Report No. 42450-UA Ukraine Improving Intergovernmental Fiscal Relations and Public Health and Education Expenditure Policy: Selected Issues February 28, 2008 Poverty Reduction and Economic Management Unit (ECSPE) Europe and Central Asia Region Document of the World Bank TABLE OF CONTENTS PREFACE .................................................................................................................................................... 1 EXECUTIVESUMMARY ......................................................................................................................... i 1.THE NEED FOR IMPROVING INTERGOVERNMENTAL FISCAL RELATIONS AND REFORMINGSERVICEDELIVERY:SETTINGTHE SCENE ................................................. 1 A. CONTEXT........................................................................................................................................................................ 1 B. RECENTMACRO-FISCALDEVELOPMENTS.......................................................................................... ..........1 C. WHY LOOKAT LOCAL GOVERNMENTFINANCESINUKRAINE?........................................................................ 5 D. WHAT ARE THEPRESSURESFORREFORM? ............................................................................................................ 6 E. GENERALEXPENDITURETRENDSAT THELOCALLEVEL................................................................................ 12 F. CONCLUSION............................................................................................................................................................... 15 2.IMPROVINGINTERGOVERNMENTALFISCALRELATIONSINUKRAINE ...................17 A. INTRODUCTION ........................................................................................................................................................... 17 B. BACKGROUND............................................................................................................................................................. 17 C. IMPROVING EXPENDITURE RESPONSIBILITIES, EXPENDITURE AUTONOMY, AND SECTORAL ADMINISTRATIVE REFORMATTHELOCALLEVEL......................................................................................... 20 D. STRENGTHENINGREVENUEASSIGNMENTS........................................................................................................ 28 E. ASSESSMENTOFAND RECOMMEDATIONSONINTERGOVERNMENTALTRANSFERS............................... 35 F. STRENGTHENINGTHESUB-NATIONALBORROWINGFRAMEWORK............................................................. 45 G. IMPROVINGOTHERCRITICAL CROSS-CUTTINGISSUES................................................................................... 48 ...................................................................................................................................................... 50 3. OVERCOMING FISCAL, EFFICIENCY, AND EQUITY CHALLENGES IN PUBLIC HEALTHSPENDING ...................................................................................................................... 51 A. INTRODUCTION........................................................................................................................................................... 51 B. THEHEALTHSECTOR IN UKRAINE: STRUCTUREOFAND TRENDSINTHEMAININDICATORS...............52 C. HEALTHEXPENDITURESAND FINANCING........................................................................................................... 57 D. ISSUESAND CONSTRAINTSFACINGTHEHEALTHSECTOR.............................................................................. 64 E. HEALTHINSURANCEREFORMSUNDERDISCUSSION........................................................................................ 73 F. DISCUSSIONAND RECOMMENDATIONS........................................... ............................................................. 75 4. OVERCOMING FISCAL, EFFICIENCY, AND EQUITY CHALLENGES IN PUBLIC EDUCATIONSPENDING ............................................................................................................... 79 A. INTRODUCTION........................................................................................................................................................... 79 B. THEEDUCATIONSECTORSTRUCTURE................................................................................................................. 79 C. PUBLICEDUCATIONEXPENDITURESAND MAINEDUCATIONINDICATORS:TRENDSAND ANALYSIS .83 D EFFICIENCYISSUES.................................................................................................................................................... . 89 E. FINANCINGAND ADMINISTRATIVEISSUES......................................................................................................... 93 F. OUTOFPOCKETEXPENDITURES,EQUITY,AND GOVERNANCEISSUES........................................................ 95 G. QUALITY ISSUES......................................................................................................................................................... 99 H. CONCLUSIONSANDRECOMMENDEDACTIONS................................................................................................ 101 5. STRENGTHENINGCAPITALBUDGETINGINLOCALGOVERNMENTS ...................... 105 A INTRODUCTION......................................................................................................................................................... . 105 B. CONTEXT................. ..................................................................................................................................... 105 C. GENERALTRENDSINLOCAL CAPITAL SPENDING........................................................................................... 107 D. ISSUESINLOCALCAPITALBUDGETING............................................................................................................. 109 E. RECOMMENDATIONS.............................................................................................................................................. 117 LISTOFTABLES Table 1.1: Key Macro Indicators. 2000-06................................................................................................................... 2 Table 1.2: Major Fiscal Trends. 2002-06 ..................................................................................................................... 3 Table 1.3: Revenues as Percent o f GDP, 2004-06 ........................................................................................................ 4 Table 1.4: Major Fiscal Trends inLocal Government Budgets, 2002-06 .................................................................... 6 Table 2.1: Expenditure Responsibilities-By Function.............................................................................................. 21 Table 2.2: Average Size o fLocalGovernments inUkraine ......................................................................................... 22 Table 2.3: Own Local Taxes andFees inDescending Order inTerms o f Collection, 2005 and 2006(in millionUAH) ................................................................................................................................................................. 31 Table 2.4: Local Taxes List:Current versus Proposed by the Draft Tax Code .......................................................... 31 Table 3.1: Life Expectancy at Birthin2005, Ukraine Compared Internationally ...................................................... 54 Table 3.2: Leading Causes o f DeathinUkraine, 2005 ............................................................................................... 56 Table 3.3: Mortality Rates: Ukraine versus Comparators, 2005" .............................................................................. 57 Table 3.4: Percentage o f Total Health Spending by Programand Financing Sources, 2004 ....................................\ 61 Table 3.5: Comparing Different Estimates o f Out-of-Pocket Expenditures inUkraine............................................. 62 Table 3.6: Norms DetailingNumbero f Beds to one doctor, by department ............................................................. 65 Table 4.1: Education Net Enrollment Ratio Estimates (in percent) and Implied Gap in Coverage of Primary and Secondary Education, 2004 ....................................................................................................................... 87 Table 4.2: Decomposition o f Increase inConsolidated Education Expenditures, 2002-05 ........................................ 88 Table 4.3: Changes inUnit Costs o f Education by Level, 2002-05 ........................................................................... 88 Table 4.4: Expenditure Responsibilities inEducation, by Local Government Level................................................. 93 Table 4.5: UnitCosts o f Educationby Level, Ukraine and OECD Average, 2005 .................................................... 94 Table 4.6: A Profile o f OOP Spending on Education during the 200312004 Academic Year .................................... 97 Table 4.7: Total Household Expenditures on Educationper Student, byIncome Quintiles and Area o f Residence ( in UAH) ....................................................................................................................................................... 98 Table 4.8: Enrollment Ratio Estimates (percent o f age group) based on the HBS Health and Education Module Sample Survey ,by Age Group, Region, Income Quintile, and Gender .................................................. 99 Table 5.1: Investment Needs at the Local Level ...................................................................................................... 109 Table 5.2: Sector Breakdown o f State Targeted Programs, 2006 ............................................................................. 116 LISTOFFIGURES Figure 1.1: Regionally Ukraine Ranks Poorly inCitizen Satisfaction inEducationDelivery ..................................... 8 Figure 1.2: Regionally Ukraine Ranks Poorly inCitizen Satisfaction inHealthDelivery ........................................... 9 Figure 1.3: Demographic Trends inUkraine ComparedRegionally ............................................................................ 9 Figure 1.4 a and b: Regional Inequalities Surface When Looking at Capital Expenditures....................................... 10 Figure 1.5: There I s Some Inequality inPublic Health Spending at the Local Level ................................................. 11 Figure 1.6: Evolution of Local Budgets Spending ..................................................................................................... 12 Figure 1.7: IncreasingWages at the Local Level ....................................................................................................... 12 Figure 1.8: Capital Expenditure Shares Among Levels o f Local Government .......................................................... 13 Figure 1.9: Social Sectors Dominate Spending at the Local Level ............................................................................ 14 Figure 2.1: Current Expenditure as a Share o f Total Expenditure inLocal Governments (oblast consolidated), 2002 and 2005 .................................................................................................................................................. 25 Figure 2.2: Public Wages Expenditure as a Percent o f Total Expenditure (oblast consolidated), 2002 and 2005 .....26 Figure 2.3: RevenueComposition o f Total Local Budget (net o f Treasury loans), 2001-05 ................................... 29 Figure 2.4: Rayon and City per Capita Revenues Before and After Equalization, and After Other Transfers and Subventions ............................................................................................................................................. 36 Figure 2.5: Equalization Power Changed Little Between2001 and 2005 (by oblasts, consolidated budgets) ...........37 Figure 2.6: Year-on-Year Change in the per Capita Equalization Received by Local Governments (oblast consolidated) (inpercent) ....................................................................................................................... 38 Figure 2.7: Year-to-Year Change inper Capita Capital Acquisition by Oblasts (consolidated), 2002-05 ..................41 Figure 2.8: Wealthier Oblasts and Rayons Tend to Receive Higher per Capita Funding for Capital Investments .....42 Figure 2.9: Treasury Loans Received, 200-05 (oblast consolidated) ......................................................................... 48 Figure 2.10: IncreasingTreasury Loans andDecreasingPrivate Loans, 2003-05 ...................................................... 49 Figure 3.1: Life Expectancy inUkraine: A Story o f Poor Performance, 1960-2005 ................................................. 55 Figure 3.2: Life Expectancy Vanes Significantly Across Oblasts ............................................................................. 55 Figure 3.3: HealthExpenditure inUkraine and Selected Countries, 2002-06 ............................................................ 58 Figure 3.4: Ukraine Ranks Close inAverage inHealth Spending as a percent o f GDP WhenNormalizedby Level o f Per Capital Income" ................................................................................................................................ 58 Figure 3.5: Growing Pressuresfrom Recurrent Public Spending inthe Health Sector, 2000-05 ............................... 59 Figure 3.6: Local Governments Are the MainPlayers inPublic Health Spending, 2001-06 ..................................... 59 Figure 3.7: Some Inequality Exists inPublic HealthSpending at the Local Level (2001 and 2005) ......................... 60 Figure 3.8: Out-of-Pocket Expenditures by Category, by Oblast............................................................................... 62 Figure 3.9: Sources for HIV Funding inUkraine, 2003-05 ........................................................................................ 64 Figure 3.10: Average Lengtho f Stay by Oblast, 2004 ............................................................................................... 67 Figure 3.11: The HealthSysteminUkraine Remains Hospital Centered, 2002-06 ................................................... 68 Figure 3.12: Informal Payments inUkraine versus Selected EuropeanGroups o f Countries .................................... 68 Figure 3.13: OOP o f Poorest Income Quintile versus Public Health Spending by Oblast, 2005 ............................... 69 Figure 3.14: Outpatient Utilizationby Income Quintile ............................................................................................. 70 Figure 3.15: Outpatient Visits by Facility Type and Income Quintile........................................................................ 70 Figure 3.16: Percent o f Households Spending More than 40 Percent o f Non-Subsistence Expenditure on OOP, Ukraine and Selected Countries .............................................................................................................. 71 Figure 3.17: Percent o fHouseholds inUkraine with Catastrophic OOP, by Quintile ................................................. 72 Figure 3.18: Redistributive Effect o f OOP Financing: A Regional Comparison ........................................................ 73 Figure 4.1: Ukraine Public Education Expenditure as Percent o f GDP (2001-06) and Regional Comparison" ........83 Figure 4.2: Public Expenditureson Educationby Level as Percent o f GDP. 2002-06............................................... 84 Figure 4.3: Public Expenditures on Education in2001-06. by Level o f Government and Level o f Education..........84 Figure4.4: Per Pupil Public Expenditures on EducationAcross Oblasts (consolidated budget inUAH). 2005........85 Figure 4.5: Evolution o fEnrollments by Level. 1990-2006....................................................................................... 86 Figure 4.6: Evolution o f Average Earnings inthe Education Sector and Other Sectors (inUAH) 2002-07 ..............89 Figure 4.7: The Student to Teacher Ratio in Ukraine is Acutely Low Compared Internationally (primary + secondary) ............................................................................................................................................... 90 Figure 4.8: Declining Student/Teacher Ratio inUkraine ........................................................................................... 91 Figure 4.9: The School Age Population Will Continue to Decline Over the Next 20 Years...................................... 91 Figure 4.10: Percent o f Respondents That Think Unofficial Payments Are Needed for Education: Ukraine and Comparator Groups ................................................................................................................................ 98 Figure 4.11:Percent o f Respondents Dissatisfied with Public Education: Ukraine and Comparator Groups ...........101 Figure 5.1: Evolution o f Local Budget Capital Expenditures as Percent o f GDP (breakdown by (fixed) capital investments and "transfers to enterprises and government agencies," 2001-06 .................................... 107 Figure 5.2: Share o f "Transfers to Enterprises and Agencies" inLocal Capital Expenditures2002-06...................107 Figure 5.3: The Growing Share o f LocalBudgets inTotal (fixed) Capital Investments, 2002-06........................... 108 Figure 5.4: (Fixed) Capital Investments by Level o f Local Government, 2002-05 .................................................. 108 Figure 5.5: Year-to-Year Change inper Capita Capital Acquisition by Oblasts (consolidated), 2002-05 ...............110 Figure 5.6 Tranfers to Enterprises and Agencies by Oblasts, 2005 .......................................................................... 112 Figure 5. 7: ExecutedCapital ConstructionProjects by Oblast as Percent o f Allocated Funding ........................... 113 Figure 5.8: Executed Capital ConstructionProjects by Rayons (and Cities) as Percent o fAllocated .................... 113 LIST OFBOXES Box 1 1. MainMessages o f the Public Finance Review (PFR) PhaseI....................................................................... . 4 Box 2.1: IntergovernmentalAdministrative andPoliticalRelations InUkraine ........................................................ 18 Box 2.2: Ukraine Comprehensive Treasury Single Account for Local Governments................................................ 19 Box 2.3: Examples o f Norms for Health and Education Facilities that Dictate How These Facilities and Local Governments FormTheir Budgets .......................................................................................................... 23 Box 2.4: The Inefficient Process o fBudgeting Based on InputNorms ...................................................................... 24 Box 2.5:: Introducing a Property Tax inUkraine- Basic Principles and Recommendations ...................................... 33 Box 3.1: HealthSector Structure and Administrative Accountabilities ..................................................................... 53 Box 4.1: Education Sector Structure and Administrative Accountabilities ................................................................ 81 Box 4.2: Financing o fHigher Education inUkraine: Contract Education................................................................. 82 Box 4.3: Norms for EducationFacilities .................................................................................................................... 92 Box 4.4: A BroadBrushTreatment o f Formal and Informal Out- of- Pocket-Payments inEducation ...................... 96 BOX5.1: KeyRecommendations ofthe PFR-I onCapitalBudgeting at the NationalLevel .................................... 106 Box 5.2: The Euro Cup 2012: A Challenge and an Opportunityto Strengthenthe Framework o f Capital Budgeting ............................................................................................................................................................... 111 Box 5.3: Distinguishing OperationalMaintenanceand Routine Repairs from Capital Renovations ....................... 115 Box 5.4: State TargetedProgramsinUkraine .......................................................................................................... 115 ANNEXES PREFACE The Programmatic Public Finance Review (PFR) (intwo phases) has the objective o f addressing key issues in public finance that face Ukraine today and recommending options for reform within a consistent, and pro-growth, macro-fiscal framework. The PFR covers the following selected areas and issues: (i) identification o f options for broadening the tax base and improving compliance to allow the gradually reducing the tax burden; (ii) the identification o f options for achieving expenditure savings in selected areas of the budget; (iii) estimation of the likely fiscal impact of key reforms to the pension the system; (iv) the identification of weaknesses in the process of capital budgeting and the provision of options for its strengthening; (v) the identification of weaknesses (and incomplete reforms) in the intergovernmental fiscal framework and o f options for its improvement; (vi) the identification o f options for achieving higher levels of efficiency in public spending on health and education; and (vii) the evaluation of the process o f local capital budgeting.The first four objectives were covered inPFR I, and the latter three are covered inthis report (PFR II). The present report (PFR 11) was prepared by a World Bank team comprising: Pablo Saavedra (team leader and lead author of Chapter 1 and Chapter 2 on intergovernmental fiscal relations), with contributions from Michel Noel and Svetlana Budagovskaya; Pablo Gottret (lead author o f Chapter 3 on health) with contributions from Adam Leive, Ajay Tandon, Katerina Maynzyuk, and Igor Oliynyk; Michael Mertaugh(lead author o f Chapter 4 on education) with contributions from Katerina Maynzyuk and Olena Bekh; and Oleksiy Balabusko (lead author o f Chapter 5 on local capital budgeting) with contributions from William Dorotinsky and Pablo Saavedra. Additional contributions to the report were received from RuslanPiontkivsky, Maris Jesse, Maria Koreniako, and Inna Lunina (consultant). Research assistance was provided by Yulia Makarova. Organizational assistance and document production was provided by Judy Wiltshire, Tetyana Komashko, Susana Padilla, and Yulia Kumetsova. This report was co-financed by the Department for InternationalDevelopment (DFID). The peer reviewers were Deborah Wetzel, h a n d Rajaram, and Mukesh Chawla. Martin Raiser was adviser to the team and provided detailed comments on intermediate outputs and on earlier versions o f the report. Asad Alam provided valuable guidance and advice throughout the preparation o f this report. The team is grateful to all government officials from the Ministry of Finance, the Ministry of Education, the Ministry of Health, the Ministry o f Economy, the State Treasury of Ukraine, and the State Tax Administration who cooperated with the team and provided comments on intermediate outputs of this report, including those inthe discussion meetingsof February, March, and June 2007. This report has also benefited from comments from World Bank colleagues made in internal presentations and informal discussions, including comments received from Amin Fidler, Arup Banerji, Edmundo Munugarra, Adam Wagstaff, HidekiMori, Katerina Petrina, AnushBezhanyan, MenahemPrywes, and Michel Noel. EXECUTIVESUMMARY 1. Ukraine spends a large amount of its budget resources on health and education (mainly through local governments). Moreover, these sectoral fiscal envelopes cannot be increased without harming the economy with an added tax burden. Ukraine's public fiscal envelope at 43 percent o f GDP i s too large for an emerging economy, i s biased towards consumption and current transfers, and i s low in public (fixed) capital investments (at 2.3 percent o f GDP in 2006).' On the hnctional fiont a significant share of spending is accounted by "social spending" (around 25 percent o f GDP), including health care, education, pensions and other social protection transfers. The budget spends 6.3 percent o f GDP on education and 3.7 percent o f GDP on health, and additionally, out-of-pocket spending (mainly done through public facilities) i s relatively high at around 1.5 and 2.5 percent of GDP in education and health, respectively. Table 1 provides an overview o f key macroeconomic and fiscal trends in TJla-aine and shows that the substantial social spending has been financed by a growing tax burden on the economy, which, combined with poor service delivery outcomes, may become a factor inlimitingfuture growth and competitiveness. Table 1: Key Fiscal (and macro) Indicators, 2002-06 2002 2003 2004 2005 2006 RealGDP (change in percent) 5.2 9.6 12.1 2.7 7.1 CPI, e.0.p.(change in percent) -0.6 8.2 12.3 10.3 11.6 PPG Debt (percent of GDP) 36.5 29 24.7 17.7 15 Fiscalbalance(percent of GDP) 0.5 -0.9 -4.4 -2.3 -1.3 Total revenues,as % of GDP 35.7 35.9 35.0 39.6 41.7 ohtotaltax revenue,as%ofGDP 29.2 33.4 35.9 Total expenditure,as % GDP (consolidated budget+ socialinsurancefunds) 35.5 36.8 39.4 42.0 43.0 Social insurance funds payments 9.0 9.2 11.4 15.2 15.4 Publicwage bill 7.6 7.4 7.6 8.0 Capital expenditures 4.3 6.0 4.1 4.6 olw (fixed)capitalinvestments 2.8 2.8 2.0 2.3 olw local (fixed)capitalinvestments 1.3 1.3 0.9 1.3 Expenditures in health, as % of GDP 3.3 3.6 3.5 3.5 3.7 olw local health expenditures, as % of GDP 2.6 2.7 2.5 2.7 3.0 Expenditures ineducation, as % of GDP 5.4 5.6 5.3 6.1 6.3 olw localeducation expenditures, as % of GDP 3.1 3.5 3.2 3.8 4.0 Local expenditures as % of total consolidated budget 40.6 37.8 43.2 Local expenditures as %of total general government(consolidated+ social insurancefunds) 34.2 32 30.2 29.8 30.1 Memo: GNI per capita (USD,Atlas method) 780 970 1,260 1,520 1,910 Sources: SSC; NBU; MoF; State Treasury; Bank staffcalculations. ' While fixed capital investmentsrecoveredslightly in2006, they remains low comparedto the new EUmembersand other middle incomecountries, with which Ukrainecompetes 1 2. However, for the large amount of resources the budget spends on services such as health and education, Ukrainians do not obtain good value. The health sector outcomes' are quite poor as shown by some indicators: while maternal and child mortality rates have improved in recent years, life expectancy remains below the pre-transition levels at 67.1 years, and male mortality (currently 61.7 years) has been worsening. At the same time, the incidence o f illnesses such as tuberculosis and HIV/AIDShas been growing. On balance, most o f the health indicators for Ukraine under-perform those for the new EUmember countries. Similar data on education performance are not available, since it was only in 2007 that Ukraine started to take part in international comparable standardized tests o f student performance. Nevertheless, there i s indirect evidence o f increasing dispersion in the quality o f learning combined with shortages o f skilled labor as reported by businesses. Moreover, Ukraine's score i s among the lowest o f all transition economies when it comes to rating the quality o fhealth and education services, as shown inthe 2007 EBRD-World Bank Life inTransition Survey. 3. In part, this is due to the acute inefficiency of service provision in these sectors, which generates under-spending on quality-enhancing expenditures and investments. In the education sector, Ukraine has acutely low studentlpupil ratios (at 9.4 for primary and secondary education, one o f the lowest ratios in the world), extremely small class sizes, and an oversized network o f facilities combined with a low teaching hours load per teacher (60 percent o f the levels inthe United States and 85 percent o f the OECD average). Inthe health sector, Ulcraine has a greater number o f health system inputs per 100,000 inhabitants compared to the new EU member states and the EU countries in general, including more hospitals, beds, doctors, nurses, and non-medical staff. This i s combined with an inflated average length o f stay in public facilities that i s associated with low utilization rates. Maintaining such inefficient network sizes and staffing leaves few resources to spend on the training and re-training o f teachers and medical personnel, the provision o f adequate educational and treatment materials and medicines, the renovation o f equipment and laboratories, and other infrastructure needs. 4. Local governments are essential in providing and financing services such as health and education, and in channeling public investments across the country: and thus reforms in these sectors need to go hand-in-hand with intergovernmental fiscal reforms. Indeed, local governments spend 80 percent and 64 percent o f the total budgets o fhealth and education, respectively, and provide the bulk o f the services in these sectors. Local governments are also important to other sectors inneed of reform, such as housing, communal heating, water, sanitation, transport, and social assistance. Local governments are also key in channeling fixed capital investments across sectors and across the country (spending 55 percent o f the total national budget inthis area). 5. Despite their large role in allocating spending, the ability of local governments to determine spending structure and to allocate expenditures within sectors is quite limited. The intergovernmental fiscal system (which i s the backbone o f local finances) didundergo positive reforms in 2001-02, but the reform process remains incomplete. Improvement across the four key pillars o f the intergovernmental fiscal framework i s necessary to enable the system to receive and allocate resources more efficiently, facilitating the proper financing o f local services and infrastructure provision. Some examples underscore this situation: (i)expenditure responsibilities overlap to some extent, creating inefficiencies and duplication at the local level, and many o f the lower tier government units are too small to deliver basic social services effectively; (ii)incentives to increase local government's "own" tax revenues are small and revenue autonomy i s very limited, and consequently local governments are over- dependent on shared revenues and transfers; (iii)there is a lack ofpredictability andtransparency ~~ ~ Ukraine has a health sector dominated by public provision; private health facilities are marginal in the health system of the country. Local governments funnel close to 43 percent o f the consolidated budget o f Ukraine (or one-third o f the general government budget which includes the consolidated budget plus pensions and the other three social insurance funds). They also spend 48 percent o f the total public sector wage bill. 11 combined with perverse incentives in intergovernmental transfers that affects the planning horizon o f local governments and their predisposition to overcome inefficiencies in service provision; and (iv) the framework for sub-national borrowing i s underdeveloped. Moreover, the parallel vertical structure o f management o f social services such as health and education continues to dictate how budgets are formed at the facility level and aggregated at the local government level (mainly on the basis o f inputsrather than demand), which dramatically reduces local authorities' budget flexibility for efficiency gains. 6. Concerns and fiscal pressures that merit reforms in local finances and in health and educationexpenditurepoliciesalso arisefrom other sources: 0 Ukraine's demographictrends present significantfiscal challenges in service delivery. Aside from the impact o f an aging population on pensions and the labor market, a rapidly aging profile signals the need for measures to anticipate and cushion fiscal pressures on the public health care system (including long term health care). Inthe education sector, the demographic effect creates a different need: the shrinking school population would require the rationalization o f the network and staffing to avoid further inefficiencies. 0 Regional inequalities in public spending and inequalities in and barriers to access to basic services among households are considerable. The differences in per capita total public spending across oblasts (regions) are small yet visible, but they mask larger inequalities in per capita public capital investment allocations. Highout-of-pocket payments inhealth and education generate significant inequities, which are combined with poorly targeted spending in these sectors. 0 Fiscal pressures arising from needed investmentsmay overwhelmlocal governments' fiscal capacity. This report estimates roughly (though conservatively) that public sector investment needs at the local government level (including those inservice delivery) are aroundUS$29 billion over the next 10 years (roughly 30 percent o f total estimated investment requirements in the country). The private sector i s likely to provide only a smaller portion o f these investments at the local level, and only if the regulatory framework for public-private partnerships (PPPs) improves significantly. Thus, even considering some private investments, local government spending on (fixed) capital investments would need to grow in real terms by at least 70-90 percent from its current level. District heating i s probably the service with the most pressing needs in terms o f physical investments, but water and sanitation services and regional and rural roads (and other transport infrastructure) also require significant refurbishing to avoid further deterioration. Investment needs are also significant in social services: these needs include the refurbishing o f deteriorated schools and hospitals, the purchasing of new equipment, laboratories, and the like. 7. Ukraine's budget does not needto spend more in service provision,but needsto spend better (more efficiently). Inthis context, the government needs to improve resource allocationby strengthening its intergovernmental fiscal system and by creating sufficient fiscal savings within the current fiscal envelopes o f the health and education sectors in order to re-allocate these savings towards quality- enhancing expenditures and investments withineach sector. 8. Creating fiscal space for local investmentsacross sectors is necessary, but to avoid wasteful spending, these efforts needto be coupledwith the strengtheningof the framework for local capital budgeting.The weaknesses o f the capital budgeting system at the national level are mirrored at the local level, and thus both the incentives and the capacity for the planning, evaluation, and selection processes that would allocate capital investments strategically are absent. ... 111 9. This report focuses on raising selected issues in the following areas: (i)improving the intergovernmental fiscal framework; (ii) overcoming fiscal, efficiency, and equity challenges in public health care spending; (iii) overcoming fiscal, efficiency, and equity challenges in public education spending; and (iv) strengthening local capitalbudgeting. (i)ImprovingtheIntergovernmentalFiscalFramework 10. Despite highly positive reforms in 2001-02, the unfinished reform agenda for the intergovernmental fiscal framework hampers the ability of local governments to deliver public services efficiently. Ukraine i s a unitary state with, broadly, three tiers o f local government, namely, oblasts, rayons and cities, and villages and settlement^.^ Local budgets are presently funneling 43 percent o f the total consolidated budget of Ukraine, including 48 percent o f wages and 55 percent o f public fixed capital investments. Although the implementation o f the Budget Code in 2001-02 helped make the decentralization process more coherent and significantly improved intergovernmental relations, Ukraine remains highly centralized fiscally, administratively, and politically. Looking forward, the reform agenda encompasses (further) clarifying expenditure responsibilities and strengthening local spending autonomy, improving revenue assignments, streamlining the equalization transfer system and reforming capital transfers (subventions), and improving the framework for local government borrowing. Expenditure Responsibilities and Autonomy 11. Efficiency in expenditure responsibilities among local governments is undermined by some overlapping of functions, significant losses of economies of scale in service provision (in certain services), and under-funded mandates. The 2001-02 budget reform helped to clarify local government expenditure responsibilities but failed to address three critical issues. First the ownership o f assets (Le., service facilities) inherited from the former system (and some functions they performed) did not change under the new system. Consequently, some overlapping o f functions remains, creating an inefficient use o f resources and hampering planning capabilities. Since fiscal transfers from the center are often facility based, there i s little incentive for consolidation (see below). Second, the average size inthe third (lowest) tier o f local government with council representation (1,400 inhabitants, approximately) i s too small to efficientlyplan and allocate resources inhnctions such as health and education (some o f which are currently performed at t h s level). Third, unfunded mandates have grown in the system since 2002, particularly in the areas o f social protection.' 12. Local governments have little discretion for expenditure allocations and administrative decision-making in services such as health and education, which undermines local accountability. Ukraine has a blended system o f de-concentrated and delegated expenditure responsibilities. Despite some notional spending autonomy at the local level, line ministries continue to exert significant top-down control over spending, thereby undermining local autonomy and accountability. The budget formation at the service facility level (inhealth and education) and its aggregation at the local budget are based on line ministries' "norms" that have their roots in the Soviet system. These norms dictate staffing levels and other resources needed on the basis o f facility characteristics (for example, number o f doctors based on the existing number o f beds in health facilities, or non-teaching staff based on the number o f square meters o f a school) and on the existing oversized network o f service facilities, rather than on the actual demand for services. Because o f the mandatory nature o f these regulations (and there may be harsh The top tier has 24 oblasts (regions), the Autonomous Republicof Crimea, and two cities with special status: Kyiv and Sevastopol.The secondtier is made up of 488 rayons (or districts) and approximately 177 cities.The third tier containsabout '12,000smalltowns,atsettlements and villages with alocal councilor "Rada." Socialassistance the local level isnot coveredspecificallyinthis report. A detaileddiscussionofthis issue canbefound in the DFID report "Facilitating Reformin SocialServices" (FRSSU) 2007. iv consequences for non-compliance) there i s a severe problem o f inflexibility in local budget allocations, which translates into high recurrent spending, particularly for wages, heating, and the like, leaving few resources for capital investments and quality-enhancing expenditures. Poorer local governments are particularly affected, which leads to inequalities in per capita public investments across localities. The problem i s exacerbated by the legislation that prohibits closing health facilities and schools, and by the difficulties local governments face infiring personnel. 13. Improvingthe structure and efficiency of expenditure responsibilitiesin local governments requiresthe implementationofthe followingset of selectedrecommendations: Strengthen the rayon (and city) level as the core local government. This second tier of local government has the right size in the context of Ukraine, and its administrative capacity has increased significantly (although there i s scope for further improvement). 