Report No. 78143-GE Georgia Public Expenditure Review Strategic Issues and Reform Agenda Volume 1 June 11, 2014 Poverty Reduction and Economic Management Unit Europe and Central Asia Region Document of the World Bank Standard Disclaimer: This volume is a product of the staff of the International Bank for Reconstruction and Development/ The World Bank. The findings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. Copyright Statement: The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development/ The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, http://www.copyright.com/. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org.   GEORGIA –GOVERNMENT FISCAL YEAR January 1 – December 31 CURRENCY EQUIVALENTS (Exchange Rate Effective as June 11, 2014) Currency Unit = Georgian Lari (GEL) US$1.00 = 1.7557 GEL ABBREVIATIONS AND ACRONYMS ADB Asian Development Bank MOE Ministry of Energy BDD Basic Data and Directions MES Ministry of Education and Science CAD Current Account Deficit MOF Ministry of Finance CIS Commonwealth of Independent States MOLHSA Ministry of Labor, Health, and Social CIT Corporate Income Tax Affairs CPS Country Partnership Strategy MOU Memorandum of Understanding DPO Development Policy Operation MTEF Medium Term Expenditure Framework EBRD European Bank for Reconstruction and NAEC National Assessment and Examination Development Center EC European Commission NBG National Bank of Georgia ECA Europe and Central Asia NDS National Development Strategy E-GP Electronic Government Procurement NPLs Non-Performing Loans EIB European Investment Bank NQF National Qualification Framework EMIS Education Management Information System OECD Organization for Economic Cooperation and ERC Education Resource Center Development ESCO Electricity System Commercial Operator OOP Out-of-pocket Payment ETA Electricity Transmission Agreements PEFA Public Expenditure and Financial EU European Union Accountability FDI Foreign Direct Investment PER Public Expenditure Review GDP Gross Domestic Product PF Partnership Fund GEL Georgian Lari PIM Public Information Management Geostat State Department of Statistics of Georgia PISA Program for International Student GER Gross Enrollment Rates Assessment GR Georgian Railways PIT Personal Income Tax GOGC Georgian Oil and Gas Corporation PPP Public Private Partnership GSE Georgian State Electrosystem ROA Return on Assets HVTL High Voltage Transmission Lines ROE Return on Equity IDA International Development Association SIMS Social Information Management System IMF International Monetary Fund SNG Subnational Government JSC Joint Stock Company SOEs State-Owned Enterprises LEPLs Legal Entities of Public Law SSA Social Services Agency LG Local Government STR Student Teacher Ratio LLC Limited Liability Company TSA Targeted Social Assistance MIP Medical Insurance Plan VAT Value Added Tax MOESD Ministry of Economy and Sustainable VET Vocational Education and Training Development UHC Universal Healthcare System   Vice President: Laura Tuck Country Director: Henry G. Kerali Acting Sector Director: Roumeen Islam Sector Manager: Ivailo V. Izvorski Task Team Leader: Mariam Dolidze ‹  Contents PREFACE…………….. .............................................................................................................................................V EXECUTIVE SUMMARY ...................................................................................................................................... VI CHAPTER 1. MACRO-FISCAL CHALLENGES.................................................................................................1 A. INTRODUCTION ................................................................................................................................................1 B. MACROECONOMIC DEVELOPMENTS AND VULNERABILITIES ............................................................................2 C. FISCAL DEVELOPMENTS AND SUSTAINABILITY ................................................................................................6 D. THE WAY FORWARD ......................................................................................................................................16 CHAPTER 2. EDUCATION EXPENDITURES ..................................................................................................27 A. INTRODUCTION ..............................................................................................................................................27 B. GOVERNANCE STRUCTURE OF THE EDUCATION SYSTEM ...............................................................................28 C. LEVEL AND COMPOSITION OF EDUCATION EXPENDITURES ............................................................................30 D. EDUCATIONAL ATTAINMENT AND PERFORMANCE .........................................................................................36 E. THE WAY FORWARD ......................................................................................................................................42 CHAPTER 3. STATE OWNED ENTERPRISES: GOVERNANCE AND FISCAL PERFORMANCE ........43 A. INTRODUCTION ..............................................................................................................................................43 B. LEGAL AND ORGANIZATIONAL FRAMEWORK OF SOES..................................................................................44 C. OPERATIONAL AND FISCAL IMPLICATIONS OF SOES ......................................................................................46 D. QUASI FISCAL OPERATIONS AND CONTINGENT LIABILITIES ..........................................................................50 E. THE WAY FORWARD ......................................................................................................................................52 CHAPTER 4. INTERGOVERNMENTAL FISCAL RELATIONS ...................................................................54 A. INTRODUCTION ..............................................................................................................................................54 B. INSTITUTIONAL AND ADMINISTRATIVE STRUCTURE ......................................................................................55 C. LEVEL AND COMPOSITION OF SNG EXPENDITURES .......................................................................................56 D. REVENUE SOURCES ........................................................................................................................................60 E. THE WAY FORWARD ......................................................................................................................................66 ‹‹  List of Tables Table 1: Selected Social Programs Introduced or Scaled-Up in 2013-2014.................................................................vi Table 2: Policy Options for Consideration to Manage Fiscal Expenditure Pressures...................................................ix Table 1.1: Georgia: Selected Economic Indicators........................................................................................................3 Table 1.2: Medium-Term Fiscal Framework.................................................................................................................7 Table 2.1: Size of the Education Sector, 2013.............................................................................................................29 Table 2.2: General Government Expenditures on Education.......................................................................................30 Table 2.3: Expenditures, Student Population in the Public System and Unit Costs, 2012 ..........................................31 Table 2.4: Education Expenditures by Level of Education .........................................................................................31 Table 2.5: State Budget Education Expenditures by Major Programs.........................................................................32 Table 2.6: State Budget Allocations to the MES by Economic Categories .................................................................32 Table 2.7: Per Capita Finance Coefficients .................................................................................................................33 Table 2.8: Preschool Indicators by Region, 2012 ........................................................................................................34 Table 2.9: Weekly Teaching Hours .............................................................................................................................35 Table 2.10: PISA Performance by Preschool Attendance ...........................................................................................38 Table 2.11: Number of Schools, Students and Teachers by Class Size, 2012 .............................................................39 Table 2.12: Teacher Pay Scheme.................................................................................................................................41 Table 3.1: Georgia Partnership Fund Portfolio............................................................................................................46 Table 3.2: Financial Information on the SOEs, 2012 ..................................................................................................49 Table 3.3: Market Risk Exposures...............................................................................................................................51 Table 4.1: Variation in Per Capita Expenditures Across Different Functions, 2012 ...................................................59 Table 4.2: Non-Tax Revenues, 2012 ...........................................................................................................................63 Table 4.3. Equalization Impact of Transfers, by Type of LSG, 2012..........................................................................65 List of Figures Figure 1.1: GDP, Exports, and Imports Growth ............................................................................................................2 Figure 1.2: Financing the Current Account Deficit .......................................................................................................4 Figure 1.3: Composition of Total External Debt ...........................................................................................................4 Figure 1.4: Revenues, Expenditures and Fiscal Deficits ...............................................................................................6 Figure 1.5: Composition of Total Public Debt...............................................................................................................8 Figure 1.6: Tax and Non-Tax Revenue, 2003-17 ..........................................................................................................9 Figure 1.7: Composition of Tax Revenues, 2013 ..........................................................................................................9 Figure 1.8: Productivity of Major Taxes .....................................................................................................................10 Figure 1.9: Functional Decomposition, 2006-2014 .....................................................................................................11 Figure 1.10: Economic Decomposition, 2006-14 ........................................................................................................11 Figure 1.11: Public Employment Vs. Size of the Government....................................................................................14 Figure 1.12: Monthly Average Real Wages in Private and Public Sectors, 2005-12 ..................................................14 Figure 1.13: Monthly Average Wages by Sectors in Private and Public Sectors, 2013 ..............................................14 Figure 1.14: State budget capital expenditure, 2013....................................................................................................15 Figure 2.1: Public Spending on Education in Selected ECA Countries ......................................................................30 Figure 2.2: Student Population Attending Public Schools , 2002-30 ..........................................................................35 Figure 2.3: Public Expenditures Per Pupil...................................................................................................................35 Figure 2.4: Net Enrolment Rate in Preschool, 2011 ....................................................................................................37 Figure 2.5: Breakdown of Unit Costs by Region.........................................................................................................37 Figure 2.6: Gross Enrolment Rates by Level of Education .........................................................................................38 Figure 2.7: PISA Reading Score and GDP Per Capita ................................................................................................40 Figure 2.8: Share of Students Scoring Level II or Above and 4 or Above in PISA (2009+).......................................40 Figure 2.9: Trends In Enrolment and Number of Vocational Education Institutions ..................................................41 Figure 3.1: Institutional and Governance Structure of SOEs in Georgia.....................................................................45 Figure 3.2: Georgia General Government Privatization Proceeds, 2001-14................................................................47 Figure 3.3: Sensitivity Analysis of Operating Activities on Profit ..............................................................................49 Figure 4.1: Decentralization of Expenditures and Country Size.................................................................................56 ‹‹‹  Figure 4.2: Subnational Expenditures..........................................................................................................................56 Figure 4.3: Decentralization of Social Services and the Size of Subnational Expenditures ........................................57 Figure 4.4: Economic Composition of LSG Expenditure 2007-11..............................................................................57 Figure 4.5: The Functional Composition of Sub-National Expenditures.....................................................................58 Figure 4.6: Expenditure Disparities in Regions, 2012 .................................................................................................59 Figure 4.7: Economies of Scale in District-Municipality Expenditures ......................................................................60 Figure 4.8: Revenue Structure of Subnational Governments.......................................................................................60 Figure 4.9: Conventional Sources of Local Revenue, 2009 ........................................................................................61 Figure 4.10: Local Government Involvement in Property Taxation............................................................................62 Figure 4.11: Decentralization of Social Services and Transfer Dependence ...............................................................63 List of Boxes Box 3.1: Information on Large Enterprises under the PF ............................................................................................48 Box 3.2:.OECD Principles and Guidelines on Disclosure...........................................................................................52 Box 4.1: International Experience With Local Property Taxes ...................................................................................62 Box 4.2: Developing formula-based grant mechanisms ..............................................................................................64 Spotlight Spotlight 1: Social Protection and Health Expenditures..............................................................................................18 Figure S.1: Social Protection Spending Across ECA countries...................................................................................19 Figure S.2: Outlook of the Social Protection ...............................................................................................................20 Figure S.3: Comparison of Pension Benefit to Average Wage....................................................................................21 Figure S.4: Projected Pension Spending Under Alternative Scenarios........................................................................21 Figure S.5: Minimum pension in Selected Countries ..................................................................................................22 Figure S.6: TSA Coverage and Spending Per Beneficiary ..........................................................................................23 Figure S.7: Structure of Private Insurance Products ....................................................................................................25 Figure S.8: Profitability of Insurance Sector ...............................................................................................................25 Table S.1: Major Social Benefit Programs ..................................................................................................................19 Table S.2: Fiscal Implications of the TSA Program under Three Assumptions ..........................................................24 ‹˜  PREFACE This public expenditure review (PER) summarizes work directed at identifying policy options to manage fiscal consolidation in Georgia. The programmatic fiscal analysis has been conducted by World Bank staff in close coordination with government counterparts, and has used intermediate outputs of ongoing analyses, just-in-time policy notes, and missions as inputs into a continuous dialogue with the government on its public expenditure policy. The programmatic fiscal work has complemented the dialogue on the programmatic Development Policy Operation (DPO) series and the preparation of the Country Partnership Strategy (CPS). This PER synthesis report accompanied by a background report on Diagnostics of Public Investment Management, was prepared by a Bank team comprising Mariam Dolidze (Task Team Leader), Mona Prasad, and Congyan Tan. The primary authors of the background papers are Mariam Dolidze and Congyan Tan (Chapter 1); Daniel Kutner and Nino Kutateladze (Chapter 2); Mediha Agar, Cevdet Denizer, and Archil Mamatelashvili (Chapter 3); Jorge Martinez and Andrey Timofeev (Chapter 4); Simon Groom and Giorgi Abdushelishvili (PIM background report) and Anita Schwarz, Tomas Damerau, Carola Gruen, and Owen Smith (Spotlight 1). The report benefited from comments and suggestions received from Adrian Fozard, Alberto Rodriguez, Andy Mason, Daniel Dulitsky, Dandan Chen, Ahmed Eiweida, Ximena Del Carpio, Lili Liu, Oleksii Balabushko, Jonas Frank, Elin Bergman, Jay-Hyung Kim, Joseph Melitauri, Sandro Nozadze and Galina Alagardova. Tamuna Namicheishvili provided logistics and processing support. Sarah Babirye, Tamuna Namicheishvili and Ewelina Lajch formatted the report. The team has benefited from the guidance of Henry G. Kerali (Country Director), Ivailo V. Izvorski (Sector Manager) and Rashmi Shankar (Lead Economist and Sector Leader). The peer reviewers are Raju Singh, Jay-Hyung Kim, Elene Imnadze and Tobias Linden. The team is grateful to government officials from the Ministry of Finance, Ministry of Labor, Health, and Social Affairs, Ministry of Regional Development and Infrastructure, Ministry of Education and Science, the National Bank of Georgia, Municipal Development Bank, Partnership Fund, State Audit Office, Parliament, Local Authorities of Tbilisi, Rustavi, Mtskheta, Marneuli and other branches of government for their cooperation in conducting the analysis, and for their comments and suggestions on various drafts of the report. ˜  EXECUTIVE SUMMARY 1. Generating growth and creating jobs within a sustainable fiscal framework is Georgia’s biggest macroeconomic challenge. Although Georgia registered rapid growth of 5.7 percent a year during 2010-13, unemployment remains high at 15 percent. New growth companies, especially in tourism and other service sectors, did not generate enough formal – or even informal -- employment. Fiscal policy played a crucial role in Georgia’s recent growth performance with a fiscal stimulus driven post-crisis recovery which increased deficit and debt levels followed by fiscal consolidation during 2010-12 when recovery took hold. The weak execution of the budget in 2013 and policy uncertainty were largely responsible for the growth slowdown during the year. Tackling the growth and jobs agenda in Georgia will require significant investment in human and physical capital and the government has a large role to play here. Additional spending, where it is needed, should be undertaken within the fiscal consolidation agenda of the government, designed to help restore the macroeconomic buffers needed to secure stability and sustain confidence in the future. 2. The change in government in 2012 marked a shift in fiscal policy with prioritization of recurrent social expenditures over capital spending, thereby, increasing budget rigidity. During 2012-13, the government raised the benefit levels under the targeted social assistance (TSA) and pensions and introduced universal health care (UHC) (Table 1). As a result, the fiscal deficit is likely to increase from 2.6 percent of gross domestic product (GDP) in 2013 to 3.7 percent in 2014. Over the medium term, an aging population and the need to improve health outcomes and coverage of the poor in social assistance programs will keep social expenditures high at more than 9 percent of GDP. The share of capital expenditures will level off, meanwhile. Such an outcome will reduce the government’s flexibility in trimming current expenditures in the future. Table 1: Selected Social Programs Introduced or Scaled-Up in 2013-2014 (In GEL million and percent of GDP) 2012 2013 2014 Additional Additional GEL GEL Cost GEL Cost Million Million % of GDP Million % of GDP Total expenditure, state budget 7,668 7,590 -0.3 8,482 3.0 Total selected programs 2,485 3,005 1.9 3,852 2.9 Health care programs 333 435 0.4 606 0.6 Social programs other than pensions 330 490 0.6 644 0.5 Pensions 1,068 1,149 0.3 1,326 0.6 General education 372 470 0.4 439 -0.1 Kindergartens 1 74 80 0.0 125 0.2 Wage bill of the Ministry of Defense 308 381 0.3 428 0.2 Agro credits (interest subsidies) 0 0 - 64 0.2 Vouchers to farmers 0 0 - 50 0.2 Tax subsidies to low-income population 2 0 0 - 170 0.6 Source: MOF and staff calculations.  1 Staff estimates assuming 50 percent additional cost due to cuts of 20 percent parental contributions and an enrollment increase from 48 percent to 66 percent (UNICEF survey, 2013). 2 Estimates of the Ministry of Finance. ˜‹  3. Spending on pensions has been the largest component of social expenditures and ad-hoc increases in pension benefits has put pressure on the government’s fiscal position. In 2012, pensions accounted for 56 percent of total social spending and 4 percent of GDP. From 2012 to 2013, benefit levels were raised from GEL100 per month to GEL150 per month and pension coverage was expanded to all pensioners. Under a scenario where future benefits are maintained at GEL150, a simulation shows that pension expenses will decline to 3.5 percent of GDP by 2016. But if the benefit level increases, pension expenditures are projected to be significantly higher. Under a scenario where future benefits are indexed to wage increases, the pension expense will be 4.3 percent of GDP by 2016 and reach as high as 9 percent of GDP by 2070. These results indicate that the sustainability of pensions will be dependent on the path of future benefit levels. By law, pensions are not indexed – neither to inflation nor to wages -- but in recent years, ad-hoc increases in the pension benefit levels have outstripped wage growth. If future benefit levels continue to increase faster than nominal wages, pension spending could well exceed the regional level. In line with international best practices, the government needs to implement a retirement savings program which complements the public pension system. 4. Universal health care (UHC) is a welcome measure but further refinement is needed to translate new statutory entitlements into effective coverage. Following the elections of October 2012, the government announced that all Georgians would be eligible for state-funded health insurance and “universal coverage” would be provided. In addition to the existing medical insurance program (MIP) targeting the poor and other vulnerable groups of population, the basic universal health service package was introduced in February 2013 and the package was expanded to include a wider range of services in July 2013. The extension of entitlements has been backed up by a 90 percent increase in levels of public funding for health from 2012 to 2013 (from 333 million GEL to 634 million GEL). Overall, public spending on health increased from 1.7 percent of GDP to 2.1 percent in 2013. Preliminary assessment of UHC Program 3 shows some positive initial results. Almost all beneficiaries (96.4 percent) are satisfied or with in-patient and/or emergency out-patient services. Meantime, coverage of out-patient drugs for people with chronic conditions, on which there is currently high out-of-pocket (OOP) spending, remains inadequate. The UHC does not provide drugs at the primary care level for non-emergency cases for formerly uninsured UHC beneficiaries. The new UHC program is being administered by the Social Service Agency (SSA) 4. From April 1, 2014 all MIP beneficiaries were transferred to UHC but maintaining the MIP service package until the expiration of the contract with the beneficiary. The Agency is tasked to perform the function of the single purchaser. This move and increased volume of claims to handle put pressure on the Government to ensure that appropriate controls and incentives for cost- containment and for efficient use of resources are in place. A longer-term development of provider payment methods and further strengthening of capacity at SSA should be considered to replace transitional provider payment methods and ensure efficiencies in purchasing. 5. The government underspent its health budget in 2013 because of uncertainties surrounding methodology and implementation of UHC. The UHC administration is guided by a capitation payment for PHC services, while reference prices generated based on standardized formula, are used for the reimbursement for in-patient care. Meantime, there was an open-ended reimbursement of provider claims, without realistic costing information, negotiated contractual arrangements, or oversight. With a stronger provider and patient awareness of the new program, health expenditures are likely to increase more than what the government is projecting. The government’s medium-term expenditure framework foresees only a marginal increase in the state health budget, from 1.6 percent of GDP in 2013 to 2.1 percent in 2014, as it is based on 2013 performance. However, the actual spending could be higher because the government is still grappling with issues related to the costing of universal healthcare.  3 “ A review of Universal Health Coverage Reforms Introduced In Georgia Since February 2013”-WHO, USAID, WB-May 29, 2014 ͶThe SSA is a Legal Entity of Public Law (LEPL) under the Ministry of Labor, Health and Social Affairs. ˜‹‹  6. In addition, the ad-hoc compensation policies in the public sector have resulted in a high wage bill which stood at 10 percent of GDP in 2012. In 2012, the public sector, including the central administration, local authorities, legal entities of public law (LEPLs) and state owned enterprises (SOEs), employed 54 percent of the formally hired workers and accounted for 20 percent of total employment. The average annual earning per public employee in 2013 was equivalent to about US$4,470 although there are large variations within this group. A public school teacher earns an average annual salary of GEL 6,000 (about US$3,400) while the CEO of the Partnership Fund earned GEL 323,500 (about US$184,000) in 2013. Weak monitoring and inadequate compensation rules have contributed to these wide disparities in public wages. 7. Public finances in Georgia are likely to come under pressure over the short- to medium- term in the context of large increase in recurrent expenditures and the limited scope to raise revenues. Georgia’s tax to GDP ratio at about 25 percent is based on an efficient tax system which ranks superior to many countries in the ECA region. The government has made remarkable progress in the areas of simplification of the taxation system, eradication of border smuggling and enhancement of administrative tools which improved tax compliance. After the 2008 crisis, the tax to GDP ratio stabilized at around 24-25 percent with six types of taxes, low rates and a non-contributory social safety system. A constitutional ban on increasing tax rates limits upsides on fiscal revenues. In July 2011, the parliament had established fiscal rules for a number of fiscal indicators through amendments to the constitution. One of the provisions of this amendment was that the introduction of any new general state tax, except excises, or an increase in the upper rate of any existing general state tax would be possible only through a referendum. This legal clause has been enacted from January 1, 2014. This limitation of tax revenues along with the increase in social benefits is likely to put pressure on the government’s finances. 