministry of 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 to provide certain public services. These special budget institutions receive revenues and are authorized to spend them. The complete list of these institutions is presented in the Public Financial Management and Control Law.

214

• SSK, Bağkur, Emekli Sandigi • High Audit Board (YDK) on the accounts of SSK, Bağkur and TRT

(Turkish Radio and Television Board ) , AOÇ ( Atatürk’s Farm) , General Directorate of Credit and Dormitories.

• National Lottery Directorate • IMF fiscal tables of June 2006 • Ministry of Internal Affairs, General Directorate of Local Authorities

9. As a general principle the GFS consolidation methodology has been used while producing the functional classification of the general government. A detailed analysis on the methodology and assumptions for producing the cross classification is available in Ferhat Emil, 2005, Turkey-Economic and Functional Classification of General Government Expenditures, background analysis note for the CEM: promoting Sustained Growth and Convergence with the European Union. Table A1 below shows the results of this study.

215

Table A1: Classification of General Government Expenditures across Economic and Functional Categories, 2003-2004, in percent of GDP

(% of GDP) Personnel

Expenditures

SSI Premium Payments

Goods and Service

Purchase

Interest Payments

Current Transfers

Capital Payments

Capital Transfers

Lending

Contingency

Total

2003 2004 2003 2004 2003 2004 2003 2004 2003 2004 2003 2004 2003 2004 2003 2004 2003 2004 2003 2004

General Public Services 0.9

0.9

0.1

0.1

0.5

0.5

16.6

13.3

0.6

0.6

0.4

0.5

0.2 0.2 0.4 0.2 0.0 0.0

19.8

16.4

Defense 0.8

0.8

0.1

0.1

1.4

1.2

0.0

0.0

0.1

0.1

0.0

0.0

0.0 - - - - -

2.5

2.2

Public Order and Safety 1.2

1.1

0.2

0.2

0.6

0.6

- -

0.0

0.0

0.2

0.2 - - - - - -

2.1

2.0

Economic Affairs 1.1

1.1

0.2

0.2

0.5

0.5

0.2

0.1

1.1

1.0

1.5

1.4

0.1 0.1 0.3 0.3 - -

4.9

4.7

Environmental Protection 0.1

0.1

0.0

0.0

0.0

0.1

- -

0.0

0.0

0.1

0.1

0.0 0.0 - - - -

0.2

0.2

Housing and Community Amenities 0.3

0.2

0.0

0.0

0.2

0.1

- -

0.0

0.0

0.6

0.4

0.0 0.0 0.0 0.0 - -

1.1

0.7

Health 1.3

1.5

0.1

0.1

3.4

3.8

0.0

0.0

0.2

0.3

0.2

0.3 - - - - - -

5.4

5.9

Recreation, Culture and Religion 0.2

0.3

0.0

0.0

0.1

0.1

- -

0.1

0.1

0.1

0.1

0.0 0.0 - - - -

0.5

0.6

Education 2.6

2.5

0.3

0.3

0.6

0.6

- -

0.5

0.5

0.7

0.5

0.0 0.0 - - - -

4.6

4.5

Social Protection 0.1

0.1

0.0

0.0

0.1

0.1

0.0

0.0

8.5

8.5

0.1

0.1

0.0 0.0 - 0.0 - -

8.8

8.9

Total 8.6

8.5

1.1

1.1

7.5

7.6

16.8

13.5

11.1

11.0

3.8

3.6

0.3 0.4 0.7 0.5 0.0 0.0

50.0

46.2 Source: World Bank Staff calculations based on data from MOF

216

ANNEX II Classification of Central Government Expenditures across Functional and Economic Categories, 2004-2008

Table 3A: Economic and Functional Cross Classification, Central Government, 2004

(YTL million)

General Public

Services Defense

Public Order and

Safety

Economic Affairs

Environmental Protection

Housing and Community Amenities

Health Recreation, Culture and

Religion Education Social

Protection Total

Personnel 2,977 3,626 4,624 3,724 51 11 2,774 904 10,140 117 28,948 Social Security Cont. 316 574 649 626 8 2 398 138 1,299 16 4,024 Goods and Services 1,421 4,948 2,098 823 11 2 1,241 283 1,734 122 12,684 Interest expenditures 56,488 0 0 0 0 0 0 0 0 0 56,488 Current Transfers 12,182 260 4 4,192 22 63 9 330 1,913 19,389 38,364 Capital Expenditures 461 14 208 4,718 6 275 541 136 1,519 172 8,050 Capital Transfers 133 0 0 218 9 22 0 30 13 12 437 Lending 1,448 0 0 1,158 0 1 0 0 0 62 2,669 Reserves 37 0 0 0 0 0 0 0 0 0 37

Total 75,464 9,422 7,582 15,459 106 376 4,963 1,821 16,619 19,891 151,702 1/Adjusted by the revenue shares of the local administrations and funds based on their functional distribution in 2006-2008 Table 3B: Economic and Functional Cross Classification, Central Government, 2005

(YTL million)

General Public

Services Defense

Public Order and

Safety

Economic Affairs

Environmental Protection

Housing and Community Amenities

Health Recreation, Culture and

Religion Education Social

Protection Total

Personnel 3,264 3,967 5,221 3,030 55 9 3,778 1,040 11,362 130 31,856Social Security Cont. 364 655 754 499 9 1 549 163 1,521 18 4,533Goods and Services 1,210 5,261 2,116 968 12 1 2,338 299 1,896 187 14,290Interest expenditures 45,680 0 0 0 0 0 0 0 0 0 45,680Current Transfers 15,002 283 41 6,365 37 122 10 453 1,974 23,825 48,112Capital Expenditures 479 27 210 5,869 6 236 665 254 1,856 82 9,684Capital Transfers 250 0 203 633 63 310 2 34 24 25 1,546Lending 541 0 0 1,152 0 2 0 0 0 90 1,784Reserves 0 0 0 0 0 0 0 0 0 0 0

Total 66,790 10,193 8,545 18,516 181 682 7,342 2,244 18,634 24,357 157,4831/Adjusted by the revenue shares of the local administrations and funds based on their functional distribution in 2006-2008

217

Table 3C: Economic and Functional Cross Classification, Central Government, 2006

(YTL million)

General Public

Services Defense

Public Order and

Safety

Economic Affairs

Environmental Protection

Housing and Community Amenities

Health Recreation, Culture and

Religion Education Social

Protection Total

Personnel 3,922 4,363 5,832 3,123 62 12 4,180 1,337 13,046 144 36,021 Social Security Cont. 431 689 818 525 9 2 614 197 1,670 19 4,975 Goods and Services 1,565 6,616 2,481 1,169 15 7 2,318 400 2,833 316 17,721 Interest expenditures 46,260 0 0 0 0 0 0 0 0 0 46,260 Current Transfers 16,394 244 6 5,695 1 1,636 12 191 812 24,118 49,108 Capital Expenditures 816 43 1,063 6,382 30 316 1,005 412 2,241 143 12,452 Capital Transfers 565 0 0 411 6 783 6 33 30 0 1,834 Lending 965 0 0 2,259 0 2 0 4 921 106 4,256 Reserves 1,695 0 0 0 0 0 0 0 0 0 1,695

Total 72,614 11,953 10,200 19,563 125 2,759 8,135 2,574 21,552 24,846 174,322

Table 3D: Economic and Functional Cross Classification, Central Government, 2007

(YTL million)

General Public Services Defense

Public Order and

Safety

Economic Affairs

Environmental Protection

Housing and Community Amenities

Health Recreation, Culture and

Religion Education Social

Protection Total

Personnel 4,189 4,660 6,257 3,288 67 13 4,467 1,449 13,993 153 38,538 Social Security Cont. 776 848 1,086 612 12 3 833 257 2,515 28 6,972 Goods and Services 1,321 6,757 2,185 1,094 10 7 416 265 2,120 267 14,441 Interest expenditures 44,084 0 0 0 0 0 0 0 0 0 44,084 Current Transfers 17,343 254 6 5,782 1 1,710 5,013 201 850 23,419 54,580 Capital Expenditures 807 35 1,111 6,501 31 329 961 442 2,493 206 12,915 Capital Transfers 590 0 0 438 7 819 8 35 28 0 1,925 Lending 847 0 0 1,381 0 2 0 0 951 110 3,291 Reserves 1,743 0 0 0 0 0 0 0 0 0 1,743 Total 71,700 12,555 10,645 19,097 128 2,883 11,698 2,649 22,951 24,183 178,489

218

Table 3E: Economic and Functional Cross Classification, Central Government, 2008

(YTL million)

General Public

Services Defense

Public Order

and Safety

Economic Affairs

Environmental Protection

Housing and

Community Amenities

Health Recreation, Culture and

Religion Education Social

Protection Total

Personnel 4,350 4,840 6,500 3,426 70 13 4,640 1,506 14,536 160 40,040 Social Security Cont. 819 865 1,127 637 13 3 866 268 2,605 30 7,232 Goods and Services 1,415 7,026 2,273 1,140 11 7 432 274 2,213 293 15,085 Interest expenditures 38,601 0 0 0 0 0 0 0 0 0 38,601 Current Transfers 18,035 265 6 6,114 1 1,779 5,514 209 884 26,251 59,057 Capital Expenditures 755 33 1,126 6,728 31 330 1,047 470 2,616 234 13,369 Capital Transfers 611 0 0 453 7 851 9 36 30 0 1,996 Lending 907 0 0 1,435 0 2 0 0 989 115 3,448 Reserves 1,783 0 0 0 0 0 0 0 0 0 1,783

Total 67,275 13,029 11,033 19,933 132 2,986 12,507 2,763 23,873 27,082 180,613

Source: World Bank Staff calculations based on data from MOF

219

Table 4A: Economic and Functional Cross Classification, Central Government, 2004 (% of GDP)

(% of GDP) General Public

Services Defense

Public Order and

Safety Economic

Affairs Environmental

Protection

Housing and

Community Amenities Health

Recreation, Culture

and Religion Education

Social Protection Total

Personnel 0.7 0.8 1.1 0.9 0.0 0.0 0.6 0.2 2.4 0.0 6.7 Social Security Cont. 0.1 0.1 0.2 0.1 0.0 0.0 0.1 0.0 0.3 0.0 0.9 Goods and Services 0.3 1.1 0.5 0.2 0.0 0.0 0.3 0.1 0.4 0.0 2.9 Interest expenditures 13.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 13.1 Current Transfers 2.8 0.1 0.0 1.0 0.0 0.0 0.0 0.1 0.4 4.5 8.9 Capital Expenditures 0.1 0.0 0.0 1.1 0.0 0.1 0.1 0.0 0.4 0.0 1.9 Capital Transfers 0.0 0.0 0.0 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.1 Lending 0.3 0.0 0.0 0.3 0.0 0.0 0.0 0.0 0.0 0.0 0.6 Reserves 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

Total 17.5 2.2 1.8 3.6 0.0 0.1 1.2 0.4 3.9 4.6 35.2 1/Adjusted by the revenue shares of the local administrations and funds based on their functional distribution in 2006-2008.