0 Start the process of moving functions from villages and settlements up to rayons (and cities) as part of a comprehensive policy o f local government reform. The functions that should be switched are mainly those inhealth and education, including primary education and pre-schools, and local clinics and obstetric centers. This does not necessarily implythe closing o f facilities but rather passing control to the rayons so that they can plan and allocate resources (and manage provision) more efficiently.6 0 Arrange for small town, village, and settlement local governments to keep some local functions (including garbage collection, firefighting, sports and cultural activities, and routine street maintenance), and for the local councils of these localities to keep their current political representation. 0 Promote the consolidation of the third tier of local government. In doing so, close attention shouldbe paid to incentives by ensuringthat the resultingsavings remain at the local government level. However, this report cautions against the creation as part o f territorial and administrative reforms of a new local government below the rayon level with significant expenditure responsibilities as being too costly and inefficient. Consolidation should take place at the rayon level. 0 Stop the passing o f new social programs to be implementedby local government if they are not fully funded as verified by a detailed ex ante examination of costs (or as verified by pilot programs that establish actual delivery costs at the local level). 14. Granting a more adequate level of expenditure and administrative autonomy while at the same time strengthening efficiency and accountability requiresthe implementation of the following recommendations: 0 Grant rayons more autonomy to allocate and re-allocate staff and other resources, and to merge and close service facilities within their jurisdictions by lifting sectoral "norms" and legal restrictions. 0 Place funding and strategic administrative decision-making regarding service facilities in the hands of the rayons and cities.' Infact, this is the case sporadicallyinsomerayonsinwhich villageshavegivenuptheir responsibilities.The legalframework for this process already exists inArticle 92 (and inArticles 89 and 93) of the BudgetCode, althoughminor amendments would be needed 'At present, providing too muchpower at the facility level would create vested interests and would become an obstacle to the neededfacility rationalization of the network, but more day-to-daymanagerialdiscretioncould be handed to facility managers. Over time, however, with a more optimal size of the networkof service facilities, facility managers shouldbe more accountable for their performance and consequentlythey would needdiscretioninhiring andfiring personnel. V Revenue Assignments andAutonomy 15. The shared revenues (the main source of localrevenues) lack predictability and encompass some inequalities.The bulk o f local governments' revenues come from shared taxes such as the PIT and the land tax (and transfers such as the equalization transfer and capital subventions/transfers- discussed later). Over the last years the central government has often changed the base and rate and has granted tax exemptions inall the shared taxes. Butthe compensations for revenue losses to local governments that are due to these central tax policy changes are ad hoc, reducing predictability and transparency and preventing medium-term planning at the local level. In addition, the way in which the shared PIT i s allocated to local governments may be generating inequalities: a share o f this tax i s allocated to the locality in which the individual i s employed (the employer's locality) not the locality in which the individual resides, and since businesses tend to locate in more urban and affluent areas, poorer districts may be adversely affected. 16. Local governments' limited autonomy over "own" local taxes is coupled with the high administrativeand compliancecosts of these taxes. Only a small part, on average less than 3 percent o f total local revenue, comes from "own" local taxesa8 Given the small revenue mobilization o f own local taxes and the rigidities embedded in how all revenues are spent, local governments are increasingly dependent on transfers from the central government. The rate for most taxes on the list o f own local taxes i s fixed centrallyg which undermines local tax accountability, although local governments can choose which "own" local taxes to levy intheir jurisdictions from a specified list. Because taxes are collected by the State Tax Administration on behalf o f local authorities, there i s little incentive to worry about collection costs (and local governments typically choose all o f the taxes on the list), but several local taxes may actually have negative net yields. The communal tax overlaps, inprinciple, with the enterprise profit tax and the single tax. The estimation o f tax liability in the market fee i s complicated, leading to highadministrative and monitoringcosts. Currently, a property tax does not exist inUkraine and existing proposals for its establishment deviate significantly from best international practice. In general, local taxes and fees are needlessly complicated intheir structure and have a highfrequency o f payments, which contributes to the hightax compliance costs inthe country. 17. Providing greater predictability and fairness to the local revenue stream and reducing the administrative and compliance costs generated by local taxes requires the implementation of the following set of selectedrecommendations: 0 Establish clearer and stable criteria (in the regulation) for compensating local governments for revenue drops caused by central tax policy changes, and stop the granting o f unilateral tax exemptions inthe land tax and the PIT. 0 Consider allocating the PIT to the jurisdiction in which the taxpayer resides rather than to the jurisdiction inwhich the employer has its establishment." Create the legislative framework for a property tax in line with good international practice (particularly as regards setting the tax base and rates as well as establishing a proper fiscal cadastre together with the mechanisms for its administration). Allow local governments to change the rate o f own local taxes (and fees), including through flexible bands at the beginning. Granting local governments the ability to change the tax bases on "Own" local taxes are those that are on the list o f taxes levied by the State Tax Administration exclusively for local governments. They are composed o f minor taxes and fees. More recently, in a few local fees there is a certain amount o f flexibility, but local governments still have no autonomy for setting rates in own local taxes in general. lo This would also influence the estimation o f revenue capacity in the equalization transfer, and consequently it should be considered intandem with its potential effects on recalculating this factor over the next two years. vi own local taxes (that is, to choose taxes outside o f the established list) is not recommendable at thispoint inthe reformprocess. Revisit the way inwhich the communal tax i s levied. An easier way o f collecting revenues might be to increase the share o f local governments inthe single tax. Revisit the way o f levying the market fee. Simply let local governments decide on a fixed fee in UAH per square meter leased per month (within a range or band, and with automatic annual indexation for inflation) without providing any additional criteria. Avoid the overlapping created by the "trade and service object placement duty." There are two ways to eliminate the duplication created by this tax: (i) by creating a special section under the marketplace fee to tax owners o f trading places; and (ii) by levying the communal tax on legal entities and individuals under the Simplified Tax System (STS). Eliminate local taxes that generate higher administrative costs than revenue yields." Examine these recommendations in the context o f the overall tax burden o f the country. The aim to generate a higher share o f "own" taxes inthe revenue structure o f local governments so they do not become over dependent intransfers and subventions from the state budget. Intergovernmental Fiscal Transfers and Subventions 18. The equalization transfer has a reasonable equalization power (which has not changed significantly since 2001 despite dozens of factors added to the formula over the last years), but it is underminedby poor predictability and transparency, ill-designedincentives,and the effect of other transfers. The equalization transfer (from the central government to oblasts and rayons) i s the main fiscal transfer inthe system." As Figure 1shows, the transfer system does have some equalizing power; nonetheless, the following shortcomings should be highlighted. The large number o f factors and adjustment coefficients to the formula that determines the size o f the transfer has created complexity and transparency concerns and has opened the door to political bargaining, as local governments lobby for the inclusion o f factors to fit their own specific characteristics. Moreover, critical components o f the formula have undergone discretionary changes that, even if well-intentioned, have affected the formula's predictability and 0bje~tivity.l~Many o f the hundreds o f factors and coefficients in the estimation o f "expenditure needs"14 provide "wrong" incentives for local governments, as they help to perpetuate the current inefficient levels o f service provision. For example, several coefficients tie financing to facility network characteristics and staffing levels and thus finance implicitly the existing network o f service facilities and staffing as opposed to funding a pure demand for services.1s There are several factors inthe formula accommodating notional finding for specific programs across sectors that in fact have n o place in a block ("non-earmarked") grant such as the equalization transfer. The equalization formula between rayons and third tier local governments reflects the same shortcomings. I'Current candidates for curtailment include: (i)the fee for an apartment occupation permit; (ii) fee on horse racing bets; (iii) the the fee on the racingprizes; (iv) the fee for television and film shootingpermits; (v) the fee/permit to conduct local auctions; and (vi) the fee imposed on pet (dog) owners. The draft tax code follows some o f these suggestions but falls short inthe streamlining. The equalization transfer uses a gap filling/ "fraternal" approach to equalization where local governments receive or contribute the difference between a notionalestimation o f expenditure needs capacity and revenue capacity. l3Transparency in a transfer means not only that the transfer needs to have criteria but also that all local governments shouldbe able to understand the workings o fthe formula and to calculate/forecast the amount that they can expect over the following years. l4The "expenditure needs" side o f the formula contains components to estimate a notionalneed o f resources for the following six areas: health, education, welfare and social protection, culture and the arts, sports, and administration. For example, there i s an adjustment coefficient that allocates a greater amount o fresources to local governments that have a higher ratio o fwages to total expenditures inthe health sector: this provides an incentive for keepinghighlevels o f unneeded staff. There are a variety o f factors and adjustment coefficients designed to allocate fundingper pupil in education, but they are attached to the type o f institution the pupilattends. Yet another factor grants two-thirds more resources per student for students insmaller class sizesthan the average. vii igure 1: The Equalization Formula: Equalization I wer and Trends. EqualizationPower of the Equalization PowerTrend (By Oblast Consolidated) EqualizationTransfer (by 450 rayons and cities) 400 350 * 2005 2500 L Total `Z300 Revenues n 250 p' 200 53 -2001 150 100 50 0 `ote: Rayons and cities are ordered from higher incomes to lower incomes (from left to right).This study used PIT per capita collections as a proxy for income level by rayon. Oblasts, ARC and cities o f special status are ordered from higher to lower level (from left to right) per capita regional (oblast) value added. Sources: MoF; State Treasury of Ukraine; State Statistics Committee of Ukraine; World Bank staff calculations. 19. The equalization transfer formula should be revised, aiming for a comprehensive simplification and elimination of ill-designedincentives;the following are key recommendations: Streamline the number o f factors and adjustment coefficients in all six components o f the "expenditure needs" estimation (namely, health, education, social care services, culture and the arts, sports, and administration). The goal for each component should be to have the following two characteristics: (i) more than four to six factors based purely on demand driven factors not and not related in any way to facility characteristics (in the case of education they should reflect enrollment as well as school age population in a locality; in the case o f the health component a small number o f age groups would be adequate); and (ii) the financial allocation per head (or per pupilineducation). Start the process of estimating appropriate per head standard costs for different services to improve the equalization formula (called factor "N' in the legislation of the equalization formula). Eliminate factors gearedto fundingspecific programs ineach sector within the equalization grant, which i s by design anon-earmarked block grant. Condense the dozens of adjustment coefficients accounting for mountainous regions into a single coefficient. Moving forward, however, start collecting and analyzing data on the differences in the costs ofdelivering services across local governments. Stop making changes to the way in which revenue capacity i s calculated. For the present, and until detailed data on tax bases are collected to perform forward looking estimations, the estimationshould adhere to that proposed by the Budget Code o f 2001. Create a proper coefficient of fiscal effort (based on the estimation of collections over potential tax bases ineach local government), but inthe meantime use the mechanism articulated originally inthe Budget Code (that is, freezingrevenue capacity for a period of time: namely, two to three years). Keep the "alpha" coefficient at the same level for all local governments. ... V l l l Simplifythe equalizationtransfer from rayons to villages and settlements, inparallel with making the rayon the core local government. 20. Capital transfers to local governments (called "capital subventions") in Ukraine are not transparent and lack consistent and stable criteria for allocation. These allocations are the main source of capital expenditures at the local level. But there i s a lack of predictability for receiving these funds and also a tendency to favor wealthier regions (Figure2). Since there is no certainty, not even for a minimum amount in a given year, local governments are unable to plan investments and renovations consistently and would notbe able to implement a MTEFinthe near future. Figure2: RegionalInequitiesin CapitalExpenditures Combinedwith Poor Predictability 2 200 PublicCapitalExpendituresby Oblas120024005 Predictibility: Percentage Change in Capital 3 E 180 200 Expenditures Allocation Year to Year (By Oblast g 160 m 2 140 2120 .e 3 100 32n - Q8100 80 k 8 50 60 8n 40 0 0 ._ 21 - -2004-2005I 0 I..*-. 2002-2003 +2003-3004 rlote: Oblasts, ARC and cities o f special status are ordered from higher to lower (from left to right) per capita regional (oblast) value added. Sources: MoF; State Treasury o f Ukraine; State Statistics Committee o f Ukraine; Bank staffcalculations. 21. The system of capital investment subventions/transfers requires greater efficiency, transparency, and fairness. As a first step towards achieving these objectives the government should establish in the legislation a formal criteria-based and stable system of capital investment transfers/subventions.These transfers should reach the first and second tiers of local governments. The following specific steps can be considered: Establish a pool of resources for capital subventiodtransfers. This pool (which should be increased from the present level o f local capital spending) could be a notional share o f the consolidated budget (or the total local budgets). It could then be further divided into sub-pools (for at least two types o f capital transfers) as suggestedbelow. Consider establishing a minimumtransfer for (fixed) capital investments to oblasts and to rayons and cities based simply on the number of inhabitants ineach oblast or in each rayon and city (in the case o f the second tier),16 to be used to cover repairs and renovation needs in the current infrastructure. The objective would be to help stop the acute deteriorationo f assets. Consider establishing a subventiodtransfer (potentially larger than the first proposed) granted for critical ongoing and new projects on a competitive basis. This second transfer could be tied to the preparation of multi-year local investment programs (or to individual projects), and would follow clear criteria for project evaluation and selection. These criteria should be based on factors such as national priorities, positive externalities acrossjurisdictions, and clear economic l6Perhaps including a "soft)' (that is, close to 1) adjustment coefficient recognizing the higher cost o f investments in mountainous areas. ix benefits (as articulated in a cost-benefit analysis). Inthe case of this second transfer, rayons and cities could apply either through the oblasts (with each oblast screening and consolidating proposals) or through direct applicationto the central level. Local GovernmentBorrowing 22. The local government borrowing framework i s undermined by holes in the regulation and problematic incentives. Local governments and communal services enterprises may need to borrow in order to finance their large infrastructure rehabilitation and development investment needs over the next 10 years. But the current local government borrowing framework has several limitations. First,it does not include debt refinancing as a legitimate purpose o f local government borrowing, which inmany cases i s a rightful operation that allows more flexible (longer-term) financing of capital investments. Second, no framework exists for local government bankruptcy, which creates moral hazard and may encourage reckless lending. Third, the legislation does not provide the authority to cities and rayons (or oblasts) to pool their resources and borrow under joint undertakings. Fourth, short-term debt arising from "accounts payable" with contractors and providers i s not included inthe concept o f local government debt and thus i s not subject to monitoring by the Ministry of Finance (MoF), raising the risk of excessive debt buildup. Fifth, the MoF letter (order) of September 2004 allows local governments to open regular bank accounts outside the Treasury Single Account (TSA), opening the risky possibility of local governments pledging large revenue streams; this "hole" in the legislation creates potential fiscal risk and uncertainty. Sixth, Resolution 208 stipulates that all debt securities are subject to mandatory credit rating, but this complicates the process and may be unnecessary for many local governments. Seventh, there are no limitations on related party lending between banks owned by municipalities and the municipalities themselves. It should be noted positively that the MoF's ex ante controls for borrowing are prudent and justified in the short and medium terms given the limited development of the market and the limited experience of the local governments inthis area. 23. Strengthening local government borrowing framework requires the implementation of the following recommendations: Implementa local government regulatory framework for bankruptcy. Apart from reducing moral hazard in sub-national borrowing, this would provide adequate incentives for lenders to be conservative intheir analysis. Allow debt refinancing as a legitimate purpose ofborrowing. Allow local governments to pool their resources injoint undertakings and to borrow jointly to take advantage of synergies and economies of scale inproject financing. Revise legal provisions regarding the issuance o f local government guarantees for projects with minority private sector participation and allow partial credit guarantees. In parallel, the government may have to introduce comprehensive amendments to the Public-Private Partnership (PPP) Law to regulate all types of PPP transactions. Revise the current regulations and planned amendments to include accounts payable as debt. This i sjust a precautionary measure, since the TSA provides tight spendingcontrols. Improve the ex ante MoF controls by setting rules to avoid unnecessary delays and to foster transparency in the authorizationprocess for borrowings. To avoid delays, for example, the MoF should formally notify local governments o f its decision to deny authorization within a clear time frame (for example, 15 working days). Decisions to deny authorization (and justification for the denial) should be made public. Over the medium term, and only with more maturity in the system, measures prior to registration could be revised, but registration needs to remain inplace, as do prior controls for borrowings inforeign currency (and obviously for borrowings requiringa X sovereign guarantee or counter-guarantee). In addition, the MoF could implement a periodic controVmonitoring of compliance with overall debt ceilings. Clarify the situation regarding revenue pledges so as to prohibit pledges of shared taxes (such as the PIT, the land tax, and the single tax) and revenues coming from the equalization transfer and other subventionsh-ansfers from the central government. Only "own" local taxes and fees should be allowed to be pledged. The rationale behind this recommendation i s that local budgets still face serious rigidities and pressures infulfillingtheir mandates. Consider abolishing Resolution 208 in order to let borrowers and local governments decide whether or not to obtain aratingbased on market incentives. Allow lending to local governments only to banks whose ultimate beneficial owners are fully disclosed. Other Issues Concerning the Intergovernmental System 24. There are other issues concerning the intergovernmental fiscal framework that require urgent measures.First, the current ad hoc government practice of granting selected local governments interest-free Treasury loans softens budget constraints, hampers the development of the sub-national borrowing market, and harms transparency, for the following reasons among others: (i) treasury loans are open to political pressures and patronage; (ii) they might encourage the expectation of being written-off by the MoF; (iii) constitute a ready instrument to bail out local governments in case of default; and they (iv) they can lead to moral hazard in the local government debt market. Second, the overall pool of resources for local governments' transfers and subventions i s not fixed. Every year during the budget process the central government establishes the pool of total resources for each transfer and special subvention for local governments, which reduces predictability. However, the decision regarding whether or not to fix a transfer pool needs to be made within the context o f fiscal responsibility rules for local government and the downsizing of the government as awhole. 25. The followingrecommendationsshouldbeconsidered: Stop granting medium-termTreasury loans to local governments and phase out short-term (three month) loans for liquidity. As an option the government could consider keeping short-term loans only in special cases, such as for smaller local governments (but establishing a clear threshold) that would not have access to market resources and under specific liquidity emergencies. Moreover, these loans would need to be fully integrated into the intergovernmental fiscal framework (for example, by reducing other transfers receivedby local governments if short-term loans are not repaid intime). Establish a clear rule for the "minimum" pool o f funds for transfers and subventions for local governments. This rule might be articulated through a specific share o f the total consolidated budget,but inthe context of hardbudget constraints for local governments. The idea would be to have a "floor" and a "ceiling" for the level of total funding for local governments. Having a share as an indicator o f the pool o f resources would nevertheless imply that local government spending may be downsized in tandem with the needed reduction o f the overall fiscal envelope of the country. 26. Suggested sequencing: Since the intergovernmental fiscal framework i s mostly established through the Budget Code, the ongoing revision to this law presents an excellent opportunity to include (during 2008) most of the recommendations suggested that require legislative changes, including those related to the following: (i) expenditure responsibilities (that is, moving up responsibilities to the rayon level and granting rayons and oblasts stronger administrative powers); (ii) revenues; (iii) shared capital transfers; and (iv) the local government borrowing framework. Recommendations on "own" local taxes xi can be legislated through the Tax Code under preparation and recommendations on the equalization transfer (and detailed recommendations on capital transfers) through amendments to the corresponding Cabinet o f Ministers Resolutions (all during 2008). Revisions and changes to sectoral regulations should start immediately (so that pilot testing, ifneeded, can be undertaking during2008). Implementation could start in 2009 (see overall sequencing o f recommendations across chapters in Table 2- at the o f the Executive Summary) (ii)OvercomingFiscal,Efficiency,andEquityChallengesinPublicHealthSpending 27. Health sector outcomes are poor despite significant available resources, which underlines problems of inefficiency as well as inequities in spending and in access to health services. Ukraine's national health system i s financed by general revenues from the budget, and the overall provision o f health care i s mainly carried out through public facilities." While maternal and child mortality rates have improved in recent years, life expectancy remains below pre-transition levels at 67.1 years (and with a widening gap between male and female mortality, at 61.7 and 72.7 years, respectively). Non- communicable diseases continue to be the dominant causes o f the disease burden and death (particularly cardiovascular diseases). A number o f health status indicators have deteriorated (for example, the mortality o f males aged 25-39 and the incidence o f tuberculosis and HIV). On balance, most o f the indicators for Ukraine under-perform those for the EU-10 countries.I8 Inthe coming years Ukraine will face additional challenges such as taking care o f a large (and aging) cohort o f now middle-aged people with highlevels of morbidity, which mightplace significant fiscal pressure on the health system and on long term care." 28. Public spending is inefficiently allocated across levels of health care. Public spending inthe health sector as a share o f GDP i s moderate (at 3.7 percent o f GDP as o f 2006) compared regionally and with new members o f the EU, but is above average compared with countries at the same level o f per capita income. Nevertheless, when out-of-pocket payments (OOP) are added, total spending on health care reaches a level between 6 and 7 percent o f GDP. Often, there i s no clear distinction between primary and secondary level health care, and the overabundance o f specialists compared to physicians working at the primary care level is striking. Correspondingly, most public health spending is concentrated on hospitals and specialized health facilities (close to 70 percent o f the health budget) and only a small fraction onprimary health care (and prevention). 29. Public health spending (mostly undertaken at the local level) is severely constrained by the Ministry of health "norms" that generate further (allocative) inefficiencies. Even though most health spending takes place through local governments (around 80 percent), they have very little control over expenditure allocations. Health "norms" mainly contained in Order 33 issued by the Ministry o f Health (MoH) generate severe rigidities in the delivery o f health services by dictating the number o f doctors, nurses, and other staff to be hired on the basis o f fixed network parameters such as number o fbeds. These regulations, inturn, increase recurrent spending (wages and salaries constitute the largest share o f public health expenditures and drive the increase in the budget envelope in this sector), squeeze the share o f quality-enhancing expenditures, and reduce the responsiveness o f the system to local health care needs. "OtherprivatefinancingsourcesareminimalinUkraine.Voluntaryhealthinsuranceandsicknessfundsconstituteaverysmall share o f total health expenditures. Less than 2 percent of the total population i s covered by voluntary health insurance and care is still providedmostly at public health facilities, since private sector health care is extremely limited in size. The new EUmember countries, hereinafter called the EU-10, are: Cyprus, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Malta, Poland, the Slovak Republic, and Slovenia. As the World Bankregional report "From Red to Gray" (2007) argues, while actual increases inpublic health spending owing to an aging population may be marginal, other factors, such as the health status o f the elder cohorts combined with the demographic profile, can interact to produce serious fiscal pressures. xii 30. Health provision (technical) inefficiencies are acute. Ingeneral, Ukraine has a greater number o f health system inputs than EU countries. Ukraine has a large number of hospitals: 5.6 per 100,000 people in 2005. This i s in contrast to averages o f 2.6 per 100,000 for EU-10 countries and 3.2 per 100,000 for other EUmembers. Inaddition, inUkraine there were 868 hospital beds per 100,000 people in 2005, compared to 644 for EU-10 countries and 571 for other EU countries. Regarding health personnel, there were more doctors inUkraine (302 per 100,000 in 2005) than in EU-10 countries (261). And Ukraine had more nurses per 100,000 (782) than did EU-10 countries (548) or other EUmembers (750). The average lengthof stay (ALOS) in Ukraine was about 15 days, higher than the CIS average and much higherthan the EUaverage o f 9 days. There i s a constitutional provision against closing health care facilities, and consequently local governments are bound to financially maintain an oversized network o f health facilities, and ALOS are artificially inflated to compensate for lower utilization rates in order to maintain spending allocations on materials and other inputs. Local governments face severe constraints infiring personnel, which adds rigidities to the system. 31. There is budget fungibility in aid resources for health. Donor assistance in health care is focused primarily on HIV/AIDS, tuberculosis, and maternal and child health (and donor support has been increasing). However, the government's own spending on these programs has been shrinkingin parallel with increasing donor finding. Difficulties for channeling donor funding to local governments are significant and some donors (particularly through NGOs) are creating parallel systems o f local funding underminingthe country's intergovernmental fiscal framework. 32. Out-of-pocket payments (OOP) in health are substantial in Ukraine and carry significant inequalities; While OOP estimations vary according to the source and methodology, they amount (conservatively) to between 2.3 and 3 percent o f GDP.20 OOP creates inequalities inhealth sector access in Ukraine. The size and the likelihood of direct OOP financing rise considerably with the level of specialization of care needs, and the largest share of OOP expenditure in every oblast is comprised of prescription drugs. Inaddition, informal, unofficial payments are common inUkraine. There i s evidence that the increasing incidence o f informal payments i s a deterrent to health care utilization, particularly for the poor. Providers extract high OOP partly to compensate for low and highly tied government health spending. The poorest quintile i s the least likely to use outpatient, inpatient, and preventive services. Catastrophic health expenditures2' are highand heavily concentrated among the poorest income quintiles (see Figure 3). Compared regionally and with OECD countries, OOP are more regressive inUkraine and have the potential to pusha large number of people below the poverty line. Inequalities generatedby OOP may also have a regional-horizontal dimension whereby people at the same level of income pay higher shares o f OOP insome regions than inothers. Public health expenditures exacerbaterather than offset the inequalities created by OOP. Indeed, 70 percent of total public health expenditures accrue to hospitals, specialized institutions, and sanatoriums, but the poorest quintiles o f the population use these facilities much less often than the richer quintiles. Public health expenditures also have regional and local inequalities, as they are positively correlated with higher per capita oblast and rayon income levels (see Figure3). The full value of OOP expenditures is composedof official user charges, drugs purchasedoutside of healthclinics, under-the- table payments, and costs of transportation, food, andlodging. 21Catastrophic spendingis calculatedusingthe methodologyproposedby Xu (2005) andis definedas OOP spendinggreater than 40 percent ofnon-subsistencespending. ... X l l l Figure 3a,b,c: Public Health Spending Inequalities across Local Governments and Inequities Arising from ,,I Per Capita Public Health Spending By rayons 1 Percentageof Ukranianhouseholdswith and Cities I catastrophicOOP spendingby incomequintile I 25.0% T 300 -Average I 20.0% 15.0% 10.0% 5.0% 0.0% lhomquinlik Richest 0.020 0.015 - A -- 0.010 - poor 0.005 0.000 -0.005 -0.010 -0.015 Ukraine -0.020 ote: 3 (a) Rayons and cities are ordered from higher incomes to lower incomes (from left to right). This study used PIT per capita collections as a proxy for income level by rayon. Sources: MoF; State Treasury o f Ukraine; State Statistics Committee o f Ukraine; World Health Survey, 2003; Bank staff calculations. 33. A key contentionof this report is that additionalbudget resources will not enable Ukraine to remedy its health care system problems until the shortcomings of the current system are resolved. By the same token, the current health insurance reforms under debate, which also focus on raising additional sources o f financing for the health sector, would not solve the fundamental distortions, inefficiencies, and inequities that underlie the health system but might create an additional tax burden and cause wasteful spending. Thus, priority reform efforts (in a first stage) need to be focused on addressing the inefficiencies of the current system. 34. The fiscal savings resulting from greater efficiency in spending are potentially large and could be re-allocated within the health sector towards quality enhancing expenditures and investments.A simple estimation suggests that reducing the number o f hospital beds (and with them the niimhernfnhvsicians and niirses) tn the FTT-1 n nr FTT 1 ~ v ~wniild vield cavinoc nfn 75 and l c ?dnPrrPnt xiv of GDP (per annum), respectively. The savings could reach up to 0.4-0.6 percent o f GDP (per annum) if facilities were merged with corresponding reductions in staff and utility expenses. Even gradual adjustments towards this goal would free enough resources to gradually meet the investment needs estimated inthis report. The following are selectedoptions for reformand recommendations: Revise and eliminate (or reformulate) the norms contained in Order 33 of the Ministry of Health for budget formation at the facility level. The guiding principle for the revision should be to identify all norms that dictate health spending and the allocation o fresources such as medical and non-medical staff and materials on the basis of existing inputs (e.g., beds) and eliminate them. If the government desires to replace the norms, thenthe right vehicle is the establishment ofpurely demand driven factors as a guideline. These factors could evolve by pricing unit costs of each service in line with the "N'factor of the equalization formula, that is, a minimum standard per unit of service (in UAH terms). These units costs, in turn, would help to attempt more sophisticated financing mechanisms in the future such as contract agreements for service packages (with outcomes as measure of results) or other schemes that would allow the government to pay services and obtain better health outcomes for the population as a retum. Nevertheless, eliminating the norms currently dictating budget formation at the facility level i s the first step inthe sequence o fmodernizing the financing o fhealthcare services. 0 Turn norms that assure a minimum standard of quality (not related to financing rigidities or staffing) into criteria for the accreditation ofhealth facilities. 0 Givelocal governments the power to start rationalizingthe network o fhealthfacilities. Although the constitutional prohibition of facility closure is an impediment that would be difficult to overcome, a revision should be attempted. In the meantime, the government could begin the processthrough the "merging" o f facilities, which i s not prohibited. 0 Address the inequalities o f the system by offering poor households some financial protection from the costs of illness. The following are potential mechanisms that couldbe considered: o Allocate greater expenditures to facilities that the poor prefer to use (such as rayon level polyclinics and rayon hospitals). This would help these facilities to purchase the necessary quality-enhancing supplies so that poor households would not be forced to finance them through OOP. Capital expenditure, which i s often quite low and indicative o f quality, could also be directed from central government funds to these institutions. o Start a program for purchasing prescriptiondrugs that targets the poor usingthe fiscal savings resulting from the abolition of input norms. The health care facilities mentioned above could be a vehicle for implementingthis targeting. 0 Provide adequate incentives for shifting a larger proportion o f doctors into primary care and for integrating the family medicine model more strongly into Ukraine's health system. A key component o f any healthreformwould be better integration o f and coordination between primary, secondary, and tertiary care to allow for more cost-effective service provision combined with the establishment o f a stronger gate keepingrole at the primary care level. 0 Increase the prevention program expenditures that relate to HIV/AIDSand tuberculosis. 0 Arrange for additional financing from external donors to be undertaken through mechanisms geared to increasing the current fiscal envelope o f selected activities in coordination with the government. To avoid having donor resources simply replace budget funding, donors should require the government to maintain the current level o f spending as a percent o f GDP in target areas (as o f 2007) as a minimum floor, and then to provide financing on top of that minimum floor level through a matching grant (or lending) scheme. 