8. The prospect of reduced capital expenditures going forward has increased the urgency of instituting a sound public investment management system. Capital expenditures were close to 6 percent of GDP in 2013 compared with the annual average of 8.7 percent of GDP during 2007-11. Over the medium-term, capital expenditures are expected to be sustained at the 2013 level. Public capital expenditures have been a significant driver of Georgia’s growth in the past and have a crucial role to play going forward. This is because the country faces a large infrastructure gap in the areas of roads, water and energy. To maximize the effectiveness of the lower level of public investments, the establishment of an effective public investment management (PIM) system will be crucial, especially at project identification and appraisal stage. It is noteworthy that demand for a more robust PIM system has not been strong in Georgia. The likely drivers of change, such as external audit, parliamentarians and civil society have been relatively weak or inactive to pursuing issues related to the quality of public expenditure in general and of public capital investment in particular. 9. To keep government finances on a sustainable path along with sustainable growth and job creation, this public expenditure review offers reform options in the areas of macro-fiscal policy, social protection and health, education, the state owned enterprise sector and inter-governmental fiscal relations. These policy options have a three-fold objective: (i) Increase savings in selected budget items to help build fiscal buffers, (ii) Strengthen the efficiency of expenditures to improve outcomes, especially in the areas of education, capital spending and inter-governmental finances, and (iii) Enhance fiscal management by eliminating off-budget expenditures and improving the availability of data, especially on local governments and state owned enterprises (Table 2). These reforms will help Georgia reduce its fiscal deficit from 3.7 percent of GDP in 2014 to 2.5 percent in 2017. ˜‹‹‹  Table 2: Policy Options for Consideration to Manage Fiscal Expenditure Pressures Policy Area Option for Consideration Sequencing Macro-Fiscal Consolidate public expenditures and increase revenues at the sub-national level, Medium term with a target to reduce the fiscal deficit from 3.7 percent of GDP in 2014 to 2.5 percent in 2017. Establish rule-based compensation policies and regulations for public employees and civil servants to contain increases in salaries and bonuses. Strengthen public investment management and external oversight. Sustain capital expenditures at 6 percent of GDP. Social Protection Limit the medium-term growth of the basic pension benefit to no more than the Medium-term and Health rate of inflation. Medium-term Strengthen the capacity of the Social Service Agency. Introduce pharmaceutical component for UHC beneficiaries to cover outpatient Short-term non-emergency treatment with focus on the lowest quintile of the population. Education Preschool Education: Use the facilities of existing underutilized primary schools to hold classes for preschool education, train primary school teachers to teach preschool children, Short-term and target children of disadvantaged backgrounds and ethnic minorities to enroll in preschools. Improve the quality of service delivery by making the Ministry of Education Medium-term and Science (MES) responsible for setting and monitoring quality standards. General Education: Increase the number of working hours of teachers to bring it closer to OECD Medium-term standards. This would result in a decline in the number of teachers and the savings could be used to improve the salary scale but with improvements in teacher quality. Medium-term Increase the class size. This again would require a reduction in the number of teachers. Medium-term Raise starting salaries of teachers without making the gradient steeper to improve teacher quality. Medium-term Vocational Education and Training (VET): Medium-term Add more general education content to VET programs. Encourage public private partnerships in the financing and delivery of Short-term vocational education. Reduce drop-out rates from VET programs by providing career guidance to help pupils make better informed choices. State Owned Establish an inventory of SOEs to get a consolidated picture of the Short-term Enterprises government’s fiscal position. Short-Term Establish a clear dividend policy for SOEs. Mainstream quasi-fiscal operations like provision of subsidized utility services Medium-term and investment projects undertaken by SOEs into the budget to increase transparency and distributive capacity of the government. ‹š  Policy Area Option for Consideration Sequencing Intergovernmental Improve reporting of subnational finances to increase transparency and fiscal Short-term fiscal relations management of SNGs. Revise the formula for the calculation of the equalization grant to make it simple Medium-term and to better reflect the overall pool of resources and expenditure needs of SNGs. Remove the suspension of taxes on movable property and lower the income threshold for exemptions on immovable property. Medium-term Allow SNGs to share the tax base for certain taxes (such as the personal income tax) and be able to set their own tax rates on top of the national tax rate applied Medium-term to the same tax base. Both taxes (the central and local components) could be administered by the central tax administration. Education Expenditures 10. Poor performance of the Georgian students in international assessments points to the inefficiencies in the education sector. Georgia’s performance on the reading scale of PISA 2009 was below what would be expected given its income level. The gap between Georgia and the OECD countries on the reading scale is approximately three years of schooling (each year accounts for 40 points). In addition, only between 30 percent and 40 percent (depending on the subject) of Georgian 15 year-olds scored Level II or above, which is a threshold usually referred to as functional literacy. 11. Government spending on education in Georgia is low – compared to countries with similar per-capita incomes and relative to both the shortage of human capital and the country’s ambitions. Despite a real increase of nearly 47 percent in education expenditures since 2006, Georgia’s education budget amounted to 2.9 percent of GDP in 2012, below all but a few countries in Europe and Central Asia and lower than comparators at a similar level of income per capita. And even though teacher salaries are one of the lowest among public employees in Georgia, teacher and administrator pay accounts for more than 70 percent of the education sector budget, leaving little space for expenditures directed at curriculum improvements, trainings for teachers, grants for research and development, scholarships for needy students and capital investments to enhance school facilities. 12. The per-capita financing system in the education sector is transparent and leads to an efficient allocation of resources. Schools receive direct transfers of funds from the Ministry of Education and Science (MES) based on the number of students for a given year. These transfers cover salaries, utilities and routine maintenance costs. The allocation formula was adjusted over time to address the problem of underfunded schools which required additional transfers on top of the per capita financing. The system is functioning well currently and is used to determine the education sector budget, including vocational education. 13. An efficient distribution of resources cannot offset substantial sector inefficiencies, such as poor teacher quality, and excess number of teachers from the pre-school through the secondary level. Teaching is not a sought after occupation in Georgia because of the very low pay scale and this has impacted the quality of teachers. At 8.7, the student teacher ratio of preschools in Georgia is considerably lower than the OECD or EU 21 countries (at 12.2 and 11.8, respectively). As a consequence, several regions in Georgia run very high costs per student. With an aging population, there has been a decline in the number of students and a consolidation in the number of schools in Georgia. However, at 241 students per school, the average school size remains small and drives up costs. There are too many teachers teaching too few hours suggesting that there is a glut in the education system. Based on an academic year of 40 weeks, the average yearly teaching hours sum up to less than 600 while the average teacher in an š  OECD country teaches 790 hours in primary, 709 hours in lower-secondary, and 664 hours in the upper- secondary level. In countries such as the US and Chile, these figures go beyond 1000 hours per year. 14. The vocational education and training program has been completely overhauled but only a fourth of the VET graduates are able to find jobs in their mainstream education disciplines. The VET system underwent significant changes with the adoption of the Law on Vocational Education in 2007. The VET strategy which was developed resulted in a consolidation and subsequent expansion of VET institutions, introduction of a voucher system for VET institutions and significant public investments in the VET sector. The government has also eased the admission requirements for certain VET programs to make them more accessible. However, the system is plagued with high drop-out rates which lead to a wastage of resources. In addition, surveys indicate that only one-fourth of the graduates from VET institutions are able to find a job in the area in which they have educational qualifications. Therefore, there is room to make substantial improvements in the curriculum and the quality of teachers. State Owned Enterprises 15. According to business survey statistics, the state owned enterprise sector in Georgia is relatively small and its contribution to GDP amounted to 7 percent in 2012 and accounted for about 10 percent of total formal employment. The privatization program that was underway during 2004-10 resulted in a significant reduction in public enterprise employment from 144,000 in 2004 to 60,000 in 2012. Although the share of SOEs is small in terms of output, their contribution to investment in fixed capital is oversized at 24 percent of total corporate investment in 2012. This is essentially because of the sectors in which the large SOEs operate–energy and transport–and the large infrastructure requirements in these sectors. Some of the government’s capital expenditures in these areas are undertaken by the large SOEs. Majority of the remaining SOEs operate in the health sector, municipal services, agriculture, tourism and manufacturing. 16. No single government entity has a comprehensive picture of SOE performance. More than 400 SOEs operate in Georgia under different institutional arrangements and ownership structures. However, with the exception of the five largest SOEs, information on them is scant. According to the law on state property in Georgia, the MOESD is primarily responsible for the management and supervision of SOEs but in practice, this oversight function is spread across several ministries. As a result, Parliament does not receive any information on these companies as part of the budget preparation and execution process. Transfers from SOEs to the budget through dividends varied widely over the years based on performance of the SOEs and their negotiations with the government. While there are no explicit subsidies to the SOEs, the government has assumed responsibility for some of their liabilities. 17. The five largest SOEs in Georgia comprise 80-90 percent of the total assets of SOEs and are grouped under the extra budgetary Partnership Fund. Their combined profit amounted to GEL 63.5 million in 2012 or about 0.2 percent of GDP – and four out of the five were profitable. Based on negotiations with the government, these companies paid dividends of GEL 6.7 million in 2012 to the Fund. Their consolidated financial data also showed a positive return on assets (ROA) and return on equity (ROE) in 2012. The only exception was the Georgian State Electrosystem (GSE) which was not profitable. Sensitivity analysis impacting the operating income and other costs and income of the SOEs reaffirms the robustness of their operations. 18. The government undertakes significant quasi-fiscal operations through these large SOEs which potentially understate the fiscal deficit and overall debt. Exact costs are not available but the government provides subsidized utilities to the population (gas, electricity, transport) through its 5 largest SOEs for which it does not compensate them. In addition, these SOEs undertake investment projects on behalf of the government, also without any compensation. The government also imposes other costs on š‹  SOEs like wage increases. Revenues and costs related to carrying government mandates need to be mainstreamed into the budget and this will potentially increase the fiscal deficit. Additional costs could surface from SOEs for which data is not currently available, especially those under the (MOESD), several of which are bankrupt. Some of the largest SOEs are highly indebted and assumption of SOE debt would also increase government debt. Intergovernmental Fiscal Relations 19. The current administrative structure in Georgia, with just one layer of sub-national government, is in line with international practices. With a territory of 69,700 square kilometers and a population of around 4.4 million, Georgia’s size warrants one subnational tier of government or at most two, if the autonomous district of Adjara is included. The current structure avoids duplication of functions and is in line with the division of responsibilities between the center and the sub-national governments (SNGs), with several public services being provided directly by the center. The population covered by the local governments in Georgia varies widely and the median population of 35,000 is much higher than the median population of 19,000 for the bottom tier jurisdictions in most European countries. This helps Georgia achieve economies of scale in the provision of local services. 20. Despite limited expenditure responsibilities, SNG spending in Georgia is reaching levels similar to those in countries where subnational governments are responsible for more resource intensive services. SNGs in Georgia are primarily responsible for the provision of housing, utilities and communal services while the more resource intensive services like education and health are with the central government. SNG expenditures are at 7 percent of GDP, and have accounted for about a fifth of consolidated government expenditures. They have been increasing steadily since 2008, both in real terms and relative to GDP. There is also significant inter-regional disparity in spending in Georgia, primarily attributable to lumpy capital expenditures. Another crucial factor affecting expenditures across regions is the size of the population. 21. The low level of revenue autonomy at the local level acts as a disincentive for effective fiscal management. Intergovernmental fiscal transfers have accounted for nearly 70 percent of total revenues of SNGs. Since 2008, the only own sources of subnational revenues have been property taxes and non-tax revenues, such as rents, fines and penalties, sale of goods and services and other local government collections. While local governments can determine property tax rates, local accountability is limited by the fact that local governments are passive recipients of revenues collected by the national tax authorities. Therefore, there is limited room to hold SNGs accountable for revenue performance. Fiscal transfers from the center help close the gap between expenditure needs and revenues of local governments. While there is a complicated formula for the calculation of equalization grants to SNGs, it lacks transparency and has proved difficult to replicate. In addition, the center also doles out unclassified special purpose transfers, which are largely residual in nature and are mostly used for capital expenditures. Therefore, there is no incentive for local governments to reduce expenditures as it would result in lower transfers from the center. š‹‹  CHAPTER 1. MACRO-FISCAL CHALLENGES A. INTRODUCTION 1.1. Generating growth and creating jobs within a sustained fiscal framework is Georgia’s biggest macroeconomic challenge. Although Georgia’s economy registered a rapid growth of 5.7 percent per annum during 2010-13, unemployment remained high at 15 percent. Translating the high growth to strong net employment remains a challenge, as older industries continue to decline, shedding jobs, while new industries have yet to create enough jobs to absorb the unemployed. The new growth sectors such as tourism and other services have not been able to generate formal employment as effectively as a robust tradable sector would have. On the supply side, businesses complain that job seekers do not have adequate skills, in large part because the existing education curriculum is not well aligned with the demand of the private sector. 1.2. Addressing the jobs and growth challenge will require policy actions to address the education and infrastructure gap while continuing fiscal consolidation. Improved roads, better water supply and a reliable power sector are crucial for facilitating growth in the tradable sector, and there is a substantial role that the government can play in this regard. The skills mismatch in the labor market and the country’s relatively low education budget also call for more spending on education at all levels starting from pre-school 5 Social spending will protect those who do not benefit from growth However, all these expenditures need to be made in accordance with the fiscal consolidation agenda of the government, designed to help restore the macroeconomic buffers needed to secure stability and growth in the future. High fiscal and current account deficits will pose a threat to Georgia’s fiscal sustainability. The fiscal deficit rose after the 2008 global financial crisis and conflict with Russia, reflecting a prompt and adequate fiscal stimulus to the economy. The current account deficit remained over 10 percent of GDP after the crisis, in line with the rise in the fiscal deficit and fast import growth fueled by lending expansions. With fiscal consolidation a top priority, Georgia’s fiscal and current account deficit were reduced to 2.5 percent and 5.9 percent respectively by the end of 2013. Due to the government’s continued commitment to fiscal consolidation, public finances need to be efficiently managed to maximize its impact on growth and job creation. 1.3. The recent increase in recurrent expenditures poses the challenge of a more rigid budget and less fiscal space in the future and this makes the need for policy actions particularly pressing. The new government which came to power in late 2012 prioritized social objectives and enacted a set of new policies to improve social benefits. During 2012-13, the government raised the benefit levels under the TSA and pensions, and introduced universal health care (Spotlight 1). The immediate effect of these actions will be reflected in a deterioration of the fiscal deficit to 3.7 percent of GDP in 2014. Over the medium term, an aging population and the need to improve health outcomes and coverage of the poor in social assistance programs will keep social expenditures high. Meanwhile, capital expenditures need to be maintained at current levels to ensure growth and address the infrastructure gaps. These increased benefits posed considerable risks to Georgia’s fiscal sustainability. On the revenue side, tax collections remain buoyant thanks to improved tax administration but there is a constitutional provision which bars the government from raising tax rates. Therefore, Georgia faces a situation of increasing recurrent expenditures which are difficult to curtail and limited upsides to revenue collections. In such a scenario, further fiscal consolidation to create fiscal buffers and efficient management of public finances will be the key for the country to sustain medium- to long-term economic growth and job creation.  ͷEvidence shows that estimated benefit-to-cost ratio of increasing enrolment in preschool from 25 percent to 50 percent in low- and middle-income countries range from 6.4 to 17.6. 1  1.4. In order to address the macro-fiscal challenges facing the country and achieve fiscal consolidation targets, this chapter proposes the following options for consideration: x Establish rule-based compensation policies and regulations for public employees and civil servants to contain increases in salaries and bonuses. x Strengthen public investment management and external oversight. x Sustain capital expenditures at 6 percent of GDP. 1.5. The rest of the chapter is organized as follows. Section B provides an overview of the macroeconomic developments in the recent past and highlights the main vulnerabilities that the country faces. The next section focuses on fiscal developments and explains the main sources of revenues, expenditure composition and efficiencies in spending. Section D gives the way forward. B. MACROECONOMIC DEVELOPMENTS AND VULNERABILITIES 1.6. The government used counter-cyclical fiscal policy to spur growth in the aftermath of the global financial crisis and conflict with Russia in 2008. The twin shocks led to a significant deterioration of economic performance: FDI inflows collapsed in 2009 ending up with US$658 million, less than 40 percent of the 2007 amount. Total investments plummeted by 57 percent during 2007-09. Goods exports also fell by about 9 percent during the period. With the levels of FDI, investments and exports shrinking, economic growth reduced to 2.3 percent in 2008 and output contracted by 3.8 percent in 2009—a sharp disruption to a stellar growth streak which averaged 9 percent per annum in previous years. 1.7. Aided by the fiscal stimulus, the economy rebounded strongly during 2010-12. To mitigate the downturn and restore confidence during the crises, the authorities responded with a fiscal stimulus which was financed through a large international crisis assistance package. As a result, public expenditures increased from 34 percent of GDP in 2007 to 38.4 percent in 2009, primarily driven by public investments. With high public spending Figure 1.1: GDP, Exports, and Imports Growth bolstering the recovery, private investment, (In percent) exports, tourism, and bank lending picked up during 2009-12. In particular, exports of goods and services improved to 36.5 percent of GDP in 2011 from 29.8 percent within two years, with merchandise exports up 39 percent and the number of tourist visits 37 percent higher per year. Total investment more than doubled in 2009-12 (Figure 1.1). Source: Georgian authorities and staff estimates. 2  1.8. Economic growth weakened again in 2013 due to political uncertainty. Georgia grew by 3.2 percent in 2013, only half of the average growth rate in 2010-12. As the Presidential elections were held in October 2013 and a new Prime Minister was appointed in November, election-induced policy uncertainty remained high for most of the year. This led to a 14.5 percent contraction in investments, particularly those financed by the government. In contrast, private consumption showed remarkable resilience in 2013 and increased by one percent in real terms, supported by falling inflation, growing wages, and continued expansion in consumer credit (Table 1.1). Table 1.1: Georgia: Selected Economic Indicators (In percent of GDP, unless otherwise indicated) 2010 2011 2012 2013 2014p 2015p 2016p 2017p Actuals Projections Real GDP growth (percent change) 6.3 7.2 6.2 3.2 5.0 5.0 5.0 5.0 Consumer Price Index (percent change) 7.1 8.5 -0.9 -0.5 3.5 5.0 5.0 5.0 GDP per capita (in U.S. dollars) 2,614 3,220 3,515 3,597 3,716 4,031 4,397 4,844 Gross investment 21.6 26.2 28.9 24.8 25.5 26.4 26.9 27.4 Public 7.4 8.0 6.8 5.2 5.9 6.0 6.1 6.1 Private 14.2 18.2 22.1 19.7 19.6 20.4 20.8 21.3 Gross national savings 11.3 13.4 17.3 18.9 17.2 18.5 20.0 21.1 General Government Operations Revenues and grants 28.3 28.2 28.9 27.6 27.1 27.3 27.3 27.2 Tax revenues 23.5 25.2 25.1 24.8 24.7 25.0 25.1 25.0 Non-tax revenues 4.8 3.0 3.7 2.8 2.4 2.3 2.2 2.2 Expenditure and net lending 34.8 31.7 31.7 30.1 30.8 30.3 30.0 29.7 Current expenditure 26.4 23.0 24.8 24.9 24.9 24.3 23.9 23.6 Social expenses 7.8 6.8 6.7 8.6 9.6 9.6 9.5 9.3 Capital expenditure and net lending 8.4 8.8 6.8 5.2 5.9 6.0 6.1 6.1 Overall fiscal balance -6.5 -3.5 -2.8 -2.5 -3.7 -3.0 -2.7 -2.5 External Sector Current account balance -10.3 -12.7 -11.7 -5.9 -8.4 -7.9 -6.9 -6.3 Exports of goods and services 34.9 36.5 38.1 44.7 45.3 46.4 47.1 47.9 Imports of goods and services 52.7 55.5 57.8 57.7 59.2 59.4 59.8 60.0 FDI (net) 5.8 6.2 3.9 5.6 6.4 6.5 6.5 6.3 Gross international reserves (Months imports of goods and services) 4.4 4.2 3.7 3.3 3.0 3.1 3.4 3.5 (In millions of dollars) 2,264 2,818 2,873 2,823 2,914 3,362 3,739 4,183 Total public debt 38.7 33.6 32.3 32.2 34.1 34.2 34.0 33.7 Unemployment rate (percent) 16.3 15.1 15.0 14.6 - - - - Source: Georgian authorities; Bank and Fund staff estimates and projections. 1.9. The current account balance improved in 2013 for the first time since 2009, reflecting subdued domestic demand and a pick-up in exports to CIS countries. The current account deficit narrowed to 5.9 percent of GDP in 2013 largely on account of subdued demand for imports related to both consumption and investments, especially those financed by the government. With the removal of the Russian ban on Georgian products, exports to Russia more than quadrupled and its share increased to 7 percent in 2013 from 2 percent a year ago. Supported by a significant increase to CIS countries, Georgia’s exports grew by 17 percent in 2013. Imports, on the other hand, declined by 3 percent, due to low investments and weak budget execution. Robust growth in service exports, mainly from transport and 3  tourism also helped to strengthen its external position. However, with economic rebound in sight, imports will pick up and the current account deficit is expected to widen to about 8 percent by 2014. 1.10. However Georgia remains vulnerable to external shocks and regional uncertainties could dampen the country’s export performance and reduce remittances. CIS countries accounted for nearly 56 percent of Georgia’s goods exports in 2013 with Azerbaijan at 24 percent, Armenia 11 percent, Russia 6.5 percent and Ukraine 6.6 percent. In addition, Turkey accounted for another 6 percent of Georgia’s exports. Although a lower export growth has been assumed for 2014, a protracted crisis in some of these countries could further reduce Georgia’s exports through lower demand and loss of price competitiveness (since the Russian ruble and Turkish lira are depreciating). The majority of tourist arrivals in Georgia in 2013 were from Turkey (30 percent of the total), Armenia (24 percent), Azerbaijan (20 percent), Russia (14 percent) and Ukraine (2 percent). Hence, there is a downside risk for lower tourism receipts as well, which accounted for 58 percent of Georgia’s service exports in 2013. Nearly 55 percent of private money transfers to Georgia in 2013 originated in Russia, most of which were remittances. However, remittances are small in Georgia and accounted for only 5 percent of GDP in 2013. The adverse impact on exports could be cushioned to some extent with increased exports to EU countries (assuming that EU does not get impacted by these developments) after the DCFTA is in place. In addition, for oil and gas imports, Georgia is largely dependent on Azerbaijan which makes it less vulnerable to regional uncertainties. Figure 1.2: Financing the Current Account Deficit 1.11. Financing of the current account deficit (In percent of GDP) in 2013 relied largely on FDI but the stock of external debt remains high. FDI inflows have provided a stable source of financing for about half of the current account deficit till 2012. With the current account deficit at 5.9 percent and net FDI at 5.6 percent of GDP, Georgia’s reliance on external loans was low in 2013 (Figure 1.2). Meanwhile additional borrowing from the private sector helped the government repay its Eurobonds and IMF credit in 2013. Therefore total external debt improved marginally by 0.5 percentage points to 84.6 percent of GDP (Figure 1.3). However with external debt averaging at more than 80 percent of GDP a year over the medium-term, Georgia faces refinancing Source: NBG. and foreign exchange risks. Figure 1.3: Composition of Total External Debt (In percent of GDP) 4  1.12. Inflation was low in 2013 but it picked up in 2014 and the National Bank of Georgia raised interest rates. During 2013, with headline inflation significantly below the target of 6 percent, the National Bank of Georgia (NBG) cut its policy rate by 150 bps to 3.75 percent. The economy experienced disinflation in 2013, with a 0.5 percent decline in the average price level. The de facto stabilized arrangement against the dollar over the past couple of years has contributed to low inflation. In addition, a reduction in administered prices on Source: MOF. energy and utilities, and a decline in the cost of health services also reduced overall prices. In the first quarter of 2014, consumer price rose by 3.3 percent year on year, driven by increases in food price. In response, the NBG raised interest rate by 25 bps to 4 percent in February 2014 over concerns of mounting inflationary pressure. The exchange rate is broadly in line with fundamentals (0.3 percent overvalued as per the macro balance approach). Driven by lower inflation, the real effective exchange rate depreciated by 1.5 percent during 2013 but it continued to rise in 2014. Monetary policy is geared towards inflation targeting although in the context of a highly dollarized economy and weak capital markets, there is limited transmission of interest rate changes. The exchange rate is broadly in line with fundamentals (0.3 percent overvalued as per the macro balance approach). Driven by lower inflation, the real effective exchange rate depreciated by 1.5 percent during 2013 but it continued to rise in 2014. Monetary policy is geared towards inflation targeting although in the context of a highly dollarized economy and weak capital markets, there is limited transmission of interest rate changes. 