Table 4B: Economic and Functional Cross Classification, Central Government, 2005 (% of GDP)

(% of GDP) General Public

Services Defense

Public Order and

Safety Economic

Affairs Environmental

Protection

Housing and

Community Amenities Health

Recreation, Culture

and Religion Education

Social Protection Total

Personnel 0.7 0.8 1.1 0.6 0.0 0.0 0.8 0.2 2.3 0.0 6.5 Social Security Cont. 0.1 0.1 0.2 0.1 0.0 0.0 0.1 0.0 0.3 0.0 0.9 Goods and Services 0.2 1.1 0.4 0.2 0.0 0.0 0.5 0.1 0.4 0.0 2.9 Interest expenditures 9.4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 9.4 Current Transfers 3.1 0.1 0.0 1.3 0.0 0.0 0.0 0.1 0.4 4.9 9.9 Capital Expenditures 0.1 0.0 0.0 1.2 0.0 0.0 0.1 0.1 0.4 0.0 2.0 Capital Transfers 0.1 0.0 0.0 0.1 0.0 0.1 0.0 0.0 0.0 0.0 0.3 Lending 0.1 0.0 0.0 0.2 0.0 0.0 0.0 0.0 0.0 0.0 0.4 Reserves 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

Total 13.7 2.1 1.8 3.8 0.0 0.1 1.5 0.5 3.8 5.0 32.3 1/Adjusted by the revenue shares of the local administrations and funds based on their functional distribution in 2006-2008

220

Table 4C: Economic and Functional Cross Classification, Central Government, 2006 (% of GDP)

(% of GDP) General Public

Services Defense

Public Order and

Safety Economic

Affairs Environmental

Protection

Housing and

Community Amenities Health

Recreation, Culture

and Religion Education

Social Protection Total

Personnel 0.7 0.8 1.1 0.6 0.0 0.0 0.8 0.2 2.4 0.0 6.6 Social Security Cont. 0.1 0.1 0.2 0.1 0.0 0.0 0.1 0.0 0.3 0.0 0.9 Goods and Services 0.3 1.2 0.5 0.2 0.0 0.0 0.4 0.1 0.5 0.1 3.3 Interest expenditures 8.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 8.5 Current Transfers 3.0 0.0 0.0 1.1 0.0 0.3 0.0 0.0 0.1 4.4 9.1 Capital Expenditures 0.2 0.0 0.2 1.2 0.0 0.1 0.2 0.1 0.4 0.0 2.3 Capital Transfers 0.1 0.0 0.0 0.1 0.0 0.1 0.0 0.0 0.0 0.0 0.3 Lending 0.2 0.0 0.0 0.4 0.0 0.0 0.0 0.0 0.2 0.0 0.8 Reserves 0.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.3

Total 13.4 2.2 1.9 3.6 0.0 0.5 1.5 0.5 4.0 4.6 32.2

Table 4D: Economic and Functional Cross Classification, Central Government, 2007 (% of GDP)

(% of GDP) General Public

Services Defense

Public Order and

Safety Economic

Affairs Environmental

Protection

Housing and

Community Amenities Health

Recreation, Culture

and Religion Education

Social Protection Total

Personnel 0.7 0.8 1.1 0.6 0.0 0.0 0.8 0.2 2.4 0.0 6.5 Social Security Cont. 0.1 0.1 0.2 0.1 0.0 0.0 0.1 0.0 0.4 0.0 1.2 Goods and Services 0.2 1.1 0.4 0.2 0.0 0.0 0.1 0.0 0.4 0.0 2.4 Interest expenditures 7.4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 7.4 Current Transfers 2.9 0.0 0.0 1.0 0.0 0.3 0.8 0.0 0.1 3.9 9.2 Capital Expenditures 0.1 0.0 0.2 1.1 0.0 0.1 0.2 0.1 0.4 0.0 2.2 Capital Transfers 0.1 0.0 0.0 0.1 0.0 0.1 0.0 0.0 0.0 0.0 0.3 Lending 0.1 0.0 0.0 0.2 0.0 0.0 0.0 0.0 0.2 0.0 0.6 Reserves 0.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.3

Total 12.1 2.1 1.8 3.2 0.0 0.5 2.0 0.4 3.9 4.1 30.0

221

Table 4E: Economic and Functional Cross Classification, Central Government, 2008 (% of GDP)

(% of GDP) General Public

Services Defense

Public Order and

Safety Economic

Affairs Environmental

Protection

Housing and

Community Amenities Health

Recreation, Culture

and Religion Education

Social Protection Total

Personnel 0.7 0.7 1.0 0.5 0.0 0.0 0.7 0.2 2.3 0.0 6.2 Social Security Cont. 0.1 0.1 0.2 0.1 0.0 0.0 0.1 0.0 0.4 0.0 1.1 Goods and Services 0.2 1.1 0.4 0.2 0.0 0.0 0.1 0.0 0.3 0.0 2.3 Interest expenditures 6.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 6.0 Current Transfers 2.8 0.0 0.0 0.9 0.0 0.3 0.9 0.0 0.1 4.1 9.1 Capital Expenditures 0.1 0.0 0.2 1.0 0.0 0.1 0.2 0.1 0.4 0.0 2.1 Capital Transfers 0.1 0.0 0.0 0.1 0.0 0.1 0.0 0.0 0.0 0.0 0.3 Lending 0.1 0.0 0.0 0.2 0.0 0.0 0.0 0.0 0.2 0.0 0.5 Reserves 0.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.3

Total 10.4 2.0 1.7 3.1 0.0 0.5 1.9 0.4 3.7 4.2 28.0 Source: World Bank Staff calculations based on data from MOF

222

ANNEX III

Estimating Structural and Cyclical Budget Balances

1. The structural budget balance reflects a fiscal policy adjusted for the effects of 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 of output and therefore does not include the effect of cyclical developments in economic activity. In contrast, the actual budget balance does include the effect of the cyclical component of economic activity and therefore fluctuates around the structural budget balance. The methodology for estimating structural budget balance consists of 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 of robust revenue and expenditure elasticities with respect to economic activity and the estimation of potential output which is 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 of 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 of the economic activity is calculated as the difference between trend and actual output. Although, this approach is simple and transparent, it suffers from end point bias. That is 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 is to estimate potential output based on a production function. The production function approach is relatively more complex since it requires judgments on the rate of technological change and the rate of structural unemployment. However, it is less susceptible to the end point bias for recent observations. The production function approach is used by many international institutions such as OECD and the EU as well as by many countries. It is also used by the Turkish government since 2003, in its annual Pre-Accession Economic Programmes (PEP) submitted to the EU. Considering the methods to determine the sensitivity of 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 of actual output from potential using estimated elasticities linking revenues and expenditures to economic activity. Therefore, the structural budget balance is defined as follows:

Pi

ii

Y

OERGETaxsb

∑ +−=

**

*

sb* : Structural budget balance (as a ratio of national income),

223

Taxi* : The structural value of tax revenues in category i,

GEi* : The structural value of government expenditures in category i,

OER : Other expenditures and revenues which are not affected by the growth cycles,

YP : Potential output.

The relationship between structural tax revenues and expenditures and their actual values can be shown as follows:

;;** ii

YY

GEGE

YY

TaxTax P

i

iP

i

iβα

⎥⎦

⎤⎢⎣

⎡=⎥

⎦

⎤⎢⎣

⎡=

Taxi : Actual tax revenues in category i, GEi : Actual government expenditures in category i, Y : Actual national income, αi : The output elasticity of category i taxes, βi : The output elasticity of category i government expenditures.

4. In this study, the potential output is calculated by using the Production function Method. Tax revenues are divided into 3 categories showing cyclical movements. These categories are taxes levied on income, corporate taxes and indirect taxes. Following the standard practice in the literature and in international applications, the elasticity of indirect taxes is taken as unit elasticity. On the other hand, the elasticity of the other two taxes was estimated from regression analysis. In the case of 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 of 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 still 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 of the structural budget balance;

Pi

i

p

Y

OERGETaxY

Y

sb

i

)(3

1*

++⎥⎦

⎤⎢⎣

⎡

=∑=

α

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.

*** sbbcb −=

224

cb** : Cyclical budget balance (as a ratio of national income) b : Actual budget balance (as a ratio of national income)

6. In the calculation of the potential output with a production function, it was assumed that output is determined as a function of labor, capital stock and TFP. A Cobb-Douglas function with constant returns to scale is used as production function and TFP is calculated as Solow residual.

tttt TFPKLY αα −= 1

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 of labor-force was estimated by taking into consideration the historical trend of the participation rate in the labor-force and then the estimated labor-force supply series was corrected by nairu (non-accelerating inflation rate of unemployment). Concerning the estimation of nairu, under the variable nairu assumption, the unemployment rate was put through the Hodrick-Prescott filter and the nairu series was calculated134.