0 Over time, and under close control of the State Treasury, local governments shouldbe allowed to pay additional incentives (such as bonuses) to attract doctors and to rewardperformance. xv 35. Sequencing: The first priority i s to start the revision o f health norms and refurnish Order 33 eliminating those norms that create budget rigidities by early 2008. Piloting o f measures to replace the norms with an appropriate "N" factor in the equalization formula (that is, a standard unit cost for each service), ifneeded, could be undertaken during2008 and 2009. The preparationo f the 2009 budget could encompass most o f the recommendations above. (iii)OvercomingFiscal,Efficiency,andEquityChallengesinPublicEducationSpending 36. Public spending on education (mostly undertaken through local governments) i s high (and is increasing), and is severely constrained by the own "norms" of the Ministry of Education and Science (MoES) which generate allocative and technical inefficiencies within the sector. In recent years, public expenditures on education in Ukraine have been increasing, outpacing in growth all other functional categories and reaching 6.3 percent o f GDP in 2006 (from 5.3 in 2004).22 This level i s high compared regionally and in comparisonwith countries with a similar per capita GDP. Local governments spend more than 64 percent o f the public education budget o f the country. While in theory local governments are in charge o f all o f the functions o f education except for higher education, the MoES continues to retain full control over norms governing staffing arrangements, teaching hours, non-teaching staffing ratios, class sizes-all based on the oversized network o f schools and school infrastructure characteristics instead o f on the actual demand for the service (e.g., enrolled children or school age population inthe jurisdiction). Therefore, these norms, inactuality, dictate how budgets are formed at the school level and, consequently, at the local government These arrangements further enlarge recurrent spending (most o f the upward trend in public spending and unit costs in the sector has been driven by wage increases) and do not allow enough fiscal space to reallocate resources to direct quality- enhancing expenditures such as the training and re-training o f teachers, better teaching aids, upgraded computers, internet access, and laboratories and to needed investments to improve the current infrastructure and to modernize the sector in general. Moreover, rising spending levels have exacerbated the existing inefficiencies rather than financing the quality-enhancing expenditures cited above and making the system more responsive to the labor market needs. 37. Ukraine shows alarming levels of (technical) inefficiency in the use of resources in education provision. Studendteacher ratios have been falling dramatically recent years, reaching 9.4 (average for 2005) for basic and secondary education. This ratio i s one o f the lowest in the world (almost half o f the ratio for OECD countries) (see Figure 4) and i s the result o f poor adjustment to falling student numbers. While the school age population fell by 25 percent over the last eight years, the number o f teachers declined by only 6 percent. At the same time, total teaching hours per teacher are only 60 percent o f the levels in the United States and 85 percent o f the OECD average. Currently, around 15 percent o f the teaching staff are between the ages o f 50 and 55, and there i s a large portion (close to 30 percent) aged between 41 and 50. However, the potentially positive effect o f attrition as retiring teachers leave i s being offset by the rehiring of retired teachers in order to fulfill the current staffing norms. The ratio o f non- teaching staff to teaching staff i s highly inefficient (at 0.7 for regular schools, 1.5 for pre-schools, and between 2 and 3 for special schools)24 and it i s driven by the same input norms. Other constraints embedded inthe system, such as the prohibition to close schools and the burdensome processes for hiring and firing make it difficult for local governments to improve efficiency through the consolidation o f school networks and the reallocation and restructuringo f staffing. Total enrollments will continue to fall 22Public spendingon education absorbs ahighproportionofpublic expenditures: 20 percentofthe consolidatedbudget and 15 percentof the generalgovernment budget. It shouldbeborne inmindthat the consolidatedbudgetplusthe extrabudgetary social insurance funds together form the notion of generalgovernmentbudget. 23These norms have their roots inthe Soviet system ofplannedeconomyand they have not changed, despitefinancing changes intheBudgetCodereform in2001-02. 24Estimationsbasedon the HBSsurvey and on the World Bank educationfacility survey(2006). xvi over the next decades, which will lead to even greater inefficiency and will create greater pressures for r'eform. Figure 4: High Spending and Low Efficiency in Public Education Spending Combined with a Shrinking School Age Population PublicExpenditurein Educationas % of GDP 35 , I StudentsReacher Ratio 2005 25 ~,000,000 ___ _-_______- __- -_ --..- -- 5,000,000 -5 to 14 years 4,000.000 15 to 17 3,000,000 -0-18 to 23 2,000,000 1,000,000 0 1 I 2005 2010 2015 2020 2025 I ources: OECD; Eurostat; WDI; United Nations Population Division; SSC; State Treasury of Ukraine; Bank staff calculations 38. Out-of-pocket payments (OOP) combined with poorly targeted public spending generate inequalities, problems with access, and governance concerns in education. This report estimates that OOP in education may total 1.5 percent o f GDP, o f which around half is payments in pre-schools, primary, and secondary levels (the remaining portion is inhigher education). But unbundlingformal and informal payments presents significant challenges. Overall, household outlays for education in pre- schools and primary and secondary schools are about as high in the poorest income quintile as in the richest quintile. Additionally, OOP payments seem to rise with the quality level o f schools (such as gymnasiums and lyceums), and such elite schools are more likely to be attended by children from richer households (the income o f parents with children in these schools tends to be five times that of parents xvii i with children in regular schools), which indicates potential inequalities in access. Moreover, the government (through the equalization formula and through specific allocations) spends significantly more per student in elite schools than in regular schools. With the aim o f procuring a higher quality o f education, parents make substantial formal and informal payments but do so in an environment with little oversight and accountability. Reliance on parental contributions and other non-budget revenues often creates perverse incentives, such as the incentive for teachers not to cover the complete curriculum in a class in order to create a demand for paid, extra-tutorial instruction and other educational a c t i v i t i e ~ . ~ ~ Overall, education coverage i s lowest for low income households and in rural areas. Inhigher education, there also significant equity and governance concerns about informal OOP geared to obtaining admission to universities, particularly in an environment where admission examinations vary significantly across institutions and where budget support to students i s poorly targeted. 39. The deterioration of education quality is a concern in Ukraine. Until recently there were no standardized performance tests that would allow an assessment o f the quality o f educational outcomes. In 2007, Ukraine implemented the first internationally recognized standard test (Trends in International Mathematics and Science Study - TIMMS), which would, over time, allow an assessment o f educational outcomes. Nevertheless, at this point there i s indirect evidence o f increasing dispersion in the quality o f the learning environment as a result o f increased reliance on non-budget financing, which i s consistent with widespread reports o f deteriorating quality. Many important educational quality-enhancing inputs, such as teachers training (and re-training), educational equipment and materials, programs for poor students and at-risk students, and school maintenance, are usually under-funded. Moreover, when it i s compared regionally, there appears to be an unusually highlevel o f dissatisfaction with the quality o f the provision o f education inUkraine.26 40. Ukraine does not need to spend more on public education, but it does need to spend more efficiently to improve performance and quality. A simple estimation suggests that if Ukraine could attain the average level o f studenvteacher ratios for pre-school, primary and secondary education o f the OECD countries it could save approximately 0.7 percent o f GDP (per annum)." The savings could reach up to 1 percent o f GDP (per annum) if the merging o f facilities also took place with a reduction in expenses in utilities and in non-teaching staff. Significant savings could also be achieved in higher education through the targeting o f budget resources (for tuition and stipends) to deserving students from poorer households. Gradual adjustments towards this target would free up significant fiscal savings to be used for quality-enhancing expenditures, which are presently under-funded. It should be emphasized that, in view of Ukraine's negative demographic profile, reform is urgent. Furthermore, a delay in reforms would make gradual adjustments impossible, and more drastic reforms later would come at a higher political cost. The following are selected recommendations: Revise, reformulate (or eliminate) norms dictated by the Ministry o f Education for budget formation at the facility level.28The guidingprinciple inthe revision o f these norms should be the elimination or reformulation o f norms that impose staffing rigidities based on physical measures of schools (for example, for managerial and non-teaching positions, and for teaching positions). At the same time, the norms on teaching hours deliveredby teacher need to be brought to higher efficiency levels together with class size norms. If the government desires to replace the norms, then the right vehicle i s the establishment o f purely demand driven factors as a guideline. These 25Moreover, the current practice of imposing a unique central teacher salary scale (and allowing informal supplementation throughparentalcontributionsand school-basedrevenueinitiatives) leadsto problems suchas unfilled teaching positionsin some subjectareas andinequalityconcernsregardinginformal payments. ''Estimationsare 26See EBRD-World Bank Life in Transition Survey(LITS) 2007. basedondisaggregateditem costs andthe aggregate cost structure ofpre-school, primary, and secondary levels. There are anumber ofMoES orders andresolutionsthat needto berevised, including MoES Order N g 1/9-234dated 19.06.2001, andrelatedVR resolutionN255 1-111, amongothers. xviii factors could evolve by pricing unit costs o f each service in line with the "N' factor o f the equalizationformula, that is, a minimumstandardper unit o f service (inUAHterms). Allow cities and rayons discretion on school closures. In the meantime, while the legislative hurdlesregarding this objective are lifted, allow the mergingof schools without any restrictions while continuingto investinbuses (and programs) to transport childrento and from schools. Liftall current central ministryhurdles to rationalizingteaching andnon-teaching staffs for local governments. Allow the natural (and painless) process of attrition to help with the rationalization process without rehiringretiredpersonnel, and minimize newhiringto meet only urgent knowledge gaps. Over time and under close control of the State Treasury, allow local governments at the oblast and rayon and city levels to pay (on top of wages) bonuses or location (or subject) premiums as incentives. Strengthen the monitoring o f admissions to gymnasiums and lyceums to curb governance concerns and ensure fairness for children from low income households, and strengthen the regulation regarding extracurricular classes and activities in all schools to curtail inappropriate incentives for teachers and to increasetransparency. Improve the targeting of budgetary expenditures for tuition costs and stipends inhigher education by limitingthem to students from poor households (who also attain high scores in standardized entry tests). Accelerate the introduction o f standardized tests across the country to be used for the application stage to highereducation and to measure overall performance after secondary education. Additionally, consider experimenting on a pilot basis with a voucher system for "budget financed" students inhighereducation (that is, for those students who receive tuition funding paid directly to universities from the budget), whereby the students would take funding with them to an institution o f their choice. This might stimulate competition among universities for such students and improve the market relevance o f their teaching. Sequencing; The first priority i s the revision of the norms contained inMoES orders immediately and the elimination the norms that that constraint the efficient use o f resources. Piloting of measures to replace the norms with an appropriate "N' factor inthe equalization formula (that is, a standard unit cost for each service), if needed, could be undertaken during 2008 and 2009. Most recommendations above could begin implementation in 2008 and 2009. Measures related to targeting could be implemented by 2010. (iv) StrengtheningLocal CapitalBudgeting 42. Local governments have a large role in infrastructure investment in Ukraine. One-third o f total capital expenditures are made through local budgets and 54 percent of the resources exclusively for (fixed) capital investmentsare spent at the local level. Total local capital expenditures, as classifiedinthe Ukrainian budget, represented just above 2.6 percent of GDP in 2006. However, as with the national consolidated budget, capital expenditures are comprised of (fixed) capital investments and "transfers to enterprises and government agencies." Under the latter category, close to two-thirds are subsidies to state- owned enterprises and spending units o f the government (i.e., Ministry o f Agriculture, Ministry o f Energy) that support subsidy programs. This leaves public investments infixed capital at the local level at close to 1.25 percent o f GDP in 2006 (up from 0.89 in 2005). Creating fiscal space for national and local investments across sectors i s a necessity, but to avoid wasteful spending these efforts should be coupled with strengthening the framework for local capital budgeting. The first volume o f the Public xix Finance Review studied associated weaknesses at the national level, and these weaknesses are mirrored by the same wealmesses at the locallevel. 43. But the processesof planning, project evaluation, and project selection are weak, unstable, non-transparent, and shortsighted. Planning i s undertaken with little medium-term perspective. The two major sources o f capital investments- namely, the intergovernmental capital subventionshransfers for local investments and the development budget o f the oblasts- are determined in the annual budget process. There i s a lack o f predictability in the allocation o f capital investments to local governments, which severely hampers any planning attempts. The project evaluation process lacks any consistent and solid economic evaluation or cost-benefit analysis framework. The selection o f capital projects i s highly discretionary. In theory, capital allocations from the state budget should give preference to funding for ongoing projects over new ones; in practice, this i s not the case. Selection criteria change annually, and often the selection o f projects for capital transfers i s based on suggestions from members o f Parliament in the context o f general bargainingover the budget, rather than on any cost-benefit criteria. 44. Regional inequalities in capital spending are partly the result of systemic problems. A significant portion o f the capital investment allocations are made on the basis o f co-financing (through capital subventions), which has negative consequences for regional equity. Generally, schemes such as co-financing are positive, as they encourage ownership, local funding commitment, and project sustainability at the local level. InUkraine, however, given that local budgets are heavily biased towards recurrent spending, and that in poorer regions this bias i s more accentuated, capital subventions from the center tend to end up inthe richer regions, thereby exacerbating regional inequalities. 45. Under-execution of capital spending in consolidatedoblast budgets is not as low as at the national level, but remains a problem in local budgets. However, at the city and rayon level, under- execution i s in some cases extremely high, especially in social infrastructure and communal services projects. Under-execution o f capital projects i s likely to stem from a number o f issues, including the following: (i) the absence o f multi-year budgeting and multi-year investment programs at the local level, which hampers proper planning beyond each annual cycle; (ii) a lack o f predictability in the level o f funding inany given fiscal year; (iii) preference given to new projects over ongoing projects; (iv) delays in capital project funding during the fiscal year (for example, resources for fixed investments tend to arrive in the TSA o f local governments towards the second half o f the fiscal year); and (v) different sources o f funding for a single project and the resultingco-financing risk. Moreover, asset maintenance i s severely under-funded and badly organized. There i s no specific budget line for maintenance, which leaves funding to ad hoc decisions, largely driven by crisis management. There are no data on current assets and their physical condition at the regional level and Treasury reports on capital spending by functional categories are not available. 46. The "targeted programs" are large in number, they reduce local flexibility in project selection, and they not longer serve the purpose for which they were created more than a decade ago. The state targeted programs have multiplied in number to the point o f being almost out o f control. These programs, established just after Ukraine's independence, had the role o f financing specific social and development programs that contained a strong component o f capital investment and had the task o f securing funds for certain favored programs in the absence o f a structured budgetary process and in the midst o f myriad unorganized off-budget funds. Their objectives range widely, from social programs to law enforcement. But the actual funding i s captured by "economic" programs (mainly subsidies for the coal and agriculture sectors), while other programs are kept alive but are under-funded. N o comprehensive re-evaluation o f the economic feasibility and achievements o f the existing targeted programs have been undertaken in recent years. Many o f these programs may no longer have an economic rationale and have been structurally out o f place inthe reformed budget environment since the introduction of the 2001-02Budget Code. xx 47. This report puts forward various recommendations geared to strengthening local capital budgetingthat are consistent with recommendationsprovided for the national level (in PFR I) and couldbe implementedfor the 2009 budget. They are as follows: Harmonize the local budgeting process with regional investmentpriorities and national strategic goals. Establish, in coordination with line ministriesand agencies, multi-year investment programs for local governments (in the first and second tier), which could later compete for national investment resourcesas proposed insection (i) this Executive Summary. of Create a relatively standard framework (including a template) for project evaluation. This framework should be comprised of a strengthened cost-benefit analysis methodology and an evaluation method with clearly established criteria and a rating scheme to allow comparability among the investments examined. The framework could have different levels of detail for the evaluation of eachproject within a program, dependingon the level o f local government, the size of the investment (inUAH),and the type of capital expenditure (renovation or new project). The MinistryofEconomy (MoE) shouldtake the leadincreating this framework and intraining local authorities for its implementation (and should later enforce its application). 0 Establish a consistent selection process across sectors and regions to ensure that projects and programs are put forward for central budget fimding only if they rank high in the evaluation process. 0 Enact legislation (preferably within the context of amendments to the Budget Code) to establish the principle of granting funding preference to ongoing projects. Additionally, annual budget documents should be specific about ongoing projects and the funding they receive in a specific year attachedto the remainingtime frame and the fundingneeded for completion. 0 Reserve the co-financing scheme as a secondary mechanism o f allocation that could run parallel with the criteria-based subventions/transfers or could be tailored so that poorer regions and rayons are not at a disadvantage. 0 Start collecting data on current assets and their condition at the regional and local levels and start feeding a national system for this purpose. Local governments under their own initiative (but in coordination with the MoE) could begin building these databases at the village and rayon (and city) levels, and oblast governments could consolidate the information under the tutelage of the oblast department o f economy. 0 Improve budget reporting (and allocation) by reporting data on maintenance spending separately from capital renovations, and by establishing an improved functional reporting o f fixed capital investments that incorporates both construction(and renovations) and asset acquisition. 0 Closely monitor the implementation process for capital projects at the local level, particularly in sectors where under-execution i s more acute. Moreover, establish working groups with local authorities from line ministries such as Health, Education, Housing and Communal Services, where project implementationproblems and under-execution are more severe. 0 Re-evaluate the existing targeted programs and curtail those that no longer retain a strong economic rationale or are redundant inthe current budgetary environment. Conclusion 48. Local governments and the intergovernmental fiscal system are essential to financing and providing basic services such as health and education, and to channelingpublic investmentsacross the country. Thus, while measures at the national level are required, sectoral reforms in health and education are intrinsically related to local governments' finance reform. The intergovernmental fiscal xxi framework requires further improvements to allow local governments to receive and allocate resources more efficiently towards service delivery (and infrastructure). The recommendations provided in each specific area o f this framework are aimed at this objective but they acknowledge the progress experienced through the 2001-02 budget reform, which implies that the changes suggested here are aimed at improving the current intergovernmental fiscal system and not at changing it completely. Moreover, the measures suggested embed clear principles o f efficiency (including through appropriate incentives), flexibility, local accountability, transparency, stability, and predictability, and thus they stand valid even inthe context o f a territorial reform, ifsuch a reformwere to take place inthe future (although, it should also be noted that a major territorial reform does not appear to be urgent in Ukraine, especially if the rayon and city level are strengthened as the core local government). 49. Inthe education and health sectors, Ukraine does not need to spend more but it does needto spend more efficiently to improve performance and quality. Gradual but steady adjustments in the efficiency o f education provision (through the abolishing o f the inputnorms that drive inefficient levels o f staffing, through attrition with limited rehiring, and through the consolidation o f schools, targeted spending inhigher education, and other measures suggested inthis report) would free up significant fiscal savings to be used for quality-enhancing expenditures and investments such as the training and retraining o f teachers and the provision o f better teaching aids, new laboratories, and IT equipment, which are presently under-funded. Inthe health sector, additional financing would not enable Ukraine to improve quality and performance until the system's current shortcomings are resolved (for example, the overstaffing driven by input norms, the oversized network o f facilities, the bias towards specialization, the access problems and inequalities, and other technical and allocative inefficiencies). Inthe same way, the health insurance reforms under debate, which also focus on raising additional sources o f financing for the sector, would not solve the fundamental distortions that underlie the health system but rather would increase the tax burden on the economy and place additional restrictions on local revenue streams. Thus, at the first stage the focus o f reforms, should be on addressing the underlying problems o f the efficiency o f the system. In doing this, the government could generate the necessary fiscal savings in the health budget and re-allocate them within the sector towards quality-enhancing expenditures and investments (such as medicines, training, modem equipment, and modem facilities), which are urgently needed and are acutely under-funded. 50. By improving expenditure efficiency in these sectors, and through the specific measures and incentives suggested on the intergovernmental system, the government would also help address the regional inequalities in service delivery and infrastructure provision. This measure would help prepare the service delivery platform for the challenges arising from demographic trends and the fiscal pressures arising from needed investments across the board. Regarding the latter, while such investments are indeed a necessity, to avoid wasteful spending, these efforts should be coupled with strengthening the framework for local capital budgeting. The government needs to resolve the problems o f the poor capacity for the planning, evaluation, selection, allocation, and implementation o f capital investments at the local level through creating a framework that brings about stability (through criteria-based allocations and a multi-year strategy), incorporates cost-benefit analysis in evaluation and selection o f projects, curtails inefficient targeted programs, and links the budgeting process to regional investment and national priorities and programs. xxii Table 2: SuggestedSequencingof key Sets of Redzommendation Sets of recommendations Through Medium Responsible (Immediate Amendments to Term Agency application) the budgetcode -inparallel or (short to mediumterm) Revise and eliminate (or reformulate) the norms M o H contained inOrder 33 o f the Ministryo fHealthfor budget formation at the facility level. Reviseand eliminate norms dictatedbythe M o E S MinistryofEducationfor budget formation at the facility level (including for teaching and non- teaching positions based on school network characteristics, for teaching hour loads, and for class sizes) Start the process o f estimating and implementing M o F with appropriate per head standard costs for different line services to improve the equalization formula ministries (called factor "N" inthe legislation o fthe equalization formula). Streamline the number o f factors and adjustment M o F with coefficients inall six components o fthe line "expenditure needs" estimation, and start the ministries application o f the other recommendations provided for improving the equalization formula from the state to oblasts andrayons and cities, and from rayons to villages Stop the passing o f new social programs to be M o F implemented by local government ifthey are not fully fundedas verified bya detailed ex ante examination o f costs Start the process o f moving functions from villages MoF, and settlements up to rayons (and cities) as part of MoES, and a comprehensive policy o f local government M o H reform. And grant rayons more autonomy to allocate and re-allocate staff and other resources (including through rationalizing personnel without hurdles), andto merge and close service facilities within their jurisdictions Implement the recommendations providedgeared M o F to providing greater predictability and fairness to the local revenue stream and to reducing the administrative and compliance costs generated by local taxes -including through the new tax code. Establish inthe legislation a formal criteria-based x M o F in and stable systemo f capital investment coordination transferslsubventions. with M o E Implement recommendations geared to x M o F strengthening the local government borrowing framework. But also phase out Treasury loans. Provide adequate incentives for shifting a larger x M o H proportion o f doctors into primary care and for b integrating the family medicine model more strongly into Ukraine's health system xxiii poor households some financial protection from the costs o f illness (specific recommendation are provided) Arrange for additional financing from external X donors for HIV and TB to be undertaken through mechanisms geared to increasing the current fiscal envelope, including with sub-national support using the state treasury system Allow the natural (and painless) process o f attrition x to help with the rationalization process without rehiringretired personnel, and minimize new hiring Allow local governments at the oblast andrayon X and city levels to pay (on top o f wages) bonuses or location (or subject) premiumsas incentives. Improve the targeting o fbudgetary expenditures x for tuition costs and stipends inhigher educationby limiting themto studentsfrom poor households Consider experimenting on a pilot basis with a x M o E S voucher system for "budget financed" students in hgher education whereby the students would take funding with them to an institution o f their choice Legislate and implement geared to strengthening x M o E in local capital budgeting b coordination xxiv 1. THE NEEDFORIMPROVINGINTERGOVERNMENTAL FISCALRELATIONSAND REFORMINGSERVICE DELIVERY:SETTINGTHE SCENE A. CONTEXT 1.1 Ukraine's macroeconomicperformance in recent years has been solid, with growtharound 7.5 percent per annum on average between 2000 and 2007. This performance has been supported by three factors: (i) level o f utilization with excess capacity, which acts as a productivity reserve despite the significant obstacles to private sector investment and rising real wage costs; (ii) the positive external environment, with metal prices that are well above their historical averages, a significant appetite among international investors for emerging market risks, and a strong regional business cycle driven by the sustained expansion in both Russia and the new EU member states; (iii) the gradual progress made in structural reform, in particular the stabilization efforts in 2000 combined with increasing transparency in the corporate sector since 2005, and a development vision (despite political volatility) that includes closer integration with Europe as its key policy direction. 1.2 Despite this strong performance, Ukraine's medium and long term growth prospects are constrained, inter alia", by weaknesses in the investment climate, a large and inefficient public sector and a correspondingly high tax burden, and weaknesses in the delivery of public services. The weight o f these obstacles i s the subject o f another investigation (the World Bank is startingwork on a new Country Economic Memorandum). However, the dual constraint o f a large public sector and low quality public services are a key concern for the present study. While the first volume o f the Public Finance Review (PFR I)investigated the scope for spending reductions in government subsidies and social transfers and privileges and improvements in the efficiency of revenue mobilization, the present volume analyzes inefficiencies in public service delivery with a particular focus on local government spending (both recurrent and capital spending), health and education. 1.3 This introductory chapter examines the macro-fiscal context o fpublic finance reforms inUkraine, motivates the case for reform, analyzes the specific pressures that reforms need to address, and closes with an examinationo f spending trends at the local level. B. RECENTMACRO-FISCALDEVELOPMENTS 1.4 Followinga blip in 2005, growthhas regainedmomentum,runningpresently at around 7.5 percent per annum?' Between 1999 and 2004 Ukraine's economy grew by more than 50 percent, but growth decelerated to 2.7 percent in 2005. Contributing to the 2005 slowdown were a substantial 29There are a number o f risksassociated with external factors and macroeconomic policy that are discussed in other World Bank documents, including the Ukraine Economic Update (2007). 30Trade, manufacturing, and construction are the fastest-growing sectors in2007, expanding by 16.1 percent, 13.1 percent, and 11.9percent year on year, respectively (January-July). Growth inagriculture slowed to 5.2percent year on year. 1 deceleration ininvestment demand and negative net export growth, the latter partly driven by a temporary decrease in international metal prices (Ukraine's main exports are in the metals industry) and a real exchange rate appreciation following the one-step revaluation o f the UAH in April 2005. But from the second quarter o f 2006, growth began to pick up again, fueled by a resurging investment demand, an upswingin steel prices, and overall favorable terms o f trade despite the rise in gas imports. Moreover, consumption in 2006 as in previous years continued to be strong owing mainly to wage growth3l and a rapid expansion inconsumer credit. Table 1.1: Key Macro Indicators, 2000-06 2000 2001 2002 2003 2004 2005 2006 Real GDP (change inpercent) 5.9 9.2 5.2 9.6 12.1 2.7 7.1 CPI, e.0.p. (change inpercent) 25.8 6.1 -0.6 8.2 12.3 10.3 11.6 Current Account Balance (percent of GDP) 4.6 3.7 7.5 5.8 10.6 2.9 -1.5 Foreign Exchange Reserves (US$ billions) 1.5 3.1 4.4 6.9 9.5 19.4 22.3 Fiscal Balance (percent of GDP) -1.3 -1.6 0.5 -0.9 -4.4 -2.7 -1.3 PPGDebt (percent of GDP) 47 38.6 36.5 29.0 24.7 17.7 15.0 Memo: Nominal GDP (inbillions of US$) 31.3 38.0 42.4 50.1 64.9 86.1 106.1 GNIper capita (US$, Atlas method) 690 720 780 970 1,260 1,520 1,910 Sources: State Statistics Committee; NBU; IMF; World Bank staff calculations. 1.5 While GDP growth ratesfor 2006 andthe first halfof 2007 have beenon the back of strong metal prices, they also reveal the considerable resilience of the Ukrainian economy in the face of increasingenergy prices. Indeed, the economy was able to internalize the sharp increases in gas imports in 2006 (i.e., the price went from US$62 to US$ 95/1,000 cm) and in 2007 (when the price went from US$95 to US$130/1,000 cm). While the current account moved into deficit in 2006, Ukraine had record inflows o f FDI and private lending. Capital inflows, particularly FDI surged, reflecting improved perceptions o f the investment opportunities; in fact, they doubled the current account deficit in 2006.32 Moreover, reserves have continued to grow and are currently above four months o f imports. The economy has taken the political instability o f recent years inits stride, with two parliamentary elections inthe space o f 18 months, and three government changes since early 2005. Investors take comfort in the country's overall prudent macroeconomic management, Ukraine's continued external competitiveness in regional comparison, and its under-exploited potential as a large domestic market. On the other hand, inflation (end o f period) stayed at double digits and concern on this front are growing (inflation was mainly driven by the pass-through o f higher energy prices to higher utility tariffs for households). Moreover, since May 2007 there has been some re-pricing o f Ukrainian risk, with particular concerns about widening fiscal and external imbalances and the effect o f tightening international liquidity on the banking sector. 1.6 Ukraine's direct public debt has continuedits declineand stood at 15 percent of GDP at the end of 2006. The steady decline since 1999 (when the debt to GDP ratio was at 61 percent) also reflects the cautious approach o f the authorities to debt sustainability. As raised in the PFR I, Ukraine also has a stock of contingent liabilities amounting to close to 40 percent o f GDP. However, the largest component of these liabilities (with 29 percent o f GDP) corresponds to the implicit contingent liabilities, called "lost savings," o f the USSR's Savings Bank, which are not likely to become explicit debt all at once. To date, Ukraine's fiscal authorities have been managing this issue prudently, minimizing any impact on the budget and the economy. 31 In2005 the rise inpensionoutlaysalso helpedto fuel consumption, andthe effect continuedin2006, since the cutback in ensionexpendituresas apercent of GDP was smaller than the2004-2005 hike. p2In2006FDIinflows to Ukrainewere US$5.3 billion. 2 1.7 After the expansionary fiscal policy of 2005, the fiscal envelope was kept almost constant in 2006, the fiscal balance remained under control, and public debt continuedto decrease. The fiscal deficit was at 1.3 percent of GDP in 2006, coming down from 2.4 percent in 2005. The overall fiscal envelope remained close to 43 percent o f GDP. The government has been able to undershoot fiscal deficit targets, although this was achieved at the expense of an increase inthe effective tax burden on the economy in2005 and 2006 (see Table 1,2).33 Table 1.2: Maior Fiscal Trends, 2002-06 2002 2003 2004 2005 2006 Total revenues, as % o f GDP 35.7 35.9 35.0 39.6 41.7 Total expenditure, as % GDP 35.5 36.8 39.4 42.0 43.0 Pensionpayments 9.0 9.2 11.4 15.2 15.4 Public wage bill 7.6 7.4 7.6 8.0 Capital expenditures 4.3 6.0 4.1 4.6 Capital investments (fixed) 2.8 2.8 2.0 2.3 Expenditures inhealth, as % o f GDP 3.3 3.6 3.5 3.5 3.7 Expenditures ineducation, as % of GDP 5.4 5.6 5.3 6.1 6.3 Fiscalbalance, % of GDP 0.2 -0.9 -4.4 -2.4 -1.3 Sources: MoF; State Treasury; World Bank staffcalculations. 1.8 Public spending in Ukraine is high and continues to be dominated by consumption and current transfers. Public expenditures have grown steadily since 2001, reaching 43 percent o f GDP in 2006 (see Table 1.2). The consumption and transfers orientation of the budget continues to crowd out needed public investments and thereby reduces public spending efficiency. The hike in current spending in2005, drivenbypensionpayments andpublic wages, stabilizedin2006 butthe threat is reemergingin 2007 in the midst of the early parliamentary elections. However, the small increase in the outlays on pensions as a percent of GDP forecast for 2007 i s due to administrative transfer of spending programs from the state budget to the Pension Fundon top o f the programs this fund already had in2005.34Infact, the measures taken by the government in early 2006 (such as tighter indexation) helped to reduce outlays by at least 1percent of GDP according to Bank estimates and, together with higher revenue mobilization, helpedto close the pensionfunddeficit. Onthe other hand, the public wage billhasbeentrendingupward since 2005 and reached 8 percent o f GDP in 2006. Based on the adjustments made to the budget in the first half of 2007, it i s expected that this ratio will continue to rise to close to 8.2 percent in 2007, but it may go beyond that by the end o f the fiscal year if additional increases are implemented, as proposed by the government. 1.9 Conversely, capital spending is low and is plaguedwith subsidies. The level of overall capital spending went down in 2005 and recovered only marginally in 2006 (see Table 1.2). But real public investments in fixed capital are meager and are still below the 2004 level, and well below the levels spent by new entrants to the EU and other emerging economies over the last few years. A large portion of capital expenditures is dominated by subsidies to enterprises and government agencies that manage subsidyprograms, particularly inthe coaland agriculture sectors. 1.10 Functionally, general government spending continues to be dominatedby the social sectors. Together, pensions, social protection and social assistancerepresent roughly 20 percent of GDP, followed ~~ ~ 33Budgetperformance has also beensupportedby conservative inflation andconsequentlyrevenueincrease forecasts, taking advantage ofthe largegap betweenthe CPI andthe GDP deflator. Shouldthis gap close, as one would expect, revenue performancewould be less buoyantand spendingpressurescorrespondinglyhigher. 34The PensionFundspends ina number of specialpensionprogramsmandatedby the government. This was the case in 2007 with the transfer o fmilitary andjudges pensions(formerly paidby the State Budget) to the PensionFund. 3 by education and health, which spent 6.3 percent o f GDP and 3.7 o f GDP, respectively, in 2006. The share of spending inthese areas has grown continuously over the years. 