1.13. Credit growth strengthened during 2013 and the financial sector showed signs of recovery. The reduction in policy rates helped reduce the lending rates of the commercial banks, increasing credit growth to 26 percent in 2013 from 10.5 percent in 2012. The introduction of an agricultural lending interest rate subsidy program though the Agricultural Development Fund helped stimulate lending to agricultural producers while the other main borrowers were from the trading sector and hotels and restaurants. The ratio of non-performing loans (NPLs) to total loans declined from 9.6 percent at the beginning of 2013 to 7.5 percent by the end of the year. This is a significant improvement from an NPL ratio of 17.9 percent at the end of 2009. The banks’ return on assets rose to 2.6 percent in December 2013 compared with 1 percent a year earlier. The banks are adequately capitalized with a capital adequacy ratio of more than 17 percent at the end of 2013. However, the large share of loans in foreign currency (about 60 percent of the total) gives rise to foreign exchange risk. 1.14. A decade of strong economic growth in Georgia has not succeeded in making a significant dent on the unemployment rate, although, poverty has fallen in recent years. The government has been successful in attracting foreign investors to the country and the economy has experienced sustained growth with the exception of 2008-09. Net job creation, however, remained weak and by 2013 the overall unemployment rate remained at about 15 percent, one of the highest in the ECA region. The labor force participation ratio remained low at 66 percent in 2013 and is assumed to decrease further due to the ageing population. In contrast, the poverty rate, according to the national poverty line, fell from 21 percent in 2010 to 14.8 percent in 2012 and the mean consumption of the bottom 40 grew at 5.3 percent annually, exceeding the growth enjoyed by the population overall (3.6 percent). This was mainly attributable to government transfers, food disinflation and increased earnings. Georgia did not register any significant changes in poverty between 2006 and 2010. During these years, poverty stayed between 17 percent and 21 percent. Inequality in Georgia is higher than in the ECA region on average with a Gini coefficient of 40.7 in 2011. 5  C. FISCAL DEVELOPMENTS AND SUSTAINABILITY 1.15. In the post-crises period of 2010-12, the government was committed to fiscal consolidation to restore the much needed macroeconomic buffers which were used up during the crisis. The stimulus provided by the government along with lower revenues due to economic contraction widened the fiscal deficit to 9.2 percent in 2009. As the economy rebounded in 2010-12 and grew by an average of 6.5 percent a year, fiscal consolidation was undertaken by reducing current expenditures. Capital expenditures were maintained at 8.5-9 percent of GDP since private investments and FDI remained weak. These efforts resulted in a deficit reduction of 6.4 percentage points of GDP to 2.8 percent in 2012. External public debt declined to 27 percent of GDP in 2012 after peaking at 33.6 percent in 2010. 1.16. The fiscal consolidation process continued into 2013, on the back of lower capital expenditures. The fiscal deficit narrowed to 2.5 percent of GDP in 2013 and could have been lower if the planned revenues had materialized (Figure 1.4). Revenue collections fell short of expectations due to subdued business and state activities, and were lower than the budgeted amount by 2 percent of GDP. The government suspended a number of infrastructure projects which resulted in under-execution of capital projects by 16.5 percent. The government plans to increase capital expenditures and social spending in 2014 which will together account for nearly 1.5 percent of GDP. However, revenue mobilization is not likely to keep pace with increased expenditures. The government is committed to narrowing the deficit from 3.7 percent of GDP in 2014 to under 2.5 percent in three years but the details have not been fleshed out. Figure 1.4: Revenues, Expenditures and Fiscal Deficits 1.17. The low central government fiscal (In percent of GDP) deficit however provides only a partial view of the overall fiscal position. A review of the fiscal risks including implicit contingent liabilities arising from the operations of the five largest SOEs under the partnership fund (PF) reveals that they undertake quasi-fiscal activities which should ideally be routed through the budget. Since they are consolidated under the PF, they do not contribute to the budget through dividends but through off-budget expenditures. Four of the five largest SOEs are run profitably. 6 1.18. Public finances in Georgia are likely to come under pressure over the short- to Source: MOF. medium-term in the context of the large increase in social expenditures and limited scope to raise revenues. The change in government in 2012 marked a shift in fiscal policy with an increased focus on social spending. Unlike the previous government which focused on capital expenditures, the new government prioritized expenditures in the areas of pension, health, and targeted social programs. These expenditures were scaled-up and other indirect subsidies like vouchers for the small farmers and personal income tax (PIT) refunds for low-wage earners were introduced. From February 2013, all citizens who were not enrolled in the targeted medical insurance program (MIP) were eligible to enroll in the state funded non-contributory Universal Health Care (UHC) program. Other expenditure increases included higher pension benefits and increased (doubled) social assistance benefits from mid-2013. Vouchers for small farmers worth GEL50 million and refund of income tax to low-  6 Georgia: Partnership Fund, IMF, Fiscal Affairs Department, July 2012, Georgia: Institutional Framework for Analyzing and Managing Fiscal Risks, IMF, Fiscal Affairs Department, September 2013. 6  income taxpayers of around GEL 170 million during 2014 will also add to the government’s overall spending. These increases amounted to 1.6 percent of GDP in 2013 and are likely to increase to 2.1 percent in 2014. In 2014, due to lower revenues and higher expenditure commitments, the government is likely to run a larger fiscal deficit of 3.7 percent of GDP. To return to the consolidation path the government needs to undertake both expenditure restraint and increases in revenues. Over the medium-term public expenditures will need to be cut by 1.1 percent of GDP (Table 1.2). The constitutional ban on raising tax rates limits the upsides on revenues. Table 1.2: Medium-Term Fiscal Framework 2006 2007 2008 2009 2010 2011 2012 2013 2014p 2015p 2016p 2017p Percent of GDP Revenues and Grants 26.7 29.2 30.7 29.3 28.3 28.2 28.9 27.5 27.1 27.3 27.3 27.2 Tax revenues 22.9 25.8 24.9 24.4 23.5 25.2 25.5 24.7 24.7 25.0 25.1 25.0 grants 1.2 0.6 3.2 2.2 2.3 0.9 1.0 0.8 0.7 0.7 0.6 0.6 Other revenues 2.6 2.8 2.5 2.7 2.5 2.1 2.4 2.0 1.7 1.6 1.6 1.6 Total Expenditure 29.7 33.9 37.1 38.5 34.8 31.7 31.7 30.0 30.8 30.3 30.0 29.7 Current Expenditure 20.7 25.0 28.3 30.1 26.4 23.0 23.2 24.0 24.8 24.4 23.9 23.6 Compensation of employees 4.1 4.0 5.3 5.8 5.4 4.7 4.6 5.2 5.1 5.0 4.9 4.8 Purchases of goods and services 5.6 9.3 8.5 6.1 5.5 5.0 5.0 3.8 3.7 3.5 3.6 3.5 Interests 0.8 0.6 0.6 1.0 1.0 1.2 1.0 0.9 1.1 1.0 0.8 1.0 Subsidies 2.8 2.3 2.7 2.3 1.8 1.8 2.0 2.0 1.8 1.8 1.7 1.5 Grants 0.0 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 Social benefits 5.1 5.0 7.2 8.4 7.8 6.8 7.1 8.5 9.6 9.6 9.5 9.4 Pension 3.4 3.0 4.1 4.8 4.5 4.1 4.1 4.3 4.2 4.1 4.0 3.8 Social Assistance 0.4 0.4 1.5 1.5 1.2 1.0 1.3 2.3 3.3 3.0 2.6 2.7 Health Expenditure 1.3 1.5 1.6 2.0 2.2 1.6 1.7 1.9 2.1 2.5 2.9 3.0 Other expense 2.3 3.7 3.9 5.2 4.8 3.5 3.6 3.6 3.5 3.4 3.3 3.2 Education 3.0 2.7 2.9 3.2 2.9 2.8 2.9 3.0 3.0 3.0 3.0 3.1 Capital Expenditure 9.0 8.9 8.8 8.4 8.4 8.8 8.4 6.0 6.0 6.0 6.1 6.1 Capital 7.8 8.6 8.0 8.0 7.4 7.7 7.3 5.1 5.5 5.8 5.8 5.9 Net Lending 1.2 0.3 0.8 0.4 1.0 1.1 1.1 0.9 0.5 0.2 0.3 0.2 Overall fiscal deficit -3.0 -4.7 -6.4 -9.2 -6.5 -3.5 -2.8 -2.5 -3.7 -3.0 -2.7 -2.5 Real 2008 Lari, million Revenues and Grants 4,431 5,444 5,854 5,372 5,515 5,902 6,412 6,293 6,544 6,955 7,372 7,786 Tax revenues 3,800 4,810 4,753 4,479 4,576 5,267 5,658 5,661 5,964 6,369 6,778 7,156 grants 199 112 617 397 444 192 230 176 169 178 162 172 Other revenues 431 522 484 497 495 443 524 456 410 408 432 458 Total Expenditure 4,929 6,320 7,081 7,067 6,790 6,635 7,029 6,866 7,429 7,726 8,092 8,488 Current Expenditure 3,435 4,661 5,396 5,525 5,155 4,802 5,155 5,500 5,980 6,210 6,445 6,751 Compensation of employees 680 746 1,008 1,070 1,053 976 1,020 1,186 1,230 1,274 1,323 1,383 Purchases of goods and services 929 1,734 1,614 1,128 1,070 1,040 1,101 859 886 892 972 1,010 Interests 133 112 121 175 194 247 215 202 254 255 216 286 Subsidies 465 429 512 429 357 366 436 466 445 459 459 429 Grants - 19 12 15 12 11 14 13 13 19 18 36 Social benefits 846 932 1,379 1,537 1,526 1,421 1,575 1,951 2,307 2,446 2,565 2,691 Other expense 382 690 750 963 941 741 794 823 845 866 891 916 Capital Expenditure 1,494 1,659 1,685 1,542 1,635 1,833 1,874 1,367 1,449 1,516 1,647 1,737 Capital 1,294 1,603 1,524 1,468 1,448 1,605 1,625 1,168 1,328 1,478 1,566 1,689 Net Lending 199 56 161 73 187 228 249 199 121 39 81 48 Overall fiscal deficit -498 -876 -1,227 -1,694 -1,275 -733 -617 -573 -885 -771 -720 -703 Source: Georgian authorities; Bank staff estimates and projections. Note: p=projected 7  1.19. Georgia’s public sector debt is much lower than other ECA countries and is dominated by bilateral and multilateral concessional debt. Public debt at 34.5 percent of GDP in 2013 is dominated by external debt. The government’s debt level declined during 2010-13 due to fiscal consolidation policies (Figure 1.5). Most of the public external debt stock has been accumulated during 2008-10 with a substantial portion of it maturing between 2012 and 2017. This implies that during this period the government’s financing requirements will be much higher with the need to refinance or repay. Due to the low fiscal deficit in 2013, the overall financing needs were lower and the government was able to repay the IMF loan of approximately US$140 million as well as a part of its Eurobonds of approximately US$65 million (total of 1.3 percent of GDP). At the end of 2013, the government’s external debt reduced by US$155 million, around 1 percent of GDP, and stood at 27 percent of GDP – a record low. However, domestic borrowing increased by 9.4 percent mostly via issuance of treasury bills. The government has no state guaranteed loans outstanding, although debt issued by two of the state enterprises in recent years (around 4 percent of GDP) could count as an implicit contingent liability. Figure 1.5: Composition of Total Public Debt (In percent of GDP) Source: MOF and Staff estimates. Government Revenues 1.20. Georgia’s tax to GDP ratio at about 25 percent is based on an efficient tax system which ranks superior to most of the countries in the ECA region. Tax collections increased dramatically between 2003 and 2007 from 14.5 percent of GDP to 25.8 percent (Figure 1.6). This was achieved through substantial simplification of the taxation system, eradication of border smuggling and enhancement of administrative tools which improved tax compliance. After the 2008 crisis, the tax to GDP ratio stabilized at around 24-25 percent with 6 types of taxes, low rates and a non-contributory social safety system (Figure 1.7). In a cohort of 20 central and eastern European countries, only Montenegro and Croatia exceed Georgia’s performance on valued-added tax (VAT) productivity, only Montenegro is ahead in corporate income tax (CIT) productivity and only Hungary is better than Georgia on personal income tax (PIT) productivity. 7 This has led to a relatively low tax burden on enterprises and has contributed to the improvement of the business environment.  7 Productivity is measured as the total VAT (PIT, CIT) revenues divided by the highest VAT (PIT,CIT) nominal rate, IMF Country Report No. 11/93, April 2011 8  1.21. However, the personal income tax productivity is expected to deteriorate in 2014 because of the introduction of tax exemptions to low-wage earners in 2013. The government enacted an amendment to the tax code to exempt low-income (below GEL500 per month) workers from personal income tax applied to the initial GEL300. The maximum amount of GEL360 is being returned to each worker’s bank accounts in 2014. The government’s initiative on tax-free minimum wages is likely to cost about 0.6 percent of GDP in 2014. The large number of beneficiaries, about 80 percent of formal employees, suggested that the eligibility threshold of GEL500 per month needs to be decreased. Tax breaks in Georgia are limited and are estimated at 1 percent, 3 percent and 17 percent of tax potential for PIT, CIT, and VAT respectively (IMF, 2011). Except for the above mentioned PIT initiative, no major tax exemption was introduced since 2010 (Figure 1.8). Figure 1.6: Tax and Non-Tax Revenue, 2003-17 Figure 1.7: Composition of Tax Revenues, 2013 (In GEL million) (In percent of total) Source: MOF. 1.22. A constitutional ban on increasing tax rates limits upsides on state tax revenues. In July 2011, the parliament had established fiscal rules for a number of fiscal indicators through amendments to the tax code. One of the provisions of this amendment was that the introduction of any new general state tax, except excises, or an increase in the upper rate of any existing general state tax would be possible only through a referendum. This legal clause has been enacted from January 1, 2014. Local taxes are however not subject to this law, and this provides some leverage at the municipal level to raise revenues. 9  Figure 1.8: Productivity of Major Taxes VAT ,:jlEO .... •:::. 16 19 20 22 H~heet ncrn1na1 rate Of I8X CIT •GEO ·- ...~A~ 10 ·II:Jll"- 20 ao Hlg!leot nomna1rate or faX 40 50 PIT 0 10 15 20 25 Hlg'le6t nomna l ral8 Of I8X Source:  ‘—–”›‡’‘”–‘ǤͳͳȀͻ͵ǡ’”‹ŽʹͲͳͳ 10  1.23. While fiscal consolidation over the medium-term will mostly depend on expenditure consolidation, the improvements in taxation at the local level could contribute to about 1 percent of GDP in revenues. The revenue performance can be improved by fully utilizing the ‘tax handles’ that have been traditionally given to the local authorities throughout the world: user fees, property taxes and betterment levies. The current suspension of taxes on movable property in Georgia is not justifiable. In the same vein, the current over-generous exemptions on immovable properties need to be reconsidered. The experience of other countries suggests that these conventional local taxes can provide significant financing for local expenditures, especially in the more affluent cities. A widely applied property tax has the potential of increasing revenue collections for local governments by 1 percent of GDP. 1.24. The non-tax revenues, which account for about 2 percent of GDP accrue from local fees and fines, sale of goods and services, rent and dividends. At the same time, the grant support from the external partners is about 0.5 of GDP. Going forward, Georgia’s graduation from IDA will lower the grant component of non-tax revenues. However, better monitoring and management of SOEs could potentially increase non-tax revenues. A clear and transparent dividend policy, common privatization rules and adherence to international accounting standards will be the key to generate sustainable flows from SOEs. Government Expenditures 1.25. The composition of Georgia’s public expenditures highlights the realignment away from defense and towards health and social sector spending. Decomposition of the general government expenditure shows that during 2006-08 defense spending was prioritized over all other categories (Figure 1.9). With military expenditures at 8.5 percent of GDP, Georgia was second only to Israel which was at 7.1 percent of GDP. Since 2009 the sharp cuts in the defense budget released resources and by 2011 the defense spending had dropped to 3.6 percent of GDP allowing social and economic sectors to benefit. In the post-crisis downturn the authorities undertook a fiscal stimulus and raised public investments. The social sector also benefited but to a lesser extent. In 2013 and 2014 the new government prioritized social expenditures over other objectives and increased its social obligations. Figure 1.9: Functional Decomposition, 2006-2014 Figure 1.10: Economic Decomposition, 2006-14 (In GEL million) (In GEL million) 10,000 3000 9,000 2500 8,000 7,000 2000 6,000 1500 5,000 4,000 1000 3,000 2,000 500 1,000 0 - 2006 2007 2008 2009 2010 2011 2012 2013 2014 2006 2007 2008 2009 2010 2011 2012 2013 2014 p Public Administration Defence and public order compensation of employees good and services economic sectors and environment Health interrest payments subsidies and grants Education Social assistance social benefits other exp Other social sectors capital and net lending Source: MOF 1.26. Within current expenditures, there has been a shift from consumption of goods and services towards social benefits in the recent past (Figure 1.10). The spending on social benefits reached 28 11  percent of total public expenditure in 2013 as compared to 22 percent in 2012 and 15 percent in 2007. This is likely to increase further to 31 percent of the total expenditure by the end of 2014. The 25 percent rise in the compensation of military staff in 2012 is likely to increase the share of the wage bill to 17 percent of total expenditures in 2014. Other current expenditures have however been shrinking as a result of fiscal consolidation measures undertaken by the government. Notably, government spending on goods and services fell from 16.6 percent in 2011 to 12.2 percent in 2013, as the government took measures to consolidate its current spending. 1.27. The recent hike in social assistance benefits will have far-reaching fiscal implications. In July 2013, the government doubled the monthly allowance under the TSA to GEL60 for the first household member and GEL48 for each additional member compared with GEL 30 and GEL24 previously. As a result, total spending on the TSA program in 2013 was GEL213 million or 0.8 percent of GDP, about 70 million more than if the benefit level had remained unchanged. This difference represents 0.3 percent of Georgia’s 2013 GDP. In 2014 the fiscal cost of this measure will be an additional 0.4 percent of GDP. The raised benefit levels could possibly create work disincentives and this is something that the government needs to be cautious about, given the high unemployment rate in Georgia. There is also a need to increase coverage and improve targeting of social spending. 1.28. Universal health care (UHC) is a welcome measure but further refinement is needed to translate new statutory entitlements into effective coverage. Following the elections of October 2012, the government announced that all Georgians would be eligible for state-funded health insurance and “universal coverage” would be provided. In addition to the existing medical insurance program (MIP) targeting the poor and other vulnerable groups of population, the basic universal health service package was introduced in February 2013 and the package was expanded to include a wider range of services in July 2013. The extension of entitlements has been backed up by a 90 percent increase in levels of public funding for health from 2012 to 2013 (from 333 million GEL to 634 million GEL). Overall, public spending on health increased from 1.6 percent of GDP to 2.1 percent in 2013. Preliminary assessment of UHC Program 8 shows some positive initial results. Almost all beneficiaries (96.4 percent) are satisfied or with in-patient and/or emergency out-patient services. Meantime, coverage of out-patient drugs for people with chronic conditions, on which there is currently high out-of-pocket (OOP) spending, remains inadequate. The UHC does not provide drugs at the primary care level for non-emergency cases for formerly uninsured UHC beneficiaries. The new UHC program is being administered by the Social Service Agency (SSA)9. From April 1, 2014 all MIP beneficiaries were transferred to UHC but maintaining the MIP service package until the expiration of the contract with the beneficiary. The Agency is tasked to perform the function of the single purchaser. This move and increased volume of claims to handle put pressure on the Government to ensure that appropriate controls and incentives for cost- containment and for efficient use of resources are in place. A longer-term development of provider payment methods and further strengthening of capacity at SSA should be considered to replace transitional provider payment methods and ensure efficiencies in purchasing. 1.29. The government underspent its health budget in 2013 because of uncertainties surrounding methodology and implementation of UHC. The UHC administration is guided by a capitation payment for PHC services, while reference prices generated based on standardized formula, are used for the reimbursement for in-patient care. Meantime, there was an open-ended reimbursement of provider claims, without realistic costing information, negotiated contractual arrangements, or oversight. With a stronger provider and patient awareness of the new program, health expenditures are likely to increase more than what the government is projecting. The government’s medium-term expenditure framework foresees only  8 A review of Universal Health Coverage Reforms Introduced In Georgia Since February 2013”-WHO, USAID, WB-May 29, 2014 ͻThe SSA is a Legal Entity of Public Law (LEPL) under the Ministry of Labor, Health and Social Affairs. 12  a marginal increase in the state health budget, from 1.6 percent of GDP in 2013 to 2.1 percent in 2014, as it is based on 2013 performance. However, the actual spending could be higher because the government is still grappling with issues related to the costing of universal healthcare. 1.30. Spending on pensions has been the largest component of social expenditures and ad-hoc increases in pension benefits has put pressure on the government’s fiscal position. In 2012, pensions accounted for 56 percent of total social spending and 4 percent of GDP. From 2012 to 2013, benefit levels were raised from GEL100 per month to GEL150 per month and its coverage was expanded to all pensioners. Under a scenario where future benefits are maintained at GEL150, a simulation shows that pension expenses will decline to 3.5 percent of GDP by 2016. But if the benefit level increases, pension expenditures are projected to be significantly higher. Under a scenario where future benefits are indexed to wage increases, the pension expense will be 4.3 percent of GDP by 2016 and reach as high as 9 percent of GDP by 2070. These results indicate that the sustainability of pensions will be dependent on the path of future benefit levels. By law, pensions are not indexed to the rising nominal wages, but in recent years, ad-hoc increases in the pension benefit levels have outstripped wage growth. If future benefit levels continue to increase faster than nominal wages, the pension spending could well exceed the regional level. In line with international best practices, the government needs to implement a retirement savings program which complements the public pension system. This program would result in contributions from employees and employers as a percentage of their income, on a mandatory or voluntary basis. Under this savings program, individuals with high income would receive high benefits when they retire, in proportion to what they contributed. Such programs would address the issue of benefit levels not keeping up in nominal terms and it would also alleviate pressures on the fiscal side. 1.31. Public expenditures on education are relatively low and inefficient, and this has impacted learning outcomes. Education expenditures accounted for about 2.7-3.0 percent of GDP during 2007-13 which is significantly lower than the CIS and EU new member states’ average of 5 percent. General education accounts for nearly 65 percent of sector spending while pre-school, VET and tertiary education accounts for the rest. Sector specific inefficiencies include too many teachers teaching very few hours and the poor quality of teachers which is a result of the low pay scale. This has had an adverse impact on student performance as measured by PISA scores. In addition, businesses complain of skill mismatches and the government is trying to address these concerns through improvements in vocational education and training (VET). 1.32. The public sector employs over half of the formal workers and public wage expenditures accounted for about 10 percent of GDP in 2012. As of 2012, the public sector, including central administration, local authorities, LEPLs and state owned enterprises (SOEs), employed 54 percent of the formally hired workers, and accounted for 20 percent of total employment (including self-employed). Compared with peers, Georgia has a relatively large public sector employment (Figure 1.11). Wage expenses for public sector employees stood at 4.4 percent, 0.8 percent, 3.1 percent and 2.1 percent of GDP in the central government, local governments, LEPLs and SOEs, respectively. The average annual earning per public employee in 2013 was about US$4,470 although there are large variations within this group (Figure 1.12 and Figure 1.13). A public school teacher earns an average annual salary of GEL 6,00010 while the CEO of the Partnership Fund earned GEL 323,500 in 2013. 11 Some 65,000 teachers, 50,000 workers in cultural, religious, scientific, educational and other public establishments, and thousands of employees in the over 400 malfunctioning SOEs are often reported as performing part time, low-income work, working much below their potential capacity. However, the top management in public institutions like the public registry, revenue services, data exchange agency, social service agency, financial analytical agency, municipal development fund,  10 Most of teachers are eligible for minimum income tax-free refund (threshold is GEL6,000) 11 www.declaration.gov.ge, 2013 13  and large SOEs receive high salaries and are eligible for bonuses and premiums. Weak monitoring and inadequate compensation rules have contributed to these wide disparities in public wages. Figure 1.11: Public Employment Vs. Size of the Government (In percent of total and percent of GDP) Source: OECD library: http://www.oecd-ilibrary.org/sites/gov_glance-2011-en/03/04/index.html?itemId=/content/chapter/gov_glance-2011-10-en Note: OECD employment (2011), OECD size of government (2009), Georgia employment and size of government (2012) Figure 1.12: Monthly Average Real Wages in Private Figure 1.13: Monthly Average Wages by Sectors in and Public Sectors, 2005-12 Private and Public Sectors, 2013 (In GEL, constant 2005 prices) (In current US$) Source: Geostat, staff calculations. 1.33. Public capital expenditures were low during 2012-13 compared with the previous five years when it averaged 8.7 percent of GDP a year. Public capital expenditures were a significant driver of Georgia’s economic growth and played a crucial countercyclical role during the crisis of 2008-09. However, during 2012-13, due to election-induced uncertainty and consequent delays in the budget execution, capital expenditures were lower than in previous years. On average, they accounted for 6 14  percent of GDP during 2012-13 or 25 percent of total general government expenditure compared with 31 percent in 2011. The implementation of several large projects was delayed in late 2012 and in 2013. These included both domestic and donor-financed projects, such as “sustainable urban transport development project (ADB)” with only 15 percent execution rate, “water infrastructure rehabilitation project (EIB, EU)” with 41 percent execution rate, and the domestically funded project for “rehabilitation and equipment of health care facilities” with an execution rate of 22 percent. Some projects were fully suspended, the largest of which was the Tbilisi by-pass railway construction. The government is likely to compensate for the above delays in 2014, at least partially. However, over the medium-term, capital expenditures are likely to stabilize around 5-6 percent of GDP a year. 1.34. Nearly 80 percent of public investments were in infrastructure projects in 2013 with education and health facilities accounting for another 11 percent. In 2013, more than 60 percent of public capital expenditure was in roads and municipal infrastructure and another 18 percent was in hydropower rehabilitation and transmission. Public investments in education projects accounted for 8.7 percent of the total capital expenditures during the year while health projects accounted for another 2.3 percent. Both education and health related capital expenditures are domestically funded (Figure 1.14). Figure 1.14: State budget capital expenditure, 2013 (In percent of total) Source: MOF. Note: Excluding net lending. 1.35. Subnational governments spend less than a quarter of the total public expenditures and are very dependent on revenue (inter-governmental) transfers. As of 2013, the subnational share in total government expenditures was 21 percent, which puts Georgia in the middle range of expenditure decentralization level among its peers with similar sizes. The local governments have responsibilities for expenditures on pre-school education, public administration and utilities, economic activities and capital investments. In terms of functionality, housing and communal amenities accounted for 40 percent of the local government expenditures. On the revenue side, a prominent feature is the low level of revenue autonomy at the subnational level. About 70 percent of local revenues are transfers from the national budget, much higher than the ECA average of 50 percent. Higher decentralization is likely to improve revenue collection (Chapter 4). 