[ ] ptttttt

ptt

pt

p TFPKnairuPARTPOPTFPKLY **)1(**15 11 αααα −− −== K: Capital stock TFPP: Potential total factor productivity YP: Potential GNP LP: Potential labour force POP15: Working age population PART: Participation rate nairu: Non-accelerating inflation rate of unemployment

8. Regression methodology is used for the estimation of tax elasticities. First of all, following the standard practice in the literature and in international applications, the elasticity of indirect taxes is 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 CIT would normally be expected and therefore, using Engel-Granger methodology, a co-integrating relation is tested between CIT 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 is to say all variables are integrated of order one and therefore a con-integration relation can be tested between income and tax variables.

9. Considering first CIT, the time path of 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 is tested by regressing log of CIT on a constant, log of income and two time dummies in order to account for the broken time path and then residuals are tested

134 For a detailed discussion of the topic see Denis C., K.Mc Morrow, W. Röger, “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.

225

for stationarity. The results show a strong and statistically significant relation between income and CIT with an estimated elasticity of 1.57. The stationarity of the residuals indicate a co-integrating relation between the two variables.

10. Similarly, the time path of PIT as ratio to GNP is presented in the Figure A3.3 below. The figure shows that the ratio of 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 of the sample is hard to be explained by the developments in GNP. Therefore, we chose the more stable second part of 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 of 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 of 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 of 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 is tested for two different samples, 1984-2003 and 1984-2005.

11. First, a co-integration is tested by regressing log of PIT on a constant, log of 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 of 1.50. The stationarity of the residuals indicate a co-integrating relation between the two variables. A second co-integration relation is tested by regressing log of PIT on a constant and log of 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 of 1.46. The stationarity of the residuals indicate a co-integrating relation between the two variables. Both equations support existence of 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 of 1.5 estimated for the full sample in the calculation of structural budgets.

12. Finally the time path of 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 of SSC on a constant, log of 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 of 0.82. The stationarity of the residuals indicate a co-integrating relation between the two variables.

226

Table A3.1: Unit Root tests

Null Hypothesis: Variable has a unit root

(Sample: 1975-2005, Lag selection: automatic based on SIC, MAXLAG=4)

Level with constant 1st difference with constant

t-Statistic Prob.* t-Statistic Prob.*

CIT 0.357518 0.9775 -5.672049 0.0001

PIT -1.425974 0.5563 -5.541518 0.0001

SSC 2.134058 0.9998 -3.093936 0.0382

GNP 0.950752 0.9948 -5.674495 0.0001

*MacKinnon (1996) one-sided p-values

Test critical values:

1% level -3.808546

5% level -3.020686

10% level -2.650413

Level with trend 1st 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 p-values

Test critical values:

1% level -4.498307

5% level -3.658446

10% level -3.268973

Source: World Bank Staff calculations

227

Corporate Income Tax

Figure A3.1: CIT as a Ratio to GNP

0.5

1.0

1.5

2.0

2.5

1975 1980 1985 1990 1995 2000 2005

CIT/GNP*100

Source: World Bank Staff calculations Table A3.2: Regression of CIT on Income Dependent Variable: LOG(CIT) Method: Least Squares Date: 05/18/06 Time: 10:13 Sample (adjusted): 1975 2005 Included observations: 31 after adjustments

Variable Coefficient Std. Error t-Statistic Prob.

C -11.08901 1.947756 -5.693223 0.0000LOG(GNP) 1.570558 0.174202 9.015715 0.0000

T1_CIT 0.047138 0.008424 5.595815 0.0000T3_CIT 0.053354 0.015482 3.446170 0.0019

R-squared 0.918644 Mean dependent var 6.966399Adjusted R-squared 0.909605 S.D. dependent var 0.654503S.E. of regression 0.196781 Akaike info criterion -0.293532Sum squared resid 1.045519 Schwarz criterion -0.108501Log likelihood 8.549744 F-statistic 101.6256

228

Durbin-Watson stat 1.558320 Prob(F-statistic) 0.000000

Source: World Bank Staff calculations Figure A3.2: Actual, Fitted and Residuals

-.4

-.2

.0

.2

.45.56.06.57.07.58.08.5

1975 1980 1985 1990 1995 2000 2005

Residual Actual Fitted

Source: World Bank Staff calculations

229

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

230

Personal Income Tax

Figure A3.3: PIT as a Ratio to GNP

3

4

5

6

7

8

1975 1980 1985 1990 1995 2000 2005

PIT/GNP*100

Source: World Bank Staff calculations Table A3.4: Regression of PIT on Income Dependent Variable: LOG(PIT) 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 0.0000LOG(GNP) 1.502484 0.136301 11.02326 0.0000

T1_PIT -0.142184 0.043240 -3.288240 0.0039

R-squared 0.875356 Mean dependent var 8.430307Adjusted R-squared 0.862236 S.D. dependent var 0.339875S.E. of regression 0.126150 Akaike info criterion -1.176568Sum squared resid 0.302362 Schwarz criterion -1.027790Log likelihood 15.94225 F-statistic 66.71730Durbin-Watson stat 1.730146 Prob(F-statistic) 0.000000

Source: World Bank Staff calculations

231

Figure A3.4: Actual, Fitted and Residuals

-.2

-.1

.0

.1

.2

.3

7.6

8.0

8.4

8.8

9.2

84 86 88 90 92 94 96 98 00 02 04

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 -3.831511

5% level -3.029970 10% level -2.655194

*MacKinnon (1996) one-sided p-values.

Source: World Bank Staff calculations

232

Table A3.6: Regression of PIT on Income Dependent Variable: LOG(PIT) 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 0.0001LOG(GNP) 1.460238 0.138625 10.53372 0.0000

R-squared 0.860421 Mean dependent var 8.405536Adjusted R-squared 0.852667 S.D. dependent var 0.346960S.E. of regression 0.133177 Akaike info criterion -1.099630Sum squared resid 0.319252 Schwarz criterion -1.000057Log likelihood 12.99630 F-statistic 110.9593Durbin-Watson stat 1.570086 Prob(F-statistic) 0.000000

Source: World Bank Staff calculations

233

Figure A3.5: Actual, Fitted and Residuals

-.2

-.1

.0

.1

.2

.3

7.6

8.0

8.4

8.8

9.2

84 86 88 90 92 94 96 98 00 02

Residual Actual Fitted

Source: 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, MAXLAG=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

234

Social Security Contributions

Figure A3.6: SSC as a Ratio to GNP

.02

.03

.04

.05

.06

.07

1975 1980 1985 1990 1995 2000 2005

SSC/GNP

Source: World Bank Staff calculations Table A3.8: Regression of 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 0.5603LOG(GNP) 0.815429 0.253353 3.218549 0.0036

SSC_T1 0.062682 0.012469 5.027038 0.0000

R-squared 0.960094 Mean dependent var 8.027606Adjusted R-squared 0.956902 S.D. dependent var 0.684470S.E. of regression 0.142097 Akaike info criterion -0.963658Sum squared resid 0.504788 Schwarz criterion -0.820922Log likelihood 16.49122 F-statistic 300.7369Durbin-Watson stat 0.674616 Prob(F-statistic) 0.000000

Source: World Bank Staff calculations

235

Figure A3.7: Actual, Fitted and Residuals

-.4

-.2

.0

.2

.4

7.0

7.5

8.0

8.5

9.0

9.5

78 80 82 84 86 88 90 92 94 96 98 00 02 04

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 -3.711457

5% level -2.981038 10% level -2.629906

*MacKinnon (1996) one-sided p-values.

Source: World Bank Staff calculations

236

ANNEX IV

Tax performance: A Regression Model135 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 of 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 is used for estimating tax revenue, where T/Y is the tax ratio and Xi (i = 1....n) represent various independent variables expected to influence the tax ratio, while U is the error term:

T/Y = 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 of trade to GDP (imports plus exports over GDP), the share of the mining sector and the agricultural sector in GDP, and population growth. An overview of the variables applied in previous empirical studies is 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 ECA and 6 non-ECA comparators countries over the period 1995-2004. The choice of sample is motivated by the need to obtain a data set composed of countries with similar characteristics to ECA and comparator countries, as well as data availability of 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 MOFs in respective countries. A set of 57 countries was used, comprising three groups: 10 in the lower middle income group136, 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 137 above$10,065 GNI per capita (31 countries in our sample)

237

Table 1 Overview of empirical findings in tax effort studies

Variable

Pian

cast

elli

(200

1)

Torg

ler

(200

4)

Wol

deM

aria

m

(199

7)

Grig

oria

n (2

005)

.Tee

ra, 2

004

Elto

ny (2

002)

1. Economic development GDP per capita + +/- + + +/- + Population density + Population growth - Urban Population +

2. Economic structure Agriculture, value added as % of GDP - + - - - - Manufacturing, value added as % of GDP + + +/- Mining, value added as % of GDP - - Services, value added as % of GDP GDP +

3. Openness Import, as % of GDP +/- + Export, as % of GDP + - Trade (Export + Import as % of GDP) + - - +

4.. Control variables External Debt, as % of GDP - Consumer Price Index - Inequality - Aid, as percentage to GDP + Share of Fuel in total export +

5.Institutions Shadow economy, as % of GDP - - +/- Index governance + Regulation to entry - Composite Institutional quality + Tax morale +/- Method of estimation

Fix

effe

ct m

odel

OLS

Fix

and

Ran

dom

m

odel

s

Fix

effe

ct m

odel

Fixe

d ef

fect

s mod

el,

Het

eros

keda

stic

ity-

cons

iste

nt st

anda

rd e

rror

s

Fix

effe

ct m

odel

Source: World Bank Staff calculations 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 of the fixed effects model.

238

However, in the next step the normal distribution of the error term was rejected and diagnostic tests revealed problems of cross-sectional correlation. To deal with the problem of cross-sectional correlation, the Prais-Winsten estimators138 were employed.