1.11 Ukraine has a relatively high tax burden which discourages private investment. In 2006, Ukraine's tax burden stood at 36 percent (or 41.7 percent considering non-tax revenues, most of which are infact different forms o f fees on production and economic transactions) (see table 1.3). This average tax burden i s slightly higher than that o f the new EUmember states and i s well above that o f fast-growing economies elsewhere with similar per capita incomes.35 Table 1.3: Revenuesas Percentof GDP, 2004-06 2004 2005 2006 TOTAL REVENUES 35.0 39.4 41.7 Indirect Taxes 10.1 13.4 14.9 VAT 4.3 7.7 9.4 Excises on goods and services 4.3 4.3 4.2 Taxes on international trade 1.5 1.5 1.3 Direct Taxes 19.1 20.0 20.9 Personal income tax 3.8 3.9 4.3 Payroll taxes 8.4 8.8 9.3 Enterprise profit tax (CIT) 4.7 5.3 4.9 Property and land taxes 0.7 0.6 0.6 Other direct taxes 1.5 1.3 1.9 Total Tax Revenues 29.2 33.4 35.9 Non-Tax Revenues 5.8 6.0 5.7 Sources: MoF; State Treasury; World Bank staffcalculations. 1.12 Close to 60 percent of the tax structure relies on direct taxes, which tax labor and capital. The three larger direct taxes are the payroll taxes (social insurance contributions), the corporate income tax, and the personal income tax, which represent 26, 14, and 12percent of the tax structure, respectively. However, a very positive development began in 2005 with a dramatic rise in VAT collections, which continued more moderately in2006. The VAT surged from 4.3 percent o f GDP in2004 to 7.7 percent in 2005, to 9.4 percent in 2006. The VAT represents (as o f 2006) 26 percent of the tax structure, and has become the second largest source o f revenues (just after the payroll taxes). Box 1.1 summarizes key findingsandrecommendations ofPFRI. Box 1.1: MainMessages of the PublicFinanceReview (PFR) Phase I The government o fUkraine needs to create sufficient fiscal space to finance the needed productive investments and its reform program (e.g., pension reform, public sector modernization, and social sectors reform), and this must be carried out within a consistent macroeconomic framework. The fiscal analysis presentedinthe PFR (phase I) shows that there is scope for a more efficient and smaller government inUkraine. To achieve this, the government should change the consumption (and current transfers) orientation o f its budget, which is hampering fbture growth, and move towards more productive and efficient spending, with emphasis on needed investments. While Ukraine has a moderate debt to GDP ratio, it should issue new debt prudently given its large stock o f implicit and explicit contingent liabilities. This report also highlights the urgent needfor setting budget policy ina strengthened medium- term framework to allow better planning and continuity. 35 Itshouldbenotedthat the measureof the tax burdenemployedhere relates to official GDP. The shadoweconomy inUkraine i s estimatedto belargeandincludesunder-declaredrevenuesby registeredtaxpayers, and so the realtax burdenis probably significantly lower than estimatedabove. However,the largeshadow economydistortscompetitionand sustainsahighlevel of administrativecorruption, which hurts Ukraine's developmentprospectsjust as muchas the high official tax burden. Indeed, one o f the reasons the shadow economy i s so largei s preciselythe cost of doingbusinessofficially. 4 The broad goals on the expenditure side o f the budget should be to have higher productive spending on infrastructure and more efficient and targeted social spending, gradually reducing the overall size o f public spending. The fiscal space for this can be found through the following sources: (i) by the phasing-out o f the most inefficient subsidies inthe coal, agriculture, and other sectors o f the economy; (ii)through better targeting o f social assistance programs inthe budget (including privilegeprograms); and (iii) through parametric changes inthe PAYG pension system (including gradually increasing the retirement age for women and lengthening the contribution period, and tightening and means-testing the eligibility criteria for pension increases to the minimum subsistence level). Creating fiscal space for investments in Ukraine also means that capital budgeting processes such as planning, selectiodevaluation, allocation, and monitoring, needto be unproved inorder to avoid wasteful spending. On the revenue side, the broad medium-term objective should be to gradually lower the average tax burden through lower rates on direct taxes, particularly on payroll taxes (social insurance contributions). But this policy should be preceded by specific reforms geared to broaden the base and improve compliance in order to allow a manageable transition. Some measures that would help with this objective are the following: (i) reforming the VAT in the agriculture sector and fixing the VAT's poor relationship with the Simplified Tax System (STS); (ii) restructuring the STS to stop the base erosion generated through its abuse (and allowing only truly small business and entrepreneurs into this system); (iii) continuing to curtail tax exemptions; (iv) strengthening VAT (and general tax) administration through improved registration, monitoring and auditing processes; and (iv) accelerating the process o f the unificationo fthe common administrative functions o fthe four social insurance funds under the PensionFund. C. WHY LOOKAT LOCAL,GOVERNMENT FINANCESINUKRAINE? 1.13 Understanding the incentives for and obstacles to improved public service delivery is crucially linked to an analysis of local government finances. Local governments account for a large share o f spending, including most o f the spending o f social services such as health and education as well as municipal utilities. However, the ability of local governments to determine their own budgets and to take measures to improve spending efficiency i s impaired by current budget rules and a lack o f accountability and autonomy in spending decisions. This i s why this volume takes a particularly close look at local government finances. 1 1.14 Local government spending has been growing in real terms over the last three years, and currently localgovernmentsaccount for close to 43 percentof the consolidatedbudget of Ukraine.36 However, when looking at the share o f local government in general government spending, which includes pensions and social insurance payments, it becomes clear that local budgets arejust keepingpace with the growth in the overall fiscal envelope (see Table 1.4). Local governments also account for close to 60 percent of fixed public capital investment. The shares o f local government spending inkey public services are large and have grown significantly over time. Moreover, local governments implement policies and programs across all key social and infrastructuresectors. 36Ukraine's legislationdoes not havea "general governmentbudget" but rather a"consolidatedbudget" and four off-budget socialinsurancefunds. Annex 1.1 contains agraphic explanationof the budgetstructure inUkraine. 5 Table 1.4: Major Fiscal Trends inLocal Government Budgets, 2002-06 2002 2003 2004 2005 2006 Local budgets spending as percent ofGDP _- -- 11.9 12.1 14.14 Localbudget expenditures, % oftotal expendituresinthe consolidatedbudget (state +local) -- -_ 40.6 37.8 43.2 Local budget expenditures, % oftotal general governmentexpenditures (with socialinsurance andpensionfunds) 34.2 32.0 30.2 29.8 30.1 Local Budget expendituresinfixed capital investmentsas % oftotal fixed capital expenditures(net of "transfers to enterprises" 41.0 48.0 46.0 45.1 54.5 Localbudget spending on education, as % of total spendingon education 57.8 61.7 60.7 63.0 64.1 Local budget spending onhealth, as % oftotal spending onhealth 79.0 75.8 71.7 77.3 79.3 Sources: MoF; StateTreasury; World Bank staffcalculations. 1.15 Reforms andimprovementsinsectors such as health and educationare intrinsicallyrelated to localgovernments andthe intergovernmentalfiscal framework.Local governments spend some 64 percent and 80 percent o f the total budget for the education and health sectors, respectively (see Table 1.4). Moreover, these shares have been growing over the last years. Local governments also provide the bulko fthe services inthese sectors, except for highlyspecializedmedicaltreatments inthe case o fhealth, and for higher education. Improving the quality and overall performance o f these sectors i s critical to improving human capital (and productivity) and to raising the living standards in the country-both key objectives stated by successive governments over the last seven years. But since services are financed and provided by local governments, any reforms in these sectors would have to rely significantly on improved intergovernmental financing and on local governments having the resources and the administrative capacity to implement reforms effectively. 1.16 Local governments andtheir financing are importantto other sectors as well. Critical sectors such as housing and communal services, which include such services as heating, water, and sanitation, depend almost entirely on local governments. Expenditures on transport, mainly through regional and local road construction and repairs, are also significant at the local level, although they represent a moderate share o f total spending in the sector. Finally, a sizable portion o f social assistance depends heavily on local governments (see section E o f this chapter for a detailed discussion o f trends in local spending across sectors). D. WHAT ARE THEPRESSURESFORREFORM? 1.17 This section looks at the key sources o f the pressures for the reform o f local government finances and of the structure o f service delivery (the latter with a focus on health and education). The main sources o f these pressures include those arising from: (i)fiscal needs and constraints; (ii)incomplete intergovernmental fiscal reforms; (iii)inefficiencies and the under-performance o f services; (iv) demographic trends; (v) regional disparities in public spending; and (vi) concerns regarding poor governance and access to services. 6 PressuresArisingfromFiscalNeeds and Constraints 1.18 Given the level of public (and private spending) in the delivery of social services, in particular in education, the greatest challenge will be to improve spending efficiency rather than to increase the overall level of spending. The level of public spending on education has been increasing, reaching 6.3 percent of GDP in 2006 (from 5.3 in 2004) and represents close to 20 percent of the consolidated budget (and 15 percent of the general government budget). This level i s highcompared with countries with a similar per capita GDP and with other countries in the region (see Chapter 4). Public health spending as a share of GDP i s more moderate at 3.7 percent o f GDP, but out-of pocket spendingi s high(see Chapter 3). 1.19 Nevertheless, the required modernization of the public sector and improvements in the quality of public services will require additional resources. This report estimates roughly that the investment needs at the local government level may reach US$29 billion over the next 10 years (see Chapter 5). At the local level, the private sector i s likely to invest only a small portion of the total needs, perhaps in areas such as communal services (housing and district heating), and that only ifthe regulatory framework for sub-national PPPs improves significantly. Thus, most o f the financing will need to come from localbudgets. 1.20 Restructuring and modernizing service delivery for the future will also require significant resources. These resources will be needed for the training and re-training o f teachers, doctors, and other technical staff at all levels if quality i s to improve in education and health. This i s particularly important given the challenges facing Ukraine interms o f strengthening human capital to improve productivity and reducing any future risks related to health status and the aging population (the latter issue i s discussed further inthis section). Resources are also needed to modernize the operations and service provision of social sectors within the larger context of public sector reforms and to restructure staffing levels (see Chapters 3 and 4). PressuresArisingfrom IncompleteIntergovernmentalReforms 1.21 The intergovernmental fiscal system did undergo positive reforms in 2001-02, but the reform process remains incomplete. Revenue autonomy is weak in Ukraine. Purely local taxes, called "own" local government taxes in this report, are meager and there i s a lack o f incentives to for local governments to collect taxes. In theory, the intergovernmental framework, including a large portion o f shared revenues (without earmarked spending) and a block (untied) equalization transfer could provide significant expenditure autonomy at the local level. In practice this i s not the case, due to a number of institutional constraints limitingspending autonomy. These issues are examined and discussed indetail in Chapter 2. 1.22 The framework for intergovernmental fiscal relations therefore needs further refinement to deliver the intended impact on improved public service delivery. Required reforms cover all four pillars of the intergovernmental system: namely, expenditure responsibilities, revenue assignments, intergovernmentaltransfers, and sub-national borrowing. Pressures ArisingfromExpenditureInefficienciesandUnderperformingPublic Services 1.23 Ukraine makes inefficient use of the large amount of resources it devotes to the education sector. Some simple ratios reveal this situation: for example, the average studendteacher ratio i s close to 10, one of the lowest inthe region and the world; the ratios o f non-teaching staff to teaching staff are also 7 highat all levels ofeducation (see Chapter 4). Staffing levels are highand the network is oversized given changing demand profiles andthese situations have not been adjusted significantly since the Soviet era. 1.24 These inefficiencies are coupled with evidence of the poor performance of the education system and a mismatch between the needs of the labor market and the training programs provided. The education system lacks an acceptable measurement system for performance and quality to guide education policy and reform. International comparison shows that the level o f dissatisfaction o f the populationwith the educational system i s strikinglyhighinUkraine (see Figure 1.1). Figure 1.1: Regionally Ukraine Ranks Poorly in Citizen Satisfaction inEducation Delivery t Very unsatisfied IUnsatisfied b *O s 10 0 1 4 !p 3I! 3 u .- U ource: EBRD-WorldBank Life inTransition Survey (LITS) 2007. 1.25 Inefficiencies in the provision of health care are also acute. Ukraine inherited an oversized network o fhospitals and health care facilities at all levels o f government. A large portion o f the resources i s spent on staff salaries and utilities at the facility level (close to 97 percent in some cases), leaving little room for supplies, adequate equipment, and renovations and repairs o f the dilapidated infrastructure. Moreover, this does not leave resources for the appropriate training and re-training o f medical staff. The Ukrainian health care system i s hospital centered, and this consumes over 70 percent o f the public health spending. Inefficiencies, as shown by some general ratios, are clearly visible: the number o f hospital beds per 100 000 population is 716 (30 percent over the EU-10 average); and the average length o f stay (ALOS) in hospitals i s 12.3 days (one o f the highest in the region) compared with 7.5 for the EU-10in 2003. 1.26 The health care system i s underperforming. While some health indicators have been improving over the last five years (for example, infant mortality and maternal mortality), others have remained stagnant or have deteriorated, (for example, the mortality o f males aged 25-39, and the overall incidence o f tuberculosis and HIV/AIDS).On balance, most o f these indicators for Ukraine remain well below those o f the new EU member countries and in some cases even below the lower middle income country averages. Moreover, in comparison with other countries, the level o f dissatisfaction o f the population with the health system in Ukraine i s low (see Figure 1.2). Chapter 3 examines in detail the shortcomings inefficiency, performance and equity inthe health care system. 8 Figure 1.2: Regionally Ukraine Ranks Poorly in Citizen Satisfaction inHealth Delivery 30 1 EVery IUnsatisfied unsatisfied 20 10 0 Source: EBRD-World BankLife inTransitionSurvey(LITS) 2007. PressuresArising from Demographic Trends 1.27 Ukraine's demographic profile has important implications for social spending and social sectors reform.Aside from the impact o f an aging population on the pension system (addressed in PFR I) relatedpressuresonthelabormarket,demographictrendsinUkrainesignalanurgentneedfor and reforms geared to anticipate and cushion fiscal pressures on public health care and long term health care (Figure 1.3). In the education sector, the effect would be different because o f a shrinking school population but reforms would still need to be in place to avoid further increases in inefficiencies in service provision, and to grasp the potential for fiscal savings. Figure 1.3: Demographic Trends in Ukraine Compared Regionally Proportionof Population65t intotal population2000-2025 EstimatedPercentageChange in PopulationBetween 2005 and 2025 20 30 20 15 10 s -100 10 5 -20 0 -30 I ource: World Bank RegionalReport"From Redto Gray," 2007. 9 1.28 The country needsto be readyto meet the increasinghealth care requirements/costsarising from a growing elderly population. A recent regional report produced by the World Band' points out that while direct increases in public health spending related specifically to an aging population are likely to be marginal, other factors, such as the health status o f the elderly, can interact with the demographic profile and produce serious fiscal pressures. In Ukraine, with the recent trend o f the high incidence o f different illnesses in male (and to a lesser extent female) adults, concerns are growing as regards the increasing vulnerability to diseases o f this cohort. Additionally, higher public spending on long-term care seems inescapable inthe case o f Ukraine; a large part o f this public spending will shift to social protection programs for the elderly, which are also handledby local governments. This means that the government needs to tackle the inefficiencies in the health care system and prepare the system for these imminent pressures while improving the national system's performance. 1.29 The school age- populationhasbeenshrinkingover recentyears andwill continueto shrink dramatically. The inefficiencies o f the education system, as briefly highlighted earlier, can only become worse as the number o f students inall age groups (and at all levels) will continue to decrease (see Chapter 4). Thus, rationalization o f the school network and o f staffing i s urgently needed to create fiscal space for financing other requirements arising from the modernization o f the education system (urgently needed to improve Ukraine's competitiveness inthe medium and long terms). PressuresArisingfromRegionalDisparitiesin PublicSpending 1.30 Inequalitiesintotal publicspending across oblasts (regions) are small, but they mask larger differences in public capital investments. Inequalities in capital spending play against the poorer regions38inthe country, and this trend has worsened over the last four years. InFigure 1.4 a-b oblasts are ordered from left to right and indescending order in terms o f oblast-regional value added (as a proxy for regional income), showing total and capital expenditures allocation per capita. The drivers behind these trends are discussed inChapters 2 and 5. Figure 1.4 a and b: Regional Inequalities Surface When Looking at Capital Expenditures (left to right). Sources: MoF; State StatisticsCommitteeof Ukraine; World Bankstaffcalculations 37 World BankRegionalReport"From Redto Gray" (2007). 38This report uses value added as a proxy for oblast (regional) income. 10 1.31 There are visible inequalities across oblasts and rayons in public health spending. Inboth cases the inequalities work against poorer local governments. While inequalities seem moderate across oblasts, they are more marked across rayons (see Figure 1.5 where Oblasts and rayons are ordered from left to right in descending order by income). Moreover, these spending inequalities have not improved over the last five years. Our analysis does not reveal large inequalities in public per capita spending between more rural or less rural oblasts. The issues and problems driving these trends are discussed in Chapters 2 and 3. Figure 1.5: There I s Some Inequality inPublic Health Spending at the Local Level 400-1 By rayons and Cities 0 .-3. - -9 V Jote: Oblasts, ARC and cities o f special status are ordcIfromhigherto lower percapitaregional(oblast) value added (left to right). Note: Rayons and cities are ordered from higher income to lower income (from left to right). This study used PIT collections per- capita as proxy for income level by rayon. Source; MoF; State Statistics Committee o f Ukraine; World Bank Staff calculations. 1.32 Inequalities in public education spending across oblasts are smaller but mask excessive pressures on current spending for poorer localgovernments. Per pupil allocations across oblasts do not show a significant or marked trend o f lower allocations to poorer regions. These trends, however, mask problems such as the high pressure on recurrent spending on education plaed on all local governments but, more acutely, on poorer oblasts and rayons, which in turn translates into less capital spending (as discussed earlier) and less spending on quality-enhancing inputs. These issues are discussed indetailinChapters2 and4. PressuresArisingfrom Poor GovernanceandAccess Concerns 1.33 There is a large amount of out-of-pocket payment (OOP) in public facilities, both formal and informal. In the case o f health this issue is particularly acute as rough estimates o f formal and informal OOP suggest that they may be as highas total public expenditures inthe sector (this i s examined in detail in Chapter 3).39 This occurs despite the constitutional guarantee of free basic health care in public facilities. In the case o f education the problem i s less acute but i s nevertheless pervasive (this is examined in Chapter 4). High OOP in public facilities implies constraints on access, which i s likely to affect the poor disproportionately. ''These informal expenditures include payments to professional health staff, self-provision o f supplies and medicines and in some cases equipment, and the excess opportunity cost o f under-utilized real estate and other assets. 11 E. GENERALEXPENDITURETRENDSAT THELOCALLEVEL 1.34 Local budget spending has grown over the last three years but the bulk of the outlays continueto be at the oblast level.The share o f different tiers o f local government intotal local spending has remained stable since 2002. Nevertheless the shares o f rayons (and city) level and village level budgets grew marginally in2006 (see Figure 1.6). Figure 1.6: Evolution of Local Budgets Spending - - 1 l6L o c a l BudgetsasPercentofGDP ] 2oo4-061 Shares of Spending Among Tiers of Local Governments 14 1W% go% 12 selllemenlr and 80% -2 villages 10 S 70% W ?" P ORayon f 8 60% 3 E2 6 (Y 3c? 50% IOblast 40% 4 2 P 30% 2 20% 10% 0 0% ... 2004 2005 2006 2002 2004 2005 2006 ources: MoF; State Treasury; WE3 staff calculations. 1.35 Local governments cover half of the total public wage bill. This means that wages (as a spending item) are significantly more relevant at the local level than at the national level, and, consequently, local governments are more sensitive to wage increases. In fact, with the minimumwage increases o f 2005, local budgets' wage outlays increased from 3.3 percent o f GDP in2004 to close to 3.9 percent in 2005, placing added pressures on local finances (see Figure 1.7). Close to 60 percent o f local wages are covered at the oblast levels and 32 percent at the rayon and city level. While the share in total wages o f townships, settlements, and villages i s less than 9 percent, wages at this level represent a large share o f their budgets (see below inthis chapter). Figure 1.7: Increasing Wages at the Local Level Public Wages by Level d Local Governments2002.2005 ZO2.X8 PublicW ~ Expndkunrty Loca Gowmmmtr M 100% 90% 80% -- u) %, 70% z ;1.5 60% -I -J 0 2 50% - IOblasl c .- 0 2 1 2 40% c n. $ 30% 0.5 20% 10% 0 G% 2002 2003 2004 2005 2002 2003 2004 2005 Sources: MoF; State Treasury; World Bank staff calculations. 12 1.36 Oblasts are the dominant player in capital expenditures at the local level but a small increase in rayon and village level investments is observed over time. Oblasts spend close to 73 percent o f the total resources allocated for fixed capital investments at the local level, althoughtheir share has been marginally diminishing(see Figure 1A). These shares contrast with comparatively larger shares o f rayons in current spending, and highlight the stark dominance o f regional government in capital investments at the local level. Chapter 5 examines the most relevant issues concerning local capital spending. Figure 1.8: Capital Expenditure Shares Among Levels of Local Government 100% Townships, SO% setllements. and villages 80% 70% 0Rayon 00% 50% I Oblast 40% 30% 20% 10% 0% 2002 2003 2004 2005 3urces: MoF; State Treasury; World Bank staff calculations. 1.37 From the functional spending perspective, health and education have increased their proportion of local spending. Figure 1.9 shows the growing dominance o f these sectors, while social protection and social assistance show a steady decline. The item "economic activities" declined in its proportion o f the budget in 2005. This item comprises spending on transport (mainly road construction and street repairs, representing close to one-third o f the spending under "economic activities"), spending on agricultural programs, and other spending on programs related to the productive sectors.40Spending on housing and communal services has increased marginally over the years but not at the pace expected considering the large need for investment inthis area (see below inthis section). 40Economic activities also include trade and labor programs, agriculture and extension programs, construction, and support to energy sector and other sectors. I Figure 1.9: Social Sectors Dominate Spending at the Local Level Functional Distribution of Total Local Budgrt 35 0Educaticn 30 fl `aHealthprctection E :: 25 I protectionand Social social asetetance H nEoonomicActivities ' 0 Dservioee and Communal Houslng 5 O A l l Other 0 2001 2002 2003 2004 2005 Sources:MoF; State Treasury; World Bankstaffcalculations. 1.38 The total local budget picture masks differences in the relative importance of spending allocations(functional) and trends at different levelsof localgovernment, as describedbelow: At the oblast level, spending is more spread out across several functional items with no categorygoing beyonda quarter of total spending. Health and education expenditures are at the top, with close to 22 percent each (of oblast total spending). Social protection, economic activities, and housing and communal services, are at 15, 9 and 7 percent of total spending, respectively. It should be noted that although housing and communal services are managed at the rayon and city levels, oblasts spend much more in absolute terms and as a percent of their budgets on this item.41Regarding trends, small increases are observed inthe relative shares of spending on education and housing and communal services, together coupled with a decline in the share o f social protection and social assistance in oblast budgets. At the rayon and city level, education is by far the largest item, taking up 41 percent of the budget. Social protection and social assistance and health care are next, with 28 and 19 percent of the budget, respectively. A small increase in social protection can be observed (above its historical average o f 25 percent of rayon budgets) which may be explained by the passing down of some programs from the oblast level to the rayon and city level. Housing and communal services represents a surprisingly small and marginally declining portion o f the rayon budget (at close to 2 percent o f the rayon budget in 2005). Housing and communal services are managed at the rayon and city level, and they would be expected to have a larger share devoted to this functional budget item, particularly because communal services provide district heating and water and sewerage, which are under serious stress owing to under- investment and infrastructure deterioration. At the level of townships, settlements, and villages, spending on education and administration are the largest items. Even though the budgets o f the third tier of local governments are comparatively small, they are burdened significantly with overhead costs. Health care i s third in importance. A small decline can be observed in education spending. This is mainly due to the upward delegation o f some schools from villages to rayons (this issue i s discussed in Chapter 2). Within the category "other" there i s an increase in the relative share of spending on economic activities. 4'This i s due to the active financing from oblasts in this areaat the rayonandcity level. 14 F. CONCLUSION 1.39 Local governments and the intergovernmental fiscal system are essential in financing and providing basic services such as health and education, and in channeling public investments throughout the country. While measures at the national level are required, sectoral reforms in health and education are intrinsically related to local government reform since local governments spend the larger share o f public resources and provide the bulk o f the services in these sectors. The government faces significant challenges inthis agenda, including increasing fiscal needs for fixed capital investments at the local level and strengthening (and completing) intergovernmental fiscal reforms to improve resource allocation. Moreover, given the level o f public (and private spending) inthe delivery o f social services such as health and education, a key challenge will be to improve spending efficiency rather than increase the overall level o f spending. In parallel, sectoral reforms (combined with reforms in the local finances) need to geared to overcoming the under-performance o f the public health and education systems, to prepare these sectors for the impact o f demographic trends inthe country and to resolve inequalities inpublic spending and in access to services. The following chapters deal in detail with these issues and provide specific recommendations for overcoming the challenges in selected areas o f intergovernmental fiscal relations, public health spending, public education spending, and local capitalbudgeting. 15 16 2. IMPROVINGINTERGOVERNMENTAL FISCAL RELATIONS INUKRAINE A. INTRODUCTION 2.1 Ukraine is a unitary state with, broadly, three tiers of local government, namely (1) oblasts, (ii) andcities,and(iii) andsettlements.Localbudgetsarefunnelingcloseto40percentof rayons villages the consolidated budget, and in some categories of spending such as wages and fixed investments, more than half of the consolidated budget (see Chapter 1). However, this i s not an indication o f the true level of fiscal decentralization or local spending autonomy inUkraine, as this chapter demonstrates. 2.2 This chapter focuses on the fiscal dimension of local governments inUkraine. The objectives of the chapter are: (i)to analyze in detail the current intergovernmental fiscal framework with respect to both equity and efficiency; (ii) to provide options for the reform o f this framework, and (iii) provide to recommendations for improving the efficiency and performance of the functions undertaken at the local level (from the fiscal and administrative perspectives). The political and administrative dimensions of increasing local government autonomy and accountability issues are not elaborated in detail in this chapter, but are taken into account inthe development o fpolicy recommendations. 2.3 Ukraine's fiscal framework and its intergovernmental fiscal (IGF) relations improved significantly with the implementationof the Budget Code in 2001-02. Butthe reformdidnot extend to all aspects of IGFrelations. Moreover, there has been some backsliding over the last four years. Thus, the reform remains incomplete, which hampers the ability o f local governments to finance and deliver services efficiently. This chapter i s organized as follows: Section B provides a brief background to the evolution of IGF relations in Ukraine and to the current framework. Section C assesses problems associated with expenditure responsibilities, and puts forth options and recommendations for reform. Similarly, Sections D, E, and F assess key problematic issues and provide recommendations for improving revenue assignments, intergovernmental transfers, and sub-national borrowing, respectively. Section G covers other cross-cutting issues inthe fiscal framework. B. BACKGROUND 2.4 Ukraine is a unitary state with three tiers of local government. The top tier has 24 oblasts (regions), the Autonomous Republic of Crimea, and two cities with special status: Kyiv and Sevastopol. The second tier is made up of488 rayons (or districts) and approximately 177 cities (Annex 2.1 provides a detailed graphic view of Ukraine's local government structure). The third tier contains about 12,000 small towns, settlements and villages with a local council or "Rada." Despite reforms in local self- governance (more active on the fiscal front), Ukraine remains highly centralized politically and administratively. Box 2.1 presents an overview o f the political and administrative relations between the central government and the three tiers of local government. 17 1Box 2.1: Intergovernmental Administrative and PoliticalRelationsInUkraine 2.5 In 2001-02, the budget reformradically changedthe country's intergovernmentalfinances. Before the enactment (in 2001) and the full implementation (in 2002) of the Budget Code, Ukraine's intergovernmental fiscal relations functioned under a system o f hierarchical or nested budgets inherited from the Soviet era (sometimes colloquially called the "matryoshka" style). The budget allocations were passed from the central government to the oblasts, which then had the discretion o f allocating these resources to the rayons and cities. The rayons and cities were in charge of allocating resources to small towns, villages and settlements. The 2001-02 budget reform changed local government expenditure responsibilities, revenue assignments, transfers, and sub-national borrowing powers. A key feature o f the reform was that local governments in the second tier (rayons and cities) were no longer subordinate to 18 oblast governments for budget and transfer allocations. Currently, around 692 local governments, including oblasts and rayons and cities (basically, the first and second tiers o f local government), separately prepare their own budgets within their jurisdiction and process them directly with the Ministry o f Finance. 2.6 The Budget Code reform helped to clarify local government expenditure responsibilities by providing guidance as to which functions should be performed by each level of local government. Moreover, the establishment o f the three levels o f sub-national government as independent budgetary units4*helped, at least in theory, to set the framework for a certain level o f expenditure autonomy for oblasts, rayons and cities, and ~illages.''~ In practice, however, Ukraine has a mixed system o f de- centralized and delegated expenditure responsibilities with a strong influence exerted by line ministries on local spending decisions. 2.7 The reform also clarified revenue sharing arrangements and "own" local taxes. All local governments were to share a fixed percentage o f revenues from the personal income tax (PIT), the land tax, and the single tax (under the simplified tax system) collected in their jurisdi~tions.~~Currently, oblasts receive 25 percent o f the PIT and land taxes. Rayons and cities receive 50 percent o f the PIT and 15 percent o f local land taxes:' Towns, villages, and settlements receive 25 percent o f the PIT and 60 percent o f local land taxes. Revenue sharing arrangements between the central government and all local governments are also in place for the single tax. In addition, the Budget Code established a list o f local taxes (hereinafter called "own" local taxes) from which a local government can pick the specific taxes that it wants to levy on its jurisdiction. The rates for these local taxes are fixed by the Budget Code and the related regulations, and local governments cannot change them, although some have bands within which local government can set a rate. The collection o f all taxes, including own local taxes, i s carried out entirely through the State Tax Administration (STA), as local governments do not have independent collection agencies. Similarly, all payments by a local government are made against its Treasury Single Account (TSA) (see Box 2.2). Box 2.2: Ukraine Comprehensive Treasury Single Account for Local Governments Reforms inthe Treasury system o f Ukraine have been significant over the last 6 years. Intandem with the budgetary reforms in 2001-02 a new integrated Treasury system has improved control and reporting across government budgetary units and agencies. Over time, the implementation o f the TSA has expanded to cover all transactions o f local governments. All revenue receipts for a local government are sent from the centralized tax administration and accumulated in a TSA. Similarly, all payments made by a local government have to be made against the same Treasury account, and based on a pre-assignedbudgetary code and amount o f resources approved under the annual budget for the specific locality. The existing tight controls help to avoid the type o f leakages o f resources suffered intransfers o f resources to local government in countries where a nationwide centralized treasury system is not in place. While there are a number o f areas in the Treasury system that need improvement at the national and the local levels (for example in the flexibility o f reporting local spending in different ways and its integration to other government agencies), the Treasury systemrepresents a major advance inpublic financial management inUkraine. 42Although the third tier still receives transfers allocated by rayons and cities, and there is some supervision applied by rayons and cities to the budget process o f local governments inthe third tier. 43This report refers to expenditure autonomy as the level o f autonomy indecision making at the local level -in the allocation o f resources within and across functions to address local needs. 44Crimea, Kyiv and Sevastopol have additional earmarked revenues. 45Cities inthe Autonomous Republic o f Crimea receive 75 percent o f the personal income tax and all o f the landtax. 19 2.8 The transfer systemis currently dominatedby a formula-based equalization block transfer. This is a significant improvement over the former ad hoc transfer arrangement for local governments. Some features of the equalizationtransfer are noteworthy. First, the equalization formula has a horizontal gap-filling approach to equalization: that is, in each o f the 692 local governments (the first and second tiers of local government separately) the amount o f the equalization grant i s calculated as the difference between estimated expenditure needs and forecasted revenue capacity. Thus, if a local government has greater expenditure needs46than its forecasted fiscal (or revenue) capacity4' can cover, it will receive an equalization transfer: otherwise, the local government has to pay into a common distribution pool o f resources.48Second, the Ukrainian equalization system also has a strong and growing vertical funding component, since a substantial portion of the pool is contributed by the central g~vernment.~'Third, the factors that determine expenditure needs (though not all of them) are derived from demographic or demand-based factors in each jurisdiction for a number of functions delegated to local governments. However, they are also adjusted for a variety o f characteristics o f the local government and its network of facilities. Fourth, the forecasted revenue capacity was originally constructed based on a moving average of past collections, although this estimation has been changing over the last two years. And fifth, as mentioned earlier, the grant i s an unconditional (or block) grant, and thus, in theory, local governments have full spending discretion. In addition, there are other (separate) special grants or subventions for financing social assistanceSoand subventions/transfers for financing capital spending. 