15  Efficiency of Fiscal Expenditures 1.36. To maximize the effectiveness of the lower level of public investments, the establishment of an effective public investment management (PIM) system will be crucial. On this front, Georgia has made some progress, especially on capital budgeting, and the new government is committed to implementing deeper reforms in this area. Efforts have been made to enhance the information content of budget documents pertaining to capital expenditures and to initiate more systematic processes to raise the overall efficiency and effectiveness of public investment. Nevertheless, more needs to be done to strengthen the PIM process, especially at project identification and appraisal stage (See background paper on PIM diagnostics). 1.37. There is a need to introduce performance-based compensation policies in the public sector to improve the quality of service delivery and to keep the wage bill under control. According to the 2012 Geostat business survey, public administration is the best paid sector in the nation, second only to financial intermediation (banking sector). However, the high salary does not normally translate into good quality of service delivery barring a few exceptions. The compensation policy should be strengthened and made performance-based. Bonuses account for nearly 10 percent of the total wage bill every year but they are largely unregulated with no institutional structures in place to determine them. They are completely at the discretion of the institution head who determines the amount, frequency and number of beneficiaries. Anecdotal evidence suggests that similar inappropriate uses of funds are common in SOEs and this affects dividend pay-outs. The bonus fund needs to be regulated as a part of the formal compensation scale using clear criteria and rules. 1.38. The 2012 PEFA assessment of Georgia showed significant improvements and the government is well-poised to undertake improvements in expenditure efficiency. The basic set of systems has been put in place for strategic budget planning, budget formulation and execution. The integrated public financial management system was launched in 2013. The introduction of international good practice in the budget cycle of the government is well advanced, including robust systems for budget preparation, adequate chart of accounts, reliable execution (including accounting and reporting) and sufficient controls. Significant progress has been achieved on program-based budgeting, furthering the government’s objective of greater results-focus in fiscal planning. The legal framework governing public procurement was further amended, with Electronic Government Procurement (E-GP) introduced in 2011 which was linked to the Treasury’s information system, thus providing for full information sharing. All the above reform initiatives were implemented to address the weaknesses identified by the 2008 PEFA assessment in the areas of external control system, personnel and payroll, public procurement, and reporting of high quality consolidated financial statements. D. THE WAY FORWARD 1.39. Intensive monitoring of social expenditures will be crucial to maintain fiscal sustainability and to achieve better social outcomes. The increase in social benefits and assistance was a core part of the new government’s election manifesto. The government has followed through on its election promise and has raised social benefits considerably. However, implementation capacity in the social sectors is constrained and needs to be enhanced. 1.40. Further cuts to capital expenditures need to be guarded against as this would impact growth. The increased current expenditures are likely to generate fiscal pressures over the short- to medium-term. With limited upsides on revenues and difficulties in scaling back recurrent expenditures, 16  the government could resort to lower spending on public investments. However, since Georgia has a large infrastructure deficit, such a measure could impact short- and long-term growth. 1.41. The Ministry of Finance would need to work closely with the respective public agencies that are responsible for implementation of the key social and infrastructure programs. Since Georgia follows program based budgeting, it will be important to evaluate the performance of social benefit programs during each budget cycle to improve the efficiency and effectiveness of these expenditures and also to ensure fiscal sustainability. 17  Spotlight 1: Social Protection and Health Expenditures 1. Spending on social benefits including health is likely to account for about 32 percent of the budget in 2014. Most of the social expenditures are managed by the Social Service Agency (SSA) under the Ministry of Health, Labour and Social Protection (MOHLSA). In 2012 spending on social benefits absorbed 27 percent of the general government expenses (Table S.1). As a result of recent policy initiatives the social budget is expected to increase in nominal terms by 51 percent by 2014 from 2012. It is set to further expand by about 12, 9, and 10 percent in each of the subsequent years until 2017. 12 2. Georgia’s spending on social benefits is largely driven by three key programs—pensions, TSA and health. Georgian authorities reformed the system of social benefits in 2006 and abolished numerous fragmented and overlapping social assistance categories. According to the 2014 budget law, pensions will absorb about 47 percent of the total social benefit expenditures. After pensions, the TSA and health protection programs consume around 22 percent each, while the remaining 10 percent of social benefits budget is managed by the other line ministries for case-by-case support to their employees or by local municipalities to supplement the centralized social assistance programs. 3. The simplicity of the benefit structure combined with the effective mean-testing formula and universal flat pension system allows the relatively small team of the SSA to effectively manage over 30 percent of the general government budget. 13 The operational system has been strengthened in 2014 by activating a sophisticated Social Information Management System (SIMS). The SIMS is set to improve the targeting effectiveness of social assistance and enable beneficiaries to access benefits faster. It is also able to exchange data with other government information systems, such as revenue services and property registration databases and hence will help quicken verification of data provided by applicants on their socio-economic status. The SIMS consists of 3 pillars with various sub-modules: pensions, child protection and guardianship, and targeted social assistance. 4. The administration of the targeted medical insurance program was straightforward due to limited role of the state while the implementation of the Universal Health Care program requires more sophisticated methodology and better capacity. The operations of SSA were limited to the disbursement of the pre-defined flat premiums to the private insurance companies that were responsible to ensure service provision to the targeted vulnerable groups of population (around 38 percent of population) including those identified through the TSA database. While simple in administering, the system had a number of weaknesses related to the oversight of quality standards of service provision and suspicion over the high profit margins of the insurance companies. Following the elections of October 2012, it was announced that all Georgians would be eligible for state-funded health care – that is, “universal coverage” would be provided. At the end of February 2013, the uninsured population was extended a package that included basic primary health care services paid via capitation to primary care providers by SSA and coverage of emergency/trauma care at hospitals paid via fee-for-service reimbursement by SSA. In July 2013, this package was expanded to include a comprehensive range of services very similar to the insurance already available to the poor and the pensioners. These packages mostly exclude outpatient drugs component, on which very high out-of-pocket (OOP) spending was registered. These services are also reimbursed on a fee-for-service basis by SSA according to the claims submitted by providers. As identified during the implementation of USAID-funded health project the new system has challenged the capacity of SSA to handle the high volume of claims.  12 BDD 2014-2017, pg 50 13 1,704 employees including all level staff in 2013 18  Table S.1: Major Social Benefit Programs (In GEL million) 2010 2011 2012 2013 2014 b 2015p 2016p 2017p Total expenditure 7,222 7,728 8,287 8,063 8,972 9,509 10,246 11,047 Social benefits total, of which: 1,624 1,656 1,858 2,295 2,811 3,150 3,434 3,760 Pensions, old age 924 987 1,062 1,149 1,326 2,130 2,272 2,352 Targeted social assistance (TSA) 232 220 271 472 624 Targeted medical insurance (MIP) 147 121 153 240 278 NA NA NA Universal health care (UHC) 0 0 0 69 200 650 705 720 Disease-specific and preventive health care 182 156 179 114 116 Other social benefits (by ministries, LGs, etc.) 139 171 169 252 267 370 457 688 Source: MOF, State Treasury Service 5. Unlike all other countries in the ECA region, Georgia’s social protection system is financed entirely out of general revenues as employers’ social contribution tax was abolished in 2007. In particular, Georgia has a non-contributory public pension scheme in place which provides a flat universal pension to all elderly at the rate of about 19 percent of 2013 average wage. Other social safety programs consist of the major assistance package targeted to poor (TSA) and small scale categorical programs for various vulnerable groups. The recently introduced Universal Health Care (UHC) program is funded entirely from the tax and non-tax collections as well. With this in place, Georgia's non-contributory social safety spending stands out and is about three times higher than the rest of the region (about 6 percent excluding health versus ECA’s average at 2 percent, Figure S.1). Figure S.1: Social Protection Spending Across ECA countries (In percent of GDP) Tajikistan 11 Kazakhstan 12 Social Assistance Armenia 12 Albania 13 Labor Market FYR Macedonia 11 Social Insurance Kyrgyz Republic 11 Moldova 10 Estonia 11 Lithuania 09 Serbia 10 Ukraine 11 0% 5% 10% 15% 20% 25% Source: ECA SPeeD. Pensions, as are not contributive in Georgia and hence should not be classified as SI, although their social functions are similar to what social insurance programs cover in other countries. In most ECA countries social insurance includes pension and disability programs based on social insurance contribution payments. Labor market programs include both passive (unemployment) benefits and active labor market programs (ALMPs). 6. Georgia’s social benefits are comparatively lower but there is limited fiscal room to increase it further. Total social protection spending including health care programs is set to reach 9.5 percent of GDP in 2014, which puts Georgia behind most of the ECA countries in terms of support to the population. However in the absence of a supplemental mechanism of health and social insurance, a further increase in social benefits cannot be ruled out and this will pose risks to fiscal sustainability. In the event of lower growth (2 percent growth in 2014, 3 percent growth in 2015, returning to medium term 5 percent from 2016) the cost of social protection may reach 10.5 percent of GDP compared to 7.1 percent in 2012 and 9.5 percent in 2017 under the baseline scenario. That will consume over 45 percent of tax revenues as compared to 28 percent in 2012 and 38 percent in 2017 (under the baseline scenario, Figure S.2). 19  Figure S.2: Outlook of the Social Protection (In percent of GDP, in percent of tax revenues) 11% 50% 10% 40% 9% 8% 30% 7% 6% 20% 2012 2013 2014 2015 2016 2017 2012 2013 2014 2015 2016 2017 Growth shock (2% growth) % of 2% growth taxes Baseline Baseline Higher social needs (5% higher) Higher social needs (5% higher) Source: Staff calculations. Pensions 7. The existing universal and non-contributory pension provision system in Georgia protects the elderly against poverty, but does not provide adequate income replacement to retirees. The pension system consists of a flat benefit of GEL 150 paid out of general revenues to all men 65 years of age or older and all women 60 years of age or older, which provides a replacement rate of around 18 percent of average wages. Figure S.3 shows the relationship between the benefit and average wage from 2004 to 2013. 14 Pension expenditures amounted to 4.2 percent of GDP in 2013. The new level of basic pension, raised by 20 percent in September 2013 by the Government, stands close to the subsistence level for a working age male (GEL 154). If not indexed, the basic pension could become fiscally unsustainable, which suggests that it may be raised again in the future by an unpredictable amount and with unpredictable timing. In order to avoid such fiscal uncertainty, the government would need to carefully consider a basic pension valorization and indexation rule which is fiscally and socially consistent with the eventual introduction of the earnings related pension pillar in the future.  14 The average benefit used for the 2013 calculation includes the increase in September 2013. However, wage data for 2013 were only available until end of third quarter. As a result, the benefit to wage ratio is likely to be overstated for 2013. 20  Figure S.3: Comparison of Pension Benefit to Figure S.4: Projected Pension Spending Under Average Wage Alternative Scenarios (In percent) (In percent of GDP) 25% 10% 9% 8% 20% Percentage of GDP 7% Pension/Average Wage 6% 15% 5% 4% 10% 3% 2% 5% 1% 0% 2012 2016 2020 2024 2028 2032 2036 2040 2044 2048 2052 2056 2060 2064 2068 2072 2076 2080 0% 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 125 lari, no index 150 lari, no index 150 lari wage-indexed Source: Wage data from GeoStat. 8. Benefit levels were raised by 50 percent during the twelve months starting September 2012. Social benefits increased from GEL 100 per month prior to September 2012 to GEL 125 per month for those above the age of 67 in September 2012 and by April 2013, benefits were raised to GEL 125 for all pensioners, irrespective of age, followed by another increase to GEL 150 per month for all pensioners in September 2013. Figure S.3 shows that the growth in the pension benefit level outstripped wage growth significantly in recent years. On the other hand, it is true that pensions are low in Georgia. Even with the September 2013 increase, the pension will reach about 18 percent of average wage. Pensions have been kept low as a means to keep them fiscally affordable. 9. Going forward, affordability of the pension system depends entirely on the indexation of future pensions. Figure S.4 shows projected pension spending as a share of GDP prior to the September 2013 increase with no further increases in benefit levels, pension spending with the September 2013 increase but no further increases, and pension spending if after September 2013 pensions become linked to wage growth. While the increase from GEL125 to GEL150 has a small short and medium-run impact on pension spending, in the long run the level is virtually unchanged. If, on the other hand, pension spending is indexed to wage growth, pension spending will grow as a percent of GDP in line with the growth in the number of elderly in the population. The projection shows pension spending is expected to rise from 4 percent of GDP to 9 percent of GDP over the next 65 years. While a huge jump from today’s spending, the projected levels are not out of line with regional neighbors. Pension spending among the transition countries of Europe and Central Asia and among the EU member countries currently averages between 9 and 10 percent of GDP, and in almost all cases is expected to rise in the future. However, the challenge is that there is a possibility of further ad-hoc increases in pension benefits which could grow faster than wages as they have for the past two years. If this trend continues, the spending could well exceed the regional averages. 10. While the pension benefit in Georgia is by no means the lowest in Europe or the region, or even among high-income countries, this is the only old age benefit currently in Georgia while in most of these other countries many elderly receive more than just pensions. Figure S.5 shows the Georgian pension benefit as a percent of average wages in 2013 compared to minimum pension benefits in other countries. A benefit level equal to 17-19 percent of average wages might be sufficient for low earners or for those who had limited earnings during their working years, but high and middle income earners will experience a sizable drop in living standards if this benefit is their only source of income. A 21  risk of having only this single benefit is that high and middle-income earners would expect the government to increase the size of the benefit over time. Figure S.3 already showed that at 18 percent of average wage, the long run cost of this benefit will be what average countries currently spend on pensions – about 9 percent of their GDP. A higher benefit is likely to be unaffordable. Figure S.5: Minimum pension in Selected Countries (In percent of average wage) Source: OECD 11. This spotlight recommends implementing a retirement savings program that complements the public pension system, in line with what the Government is suggesting in both the medium-term document of the government (BDD 2014-17) and the draft National Development Strategy of Georgia (NDS 2014-2020). These savings would come from contributions paid by employees and employers as a percentage of their income. Therefore, higher income individuals would receive higher benefits when they retire, in proportion to what they contributed. The savings program could be implemented on a mandatory or a voluntary basis. International experience shows that if individuals are left to voluntarily save on their own, only a limited number of high-income individuals actually participate. Best practice now suggests the use of automatic enrollment, where employees are automatically enrolled in a savings scheme unless they choose to opt out. A much higher percentage of people actually participate, and the system remains voluntary. Tax incentives are typically provided on these savings, with the contributions not subject to income tax until they are withdrawn at retirement age. Earlier withdrawals, while possible, can be discouraged with large tax penalties. The difference between mandatory savings programs and voluntary savings with automatic enrollment becomes quite small, particularly in countries with large informal sectors. TSA 12. Targeting of the TSA program has improved considerably in the last years, but gaps remain in coverage. The TSA program supports the income of about 10 percent of the population - 454,000 beneficiaries living in 150,000 families as of end 2013. Georgia’s targeting approach is fairly good and effective at excluding those that should not benefit from the TSA program. In fact, 45 percent of the benefits reach the poorest decile and an additional 18 percent goes to the next decile of poorest families. The leakage of resources is low relative to other countries, with only 6 percent of the funds leaking to the richest three deciles of the population. The issue with the TSA program in Georgia is related to its relatively low coverage. Estimates show that about 300,000 individuals who may qualify do not receive the benefits. As shown in Figure S.6, both the number of registered applicants and the actual beneficiaries barely changed in the last 18 months. Currently, only one person out of four 22  registered people receive the TSA benefit. At an absolute poverty rate of 17.7 percent, about 730,000 Georgians are considered poor. While coverage is a function of the budget envelope, the MOLHSA is committed to revise the TSA mean tested formula to replace subjective parameters with more objective ones and improve coverage through even better targeting. Figure S.6: TSA Coverage and Spending Per Beneficiary (Million people, GEL) Source: SSA, staff calculations. 13. The benefit increase in July 2013 has significant fiscal implications. As expected, in the first months since the implementation of the increase, the budget allocated to the TSA program doubled – increasing from the range of GEL10-12 million to approximately GEL23 million per month. Without any increase in the number of beneficiaries, the Government allocated GEL213 million by the end of 2013, about 70 million more than if the benefit level had remained unchanged. This difference represents 0.3 percent of Georgia’s 2013 GDP. In 2014 the fiscal cost of this measure implies an estimated additional 0.4 percent of the GDP. 14. Three possible scenarios (0, 1, and 2) project the potential fiscal implications of the recent benefit increase in the TSA program from 2013 to 2016. All scenarios start with the average number of beneficiaries at 430,000, assume a constant absolute poverty rate of 17.7 percent, and apply the same estimated GDP levels until 2016. 15 Table S.2 summarizes the assumptions and main results. Scenario 1 closes the current coverage gap at the new benefit level, but its cost exceeds the baseline by 1.27 percent of Georgia’s 2016 GDP. Assuming automatic increases of 5 percent due to the new benefit level, scenario 0 stays far from closing the coverage gap - reducing it only by 10 percentage points from 40 to 30 percent. In contrast, scenario 2 illustrates a greater coverage improvement at a bearable fiscal cost slightly under 1 percent in 2016. If the eligibility conditions were revised, as the government’s socio-economic development strategy suggests, a sudden shock of new entrants to the TSA program could be expected. Under these circumstances, scenario 2 would be of interest and for further consideration.  15 GDP levels are based on GoG (2014) (in nominal million GEL) at 27,238 in 2013; at 30,113 in 2014; at 33,515 in 2015; and at 37,303 in 2016. 23  Table S.2: Fiscal Implications of the TSA Program under Three Assumptions (In percent, except where otherwise indicated) Assumptions Implications Benefit Level Cost of (For First and Additional Family Population Covered GDP by Members, Respectively) Year-On-Year Coverage Increase by 2016 2016 Scenario 0 GEL 60 & GEL 48 5% 12% (498,000) 0.85% Scenario 1 GEL 60 & GEL 48 20% 18% (743,000) 1.27% Scenario 2 GEL 60 & GEL 48 20% first year and 5% each year after 14% (568,000) 0.97% Source: Staff calculations. 15. Georgia’s TSA program could potentially lead to work dis-incentives. A previous Word Bank PER (2012) report analyzed the demographic profile of TSA beneficiaries finding that, in 2009, TSA benefits did not discourage people from seeking employment. That study found no differences in labor force participation between TSA beneficiaries and non-beneficiaries. More recently, the same analysis was revised using new information (Kits et.al. 2014). 16,17 Preliminary results suggest that the TSA program generates work disincentives around the eligibility threshold, although these disincentives are concentrated among young women. Among men, there is no statistically significant effect. These results suggest that the disincentive effects found from the TSA program in Georgia may be mediated through appropriate mechanisms for supporting working women, especially when they are young. As previously posited, there is likely a link between having care related responsibilities and no access to care centers that are cost effective. More research is needed to thoroughly investigate what induces disincentives, among women receiving TSA, to participate in the labor force. Other studies have shown that linking benefits such as those received through the TSA, to activation programs can achieve a greater impact on poverty reduction while minimizing work disincentives. Health 16. New initiatives in the health sector also have significant budgetary implications. Table S.1 shows Georgia’s health budget during 2010-17. Health spending is projected to double in nominal terms, and increase by 75 percent when measured as a share of the budget or as a share of GDP. This is to be expected given the near tripling of the population covered by state-funded health programs over this period (despite some offsetting savings from scaling down the priority disease programs that will be folded into the new initiative). The ratio of health spending to GDP is expected to rise to 2.1 percent in 2014, still lower than the regional CIS average (approximately 3.5 percent). The affordability of this increase from a broader fiscal perspective will depend on revenue trends and other sectoral spending patterns. 17. The estimates of the fiscal impact of the UHC were inaccurate in 2013 and this has raised concerns on the ability of the government to realistically project its medium-term health expenditures. Budget execution during 2013 was much lower than anticipated vis-à-vis the original budgeted amount. There are several possible reasons for this: (i) the original 2013 budget assumed the achievement of universal coverage by February, whereas hospital coverage was not included in the package until July, and thus figures do not reflect a full year of coverage; (ii) public awareness of the new program is picking up and this could result in additional demand for health care from the previously uninsured; (iii) it may be too early for provider “gaming” of the new program to take  16 The study used a newly designed survey of approximately 2000 households and administrative data and combined it with a regression discontinuity design in order to exploit the sharp discontinuities in treatment around the proxy means score threshold of the TSA, before the recent policy changes in July 2013. 17 Source: Kits, B., Santos, I., Smith, O., Isik-Dikmelik, A. (forthcoming) “The Impact of Targeted Social Assistance on Labor Market Outcomes in the Republic of Georgia. A Regression Discontinuity Approach” , Washington DC, The World Bank 24  hold; (iv) reimbursement rates may be low. Therefore, it likely that spending will rise during 2014 as a result of the full-year of implementation and stronger provider and patient awareness of the new program. Precise forecasts of health spending trends are very difficult to make, but over the medium- term (2 to 4 years), government health spending in a range near 2.5 to 3 percent of GDP should be expected if the program is to restore and strengthen outpatient drugs component in the package and thus tackle the problem of very high OOP spending. Currently the government forecast indicates a downward trend of the health budget reaching 1.7 percent of GDP by 2017. 18. The UHC program will also have implications for the private insurance business development. The insurance sector is small in Georgia but it has been on the rise since 2004. The government had introduced the state-funded targeted medical insurance program in 2007 and had implemented it through private insurance companies until 2013. The volume of net insurance premium attracted by the sector increased considerably from GEL150 million in 2008 to GEL333 million in 2012. The structure of insurance products in Figure S.7 shows that the state is the main driver for the sector with a bit less than 50 percent contribution to the total premium. The dependence on MIP is even more profound if company by company statistics are considered. The introduction of the single purchaser and exclusion of the private insurance companies from the UHC will most likely limit the market for the sector. One of the key rationales for the government to move to a single purchaser system from 2014 is the opportunity to make significant savings as the private companies were running high profit margins as compared to the international standards, which is close to 10 percent. The data released by the regulatory agency confirms that in earlier years the net profits were in the range of 30-40 percent for the sector as a whole. The profit margin had reduced to 18 percent in 2012 (Figure S.8). 18 Figure S.7: Structure of Private Insurance Products Figure S.8: Profitability of Insurance Sector (In percent of total) (In GEL million, in percent) Source: NBG. 19. Going forward, there are many uncertainties on program implementation. A key issue is the capacity of the SSA to act as a purchaser of health care services under the program. The first months of implementation have been characterized by an open-ended reimbursement of provider claims, without adequate costing information, standardized pricing of services, contractual arrangements, or oversight. SSA’s claims management capacity is currently improving, but from a low baseline. It has recently been decided that private insurance companies would no longer have a role in purchasing care for the state- funded programs after their current contracts expire during the course of 2014. In principle, an active single purchaser such as the SSA can keep costs down on the basis of its monopsony power, but SSA’s  18 National Bank of Georgia, 2012 Annual Report 25  capacity to fulfill this potential is uncertain at present. Moreover, attempts to create such an institution in Georgia have failed in the past. A final risk is that the quality of care in Georgia is often inadequate, and unless there are concerted efforts to address this issue, efforts to translate the universal health coverage program into actual improved health outcomes will meet only limited success. 