Table 2 Panel Regression Outcome (Prais-Winsten estimation), 1995-2004

EQ1 EQ2 Base

EQ3 Sensitivity

EQ4 Sensitivity

EQ5 Sensitivity

EQ6 Endogenity

EQ7 Endogenity

GDP per capita .001* (.000)

.001* (.000)

.001* (.000)

.001* (.000)

.000* (.000)

.001* (.000)

.001* (.000)

Trade .059* (.007)

.056* (.006)

.056* (.007)

.051* (.007)

.046* (.010)

.056* (.009)

.025** (.013)

Agriculture -.137* (.048)

-.113* (.046)

-.112* (.051)

-.149* (.062)

-.185* (.049)

-.104* (.039)

-.224* (.033)

Manufacturing .098 (.064)

.161* (.082)

.222* (.091)

Population growth -1.881* (.320)

-1.767* (.309)

-1.729* (.329)

-1.720* (.342)

-1.356* (.373)

-1.643* (.318)

-1.781* (.306)

Dummy lower middle income

-3.368* (1.232)

-3.135* (1.028)

-1.324* (1.119)

-.719* (1.119)

-3.249* (1.161)

0.734 (0.100)

Dummy high income .438 (.432)

1.028* (.410)

1.618* (.248)

1.691* (.375)

1.467* (.539)

1.598* (.518)

Oil dummy -5.437* (1.285)

-4.742* (.927)

-5.356* (.966)

-5.147* (.966)

-6.863* (.928)

-5.410* (1.320)

-3.817* (1.312)

Trend -.151 (.099)

Population rural -.008* (.001)

-.008* (.001)

-.010* (.001)

CPI -.039* (.015)

-.031* (.019)

External Debt -.046* (.012)

GDP per capita*Manuf. -.00002* (2.391)

-.000* (1.750)

GDP per capita*Agri. .000* (5.690)

GDP per capita*Trade 2.740* (4.150)

Const 324..56 (198.52)

20.96* (.867)

22.26* (.867)

25.72* (1.754)

29.59* (2.271)

17.38* (1.386)

16.51* (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 *significant at five percent level ** significant at ten percent level Standard errors reported in brackets Note: external debt variable only available for developing countries. Source: World Bank Staff calculations 138 The method is an alternative to feasible generalized last squares for fitting the linear cross-sectional time –series models when disturbances are not assumed to be independent and identically distributed, and it is preferable to the feasible GLS when the number of observations and time span are limited.

239

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 is associated with a higher tax ratio. The structure of the economy seems also to matter. The tax ratio is negatively related to the share of agriculture in GDP and positively related to the share of manufacturing sector in GDP, but the latter variable proved to be statistically insignificant (equation 1). The insignificance of manufacturing is somewhat surprising, although it may potentially be explained by a negative correlation with agricultural share or the fact that the manufacturing share varies across countries dependent on the stage of 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 term may indicate that as countries develop the importance of manufacturing as a source of tax revenue declines. Moreover, the agriculture sector is much more difficult to tax for less developed countries, as indicated by the significantly positive coefficient on the agriculture-GDP per capita interactive term. A faster rate of population growth leads to a lower tax ratio, while openness is associated with a higher tax ratio. Inclusion of dummy variables controls for differences in stage of 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, it 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 of a country’s taxable capacity vis-à-vis its natural resource base a dummy variable139 for important oil producer countries was included. The easiness of taxing natural resource extraction is likely to generate more tax revenue than non-fuel activities. The coefficient for oil dummy has the predicted negative signs and is statistically significant in all equations. Finally, the trend variable is generally negative, indicating that, all else equal, tax ratios are on a downward trend as a result of global tax competition. However, it is not statistically significant (equation 1). A similar exercise was carried out with respect to the indirect and direct taxes collection across countries.

139 OIL dummy takes value of 1 if the share of fuel (and related products) in total merchandize exports exceeds 40% is negative and insignificant

240

ANNEX V

DRAFT TAX BENCHMARKS FOR TAX EXPENDITURES

A tax benchmark identifies the normal tax arrangements that apply to similar classes of taxpayers or types of activity. Tax expenditures are defined as deviations from the benchmark. The principal criterion of benchmark design is that the benchmark

• represents a consistent tax treatment of similar activities or classes of taxpayers and neither favors nor disadvantages similarly placed activities or classes of 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 of taxation o simplify or make feasible tax administration with respect to a class of taxpayers or

type of activity • ensures that tax expenditure report provides sufficient information for policy formulation.

In addition, each country will have its own purposes or demands for a tax expenditure report so that the benchmark should reflect such purposes or demands. The following benchmarks summarize some of the basic features that could be consider forming the benchmarks of 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 on world wide income, non-residents on income sourced from Turkey • Taxable net income is gross income reduced by the allowable cost s of 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 is 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 of income) (as given in the law)

o Final withholding on specified types of 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)

241

• Exemptions provided for persons subject to diplomatic privileges on a reciprocal basis • Exemption provisions for central and local governments and non-government

organizations of a public nature in the context of providing services of a public nature on a non-commercial basis

Corporate Income Tax Benchmark

• Tax period: calendar year /end of the financial period of 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 of 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 of surplus of cooperatives as part of tax integration • Up to 50% deduction of repatriated profits as part of 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 of a public nature in the context of providing services of a public nature on a non-commercial basis

VAT Tax Benchmark

• Tax period: month for domestic deliveries (supplies) and quarter for imports • Valuation of domestic deliveries and imports (given in the law) • Timing of recognition of 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 of 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 of input VAT and refund of excess VAT only in case of

exports, and supplies at rates below standard rate; excess input VAT carried forward. • Exemptions related to normal customs treatment of transit trade, free zones, bonded

warehousing, temporary importation, change of residence, etc • Deferment mechanisms (or equivalent structures) for VAT on imported and domestically

supplied machinery and equipment

242

• Exemption provisions for government and its agencies and non-government organization providing services of a pubic character and on a non-commercial basis (as specified in the law)

• Exemption for self-employed businessmen, tradesmen and farmers (as specified in the law)

• Exemption of mergers and acquisitions, delivery and rental of non-commercial immovable property; financial sector; delivery of financial instruments.

• Deliveries to persons with diplomatic privileges are fully exempted (zero rated) Special Consumption Tax Benchmark

• Destination based single stage tax • Tax period: monthly • Valuation of imports and domestic deliveries at factory gate, dealer or retail level as

specified in law for the different specified commodities • Timing of payments (as specified in the law) • Tax rates as specified for lists of goods

o List I: tax rates specified for petroleum products o List II, tax rates specified for vehicles o List III

tax rates specified for tobacco products tax rates specified for alcoholic products

o List IV. Tax rate at 6.7% on specified luxury items • Exemptions for exports and for persons with diplomatic privileges

• Exemptions related to normal customs treatment of transit trade, free zones, bonded warehousing, temporary importation, change of residence, etc

• Exemption for donations to government and its agencies

243

ANNEX VI

Towards Estimates of Tax Expenditures in PIT and CIT

The following estimates are based on the summary PIT and CIT tax return for 2003. Based on analysis of tax expenditure tables and some added data, there are a number of questions, comments and observations towards estimates of tax expenditures in PIT and CIT.

PIT

1. Little can be gathered from the 2003 PIT return. Out of the YTL 6.393 billion in income, some 29.2% is 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 ITL Articles 89/5-8) These represent small tax expenditures of about YTL 5 million for item 26 and YTL 5.1 million for item 27 respectively as the bulk of these deductions would occur as a result of premiums paid by employees and would be captured under the PAYE. In addition, and importantly, most of the PIT tax expenditures (some 100 provisions of 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 is considerably richer in tax expenditure information, but without (i) use of a tax model and (ii) consideration of the inter-temporal impacts on loss and exemption/deduction/allowance carryovers, it still has limitations. For example, a deduction taken in one period can result in a loss of 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 of the exemption or deduction, and (c) the difference in timing of accrual and realization of the exemption or deduction:

244

a. Where there are different tax rates in different tax brackets, the income level of the beneficiary of the exemption or deduction affects the marginal tax rate (MTR) applied to the exemption or deduction. Ideally, a weighted average MTR is required to apply to the amount of the exemption or deduction. This is clearly an issue in estimating tax expenditures in the PIT, including any flow through of dividends and other investment income from the company to personal levels. This is typically solved by calculating tax expenditures using a micro simulation tax model based on a representative sample of taxpayers so that the appropriate weights and MTRs are applied depending on the distribution of taxable income.

b. The size of exemptions and deductions affects tax expenditure estimates in two ways.

i. The MTR is affected by how many tax brackets are covered by the exemption or deduction. This is most important if tax expenditure items are considered collectively rather than individually or one-at-a time. The best-practice tax expenditure convention is 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 is applicable. Again the use of micro simulation tax models with a sample of representative taxpayers in terms of taxable incomes and deductions allows the accurate estimation of each tax expenditure item.

ii. An exemption or deduction may make a taxpayer move from being taxable to non-taxable such that only part of the exemption or deduction has current cash flow consequences. The tax value of 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 exemption/deduction 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 of 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 of Turkey). From a tax flow perspective, the potential tax value of an exemption or deduction in the current year is 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 of the new deductions in the year plus the carry forwards of deductions taken but not used in prior years. This issue of the timing of when tax expenditures are realized is a major and difficult problem. It is 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 of taxable income. This problem is 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 is a loss,” which generate either traceable carry forwards or general (untraceable) loss carry forwards, and as “exemptions deducted when there is a profit,” which have to be carried forward separately when there is not a profit. The actual cash flow tax expenditure cost of the former is difficult to determine in any current year. In addition to some of 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 of the exemption or deduction in an earlier year

245

affects the loss carry forwards into the current year and hence affects the current tax value of current exemptions or deductions. The tax expenditure treatment of exemptions that do not cause a loss carry forward (but unused amounts are carried forward separately) are treated subsequently below.

246

Source: Tax Revenue Administration and World Bank Staff

Table 1: Towards some preliminary CIT tax expenditure estimates and identification of issues

30% of exemption or allowance Questions, issues, concerns

YTL millionsExemptions to be deducted even when there is loss

Income from Investment Funds (CTL Article 8/4)

1,671 Correct tax rate? Ideally should be taxed at rate of person holding investment units or shares.