2.9 The Budget Code reform also provided a structure for local government borrowing (although not an ideal one). Prior to the reform, borrowings were undertaken by local governments (mainly large cities) with little regulation and controls. The Odessa city default in 1999 was in part a result of this lack o f oversight and contributedto a sharp decline inlocal borrowing, which, however, has recently reversed slightly. C. IMPROVINGEXPENDITURERESPONSIBILITIES,EXPENDITURE AUTONOMY, ANDSECTORALADMINISTRATIVE REFORMAT THE LOCAL LEVEL Assignment of ExpenditureResponsibilities 2.10 Some overlapping of expenditure responsibilitiesin service provision remains, creating an inefficient use of resources and hampering planning capabilities. The ownership of assets such as schools and health facilities inherited from the former system did not change under the new system. 46 Expenditure need represents the potential demandfor spending that a local government faces to provide a particular standard level of service(s) to its population. It refers to the level of spending in a sub-unit,given its assigned responsibilities and its ?,articular characteristics (demographic, geographic, etc.). Fiscal capacity is the potential o f a sub-unitto mobilize revenues. It represents its potential to raise any tax or non-tax revenue, given its assigned tax bases and its level o f income. Since wealthier regions have the potential to raise more revenues, fiscal capacity is closely linked to the level o f income o f each locality. A commonly used proxy to account for fiscal capacity i s regional per capita income. 48 This system has the funding approach known as "fraternal" or "Robin Hood-" InUkraine the most basic shape o f the formula for the allocation to a local government is the following: = a(5-oi);where Tiis the transfer to local government i; Vi i s the estimated expenditure needs for local government i and Di its estimated revenue capacity. The formulation for the amount received or paid also depends on an equalization coefficient a and a variety o f adjustment coefficients that the central government regulates. 49 This share has been increasing over time since 2002. The local government role and local government funding in social assistance are not covered in this report. A detailed discussion o f this issue can be found inthe DFID report "Facilitating Reform in Social Services" (FRSSU) 2007. 20 Consequently, an overlapping of functions has remained to some extent. For example, there are education facilities that provide the function of general secondary education at the village level while this h c t i o n is also provided at the rayon level. In addition, overlapping i s found in some education institutions with special status such as lyceums, boarding schools, and technical training institutions that operate at the rayon level and insome cases at the oblast level (see Table 2.1). Moreover, the systemo f financing local governments through transfers and the budget formulation at the local level perpetuate this problem, as these arrangements link funding to the facilities owned by each local government (see Section E on transfers). ExpenditureFunctions Levelof Local Budget Oblasts, the ARC Rayons and Cities Townships, and Cities of Settlements,and SpecialStatus Villages Local Administration X X X Pre-school X X Primary and Secondary X X Technical X X Higher Primary Care Functions X X Secondary X X Tertiary X Social Protection X X Transport X X X Housing and Communal X X X Services Physical Education and Sports X X Culture and Arts X X X 2.11 Unfundedmandates have growninthe system since 2002. Despite the goals envisaged by the budget reform, a growing amount o f unfunded mandates i s burdening local finances. Mandates in the areas of social protection and social assistance are particularly large and are not fully f~nded.~' The central government continually approves new entitlements and programs in these areas, and although specific funding i s allocated for them, in most cases it does not match the responsibilities that these mandates carry. This situation i s likely to worsen over the mediumterm inview o f the fiscal pressuresto be expected on long-term care for the elderly. 2.12 The averagesize oflocal governmentsinthe third (lowest) tier seemstoo smallfor the efficient management of some of the functions. Outside of the first and second tiers, local governments with councils (Radas) comprise over 12,000 smalYrura1towns, villages, and settlements of varylng sizes. The average population size inthe thirdtier (withcouncilrepresentation) is around 1,400 (see Table 2.2 ). While international experiencevaries significantly interms ofthe adequate size of a local government, for a country with the level of income ofUkraine, andwithUkraine's network of facilities, its geographic characteristics, 5'See DFIDreport "Facilitating Reform in Social Services" (FRSSU) 2007. 21 and its institutional features, there are drawbacks to maintaining certain functions such as health and education at the lowest tier o f local government^.^^ Table 2.2: Average Size of LocalGovernments inUkraine Level of Local Government Average Population Size (approximate) 1" tier: Oblasts, ARC and cities of specialstatus 1,730,000 2" tier: Rayonsand cities 61,000 Ydtier: small rural towns, settlements, villages 1,400 2.13 Potential efficiency gains from economies of scale in service provision are lost below the rayon level. The network o f service facilities in the third tier i s large and in many cases overstaffed, but these facilities are less utilized compared to rayon level facilities. Moreover, maternity facilities and other health and education facilities below the rayon level, aside from being underutilized, lack the proper equipment and resources to perform their functions. At the same time there are health and education facilities, also fully staffed, that offer the same type (level) o f services in close proximity. Inmany cases localities (inthe thirdtier) are so small that they lack the ability or the resources to manage such services as health and education. Infact, some villages inUkraine have been devolving their functions o f managing pre- schools andsecondary schools, as well as some health care facilities, to rayons, arguing that they cannot cope with the costs. Given better service standards, people are more likely to trust and attend rayon center health care facilities, which contributes to underutilization at the village level. T h i s situation i s coupled with the large fixed recurrent costs o f sustaining the administrative overheads o f the councils o f villages and settlements, which leads to underinvestment problems, including a drastic deterioration o f facilities and equipment. This i s more o f an issue o f assigning the functions to the "right" level o f local government and less o f one o f decentralization versus centralization. 2.14 During2005 (and early 2006) proposals were under discussion for creating "communities" below the rayon level as a third tier of local government under a new territorial structure, but these proposals did not seem adequate in several critical aspects. On the positive side, these proposals argued for consolidation below the rayon level, which i s necessary. On the negative side, however, these proposals - argued for providing significant expenditure responsibilities, similar to those o f the rayons, to these newly created "communities" with populations o f around 5,000 people. This proposal has a number o fpotential drawbacks, as listed below: The costs of administration would increase: the overhead (fixed costs) o f keeping administrative offices and councils in around 9,000 "communities" (as envisaged by the proposals), all with significant expenditure functions, would be very high.Moreover, the cost o f an entirely new administrative level would be an extra burden for the existing rayon administration. The problems o f inefficiency and the loss o f economies o f scale would worsen, particularly if these territorial units (communities) were assigned functions such as secondary education, health, etc., as envisaged in the proposals. Even under the current system, Ukraine has already over-decentralized some functions (such as health and education) to villages and settlements that do not have the resources or the administrative capacity to efficiently plan and manage these functions. 52A similarargumentis made for the case of social assistance functions by the DFIDreport"Facilitating Reformin Social Services" (FRSSU)2007 22 Recurrent wage expenditures would inevitably increase sharply. Local governments already spend half o f the general government wage bill (and this i s growing, as discussed in Chapter 1); and adding a tier oflocal government would further increasethese costs. 0 Problems with administrative capacity below the rayon level would persist. It would be administratively costly to have a full set o f transfers and revenue assignments from the central government to these "communities," as was suggested. ExpenditureAutonomy, AdministrativeAutonomy, and UnreformedSectoralPolicy (with a focus on healthand education) 2.15 This sub-section raises issues related to the incomplete sectoral reforms that significantly affect local expenditure and budgeting autonomy, creating acute inefficiencies and consequently undermining the provision o f services. While this section focuses only on health and education, reforms need to be reinvigorated in all other sectors inwhich local governments have functions. 2.16 The norms contained in the orders that govern the functioning of education and health serviceswere not changed duringthe budget reformdespitetheir obvious budgetimplications. The ministries o f these sectors have established orders containing mandatory guidelines for staffing and facility services, also called "norms," which have their roots in the Soviet system. They are geared to maintain quality standards in service delivery by dictating to service facilities how to allocate staffing and other resources based on input coefficients and on the existing oversized network o f schools and health facilities. Box 2.3 contains some examples extracted directly from Order No. 1/9-234 and VR Resolution N2551-I11inthe case o f education and Order No. 33 inthe case o f health, although there are a number o f other orders and regulatory acts that also impose spending rigidities at the facility level on local governments (these are also discussed inChapters 3 and 4). Box 2.3: Examples of Norms for Health and Education Facilities that Dictate How These Facilities and Local Governments Form Their Budgets 1Therapist per 25 beds inDistrict hospitals 1Obstetrician-gynecologistper 20 beds inRayonhospitals; 1post per 15beds inDistrict hospitals- 1Infectious disease doctor per 25 beds inoutpatient aid inRayonhospitals- 1 Surgeonper 20 beds (adults) and 15 beds (children) inRayonhospitals 1 Dentist per 20 beds inchildren's hospitals 1Nurse (gynecology) per 25 beds inRayonhospitals- 1Nurse per 20 beds inchildren's hospitals 1Nurse (infectious diseases) per 13beds inRayonhospitals 1 Dietarian nurse per 500 portions served a day 1Nurse assistant cleaner post per 8 beds (depending on department) inchildren's hospitals 0.5 Statisticians per 20 posts o f doctors inpolyclinics 1Cook per 30beds ina healthfacility 1Deputyschool director for 11class groups 1Cleaner per 500 square meters (0.5 per each 250 square meters) 1Coat roomattendant (inschools) per 200 coat spots- Source: Extracted from OrderNo. 33 Ministryo f Health and Order No. 1/9-234 Ministryo f Education 2.17 Becauseof the mandatory nature of these regulations(and there are consequencesfor non- c~mpliance)~' service facilities in these two sectors form their budgets accordingly and submit all ''For example, there was a case o f a school director who was brought to trial because he hired two technical staff instead o f the- four mandated by the "normative" and used the savings to improve the infrastructure of his school. (This was extracted from the minutes o f the VR hearings with the Ministryof Education on January 7, 2007.) 23 them to the correspondinglocal government, which creates a large degree of inflexibility in budget allocations (see Box 2.4). The local government in turn pulls together the budgets o f all facilities, aggregating the problem across thejurisdiction. This problem i s particularly acute given the existing large and underutilizednetwork of service facilities. Inthe end, this severely constrains local budget autonomy and flexibility. Box 2.4: The Inefficient Processof Budgeting Based on Input Norms Ministry of Finance Budget Allocation $$$ Through: Input Normsto Input Normsto (1) SharedRevenues (PIT, Land) form school formhealth (2) EqualizationTransfer, (based mostly budgets, which facility on demographicheed constrainbudget budgets, which factors (although problemsexist-) flexibility (Order constrain (3) LocalTaxes NO119-234;VR budget resolutionNo flexibility 2551-111) (Order No 33) 17 Submission complyingwith Norms(Healthand Education) I I ~ An Average RayonBudget: "Norms" fulfillment large + networkof facilities =non- flexible crowded by recurrent spendinglocalbudgets;little spendingautonomyis left to LG. Only marginalneededinvestments are possible. 2.18 The systemic mismatch generated by the juxtaposition of a formula-based intergovernmental transfer system and an input norm system of budgeting negatively affects the financing of services at the local level, creating problems of under-funding. Local governments not only have to comply with the large list of expenditure mandates (some o f them under-funded), but they must also fund these services according to supply (input) norms without having the flexibility to meet 24 demand efficiently. The latter generates the overspending o f current expenditures, particularly wages, heating, and the like (see Figure 2.1) and leaves very little for capital expenditures. While Article 34 of the Budget Code allows spending units (that is, line ministries or sectors) to have an influence on the budget preparation, the way inwhich this process is plagued by health and education noms goes against the precept of the "effective and efficient" use of resources stipulated in Article 36 of the Budget Code and of any notion of efficiency. Figure 2.1: Current Expenditure as a Share of Total Expenditure in Local Governments (oblast consolidated), 2002 and 2005 I .-0 2002 .. --C- 2005 ote: Oblasts, ARC and cities o f special status are ordered le!? to right from higher to lower per capita regional value added as a proxy for income. Sources: MoF; State Statistics Committee o f Ukraine-; Bank staff calculations. 2.19 This input based budget formation process also generates acute inefficiencies in service delivery. Thus, Ukraine has an average studentheacher ratio close to 10, one o f the lowest levels in Eastern Europe, which i s exacerbated by the prospect of a declining school age population over the coming years. Ukraine's non-teaching to teaching staff ratio i s close to 2 for general education schools, one o f the highest in the world. In the health sector the country has 716 hospital beds per 100,000 population(this i s 30 percent over the EU-10 average), butthis is coupledwith underutilization. 54 2.20 The situation i s worsened by legal provisions prohibiting the closure of health care facilities and schools, and the difficulties local governments face in firing personnel. Article 49 o f the Constitution prohibits any closure of heath care facilities at any level of government, and other Verhovna Rada resolutions prohibit closures of schools. Consequently, local governments are bound to maintain a large network o f service facilities despite a shifting structure o f demand and low levels o f utilization in both education and health. Evidence collected for this study during field visits, interviews, and round table discussions with local officials suggests that local officials are aware o f these issues and would like to proceed with the closure and merging of service facilities in order to allow a more optimal use of resources. '`Chapters 3 and 4 address health financing and education issues indetail. 25 2.21 In theory, local governments can hire and fire personnel, but in practice the process of firing is extremely difficult. First, this process is lengthy; the timetable to be followed could span close to a calendar year. Moreover, the process itself i s cumbersome with clearances fiom line departments in higher levels of government and many other time-consuming steps to be followed. Inthe end, the process i s highly impractical for local governments, and consequently staffing changes are very rare. 2.22 Partly because of these unfinished reforms, local governments have little to spend on investments, maintenance, and quality-enhancing inputs. Thus, service facilities in health and education, and in general infrastructure across sectors in local government, are deteriorating. Moreover, there are not enough resources for the needednew health equipment and teaching aids, and there are no resources for the training and re-training of critical personnel. Under these circumstances, quality can hardly be enhanced. 2.23 This overall situation also has equity implications.Given the large expenditure mandates, the mandatory norms that generate overstaffing, and the oversized network of facilities, poorer local governments tend to spendmore on current expenditures and less on investments that could enhance the quality of service provision for the poor. Figure 2.2 shows that lower income oblasts tend to spend more on wages as a portion of total spending compared to wealthier oblasts (Figure 2.1 shows the same problem from the total recurrent spending perspective). This automatically constrains lower income local governments to spend in other areas. The situation i s even more acute at the rayon level, where wages represent ahigherproportion of total spending. Inthe case o f health and education, this situation (together with other circumstances, as discussedinChapters 3 and4) also pushes service facilities to ask for out-of- pocket formal and informal payments from the population served. Figure2.2: PublicWages Expenditureas a Percentof TotalExpenditure(oblast consolidated),2002 and 2005 --)- 2005 ote: Oblasts, ARC and cities o f special status are ordered from higher to lower per capita regional (oblast) value added (left to right). Sources: MoF; State Statistics Committee of Ukraine; Bank staffcalculations. 26 Recommendations 2.24 The following i s a set o f selected options for reforms and recommendations gearedto improving the structure and efficiency of expenditure responsibilitiesfor local governments: Strengthen the rayon as the core local government. At present, the rayon has the right size for a core local government in the context of Ukraine. The administrative capacity o f rayons and cities has increased significantly since the reform and should be further strengthened with adequate administrative tools and more decision-making autonomy (this i s further discussed in this section). Start a comprehensive process of moving functions up from villages and settlements to rayons. The functions that shouldbe moved are mainly inhealth and education, including general secondary education, pre-schools, and local clinics and obstetric centers. This does not necessarily imply the closure of facilities but rather the passing of control to the rayons so that they can plan and manage service provision more efficiently. In fact, this has already ocurred sporadically in some rayons in which villages gave up their responsibilities, but it should be pursuedthrough incentives as a comprehensive policy by the government. The legal framework for this process already exists in Article 92 (and in Articles 89 and 93) o f the Budget Code, and only minor amendments would be needed. A suggested sequencing would be to amend the Budget Code in2007-8 and to have full implementation in2009. Arrange for small town, village, and settlement councils to keep some local functions and their current political representation. Functions such as firefighting, street cleaning, sports and cultural activities, routine street maintenance, and other core maintenance functions o f a locality seem appropriate for these local governments. At the same time, it i s important to maintain local councils (although perhaps with a smaller number of representatives) and to keep their current political representation inthe rayon councils. Promote the consolidation of the third tier of local government (rural towns, villages, and settlements). This effort should include a legal framework and incentives for pooling joint functions and resources between two or more local governments at this level. Regarding the latter, local governments should be allowed to keep the savings generated by joint undertakings without beingpunishedwith a budget cut inthe next fiscal year. Avoid creating (through administrative reform) a new local government below the rayon level with significant expenditure responsibilities. The upward consolidation o f the fiscal and service delivery functions o f local governments should be pursued through strengthening the rayon as the core local government. Beyond that, joint undertakings and other consolidation mechanisms below the rayon level could be pursuedbut without granting additional functions. 2.25 The following is a set of selected recommendations geared to granting greater expenditure and administrative autonomy to local governments while at the same time suggesting changes in key un- reformed issues that negatively affect expenditure efficiency and local finances: Grant rayons more autonomy to allocate and re-allocate staff and other resources, and to merge and close service facilities within their jurisdictions without the influence of sectoral ministries. Thus, aside from revising the norms that dictate staffing (discussed later), consider revising Article 49 of the Constitution (on health) and VR resolutions (on education) that prohibit facilities' closures. This would allow the optimization of the quantity o f health and education 27 facilities and a better resource allocation. This step i s critical to breakmg the cycle o f inefficient spending and providing the fiscal space for needed investment in equipment, upgraded facilities, and the training and re-training o f personnel in both sectors. Moreover, these reforms (together with other policies ineach sector (as discussed inChapters 3 and 4) would free up resources that couldbe re-allocated within the same sector to enhancing the quality o f education and healthcare. 0 Place funding and strategic administrative decision-making regarding service facilities in the hands of the rayons and cities. This should be carried out at least over the medium term, while the reduction o f the size o f the network ("network optimization") takes place. This i s critical inthe context o f Ukraine: if the reform i s aimed at optimizing (and reducing) the network o f facilities, as it should be, key expenditure decisions should be taken by the rayons and cities (or oblasts, accordingly). Providing too much power at the facility level, for now, would create vested interests and become an obstacle to the needed optimization o f the network. 0 But place greater (day-to-day) managerial discretion in the hands of service facilities. Facilities should be more accountable for their performance. Consequently they need to have more discretion in hiring and firing non-teaching and nonmedical personnel, in managing their inputs and routine investments, and in dealing with other operational issues. Over time, as the network i s reduced, they may acquire larger administrative responsibilities. 0 Revise all of the norms of the Ministry of Education that dictate staffing needs based on inputs and then eliminate those that impose rigidities (and inefficiency) in facilities and in local government budgets. Some examples o f these norms are Order No. 1/9-234 and VR ResolutionN2551-111. This situation acts as a bottleneck ineducation financing at the local level and creates perverse incentives inthe system. Its elimination would permit a more optimal use o f resources; the savings could be used for investments and higher salaries within the sector. The remaining normatives that do not impose any staffing rigidities based on inputs but that rather assure a basic level o f quality could be transferred to a more efficient and quality-enhancing accreditation procedure for schools and education facilities. 0 Revise Order 33 of the Ministry of Health and then eliminate all input based norms that impose budget rigidities on health facility budgets and consequently on local government budgets. This presents a bottleneck in health financing and local government budgeting and creates serious inefficiencies and perverse incentives inthe system. Its elimination would permit a more efficient use o f resources and greater flexibility, which in turn would allow higher budget allocations for investments and higher salaries withinthe health sector. This i s inline with actions proposed by the government program in relation to creating the regulatory conditions for budgetary (decision-making) decentralization. The remaining normatives that do not impose any staffing rigidities based on inputs, but that rather assure a basic level of quality based on existing demand, could be transferred to a more efficient and quality-enhancing accreditation procedure for health facilities and health providers. D. STRENGTHENINGREVENUEASSIGNMENTS IssuesinRevenueAssignments 2.26 Tax autonomy is very limited in Ukraine. The bulk o f the revenues for local governments come from transfers (such as the equalization transfer) and shared taxes (such as the PIT and the land tax). Only a small part, on average less than 3 percent o f total local revenue, comes from own local taxes 28 (see Figure 2.3).55 Giventhe small revenue mobilization of "own" local taxes, and given other constraints embedded in the system, local governments are increasingly dependent on transfers from the central government, which undermineslocal accountability. Figure 2.3: Revenue Compositionof Total LocalBudget(net of Treasury loans), 2 0 0 1 4 5 6o 50 I Taxes Shared 40 0Equalization transfer received 30 0Localtaxes 20 BOther 10 0 2001 2002 2003 2004 2005 Note: The category other includes special transfers and subventions to cover mandates o f the central government (includ ng Issues RegardingShared-Revenues 2.27 There are no fully structured ways of compensating local governments when their revenues from shared taxes drop because of tax policy changes at the central government level. An example of this was the revenue forecasting problem arising with the implementation o f the flat rate reform for the PIT in 2004.56 PIT collections dropped by 1.3 percent of GDP in 2004 from 2003; they recovered moderately in2005 and 2006, but are still below the pre-reform level as a percent o f GDP. And while the PIT reform can be viewed as an overall positive tax reform, the revenue shortfalls significantly affected local governments. The government reacted by changing certain important features of the equalization formula (i.e., the estimation o f fiscal capacity) and passed on additional ad hoc compensations to local governments. While the central government's intent was to compensate local governments, the unstructured manner employed distortedimportant components o fthe intergovernmentalfinance system. 2.28 Some tax exemptions granted by the central government in shared taxes also affect local government budgets, frequently without compensation. This occurs despite specific provisions contained in the Budget Code (Article 103) that oblige the central government to compensate local budgets for shortfalls in assigned revenues coming from central policy changes. The most common example is the exemptions granted (and currently inplace) for the landtax. Tax expenditures on the land tax are close to 0.3 percent o f GDP (as o f 2005) as estimated by PFR I, butthere is ahighprobability that these figures are on the conservative side, giventhe large price increases for landover recent years. 2.29 There are no structured ways of compensating local governments when other central government policies affect local budgets. For example, the large wage increase in 2005 and 2006 (and 55Local governments also obtain additional revenues from the sale o f property transactions, mainly land. These transactions often lack transparency and are not properly reflected inthe budget. 56In2004 Ukraineimplemented a flat rate of 13 percent for the PIT (which increasedto 15 percent in2007), coming from a - progressive schedule o f taxation. 29 potentially in 2007) significantly affected local budgets' resource allocations. Given the large portion o f wages in total local government spending, the salary increases mandated by the budgets laws in 2005 and 2006 generated large shortages. The central government reacted in an emergency mode by sending compensatory funds to oblasts instead o f directly to rayons. Thus, oblasts were left to decide how to redistribute these fbds among second tier local governments, returning to the old practice which i s prone to political bargaining at the local level. 2.30 The ad hoc way in which these compensations are carried out harms predictability and transparency and hampers the establishment of a strong medium term budget framework in the country. This situation is bound to create inefficiencies, since the predictability o f resources, a necessary condition for good planning, i s undermined. The intergovernmental system i s not fully prepared for the changes that often take place inthe country's public finances. 2.31 The way in which the shared PIT is allocated to local governments may be generating inequalities.It i s obvious that poorer local governments collect a smaller percapita PIT and consequently receive a smaller share o f this tax. But, more important, the PIT i s paid in the locality in which the individual i s employed (that is, where the employer is located) and not where the individual resides. Often, companies and businesses tend to be located in more urban and resource affluent areas. Consequently, poorer districts may be severely affected in their capacity to provide services for their residents who may actually be contributing to a different jurisdiction. The equalization transfer compensates to some extent for this but at the risko f increasing overdependence. 2.32 Local governments also obtain additional revenues from transactions related to property sales (mainly land), which often lack of transparency and are not properly reflectedin the budget. These transactions arise from sales (to thirdparties) o f land that belongs to local governments. Since these transactions are often not reported appropriately inthe budgetary process it i s difficult to assert estimates, but in some local governments where there is considerable demand for land and high land prices these resources may reach up to 4-6 percent o f annual allocations (although with large variances from year to year). Issues Regarding "Own"Local Taxes 2.33 Although there is a certain amount of flexibility (through bands in some taxes), local governmentshave no autonomy for setting rates in own local taxes levied. Moreover, many o f these taxes and fees are innominal amounts (inUAH) and have undergone little change since 2001 despite the annual double digit inflation inUkraine over recent years. 2.34 The administrative costs of collecting certain local taxes are likely to be higher than their yield. There are a number o f small taxes the Tax Administration collects onbehalfo f local governments. But many o f these taxes may be administratively too costly to collect. Since local governments do not bear the costs o f the collection o f these taxes, they freely request the State Tax Administration to levy and collect all o f the listed taxes in their jurisdiction (on their behalf). This generates a "moral hazard" problem, which inturn creates losses to the government as a whole. Apparently no data i s collected on the administrative costs corresponding to each own local tax levied. However, examining the figures on revenue collection for each o f these taxes (see Table 2.3) and considering their potential compliance costs, there is clearly room for eliminating several taxes. The Draft Tax Code o f June 2007 rightly proposes the elimination o f some local taxes, but it falls short inpicking all o f the suitable candidates for elimination (as shown inTable 2.4). 30 Table 2.3: Own LocalTaxes and Fees inDescendingOrder inTerms of Collection, 2005 and 2006(in million UAH) 2005 2006 Market fee 345.920 374.040 ~~ Communal tax 152.840 156.400 Tax for advertising 29.340 40.300 Licence fee for placing of trade and services objects 33.02 31.74 Fee on the jockies at the hippodromes 13.920 16.980 Fee for parking cars 13.890 15.690 Fee for development of the Crirnea recreationalcomplex 5.9 5.25 Resort fee 0.980 1.070 fee for permissionto conduct local auctions and lotteries 0.21 0.26 fee imposed pet (dog) owners 0.29 0.26 fee for apartment occupation permit 0.220 0.220 fee for permissionfor TV and film shooting 0.029 0.04 fee on the racing prizes 0.001 0.003 fee to bettingon horseraces 0.001 0.001 Source: State Tax Administration. Table 2.4: Local Taxes List: Current versus Proposedby the Draft Tax Code Current "Own" Local Taxes and Duties "Own" LocalTaxes andDutiesProposed(by the DraftTax Code) Tax for advertising Advertisingfee Communaltax 1 Fee for permission to conduct TV and film I Communaltax Fee for conductingTV and film shooting Sources: Current tax legislationand Draft Tax Code. 2.35 The communal tax overlaps with the enterprise profit tax (EPT), potentially raising administrative and compliancecosts. The communal tax applies to all private legal entities operating in the country. The liability i s calculated by the reported annual wage bill o f a company times a rate. This tax is innature an income tax on business activities but it has a separate structure and collection procedure from the EPT, which potentially increases the administrative costs (for the STA) and the compliance costs (for taxpayers). Moreover, the methodology for establishing the base for this tax i s somewhat odd. Thus, the proposed Draft Tax Code (as o f June 2007) establishes a presumptive (and unified) way of estimating the wage bill o f a firm and its liability, resulting from the average number o f employees times the minimumwage multiplied by 0.5 percent. The draft tax code also expands the list o f liable taxpayers to the regional and local branches o f companies (so that each would be paying the tax for its individual wage bill), non-resident representatives, and to individual entrepreneurs. But under the current tax legislation and the Draft Tax Code, small traders (individuals) and small businesses do not pay this tax if they are under the SimplifiedTax System (STS). 31 2.36 The marketplacefee has a rather complicatedwa of establishingtax liability, which may be leadingto higher administrative and monitoring costs? The rate i s determined by a large number o f special characteristics o f the space leased in a market; thus, aside from square meters, it i s determined by features such as use, hours o f operation, type o f market, seasonality, and the like. Once the rate is selected (within a band) it i s multiplied by the established minimum wage o f the country. This i s an unduly complicated process for establishing liability for this type o f tax, as the Tax Administration follows it for every single trader. The Draft Tax Code (of June 2007) envisages that the sole taxpayer liable for this tax would be the legal entity or individual that owns or has the administration o f a market instead o f each trader. Thus, this legal or natural person that has the responsibility for collecting revenues from all traders renting a space in a market facility. While this proposal has the potential to reduce administrative costs, the features used to calculate tax liability are not changed. Inlocal taxes, simpler i s better, especially in the context o f the high tax compliance costs o f Ukraine. Even if local governments are left to choose the characteristics they want to consider, past experience shows that local governments select the most complicated schemes offered by the legislation (i.e., with more features to determine tax liability) because they do not bear the burden o f actually collecting these taxes. 2.37 Currently, a property tax does not exist in Ukraine. Moreover, the way in which it is proposed in the Draft Tax Code (as o f M a y 2007) departs significantly and negatively from best international practice in terms o f tax base and tax rate determination and the procedures for assessing tax liability. Box 2.5 provides general recommendations and guidelines for implementing a property tax in Ukraine. 2.38 The "trade and service object placement duty" appears to overlap with other local taxes, thereby creating higher administrative and compliance costs. This tax covers taxpayers that are under the communal tax and individual traders that have ownership o f their business establishments. Inaddition, it seems to be aimed at taxing taxpayers under the STS that are not subject to the communal tax. 2.39 The planned introduction of a "construction duty" is unnecessary in the context of the existing system of taxation. This tax would not only increase administrative and compliance costs, but would increase the tax burden significantly and unnecessarily inthe context o f the current system and the reforms envisaged (e.g., the introduction o f a property tax). Recommendations 2.40 The following set o f selected options for reform and recommendations is geared to providing greater predictability and fairness to the local revenue stream: Establish clearer criteria for compensating local governments for revenue falls caused by changes in central tax policy. First, the government should develop the capacity to swiftly estimate the potential revenue losses o f any given tax policy decision. Second, it should establish clear compensation criteria: the criteria could perhaps be based on the estimated loss as a percent o f GDP. Third, the compensation, amounting to the same percent o f GDP, could be undertaken through a special (one-time) grant directly to the local governments in the first and second tiers independently, usingthe proportions o f tax sharing among levels of government as established in the Budget Code, and then redistributing within each tier o f local government on the basis of historical figures. The suggested sequencing would be: to undertake the necessary clarifications/amendments to the budget code in 2007-8 and to implement them in the 2009 budget. ''Something ~~ similar (interms oftax base and tax liability calculation) takes place with the fees for parking areas for vehicles. 32 Stop the granting of tax exemptions that affect local budgets (and the consolidatedbudget in general), and curtail existing tax expenditures. Exemptions to land taxes damage local budgets. 0 Establish clear criteria for compensating local governments for expenditure increases that have been directly caused by changes in central expenditure policy. Since, over the medium term, the size of the government should decrease, compensation could be undertaken as a proportion o f the local expenditure item in the consolidated budget. The suggested sequencing would be: to undertake the necessary clarifications/amendments to the Budget Code in 2007-08 and to implementtheminthe 2009 budget. Consider allocatingthe PIT to the jurisdiction inwhich the taxpayer resides rather than the jurisdiction in which the employer has its establishment. This may lead to some administrative/reportingchanges, but these shouldbe moderately easy to resolve through changes in the information systems and in certain routine procedures in the STA. This would also influence the estimation of revenue capacity in the equalization transfer, and consequently it should be considered intandem with its potential effects inre-calculating this factor over the next two years. The suggested sequencingwould be: to start evaluatingthe best administrative ways of handling this change during 2007-08, to pilot test examining the potential changes in the equalizationtransfers in2008, and to implement in2009. 