20. Several measures are needed to address the key implementation challenges and to ensure that the health budget is efficiently spent. First, there is a need to strengthen SSA’s implementation capacity. The current environment, with open-ended reimbursement on a fee-for-service basis with widely-varying prices determined by providers, should be replaced with contractual arrangements that define ex-ante some combination of prices, service volumes, and budget ceilings at the provider level. This would serve to impose a budget constraint at the provider level to help contain cost growth going forward. Initial steps have been taken to develop diagnosis-related groups as a provider payment mechanism, but implementation will take a long time and information requirements are substantial. As noted, improving quality of care is also essential to ensure better value for money of Georgia’s health care spending. Lastly, these efforts should be combined with outreach and information for the public, along with efforts to monitor and evaluate the reforms as they unfold. 26  CHAPTER 2. EDUCATION EXPENDITURES A. INTRODUCTION 2.1 Poor performance of the Georgian students in international assessments points to the inefficiencies in the education sector. Georgia’s performance on the reading scale of PISA 2009 was below what would be expected given its income level. The gap between Georgia and the OECD countries on the reading scale is approximately three years of schooling (each year accounts for 40 points). In addition, only between 30 percent and 40 percent (depending on the subject) of Georgian 15 year-olds scored Level II or above, which is a threshold usually referred to as functional literacy. 2.2 Government spending on education in Georgia is low – compared to countries with similar per-capita incomes and relative to both the shortage of human capital and the country’s ambitions. Despite a real increase of nearly 47 percent in education expenditures since 2006, Georgia’s education budget amounted to 2.9 percent of GDP in 2012, below all but a few countries in Europe and Central Asia and lower than comparators at a similar level of income per capita. And even though teacher salaries are one of the lowest among public employees in Georgia, teacher and administrator pay accounts for more than 70 percent of the education sector budget, leaving little space for expenditures directed at curriculum improvements, trainings for teachers, grants for research and development, scholarships for needy students and capital investments to enhance school facilities. 2.3 The per-capita financing system in the education sector is transparent and leads to an efficient allocation of resources. Schools receive direct transfers of funds from the Ministry of Education and Science (MES) based on the number of students for a given year. These transfers cover salaries, utilities and routine maintenance costs. The allocation formula was adjusted over time to address the problem of underfunded schools which required additional transfers on top of the per capita financing. The system is functioning well currently and is used to determine the education sector budget, including vocational education. 2.4 An efficient distribution of resources cannot offset substantial sector inefficiencies, such as poor teacher quality, and excess number of teachers from the pre-school through the secondary level. Teaching is not a sought after occupation in Georgia because of the very low pay scale and this has impacted the quality of teachers. At 8.7, the student teacher ratio of preschools in Georgia is considerably lower than the OECD or EU 21 countries (at 12.2 and 11.8, respectively). As a consequence, several regions in Georgia run very high costs per student. With an aging population, there has been a decline in the number of students and a consolidation in the number of schools in Georgia. However, at 241 students per school, the average school size in general education remains small and drives up costs. There are too many teachers teaching too few hours suggesting that there is a glut in the education system. Based on an academic year of 40 weeks, the average yearly teaching hours are less than 600 while the average teacher in an OECD country teaches 790 hours in primary, 709 hours in lower-secondary, and 664 hours in the upper-secondary level. In countries such as the US and Chile, these figures go beyond 1000 hours per year. 2.5 The vocational education and training program has been completely overhauled but only a fourth of the VET graduates are able to find jobs in their mainstream education disciplines. The VET system underwent significant changes with the adoption of the Law on Vocational Education in 2007. The VET strategy which was developed resulted in a consolidation and subsequent expansion of VET institutions, introduction of a voucher system for VET institutions and significant public investments in the VET sector. The government has also eased the admission requirements for certain VET programs to make them more accessible. However, the system is plagued with high drop-out rates 27  which lead to a wastage of resources. In addition, surveys indicate that only one-fourth of the graduates from VET institutions are able to find a job in the area in which they have educational qualifications. Therefore, there is room to make substantial improvements in the curriculum and the quality of teachers. 2.6 This chapter reviews the fiscal implications of the education sector in Georgia, with an emphasis on preschool and general education and vocational education and training. The analysis in this chapter leads to the following measures for the government to consider: x Preschool Education: o Use the facilities of existing underutilized primary schools to hold classes for preschool education, train primary school teachers to teach preschool children, and target children of disadvantaged backgrounds and ethnic minorities to enroll in preschools. o Improve the quality of service delivery by making the (MOES) responsible for setting and monitoring quality standards. x General Education: o Improve the monitoring of student performance through regular national and international assessments. o Increase the number of working hours of teachers to bring it closer to OECD standards. This would result in a decline in the number of teachers and the savings could be used to improve the salary scale but with improvements in teacher quality. o Increase the class size. This again would require a reduction in the number of teachers. o Raise starting salaries of teachers without making the gradient steeper to improve teacher quality. x Vocational Education and Training (VET): o Add more general education content to VET programs. o Encourage public private partnerships in the financing and delivery of vocational education. o Reduce drop-out rates from VET programs by providing career guidance to help pupils make better informed choices. 2.7 The rest of the chapter is organized as follows. Section B provides an overview of the governance set-up. The next section analyses the level and composition of public spending on education in Georgia. Section D assesses the impact of public spending on enrolment rates and educational performance while section E suggests the way forward. B. GOVERNANCE STRUCTURE OF THE EDUCATION SYSTEM 2.8 The education sector in Georgia is dominated by public schools and universities but with a growing share of the private sector in higher and vocational education. The system of education in Georgia consists of pre-school, general (primary, basic and secondary) and tertiary and vocational education. In 2013, there were 2,118 public and 362 private schools teaching nearly 700,000 students. For the academic year 2012-13, nearly 91 percent of the students were enrolled in public schools at the pre- school and general education level while the rest were in private schools. Although the number of private institutions at the tertiary level is double that of the public institutions, 72 percent of the students were 28  studying in public institutions because of a larger number of vacancies. Vocational education is the only area where the number of students in private institutions exceeds that in the public ones (Table 2.1). Table 2.1: Size of the Education Sector, 2013 Number of Number of Enrolment in Enrolment in Public Private Public Private Institutions Institutions Institutions Institutions General education 2,085 247 506,659 52,756 Vocational education 14 77 8,457 12,619 Higher education 19 38 84,338 32,209 Source: Geostat and National Center for Education Quality Enhancement. 2.9 The provision of education is highly centralized in Georgia with the exception of pre-school education. The MES is the key player in education policy-making and has the primary role for sector oversight and management. There is also a network of 69 education resource centers (ERCs) which support the implementation of education sector policies at the general education level and also assist schools on financial and logistical issues. Each ERC covers between 5 and 20 schools. In addition, there are school boards of trustees comprising select teachers, parents, students and a local government representative. However, they have a very limited role to play in the functioning of schools with most of the decisions taken centrally. Local governments fund preschools entirely from their own revenues while there is a per capita financing system for the general, vocational and higher education level.19 With the exception of pre-school education, the MES ensures accountability and standards in the system through specialized semi-autonomous institutions. 2.10 The management and delivery of preschool education is with the local authorities and the role of the MES is limited. The MES has the authority to recommend to local governments a set of standards, which are however not binding. The coordination, methodological guidance, evaluation and monitoring of the learning process in kindergartens is performed by the preschool agencies which are accountable to the local authorities. A majority of these agencies have limited institutional capacity to implement and monitor quality standards proposed by the MES. Neither does the MES have tools to oversee and encourage implementation of the proposed standards. The MES is currently developing a new law on preschool education and a new consolidated education sector strategy in which it could potentially include national goals for preschool education and mechanisms to facilitate and monitor the quality of preschool education. 2.11 The MES is responsible for the vocational education policies although majority of vocational education and training institutions are privately run. The National Vocational Education and Training (VET) Council, established in 2009, is responsible for ensuring collaboration among the government, employers and other stakeholders in the sector. There were 19 public vocational education institutions (six community colleges, four higher education institutions and nine VET colleges) in 2013 and 72 authorized private service providers. The vocational education system is divided into five levels based on the National Qualifications Framework (NQF). Service delivery in publicly run VET programs is carried out by four types of institutions: VET colleges, community colleges, some general schools and some higher education institutions. VET colleges and general education institutions are entitled to offer VET programs at levels I to III, whereas Community Colleges and Higher Education Institutions can offer programs at all five levels.  19 Parental fees were abolished in September 2013. 29  C. LEVEL AND COMPOSITION OF EDUCATION EXPENDITURES 2.12 Spending on education has remained at around 3.0 percent of GDP over the past few years despite a real increase of nearly 47 percent in education expenditures since 2006. Georgia has experienced an upward trend in government expenditures on education (Table 2.2). The two exceptions were a temporary decrease during the economic crisis in 2008 and a slight decrease in 2011. The latter was the result of the decline in public allocation for preschool education by 24 percent with the introduction of parental fees in 2011. However, since 2006 the growth of the total public expenditures constantly outpaced the growth of the education budget, which explains the downward trend in education spending as a share of the total public expenditures. Public expenditures in 2012 and 2013 have been 2.8 percent and 3.0 percent as a share of GDP, respectively, which remains below other countries even with a similar level of income per capita (Figure 2.1). Table 2.2: General Government Expenditures on Education (In percent except where otherwise indicated) 2006 2007 2008 2009 2010 2011 2012 2013 GG expenditures on education (current values) 378 438 480 531 586 629.3 727 777 GG expenditures on Education (2006 prices) 378 401 400 434 448 443 516 555 Real growth of education expenditures 6.1 -0.2 8.6 3.1 -1.1 16.6 7.4 Education expenditure (% of Total GG exp.) 9.0 7.5 6.7 8.1 8.1 8.1 8.8 10.1 Total education expenditures as % of GDP 3.0 2.7 2.9 3.2 2.9 2.8 2.9 3.0 Total Gov. expenditures (current values) 4,208 5,865 7,119 6,536 7,273 7,728 8,287 7726 Total real Gov. expenditures 4,208 5,369 5,924 5,347 5,555 5,438 5,886 5,516 Government size (total exp/GDP) 30.5 34.5 37.3 36.3 34.8 31.7 31.7 30.0 Memo: GDP (current GEL) 13,790 16,994 19,075 17,986 20,743 24,344 26,167 26,800 CPI index (2006=100) 100 109 120 122 131 142 141 140 Source: MOF, MES, GeoStat, WB staff calculations. Figure 2.1: Public Spending on Education in Selected 2.13 On a per student basis, Georgia’s ECA Countries expenditures are lower than that of (In percent of GDP) comparators at all education levels (Table 2.3). In preschool education, the per student cost was approximately 12 percent of the GDP per capita in 2012 while OECD and EU countries spent about 20 percent of their GDP per capita.20 The per student cost of preschool was approximately 80 percent of that of general secondary education, which is similar to the ratios in OECD and EU countries. Expenditures per student in higher education are 30 percent higher than those of general secondary education. This is lower when compared to the EU and OECD countries, Source: World Bank, EdStats; TransMoney 2013. which spend approximately 60 percent more per Note: Latest figure available for Latvia and Croatia is 2009. MOF, Georgia. For the Russian Federation, Slovakia and Estonia tertiary education student. latest figure is 2008.  20 Public expenditures on preschool education are also below those of CIS countries. 30  Table 2.3: Expenditures, Student Population in the Public System and Unit Costs, 2012 (GEL million, student number, GEL per student, ratio and share of GDP per capita) Expenditures Students Unit Cost Unit Cost Ratio (general sec. Share of GDP Level of Education (million GEL) (2012) (GEL) (US$) = 1) Per Capita Preschool 74,5 103,027 723 438 0.8 12.4 General secondary 451,6 516,738 874 529 1.0 14.9 Vocational education 9,3 12,746 730 442 0.8 12.5 Higher education 79,6 70,922 1,122 680 1.3 19.2 Source: MOF; MES; Geostat (2012); OECD (2013), p. 177. Note: Vocational education expenditures in 2012 were particularly high because of heavy capital investment (see Section 0). There are 12,746 students attending public institutions, but there are some students who receive the state voucher and attended private institutions. Due to this, we provide an upper bound for the unit cost (assuming that no students in private institutions had the voucher). 2.14 Spending on general education accounts for about 62 percent of the total education budget, followed by higher education at 11 percent (Table 2.4). This is in line with the trend in most ECA countries. All these expenditures are on a per capita financing basis. The majority of central government expenditures are channeled through the MES to fund all levels of education except for preschools which is funded by local governments. A small share of the education budget is also administered by various line ministries, which mostly support programs in vocational and higher education in relevant areas for the ministries (i.e. health; medical education; agriculture; police academy, etc.). Table 2.4: Education Expenditures by Level of Education (In GEL million) 2006 2007 2008 2009 2010 2011 2012 2013 est Total Education expenditures 377.5 437.7 480.4 530.7 586.2 629.3 727.0 776.9 Preschool 29.2 53.7 59.9 71.1 82.1 64.5 74.5 96.9 General education 199.3 293.8 315.3 331.0 367.3 392.1 451.6 502.9 Vocational education 0.0 15.3 13.3 11.7 10.7 9.2 9.3 9.3 Higher education 23.6 48.2 52.5 65.0 64.3 68.2 79.6 78.8 Other education programs 125.5 26.7 39.4 51.9 61.9 95.3 111.9 89.0 Source: MOF. 2.15 The government has also introduced some targeted programs in specific disciplines although they form a small share of education spending (Table 2.5). The largest share of resources is spent on per capita financing of general education schools. Voucher financing of higher education institutions and infrastructure programs are a distant second and third. In parallel, over the last few years, the MES has initiated relatively small-scale but well-targeted programs placing a focus on Information and Communication Technologies (ICT) such as “My First Computer” for first graders, the physical security of schools (“Mandaturi” Program for enhancing safety and disciplinary measures in all schools), and provision of English language classes taught by native speakers (“Teach & Learn with Georgia” inviting native English speakers to teach in Georgia). 31  Table 2.5: State Budget Education Expenditures by Major Programs (In GEL million) 2006 2007 2008 2009 2010 2011 2012 2013 Total MES 358 411 458 490 538 553 627 675 General education per capita vouchers 187 201 269 300 325 316 326 410 National examination center 0 2 6 8 9 17 17 17 Science Programs and Grants 17 22 31 40 38 22 19 19 National program of school rehabilitation 73 79 25 8 14 38 72 27 Teachers' professional development 0 0 1 2 9 18 18 14 VET support programs 4 6 8 8 8 6 18 14 HED support programs 30 40 47 52 52 56 93 90 WB: APL1, APL2 10 9 8 18 13 3 0 0 Other small scale projects 37 51 63 53 70 76 64 84 Source: MOF, WB staff calculations. 2.16 Almost 90 percent of education sector spending is on current expenditures, mostly salaries, making the education sector budget highly rigid. Salaries represent nearly 75-80 percent of current expenditures although teacher salaries are one of the lowest in the public sector. The subsidies and transfers to other education entities, (such as Teacher’s Professional Development Center, Education Infrastructure Agency, and National Examination Center) finance implementation of education programs. Capital expenditures have been on the rise since 2008 given the increased focus of the government to address the existing investment needs in education infrastructure (Table 2.6). Table 2.6: State Budget Allocations to the MES by Economic Categories (In GEL million) 2006 2007 2008 2009 2010 2011 2012 2013 Total Expenditures 358 411 458 490 538 553 627 675 Current Expenditure 358 411 405 462 501 490 556 616 Compensation of employees & social contrib. 3 5 5 5 10 17 21 21 Other goods and services 1 2 12 12 38 48 62 48 Subsidies 0 0 58 56 40 31 23 21 Social provision 0 0 6 5 2 2 2 2 Other expenditure 354 404 324 383 411 391 446 522 of which: General education vouchers 187 201 269 300 325 316 326 410 Capital expenditures 1 0 49 28 36 62 71 59 Source: MOF. 2.17 Georgia follows the per capita financing method to ascertain its education sector budget. Schools receive direct transfers of funds from the MES in a transparent manner based on the number of students for a given year to cover salaries, utilities and routine maintenance costs (Table 2.7). The allocation formula was adjusted over time to address the problem of the so called “deficit schools.” Due to the initial design problems of the formula, small size schools did not receive adequate financing and required additional transfers on top of the per capita financing. Gradually, the mechanism has been refined and in 2010 substantial changes were introduced in the per student financing formula21. As a  21 There are different vouchers for basic compulsory and upper secondary education, there is a fixed component (base funding) on top of the per capita allocation, a different financing mechanism for small size schools (1-169 students) and schools with special needs students, and additional financing for consolidated schools with more than one school building and minority schools. 32  result, the share of deficit schools declined from almost 55 percent in 2010 to less than 4 percent of total schools in 2012 22. Table 2.7: Per Capita Finance Coefficients (GEL) Voucher Number of Number of Number of Students Grades 1-8 Grades 9-12 Base Funding Schools 2011/12 Students 2011/12 1-169 Calculated per National Curriculum 1211 98,170 170-190 405 486 47,000 75 13,462 191-205 400 480 45,000 71 14,068 206-225 395 474 43,000 57 12,297 226-530 390 468 39,000 401 138,647 531-735 385 462 32,000 102 63,728 736-1269 380 456 25,000 133 122,185 1270< 375 450 10,000 33 54,969 Private schools 300 300 0 229 51,371 Source: MES, USAID. Specificities of Preschool Education Expenditures 2.18 Parental contribution towards pre-school education was abolished in September 2013 and this is likely to increase public spending. Until 2013, the preschool educational institutions were co- financed to the extent of 30 percent by parental contributions collected directly by service providers. Vulnerable groups of population were however exempt from paying these contributions. In September 2013, the parliament abolished parental fees in all public kindergartens which significantly increased the enrolment rate from 46 percent in 2012 to 66 percent in 2013.23 While this was a welcome development, . In September 2013, the parliament abolished parental fees in all public kindergartens which significantly increased the enrolment rate from 46 percent in 2012 to 66 percent in 2013.24 The government could possibly use the underutilized general education facilities in some regions that are often vacant. 2.19 In the absence of a supervisory body, the cost per student, teacher salaries, and teacher qualifications vary widely across preschools in different municipalities. The per student cost was GEL 305 in Shida Kartli and GEL 1337 in Adjara in 2012, while the country average was GEL 698. In addition, each region has different number of vacancies, student-teacher ratios (STRs) and compensation schemes (Table 2.8). Compensation of teaching and non-teaching also varies significantly. For instance, the salaries of directors in urban preschool institutions range from GEL 150 to GEL 700 (excluding Tbilisi). For teachers the range is between GEL 70 and GEL 320. In addition, some regions successfully manage to recruit teachers with a teaching background and experience, while others— particularly in rural areas—face a shortage of skilled professionals.  22 Sources: MOF and Geostat. 23 Law on city of Tbilisi and law on self-governance were amended to reflect this change 24 Law on city of Tbilisi and law on self-governance. 33  Table 2.8: Preschool Indicators by Region, 2012 Vacancies (Per Non-Teaching Children 100 Spots) Teachers STR Staff Total Cost Unit Cost Tbilisi 43,084 4.3 4,200 10.3 2,197 31,550,000 732 Adjara 6,667 4.8 1,271 5.2 466 8,914,660 1,337 Guria 2,541 8.8 430 5.9 441 2,561,975 1,008 Imereti 15,696 -2.7 1,571 10.0 1,368 9,691,609 617 Kakheti 9,300 -9.8 1,205 7.7 962 4,530,870 487 Mtskheta 2,522 3.6 395 6.4 309 1,781,300 706 Racha 654 -11.8 142 4.6 112 659,461 1,008 Samegrelo 8,307 1.2 1,088 7.6 569 4,692,454 565 Samtskhe 2,213 10.9 263 8.4 166 1,448,833 655 Kvemo Kartli 6,894 23.8 668 10.3 449 4,479,376 650 Shida Kartli 5,149 25.3 580 8.9 502 1,571,997 305 Total 103,027 5.1 11,813 8.7 7,541 71,882,535 698 Source: MES. Note: Figures may vary slightly from those provided by the MOF. Vacancies are the number of places minus the number of children attending preschool divided by the number of places. 2.20 A low student-teacher ratio and a large number of non-teaching staff in preschools led to a high per student cost. At 8.7, the STR of preschool institutions in Georgia is considerably lower than OECD or EU 21 countries (at 12.2 and 11.8, respectively). As a consequence, regions like Racha, Adjara and Guria run very high costs per student. Salaries account for most of the expenditures in preschool education followed by food. Even though salaries are low in comparison with other professions in Georgia, this budget line item takes up the largest share of expenditures. Outside Tbilisi, between 54 percent and 75 percent of the expenses are on salaries. Spending on food ranges from 4 to 42 percent. Another efficiency gap is the result of a high share of non-teaching to teaching staff outside Tbilisi – some regions have nearly the same number of teaching and non-teaching staff. Specificities of General Education Expenditures 2.21 With an aging population, there has been a decline in the number of students and a consolidation in the number of schools—however, at 241 students, the average school size in Georgia remains small and drives up costs. Between 2006 and 2013, the number of schools has dropped by 15 percent, compensating the decline in students in the same period by 12 percent. While Tbilisi has relatively larger schools and accommodates almost one third of the total student population, in some regions the average school has less than 150 students and this drives up the costs. The United Nations (UN) population estimates indicate that the school age population in Georgia will decrease substantially in the medium term. If enrolment rates and the balance between public and private schools remain constant, students in public primary education (ages 6-12) will drop by 18 percent by 2030, while those in secondary education (ages 12-18) will fall by 12 percent (Figure 2.2).25 This suggests that the government will need to consider a further consolidation plan to reduce costs in the medium-term.  25 In fact, if any, the trend in the balance between public and private schools is towards the latter, which would reduce the number of pupils in public schools to an even greater extent. 34  Figure 2.2: Student Population Attending Public Figure 2.3: Public Expenditures Per Pupil Schools , 2002-30 (In percent of GDP) (No of students) 450,000 50 Elementary 45 (I-VI) 40 400,000 35 30 25 20 350,000 15 10 5 300,000 0 Russian Federation Lithuania Azerbaijan Moldova Romania Croatia Georgia Hungary Bulgaria Poland Latvia Estonia Armenia Slovak Republic Czech Republic Belarus Slovenia 250,000 200,000 CIS New Member States (EU) South Caucasu 150,000 2010 Unweighted group average 2010 Source: MES; UN Population Division; World Bank staff Source: World Bank, EdStats; MOF and Geostat, Georgia. calculations. Note: Figures for Belarus, the Slovak Republic, Croatia and Note: Estimates use implied population growth rate for relevant Azerbaijan account for expenditures per pupil across all age and assumes enrollment and share of the population served education levels. Romania’s figure is from 2009. by public system constant. 2.22 Despite the declining student population, spending per student as a percent of GDP per capita has not increased and remains below comparators. It grew from 15.0 percent to 15.5 percent from 2012 to 2013. However, comparator countries such as CIS and new EU member states spent substantially more per pupil in 2010 (Figure 2.3). Furthermore, in the EU and OECD countries, this figure ranges from 23 percent to 26 percent. 26 Table 2.9: Weekly Teaching Hours 2.23 There are too many teachers teaching too few hours suggesting that there is a glut in the education Average Number of system (Table 2.9). Based on an academic year of 40 Age Hours Teachers <20 4.3 7 weeks, the average yearly teaching hours sum up to less 20-29 13.8 5,795 than 600 while the average teacher in an OECD country 30-39 14.9 14,828 teaches 790 hours in primary, 709 in lower-secondary, and 40-59 15.2 34,590 664 in the upper-secondary level. In countries such as the 60-69 13.8 9,781 US and Chile, these figures go beyond 1000 hours per year 70-79 13.2 2,527 (25 per week). 27 Hence, there is room to cut the number of 80-89 12.0 260 90-99 11.5 5 teachers although this could be a politically sensitive issue. Source: MES. Note: Number of teachers in this table includes Specificities of Vocational Education and Training private and public schools. There are approximately Expenditures 60 thousand teachers in public schools and 7 thousand in private schools. Source: EMIS, 2.24 The VET system has seen a complete overhaul Georgia. with the adoption of the Law on Vocational Education in 2007. This law was subsequently amended and a VET strategy for 2009-12 was developed. This strategy included: (i) optimization and consolidation of vocational education institutions in 2008-09; (ii) introduction of a voucher financing system for Level I-III programs in May 2012; (iii) significant investments by the government for  26 OECD (2013), p. 177. 27 OECD, 2013, Table D 4.2, p. 402Ǥ 35  supporting the VET sector as a part of the 2012-16 program (nearly GEL 4 million); and (iv) the adoption of the National Qualifications Framework (NQF) in 2010. The investments were targeted at new school infrastructure, launch of three new vocational colleges, and the construction of the new technological university in Batumi. 2.25 The government’s 2013-2020 VET strategy aims to improve the financing and targeting of VET programs. It includes three components: (i) differentiated vouchers, which take into account the cost of individual study programs replacing the previously flat vouchers of GEL 1,000 – the new vouchers range between GEL 400-2250 and cover teacher salaries and learning material; (ii) program budget transferred to VET institutions to cover administrative costs and utilities; and (iii) targeted program financing to support implementation of the Ministry’s different priority initiatives. Public VET institutions are financed through a voucher system while private operators charge tuition fees. The voucher scheme was introduced by the government with support from GIZ in May 2012. The eligibility for vouchers is determined by the student’s socio economic background and educational standing. 