Income from Pre-emptive Rights & Issue Premiums (CTL Art. 8/5)

30 Not all current exemption generates a current tax loss, but prior years exemptions may be currently deductible

Income From Constructing Abroad (CTL Art.8/7)

126 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

C.T.L. Temporary Article 28/a 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 Law 3218)

89 Not all current exemption generates a 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 Less Developed Regions

33 Not all current exemption generates a 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 Exemptions

236

Exemptions to 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

247

6. The cash flow tax cost can be broken out into two main components.

i. The actual amount of 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 of the exemption or deduction used in the year. Alternatively, the amount of the current deduction is limited by the amount of total loss carry forwards from the current year to the next year (LCFnext). This LCFnext is the total of traced and untraced LCFs and the unused amounts of the taxable income limited deductions (such as the investment allowance). Unless the current deduction exceeds the LCFnext, its removal will result in no current tax, but merely a decrease in LCFnext.

ii. The removal of the exemption or deduction of the tax expenditure type in previous years could have resulted in lower loss carry forwards into the year, and hence the amount of the current exemption or deduction that would have been claimed in (i) above increases. The amount of carry-forwards of unused deductions or exemptions of the same type used in the year. This can be precisely determined if the amount of the unused exemptions or deductions are traceable (or are in type-specific pools carried forward). Otherwise some share of the loss or exemption carry forwards has to be attributed to the type of tax expenditure. Table 1 above shows a tax cost of some YTL 222 millions in exemptions carried forward and claimed in 2003. In addition, some share of the YTL 3.577 billon in general losses carried forward into 2003 may be arising from untraced exemptions and deductions in earlier years rather than economic losses. Where use of carry forwards of traced deductions is known, the impacts of changes in these deductions on the carry forwards of 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 critically on (i) whether he tax payer is tax payable in the year (T>0) or not and (ii) 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 of current deductions and will have used up all LCF into the year. The total LCF into the year determines the maximum carry forward of unused deductions into the year and put an upper limit on how much of prior year unused losses can affect the tax flows in the current year. Where no tax is payable in the year and losses are carried forward into the following year, then any increase in current year taxable income because of the removal of a specific deduction, may not will not result in the full value of 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 of cross tabulating the tax returns in order to analyze and estimate the tax expenditures.

Table 2: Tax expenditure (TE) cases for current deduction (D) and marginal tax rate (t in terms of tax payable in year (T), loss carry forward (LCF) and unused prior deductions of the same type (Dprior) Tax Payable, T = 0 Tax Payable, T > 0 Loss carry forward, LCF = 0

A. Partial TE for cases where (D – LCFnext) t > 0

B. Full TE = Dt

Loss carry forward, LCF

Dprior > LCF C1. Partial TE for cases where (D +LCF – LCFnext) t > 0

D1. Partial TE = (D -LCF) t

1,671

248

> 0 Dprior < LCF C2. Partial TE for cases where (D + Dprior – LCFnext) t > 0

D2. Full TE = (D + Dprior ) t

Definitions: LCF = total loss carry-forward from prior to current year of all traceable and untraced losses plus unused deductions carried forward separately LCFnext = total loss carry-forward from current to next year of all traceable and untraced losses plus unused deductions carried forward separately Dprior = Cumulative amount of unused deductions of same type as D in prior year and each successive earlier year where no tax was payable or loss carry forwards to the next year were generated up to the legal limit for carrying forward losses of the type, generally 5 years except for investment allowances. T = tax payable on the tax base after all exemptions, deductions, allowances and preferential tax credits, but before any withholding, installment or advance taxes are credited against the tax payable. Source: World Bank Staff

7. What are the implications of the above analysis of estimating tax expenditures on a cash flow basis? First, that an annual micro simulation model is necessary, but not sufficient to estimate tax expenditures where there are loss carry forwards or losses created by exemptions or deductions. Second, additional information is required to estimate “TE”, namely, total LCF, LCFnext and Dprior . LCF and LCFnext should be knowable for each tax return and should be added to any tax return report. Dprior requires the further work of linking tax returns over the years. Finally, it is useful to cross-tabulate estimates of tax expenditures by the taxpayers status in terms of 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 of unused income-limited

deductions and special deductions where any losses generated and carried forward are traced as a special pool of losses:

a. In the case of income-limited losses, the actual amount of losses used in the current year from current deductions and unused carry forwards gives the correct TE estimate. It is automatically the sum of D+ LCF – LCFnext .

b. In the case of traced deductions, the sum of the current deduction taken (D) plus the amount of any LCF of that type used in the year has to corrected for any changes in the stock of unused income-limited deductions over the current year to be an accurate estimate. Alternatively, the methodology in table 2 based on the current deduction and changes in the stock of total LCF overhe current year can be used.

9. Are these concerns about loss carry forward quantitatively important or not in tax

expenditure estimates. Table 3 provides some clues. It is based on the summary tax return data readily at hand for 1998, 1999 and 2003 for companies. It shows that tax losses are extremely high, even excluding the impacts of investment allowances, which are only taken when there is available taxable income and do not generate explicit loss carry forwards. The following is evident: a. Loses in a year vary between 21% and 54% of taxable profits b. Combined current year losses and losses carried forward vary between 83% and 99% of

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 of loss carry forwards have declined at the rate of 23% per year through use or companies folding. This means that at least 26% of the

249

original losses are lost because of 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% of taxable income. e. In addition, exemptions that are taken only when there is a profit (essential investment

allowances) have eliminated another 27% to 35% of taxable income in the year. Unfortunately, no information is available on the investment allowances created by vintage year and the outstanding stocks of these vintages in order to find out how rapidly they are absorbed. The large size of 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 of the stocks of unused investment allowances a source of 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 of deductions is “other exemptions and deductions” (line 38) generating a direct tax expenditure of about YTL 2.1 billion in 2003. Clearly it would be useful to identify what are the major sources of this large amount of exemptions and deductions?? Aside from this item, a number of other items amongst the exemptions generating losses have other issues as noted in Table 1. The treatment of 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 of 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 of companies in 2003, if divided into those with and without these exemptions, and then into the four cells in table 2 assuming the case of Dprior > LCF, . An accurate estimate would require knowing the exemption claims (Dprior ) of these companies currently with LCF >0 in 2003 over the previous 5 years. The LCF itself would be the upper bound on the amount of the exemption carried forward into 2003 . Studies of time series f tax returns is required to find Dprior and estimate usefl probabilities such as the probability of also claiming the same type of deduction in prior years and also being not-taxpaying and generating LCF.

11. Finally, based on 8 above, the tax expenditure for investment allowances is the dominant

item and clearly very important. Table 1 shows a tax expenditure of YTL 3 billion in 2003 The amount of unused investment allowances, which are carried forward indefinitely with indexation is unknown. If the amounts of deductions, losses and loss carry forwards declines in the future these investment allowances could grow in the future as the available taxable income grows, even if the investment allowance were limited or used less. It is important to get estimates of the amount of unused investment allowances by companies broken out into the cross tabulations of table 2.

250

1998 1999 2000 2001 2002 2003

1993 Year Loss 59,275,270 1994 Year Loss 165,811,023 295,132,666 1995 Year Loss 295,054,163 195,741,720 1996 Year Loss 523,329,845 377,834,460 1997 Year Loss 1,554,065,087 859,142,529 1998 Year Loss 2,130,255,976 816,868,186 1999 Year Loss 2,332,369,545 2000 Year Losses Resulting From Exemptions 470,680,130 2000 Year Other Losses 3,811,251,522 2001 Year Losses Resulting From Exemptions 1,335,827,282 2001 Year Other Losses 15,691,193,068 2002 Year Losses Resulting From Exemptions 1,421,010,998 2002 Year Other Losses 10,694,134,141 Total loss carry forwards into current year A 2,597,535,388 3,858,107,351 36,573,334,852

Corporate income -- Balance sheet profit 8,533,191,021 11,415,537,880 52,017,055,718 Corporate income -- Balance sheet loss 5,670,315,410 5,342,585,456 12,613,998,868

Loss (40 - 39) B 2,354,408,557 5,171,942,732 10,020,344,102 Profit (39 - 40) C 5,959,746,552 9,554,319,132 47,047,726,377

B/C 40% 54% 21%(A+B)/C 83% 95% 99%

Exemptions to be deducted if a profit (52) E 1,950,500,186 2,566,081,897 10,262,709,944 Total Previous Year Loss to be Deducted (43+44) F 602,261,610 877,287,536 12,661,978,627 Base for Deductions (42 - 45) G 5,599,866,084 8,861,448,776 38,210,013,322

E/G 35% 29% 27%F/G 11% 10% 33%

Current year

YTL

Table 3 Current Year Income, Losses and Loss Usage and Loss carry Forwards for 1998 - 2003

Source: Tax Revenue Administration

251

ANNEX VII

Classification of Budget Institutions

Chart I: General Budget Institutions 1. Presidency of Republic 2. Turkish Grand National Assembly 3. Constitutional Court 4. Court of Appeals 5. Court of State 6. Turkish Court of Accounts 7. Prime Ministry 8. State Planning Organization 9. Treasury Undersecretariat 10. Foreign Trade Undersecretariat 11. Undersecretariat of Customs 12. State Institute of Statistics 13. Ministry of Religious Affairs 14. Ministry of Justice 15. Ministry of National Defense 16. Ministry of Interior 17. General Directorate of Security 18. General Commandership of Gendarmerie 19. Commandership of Coastal Security 20. Ministry of Foreign Affairs 21. Ministry of Finance 22. Ministry of National Education 23. Ministry of Public Works and Settlement 24. General Directorate of Land Registry and Cadastre 25. Ministry of Health 26. Ministry of Transport 27. Undersecretariat of Marine Affairs 28. Ministry of Agriculture and Rural Affairs 29. Ministry of Labor and Social Security 30. Ministry of Industry and Trade 31. Ministry of Energy and Natural Resources 32. Ministry of Culture and Tourism 33. Ministry of Environment and Forestry 34. General Directorate of State Meteorological Affairs