2.41 The set of selectedrecommendations that follows i s geared to attaining a higher, though prudent, degree o f fiscal autonomy and, to some extent, a higher degree o f local accountability as regards taxation: 0 Create the legislativeframework for a property tax and implement it. This i s an important component o f local revenues around the world and it i s absent in Ukraine. Recommendations related to (i) the tax base and the taxpayers subject to the tax; (ii)fiscal cadastre; (iii) tax a the rates and tax liability calculation; (iv) tax exemptions; and (v) broad mechanisms for administration are contained inBox 2.5. Box 2.5:: Introducinga PropertyTax inUkraine- BasicPrinciples andRecommendations (i) Recommendationsfor the tax base and taxpayerssubject to the tax: o Establish a property tax for both buildings and land. Inthe case of Ukraine, this implies that in urbanized areas (including towns) "land tax" and "real estate tax" (as envisaged currently) should be viewed and collected as a single property tax. The agricultural land tax could continue to be charged with a special regime but under the same system o f management and within a single registry o f property rights for land and real estate. o Establish a comprehensive list of the buildings, dwellings, and other premises subject to taxation inthe text o f the code or related regulation. o Do not include personal or financial leases as part of the tax base (as currently envisaged inthe (May 2007 revisiono f the Draft Tax Code). o Levythis tax onreal estate and landownedbybothnaturalpersons (individuals) and legal entities. The rates imposed, however, may vary betweenthese two types o ftaxpayers (see below). o Make all owners o f the property jointly liable for the tax. A specific personal liability, that is, total liability divided by several owners or their shares (as envisaged by the draft tax code) is not practical. (ii) Recommendationsfor establishingajscal cadastre: 33 o Assign a notional valuelproxy for the market value o f properties through low-cost mass valuation methodologies, generally along the following lines: (a) cities and towns are commonly divided into zones in an attempt to match the notional value that the market assigns to certain zones o f a city or town; (b) once zones are set, a mass valuation tool would produce graduated schedules o f value for properties based on the property's location and use (see below); (c) after zoning, the second tier o f value definition is commonly carried out through the "use" o f the property, that is, residential, industrial, commercial, mixed, or agricultural use; (d) a third tier to apportion value is identified through the property's specific characteristics such as amount o f square meters o f buildings and land (the latter incase the property is not a co-owned building), and the like. o Overlap the fiscal cadastre with the physical cadastre o f properties using several tiers of benchmark value schedules and rates. Inflation and other factors would change over time the assigned value o f properties: thus, the fiscal cadastre would needto be updated annually, including through revising the zoning as needed. (iii) Recommendationsfor establishing tax rate and tax liability calculation: o Establish a reduced number o f rates to be applied to the notional mass value assessed for the property inthe fiscal cadastre (as explainedearlier). Itis likelythat there willbedifferent rates for legalpersons and individuals, and for the use o f the property (industrial, commercial, residential, agricultural). o Set very low proportional rates and then calibrate them using the fiscal cadastre ina pilot exercise. o Allow local governments to set their rates under certain guidelines and bands in the first years o f implementation. Additionally, in cities where value o f real estate is high, equity issues need to be considered, and appropriately targeted reduced rates or fees could be applied to certain segments o f the population (e.g., the elderly living without other income earners inthe property, and the like). (iv)Recommendations on tax exemptions: Grant exemptions to property owned by the government, embassies (based on reciprocity agreements), international organizations, religious organizations, and education and health care establishments. NGOs are commonly exempt as well (unless they are used by subsidiaries or parent companies as a tax shield). Granting tax brakes based on the number o f square meters of the real estate, as envisaged by the draft tax code (e.g., having a tax breaks for real state below a threshold o f 100 m) is inappropriate and should be avoided as it could generate a variety o f perverse incentives inthe real state market. As recommended above some special and targeted exemptions to some segments o fthe population couldbe granted. (v) General recommendations for administrative mechanisms: Clearly divide functions between the State Tax Administration and local government administration. Let local governments manage the fiscal cadastre (in coordination with a single registry o f land and real estateproperty). The nature o f this tax is local and its proper implementation and monitoring must rely significantly on local authorities' knowledge o f the local real estate market. Moreover, local governments are critical to maintaining the appropriateness o f the mass valuation process o f property, and, later, to updating the fiscal cadastre, including zoning, values assigned, and property characteristics--especially considering the fact that real estate markets are dynamic. Local governments should also set up an appeals system for value assessed with the oversight o f other regional authorities. Let the Tax Administration be incharge o f collection, enforcement, and taxpayer services. Allow local governments to change the rate on "own" local taxes. Letting local governments set the rate (even within flexible bands for all own local taxes and fees) may enhance revenue mobilization and accountability at the local level. Local governments have a better understanding of their tax bases and the potentialbehaviors with rate changes. On the other hand, granting local governments the ability to change the bases (that is, to choose taxes outside of the established list o f -own- local taxes) i s not recommendable at this point inthe reform process. - Over time, create the institutional framework for allowing local governments to actually collect their own localtaxes. The current framework, inwhich the STA handles all collections in the country, i s appropriate, particularly given the administrative capacity level o f local governments. Over the medium term, however, as local capacity improves (as has been the case 34 since the budgetreform), and incoordination and within a fullreporting framework with the STA and Treasury, local governments could begin to collect their own local taxes. This would also improve their tax autonomy and their tax accountability to their citizens. 2.42 The following set of selected options for reform and recommendations is geared to help reduce tax compliance and administrative costs: 0 Revisit the way inwhich the communal tax is levied. An easier way to collect revenues might be to implement a local (and very moderate) "piggybacking" on the corporate income tax (CIT) rate, or simply to share with local governments a small portion of this tax: for example, less than 0.3 percentage points o f the CIT rate represents more than double the revenue collection o f the communal tax. 0 Revisit the way of levying the market fee. A recommendable and straightforward way to reduce compliance and administrative costs would be to simply let local governments charge a fixedfee inUAHper squaremeter leasedpermonth (withina range or bandestablished under the code and with automatic annual indexationfor inflation) without any additional criteria. 0 Avoid the overlapping created by the "trade and service object placement duty." There are two ways to eliminate duplication in this tax while maintaining its intended purposes, for example: (i)by creating a special section under the marketplace tax to tax owners of trading places; and (ii) by levyingthe communal tax on legal entities and individuals underthe STS. 0 Avoid the introductionof the constructionduty. The implementation o f such a burdensome tax i s not recommendable in the context o f the current system; it should not even be included in the options list from which local governments can choose. 0 Eliminate local taxes that generate higher administrative costs than revenue yield. As of today, current candidates include: (i) the fee for an apartment occupation permit; (ii) fee for the betting on horse races; (iii) the fee on the racing prizes; (iv) the fee for permission for television and film shooting; (v) the fee for permission to conduct local auctions and lotteries; and (vi) the fee imposed on pet (dog) owners. The draft tax code follows some o f these suggestions but falls short inits streamlining. E. ASSESSMENTOFANDRECOMMEDATIONSONINTERGOVERNMENTAL TRANSFERS IssuesRegardingIntergovernmentalTransfers5* TheEqualization Transfer 2.43 The equalization grant has reasonable equalization power, but other transfers and subventions undermine its role?' The analysis employing data from oblast consolidated budgets and the analysis usingrayon level data show that the equalization grant is helpingto equalizejurisdictions that 58Annex 2.1 containsagraphic representationof intergovernmentaltransfers system 59Equalizationpower is the degree to which the equalizationformulawould allocate resourcesin favor o f lower-incomeregions. Higher equalizationpower meansthat lower-incomeregionswould receivemoreresources from the funding formula thanhigher- income regions. From the point of view of expenditures, higher equalizationimplies that more resources would be allocatedto regionswith higher spendingneeds relativeto regionswith lower spendingneeds. 35 have lower means after tax revenues for a given year (Figure 2.4). However, considerable variability across local governments becomes evident when other transfers and subventions to local governments are added to the sum o f the tax revenues and the equalization transfer. This is also evident from a look at the expenditure side and the inequalities in capital spending allocation that favor wealthier regions, as presented inChapter 1. Figure2.4: Rayonand City per CapitaRevenuesBefore andAfter Equalization, and After Other Transfers and Subventions 3500 I 3000 2500 Revenues 2000 1500 Revenues 1000 Equalization 500 /Revenues 0 Note: Rayons and cities are ordered from higher incomes to lower incomes (from left to right). This study used PIT per capita collections -as a proxy for income level by rayon. Sources: MoF; State Treasury o f Ukraine; State Statistics Committee o f Ukraine; Bank staff calculations. 2.44 Moreover, the transfer formula itself contains several design problems, including a high degree of unnecessary complexity. The structure o f the formula in2001 and the modifiedversion after the implementation o f the Budget Code in 2002 were already burdened by too many factors and sub- formulas. Since then, there has been further backslidingthrough the addition and changing o f factors and coefficients to make the formula the "black box" it i s today. The equalization formula encompasses around 70 main sub-formulas and literally hundreds o f factors and adjustment coefficients (for estimating both expenditure needs and revenue capacity).60 2.45 Despite the many actors and adjustment coefficients added to the formula, the equalization power of this transfer (as of 2005) had not changed significantly compared to 2001. The 2001 equalization transfer was a "half-way" reformed transfer still administrated by the oblasts for its distribution to the second tier (rayons and cities). Over the years, a large number o f factors and adjustment coefficients were added without having any significant effect on the equalization power (see Figure 2.5). This, in a way, should be expected. Instead of improving equalization power, added complexity leads to a lack of predictability and transparency for intergovernmental transfers, as further elaboratedbelow. 6oThis occurs for each of the equalization formulas: for oblasts, for rayons, and for the formula from rayons to villages and settlements. All o f these formulas are similar; and the differences lie in some o f the factors. 36 Figure 2.5: Equalization Power ChangedLittle Between2001 and 2005 (by oblasts, consolidated budgets) 450 lote: Oblasts, ARC and cities o f special status are ordered from higher to lower left to right per capita regional (oblast) value added. Sources: MoF; State Treasury o f Ukraine; State Statistics Committee o f Ukraine, Bank staff calculations. 2.46 The excessivecomplexityof the equalizationformula affects itstransparency. Transparency inany type offiscal transfer not only meansthat the transfer needs to be criteria based (which is the case with the equalization transfer inUkraine) but also that all stakeholders, particularly local governments, should be able to understand the workings of the formula and to calculate/forecast on their own the amount that they can expect over the following years. With the current level o f complexity this i s not possible, which creates a serious transparency problem inthe system. 2.47 Allowing new factors each year has opened the door to political bargaining for local governments,which is counter to the reformgoals of the Budget Code. Since 2002, dozens o f factors and adjustment coefficients have been added, mainly because o f pressures from local governments that wanted the formula to accommodate their own interests and characteristics in hopes of obtaining more funding(every year duringthe budgetprocesslocal governments lobby their interests withthe Ministry of Finance). This i s a useless and highly counter-productive practice since no local governments can, in reality, obtain any additional benefit through one additional factor inthe midst of hundreds (with roughly the same weight), and everyone loses in terms of simplicity and transparency, including the intergovernmental fiscal framework itself. 2.48 Serious predictability problems arise from the constant changing of critical features of the equalizationformula. This bringsuncertainty to local governments and further reduces predictability, as can be seen in the data examined from 2001 to 2005 (see Figure 2.6). There is a need for significantly improved revenue stability inthe equalization grants 37 Figure 2.6: Year-on-Year Change in the per Capita EqualizationReceivedby Local Governments (oblast consolidated) (inpercent) 160 I 140 120 ! 100 - ao 60 .-t-2003-2002 40 2004-2003 20 0 -- 2005-2004 -20 L Notes: Oblasts, ARC and cities o f special status are ordered from higher to lower (left to right) per capita regional (oblast) valueI added. Transfers received by the oblasts Zaporizhzhya, Odessa, and Dniepropetrovsk are not included. Sources; MoF; State Statistics Committee o f Ukraine; Bank staff calculations. 2.49 The estimation of "expenditure needs" in the formula is unnecessarily complicated. The "expenditure needs" side o f the formula contains components for estimating a notional need o f resources for the following six areas: health, education, social care services, culture and the arts, sports, and administration. The number o f factors and adjustment coefficients within each o f these components has further increased over time. Currently, for education alone, there are some 32 factors and adjustment coefficients that, through a collection o f sub-formulas, determine the notional expenditure needs in this sector within the equalization formula for rayons. A similar number o f factors and coefficients are used to calculate education expenditure needs for oblasts, the ARC, and cities o f special status. For health care, there are around 23 factors and adjustment coefficients for estimating the notional expenditure needs for rayons and cities (in addition, there are between 12 and 15 factors for oblasts, the ARC, and cities o f special status). Aside from health care and education, each o f the components encompasses a variety o f formulas, factors, and adjustment coefficients as well. Because o f the weighting implicit in each factor, and because o f the large number o f factors involved, none o f them bringsany value added to (or changes) the outcome o f the formula. 2.50 Many factors in the formula have worsened in the incentives they provide to local governments. Many of the factors and adjustment coefficients introduced since 2002 revert to the old bad practices. A number o f coefficients that alter forecasted spending needs are geared to account for specific characteristics o f the existing large (and under-utilized) network and staffing rather than to financing pure demand for services in a locality. Some o f these issues are discussed briefly in the list below for health care and education.61 Additionally, inrelation to the component on social assistance, the DFIDreport "Facilitating Reform in Social Services" (FRSSU) 2007 argues that the transfer formula (inthat component) i s excessively detailed and reinforces an inefficient provision o f social assistance services through residential institutions, which impedes the development o f community-based alternatives that are responsive to demand. 38 0 Several factors and adjustment coefficients in the health component of the "expenditure needs" estimation perpetuate inefficiency and hamper the proper functioning of the transfer. For example, there i s an adjustment coefficient that allocates a greater amount o f resources to local governments that have a higher ratio o f wages to total expenditures in the sector: this provides an incentive for keeping high levels o f unneeded staff. There i s an adjustment coefficient that accounts for the relative costs o f providing services by age and gender, and while this i s good inprinciple, the fact that there are 16 age groups for each gender in the calculation o f this single coefficient complicates the allocations, especially since there are some 32 other factors and coefficients for health care alone. 0 There are factors accounting for specific programs within the health sector that have no place in a block (non-earmarked) grant such as the equalization transfer. Over time, several factors (between three and four depending on the level o f local government) were created to account for needs arising from certain specific sectoral programs, for example, the program for pancreatic diseases. The accounting for such needs does not render any benefits in a block equalization grant, under which, intheory, local governments can use the receipts as they see fit. 0 Some factors and adjustment coefficients in the formula for education "expenditure needs" foster inefficiency. There are a variety o f factors and adjustment coefficients (more than a dozen) designed to allocate fundingper pupil,but the allocation depends on the type o f institution the pupil attends. Thus, fknding also follows implicitly the network o f schools. There i s an adjustment coefficient that awards more resources to schools and local governments with smaller class sizes, that is, fewer students per classroom (Chapter 4 provides a more detailed discussion). This provides an incentive for maintaining inefficient class size levels and an inefficient network o f establishments and staffing. On the contrary, local governments should be provided with incentives to merge schools in order to reach more efficient class sizes and thereby optimize and save resources that could be better investedelsewhere inthe sector. 0 There are factors accounting for specific programs within the education sector that have no place in a block grant such as the equalization transfer. Over time, several factors (between three and six depending on the level o f local government) were created to account for the needs arising from certain sectoral programs, such as allocations for computers, other materials, and scholarships. Again, accounting for the need for such programs does not provide any benefits ina non-earmarked transfer. 2.5 1 The dozens of adjustment coefficients added to differentiate mountainous areas from the other areas are unnecessarily complicated. Attached to the considerable number o f factors that determine a notional amount o f expenditure needs are a large number o f coefficients intended to differentiate local governments and service recipients in mountainous areas. This coupling creates more than 40 extra variables in the formula. While this differentiation is important and should be accounted for, the mechanics used could be mathematically managed ina much simpler way. 2.52 The component that estimates revenue capacity in the formula i s to some extent appropriate but has been overridden by ad hoc adjustments when tax policy changes have affected local revenues. The Budget Code establishes a formula for estimating the fiscal (revenue) capacity o f local governments that relies on adjusted moving averages o f past performance (rather than on an actual forward estimation). This approach, while not fully accurate, i s reasonable given the data constraints to constructing the estimators (namely, the lack o f precise data on all tax bases in each local government). However, during the reform o f the PIT 2004 this has not, in fact been applied and instead ad hoc procedures have beenput inplace. 39 2.53 Incentives for fiscal effort6*are currently weak in the equalization formula. The Budget Code establishes (in Article 98) that the fiscal capacity estimate would be fixed for three years; this had the goal o f providing local governments with the incentive to collect more revenues without being affected by a reduction in transfers every year. At least in the short run this would have proved a workable tool for encouraging fiscal effort. Instead, in 2005 the government introduced a revenue "incentive" factor into the formula. This factor rewards local governments that had a ratio o fre-calculated fiscal capacity for 2005 to revenue capacity o f 2004 that was higher than the national average (Thirsk 2005). The first problem with this approach i s that it rewards a poorly "estimated" change in revenue capacity as opposed to a true fiscal effort. The second problem i s that the local governments that were not benefiting from the approach, that is, those with a ratio below the national average, had their estimated "revenue capacity" artificially inflated inthe formulation, and consequently had their transfers reduced so that the overall fiscal envelope would not be increased because o f this "incentive" factor. This inappropriate way to generate fiscal effort raises fairness concerns. 2.54 Asymmetric changes in the equalization coefficient (also called "alpha") damage the objectivity of the formula. The equalization adjustment coefficient inthe equalization formula ("alpha" o r a ) was changed over the years, but starting in 2005 it was changed asymmetrically (that is, with different coefficients across local governments). This differs from changing the level o f the coefficient (that is, the level o f equalization for all local governments). Under the new asymmetric arrangement, donor local governments, that is, local governments with greater revenue capacity than their expenditure needs, have an "alpha" coefficient lower than 1, which, according to the most recent amendments to the regulation for the 2007 budget, may range between 0.6 and 1 depending on the growth o f the funds contributed to the pool by the local government. This allows donor local governments to pay less to the pool than the difference between their revenue capacity and their expenditure needs. On the other hand, recipient local governments (that is, local governments that have a lower revenue capacity than their expenditure needs) have an "alpha" coefficient o f 1. This change was geared to produce extra revenue efforts in wealthier local governments, which i s a laudable goal, but it i s not an appropriate vehicle for thispurpose. 2.55 Aside from the technical issues of the formula, there are issues with its administration. In the Ministry o f Finance (MoF), those departments and units working on the equalization formula allocation, do so almost in isolation from one another. Thus, none o f these departments or units has a comprehensive understanding o f how all o f the components and sub-components o f the formula work together. This arrangement i s far from good international practice and impedes a comprehensive view o f the existingproblems and thus hampers efforts to resolve them ina timely manner. CapitalTransfers 2.56 Transfers/subventions for capital expenditures are not transparent and lack consistent and stable criteria. The procedure for the selection o f capital projects for funding from subventionshransfers from the state budget is usually regulated by ad hoc Cabinet o f Ministers resolutions that are approved annually during the budget cycle.63Local governments typically bargain with the central government for this type o f allocation, creating a serious transparency problem. This is one o f the "holes" inthe Budget 62Fiscal effort i s the degree to which a region utilizes the available revenue potential. Inother words, it is the level at which it raises revenue, given its fiscal capacity. This is typically measured as the ratio o f the actual collection o fregional revenues (taxes, premiums, etc.) to the gross regional product. 63 The 2004 resolution, for example, was just a list o f capital projects prepared on the basis o f suggestions from members of Parliament. The 2005 resolution, which was at least based on criteria, established a mechanismthrough which the amount of the subvention for each oblast would depend on the population. 40 Code that needs to be filled intandem with improving the overall process o f capitalbudgeting at the local level (see Chapter 5). 2.57 Moreover, there is a lack of predictability for receiving funds for capital investments. Transfers for public investments vary significantly from year to year (see Figure 2.7). These differences in (year to year) funding are very large and go beyond the expected year-on-year change in project relevance across local governments. Not even a minimum funding to renovate and further avoid the deterioration o f existing investments i s assured under the current system. Since there i s no certainty, local governments are unable to undertake consistent planning. International experience shows that lack o f predictability may lead to poor capital expenditures execution at the local level, and this seems to be the case in Ukraine (see Chapter 5). When local governments are not able to plan investments properly and over the medium term owing to the lack o f predictability in resource allocation, they are likely to experience problems in usingthe resources effectively when they arrive. Moreover, lack o f planning and poor execution may in turn significantly reduce the efficiency of spending in critical areas o f basic infrastructure. Figure2.7: Year-to-Year Changeinper CapitaCapitalAcquisition by Oblasts(consolidated), 2002-05 I - * 2002-2003 - --l.~- 2003-3004 --- 2004-2005 \late: Oblasts, ARC and cities o f special status are ordered fromhigher to lower per capita regional value added (left to right). Sources: MoF; State Statistics Committee o f Ukraine; Bank Staff calculations. 2.58 Finally, capital spending allocations (mostly through capital subventions from the budget) are unequal across local governments, favoring richer oblasts and rayons (see Figure 2.8). Moreover, the difference between the highest and the lowest per capita allocation has been increasing since 2002. These inequalities are partly a reflection o f larger capital stocks inricher regions but also are a result o f following: (i)the lack o f criteria and a stable mechanism for the distribution o f capital subventions/transfers; (ii) the existing political bargaining (more powerful inricher regions); and (iii) co- financing schemes which allow more leverage to richer regions (discussed indetail inChapter 5) 41 Figure2.8: Wealthier Oblasts and RayonsTend to ReceiveHigher per CapitaFundingfor Capital Investments PublicCapblExpendituresby Oblast2002.2006 22M- ."" Rayons Capital investments on Per Capita Basis 2005 3 180. , I A 350 0 Jotes: Oblasts, ARC and citiesof special status are orderedfrom higher to lowerper capitaregional(Oblast) value added (left to right). Rayons and cities are ordered from higher income to lower income (from left to right). This study used PIT collectionsper-capitaas proxy for incomelevelby rayon. Source: MoF, State Statistics Committee of Ukraine, Bank Staffcalculations Transfers Below the Rayon Level Cfrom rayons to small towns, villages and settlements) 2.59 Transfers below the rayon level mirror the structure, complexity, and problems of the transfer system from the central government to the first and second tier local governments. The equalization formula i s too complex to be manageable at this level. In addition, capital transfers are non- transparent, and other budget support i s on an adhoc basis. Recommendations 2.60 The equalization transfer formula should be revised, aiming for a comprehensive simplification and elimination of ill-designed incentives. The equalization formula could have the same (and better) equalization power without its current complexity which damages transparency and predictability. The specific options/recommendations listed provide a sequenced approach in order to avoidradicalrevenue changes for local government^.^^ Start by streamlining the number of factors and adjustment coefficients in all six components of the expenditure needs estimation (health, education, welfare and social benefits, culture and arts, sports, and administration). This should be undertaken for the expenditure needs estimations for oblasts, rayons and cities, and for the transfer formula fiom rayons to villages. The goal for each component should be to contain not more than four factors based purely on demand driven factors, not related to the type of facility and the financial allocation per head. 0 Inparallel, start the process of estimating appropriate per head standardcosts for different services to improve the equalization formula (called factor "N" in the legislation of the equalization formula). 64Annex 2.2 presentssome guidingprinciplesfor settingintergovernmentalfiscal transfers, includingequalizationtransfers. 42 0 Eliminate factors geared to funding specific programs in each sector. It should be borne in mindthat the equalization transfer inUkraine is a block grant (withno strings attached to its use by spending sectors), andthis shouldbereflectedinthe formulation. 0 Base these factors that are kept solely on objective data over which neither the local governments nor the central government has any influence, such as demographic factors (age, gender), population served, actual demand for services, and the like. Moreover, these factors should not be attached to any type of network characteristic. 0 Condense the dozens of adjustment coefficients accounting for mountainous area factors into a single coefficient in the formula. This i s a simple mathematical change- and would streamline the formula significantly while keepingthe factor inplace for adjusting all expenditure needs. Moreover, this adjustment factor shouldbe very "soft)' (that is, close to 1). 0 Moving forward, start to collect and analyze data on the differences in the costs of delivering services. Currently, mountainous and rural areas are favored by adjustment coefficients, assuming that the unit costs of delivering services are substantially larger; this may be the case, but it could also be argued that price levels incities are substantially higher than in those areas, and consequently that there i s an automatic balancing effect. Stop changing the way in which revenue capacity is calculated. For now, and until detailed data on tax bases are collected to perform forward looking estimations (instead of estimations based on historical collections), the estimation should adhere to the principle established by the Budget Code of 2001. Revenue capacity should not be artificially inflated or reduced without a proper coefficient to account for fiscal effort (that is, the existing method i s not appropriate (see below). 0 Create a proper coefficient for fiscal effort, but in the meantime use the mechanism articulated originally in the Budget Code (that is, freezing revenue capacity for a period of time, namely two to three years). Ideally, anew coefficient for fiscal effort shouldbe constructed based on data and estimations on the bases of each tax in each locality so that they can be compared to actual collections (and changes in collections). The STA could provide part of the data needed, but this may be a complicated exercise. Alternatively, a proxy for local gross value added or total income could be generated for local governments and then used to calculate fiscal effort: gross regional value added figures already exist by oblast, thus, an extrapolation to account for "within" oblast differences may be calculated (includingthrough the use of wage figures). 0 Keep the "alpha" coefficient at the same level for all local governments. Asymmetric changes take away the good principles still left in the formula. The level of equalization can be changed from time to time, but equally at the same level of "alpha" for all local governments; although frequent changes inthe level of equalization are not recommended. 0 Simplify the equalization transfer from rayons to villages and settlements, in parallel with making the rayon the core local government. This involves, as a first step, the streamlining of the equalization formula at this level inparallel with the process suggestedabove for this transfer from the central government to oblasts and rayons. Moving forward, and synchronized with the recommendations to make the rayon the core local government level for the delivery o f most social services, the actual shape of the equalization transfer could be transformed to a simpler transfer that would account only for the number of inhabitants or service recipients ineach village 43 or settlement (with some minor adjustments). This would provide the rayons with more flexibility to manageresources and services throughout thejurisdiction. 0 Strengthen the administration of the equalization transfer and all other transfers by initially placing the responsibilitiesunder a single department in the Ministry of Finance. This would provide senior and other officials from this ministrywith a more comprehensive view of the needs and problems o f the system in a timely manner. Over time, the government should consider the creation of a grants commission to handle all transfers to local governments. The proposed commission could be linked to the Ministry o f Finance but with more independence, and devoted itself exclusively to improving the systemina transparent way. 2.61 Suggested sequencing: first, start the revision and testing during 2008 o f streamlined formulas during 2008; a streamlined formula couldbe implementedduring2009. 2.62 A formal criteria-based and stable mechanism for the allocation of capital subventions should be established. This would curb political bargaining and improve transparency andpredictability inthe allocation of resources for capital spendingto local governments. This criteria-based or formula- based system should reach the first and second tiers o f local governments directly. The following steps and suggestionscan be considered: 0 Establish a pool of resources for capital subventions/transfers. This pool (which should be increased from the present level of local capital spending) could be a notional share of the consolidated budget (or the total local budgets). It could then be further divided into sub-pools (for at least two types of capital transfers), as suggestedbelow. 0 Do not tie the proposed pool to any specific "targeted program." Currently, the so-called targeted programs (as part o f the development budget) crowd out the possibilities for more (and new) investments. Moreover, they create inflexibility incapital budgeting, which ultimately leads to inefficient spending on top of the under-execution of projects (see Chapter 5). 0 Consider establishing a (first) minimum transfer to cover the basic renovation needs of oblasts and rayons and cities. This subvention could be allocated simply on the basis o f the number of inhabitants ineach oblast or ineach rayon and city.65The objective of this allocation would be to cover repairs and renovation needs in the current infrastructure to avoid further deterioration. 0 Consider a second subventiodtransfer (potentially with a larger pool than the first) to be grantedfor critical on-going and new projectson a competitionbasis.This transfer could also be granted to broader multi-year investment programs of local governments. But clear criteria for project and program evaluation and selection need to be established in the regulation. These criteria should be based on factors such as national priorities, positive externalities across jurisdictions, and clear economic benefits (as articulated in a cost-benefit analysis for each project, even if it i s within a larger program). Inthe case o f this secondtransfer, rayons and cities could apply through the oblasts (each oblast being a project screener and consolidator) or, alternatively, the government could consider a separate direct path of application for funds for secondtier local governments. The project selection for this second transfers eachyear should be transparent for all stakeholders and should be announced before the first reading of the budget in the Parliament. The selection process could be handled by the Ministry o f Economy (under clear guidelines) incoordination with the Ministryo f Finance. `'Perhapsincluding a "soft" (that is, close to 1) adjustment coefficient recognizing the higher costs o f investments in mountainous areas. 44 2.63 Suggested sequencing: Discuss and establish the criteria (as recommended) for these capital grants in2007-08, implement a pilot in2008, and fully implement the new system in2009. F. STRENGTHENINGTHE SUB-NATIONALBORROWINGFRAMEWORK IssuesRegardingSub-nationalBorrowing66 2.64 Local governments and communal services enterprises may need to borrow in order to finance their large infrastructure rehabilitation and development investment needs over the next 10 years (see Chapter 5). This section examines the key impedimentsto a prudent development o f the sub-national debt market as well as key policy reform priorities inaddressing these impediments. Indoing so, it comments on the draft amendments to the local government borrowing framework inthe context o f the Budget Code (hereinafter called the LGB amendments) adopted by the Council of Ministers in November 2006 and subsequent revisions, as well as on other relevant laws andregulations. Issues ConcerningLegal Provisions 2.65 The current draft LGBamendmentsdo not includedebt refinancingas a legitimatepurpose of local government borrowing. Moreover, they prohibit borrowing agreements that contain provisions for the repayment of local borrowing by way of offsetting counter-claims. This would avoid the proper restructuring o f debt in some cases, which may be worth doing (for example, restructuring large infrastructureprojects under better financial conditions). 2.66 There is no framework for local government bankruptcy. This creates a variety of problems, including moral hazard for local governments and lenders since the possibility ofbailout remains open. 2.67 The BudgetCode does not providethe authority to cities (or rayons)to pooltheir resources and borrow under joint undertakings. Inthe case of some mediumsize local governments that share, or would benefit from, the use o f an infrastructure asset (such a as a bridge), joint undertakings and borrowing that would be needed or desired have no viability under the current legislation. This might hamper efficiency and market access for some local governments. 