2.26 The government has also eased the admission requirements for certain VET programs to make them more accessible. The requirement to have secondary school graduation diploma for transferring to levels IV and V of VET programs was abolished unless required by the specific study programs. Similarly, the requirement to take aptitude tests as part of the University Entrance Examinations was also eliminated for students entering VET education. To enter Level I of the VET program, students are required to pass a centralized professional skills test administered by NAEC. The transition from one level to another is regulated by the VET institutions. 103 units were created in higher education institutions, state colleges and resource centers including the 3 existing Information Centers to support the registration of VET applicants across the country. However, the link between VET programs and higher education remains weak. 2.27 High dropout rates from VET programs, however, leads to a waste of resources. Over 1,000 VET students dropped out in 2013. This includes students that dropped-out and those who were unable to complete the study 28 leading to a wastage of nearly GEL2.0 million, approximately 39 percent of the vocational education budget for 2013. Lack of maturity and information about the course and the profession were the main contributors to the high drop-out rate. Before September 2013, VET institutions were not required to reimburse the remaining funding to the MES in case students dropped out. This was changed in 2013 and currently, payments to the institutions are made on a quarterly basis instead of annual or semi-annual transfers. This allows the ministry to suspend the next quarter’s funding of the student as soon as he drops-out. This should allow for substantial savings and efficient reallocations. D. EDUCATIONAL ATTAINMENT AND PERFORMANCE Preschool Education 2.28 Georgia's net preschool enrolment, at 66 percent in 2013, is relatively low compared to the new EU member states and most of the CIS countries and far from the universal coverage in the developed EU or OECD countries (Figure 2.4). Preschool coverage in Georgia dropped sharply in the early 1990s to about 28 percent and started to recover only a decade later and reached 40 percent in 2006 and 46 percent in 2012. Preschool education was made free in 2013 which led to a spike in the enrolment rates to 66 percent indicating that affordability was a major constraint for parents. Another factor responsible for the low enrolment rate is the lack of awareness on the part of parents regarding the  28 State Audit Office of Georgia, Performance audit: Professional education in Georgia, 30.12.2013, p. 42. 36  importance of early childhood interventions in boosting non-cognitive skills.29 The gross enrolment rates show disparities among the underrepresented groups. The gap between the enrolment rates for Georgian and Azerbaijani population has grown from 20 percent in 2006 to 47 percent in 2012; the gap between the richest and the poorest quintiles grew from 19 percent to 36 percent in the same period and the gap between urban and rural children rose from 11 to 23 percent 30. Figure 2.4: Net Enrolment Rate in Preschool, 2011 Figure 2.5: Breakdown of Unit Costs by Region (In percent) (In GEL) 100 1400 90 80 1200 70 60 1000 50 40 800 30 20 10 600 0 400 Russian Federation Slovenia Armenia Belarus Slovakia Croatia Latvia Lithuania Poland Hungary Azerbaijan Ukraine Bulgaria Georgia Romania Czech Republic Moldova Estonia 200 0 CIS South Caucasus New Member States (EU) Other Trainings Learning materials/toys Food 2010/2011 Group Average Running costs Capital Expenses Source: TransMonee2013; ISET Policy Institute (2012) Source: Geostat (2012). Note: Net enrolment rate of children aged 3-6 (or 3-5, depending on the country). Figures for 2011, except for Georgia (2012) and Slovakia (2010). 2.29 The use of resources in preschool education suggests there are efficiency gains to be reaped in Georgia. At 8.7, the student-teacher ratio (STR) of preschool institutions in Georgia is particularly low. OECD and EU 21 averages are 12.2 and 11.8, respectively. Mainly as a consequence of their low STRs, regions like Racha, Adjara and Guria have very high unit costs. (Figure 2.5). Even though salaries are low in comparison with other professions in Georgia this budget line takes up a big share of the expenditures. Outside Tbilisi, between 54 percent and 75 percent of the expenses are salaries. Spending on food ranges from 14 to 42 percent. Moreover, the high share of non-teaching to teaching staff outside Tbilisi – some regions have nearly the same number of teaching and non-teaching staff – suggests preschool facilities can be more efficiently managed. 2.30 Georgia’s performance in PISA 2009+ indicates that having attended at least one year of preschool has a positive effect on academic achievement.31 Students who attended preschool attain scores significantly higher than those who did not. Nevertheless, attendance of preschool could hide other important determinants of academic performance, such as socio-economic status, which usually have a strong impact on achievement. However, even after controlling for a myriad of factors such as economic, social and cultural status, age, grade, gender, location (urban, rural and mountainous areas) and highest level of parental education, children that attended at least a year of preschool education had scores between 7 and 10 points higher than their peers on reading, mathematics and science (Table 2.10).  29 See Heckman and Masterov (2007), Heckman (2006), Cunha and Heckman (2006) and Carneiro and Heckman (2003) among others. ͵Ͳ  ǣ‘’”‡Š‡•‹˜‡‘•–‹‰ƒ† ‹ƒ…‡–”ƒ–‡‰‹‡•ˆ‘”–Š‡ƒ”Ž›‡ƒ”‹‰›•–‡‹ ‡‘”‰‹ƒǡʹͲͳ͵ 31 Sixty-four economies originally participated in PISA 2009. Ten additional partner participants, who were unable to participate within the PISA 2009 project timeframe, participated in the PISA 2009 study on a reduced and delayed timeline in 2010. This is known as the PISA 2009+ project. 37  Table 2.10: PISA Performance by Preschool Attendance Economic, Social Attend preschool Reading Mathematics Science and Cultural Status No 355.70 363.88 358.79 -0.68 Yes, one year or less 379.35 393.89 383.80 -0.19 Yes, more than one year 397.79 396.30 388.82 0.01 Source: WB staff calculations; PISA (OECD). General Education 2.31 Gross enrolment rates (GER) show that access to primary and lower secondary education is nearly universal but the enrolment rate drops in upper secondary education. At 74 percent enrolment in the upper secondary level, Georgia trails behind comparator countries (Figure 2.6) 32. In the primary and lower secondary level, the enrolment rate for females is slightly below that of males (102.8 percent vs. 104.8 percent in primary and 99.9 percent vs. 101.6 percent in lower secondary). However, female GER is 77.3 percent in the upper secondary level while for males it is only 71.3 percent. Survival rates at each education level are high. At the primary and basic education level, enrollment rates do not show significant differences between different disadvantaged groups or between ethnic groups. However, such differences begin to appear at the upper-secondary level. 85 percent of ethnic Georgians aged 15 to 17 are enrolled in schools but only 59 percent of the same age group among ethnic minorities are in secondary school. 33 Figure 2.6: Gross Enrolment Rates by Level of Education (In percent) Source: EMIS, 2012. 2.32 There is a statistically significant positive association—albeit small in magnitude—between larger school size (number of students) and scores in all competencies in the PISA assessment. Nearly 39 percent of the schools in Georgia have less than 100 children and many of them have very low student-teacher ratios (Table 2.11). These numbers indicate a suboptimal allocation of resources. The positive relationship between PISA scores and a larger school size is an indication that students benefit  32 Decline in enrollments at the upper secondary level may be explained by past policies which made school principals accountable for performance of students in the school leaving exams. This policy prompted principals to discourage low performing students to continue education at the upper secondary level. 33 Education Access and Equity in the Caucasus Region, EPPM (2011). 38  from interactions with each other. In addition, there is a possibility that better teachers are allocated to larger schools so as to ensure that more students are exposed to good teachers. In addition, student performance indicates that higher STRs are not associated with poor academic performance in Georgia. Hence, having larger schools and more pupils per teacher – up to a certain level – will not worsen student achievement. Table 2.11: Number of Schools, Students and Teachers by Class Size, 2012 School Size Number Share of Number Share Number Share Student/ (Student of Total of of Total of of Total Teacher Class size Number) Schools (%) Students (%) Teachers (%) Ratio 3-4 1-50 410 18 12,287 2 5,889 9 2.0 4-5 51-100 475 21 35,426 6 9,509 14 3.7 <15-18 101-320 1033 39 164,796 29 25,067 38 6.5 18-30 320 -1000 449 19 252,453 45 21,563 32 11.7 30-35 1000 (and above) 67 3 93,175 16 5,946 9) 15.6 Total 2,322 558, 137 65, 666 8.5 Source: EMIS, 2012 2.33 The quality of education in Georgia is monitored through national and international student assessments but participation in the latter has not been regular. The national assessments have not been carried out on a regular basis. Moreover, due to the different methodologies used in each of the assessments, it has not been possible to compare learning trends over years. International assessments are the most useful source for judging the quality of learning outcomes. Georgia joined TIMMS and PIRLS international assessments in 2006 and a PISA assessment was conducted only once in 2009. 2.34 Georgia’s performance on the reading scale of PISA 2009 was below what would be expected given its income level. PISA gauges skills and competencies as opposed to bookish knowledge of a subject. As in other education systems of the former Soviet Union countries, students in Georgia are encouraged to learn things by rote and are suitably rewarded on this parameter. The gap between Georgia and the OECD countries on the reading scale is approximately three years of schooling (each year accounts for 40 points) 34. In addition, only between 30 percent and 40 percent (depending on the subject) of Georgian 15 year-olds scored Level II or above, which is a threshold usually referred to as functional literacy (Figure 2.8). Finally, very few students score Level IV and above. In order to attain this level, students need to demonstrate sophisticated reasoning and complex thinking, such as applying knowledge in different contexts.  ͵ͶOECD average in reading literacy rate for PISA 2009+ is 493 points; 39  Figure 2.7: PISA Reading Score and GDP Per Figure 2.8: Share of Students Scoring Level II or Capita Above and 4 or Above in PISA (2009+) Source: PISA 2009+ (OECD); World Development Indicators, World Bank. 2.35 Two important reasons why Georgia trails behind other countries in PISA performance relates to learning strategies and teaching practices. After controlling for socio-economic status and parental education, an index that compiles a set of measures of learning strategies explains the reading scores gap between Georgia and the new EU member states to the extent of 26 points – equivalent to more than half year of studies. Teaching practices, after controlling for socio-economic status, explains for 2 points of the gap with the new EU member states, but 9 points of the difference with respect to the OECD. Therefore teachers, who are responsible for teaching practices and helping children learn, have an important role in improving education quality in Georgia. It is important to highlight that this has critical consequences as the literature on education and economics of education has underscored the strong causal relationship between good teachers and both learning outcomes and economic returns.35 2.36 Teaching is however not an attractive career choice in Georgia for highly skilled workers because of the low pay scale. In addition, the social prestige of the profession seems to be low according to the research undertaken by the Social Research and Analysis Institute in 2012. The average age of teachers in Georgia is 48 given that less than 19 percent of the teachers are 35 or younger, while almost 20 percent are 60 and older. In order to make the profession more attractive, the government introduced changes in the teacher compensation scheme in January 2013. 36 The base salary of teachers rose from 245 GEL, set in 2009, to 305 GEL. The change in teachers’ salaries is expected to increase the education budget by approximately 15 percent, or 0.4 percent of GDP. According to the wage scheme, more experienced teachers do not get paid substantially more than inexperienced teachers. Having more qualifications (school attainment), however, has a more significant impact on wages. Finally, teachers that teach Georgian culture, language or history; multi-grade classes; head teachers and certified teachers (ICT and subject) get additional income (Table 2.12). However, teacher certification results have been below expectations with only 25 percent of the entire teaching cohort being certified since 2011. Therefore, more extensive and well-tailored professional development of teachers is needed.  35 See Hanushek & Rivkin, 2010; Hanushek, 2011; Rivkin, Hanushek & Kain, 2005. 36 MES Decree No.4/N dated December 29, 2013. 40  Table 2.12: Teacher Pay Scheme (In GEL per month) Less than 5 Between 5 and More than 10 Qualification years 10 years years General education 290 303 308 Vocational education 325 339 343 BA 360 374 378 MA 395 409 413 Masters 430 444 448 Additions to the standard pay Multi-grade 31 Georgian culture, language, history 31 Head teacher 76 Certification 75 ICT 125 Certification (excellent scores) & ICT 1000 Source: MES. Vocational Education and Training (VET) 2.37 Enrolment in VET programs increased substantially during 2010-12 with significant public and private sector participation in the delivery of these programs. The government invested heavily in the required infrastructure for delivering VET programs starting in 2009. At the same time, several private providers were encouraged to deliver VET services more actively. As a result, the quality of the programs improved considerably and the demand has been on the rise. The enrollment doubled between 2010 and 2011 and increased further in 2012 (Figure 2.9). This was a substantial increase from the 2009 level when enrolment has dropped to nearly zero because of the closure of many vocational education institutions as a result of expenditure optimization by the government. Figure 2.9: Trends In Enrolment and Number of Vocational Education Institutions (In real terms) Source: Geostat for 2001-2009; MES for 2010-2013. 2.38 Surveys indicate that only a fourth of the graduates from VET institutions are able to find a job in the area in which they have educational qualifications.37 Therefore, there is room to make  37 GIZ, 2011 41  substantial improvements in the curriculum and the quality of teachers. The education quality enhancement center monitors performance of the VET institutions through authorization and accreditation (once in five years), annual self-assessment reports by the institutions and visits to the institutions. Currently, work is in progress to update and enhance over 200 professional standards with support from the EU and this will help strengthen the VET system. E. THE WAY FORWARD 2.39 The government’s new socio-economic development strategy emphasizes high quality of education to improve human capital. Strategic priorities include increasing labor market relevance of educational programs to meet workforce requirements, increasing access to preschool education, improving quality of general education, emphasizing vocational training, and enhancing the attractiveness of the teaching profession. These priorities are well-aligned since unemployment among higher education graduates in Georgia, particularly women, is extremely high reflecting the inadequacies of the current education system. In addition, businesses complain that job seekers do not have adequate skills, in large part because the existing educational curriculum is not in line with the demands of the private sector. 2.40 The government’s medium-term expenditure framework projects an increase in the education sector budget but this increase needs to be complemented with efficiency gains. Georgia’s public expenditure on education is comparatively low and an increase in the education budget would certainly be a step in the right direction. However, there are several complementary reforms that will be needed to effectively use the increased resources. A downsizing of schools and the number of teachers would help. Similarly, improvements in teacher quality through trainings and relevant certifications would be beneficial. Greater engagement of the private sector in determining educational curricula would help increase the relevance of the education system. And finally, 2.41 Larger class sizes and improved teacher quality will go a long way in improving education sector outcomes. The student teacher ratio in preschools and the number of work hours of teachers need to be brought closer to OECD standards. There is currently a glut in the system and this optimization will result in cost savings which could help strengthen other areas of the education system. The MES has recently (in March 2014) instituted professional standards for teachers which defines professional characteristics, knowledge and skill requirements pertaining to the teaching environment, teaching/learning process, and professional development for each category of teachers. These standards will help address the issue of poor teacher quality. 2.42 Capacity development within the MES would enable it to effectively drive and implement this huge agenda. The knowledge sharing and alignment with international practices would be beneficial for the Government to consider. At the same time support from the development partners needs to fit into the comprehensive education development strategy that is being currently drafted with the Bank support. The MOF’s guidance in developing strong program budgeting practices with attributable and measurable performance indicators will help measure the progress and ensure funding is available to support it. It is also important to keep in mind that impact of reforms on the ground and improvements in education outcomes will be visible only in medium to long-run. 42  CHAPTER 3. STATE OWNED ENTERPRISES: GOVERNANCE AND FISCAL PERFORMANCE A. INTRODUCTION 3.1. According to some measures, the state owned enterprise sector in Georgia is relatively small and contributed to 7 percent of gross GDP in 2012 and accounted for less than 10 percent of total formal employment. The privatization program that was underway during 2004-10 resulted in a significant reduction in public enterprise employment from 144,000 in 2004 to 60,000 in 2012. Although the share of SOEs is small in terms of output, their contribution to investment in fixed capital is oversized at 24 percent of total corporate investment in 2012. This is essentially because of the sectors in which the large SOEs operate–energy and transport–and the large infrastructure requirements in these sectors. Some of the government’s capital expenditures in these areas are undertaken by the large SOEs. Majority of the remaining SOEs operate in the health sector, municipal services, agriculture, tourism and manufacturing. 3.2. No single government entity has a comprehensive picture of SOE performance. More than 400 SOEs operate in Georgia under different institutional arrangements and ownership structures. However, with the exception of the five largest SOEs, information on them is scant. According to the law on state property in Georgia, the MOESD is primarily responsible for the management and supervision of SOEs but in practice, this oversight function is spread across several ministries. As a result, Parliament does not receive any information on these companies as part of the budget preparation and execution process. Transfers from SOEs to the budget through dividends varied widely over the years based on performance of the SOEs and their negotiations with the government. While there are no explicit subsidies to the SOEs, the government has assumed responsibility for some of their liabilities. 3.3. The five largest SOEs in Georgia comprise 80-90 percent of the total assets of SOEs and are grouped under the extra budgetary Partnership Fund. Their combined profit amounted to GEL 63.5 million in 2012 or about 0.2 percent of GDP – and four out of the five were profitable. Based on negotiations with the government, these companies paid dividends of GEL 6.7 million in 2012 to the Fund. Their consolidated financial data also showed a positive return on assets (ROA) and return on equity (ROE) in 2012. The only exception was the Georgian State Electrosystem (GSE) which was not profitable. Sensitivity analysis impacting the operating income and other costs and income of the SOEs reaffirms the robustness of their operations. 3.4. The government however undertakes significant quasi-fiscal operations through these large SOEs which potentially understate the fiscal deficit and overall debt. Exact costs are not available but the government provides subsidized utilities to the population (gas, electricity, transport) through its 5 largest SOEs for which it does not compensate them. In addition, these SOEs undertake investment projects on behalf of the government, also without any compensation. The government also imposes other costs on SOEs like wage increases. Revenues and costs related to carrying government mandates need to be mainstreamed into the budget and this will potentially increase the fiscal deficit. Additional costs could surface from SOEs for which data is not currently available, especially those under the MOESD, several of which are bankrupt. Some of the largest SOEs are highly indebted and assumption of SOE debt would also increase government debt. 3.5. This chapter reviews the fiscal implications of the operations of the Georgian SOEs, with an emphasis on the largest companies. The analysis in this chapter leads to the following measures for the government to consider: 43  x Short-term: Establish an inventory of SOEs to get a consolidated picture of the government’s fiscal position. x Short-term: Establish a clear dividend policy for SOEs. x Medium term: Mainstream quasi-fiscal operations like provision of subsidized utility services and investment projects undertaken by SOEs into the budget to increase transparency. 3.6. The rest of the chapter is organized as follows. Section B provides an overview of the legal and institutional framework of SOEs. The next section focuses on their operational and fiscal implications while Section D discusses quasi-fiscal operations and contingent liabilities. Section E gives the way forward. B. LEGAL AND ORGANIZATIONAL FRAMEWORK OF SOES 3.7. More than 400 SOEs operate in Georgia under different institutional arrangements and ownership structures. The assets of the five largest SOEs were transferred to the Partnership Fund (PF) which was established in 2011. These five companies include LLC Georgian Railway, JSC Georgian Oil and Gas Corporation, JSC Georgian State Electrosystems, JSC Electricity System Commercial Operator, and JSC Telasi. In addition, the PF includes other smaller SOEs (Table 3.1). Another group of 380 SOEs are under the supervision of the MOESD. The majority of these SOEs do not operate and some are bankrupt. According to MOESD estimates, only 70 of these 380 firms are viable and they operate in the health, navigation, and manufacturing sectors. A third group consists of enterprises which operate under the supervision of line ministries including Ministry of Energy, Ministry of Agriculture, and Ministry of Labor Health and Social Affairs (MOLHSA). 3.8. According to the law on state property in Georgia, the MOESD is primarily responsible for the management and supervision of SOEs but in practice, this oversight function is spread across several ministries. The MOESD is the sole body within the government responsible for the development and implementation of policies for the management of state property and its disposal. The National Agency of State Property (NASP) was established in 2012 under the MOESD and is responsible for actively managing state owned enterprises and those in which the state has a minority stake.38 It is tasked with: (i) ensuring the flow of dividends; (ii) coordinating management of state owned shares; and (iii) promoting the privatization process. However, the law allows the transfer of assets and management of state ownership to bodies other than the NASP based on certain procedures implemented by MOESD. If the state share of the transferred enterprise is larger than 50 percent, then presidential authorization is required for the transfer. Some of the enterprises currently under the line ministries were transferred in this manner and the Partnership Fund was also created through this channel. Hence, there are several entities which are responsible for oversight of the SOE sector and no institution at the center of the government has a comprehensive picture of the performance of SOEs (Figure 3.1).  38 Formerly known as Enterprise Management Agency (EMA). 44  Figure 3.1: Institutional and Governance Structure of SOEs in Georgia Source: Staff constructed based on the information provided by the authorities. 3.9. The Partnership Fund was established to promote private sector investments in strategic sectors. The Fund is a joint stock company (JSC), 100 percent owned by the State. It is governed by the Georgian Civil and Commercial Legislation (Joint Stock Company Law), as well as its founding law and charter. The Government sought to boost economic growth, generate employment, and raise the country’s welfare by attracting foreign investments in energy, agriculture, manufacturing and real estate. According to its founding law and charter, the PF is also tasked with finalizing projects that have run into implementation problems and privatizing completed projects. The Fund may contribute to the financing of private sector projects through minority equity participation, lending or guarantees. It can also initiate feasibility studies for projects that can attract foreign investors. In these projects, the respective ministry acts as the regulator. Its operations are undertaken without explicit state guarantees. According to the financial statement of PF, gross assets of these public companies amount to GEL 4.9 billion (19 percent of GDP) as of end 2012 and account for 80-90 percent of the public enterprise sector in terms of value of assets. The net assets though are GEL 2.3 billion given large external liabilities. 45  Table 3.1: Georgia Partnership Fund Portfolio (GEL million, share in percent) Total Assets Ownership/ (GEL Voting Name Million) (%) Principal Activities Partnership Fund (consolidated) 4,910.4 100 Oil and Gas sale, extraction and exploration Georgian Oil and Gas Corporation JSC 878.3 100 and rent of pipelines Namakhvani JSC 100 Construction and operation of a power plant Georgian Railway JSC 2,831.5 100 Railroad transportation Georgian Railway Property Management LLC 100 Property management and development Georgian Railway Construction JSC 100 Container transportation Barjomi Bakuriani Construction JSC 100 Passenger transportation Rail Parking LLC 100 Parking service Georgian State Electrosystem JSC 1,024.3 100 Electricity dispatching and transmission Energo Trans LLC 100 Electricity transmission Karcal Energy JSC 100 Electricity transmission Electricity System Commercial Operator (ESCO) JSC 33.8 100 Sale and purchase of electricity Purchase and distribution of electric power to Telasi JSC 287.4 24.53 industrial and customers in Tbilisi Other Construction and operation of a hydro power Nenskra JSC 100 plant Tbilisi Logistic Center LLC 100 Food services Fruit and Vegetable Export Company LLC 100 Export of fruit and vegetables Construction and leasing out of shopping Port Lazika LLC 100 mall Gardabani LLC 100 Construction and operation of a power plant Borjomi Likani International JSC 50 Hotel operation Source: Partnership Fund. 3.10. Data on the remaining SOEs is scant. The financial and non-financial information for the SOEs under the MOESD and line ministries is in the works and a meaningful assessment of the SOE sector will require the availability of such a database. With the exception of the five largest SOEs, data on the rest are of poor quality and constrains the analysis in this chapter. C. OPERATIONAL AND FISCAL IMPLICATIONS OF SOES 3.11. SOEs – both at the state and municipal level -- contributed to 7 percent to total output in 2012 and accounted for less than 10 percent of total formal employment. The statistical business survey shows that the number of employees in the SOE sector has been shrinking due to the large scale privatization process which was underway during 2005-08 (Figure 3.2). Employment fell from about 144,000 in 2004 to 64,000 in 2011 and to 60,000 in 2012. 3.12. Although the share of SOEs is small in terms of output, their contribution to investments in fixed capital stood at 24 percent of total investments by firms in 2012. This is essentially because of the sectors in which the large SOEs operate–energy and transport–and the large infrastructure requirements in these sectors. Some of the government’s capital expenditures in these areas are undertaken by the large SOEs under the PF. Majority of the remaining SOEs operate in the health sector, municipal services, agriculture, tourism and manufacturing. 