252

Chart II: Special Budget Institutions

A. Higher Education Council, Universities And High Technology Institutes 1. Higher Education Council 2. Student Selection and Placement Center 3. Universities (total of 53)

B. Administrations with Special Budgets 1- Undersecretariat for Defense Industries 2- Atatürk Culture, Language and History High Agency 3- Turkish and Middle-Eastern Public Administration Institute 4- Turkish Scientific and Technical Research Council 5- Turkish Science Academy 6- Turkish Justice Academy 7- Higher Education Credit and Dormitories Institution 8- General Directorate of Youth and Sports 9- General Directorate of State Theatres 10- General Directorate of State Opera and Ballet 11- General Directorate of Foundations 12- General Directorate of Health on Borders and Coasts 13- Electric Works Study Administration 14- Mineral Exploration and Research Institute 15- General Directorate of Civil Aviation 16- Turkish Accreditation Agency 17- Turkish Standards Institute 18- National Productivity Center 19- Turkish Patent Institute 20- National Boron Research Institute 21- Turkish Atomic Energy Agency 22- Administration for the Development and Support of Small and Medium Sized

Enterprises 23- Center for Studies on Export Development 24- Turkish Collaboration and Development Administration 25- Special Environmental Protection Institution 26- GAP Regional Development Administration 27- Privatization Administration 28- Workshop Agency for Punishment and Execution Institutions and Prisons

Chart III:Regulatory nad Supervisory Institutions

1- Radio and Television High Council 2- Telecommunication Agency 3- Capital Markets Board 4- Banking Regulation and Supervision Agency 5- Energy Market Regulation Board 6- Public Procurement Agency 7- Competition Authority 8- Tobacco, Tobacco Products and Alcoholic Beverages Market Regulation Agency

253

ANNEX VIII

BASELINE INDICATOR SYSTEM (BIS) FOR PUBLIC PROCUREMENT JOINT ASSESSMENT REPORT FOR TURKEY

In 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 procurement system is assessed. It was agreed that all indicators have the same weight and within each indicator, all sub-indicators have the same weight. For indicators with multiple sub-indicators; a “yes” or “no” answer will be provided for each sub-indicator. If over 90% of 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 is fully achieved (FA) Between 70-90% = baseline is substantially achieved (SA) Under 70% = baseline is not achieved (NA) The International Relations and Coordination with EU Department of PPA is 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 of ratings of the self-assessment by the PPA is 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: Mainstreaming Procedures into Public Financial Management = SA Indicator 4: Functional Management/Normative Body = 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

254

Indicator 8: Existence of Contract Administration and Dispute Resolution Provisions = SA Indicator 9: Effective Control and Audit System = SA Indicator 10: Efficiency of Appeal Mechanisms = FA Indicator 11: Degree of 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 of each sub-indicator with the PPA and the Bank. Following detailed discussion of ratings; the actions that might be taken by the PPA (or other relevant organizations) to improve the rating that was indicated for substantially achieved or not achieved sub-indicators are identified. Even on indicators rated as fully achieved by the PPA; the discussions enabled the Bank and the PPA to identify implementation issues that are affecting in achieving the full benefit of 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”. In 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.

255

Baseline Indicators System – Assessment Table (February 2006)

Indicator No (*)

Indicator Data Used in Assessment Self Assessment by PPA Bank’s Comments on Assessment

1 Public procurement legislative and regulatory framework achieves the agreed standards and complies with applicable obligations

1.1 Scope of application and coverage of the legislative and regulatory framework

1.1.1

S

Contracting entities at all levels, including government authorities, municipalities, regional authorities and utilities/state-owned enterprise, are covered

Public Procurement Law, Article 2 Fully Achieved Contracting entities except utilities sector are all covered by the Law. But a new draft law on public procurement of utilities sector has already been prepared by PPA.

The Bank remains concerned until the new law for Utilities Sector is enacted.

1.1.2

M

All areas of procurement: work, goods and consulting services, are included

Public Procurement Law, Article 2 Fully Achieved Consulting services procurement is regulated through a separate division in the Law Article 48-52

Public Procurement Authority (PPA) should maintain its oversight on all amendments and new laws which effects public procurement e.g. excluding certain type of services and entities from the scope of law.

1.1.3

M

Procurement using public funds, irrespective of contract value, is included

Public Procurement Law Articles 1, 2, 3

Fully Achieved

The Bank remains concerned about exemptions provided to certain type of services and entities through other laws.

1.1.4

S

The applicable legislative and regulatory framework is structured, consistent and accessible to users and all interested stakeholders

Public Procurement Law, Public Procurement Contract Law, Implementing Regulation

Fully Achieved Everybody can instantly access the website of PPA for all legislative and regulatory framework information on Public Procurement. Official Journal in which the public procurement legislations are issued is thoroughly open to the public as well.

It was agreed that in order to address the needs of international business community, translation of whole legislation into English would continue. With the help of twinning agreement; translation of main procurement legislation into English is completed. The translation is in draft form and unofficial translation is already published in PPA web-site. The translation will be finalized after a detailed review and proof-reading and official translation will be published at the PPA web-site.

1.2 Procurement methods 1.2.1

M

Stated preference for the use of open, competitive procurement unless otherwise justified in accordance with the legislative and regulatory framework

Public Procurement Law, Articles 5, 18,19,20,21,22

Fully Achieved

Open competitive procurement is not the stated preference as restricted bidding is also a standard alternative selective method to open tendering.

256

1.2.2

M

International competitive tendering methods defined for specified contracts (e.g. where monetary thresholds exist) that are consistent with international standards

Public Procurement Law, Articles 8, 13, 63

Fully Achieved In conformity with the provisions of public procurement law (PPL), on the initiative of procuring entities, unless adversely decided, procurement market is open to all bidders regardless of nationality. Procurement above specified threshold values have to be open to all international bidders in any case. (with an option for domestic preference)

The thresholds provided in Article 8 regarding international publication of notices are significantly above the EU thresholds. The PPL should be aligned with EU Directives. In 2004, Domestic preference is applied only in 2 % of the contracts awarded (13% in terms of contract value). International bidders may be prohibited to participate to procurements under a certain thresholds value. All public procurements below or above the thresholds should be open to international competitors.

1.2.3

S

Defined basis for the procurement method, if other than open competition

Public Procurement Law, Articles 18,19,20,21,22

Fully Achieved

It is agreed that data on the use of procurement methods will continue to be monitored. It is required that all procuring entities, which may use direct procurement method, are required to complete an “information form” and send to PPA for their records.

1.2.4

M

Negotiated procedures and direct purchasing only under well-defined and justified circumstances, subject to controls

Public Procurement Law, Articles 21,22, 62/ı

Fully Achieved

See above 1.2.3 These procurement procedures with restricted competition or no competition at all should be very rarely used and under very specific circumstances.

1.3 Advertising rules and time limits 1.3.1

M

Mandatory and accessible publication of opportunities for competitive procurement

Public Procurement Law, Articles 13, Annex provision 1 (4964/ Article 41) Implementing regulation

Fully Achieved

The Public Procurement Bulletin is available to all subscribers. The Bank recommends expanding its flexible availability to all interested persons. The advertisements should be made accessible to all national and foreign bidders that are not subscribers of the Bulletin. The annual subscription fee of the Bulletin for both electronic and hard copy is reduced to about $90 equivalent

1.3.2

M

Mandatory publication of result information on contract awards based on defined thresholds

Public Procurement Law, Article 47 Fully Achieved

The thresholds for the mandatory publication of contract awards are significantly high. In 2004, about 40,000 procurement notices are

257

published, whereas only 1,000 contract award notices are published. Other than the mandatory publications; the procuring entities are required by implementation regulations to complete a “procurement result form” for each procurement activity and submit it to PPA within 10 days after signing the contract, which would enable PPA to monitor results of all procurements.

1.3.3

M

Minimum time limits for submission of tenders and applications, which should be consistent with method of procurement, national conditions and, when applicable, international requirements

Public Procurement Law, Article 13, 29, 36 Implementing regulation

Fully Achieved

The thresholds provided in Article 13 regarding time limits for submission of tenders are significantly shorter than the limits provided in EU Directives. The attempts to reduce time limits should be prevented and PPL should be aligned with EU Directives. The draft amendment (prepared in June 2005) of the PPL has a provision about preliminary notices, which would result more shortening of the current time limits.

1.4 Rules on participation and qualitative selection 1.4.1

M

Fair, predictable and defined rules for participation that rely on qualifications and ability to perform the requirement

Public Procurement Law, Articles 10, 11 Implementing regulation

Fully Achieved

Implementation needs to be monitored to avoid hindering competition.

1.4.2 M

Limited and controlled use of price preferential clauses

Public Procurement Law, Article 63, Implementing regulation

Fully achieved

See above 1.2.2

1.4.3

S

Debarment process if covered, on defined basis, allowing for due process and appeal

Public Procurement Law, Articles 17, 58, Implementing regulation

Fully Achieved Any transaction and operation of public administrations in Turkey is subject for appeal to the administrative courts.

There is no provision in the law guarantees due process or appeal against debarment decisions. The data on the debarment process by various contracting entities should be monitored and consideration to create an appeal mechanism at a lower level than administrative courts, which are back logged and resulting delayed resolutions, should be given.

1.4.4

S

Rules for participation of government-owned enterprise that provide for equal treatment in competitive procurement

Public Procurement Law Article 11, paragraph 9 ( as amended by 4964/8) Implementing regulation

Fully Achieved There exists no provision of special treatment for government-owned enterprises participating in any

PPL does not restrict participation of government-owned enterprises to public tenders, except they are not allowed to participate to tenders of the contracting agencies

258

competitive procurement. which are directly related with them. In order to determine if there is any unfair competitive advantage of the state-owned enterprises, the number of contracts awarded to them should be monitored and necessary measures should be considered for equal treatment of all bidders.