2.68 There are several deficiencies inthe legalprovisionsthat regulate guarantees. For example, local government guarantees may not be issued to finance investment projects in which the budget funding exceeds 50 percent of the project size, thereby preventingthe issuance o f partial credit or partial risk guarantees in the case of co-financing structures (such as those associated with EUpre-accession funds). Also, regulations do not permit the structuring of partial credit guarantees to local government bonds based on a pre-set strike at default. 2.69 A key feature omitted in the local borrowing framework (and the planned amendments) is the issue of short-term debt arising from "accounts payable," and this carries potential fiscal risks. Accounts payable (that is, short-term debt generatedthrough contractors and providers) as a specific debt generating item i s missing in the definition of local government debt. Thus far, debt accumulation through this channel has not been substantial in Ukraine, mainly because of appropriately tight Treasury controls, but it i s a latent problem, This was one o f the issues that several countries in Latin America missed addressing at the beginningo f their fiscal decentralization reforms, and, as a consequence, local governments accumulated short-term debts through that channel to unsustainable levels. 66This reportdraws on the Ukraine's sub-nationalborrowingNote preparedby MichelNoel (2007). 45 Authorization and Controlsfor Local Government Borrowings 2.70 Controls and authorization under the Budget Code and the planned amendments are sound. In addition to the normal registration requirement, the draft LGB law establishes ex ante control o f local government borrowing by the Ministry o f Finance. This ex ante control i s justified in the short term, given the limited development o f the market and the limited experience o f the local governments in this area. 2.71 The documentationrequiredas part of the registrationprocessinthe Ministry of Financeis appropriate, but some features could be strengthened.Currently, the documentation does not require any information from local governments concerning pledges o f revenues. Under the Budget Code, the non-earmarking principle does not allow for the segregation o f specific revenues under the TSA for local governments. But at the same time, the Ministry o f Finance letter (order) o f September 2004 allows local governments to open bank accounts, which opens the door to pledging revenue streams as they are transferred from the TSA to the local government's bank account. This situation creates uncertainty and represents a "hole" inthe legislation. 2.72 Resolution208 contains a number of additional conditions for local governmentborrowing that complicatethe system unnecessarilyat this stage. For example, Resolution 208 stipulates that all debt securities are subject to mandatory credit rating, which may be a complicated process and may be unnecessary for many local governments. Market Distortions, Moral Hazard, and Transparency 2.73 The lack of full disclosure of the ultimate ownership of financial institutions undermines transparency and efficiency in local governments- borrowing. This situation i s merely a reflection o f what currently takes place in the financial system. Connected lending may distort local government choices inproject selection, leading to reduced efficiency in capital expenditures. Moreover, it increases moral hazard inlocal governments and for certain lenders, particularly inthe context o f a myriad o f local and regional banks. 2.74 The accounting format for local governmentsfalls short of the necessary standards that would allow proper credit analysis. Moreover, the fiscal data are subject to internal audits only, and these audits are infrequent. Additionally, the treatment o f guarantees in local budgets i s not up to international standards. Recommendations 0 Implement a local government regulatory framework for bankruptcy. The government shouldrevise the draft LGB law accordingly, and based on internationalbest practice. Apart from reducing moral hazard in sub-national borrowing, a bankruptcy framework would also provide adequate incentives for lenders to be conservative andjudicious intheir analysis. 0 Allow debt refinancing as a legitimate purpose of borrowing. In many cases refinancing is a rightfuloperation that allows more flexible terms for fundingcapital investments. It is the task o f the lenders to evaluate the financial feasibility and the nature o f refinancing in a local government. 0 Provide authority in the regulation that allows local governments to pool their resources into joint undertakings and to borrow jointly. This would permit important synergies and 46 would create economies of scale inthe financing and use o f infrastructureprojects. A concise but specific articulation in the Budget Law would be needed in addition, clarifying joint responsibilities and the consequences of failing to comply with them, in order to deter disputes and negligentbehavior. 0 Revise the current regulations and planned amendments related to the issuance of local government guarantees for projects with minority private sector participation. The law should allow partial credit guarantees based on a set strike at default, and relying on MPL methodology or simplified risk scoring methodology to establish guarantee provisions and guarantee fees. In addressing this issue, the government may have to introduce comprehensive amendments to the Public-Private Partnership (PPP) Law to regulate all types o f PPP transactions. 0 Revise the current regulations and planned amendments to include accounts payable and guarantees at issuance as local government debt. Including accounts payable with contractors as a debt itemwould allow for proper control andmonitoring o f true debt levels. 0 Improve the ex ante Ministry of Finance controls by setting rules to avoid unnecessary delays and transparency in the authorizationprocess for borrowings. To avoid delays inthe authorization o f borrowings, the Ministry should formally notify local governments of its decision to deny authorization within a short period of time (for example, 15 working days). Decisionsto deny authorization should be made public, and the local government should have the right to appeal under transparent procedures. Over the medium term, and with more maturity in the system, measures prior to registration could be relaxed, but registration needs to remain in place as well as prior controls for borrowings in foreign currency (and for borrowings requiringa sovereign guarantee or counter-guarantee, obviou~ly).~' In addition, the Ministry could implement a periodic control/monitoring of Ministry of Finance debt limits (for example, on a quarterly or semesterbasis). 0 Clarify the situation regarding revenue pledges to secure debt operations. This chapter recommends that the draft amendments to the BudgetCode include a specific article authorizing revenue pledges only for revenues coming from "own" local government taxes, which would mean prohibiting pledges o f shared taxes (such as the PIT, the land tax, and single tax) and revenues coming from the equalization transfer and other subventions/transfers from the central government. The rationale behind this recommendation i s that local budgets still face serious rigidities and pressures in fulfilling their mandates and cannot afford to pledge shared taxes or other grants from the center as borrowing guarantees. 0 Revise and then consider abolishing Resolution 208. This resolution creates cumbersome procedures without adding value. The regulations should let borrowers and local governments decide whether or not to obtain a ratingbasedon market incentives. 0 Allow lending to local governmentsonly to bankswhose ultimate beneficialowners are fully disclosed.The disclosures and background checks should beprovidedand administered under the 67Over the medium term, M O F authorization couldbecome automatic unless rehsed within the 15 day time limit. Over the long- term, and if local government borrowingmatures with fiscal responsibility, there could be scope to move from ex ante to ex post control o f local borrowing by the M O F initially with overall debt ceilings enforced periodically. Moreover, as markets mature, the prudential framework could be revised moving toward market regulation in selected cases, ifthat is accompanied by measures to increase fiscal autonomy as well as enhance their accountability to local constituencies. 47 broader efforts to legislate ultimate ownership inthe financial systemand under the NBUstrategy (and enforcement) onthis issue. 0 Revise the accounting framework established for local governments in line with best international practice and the EU directives on this issue. Inaddition, the regulations should establish a requirement for large local governments (above 300,000 in population or any other appropriate threshold) to provide external auditing of financial statements. G. IMPROVINGOTHERCRITICALCROSS-CUTTINGISSUES Treasury Loansfor Local Governments 2.75 The current ad hoc government practice of granting selectedlocalgovernment interest-free Treasury loans softens budget constraints. The Budget Code rightly banned Treasury loans, with the exception of short-term (up to three months) interest-free loans to finance liquidity gaps due to seasonal fluctuations in revenue collections (as stipulated in Article 73). However, the 2004 Budget Law introduced medium-term Treasury loans, which go against the framework established by the Budget Code. Both short-term and medium-term Treasury loans have increased over the last two years (see Figures2.9 and 2.10), and several outstanding arrears were written off under the 2005 amendment to the Budget Law. Figure 2.9: Treasurv Loans Received, 200-05 (oblast consolidated) 600 500 400 300 2003 E2004 200 02005 100 0 ote: Oblasts, ARC and cities of special status are ordered fromhigher to lower left to rightper capita regional value added. - Sources; MoF; State Statistics Committee o f Ukraine; Bank staff calculations. 2.76 These inter-budgetary loans also harm the development of the sub-national borrowing market and harm transparency. There are five key reasons for this: (i)treasury loans involve the MinistryofFinance inthe role of credit allocation andpricingto local governments; (ii) are open to they 48 political pressures and patronage; (iii) may encourage expectations of a future write-off by the there Ministry; (iv) they constitute a ready instrument for bailing out local governments incase of default; and (v) they can leadto moral hazard inthe local government debt market. Figure 2.10: Increasing Treasury Loans and Decreasing Private Loans, 2003-05 5000 4b 4000 -A- TotalGovernmentCentralLoan8 1 3000 received 2000 --Total Private Loans 1000 0 , I Source: MoF; State Statistics Committee o f Ukraine; Bank Staff calculations. Fiscal Envelopefor Local Governments 2.77 The overall pool of resources for local government transfers and subventions is not fixed. Every year in the budget process, in a fragmented way, the central government establishes the pool o f total resources for each transfer and special subvention for local governments. The implications are two- fold. First,there i s the implication that predictability inlocal government fundingis poor. Second, there i s the implicationthat the central government can pass down its own deficits to local governments, which, in fact, has been occurring through the years, either through under-funded mandates or thorough policies taken unilaterally that affect local budgets, and for which there i s no proper compensation (as discussed earlier inthis report). However, the decision regardingwhether to fix a pool needs to be dealt with in the context of fiscal responsibility rules for local government and the downsizing of the government as a whole. Other Issues 2.78 The laws and regulations that relate to intergovernmental fiscal relations are scattered. In addition to the guidelines established by the Budget Code, there are a number o f laws, regulations, and amendments, many of them established through the annual budget laws. Regulations that are so widely spreadreduce the clarity o f the system. Recommendations 0 Stop granting medium-term Treasury loans to local governments and phase out short-term (three month) loans over the next two years. As argued earlier, central government loans may create moral hazard and may raise transparency concerns. As an option the government could consider keeping only short-term loans for very special cases, for example, for smaller local governments that would not have access to market financing, and under justified liquidity emergencies. But a clear threshold for the size o f government allowed to receive a short-termloan 49 needs to be established inthe legislation. Moreover, these loans would need to be fully integrated into the intergovernmental fiscal framework, including offsetting such support against other transfers receivedby local governments if short-term loans are not repaidintime. 0 Establish a clear rule for the "minimum" pool of funds for transfers and subventions for local governments. This rule might be articulated through a specific share o f the total consolidated budget. The idea i s to have a "floor" and a "ceiling" for the level o f total fundingfor local governments. Determining the pool o f resources for local governments through a constant share in total government funding, would allow adjustments to be made in tandem with the needed reduction o f the overall fiscal envelope o f country. However, such a rule would be clear and avoid the ad hoc passing-down o f deficits to local governments. With each new mandate passed to local governments the floor should be increased accordingly to maintain the level o f spending. This could only be implemented inthe presence o f other important rules o f local fiscal responsibility, such as the elimination o f medium-term Treasury loans (and the phasing out o f short-term Treasury loans), the absence o f bailouts inlocal government borrowing, and continued tight budget/Treasury controls. H. CONCLUSION 2.79 The intergovernmental fiscal framework requires further improvements to allow local governments to receive and allocate resources more efficiently towards services and infrastructure provision. The recommendations provided in each specific area in this chapter are aimed at this objective but achowledging the progress this framework has shown through the 2001-02 budget reform, which implies that the changes suggested here are aimed at improving this framework and not at changing it drastically. Moreover, the measures suggested embed clear principles o f efficiency (including through appropriate incentives), flexibility, local accountability, transparency, stability, and predictability, and thus they stand valid even in the context o f a territorial reform should such a reform take place in the future. However, it should also be said that a major territorial reform does not appear to be urgent in Ulaaine, particularly if the rayon and city level i s strengthened as the core local government, and if consolidationbelow the rayon level takes place through incentives rather than through imposition. 2.80 Suggested sequencing: Since the intergovernmental fiscal framework i s mostly established through the Budget Code, the ongoing revision to this law presents an excellent opportunity to include (during 2008) most o f the recommendations suggested that require legislative changes, including those related to the following: (i) expenditure responsibilities (that is, moving responsibilities up to the rayon level and granting rayons and oblasts stronger administrative powers); (ii) shared revenues; (iii) capital transfers; (iv) and the local government borrowing framework. Recommendations on own local taxes can legislated through the Tax Code under preparation, and recommendations on the equalization transfer (and detailed recommendations on capital transfers) can be regulated through amendments to the corresponding Cabinet o f Ministers Resolutions (all during 2008). Revisions and changes to sectoral regulations should start immediately (pilot testing, if needed, can be undertaking during 2008). Implementation could start in2009. 50 3. OVERCOMINGFISCAL, EFFICIENCY,AND EQUITY CHALLENGES INPUBLIC HEALTH SPENDING A. INTRODUCTION 3.1 Although several health indicators have been improving over the last five years (e.g., infant mortality, maternal mortality), others have remained stagnant or have deteriorated (e.g., mortality o f males aged 25-39, incidence of tuberculosis and HIV). Moreover, Ukraine continues to face challenges with regard to broader health system outputs and outcomes such as improving efficiency in service provision, reversing the decline in life expectancy, reducing the incidence o f catastrophic health expenditures, and reducing inequities inaccess to effective health care coverage. On balance, most o f the indicators for Ukraine under-performthose for the new EUmember countries6' (referred to inthis report as the EU-10). 3.2 Currently, Ukraine's national health system i s financed by general revenues from the budget. Public health spending i s moderate (at 3.7 percent of GDP in 2006) for the region, although it i s slightly above average compared internationally when normalized by GDP. Moreover, and despite constitutional provisions, out-of-pocket spending (Le., "voluntary/formal" and informal payments) inpublic facilities i s fairly high: rough estimates of formal and informal out-of-pocket expenditures suggest that they may be as high as the total public expenditures in the sector.69 Thus, total health expenditure i s at least in the range of 6-6.7 percent of GDP. But spending on health i s highly inefficient as argued inthis chapter, and i s plagued with sectoral regulation that constrains budget flexibility at the health facility and local government levels. 3.3 This chapter is organized as follows: Section B provides a brief description of the health care setting in Ukraine and points out the trends in selected key indicators. Section C examines public (and private-out-of-pocket) expenditure on health care. Section D examines and discusses key problematic issues facing Ukraine's public health care system from several perspectives, including financing, supply side inefficiencies, and equity (the last also in the context of some governance concerns). While the chapter focuses on the problems of the current system, it also highlights some key considerations regarding proposals for an insurance-based system in Section E. Section F provides a discussion and selected options for reformand recommendations. The New European Union member countries, hereinafter called the EU-10, are: Cyprus, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Malta, Poland, the Slovak Republic, and Slovenia. 69 These informal expenditures include payments to professional health staff, self-provision o f supplies and medicines and in some cases equipment, and the excess opportunity cost o f under-utilized real estate and other assets. 51 B. THEHEALTHSECTORINUKRAINE:STRUCTUREOFAND TRENDSINTHE MAININDICATORS The Structure of the Health Care System 3.4 The government is the absolute dominant player in the health care provision system in Ukraine. The vast majority o f preventive and curative health services are provided and financed by the state through general revenues and through all o f the different levels o f government. Private participation inhealthcare service delivery is negligible. There are some small health insurance plansystems provided by large SOEs. Inaddition, the ministries o f Defense, Transport (including the railways health program), and Internal Affairs have their own health care facilities that provide services to their employees and family members. And there are small private insurance programs that have agreements with the few existing private hospitals. These last programs cover a marginal fraction o f individuals, and their expenditures inthose schemes are very marginal inthe context o f Ukraine public health care system. 3.5 The health care system is managed by the Ministry of Health (Mow under a fairly centralizedstructure as regards administration, budgeting, and accountability. The central M o H i s responsible for setting national health policies and manages directly certain specialized health care institutions. The central M o H also coordinates the regional (oblast) level, which i s divided administratively into 24 oblasts, the Crimean Autonomous Republic (Crimea AR), and two special status city authorities (Kiev and Sevastopol). The 24 oblasts and Kiev and Sevastopol have a health administration (which is a direct branch o f the MoH); and Crimean AR has a special ministry." These regional departments, in turn, coordinate downwards with and through rayons and cities (the second tier o f local government) and villages (the third tier) under a fairly hierarchicalsystem (see Box 3.1). 3.6 This hierarchicalstructure, with a tight centralcontrol, is inplacedespitethe fact that most health facilities are owned by local governments at different levels, that most spending is done through local governments, and that, in theory, local governments have a certain level of expenditure autonomy. Since the budget reform o f 2001-02, around 692 local governments, including oblasts, and rayons and cities (basically, the first and second tiers o f local government), prepare and process their budgets directly with the Ministry o f Finance. To fulfill their delegated functions, all local governments are assured non-earmarked shared revenues through fixed portions o f the PIT, the land tax, and the single tax (under the Simplified Tax System) within their j u r i ~ d i c t i o n(see ~ ~ Chapter 2 for a detailed discussion o f the intergovernmental fiscal framework). In addition, local budgets are augmented or reducedby a formula-based gap-filling equalization block (non-earmarked) grant, which encompasses all o f the main delegated functions (i.e., education, health, social care services, culture, sports, and administration) for the estimation o f the expenditure needs o f local governments. The formula for the equalization grant has a horizontal gap-filling approach to equalization: that is, in each o f the local governments the amount o f the grant i s calculated as the difference between the estimate o f expenditure needs and the revenue capacity. Ifthe local government has greater expenditure needs than its forecasted revenue capacity, it will receive an equalization transfer; otherwise, the local government pays into the common (see Chapter 2). Local governments also have own revenue sources but they are marginal to the structure o f total revenues. 'OUnder the Constitution o f Ukraine, the country's administrative and territorial system is comprised of the Crimean AR, and 24 oblasts, rayons, cities, city districts, towns and villages. The cities o f Kiev and Sevastopol have a special status, defined by separate laws, practically equivalent to the status of an oblast. 71Inaddition, Crimea, Kyiv and Sevastopol have additional earmarked revenues. 72The pool is also supplemented by central government funding. 52 3.7 The health care regulatory functions are concentrated at the national level. The Ministryo f Health i s responsible for the accreditation o f public and private health facilities as well as the licensing o f health professionals and o f pharmaceutical manufacturers and distributors. It determines the range o f pharmaceutical products that can be purchased by state facilities (subject to the approval o f the Ministry o f Finance). ~~ ~ Box 3.1: Health Sector Structure and Administrative Accountabilities 53 Trends in Selected Health Indicators 3.8 At 67.1 years, life expectancy at birth in Ukraine in 2005 was below the average for ECA c~untries.'~ Ukraine's life expectancy was higher than that o f Russia but remained far below the averages for the EU and the high-income OECD countries (see Table 3.1).74 Life expectancy rates inUkraine are also low when compared with the EU-10 countries. Moreover, life expectancy i s much lower for males, comparatively: for example, the male-female differential in life expectancy in Ukraine i s almost double that observed in high-income OECD countries and male life expectancy in Ukraine i s far lower than the E C A and EU-10 averages. Taking a more global perspective, and inrelation to its per capita GDP, life expectancy inUkraine i s about average for females but far below average for males (see Annex 3.1). Table 3.1: Life Expectancy at Birthin 2005, Ukraine Compared Internationally CountryIRegion Life expectancy Life expectancy I Lifeexpectancy (total) (males) (females) Russian Federation 65.2 58.8 72.0 Ukraine 67.1 61.7 72.7 ECA 70.9 67.0 74.9 EU-10 74.6 70.6 78.8 EU 76.8 73.5 80.2 High-Income OECD 79.2 76.4 82.0 Note: Datafor Ukraineare for 2005; datafor comparators are for 2004. Sources: World Development Indicators; Ministry of Healthof Ukraine. 3.9 Life expectancy in Ukraine declined sharply early in the transition, and the performanceof this indicator hasbeensluggish since then. Infact, Ukraine is one o f the few countries inthe world that had higher life expectancy rates inthe 1960s than it has today (see Figure 3.1). Ukraine's early transition period experienced a rise in mortality rates, especially for adult males. As a result, life expectancy declined sharply and has yet to recover to the pre-transition level. This decline coincided with the steep decline inoutput, and although GDP has rebounded since 2000, life expectancy has not followed suit. 73 ECA countries: Uzbekistan, Belarus, Croatia, Kazakhstan, Azerbaijan, Lithuania, the CzechRepublic,Romania, Bulgaria, Slovenia, the Kyrgyz Republic,Armenia, Tajikistan,Estonia, SerbiaandMontenegro, Turkey, the RussianFederation, Latvia, the SlovakRepublic,Poland, BosniaandHerzegovina, Albania, Georgia, Ukraine, andHungary. 74 EUcountries: Ireland,Cyprus, Belgium, Spain, Luxembourg, Austria, Italy, the UnitedKingdom, Finland,Malta, Denmark, Portugal, Sweden, France,the Netherlands, Greece, Germany, Lithuania, the CzechRepublic, Romania, Bulgaria, Slovenia, Estonia, Latvia, the Slovak Republic, Poland, andHungary.High-incomeOECD countries:the United States, Ireland,Belgium, Iceland, Spain, Luxembourg,Japan, Austria, Switzerland,Canada,New Zealand, Italy, the UnitedKingdom, Finland,Norway, Denmark,Portugal, RepublicofKorea., Australia, Sweden, France,the Netherlands, Greece, andGermany. 54 Figure3.1: Life ExpectancyinUkraine: A Story of Poor Performance,1960-2005 -0 0 ln r 0 - m $0- -IC 6 f a X O a h - 2 ln (0 1960 1970 1980 Year 1990 2000 2004 1985 1990 4% 2000 2005 ources World Development Indicators; State Statistics Committee o f Ukraine; Bank staff calculations. 3.10 Life expectancy varies across oblasts in Ukraine. Life expectancy is highest in the oblasts of Temopil and in Kyiv City. By contrast, the oblasts of Dnipropetrovsk, Odesa, Zhytomyr, and Donetsk had comparatively lower levels of life expectancy - almost five years lower than that o f Temopil or of Kyiv City. Regional income does not seem to have any significant effect on these differences (see Figure 3.2). Life expectancy inrural areas, as expected, tends to be lower than inurban areas, andthis is likely to be correlated with higher percentages of the population living below the poverty line in rural areas." More important, the oblasts with relatively high life expectancies also tend to have a greater number o f registered diseases per capita, which suggests that the coverage of the health system (andor the knowledge and awareness o f the population) could be a factor in determininghealth outcomes across oblasts. Figure3.2: Life ExpectancyVaries SignificantlyAcross Oblasts I 71 I 250.000 70 -- --200.000 69 -- 100.000 65 --50.000 64 -- -a- Life expectancy -+-Numberof reaistereddiseases/100.000 I Note: Oblasts, the ARC and cities o f special status are ordered from higher to lower per capita regional (oblast) value added from left to right. Sources: MoF; State Statistics Committee of Ukraine; Bank staff calculations. 3.11 Young and middle-aged men faced the largest increases in mortality rates following the transition in Ukraine. The age-standardized mortality rate for men aged 35-44 almost doubled between 75See also the World Bank poverty update, 2007. 55 1990 and 1995 and has remained high since then: in 2004 it was 619.7 per 100,000, which was two to four times higher than the corresponding rate for other European countries. Most o f the deaths inthis age group are caused by "unnatural" or "external causes," which include accidents, poisonings or injuries, suicides, and alcohol-related diseases. Regular alcohol consumption i s frequent, especially among the younger population. Ukraine's per capita alcohol consumption i s among the highest in the Tobacco and alcohol were the two largest disease risk factors inthe country, with tobacco accounting for 20.2 percent and alcohol accounting for 18.3 percent o f the overall disease burden for males in2002.77 3.12 Non-communicable diseases are the dominant cause of the disease burden in Ukraine. Among these, cardiovascular diseases are responsible for the largest percentage o f the overall burden for both males and females. Unintentional injuries and neuropsychiatric conditions are the second most important in the burden o f disease among males and females, respectively (Annex 3.1 contains a list o f the top ten health conditions underlying the disease burden among males and females in Ukraine). But there is also evidencethat the prevalenceof infectiousdiseases -in particular tuberculosis(TB) and HIV/AIDS-has been increasing. The HIV incidence i s nearly twice that o f Western Europe and 14 times that in Central Europe. The growing incidence o f HIV/AIDS i s serious, and while the overall prevalence rate i s still moderate, the rate o f infection has been increasing by a third each year since 1994.78 3.13 Cardiovascular diseases and malignant neoplasms are also the dominant causes of death in Ukraine (see Table 3.2). Among cardiovascular diseases, ischemic heart disease and cerebrovascular disease were the dominant causes o f death across all ages. Table 3.2: LeadingCausesof DeathinUkraine,2005 Cause Deaths (`000s) Total 782 Cardiovascular diseases 488.8 Neoplasms 91.8 External causes of death 70 Digestivesystemdiseases 31.7 Pulmonary diseases 28 Infectious and some parasitic diseases 17.2 Note: Since 2005 the registrationofdeaths is classifiedaccordingto ICD-10 codes. Source: State Statistics Committee; Ukraine (http://WWW.ukrstat.gov.ua/ ). 3.14 The deterioration in health status among middle-aged groups may exert significant fiscal pressureson the long-term care systeminthe future. As a regional report o f the World Bank7'argues, while the actual increases in public health spending that are due specifically to an aging population are likely to be marginal, other factors, such as the health status o f the elderly, can interact with the demographic profile and produce serious fiscal pressures. In Ukraine, with the recent trend o f a high incidence o f different illnesses in male (and to a lesser extent female) adults, concerns are growing regarding the conditions o f this population cohort later intime. This cohort may become more vulnerable to more diseases when it reaches the elderly years and may become more costly in terms o f care. Moreover, higher fiscal pressures in long-term care seem, inany case, inescapable inthe case o f Ukraine, 76Closeto 15.4liters per annum while the global average is closeto 6.2 liters, see Popova, S. J. Rehm, J. Patra, andW. Zatonski (2007), "Comparing Alcohol ConsumptioninCentralandEastemEuropeto Other EuropeanCountries," Alcohol and Alcoholism. 77WHO (2006). ''WorldWorldBankreport 78 Bank (2006). See "From Redto Gray" (2007). 56 giventhe demographic trends. Inthe context of the Ukrainian budget, a large part of this spendingwould shiftto social care services for the elderly andthe disabled, whichare also handledbylocalgovernments. 3.15 Although Ukraine's infant, child, and maternalmortality rates have improved over recent years, they are still two to three times higher than those of the EU countries (see Table 3.3). Nevertheless, Ukraine's performance with regard to these indicators i s better than that of Russia and the average for ECA countries. Unlike life expectancy, the transition does not appear to have had a significant impact on these indicators (Annex 3.1 shows the trends in infant and child mortality rates for Ukraine along with comparator regions/countries). Table 3.3: Mortality Rates: Ukraine versus Comparators, 2005l' Maternal mortality CountryIRegion Infant mortality rate Childmortality rate (per 1,000 live rate (per 1,000) (per 100,000 live births RussianFederation 16.8 20.6 67.0 Ukraine 10.0 13.0 35.0 ECA 25.4 30.3 45.9 EU-10 6.1 7.2 22.3 EU 5.7 6.7 16.3 High-Income OECD 4.4 5.0 9.5 /1 Datafor comparatorsare for 2004. Sources: World DevelopmentIndicators;State Statistics Committee; MoH; Bank staffcalculations. C. HEALTHEXPENDITURESANDFINANCING 3.16 Currently, Ukraine has a national health system that is mainly financed by general revenuesfromthe budget.Public healthexpenditures have increasedmoderatelysince 2002 andreached 3.7 percent of GDP in 2006 (see Figure 3.3), but out-of-pocket spending, mainly in public facilities, i s high. Bank estimates based on household surveys and other surveys for formal and informal out-of- pocket expenditures in health suggest that they may have been as high 2.8 percent of GDP for 2005,80 which i s on the high end if compared regionally. Total health spending i s marginally increased (by 0.1- 0.2 percent of GDP) by the spending of sick funds (funded and managed by NGOs and charitable foundations) and by expenditures through private insurance funds managed by large SOEs. 81 Thus, total health expenditure i s at least up to the range of 6.0-6.7 percent of GDP, which takes Ukraine above the ECA average althoughit is still below the EUaverage. 8o These out-of-pocketexpenditures includepayments to professionalhealthstaff, self-provision of supplies and medicines, and, insome cases, equipment, andtransportationcosts. 8'Insurancesystemsprovidedby large SOEs andother governmentagencies, such as the railwayshealthplan, andother small privateinsurance systems, are rather marginalto the overallpicture. Most ofthese planshave agreements with the few existing privatehospitalsandwith somepublic institutions. 57 I Figure3.3: HealthExpenditureinUkraineand SelectedCountries,2002-06 Ukraine PublicHealthExpenditureas Percentof GDP,2002-06 I Public and Private Expenditurein Health: A RegionalComparison 9.00 -r-- I^ --., 8- 4.00 8 00 3.50-- 7 00 IPrivateloutof pocket 3.00-- 6 00 5 00 2.50-- 4 00 2.00-- 3 00 0 Public Health 1S O-- 2 00 Expenditures 100 1.00-- 0 00 0.50-- 0.00 7 2002 2003 2004 2005 2006 U 3.17 From a broader cross-country perspective,Ukraine's public and total health spending as a share of GDP is just above average for its income level (see Figure 3.4). As countries grow richer, the share o f total health expenditure in GDP, as well as the share of public health expenditure, tends to rise. Hence it i s important to situate Ukraine in relation to the norm for its income level when looking at overall health expenditure patterns. Figure3.4: UkraineRanks CloseinAverage inHealthSpendingas a percentof GDPWhen Normalizedby Levelof Per CapitalIncome'' HEALTH EXPENDITURE VS INCOME i) c> b 100 250 1000 2500 '3DP per c-plt- 1000(P5000 I/: The figure is a log scale. Data for Ukraine are for 2006; data for other countries are for 2004. Sources: World Development Indicators; WHO; SSC o f Ukraine; Bank staff calculations. Public Spendingon Health 3.18 Wages and salaries constitute the largest share of public health expenditures and drive the overall spending trends in this sector. Figure 3.5 shows public spending allocations by economic classification in Ukraine. A large proportion-close to 58 percent o f overall government health expenditure-was allocated to wages and salaries. Recurrent spending in the procurement o f items, materials, and services was the second largest share. Capital expenditures, through new construction, 58 capital goods acquisition, and renovations, were below 8 percent for the whole health budget. Moreover, capital expenditures at the local levelrepresent less than 2.5 percent of health spending at the local level. Figure3.5: GrowingPressures fromRecurrentPublic Spendinginthe HealthSector, 2000-05 0Capital xpenditures (construction, acquisition of assets. renovation. other ) mi Other Q3 Recurrent spending in materials and services. 0utiilties I andsalaries Wages 2000 2002 2005 b'oources: State Treasury o f Ukraine; Bank staff calculations. 3.19 Local governments' health spending accounts for the bulk of government spending in this sector. The local government share in public health spending has been growing steadily since 2004, reaching 79.3 percent in 2006. This increase has been driven mainly by the increasing wages o f health workers, which are mostly dependent on local budgets. The expenditure shares by level o f government have remained relatively constant in recent years, except for the share of rayons, which has increased slightly. Rayons have a relatively higher percent o f their health spending in wages compared to other levels of local government and the central government. Figure3.6: Local GovernmentsAre the Main PlayersinPublic HealthSpending, 2001-06 0Central f 3 Village Sources: State Treasury o f Ukraine; Bank staff calculations. 3.20 Public health expenditure varies across local governments and i s positively correlated with average oblast and rayon income levels. Figure 3.7 illustrates this situation by ordering oblasts and rayons by income and showing oblasts and rayons for the 2005 budget allocation in per capita terms. While there i s a tendency for the richer oblasts to spend more on health, there i s considerable variation around this trend. 59 Figure 3.7: Some InequalityExistsinPublic HealthSpendingat the Local Level(2001 and 2005) I a 350 .-I___ .-,300 3 By rayonsand Cities e By Oblasts .-' 400 A I I -c2005 **" I =- 3 200 C 300 m -t'It -.-n= 150 U a 100 '4 50 I I 0 -y=150 n P 0 100 50 0 I rlotes: Oblasts, ARC andcities of special status are orderedfrom higher to lowerper capitaregional(Oblast) value added (left to right). The city o f Kyiv was taken out ofthe sampleas it is ahigh up outlier inper-capitaspending. Rayonsandcities are ordered from higher income to lower income (from left to right). This study usedPIT collections per-capitaas proxy for income levelbyrayon. Sources: MoF; State Statistics Committeeof Ukraine; Bank staffcalculations Private/Out-of-PocketFinancing 3.21 As stated earlier, the government finances close to 60 percent of total health spending, and the remainder comes mainly through private/out-of-pocket (OOP) household payments. This estimate i s based on the analysis of the National Health Accounts (NHA). Private sources make up the remainder of public health spending, and this i s almost entirely in the form of out-of-pocket (OOP) household contributions. Our estimates may be on the conservative side, as household contributions in Ukraine are likely to be underestimated because of the difficulty in accurately measuring OOP payments.82Some estimates consider OOP spending to be as high as total public pend ding.'^ External sources comprise less than 1 percent o f total health financing (see Table 3.4). Most inpatient care i s fundedby the government while the majority of outpatient care is financedby OOP. This occurs because a large share of outpatient care i s composed of pharmaceuticals. The relative lack of government spending on drugs translates into high amounts paid by households, sometimes resulting in catastrophic health paymentsnE4Private sources, again mostly households, also make up the majority o f the financing for ancillary services, while government funds account for a quarter of the total. 82 While these estimatesare based ondatatwo-threeyears lagged, the share of OOP in total healthspendingi s not likely to be smaller at present, ifsomething it mayhaveincreasedintandemwith higher disposableincomeofthe populationowing to the wage growth. 83 See Joint Report (WB, EC, SIDA, Ministry of HealthofUkraineandMinistry of Economy) "Key strategies for further development of the healthcare sector inUkraine" (2007), editedby LekhanandRudiy. 84 Catastrophichealthexpenditure is definedhereas spendingmorethan 40 percent ofnon-subsistenceexpenditure on OOP paymentsaccordingto the World HealthOrganization's classification(see also WHO HealthFinancingPolicy discussion paper) andXu,K., Aguilar,A., Carrin,G., and Evans, D.