3.13. Transfers from SOEs to the budget—largely dividends and interest payments—vary widely because dividend payouts are a subject of case-by-case negotiations between the SOE and the government. On average, transfers amounted to 0.3 percent of GDP a year. In 2007, it was GEL 30.7 million, GEL 87 million in 2008, GEL 42.5 million in 2009, GEL 11.8 million 46  in 2011, GEL 129 million in 2012 and GEL 46.5 million in 2013. SOEs are allowed to retain the profits if they are able to justify their investment needs. In addition to dividends, SOEs also repay loans and pay interest on loans borrowed from the government. This component has averaged about GEL 10 million (0.3 percent of total revenues) a year during 2007-12. Figure 3.2: Georgia General Government Privatization 3.14. There are no explicit subsidies to the Proceeds, 2001-14 SOEs, but the government has assumed (In percent of GDP) responsibility for some of the liabilities of 6.0% these enterprises. The government has made 5.2%5.2% loan repayments on behalf of some SOEs despite 5.0% there being no explicit state guarantee. The SOE 4.0% 3.8% 3.7% “Mechanizatori” borrowed GEL 80 million from a commercial bank to purchase agricultural 3.0% equipment in 2009. Although there was no 2.0% 1.6% explicit guarantee, this loan was fully repaid 1.2%1.1% 1.1% 1.0% 0.7% 0.5%0.3% through budgetary resources. 0.4% 0.1%0.2% 0.0% 3.15. The Partnership fund is chaired by the Prime Minister. It is governed by the Supervisory Board which comprises Source: Ministry of Finance. representatives of the Government of Georgia Note: Privatization defined as a disposal of non-financial assets and of the private sector and is chaired by the (GFS2001). Prime Minister.39 It is the main decision making body of the PF and reviews and approves project financing proposals and recommendations made by the PF’s Investment Board. There is also a Risk Management Committee which advises the management on projects risks. The Supervisory Board is vested with the authority to monitor the progress and implementation of approved projects. Day-to-day operations of the PF are run by an executive body (which is approved by the chairman of the Supervisory Board) that is headed by the Chief Executive who is tasked with directly overseeing and managing the implementation of approved projects. The PF has a separate Investment Board which consists of experts from the private sector. The Investment Board assesses project opportunities within Georgia and prepares project recommendations for consideration by the Supervisory Board. The PF was set up with the following assets: 100 percent of the outstanding share capital of Georgian Railway LLC, 100 percent of the outstanding share capital of the Georgian Oil and Gas Corporation JSC, 100 percent of the outstanding share capital of the Electricity System Commercial Operator JSC, 100 percent of the outstanding share capital of the Georgian State Electro System JSC, and 24 percent of the outstanding share capital of Telasi JSC (Box 3.1). 3.16. The Partnership Fund follows specific investment guidelines with the ultimate objective of attracting foreign investments in energy, agriculture, manufacturing and real estate. The Fund can support projects which include an element of private sector financing as long as the private sector parties involved in the project present a bank guarantee for an amount that is equal to at least 10 percent of the planned project investment requirement. The Fund can support projects in the agricultural sector with a minimum value of GEL 5 million. For projects in non-agricultural sectors the minimum project value should be GEL 30 million. Projects in the services sector are outside the Fund’s scope and will not be eligible for Fund support. Financial participation of the Partnership Fund is limited to 49 per cent of the project's total capital, and up to 25 per cent of the total project cost. The rest of the capital will need to be provided by the strategic investor. The newly created (2013) Co-investment fund has similar objectives and the difference between these two funds is not very clear.  39 The Board’s membership is dominated by the government and this is not in line with international best practices. 47  Box 3.1: Information on Large Enterprises under the PF JSC Georgian Oil and Gas Corporation (GOGC) is one of the largest companies in Georgia. It is the owner of the main pipeline and its revenues stood at GEL 299.5 million in 2011. The company was established in March 2006 by the order of the Minister of Economic Development of Georgia. The government-owned shares of JSC “Georgian International Oil Corporation”, JSC “Georgian International Gas Corporation” and JSC “Saknavtobi” (Georgian Oil) were accumulated in GOGC’s authorized capital. In September 2011 the corporation changed its legal form into a joint stock company. In May 2012 GOGC issued US$250m Eurobonds with a 6.875 percent coupon. They have 300 about employees. JSC Georgian State Electrosystem (GSE) provides transmission and exclusive dispatch services to about 50 eligible companies in Georgia. GSE was formed in 2002 from the merger of JSC “Electrogadatsema” and JSC “Electrodispetcherizatsia”. These two entities were themselves established in 1999 and 2000 respectively to own the transmission assets of the state utility Sakenergo and to manage the national dispatch center. Based on the transmission and dispatch tariffs established by the independent regulator (GNERC-Georgian National Energy and Water Regulatory Commission) GSE carries out technical control over the entire power system to ensure the availability of the system for uninterrupted and reliable power supply; and transfers, without the right of purchase or sale, the electricity imported or generated in Georgia to distribution companies, direct customers or the power systems of neighboring countries. In 2012 company generated revenues in the amount of GEL 54.1 million and employed more than thousand people all over the country. One of its subsidiaries, Energo Trans, is in electricity transmission business, and its operations in part are related to power purchase agreements (PPAs) and carry some risks. JSC Electricity System Commercial Operator (ESCO) started operations in the power sector of Georgia on September 1, 2006. The Commercial Operator represents one of the liquid and important links of the energy system value chain. With the contribution of ESCO capacities Georgia is an electricity exporting country and electricity energy sector has high potential for economic development and growth. JSC "Telasi" was founded in 1993. In 1998 AES Silk Road Holdings B.V., AES Corporation purchased 75 percent of the shares of Joint Stock Company "Telasi". In August 2003, AES Corporation, the ultimate owner, sold its 100 percent shareholding in Silk Road Holdings B.V to a Finnish company JSC RAO Nordic OY (INTER RAO Holdings B.V.) ultimately owned by CJSC INTER RAO UES (OJSC INTER RAO UES). The Partnership Fund owns the rest of the shares. The company’s main business activities are: purchase and sale of electric power; maintenance and operation of electric grids; power transit services; technical service of customers; administration of unified system of power and water supply, as well as refuse collection in the City of Tbilisi. Source: Partnership Fund 3.17. In 2012, these 5 largest SOEs generated profits of GEL 63.5 million and issued dividends of GEL 6.7 million to the Partnership Fund which is outside the state budget. The financial statements of these enterprises are externally audited following International Financial Reporting Standards (IFRS). This makes it possible to identify and quantify risks under certain assumptions. The consolidated financial data shows a positive Return on Assets (ROA) and ROE. The only exception is GSE which was not profitable in 2012. Individual performance of SOEs shows that with the exception of ESCO, they are all highly indebted, especially Telasi (Table 3.2). 3.18. A simple simulation shows that even with a 60 percent decline in operating income, the group of 5 largest SOEs will continue to be profitable. Operations of SOEs may be impacted by shocks to its operational performance before: (i) interest income and interest cost; (ii) other financial charges gains and losses; and (iii) taxes. Given the very limited data on these SOEs, simple simulations were undertaken to demonstrate the impact of a decline in operating income. Figure 3.3 (a) shows that the SOEs will continue to be profitable even if their operating incomes decline by 60 percent. This suggests that there is a high likelihood that they will remain profitable in the short- to medium-term. 48  Table 3.2: Financial Information on the SOEs, 2012 (GEL million) GOGC GR ESCO GSE Telasi Revenues 299 469 172 54 232 Retained earnings 100 449 5 255 NA Long term debt/equity ratio (percent) 104 75 No debt 193 324 Source: Audited financial statements of companies. 3.19. A similar simulation which reduces financial income and raises financial costs shows limited impacts on profitability. Financial income and financial costs account for a significant portion of the income of SOEs. This includes interest income and interest costs as well as foreign exchange gains and losses. Figure 3.3(b) shows the impact of a decline in their financial income, mainly interest income while Figure 3.3(c) considers the impact of a sharp rise in interest rates and depreciation of the local currency against the US dollar. Here again, the impact on profitability of these five SOEs is limited. Figure 3.3: Sensitivity Analysis of Operating Activities on Profit (Million GEL) a. Change in profit based on operating activities b. Change in profit based on financial income 70 80 (Profit of the PF, million GEL) 60 70 60 (profit of the PF , Million GEL) 50 50 40 40 30 30 20 20 10 10 0 0 -10 Source: PF 2012 Consolidated Financial Statements and own calculation F&KDQJHLQSURIÕWEDVHGRQIÕQDQFLDOFRVW 3.20. The above sensitivities as well as an 80 analysis of individual enterprises indicate that the five largest SOEs in Georgia, taken as a (Profit o f the PF, Million GEL) 70 group, are being run profitably. A similar 60 sensitivity analysis for individual enterprises gives 50 the same results with the exception of the 40 Georgian State Electrosystem (GSE). 30 20 3.21. However, certain components in the charter of the partnership fund like support to 10 unfinished projects and an explicit objective to 0 generate employment could have a negative 10% 20% 30% 40% 50% impact on profitability in the future. The increase increase increase increase increase original purpose of the PF was to provide equity Source: PF 2012 Consolidated Financial Statements and own and attract foreign investments with no recourse to calculation. government financing. However, a review of its 49  Charter reveals that the Fund has a mandate to promote employment and support unfinished projects. It is not clear if these objectives are in line with commercial considerations. Since these objectives are in its Charter, the PF could come under pressure to support non-commercial projects which could impact profitability. The Government may wish to remove these objectives from the PF’s charter. 3.22. The Government is now considering reorganizing the PF into a development bank which raises some concerns. The Government of Georgia has drafted and submitted to the parliament a law that would create a development bank by transforming the existing Partnership Fund. It is expected to be a legal entity of private law40 with the supervisory board managing the development bank and the prime minister being responsible for nominating the president of the bank. This new entity is expected to issue loans (through commercial banks) and guarantees. There could be other operations it could potentially be involved in depending on the government’s decisions. However, it will not receive deposits from the public. Also, the creation of this development bank has not been very well justified in terms of the market failure it will be addressing. D. QUASI FISCAL OPERATIONS AND CONTINGENT LIABILITIES 3.23. Quasi-fiscal activities refer to operations that result in a net transfer of public resources to the private sector through non-budget channels. Such activities distort the picture regarding the government’s true fiscal position and size, they may generate contingent liabilities, and may have undesirable redistributive effects. The identification of quasi-fiscal activities assumes that a clear distinction can be made between the transactions of state-owned enterprises that are undertaken on a commercial basis, and those whose goal is to provide a public service rather than to earn a profit. Determining whether something is a quasi-fiscal activity also assumes that there is a general commercial rate of profit and so transactions that do not achieve this rate may be identified. However, this is difficult in practice. The following could count as quasi-fiscal activities: x Charging less than commercial prices, where state-owned enterprises may provide, for example, electricity at a subsidized price to some or all consumers. x Provision of noncommercial services (e.g., social services), where state-owned companies may provide some services at less than full cost; for example, universities may not charge the full cost of tuition. x Pricing for budget revenue purposes, where state-owned enterprises may be in a monopoly position and so may be able to charge prices above what a competitive market might establish to raise revenue for the government. x Paying above commercial prices to suppliers, where local suppliers may be paid above the market rate as a form of protection for their industry. 3.24. Through the five SOEs under the Partnership Fund, the government undertakes quasi- fiscal operations which potentially understate the actual fiscal deficit. While exact costs are not available, there is a high likelihood that the cost imposed by the government on the five largest SOEs through quasi-fiscal operations exceeds the dividend that it would have otherwise received. Some examples are provided below. x The Georgian Oil and Gas Corporation (GOGC) supplies gas to households at less than commercial prices (costs GEL 29 million). The GOGC also built a high voltage line and transferred it to the government which cost GEL 36 million.  40 Operates as private company and may not be a subject of the scrutiny that public entities usually go through. 50  x At the request of the government, Georgian Railways (GR) increased the salary of its employees by 25 percent (GEL 27 million). In addition, GR offers lower railway passenger tariffs for specific routes and to vulnerable groups for which it receives no compensation. x GR is also implementing the Tbilisi by-pass project which is a government undertaking and on which it has already spent US$270 million. It halted the project in 2013 and is now reconsidering its economic and social rationale before moving forward. The initial feasibility of the project produced negative results but implementation was still undertaken. A new feasibility study is underway but this is not expected to be available before 2015. x The Georgian State Electrosystem provides electricity to the Abkhazia region for which it does not receive any compensation. In addition, its tariffs have not changed since 2006 and this is one of the main reasons for its poor performance. x The tariff charged by state power generation companies Enguri and Vardnili is below the private sector’s tariff and does not cover costs (2.5 tetri per Kwh as opposed to 3-8 tetri per Kwh for the private sector). 3.25. The government also faces explicit contingent liabilities from PPAs that it has signed with foreign investors to purchase all newly generated electricity. Hydropower is a potential source of economic growth and is a cheap alternative to thermal energy and imported electricity. To attract foreign investors in the hydro generation sector, the Government has signed Memoranda of Understanding (MOU) with foreign investors and has entered into 27 PPAs to purchase energy from the generators41. Some of this energy will be exported through high voltage transmission lines (HVTLs) through a subsidiary of GSE under the Electricity Transmission Agreements (ETAs). 3.26. However there is a low likelihood that these contingent liabilities will materialize. For PPAs two main sources of risks are identified: (i) price proxied by projected import prices; and (ii) market, or demand based on MOE projections. Two scenarios were considered to assess the risk to the government. The first scenario, which is an outlier, calculates the maximum exposure to market risk assuming no change in tariffs and further assuming that electricity consumption declines by 20 percent which was the largest drop since independence due to Georgia’s civil war in 1991-92. Under this assumption, the cost to government could average US$56.23 million per year during 2014-23 starting from US$9.3 million in 2014 and going up to US$87.4 million by 2020. The second scenario, which is the base case, uses MOE data and assumes a consumption growth rate of 5 percent, and with no price change there are net benefits every year starting in 2015 and rising to US$113 million in 2023. This scenario assumes that the newly built HPPs substitute more expensive imports and thermal electricity (Table 3.3). Table 3.3: Market Risk Exposures (In US$ million) 2015 2016 2017 2018 2019 2020 2021 2022 2023 First Scenario (full exposure) 19.5 33.3 39.7 43.8 72.9 81.7 87.4 87.4 87.4 Second scenario (Gain/loss to ESCO from trading) 21.6 6.7 22.8 49.4 30.9 65.5 33.1 67.3 113.9 Source: Assessment of risks related to power purchase agreements in Georgian Electricity Sector, PER Background paper. 3.27. The government also faces implicit contingent liabilities from SOEs which have borrowed actively from international markets.42 The bond issuance by two SOEs, Georgian Railways and GOGC, which amounted to US$750 million (about 5 percent of GDP) are not recorded as part of the  41 This section draws from a recent World Bank Policy Research Paper, “Assessment of Risks Related to Power Purchase Agreements in Georgian Electricity Sector”, January 2014. 42 JSC Partnership Fund, 2013 Consolidated Statement of Financial AccountsǤ 51  public debt stock as they do not have an explicit guarantee. While they do not represent an immediate liability (repayments start only in 2017), such issuances need to be monitored. The PF also enjoys an implicit backing from the Government. While the PF legislation states that the Government is not under any legal obligation to bail out the PF, in practice, things seem to be different. The International Monetary Fund (IMF) noted that according to the Fitch credit rating report, the Government issued a Letter of Comfort in March 2012 stating its commitment to support the PF in case of distress. E. THE WAY FORWARD 3.28. The government needs to have an inventory of its SOEs. There are significant data constraints in conducting any analysis of SOEs. The current degree of decentralization in the management of SOEs makes it difficult for the government to have a consolidated picture of SOE performance. In addition to the database, the government also needs to decide on an overall policy for SOE ownership. This policy would then dictate the government’s decisions to privatize, change legal status, close or integrate SOEs with public administration. For its privatization efforts, the government needs to have uniform privatization policies. Box 3.2:.OECD Principles and Guidelines on Disclosure From OECD Principles of Corporate Governance The corporate governance framework should ensure that timely and accurate disclosure is made on all material matters regarding the corporation, including the financial situation, performance, ownership, and governance of the company. Disclosure should include, but not be limited to, material information on: 1. The financial and operating results of the company. 2. Company objectives. 3. Major share ownership and voting rights. 4. Remuneration policy for members of the board and key executives, and information about board members, including their qualifications, the selection process, other company directorships and whether they are regarded as independent by the board. 5. Related party transactions. 6. Foreseeable risk factors. 7. Issues regarding employees and other stakeholders. 8. Governance structures and policies, in particular, the content of any corporate governance code or policy and the process by which it is implemented. From OECD Guidelines on Corporate Governance of State-Owned Enterprises State-owned enterprises should observe high standards of transparency in accordance with the OECD Principles of Corporate Governance. SOEs should disclose material information on all matters described in the OECD Principles of Corporate Governance and in addition focus on areas of significant concern for the state as owner and the general public. 1. A clear statement to the public of the company objectives and their fulfillment. 2. The ownership and voting structure of the company. 3. Any material risk factors and measures taken to manage such risks. 4. Any financial assistance, including guarantees, received from the state and commitments made on behalf of the SOE. 5. Any material transactions with related entities. Source: Extracted from OECD Principles of Corporate Governance (2004) and OECD Guidelines on Corporate Governance of State-Owned Enterprises (2005). 52  3.29. All financial links between the state budget and SOEs should be fully transparent. This would require the government to route all public expenditures through the budget rather than the SOEs. OECD guidelines recommend contractual agreements between governments and SOEs whereby costs are transparent, particularly for SOEs that operate in competitive sectors to avoid distortion in competition (Box 3.2). 3.30. The current degree of decentralization in the management of SOEs contradicts OECD guidelines for the governance of SOEs. OECD guidelines recommend the centralization of or at least strong coordination of the government’s ownership function in one agency. This agency should be directly accountable to Parliament and report annually on SOEs’ aggregate performance. The Ministry of Economy seems to be well placed to collect and maintain centralized SOEs database, while the MOF needs to develop a holistic approach in accounting, reporting and managing state assets. 53  CHAPTER 4. INTERGOVERNMENTAL FISCAL RELATIONS A. INTRODUCTION 4.1 The current administrative structure in Georgia, with just one layer of sub-national government, is in line with international practices. With a territory of 69,700 square kilometers and a population of around 4.4 million, Georgia’s size warrants one subnational tier of government or at most two, if the autonomous district of Adjara is included. The current structure avoids duplication of functions and is in line with the division of responsibilities between the center and the sub-national governments (SNGs), with several public services being provided directly by the center. The population covered by the local governments in Georgia varies widely and the median population of 35,000 is much higher than the median population of 19,000 for the bottom tier jurisdictions in most European countries. This helps Georgia achieve economies of scale in the provision of local services. 4.2 Despite limited expenditure responsibilities, SNG spending in Georgia is reaching levels similar to those in countries where subnational governments are responsible for more resource intensive services. SNGs in Georgia are primarily responsible for the provision of housing, utilities and communal services while the more resource intensive services like education and health are with the central government. SNG expenditures are at 7 percent of GDP, and have accounted for about a fifth of consolidated government expenditures. They have been increasing steadily since 2008, both in real terms and relative to GDP. There is also significant inter-regional disparity in spending in Georgia, primarily attributable to lumpy capital expenditures. Another crucial factor affecting expenditures across regions is the size of the population. 4.3 The low level of revenue autonomy at the local level acts as a disincentive for effective fiscal management. Intergovernmental fiscal transfers have accounted for nearly 70 percent of total revenues of SNGs. Since 2008, the only own sources of subnational revenues have been property taxes and non-tax revenues, such as rents, fines and penalties, sale of goods and services and other local government collections. While local governments can determine property tax rates, local accountability is limited by the fact that local governments are passive recipients of revenues collected by the national tax authorities. Therefore, there is limited room to hold SNGs accountable for revenue performance. Fiscal transfers from the center help close the gap between expenditure needs and revenues of local governments. While there is a complicated formula for the calculation of equalization grants to SNGs, it lacks transparency and has proved difficult to replicate. In addition, the center also doles out unclassified special purpose transfers, which are largely residual in nature and are mostly used for capital expenditures. Therefore, there is no incentive for local governments to reduce expenditures as it would result in lower transfers from the center. 4.4 This chapter reviews the fiscal implications of SNGs in Georgia. The analysis in this chapter leads to the following measures for the government to consider: x Short term: Improve data collection on all aspects of subnational finances to increase transparency and fiscal management of SNGs. x Medium-term: Change the formula for the calculation of the equalization grant to make it simple and to better reflect the overall pool of resources and expenditure needs of SNGs. x Medium-term: Remove the suspension of taxes on movable property and lower the income threshold for exemptions on immovable property. 54  x Medium-term: Allow SNGs to share the tax base for certain taxes (such as the personal income tax) and be able to set their own tax rates on top of the national tax rate applied to the same tax base. Both taxes (the central and local components) could be administered by the central tax administration. 4.5 The rest of the chapter is organized as follows. Section B provides an overview of the institutional and administrative framework for local governments in Georgia. The next section focuses on the level and composition of their expenditures while Section D discusses the main sources of revenues for SNGs. Section E gives the way forward. B. INSTITUTIONAL AND ADMINISTRATIVE STRUCTURE 4.6 Georgia’s administrative structure comprises two levels of government—the central government and the municipalities. The current system of local self-governments was put in place in 2006 when Georgia completely revamped its system of local self-governments. The existing two-tiers of sub-national governments was reduced to a single tier and the number of sub-national units reduced from about 1,000 to 70—5 cities, 64 municipalities and the autonomous district of Adjara.43 The size of municipalities in Georgia varies widely—ranging from 5,000 to 178,000 in the municipalities and from 48,000 to 1.17 million in the cities. The population in the cities is much larger than the median population (of 19,000) in the bottom tier jurisdictions in most European countries. The median population in Georgia’s municipalities is 35,000 and this helps in optimizing the economies of scale in the provision of local services. The government has adopted a new local self-government code in February 2014 which will add another five cities to the existing five, thereby reducing the average population size but increasing the costs.44 The newly formed cities will become effective from June 2014, after local elections have been held. 4.7 The current single-tier of sub-national government is well suited to Georgia’s size. With a territory of 69,700 square kilometers and a population of around 4.4 million, Georgia’s size warrants one subnational tier of government or at most two, if Adjara is included. The current structure of governance, with just one layer of sub-national government (SNG), is in line with international practices. It also avoids duplication of functions and is in line with the division of responsibilities between the center and the sub-nationals, with several public services being provided directly by the center. 4.8 The law on local self-governments provides for a simple governance structure. Each municipality has a directly elected local council (Sakrebulo) which elects the head of the council (Gamgebeli). The mayors of four of the five cities are also elected by the local councils whereas the Mayor of Tbilisi is directly elected by the people. These municipalities are responsible for managing sub- national service delivery and preparation and execution of local budgets. The local councils are accountable to the local population and on fiscal issues, they report to the Ministry of Finance. 4.9 The relatively larger size of municipalities, in terms of land area, may raise some issues with civic participation. With a median land area of 825 square kilometers, some municipal governments have very dispersed populations under their jurisdictions resulting in limited civic participation in local policies. For the benchmark group of countries, the median land area of the bottom tier jurisdictions is only 190 square kilometers, less than a fourth of that in Georgia.  43 Adjara has another layer of government under the district level. 44 These will be carved from existing municipalities. 