1.5 Tender documentation and technical specifications 1.5.1

S The minimum content of the tender documentation is specified

Public Procurement Law, Article 27, Implementing regulation

Fully Achieved

No comment.

1.5.2

M

Neutral technical specification with reference to international standards where possible

Public Procurement Law, Article 12, Implementing regulation

Fully Achieved

It is agreed that PPA will prepare and issue a Good Practice Guidance Note on how to prepare neutral technical specifications.

1.5.3

M

Content of tender documentation is relevant to meeting requirement and implementing the process

Public Procurement Law, Articles 27, 28, 29 Implementing regulation

Fully Achieved

No comment.

1.6 Tender evaluation and award criteria

1.6.1

M

Objective, fair and pre-disclosed criteria for evaluation and award of contracts

Public Procurement Law, Articles 24, 25 and 40 Implementing regulation

Fully Achieved

To minimize implementation problems; amendment to PPL is under consideration.

1.6.2

M

Clear methodology for evaluation of tenders based on price and other fully disclosed factors leading to award of contracts

Public Procurement Law, Article 38, 40 Implementing regulation

Fully Achieved Detailed information on the methodology for evaluation is in the implementing regulation.

To minimize implementation problems; amendment to PPL is under consideration.

1.6.3

M

Requirement to maintain confidentiality during the evaluation process

Public Procurement Law, Articles 9, 17, 36, 37, 60 Public Procurement Contract Law, Article 25 Implementing regulations

Fully Achieved

No comment.

1.7 Submission, receipt and opening of tenders 1.7.1

M

Public opening of tenders in a defined manner that ensures the regularity of the proceedings

Public Procurement Law, Article 36 Implementing regulations

Fully Achieved

No comment

1.7.2

M

Clear requirement to maintain records of proceedings and process that are available for review/audit

Public Procurement Law, Article 7 Implementing regulations

Fully Achieved The minutes of public opening announcing the bid prices and existence of required documents as part of bid are provided to all participating bidders.

1.7.3

Requirement to maintain security and confidentiality of tenders prior

Public Procurement Law, Articles 30, 60, 61

Fully Achieved

No comment.

259

M to bid opening Implementing regulations 1.7.4

M

Submission and receipt modalities of tender documents are well defined

Public Procurement Law, Articles 36, 30, 31 Implementing regulations

Fully Achieved

No comment.

1.8 Complaint review procedures 1.8.1

S

Inclusion of complaint and remedy procedures that provide for fair, independent and timely implementation

Public Procurement Law, Article 53, 54, 55, 56, 57 Regulation on Administrative Applications Against Procurements

Fully Achieved The time allowed for the contracting authority to respond to complaints submitted by unsuccessful bidders can be reduced to enhance the effectiveness of appeal mechanism. An electronic system is put in place to monitor the current status of the complaint review process.

2 Existence of Implementing Regulations and Documentation

2.1

S

Implementation regulations that define processes and procedures not included in higher-level legislation

Regulation on Implementation of services procurement Regulation on Implementation of goods procurement Regulation on Implementation of works procurement Regulation on Implementation of consultancy services procurement Regulation on Administrative Applications Against Procurements

Fully Achieved

It is agreed that PPA will continue to ensure consistency within all related regulation. The entities exempted by Article 3 of PPL prepare their own procurement regulations, with the approval of PPA. The implementation regulations are being updated regularly to provide further clarifications based on the experiences gained.

2.2

S

Model tender documents for goods, works and services

Annexes of the Implementing Regulations, standard forms to be used during each step.

Fully Achieved

No comment.

2.3

S

Procedures for pre-qualification Public Procurement Law Article 20, 21, 48 Implementing Regulations Annexes of the Implementing Regulations

Fully Achieved

No comment.

2.4

S

Procedures suitable for contracting for services or other requirements where technical capacity is a key criterion

Public Procurement Law 20, 21/e, 48, 49, 52 Implementing regulations

Fully Achieved

No comment.

2.5

S

User’s Guide or manual for contracting entities

Hard copy User’s guide Electronic User’s guide Website of Public Procurement Authority

Fully Achieved

The Manual should be updated regularly based on amendments in the procurement legislation. As part of the Twinning Agreement; the Manual will be updated to include any changes in the

260

procurement legislation. 2.6

S

General Conditions of Contracts for public sector contracts covering goods, works and services consistent with national requirements and, when applicable, international requirements

Public Procurement Contract Law Annexes of the Implementing Regulations Regulation on Inspection and Acceptance General Specifications Annex of the Implementing Regulations

Fully Achieved

No comment.

3 Mainstreaming Procedures into Public Financial Management

3.1

S

Procurement planning and data on costing is part of the budget formulation process and contributes to multi-year planning

Public Procurement Law Article 5, 62 Public Financial Management and Control Law. No 5018 Articles 9, 13, 15, 16 Amendment to Public Financial Management and Control Law No 5436 (December 2005)

Substantially Achieved Public Financial Management and Control Law have been approved by the Grand National Assembly and become effective on January 2006. The law anticipates yearly basis budget but on the other hand following three years’ revenue and expenditure prediction are envisaged.

It is agreed that further studies will be done to provide data on actual cost to support budget formulation. Piloting e-procurement project would help to collect data.

3.2

S

Budget law and financial procedures support timely procurement, contract execution and payment

Public Procurement Law Article 62(b) General Accounting Law Articles 50, 51, 83 PFMC Law Articles 26, 27, 28, 35

Substantially Achieved

No comment.

3.3

S

No initiation of procurement actions without existing budget appropriations

Public Procurement Law Articles 5, 62 General Accounting Law PFMC Law

Fully Achieved

No comment.

3.4

S

Contract execution is subject to budgetary controls to ensure sufficient funding for contract

Public Procurement Law Article 5, 62/a Public Financial Management and Control Law. No 5018 Public Procurement Contracts Law Article 24

Substantially Achieved

This is out of control of PPA and should be considered under broader Public Financial Management System.

3.5

S

Budgeting system provides for timely release of funds to make payments against contractual obligations

Annual Budget Law Article 8 Public Financial Management and Control Law. No 5018 Article 20

Substantially Achieved

The system should enable the timely release of funds after the commitment of funds when award of contracts are approved.

3.6

S

Systematic completion reports are prepared for certification of budget execution and for reconciliation of delivery with budget programming

General Accounting Law Articles 100, 101 PFMC Law Articles 41, 42

Not Achieved

See above 3.1

4 Functional Management/Normative Body

261

4.1

S

The status and basis for the functional normative bodies is covered in the legislative and regulatory framework

Public Procurement Law Article 53 Regulation for Public Procurement Authority

Fully Achieved

No comment.

4.2

S

The responsibilities address a defined set of functions that include, but not limited to: providing advice to contracting entities; drafting amendments to the legislative and regulatory framework and implementing regulations; providing monitoring of public procurement; providing procurement information; managing statistical databases; reporting on procurement to other parts of government; developing and supporting implementation of initiatives to improve the public procurement system; and providing implementing tools and documents to support training and capacity development of implementing staff

Public Procurement Law Articles 53, 56 Regulation on Administrative Applications Against Procurements

Fully Achieved

PPA is still in process of consolidating authority in the areas defined in the legislation.

4.3

S

Organisation, funding and staffing and the level of independence and authority (formal power) of the bodies is sufficient and consistent with their responsibilities

Public Procurement Law Article 53 Regulation for Public Procurement Authority

Substantially Achieved

See above 4.2

4.4

S

Responsibilities provide for separation and clarity so as to avoid conflict of interest and direct involvement in the execution of procurement transactions

Public Procurement Law Articles 53, 56

Fully Achieved

No comment.

5 Existence of Institutional Development Capacity

5.1

S

A system exists for collection and dissemination of procurement information, including tender invitations, requests for proposals and contract award information

Public Procurement Law Article 53 Electronic public procurement control system exists in website of PPA Public procurement bulletin issued periodically by PPA in hard copy and electronically.

Fully Achieved

See above 1.3.1

262

5.2

S

A sustainable strategy and training capacity exists to provide training, advice and assistance to develop the capacity of government and private sector participants to understand the rules and regulations and the how they should be implemented

Public Procurement Law Article 53

Substantially Achieved PPA provides training programs and sector programs for stakeholders at their request.

It is agreed that PPA will develop a more proactive strategy for capacity development with the help of twinning agreement (such as mandatory training, minimum qualification standards to professionalize procurement staff, on-line certification of staff, awareness campaign etc.). Training of private sector participants is currently under consideration by TOBB and PPA and will start to be implemented soon.

5.3

S

Systems and procedures exist for collection and monitoring of national procurement statistics

Public Procurement Law Article Second paragraph after 53(b) 9 Electronic public procurement control system exists in website of PPA

Fully Achieved

Second paragraph after Article 53(b) 9 provides that PPA requires mandatory supply of required statistics from all contracting entities. PPA will continue to increase the data collection rate.

5.4

S

Quality control standards are disseminated and used to evaluate performance of staff and address capacity development issues

No data

Not Achieved

See above 5.2.

6 Efficient Procurement Operations Capacity and Practice

6.1

S

The level of procurement competence among government officials within the entity is consistent with their procurement responsibilities

Public Procurement Law Article 6

Substantially Achieved There is no data collected about the level of procurement competence among government officials.

See above 5.2 and 5.4

6.2

S

The procurement training and information programmes implemented for government officials and private sector participants is consistent with demand

Public Procurement Law Article 53.b.3 Regulation for Public Procurement Authority

Substantially Achieved

See above 5.2

6.3

S

The existence of administrative systems for public procurement operations, and information databases, to support monitoring of performance and reporting to and responding to the information needs of other related government systems

Procurement statistics Complain statistics Notice of procurement statistics Updated records keeping for banned contractors (All statistical data reports are prepared for 3 month period and announced via internet and press releases)

Substantially Achieved

Bank will consider providing support to PPA for analysis and application of data to improve procurement performance.

263

6.4

S

The existence and implementation of internal control mechanisms for the understanding of procurement operations at the contracting level, including a code of conduct, separation of responsibilities as a check/balance mechanism and oversight/control of signature/approval authority

Public Procurement Law Articles 54, 55, 56 Public Procurement Contract Law Article 11 Regulation on Inspection and Acceptance Law on Court of Accounts Inspection and Control Legislations of the Administrations

Fully Achieved

By the effectiveness of PFMC Law, new performance inspection mechanisms will be institutionalized.

6.5

S

The existence of norms for the safekeeping of records and documents related to transactions and control management

General Specifications Annex of the Implementing Regulations

Substantially Achieved

Implementation issues remain.

6.6

S

Provisions exist for delegation of authority to others consistent with capacity to exercise responsibilities

Legislations of the ministries, organisations and institutions.

Substantially Achieved All Ministers may delegate their own authority to the sub level bureaucracy. This is usually ruled by the respective administrative codes of the entities.

Implementation issues remain.

7 Functionality of the Public Procurement Market

7.1 There are effective mechanisms for partnerships between the public and private sector

Not applicable Not applicable.

7.2

S

Private sector institutions are well organised and able to facilitate access to the market

TOBB nominates one member of PPA Board. TOBB’s sector board (construction) is in close relation with the PPA Board. Private sector is fairly in good position to facilitate & inform the market

Fully Achieved

No comment.

7.3

S

There are no major systematic constraints (e.g. inadequate access to credit, contracting practices, etc.) inhibiting the private sector’s capacity to access the procurement market

Major problem is about the difficulties in access to credit market. Prime reason is informalities in company’s assets (i.e. informal economy). Another reason is high provision level of non-cash credits (currently 40% of cash credit) that increases costs of letter of guarantees & performance bonds. These are the major issues for some of the small and medium sized companies, but not systematic

Substantially Achieved This is out of PPA’s control, and will be referred to TOBB for study and recommendations to resolve systematic constraints.

264

constraints for the entire private sector.

7.4

S

There are no major constraints that inhibit competition (e.g. technical, labour and other standards)

Public Procurement Law Article 12 Those technical, labour, financial standards stated in Article 10 of PPL are integral part of fair competition.

Fully Achieved

No comment.

7.5 There are clear and transparent rules for determining whether to engage international or national markets, based on a sound development and business logic

Not applicable. Not applicable.

8 Existence of Contract Administration and Dispute Resolution Provisions

8.1

S

Procedures are defined for undertaking contract administration responsibilities that include inspection and acceptance procedures, quality control procedures and methods to review and issue contract amendments in a timely manner

Public Procurement Contract Law Regulation on Inspection and Acceptance The existence of Supreme Science Board (Yüksek Fen Kurulu)

Fully Achieved

Implementation issues exist.

8.2

S

Dispute resolution procedures are included in the contract document providing for an efficient and fair process to resolve disputes arising during the performance of the contract

Public Procurement Contracts Law Draft contract annexes of the implementation regulations International Arbitration Law

Substantially Achieved Although there is clear indication at the end of the draft contract annexed to the implementing regulations about dispute resolution by courts, there is no administrative dispute resolution procedure (e.g. arbitration) for disputes arising during the performance of the contract.

More research is needed especially to determine the coverage of existing procedures. International arbitration is applicable for contracts signed with foreign firms, not with national firms. Administrative dispute resolution procedures are being used in private sector contracts, not in public sector contracts. Change in legislation may be required to facilitate efficient dispute resolution process.

8.3

S

Procedures exist to enforce the outcome of the dispute resolution process

Regulation on Administrative Applications Against Procurements Sample form of contracts

Fully Achieved

Any disputes during contract administration are either referred to local courts or international arbitration in accordance with International Arbitration Law.

9 Effective Control and Audit System

9.1

M

A legal framework, organisation, policy and procedures for internal and external control and audit of public procurement operations exists

Public Procurement Law Article 53 Law on Court of Accounts Public Financial Management and Control Law Articles 63, 64, 68

Substantially Achieved

Subject to proper implementation of PFMC Law with respect to the new control and audit arrangements.

265

and operates to provide a functioning control framework

Constitution Article 160 General Accounting Law

9.2

M

Enforcement and follow-up on findings and recommendations of the control framework provide an environment that fosters compliance

Law on Court of Accounts Article 50, 64 PFMC Law Articles 63, 64

Fully Achieved

See above 9.1

9.3

M

The internal control system provides timely information on compliance to enable management action

Decree Law on Organisation and Mandates of Ministry of Finance, no 178, Article 10, 11, 20, 43

Substantially Achieved When the effective implementation of Public Financial Management and Control Law is taken up, new provisions will be in force on this subject.

See above 9.1

9.4

S

The internal control systems are sufficiently defined to enable performance audits to be conducted

PFMC Law Articles 64, 68 Not Achieved

See above 9.1

9.5

S

Auditors are sufficiently informed about procurement requirements and control systems to conduct quality audits that contribute to compliance

Law on Court of Accounts Article 9 Decree Law on Organisation and Mandates of Ministry of Finance, no 178, Articles 10, 11, 20, 43 PFMC Law Articles 63, 64, 68

Fully Achieved

See above 5.2

10 Efficiency of Appeal Mechanisms

10.1

S

The existence and operation of a complaint review system that gives participants in the public procurement process a right to file a complaint within the framework of an administrative and judicial review procedure

Public Procurement Law Articles 54, 55, 56, 57 Regulation on Administrative Applications Against Procurements

Fully Achieved

Complaints are not resolved at the initial stage (at the contracting entities) and generally referred to PPA, which delays the contract award for one month and increases the workload of PPA. PPA should develop and provide guidance to the contracting entities to address the complaints from participants.

10.2

S

Decisions are deliberated on the basis of available information and the final decision can be reviewed and ruled upon by a body (or authority) with enforcement capacity under the law

Public Procurement Law Articles 54, 55, 56 Regulation on Administrative Applications Against Procurements

Fully Achieved

No comment.

10.3

S

The complaint review system has the capacity to handle lodged complaints efficiently and a means to enforce the remedy imposed

Public Procurement Law Articles 56, 57 Regulation on Administrative Applications Against Procurements

Fully Achieved

No comment.

10.4

The system operates in a fair manner, with outcomes of decisions

Public Procurement Law Articles 53, 54, 55, 56, 57

Fully Achieved

No comment.

266

S balanced and justified on the basis of available information

10.5

S

Decisions are published and made available to the public

Public Procurement Law Article 53 Official Journal Website of PPA

Fully Achieved

No comment.

10.6

S

Administrative review body or authority is separate from the regulatory body, executing agency and the audit/control agency

Public Procurement Law Article 53 Regulation for PPA

Fully Achieved Administrative review body is not separate from the regulatory body, but is separate from executing agency and the audit/control agency

No comment.

11 Degree of Access to Information

11.1

M

Access to information by stakeholders in the process is supported by publication and distribution of information through available media with support from information technology when feasible

A registration number is given for all procurements Stakeholders can access information about procurement process and complaints via internet Website of PPA Publication of the Legislations

Fully Achieved

No comment.

11.2

S

Systems exist to collect key data related to performance of the procurement system and to report regularly

“Procurement Control System” project has come into force. Proceedings on “Real Time” reporting system project has been going on.

Fully Achieved

No comment.

11.3 M

Records are maintained to validate data

All records are kept in digital form. Fully Achieved

No comment.

11.4

S

There is clear legal basis providing access to information to the public

Public Procurement Law Articles 5, 13, 40, 47, 53, 58 and 59 Public Procurement Contracts Law Articles 26 and 27 Secondary legislation 2004 General Communiqué on Public Procurement Daily Public Procurement Bulletin Freedom of Information Act

Fully Achieved

No comment.

12 Ethics and Anti-corruption Measures

12.1

M

The legal and regulatory framework for procurement, including tender and contract documents, includes provisions addressing the issue of corruption, fraud, conflict of interest

Public Procurement Law Articles 10, 11, 17, 58, 59, 60 Public Procurement Contract Law Articles 25, 26

Fully Achieved

No comment.

267

and unethical behaviour and states actions which can be taken with regard to such behaviour (either directly or by reference to other laws)

12.2

M

The legal system defines responsibilities, accountabilities and penalties for individuals and firms involved in fraud or corruption cases

Public Procurement Law Article 58, 59, 60 Public Procurement Contract Law Article 25, 26, 27, 28

Fully Achieved

No comment.

12.3

M

Evidence of enforcement of rulings and penalties exists

Public Procurement Law Article 58, 59, 60 Public Procurement Contract Law Article 25, 26, 27, 28

Fully Achieved

Implementation issues remain.

12.4

S

Special measures exist for the government to prevent and detect potential fraud and corruption in public procurement (e.g. procurement audits)

Public Procurement Law Article 53, 58 Turkish Criminal Code Law on Fight Against Corruption and statement of property

Fully Achieved

Implementation issues remain.

12.5

S

Stakeholders (private sector and civil society) support the criterion of a procurement market known for its integrity and ethical behaviours

Ethical rules of Professional Institutions, Non governmental Organisation, trade associations.

Fully Achieved

No comment.

12.6

S

Existence of a secure mechanism to report fraudulent, corrupt or unethical behaviour

Public Procurement Law Article 11, 53, 58, 60 Turkish Criminal Code Law on Fight Against Corruption and statement of property Law on Ethical Rules in Public Administration

Fully Achieved

Implementation issues remain.

12.7

S

Existence of Codes of Conduct/Codes of Ethics for participants that are involved in aspects of the public financial management systems that also provide for disclosure for those in decision making positions

Public Procurement Law Article 60 Law on Ethical Rules in Public Administration Turkish Criminal Code Law on Civil Service

Fully Achieved

Signature of ethical contract is required for all public employees. Disclosure of financial assets by the public staff is required every 5 years.

(*): S = Standard sub-indicator M = Mandatory sub-indicator