55  C. LEVEL AND COMPOSITION OF SNG EXPENDITURES 4.10 SNG expenditures have accounted for about a fifth of consolidated government expenditures since 2008. As of 2012 the subnational share in general government expenditures puts Georgia internationally in the middle range of expenditure decentralization levels for its size (Figure 4.1). The share is just less than a quarter but has been increasing steadily since 2008, both in real terms and relative to GDP, and also as a share of total general government expenditures (Figure 4.2). This increase has largely been accounted for by capital expenditures financed by special transfers from the central government. Thus, at almost 7 percent of GDP, subnational expenditures in Georgia are now reaching the levels of some countries where subnational governments are responsible for more resource intensive services, such as primary education (Figure 4.3). Meanwhile almost one third of the subnational expenditures have been accounted for by capital expenditures which have fluctuated significantly over the years (Figure 4.4). At 2 percent of GDP, the five-year average of subnational capital expenditures in Georgia is higher than the 1.1 percent average for low-middle income countries or the 1.5 percent average for ECA countries. Figure 4.1: Decentralization of Expenditures and Country Size (Share in total expenditure, sq.km, thousand population) Source: Author’s estimates Figure 4.2: Subnational Expenditures (In real terms and as a share of general government) 1.8 31% 30% 35% 1.6 30% 24% 23% 23% 1.4 21% 21% 19% 20% 25% 1.2 18% 18% 19% 17% 16% 20% 1.0 13% 0.8 15% 0.6 10% 0.4 0.2 5% 0.0 0% 2004 2005 2006 2007 2008 2009 2010 2011 Subnational Public Expenditures in constant 2011 Prices, bil.GEL (lhs) SNG Share in GG Expenditures (excl. CapEx) SNG Share in GG Expenditures Source: IMF GFS. 56  Figure 4.3: Decentralization of Social Services and the Size of Subnational Expenditures (Percent of GDP, education and health share of SNG) Source: Author’s estimates 4.11 SNGs in Georgia are primarily responsible for the provision of housing, utilities and communal services while the more resource intensive services like education and health are with the central government. Public services delivered by SNGs in Georgia include communal affairs, kindergarten schools, and selected health services.45 All other services are provided directly by the central government’s line ministries and agencies. However, the law on local self-governments includes certain clauses by which the ambit of responsibilities and functions of SNGs can be widened. These include: (i) a “general competence” clause allowing local governments to perform functions beyond those specifically enumerated in the law as long as they are not explicitly disallowed in the legislation; and (ii) a clause that allows for delegation of functions to local authorities by law or through intergovernmental agreements.46 Figure 4.4: Economic Composition of LSG Expenditure 4.12 Housing, culture and sports account 2007-11 for more than half of SNG expenditures. The (In percent of GDP) largest functional category is Housing and 10.0% Communal Amenities, accounting for 40 percent of total subnational expenditures Percent of GDP (Figure 4.5). Besides residential housing, this 2.2% 1.6% 1.6% 2.2% category includes water supply, outdoor 5.0% 2.6% 1.6% 1.9% lighting and other minor communal services. A 1.5% 2.5% 2.3% distant second category is Physical Culture, 0.9% 0.9% 0.0% 1.3% 1.2% 0.9% Sports, Culture and Media, accounting for 15 0.6% 0.0% 0.5% 0.9% 0.0% 0.5% 0.5% 0.6% 0.6% 0.5% percent of total subnational expenditures. The 2007 2008 2009 2010 2011 third place is shared by Economic Affairs Personnel Goods and Services (transport/maintenance of local roads and Interest Subsidies support to agricultural activities) and General Grants Social Benefits Public Services (personnel, operation and Source: IMF GFS administrative expenses), both accounting  45 Reportedly, in rare occasions local governments do assist school with emergency repairs and material supplies. 46 The Constitutional Law of Georgia on the Status of the Autonomous Republic of Ajara delegates to this region authority over budget, economy and finances, tourism, education, culture, sport, urban development, health-care, agriculture, forestry, local taxes and dues, and state property management. 57  for 11 percent of total subnational expenditures.47 Actual expenditures vary depending on the needs of specific municipalities. The cities of Rustavi and Poti spend most of these resources on the development of housing stock while Batumi and Tbilisi spend the bulk of their expenditures on urban planning, administration and community development. 4.13 In line with the limited expenditure Figure 4.5: The Functional Composition of Sub- responsibilities of SNGs, they accounted for a National Expenditures lower share of total public investments in (In percent of GDP) Georgia compared with other countries in a 10% similar income group. In the group of lower- middle income countries, subnational governments account for half of capital 2.2% 1.8% 1.4% 1.9% 2.4% expenditures but only a quarter of recurrent 5% Peecent of GDP 0.2% 0.7% 0.4% 1.1% 0.3% 0.3% 1.4% 0.8% expenditures. In Georgia the share of 0.5% 1.0% 0.5% 1.3% 0.7% 1.0% subnational governments in total public 0.6% 2.9% 2.6% 2.5% 2.7% 1.9% investments reached 38 percent in 2011, below 0% the 50 percent mark. Typically, subnational 2007 2008 2009 2010 2011 governments tend to account for a larger share Other Social policy of capital expenditures because of the capital Economic affairs intensive nature of the services provided, Physical culture, sports, culture and media especially in the social sectors (Martinez- Vazquez and Timofeev 2012). Source: IMF GFS. 4.14 There is significant inter-regional disparity in spending in Georgia. Even when aggregated to the regional level, combining large and small and urban and rural municipalities, the difference in spending levels is still very stark (Figure 4.6). Some of this disparity could be due to the recent wave of capital projects, which tend to be lumpy. But even for recurrent expenses, the difference is more than two- fold between the lowest per capita spending in Kakheti region and the highest per capita spending in the region of Racha Lechkhumi and Kvemo Svaneti. Some of these differences may reflect differences in the cost of inputs for which information is not available.48 Another possible source of disparity is the differences in population size and the inherited network of facilities. Differences are even starker when comparing individual municipalities (Table 4.1). Even among the cities, there is a four-fold difference in total per capita expenditures and a three-fold difference in per capita operational expenses.  47 It appears that “General public services” are slightly more decentralized than total expenditures, likely reflecting the higher expenditure needs of multiple administrations – both executive and legislative—at the subnational level.  48 Wage differences are likely to be an important driver of costs differences. From the available information, the wage difference between Tbilisi and the poorest region in the country (Racha Lechkhumi and Kvemo Svaneti) is 2.8 times. 58  Figure 4.6: Expenditure Disparities in Regions, 2012 (GEL per capita) Total Recurrent 600 498 300 276 500 385 400 250 209 GEL per ca;pita 206 GEL per capita 400 191 291 291 300 200 150 161 164 146 148 300 245 253 217 237 150 118 200 100 50 100 0 0 Source: MOF. Table 4.1: Variation in Per Capita Expenditures Across Different Functions, 2012 (GEL per capita) Housing Recreation, Out of and Culture, Which Communal and Economic Social Type of Jurisdictions Total Recurrent Amenities Religion Affairs Protection Other Districts Average 308 156 42 34 109 12 111 Minimum 113 54 5 8 1 0 44 Maximum 1,705 375 204 165 1,457 88 437 Coefficient. of Variation 0.75 0.39 0.79 0.78 1.89 0.96 0.52 Cities Batumi 1,452 600 619 187 341 63 243 Tbilisi 486 372 155 28 102 63 138 Kutaisi 310 207 110 56 34 8 102 Poti 639 369 126 236 90 18 170 Rustavi 392 310 139 55 42 17 139 Average 656 371 230 112 122 34 158 Coefficient of Variation 0.70 0.39 0.95 0.82 1.04 0.80 0.33 Source: World Bank Staff Estimates, MOF. 4.15 Population size is one of the key determinants of district expenditures. The estimated elasticity of expenditures with respect to population size is around 0.7. The only other statistically significant determinant of district expenditures was the incidence of indigent households (positively related), explaining an additional 2 percent of expenditure variation (in logarithms).49 Availability of own revenues from property taxes, while weakly statistically significant, does not have a substantial impact on district expenditures with an estimated elasticity of under five percent. Based on all these determinants of  49 In addition to pressure for bloated local government employment, higher incidence of poverty also increases public expenditures on child day care as the national legislation has historically exempted indigent households from paying kindergarten tuition. 59  expenditures, it was estimated that splitting in half a district of average size would increase per capita expenditures by about 30 percent. Figure 4.7: Economies of Scale in District-Municipality4.16 Per capita spending is negatively Expenditures correlated with the size of population (GEL, population) reflecting higher resource requirements to provide the same level of service and essential fixed costs irrespective of the number of people served. The municipality with the highest per capita expenses in Georgia also has the smallest population (4,800).50 This is essentially because of the low scale of operation. For example, empirical evidence shows the existence of economies of scale in the delivery of water services up to 150,000 or so customers. This comes as no surprise given the large capital expenditure in infrastructure Source: author’s calculations, MOF, Geostat. that is required to build water treatment facilities. Similarly, there is evidence of economies of scale in garbage collection services for populations ranging between 5,000 and 20,000 (Figure 4.7). D. REVENUE SOURCES 4.17 Tax and non-tax revenues account for only about 30 percent of total SNG revenues. Since 2008, the only own sources of subnational revenues have been property taxes and non-tax revenues, such as rents, fines and penalties, sale of goods and services and other local government collections. The only exception is Adjara region which retains 100 percent of its income tax collections. Non-tax revenues accounted for 15 percent of SNG revenues in 2013 and comprised local fees, penalties and fines and sale of goods and services. (Figure 4.8). For both cities and municipalities, the revenue sources are the same. Figure 4.8: Revenue Structure of Subnational Governments 4.18 While local governments can (In percent of GDP) determine property tax rates, there exist some design and implementation issues. Even for the 8% property tax, local accountability is somewhat 7% 1.17% reduced by the fact that local governments are 1.35% 6% 0.86% 1.22% passive recipients of revenues collected by the 0.00% national tax authorities—local governments are Percent of GDP 5% 1.05% 4.08% not involved in the administration and 4% 3.61% 3.58% 3.52% enforcement of property taxes. On 3% 3.10% implementation, no information on delinquent tax 2% 0.37% 0.40% 0.46% payers is shared with subnational governments 0.41% 1% 0.63% 0.69% 0.89% 0.92% 0.91% and therefore they cannot make any effort to 0.75% 0.66% 0.66% 0.81% 0.80% improve tax compliance. Regarding design, under 0% 2007 2008 2009 2010 2011 the current legislation, most physical persons do Non-Tax Property Taxes Profits and Income Taxes not have to pay property taxes due to a nationally Recurrent grants Capital grants mandated exemption determined by a rather high Source: IMF GFS. household income threshold, and in addition, the national legislation also exempts agricultural  50 Kazbegi which is incidentally located in Mtskheta Tianeti region has the highest average wage outside Tbilisi. 60  plots under 5 hectares. These measures, which are highly unusual, severely limit the revenue potential of property taxes. In addition, the taxation of movable property in Georgia has been abolished. 4.19 Rather than restoring the Soviet practice of retaining a share of the national taxes, local governments could share the tax base on national taxes to increase revenue autonomy. In particular, it would be more desirable to allow local governments to share the tax base for certain taxes (such as the personal income tax) and be able to set their own tax rates on top of (or to piggyback on) the national tax rate applied to the same tax base. Both taxes (the central and local components) could be administered by the central tax administration. This would help increase accountability and fiscal responsibility among SNGs and also increase fiscal revenues. The main challenge in this alternative would be to avoid increasing the tax burden in the country. Figure 4.9: Conventional Sources of Local Revenue, 2009 (In percent) Property taxes Non-tax revenues 1.60% 1.80% 1.40% 1.60% 1.40% Percent of GDP 1.20% 1.20% Percent of GDP 1.00% 1.00% 0.80% 0.80% 0.60% 0.60% 0.40% 0.40% 0.20% 0.20% 0.00% Azerbaijan Croatia Moldova Georgia Ukraine Armenia Czech Republic Slovakia Belarus Slovenia B&H Lithuania Romania Bulgaria Latvia Serbia FYR Macedonia 0.00% Czech Republic Turkey B &H Georgia Malta Russia Ukraine Poland Hungary Azerbaijan Moldova Belarus Croatia Slovakia Armenia Slovenia Albania Latvia Romania Serbia Bulgaria FYR Macedonia Lithuania Source: World Bank Staff Estimates. 4.20 SNGs could also optimize on property taxes to enhance revenues. The current suspension of taxes on movable property in Georgia needs to be lifted because it not only reduces fiscal revenues but also the overall progressivity of the tax system. In the same vein, the current over generous exemptions of immovable properties need to be reconsidered.51 The experience of other countries suggests that these conventional local taxes can provide significant financing for local expenditures, especially in the more affluent cities (Figure 4.9). A more widely applied property tax has the potential of increasing revenue collections for local governments by 1 percent of GDP (Box 4.1).52  51 The general incidence of property taxes tends to be progressive especially when accompanied by “circuit breaker” provisions that reduce or suspend the tax for households with low income levels. 52 See, for example, Bahl and Martinez-Vazquez (2008). 61  Box 4.1: International Experience With Local Property Taxes Property taxes are a good and commonly used source of local government revenues. First of all, the property tax is a visible tax and thus conducive to political accountability. When both property and population are homogeneous and ownership of property is widespread, a property tax complies with the benefit principle. The balance between the services received by property owners and the property taxes they pay on their real estate typically can be capitalized into property values. That is, property taxes do not have to reduce the market value of dwellings in the market if the general perception is that the quality of services provided by the local government is good. Property taxes further follow the ability-to-pay principle, although some liquidity problems may be present for those homeowners with valuable real estate assets and low income. Being “house rich and income poor” can be a problem for elderly people. Some countries use especial exemption schemes (“homestead exemptions” or “circuit breakers”) to increase equity in the implementation of property taxes. In principle, the property tax should be easy to administer since property is very visible and immobile across local jurisdictions, which should give local officials a strong “tax handle”. However, in other situations, property taxation can weaken the benefit link by moving the tax burden on to a few classes of property, such as non-residential property. Figure 4.10: Local Government Involvement in Furthermore, taxation of property in an equitable way Property Taxation requires costly revaluation of property on a regular basis. (% of countries) The Figure 4.10 reports the share of countries with a local government involvement in various aspects of property taxation. This figure is based on a survey by Bird and Slack (2004) of 5 countries from each of the following 5 groups: OECD, Africa, Asia, Latin America, Central and Eastern Europe. For each of the 5 aspects of property taxation (rate- setting, property identification, assessment, revenue collection and appropriation), they find local government involvement in over half of the 25 surveyed countries. Local identification and assessment of properties are least prevalent in Central and Eastern Europe (both identification and assessment in Hungary, only assessment in Poland, neither identification nor assessment in Latvia, Russia, Ukraine). Source: Bird and Slack (2004). Local rate-setting is least prevalent in Asia. 4.21 A surcharge on local taxes, surtax, is another option that can be considered although there may be administrative difficulties in levying a local PIT surcharge according to the residence principle. Unlike revenue sharing, surtaxes allow local governments a certain measure of discretion over the tax rate, and the ability to control revenues at the margin. However, the employer withholding PIT from its workers might have to apply different PIT rates for workers residing in different municipalities. The solution is to have a flat rate of PIT surtax with minimum and maximum rates nationally legislated. The tax base would be the same as that for the national PIT and would be paid on a residence-basis (to the jurisdiction where the taxpayer lives). 4.22 Other revenues like user fees and betterment levies could also enhance the revenue autonomy of SNGs. Local governments can make further use of fees and user charges. As can be seen from Table 4.2, there are wide differences among municipalities in per capita levels of non-tax revenues. Local governments could also be granted the potential use of “betterment levies”. This is a one-time lump-sum payments exacted up front by sub-national governments from land and housing developers— and also potentially from homeowners— as a charge for public service improvements, such as road paving, drain infrastructure, sidewalks, street lights etc., which all have a visible benefit on property values. Third, vehicle and transportation taxes are generally attractive sub-national taxes because of the strong link between the ownership of vehicles on the one hand, and the use of local services and infrastructure (particularly roads) on the other hand. A fourth possibility is the assignment at the local level of natural resource taxes, especially when resources are geographically evenly distributed. 62  Table 4.2: Non-Tax Revenues, 2012 (In GEL per capita) Revenues Sale of From Goods and Penalties Type of Jurisdictions Property Services and Fines Donations Other Total Districts Average 14 3 4 1 2 23 Minimum 1 0 0 0 0 4 Maximum 156 63 18 28 53 161 Coefficient of Variation 1.70 2.46 0.76 6.29 4.32 1.12 Cities Batumi 63 78 19 0 11 171 Tbilisi 19 37 17 0 5 78 Kutaisi 3 18 12 0 10 43 Poti 18 20 6 0 0 45 Rustavi 3 20 7 0 2 32 Average 21 35 12 0 6 74 Coefficient of Variation 1.16 0.73 0.48 2.24 0.83 0.77 Source: World Bank staff estimates, MOF. Coefficient. 4.23 The low level of revenue autonomy at the local level acts as a disincentive for effective fiscal management. Regional governments in Georgia are highly dependent on central government transfers— recurrent and capital grants. The heavy dependence on transfers from the national budget has stood at around 70 percent of total local revenues in the last five years. Even the city of Tbilisi, the wealthiest local government, receives two thirds of its revenues in the form of grants from the national budget. Given that local governments are not responsible for social services like basic education and healthcare, which are significant in budgetary terms, the current level of transfer dependence of subnational government in Georgia stands out in comparison to other decentralized countries (Figure 4.11). It also makes it difficult to hold SNGs responsible for revenue performance. Given the high level of dependence on transfers, SNGs have no incentives to reduce expenditures as this would reduce the central transfers. Figure 4.11: Decentralization of Social Services and Transfer Dependence (Share in percent) Source: World Bank Staff Estimates. Note: Transfers include sharing of revenue from national taxes at source. 63  4.24 The intergovernmental fiscal transfers help close the gap between expenditure needs and revenues of SNGs. Since 2007, the formula-based equalization grants have accounted for half of intergovernmental grants while the other half includes special purpose transfers and conditional earmarked grants. According to the budget code, special purpose transfers are essentially unclassified, residual transfers. They are mostly capital grants, the allocation of which has been ad hoc until 2013. At the same time conditional earmarked grants, which are supposed to follow delegated functions (military conscription centers, certain healthcare programs, etc.), are miniscule at less than 1 percent of total transfers. 4.25 However, the formula for the calculation of the equalization grants creates disincentives for improving spending efficiency. The Budget Code (Article 74) requires the equalization grant formula to be transparent as it is essentially intended to close the gap between estimated expenditure needs and projected revenues of each local government unit. However, the current methodology for calculating expenditure requirements is rather cumbersome and lacks transparency. Expenditure needs are estimated by apportioning the total expenditure envelope of all the municipalities in proportion to the relative index of expenditure needs computed for each locality. The index of expenditure needs is computed for each locality as an adjusted weighted average of six parameters (population size; share of indigent population; shares of population below six years old and between 6-18 years old; and population density). 4.26 Beyond its lack of simplicity and transparency, the current formula for the computation of the equalization grant has some problematic aspects. x The formula confuses the two basic dimensions of expenditure needs and revenue capacity. x Some of the local characteristics entering the formula do not reflect actual expenditure needs. x The methodology does not provide justification for the calculation of adjustment coefficients. x The computation of revenue capacity might provide negative incentives for revenue mobilization. 4.27 Notwithstanding the explicitly prescribed methodology, many municipalities have difficulty in replicating and comprehending the current grant allocation system. This is due to either a lack of access to data for the indicators used in the formula or because of confusion about how and why specific weights are assigned to these indicators in the formula. Previous analytical reports have failed to replicate the actual equalization transfers finally issued by the Ministry of Finance to the different jurisdictions. 53 Box 4.2: Developing formula-based grant mechanisms In general, the choice of variables or cost factors to be used in formula-based grants depends on the cost structure of the specific category of local functions at hand. Such factors should (Bahl et al. 2001): x Accurately reflect the targeted characteristics. The school-age population with special needs captures better the costs for education than the actual enrollment of those categories of students. x Be regularly updated in the future (every year or every two years). Data originating from ad-hoc studies should not be used. x Come from an independent source respected by all stakeholders. A national bureau of statistics often has more credibility than the Ministry of Finance or a line ministry. x Be drawn from a source that cannot be manipulated by one or more local governments. Self-reporting of  53 See Murgulia, Gvelesiani and Toklikishvili (2011). 64  data by local governments cannot be trusted if local governments are well aware of the link between what they report and the amount of resources they receive in the future. x Reflect objective drivers of costs (for example, the number of clients) rather than the chosen mode of service provision. Population density is more objective in capturing cost drivers than the actual enrollment size. Given that a variety of functions are assigned to local governments in different countries, there is a wide variation in the factors included in the formula. 54 These factors can be classified into two broad categories: 1. Indicators of service need: illiteracy rate, poverty rate, sickness rate, and infant mortality; 2. Cost indicators: price index, land area, average temperature, and mountainous areas. One of the most important issues is how to determine weights for each of the factors so that it adequately captures the disparities in service costs across municipalities. Several approaches can be used to arrive at a particular set of factor weights in a more or less objective manner. A more scientific approach is to utilize local budget data to establish how the cost of delivering standard services varies across local jurisdictions, and, in particular, how these costs are responsive to variations in socio-economic characteristics of the population. 55 The elasticity of per capita expenditures shows substantial responsiveness to population size, some responsiveness to the incidence of poverty and no statistically significant responsiveness to population density or road length. 4.28 Intergovernmental transfers have helped achieve fiscal equalization among district municipalities but not as much in the cities. The 40-fold difference between the poorest and the richest districts in own revenues per person is reduced to a six-fold difference after the allocation of the equalization grants. In 2012, the per capita own revenues of districts, on average, deviated by 100 percent from the district average. At the same time, the per capita recurrent expenditures of districts, on average, deviated by 39 percent from the district average (Table 4.3). The coefficient of variation for cities shows that the disparities among cities in own revenues actually increase after the allocation of equalization grants. This is possibly because of the 3.8-fold adjustment applied to the relative index of expenditure needs for the cities of Batumi and Tbilisi. Only the allocation of special funds (mostly capital grants) manages to bring the coefficient of variation back to the level observed for own-source revenues of the cities. But much more importantly, these transfers fail to achieve any level of equalization between the districts as a group and cities as a group which narrows to 50 percent after the allocation of special (capital) grants. Table 4.3. Equalization Impact of Transfers, by Type of LSG, 2012 (GEL per capita) Plus Transfers Own-Source Out of Which From the Central Plus Transfers Type of Jurisdictions Revenue Property Tax Budget From Funds Districts Average 55 31 152 287 Minimum 7 1 72 90 Maximum 279 196 452 1,680 Coefficient. of Variation 1.00 1.13 0.39 0.84 Cities Batumi 271 100 1,127 1,134  ͷͶ‘‡šƒ†ƒ”–‹‡œǦƒœ“—‡œǡʹͲͲ͹ ͷͷ‡…ƒ—•‡–Š‹•‹†‘ˆƒƒŽ›•‡•‹•„ƒ•‡†‘•–ƒ–‹•–‹…ƒŽƒ˜‡”ƒ‰‹‰ǡ‹–ƒ••—‡•–Šƒ––Š‡‹†‡–‹ˆ‹‡†’ƒ––‡”•”‡ˆŽ‡…– ‘„Œ‡…–‹˜‡‹’ƒ…–‘ˆ•‘…‹‘Ǧ‡…‘‘‹……Šƒ”ƒ…–‡”‹•–‹…•‘—‹–…‘•–•Ǥ 65  Table 4.3. Equalization Impact of Transfers, by Type of LSG, 2012 (GEL per capita) Plus Transfers Own-Source Out of Which From the Central Plus Transfers Type of Jurisdictions Revenue Property Tax Budget From Funds Tbilisi 179 101 559 559 Kutaisi 66 23 189 280 Poti 185 140 219 572 Rustavi 81 48 205 326 Average 156 83 460 574 Coefficient. of Variation 0.54 0.56 0.88 0.59 Source: Author’s calculations. 4.29 Subnational borrowing is allowed in Georgia with an explicit authorization from the Ministry of Finance but has not been used much. Given the limited sources of own revenues for SNGs in Georgia, they have not used the debt route to finance expenditures and are relatively debt-free. Their borrowings are mostly related to on-lending of donor funds. If subnational governments are given more taxing powers in the future, this should improve their ability to repay loans and hence their creditworthiness. However, borrowing at the subnational level is risky because local officials can easily be tempted to overspend and then move the responsibility of repayments to the central government. Therefore, there is a need to balance the access to borrowing by subnational governments with institutional mechanisms that preserve fiscal discipline. SNGs can also issue guarantees with an authorization from the ministry of finance. E. THE WAY FORWARD 4.30 The current two-tier structure of governance in Georgia is in line with international standards and should be preserved. The current structure helps Georgia reap the economies of scale in the delivery of local services. With the creation of the five additional cities, the issue of adequate representation would be taken care of but a further proliferation of SNGs needs to be guarded against since it will increase costs. In addition, given the limited ambit of responsibilities of SNGs in Georgia, there appears little justification to increase their numbers. 4.31 More effective fiscal management at the local level will require the institutionalization of an incentive mechanism to raise local revenues and reduce local expenditures. There is currently no incentive for SNGs to improve fiscal management because they lack revenue autonomy and are not accountable for revenue performance. In addition, their expenditures are taken care of largely through central transfers, nearly half of which are ad-hoc. Hence the current system of sub-national finances incentivizes local governments to show high expenditure requirements so as to get larger transfers from the center. Improved fiscal performance will require an incentive mechanism to reward fiscally prudent SNG. 66