(2005). 60 Table 3.4: Percentage of Total Health Spending by Program and Financing Sources, 2004 Program Government Private External Inpatient 98.0% 2.0% 0.0% Outpatient 28.4% 71.0% 0.6% Pharmaceuticals 0.1% 99.9% 0.0% TB 74.4% 1.3% 24.4% ARVS 25.4% 0.0% 74.6% OB/GYN 72.9% 19.2% 7.9% Other outpatient 84.7% 14.7% 0.6% Rehabilitative care 54.7% 45.3% 0.0% Long-term care 100.0% 0.0% 0.0% Ancillary 24.7% 75.3% 0.0% Preventionand public health 90.7% 3.4% 5.9% Maternal and child 11.8% 0.0% 88.2% Communicablediseaseprevention 72.2% 0.0% 27.8% Non-communicable disease prevention 0.0% 0.0% 100.0% Other 94.9% 3.9% 1.2% Health administration 97.1% 0.0% 2.9% Other 97.4% 1.5% 1.1% TOTAL 58.2% 41.2% 0.7% Sources: National Health Accounts; Bank staff calculations. 3.22 Estimates for OOP in Ukraine vary considerably. This is mainly because different sources cover different components o f OOP spending. The full value o f OOP expenditures is composed o f official user charges, drugs purchased outside o f health clinics, under-the-table payments, and costs o f transportation, food, and lodging.85 Table 3.5 illustrates the range o f estimates for OOP spending in Ukraine and the associated components included in each estimate. Government estimates o f OOP payments are considerably underestimated because they include only the official user charges at public facilities. The NHA estimates are much higher, since they include the cost o f drugs purchased outside o f clinics. InUkraine, drugs are generally purchased inpharmacies and not inclinics or hospitals. Household surveys also incorporate under-the-table payments (gratuitieshnforrnal payments) which are prevalent in Ukraine, probably owing to the low remuneration o f health workers. Estimates from the Health and Education Module Sample Survey 2004, which also include health-related transport costs, indicate overall OOP expenditure to be 9.2 billion hryvnas (Table 3.5). The World Health Survey (WHS) 2003 estimates are even higher. The W H S contains more detailed modules on OOP costs (including an "other" category which could potentially cover food and lodging) than the World Bank Survey and this may explain why the W H S estimates are much higher. Because o f the comprehensiveness o f the two surveys, estimates derived from them are likely to be more reflective o f the full burden o f OOP expenditures inUkraine than estimates available from other sources. 85An indirect but additional cost is the opportunity cost of seeking heath care, which is income forgone. 61 Table 3.5: ComparingDifferentEstimatesof Out-of-pocketExpendituresinUkraine - Source 2003 2004 Components included inOOP estimation (UAH) (UAH) - A Government 236,612,088 321,448,481 User charges NationalHealthAccounts 7,174,537,772 8,797,272,566 User charges; Purchase of drugs Healthand Education Module Sample Survey -- User charges; Purchase of drugs; Informal 2004 9'214y414'549 payments; Transport World Health Survey -- User charges; Purchase of drugs; Informal 2003 13,356,27 1,775 payments; Transport; Food; Lodging Sources: Ministry ofFinance; State Treasury; NHA; World Bank HealthandEducationSurvey 2004; World Health Survey 2003; Bank staffestimations. 3.23 The size and the likelihood of direct OOP financing rise considerably with the level of specialization. The medical facilities o f research and academic institutions are markedly more expensive than most other hospitals (mean UAH 241 compared to UAH50 average, pCO.01). Official payments at oblast hospitals are higher than at rayon and city hospitals. Furthermore, people are more likely to pay at oblast hospitals than at other facilities (55.5 percent versus 40.6 percent average across all facilities). 3.24 The extent of OOP paymentsvaries significantlyacross oblasts (see Figure 3.8). Ternopil and Vinnitska had the highest incidence o f OOP spending as a proportion o f income. The largest share o f OOP expenditure inevery oblast was drugs, 95 percent o f which were purchased at pharmacies and not at outpatient or inpatient facilities. In general, patients must purchase their own drugs and other consumables at pharmacies and then bringthem to hospitals.86 Moreover, only a very small proportion o f total OOP spending is composed o f preventive care. Figure3.8: Out-of-Pocket Expendituresby Category,by Oblast E I Note: The drugs categorydoes not includedrugspurchasedat outpatient or inpatientvisits. Source: Authors' estimatesfrom World Bank HealthandEducationSurvey 2004 data. 86This may explainwhy the average OOP expenditure for inpatient care i s relatively low comparedto outpatient care anddrugs. OOP spending on drugs reflects purchasesoutside ofthe hospital, even thoughthe consumptionmay occur for an episodeof inpatientcare. This is also most likely additionally explainedby different recallperiodsof drugs, outpatientcare, and inpatient care. 62 Other PrivateFinancingSources 3.25 Other private financing sources are minimal in Ukraine. Voluntary health insurance and sickness funds constitute a very small share of total health expenditures. Less than 2 percent of the total population i s covered by voluntary health insurance and care is still provided mostly at public health facilities, since the private sector i s limited in size. Sickness funds are public organizations voluntary in membership that primarily fund the cost of pharmaceuticals. Contribution rates are based on salary or pension income. However, similar to voluntary health insurance, both the membership in and the total financing contribution o f sickness funds are extremely limited. The Donors' Role in FinancingHealthPrograms 3.26 The role of external health financing in Ukraine is focused primarily on HIV/AIDS, tuberculosis, and maternaland child health, and it has been increasing. The Global Fund to Fight AIDS, TB, and Malaria (GFATM), the World Bank (World Bank investments are part of public spending), and the United States Agency for International Development (USAID) are among the largest donors financing both HIV/AIDS and tuberculosis in Ukraine. The Swedish International Development Cooperation Agency (SIDA) also finances maternal and child health. The donor share of total antiretroviral (ARV) financing has increased substantially over the last four years. 3.27 Government own spending on these programs seems to be shrinking as donor investment increases,reflectingbudget fungibilityto aid from external sources.That is, the government maynot spend as much as it otherwise would on HIV/AIDS treatment since donors are financing a large and increasing portion of the spending in specific programs (see Figure 3.9). Although donor resources are marginal in the overall public health budget picture, excessive dependence on donor financing for HIV/AIDS raises issues of sustainability and could be potentially detrimental to other government funded programs ifthe predictable donor support were to stop. However, in2004 the increase inpublic spending that would be neededif the government were to finance 100 percent of ARV treatment would amount to just 0.18 percent of government healthexpenditure (again a very small portionintotal public spendingin health). This suggests two things: (i) a minimal effort would be needed on the part of the government that to replace all external aid inthis area, and (ii) that current ARV treatment spending i s very small and the coverage i s very limited. This situation has also implication for how donors pledge their support, given the revealed behavior of public spendinginthese areas inthe presence o f donor funding. 3.28 There are indeed difficulties for channelingdonor funding to local governments, but donor created parallelsystems of financing underminethe intergovernmentalfiscal framework and raise sustainability concerns.The existing difficulties posedby the lack of flexibility inthe structure of local government budget allocations make donor financing for these entities difficult. This situation has led some donors (including through NGOs) to create funding mechanisms for certain local governments completely off the budget and to some extent on ad-hoc basis in terms of recipients and timing. These ring-fencing mechanisms undermine the process o f strengtheningintergovernmental fiscal relations and budget transparency, and raise serious concern about sustainability o f financing for certain health care programs at the local level. 63 Figure 3.9: Sources for HIV FundinginUkraine,2003-05 60 50 40 30 20 10 0 Private State budget Local budgets InternationalNGOs providing household donors services to spending households wrces: HBS; MoF; State Treasury; international donors; Bank staff estimations. D. ISSUES AND CONSTRAINTSFACINGTHE HEALTHSECTOR Budgeting,Administrative,andAccountabilityIssues 3.29 Even though most public spending takes place through local governments they have very little discretionary control over expenditure allocations. Despite the positive (though incomplete) reforms o f 2001-02 that changed the manner o f allocating revenues and transfers to local governments, the way in which local health facilities, and then local governments, form their budgets has changed little and continues to be dictatedby strict input-based norms established by the Ministryof Health. 3.30 Healthnorms issuedby the M o H generate severe rigiditiesinthe delivery of health services and inhibit the appropriate administration of health facilities. These norms detail the number of doctors, nurses, and other staff at each local administrative level based on hospital beds and other network parameters, as opposed to being demand based. Examples o f such norms include the requirement that there be one therapist per 25 beds in district hospitals, one obstetrician-gynecologist per 20 beds inrayon hospitals, and one cleaner per 500 square meters. Given the oversize in term of number o f beds (compared to other countries in the region as discussed in detail in the next section), this situation forces overstaffing into facilities without giving administrators and local governments the possibility of adjusting this supply to demand or need in eachjurisdiction. Table 3.6 lists the number o f beds required to staff one doctor by department. 64 Table 3.6: NormsDetailing Number of Beds to one doctor, by department Department Doctor Numberofbeds Obstetrics Obstetrician-gynecologist 15 Gynecology Obstetrician-gynecologist 20 Infectious disease ward Infectious disease doctor 25 Infectious disease ward for Infectious disease doctor for children 20 children Infectious isolation ward Infectious disease doctor 20 Infectious isolation ward for Infectious disease doctor for children 15 children Dermatology (for adults and Dermatologist-venerologist(including for 30 children) children) Neurology (for adults and Neurologist 20 children) Maternal andNewborn Pediatriciadneonatologist for healthy 25 newborns Pediatriciadneonatologist for probationary 10 obstetric ward Pediatriciadneonatologist for premature newborns 10 Source: MoH Order # 33. 3.31 These regulations,inturn, bringabout rigiditiesin budget management,with most of such budgets being consumed in recurrent spending as opposed to increasing the share of quality- enhancing expenditures. Indeed, while wages and recurrent procurement o f materials and services consume most o f the budget, other needed supplies such as drugs must be financed privately out o f pocket by patients. This situation leaves little to invest in the training and re-training o f personnel critical to improving the quality o f care. Moreover, the meager amount invested inmodem medical equipment and the upgrading o f certain facilities i s a cause for concern, as this is indicative o f the slow replacement o f medical equipment, which i s also likely to compromise quality. 3.32 The accreditation of health facilities has been less rigorous in its regulation. The system o f government accreditation o f health facilities based on set standards was established in 1997 and amended in2001. Currently, the 27 regions each have accreditation commissions that report to the main Ministry o f Health Accreditation Commission for approval decisions. However, accreditation decisions are not used to improve service delivery within facilities; facilities that did not meet accreditation requirements were not reorganized so as to receive accreditation, and the performance o f those facilities receiving the top grades was not enc~uraged.~' 3.33 There is a constitutional prohibition to closing health care facilities. Article 49 o f the Constitutionprohibits any closure o f heath care facilities at any level o f government; thus, demand issues are not thoroughly discussed, and the needed restructuring i s an uphill battle. Consequently, local governments are bound to financially maintain an oversized and largely underutilized network o f health facilities. At the local level, however, particularly inrayons and cities, there i s considerable buy-infor the restructuringhationalization o f the network process (as evidenced by our field visits), as local governments desire more flexibility in allocating resources, even if this i s undertaken within the same sectoral fiscal envelope. *'Joint Reporton the "Key Strategies for FurtherDevelopmentofthe HealthCare Sector inUkraine(2007), Lekhanand Rudiy, eds. 65 3.34 Local governments face severe constraints in firing personnel, which adds rigidities to resource allocation. Although, in theory, local governments have the ability to hire and fire personnel, the way inwhich administrative and accountability issues are set up (as shown inBox 3.1) i s a constraint to a more fluid environment in which restructuring could be feasible. To fire an employee, regardless o f the cause, local governments must first follow sectoral procedures within the hierarchy o f the Ministry o f Health at different levels o f local government, and must then wait for several months at each step o f a cumbersome process. Health Delivery and Supply Side Issues 3.35 In general, Ukraine has a greater number of health system inputs than do EU-10 countries and other EU countries. Partly as a legacy o f the Soviet era, Ukraine had a large number o f hospitals: namely, 5.6 per 100,000 people in 2005. This i s in contrast to averages o f 2.6 per 100,000 for EU-10 countries and o f 3.2 per 100,000 for other EUmembers, Inaddition, in Ukraine there were 868 hospital beds per 100,000 people in 2005, compared to 644 for EU-10 countries and 571 for other EU counties. Regarding health personnel, there were more doctors inUkraine (302 per 100,000 in 2005) compared to EU-10 countries (261). And Ukraine also had more nurses per 100,000 (782) than did either EU-10 countries (548) or other EUmembers (750). 3.36 The average length of stay (ALOS) in Ukraine was about 15 days, higher than the CIS average and much higher than the EU average of 9 days. During the 1990s one o f the factors for closing a limitednumber o f facilities was occupancy rates, which created a strong incentive for extending patients' stays in hospital and other health care facilities. During the last few years, while closure o f facilities i s no longer allowed, there have been some budgetary incentives for extending ALOS inhealth facilities (for example, the allocation o f resources for materials which can be short-changed incase o f low utilization rates). Because o f this situation utilization rates are to some extent inflated and in many cases unreliable. 3.37 The analysis of variations across oblasts provides evidence that larger ALOS may be compensating for lower utilization. Many factors may contribute to determining the ALOS in a given region. To assess whether there are any unexplained variations in ALOS by region, a probit regression was estimated for the probability o f an individual's having a length o f stay longer than the Ukraine average o f 15.2 days, using the 2004 World Bank Survey. Controlling for age, sex, body mass index, marital status, income, education level, inpatient facility type, and geographical location o f the facility, there remains a statistically significant difference between the Prechornomorsk and Polisya economic regions.'' People in the former are 9.5 percent more likely to have a longer length o f stay than the Ukrainian average compared to those in the Polisya region (see Figure 3.10 and Annex 3.2).*' Inpatient utilization i s significantly lower in the Prechornomorsk region than in the Polisya region, which would support the hypothesis that a higher ALOS may be compensating for lower utilization rates in order to maintain budgetary allocations. 88The Prechomomorsk economic region comprises the following oblasts: Odeska, Mykolaivska, Khersonska, the Autonomous Republic o f Crimea, and the City o f Sevastopol. The Polisya economic region comprises the following oblasts: Volynska, Zhytomyrska, Rivnenska, and Chemihivska. The classification o f oblasts into economic regions is based on the classification used inthe Statinformconsulting Information andAnalytic Agency. 89The full regression results and average partial effects for the regional dummyvariables can be found inAnnex 3.2. 66 Figure 3.10: Average Length of Stay by Oblast, 2004 `.I Mykolaevskd * t 4 Lu#ernopiiska anska i.) Zaporlzhska .I Odeska Vlnnitska I .I Klev city Sebastopoi city 0 - t I ZhytomirskaRivnenska t Khersonska 0 - 0 - Kharklvskai Poitavska t Zakarpatska i.) , _ . 0 Kiev ARSurnska Crimea I0 Donetska 4 0 . __ ._ Lvivska CherkaskaCherniglvska - 0 - - 0 - r - - t - ---- Dniepropetrovska t ivano-Franklv ska Khrnelnitska , * 0 Kirovo radska 0 1 Cherniveyska 0 Voly nska Sources: HBS Health and Education Module Sample Survey; Bank staff calculations. 3.38 The Ukrainian health delivery system contains a costly and counterproductive drive towards specialization. This occurs in all spheres: doctors pursuingcareers (and monetary incentives), patients seeking care, and even the structure o f public facilities and public spending in the sector, as i s further discussed below: 0 The prominence of specialized medical care is seen in the overabundance of specialists comparedto physiciansworking at the primary care level. Primary care physicians comprise only about a quarter o f all active doctors, while general practitioners make up less than 2 percent o f all physician^.^^ This professional drive towards specialization i s caused in part by the low salaries o f non-specialists. Nevertheless, even specialist salaries are less than half the average salary in Ukraine. L o w wages generate incentives to accept gratuities and informal payments. Such payments are over two-and-a-half times the official user charges paid by patients for inpatient and outpatient visits, on a~erage.~' 0 There is often no clear distinctionbetweenprimary and secondary levelhealth care.Patients can visit specialists directly without a referral and generally do so for primary health care needs. There i s a lack of established and accountable physician gate-keeping. According to a recent survey, patients first visit primary health care physicians in about 40 percent o f cases and first go directly to specialists inthe remainder o f cases.92 Moreover, more than half o f all direct specialist visits are to inappropriate providers. Accordingly, almost two-thirds o f all visits result in improper care for the patient's condition. It i s worth noting that the adoption in 2000 o f both "The Concept o f Health Care System Development inUkraine" and "Decree o f the Government on Comprehensive Measures for Integrating Family Medicine into the Health Care System" was an important step forward in remodeling primary care on the principles o f family medicine. However, there i s still much room for better integration with higher levels o f care and for the strengthening o f the physician gate-keeping function. See Joint Report on the "Key Strategies for Further Development o fthe Health Care Sector inUkraine" (2007), Lekhan and Rudy, eds. 9iCalculations based on the 2004 World Bank Health and Education Survey. 92Joint Report (2007). 67 0 Most public health spending is concentratedin hospitals and specialized health facilities as opposedto primary health care. Close to 70 percent of the spending goes to hospitals. Around 14 percent of this spending goes to polyclinics (including specialized polyclinics) and other first aid facilities (see Figure 3.1l), only one-fifth of this 14 percent goes to first aid and obstetric but centers. Figure3.11: The HealthSysteminUkraineRemainsHospital Centered, 2002-06 100% 80% 80% 70% -- pandemic institutions and activities BO% -- activities 50% -- 40% -- first-aid, emergency 30% -- 20% -- 10% -- 0% r 2002 2003 2004 2005 2006 ources: State Treasury o f Ukraine; Bank staff calculations. GovernanceIssues 3.39 Informal unofficialpayments are common in Ukraine.In this context, and highlighting the acuteness o f this situation, a recent survey report fiom the EBRD-World Bank (2007) points out that Ukraine has one of the worst profiles regardinginformal and unofficial payments inthe Eastern European region (see Figure 3.12). Figure 3.12: Informal PaymentsinUkraineversus SelectedEuropean Groups of Countries - Receiving medical treatment in the public health system Percent of respondents that think unofficial payments are needed 20 18 1 :; 16 I Sometimes 3 " 0Usually 6 E9Always 4 2 0 CiS-low inoome CIS-middle inoome South-Eastern EU members + Ukraine Europe Croatia 3urce: EBRD-World Bank Life inTransition Survey (LITS) 2007 68 3.40 care utilization. For example, basedon survey results, Balanova (2004)9 reports that 34 percent of those There is evidencethat the increasingincidenceofinformalpa9mentsis a deterrent to health individuals reporting illnesses who did not seek care gave their reason as the inability to afford the necessarypayments. Another reason for not seeking care was "not being sick enough." 3.41 The weak accountability of health facility management creates opportunities for extracting high out-of-pocketpayments (OOP) from patients. A minimumlevel of health care is guaranteed to be free, and hospitalizations with areferralor due to an emergency are also free. Outpatient and inpatient care without referrals i s subject to official user fees or charitable contributions. However, OOP householdspending i s much higher than officially reported, as was previously discussed. Often, doctors must make payments to management in order to secure their positions and purchase equipment and supplies. This situation encourages doctors to levy informal payments on patients and, as a result, excessive OOP spending i s frequent. There i s little oversight of these practices. EquityIssues 3.42 Providers extract high OOP payments partly to compensate for low and highly tied government healthspending. Figure 3.13 suggeststhat individuals inthe poorest quintile of income (in each oblast) may spend a higher share of their income in OOP if they live in an oblast with lower per capita public health spending. This may imply that individuals in these oblasts (which are also poorer oblast in terms of overall income) have to cope through OOP with the shortfalls generated by the intergovernmental finance system. Figure3.13: OOP of PoorestIncomeQuintileversus PublicHealthSpendingby Oblast, 2005 olvano-Frankivsk oMykolayiv OVinnytisa oKharkiv Chernivtsi --a_. OKhrnelnytsky Kirovohrad - Tern$-- __ AR Crirnea Zakarpgtia OLViv Luhansk oPoltava_Zaporizhaia -R%me- &esPe$Xrny r oDniP6btrov sk`- ~Cherkasy '--- -.-_ -.__ oKyiv city Kyiv 0oDonetsk ochernihiv ---...._--...-_ 2 200 250 300 350 400 2005 Oblast public health expenditure per capita (UAH) ources: StateTreasuryof Ukraine; HBS HealthandEducationModule Sample Survey; Bank staff calculations. 3.43 The poorest quintile is the least likely to use outpatient, inpatient, and preventive services. The poor have fewer outpatient visits than the higher income groups, as shown inFigure 3.14.94Inpatient services are usedleast by the poorest and the richest quintiles. For the poorer quintiles, financial barriers 93For detaileddiscussion see Balabanova, D.,M.McKee, J. Pomerleau,R. Rose, andC. Haerpfer(2004), "Health Service Utilization inthe Former SovietUnion:Evidencefrom Eight Countries," Health ServicesResearch, 39(6): 1927-1949. 94The bandsaroundthe meansmeasure a 95 percent confidenceintervalandshow that the differencebetweenthe poorest quintile andthe others is statisticallysignificant. 69 to access are more likely, whereas the richer quintiles may have lower inpatient utilization rates because they are healthier. The decision to seek preventive care also demonstrates a pro-rich bias. The poorest quintile i s also least likely to report needingdrugs and least able to purchase the neededmedication. Figure3.14: Outpatient Utilizationby IncomeQuintile 1E Bottom 20 O/O I I 2 t I 3 I 4 !J I 4 I I Top 20% I I ource: Bankstaff calculations based on HBS Health and EducationModule Sample Survey. 3.44 The poorest tend to use significantly more rayon hospital than village hospitals. The proportion o foutpatient utilization by income quintile fluctuates across facilities, with the richest quintile constituting the largest share at both public and private polyclinics as well as specialized facilities. Individuals in the poorest quintile never make up the majority in any facility and are often the smallest share. The poorest more fkequently use rayon hospital and public polyclinics. Figure 3.15 depicts outpatient visits by facility type and economic status. Most outpatient visits are to public polyclinics and rayon-level hospital^.'^ Figure3.15: Outpatient Visits by Facility Type and IncomeQuintile - Private polyclinic Oblast hospital -- Other = 0 --- 500 Sum of outpatient visits 1.000 1,500 Bottom Quintile Quintile 2 Quintile 3 Quintile4 Top Quintile ource: Bank staff calculations based on HBS Health and Education Module Sample Survey. 95The length of the horizontal bars corresponds to the total number o f visits at the respective facility in the previous month. 70 3.45 Public health expenditure appears to favor higher income groups in Ukraine. Seventy percent of total public health expenditure accrues to hospitals and sanatoriums (Figure 3.1l), and the poorest use hospitals for outpatient care less than do the other quintiles. The inpatient utilization rates of rayon and oblast hospitals for the poorest are similar to those for other quintiles. The poorest also use these facilities more than other types of facilities, while the rich use sanatoriums more frequently than the poor. 3.46 Many households incur catastrophic health expenditures because of high OOP spending. Compared to some other ECA countries, Ukraine has the second highest proportion of households- nearly one in five- facing catastrophic spending (Figure 3.16).96 Almost 10 percent of all households exhibit catastrophic spending levels that are due to drug expenditures alone. Furthemore, catastrophic OOP spending i s heavily concentrated in the bottom income quintiles (Figure 3.17). If the government were to finance the cost o f protecting all households from falling into catastrophic spending, an increase of approximately 27 percent of current government health spending would be required." If the government were to protect any household from spending more than 20 percent o f its non-subsistence expenditure on OOP healthpayments, an increaseinspending o f around 52 percent would be needed." Figure3.16: Percent of HouseholdsSpendingMore than40 Percentof Non-SubsistenceExpenditureon OOP, Ukraineand Selected Countries 35.0% 30.0% - 25.0% - 20.0% - 15.0% - io.07- - 5.0% - .- Croetle Slovakle Hungery Latvl- Eetonla Ukraine Ru=.ela Source: Basedon World Health Surveyresults, 2003. 96 Catastrophic spending is calculatedusingthe methodologyproposedbyXu (2005) andis definedas OOP spending greater than 40 percent ofnon-subsistencespending. For the purposeof cross-country comparisons, householdsubsistencespendingis definedas the equivalencescaledhousehold size times arelativepovertyline based on the food consumptionof the household with the medianrood expenaimre snare 01total expenaimre. .. . .. - . " . "Theseestimates,though,assumethatthesubsidyisperfectlytargetedanddoesnotfactorinanydemand-sideresponsethat would inducepeopleto use more services. Moreover, at this stage, such a subsidy would only be channelingmore money to the inherent problems in the health system, given the current institutional incentive structure. By tying up a large portion of the '*Theseestimateswere budgetwith salary payments, the norms indirectly support the largeinefficienciesand inequitiesof Ukraine'shealthsystem. calculatedusingthe World HealthSurvey2003 by the following: h - OOP * O h hhsize x N GHE where oophis the monthly out-of-pocketspendingby householdh, oop40his the level of out-of-pocket spendingthat would result inhouseholdh spending40 percentofnon-subsistenceincome on OOP, H+ is thenumber ofhouseholdsinthe survey spending morethan 40 percentofnon-subsistenceincomeon OOP, hhsize is the averagehouseholdsize inthe survey, Nis the total number of households inthe survey, P is the totalpopulationofUkraine, andGHE is annualgovernmenthealthexpenditure. All spending values are calculatedinUAH. The same equationis usedfor calculatingthe increasedamount to cover OOP costs exceeding20 percentof non-subsistenceexpenditure. Inthis case, H+ is now the number ofhouseholdsspendingmorethan 20 percentof non-subsistencespendingon OOP costs. These estimatescover only the cost of giving financial protectionto each householdanddo not necessarilymeanthat the OOP costs are for services for which the risksare insurable. However,ifan importantaspect of OOP costs on equity is the effect on ahousehold's disposableincome, the estimatesare usefulfor representingthe increaseingovernment spendingneededto providesimple cashtransfers to protecthouseholdincome, irrespectiveof what the OOP costs are usedto finance. 71 Figure 3.17: Percentof Households inUkraine with Catastrophic OOP, by Quintile 25.0% ._ I -Average 20.0% -- 15.0% -- 10.0% -- 5.0% -- Poorest Income quintile Richest Source: Basedon World HealthSurveyresults, 2003. 3.47 Moreover, OOP payments force many people into poverty and increase the size of the poverty gap.99Simulations o f the effect o f pre-OOP and post OOP payments on household incomes based on the 2004 household survey (see Annex 3.3) show that the impoverishing effects o f health payments are especially true for inpatient care and drugs purchased outside o f health facilities. In the simulations, inpatient care often consumes more than the monthly income o f households. Richer households may have savings to cope with these shocks but poorer ones may face more difficulty ifthey lack such resources. Purchasing drugs outside o f health facilities also forces many people into poverty across the income distribution, although the size of the shocks i s smaller and concentrated less among the better off. 3.48 Compared to both other ECA and OECD countries, OOP payments are more regressive in Ukraine. Figure 3.18 illustrates the redistributive effect (RE) o f OOP payments for Ukraine and some o f the countries depicted inFigure 3.16, usingdata from the 2003 World Health Survey (see Annex 3.4 for details on the methodology o f the regressive effect calculations). Expenditure i s measured instead of income, and countries are listed from left to right in order o f the most pro-poor to the most pro-rich RE o f OOP payments. This illustrates the pro-rich effect o f OOP financing in Ukraine. The increase in income inequality generatedby OOP payments inUkraine i s also greater than the estimates calculated for number o f OECD countries.'00 99The povertygap measureshow far on averagethe poor are below the povertyline looSee van Doorslaer, E. et al. (1999). "The redistributiveeffect of healthcare finance intwelve OECD counhies."Journal ofHealth Economics, 18: 291-313. 72 Figure3.18: Redistributive Effect of OOP Financing: A Regional Comparison 0.020 0.015 A 0.010 - - poor 0.005 0.000 -0.005 -0.010 -0.015 vI sia Ukraine -0.020 burce: World Health Survey results, 2003. 3.49 But there is also a horizontal dimension to the inequalities created by OOP in Ukraine. The analysis and results based on the World Health Survey'" also point to different treatment among households of similar income levels. While some of this may be due to the randomnature o f illness, this variation signals geographical differences inthe rates for formal and informal payments for healthcare. E. HEALTHINSURANCEREFORMSUNDERDISCUSSION 3.50 Social health insurance (SHI) reformhas been under discussion inUkraine over the last 10years and several proposals, taking different shapes and forms, have been put forward. While this report does not focus on analyzing the many proposals nor it examines this issue indetail, it takes a broad look at the latest proposal and lays out some key considerations, emphasizing useful lessons that could apply to any reforminthis direction. Backgroundto the MainProposalunderDiscussion 3.5 1 The most recent health financing reformproposals envisage a four tierblended financing system, as follows: 0 First tier: State Program for Health Security. The program i s to be relatively comprehensive, covering the primary, secondary and tertiary levels of care for the entire population at the "health supporting level". The services are to be provided by current public sector providers. Financing for the program i s to come from current state and local government budgets. This program maintains the current supply side funding coming from the central government to local lo'The analysis used the methodologyproposedin Wagstaff, A. and E. Van Doorslaer "Progressivity, horizontal equity and reranking in health carefinance: a decompositionanalysisfor TheNetherlands".(1997). Journal of Health Economics, 1997. 16: p.499-516Wagstaff and van Doorslaer (1997)--see Annex 3.4 73 governments (equalization formulas, direct capital investment and other services) and from local governments to facilities. Second tier: Compulsory Health Insurance. This complementary insurance package will be financed by employers, local governments, and pension funds, and will be managed by a State Health Insurance Organization and, eventually, by other private health insurance agencies. It will contract public facilities and primary level doctors. Third tier: Supplementary Voluntary HealthInsurance. This package would cover improvedhotel accommodations, and specific interventions not covered in the initial two tiers. The Voluntary Health Insurance i s supposed to be managed by the private health insurance agencies that may also manage the program for Compulsory Health Insurance. Fourth tier: Charity and other voluntary payments. ProblematicIssues concerningthe Proposal 3.52 One o f the greatest problems with this reformproposal (and others that have been put forward) i s that it does not clearly specify the objectives o f the new SHI system. Without explicit objectives, such as universal coverage or financial protection from catastrophic illness, it will be difficult for stakeholders to fully comprehend whether the law provides them with sufficient instruments to meet these situations and to establish clear accountabilities. 3.53 There are additional reasons to believe that the proposed SHI reform will face serious difficulties in resolving the major problems confronting Ukraine's health system. This is because the reform is focused primarily on raising financing levels, but it i s clear that the low level o f financing i s not the key problem. The system has several distortions that generate problems o f efficiency and equity in the delivery o f health services. For example, the system i s severely constrained by supply-side norms which inhibitthe possibility o f using available resources appropriately. The proposed reform does not address this and other such distortions. Without addressing these concerns, the reform is likely to compound the existing distortions and make it increasingly difficult to improve the situation at a later stage. 3.54 Another major problem i s that the boundaries o fthe entitlement for each o f the proposedsystem's tiers have not been defined, yet the sources o f financing have. This has serious implications as it could generate (i)problems o f management and accountability; (ii)problems o f sustainability as more obligations are made compulsory; and (iii)difficulties in attracting private sector interest given the absence o f transparency for both the compulsory and the voluntary insurance systems. 3.55 The proposed scheme i s likely to maintain a public managed compulsory health insurance system and, at best, a voluntary health insurance system which would mostly be a substitute for public insurance for those who could afford it. In addition, the fourth tier, as well as the first three 3 tiers, would, nonetheless, allow for the continuation o f the possibility o f informal payments made to providers, thus perpetuating a major source o f inequity and inefficiency inUkraine's health financing system. 3.56 Furthermore, the governance arrangements o f the system are not clearly defined. It i s unclear, for example, how intergovernmental fiscal relations would accommodate to the new scheme. If budgetary transfers from the central government and local governments with historical budgets continue as before, then the first tier will have no contracts and no incentives for improvement. The compulsory second tier will be managed initially by the State Health Insurance Organization. However, this institution also has the mandate to establish the initial contracts with providers as well as to set the rules o f the game for the opening o f the market to private institutions. Thus, there is the likelihood o f a clear conflict o f interest, which would make it difficult for this tier to open upthe private sector on a level playing field. 74 3.57 Inshort, the current SHIproposal, whichlacks clear accountabilityandentitlement arrangements and does not eliminate the existing supply-side rigidities and problems, would essentially throw more money at a problematic health system, potentially undermining future efforts to improve the equity and efficiency problems. Generating greater levels o f financing will not fix the underlying distortions in the system discussed inthis chapter. F. DISCUSSIONAND RECOMMENDATIONS 3.58 The health system in Ukraine suffers from inefficiencies in the management, financing, and delivery o f health care. In addition to underperforming, the current system generates inequities across oblasts and income levels. While maternal and child mortality rates have improved in recent years, life expectancy remains low and there i s evidence that many people lack financial protection from OOP. 3.59 One o f the health system's major problems i s the budgeting rigidities imposed by the central M o H via the imposition o f input-based norms that effectively determine how resources are budgeted in facilities at the local level (as discussed earlier, local governments spend close to 80 percent o f total public spending on health care). The inefficiencies created by these and other regulations in the health sector have led to a system that it i s inefficient (in terms o f staffing and networks o f facilities) and unresponsive to health care needs- that is, a system that leaves little scope for changes in expenditures based on localized health care needs. Resolving this bottleneck in health financing i s a first step toward allowing the following: a more effective use o f resources in order to enhance the quality o f service provision; a more outcome-based system; and a health system that places more emphasis on preventive and primary care while reducingthe financial burdeno f health care on the poor. 3.60 While public spending in Ukraine i s not particularly high by regional comparison, a key contention o f the analysis inthis chapter i s that additional financing would not necessarily enable Ukraine to remedy its health system problems untilthe shortcomings o f the present system are resolved. In other words, throwing,scarce public resources in a system that is inefficient and plagued with problems would further exacerbate inefficiencies o f the total fiscal envelope. Inthe same way, the current health insurance reforms under debate, which also focus on raising additional sources o f financing for the sector, would not solve the fundamental distortions that underlie the health system. 3.61 Thus, at a first stage the focus of reformneedsto be on addressingthe underlyingproblems of the current system. Moreover, indoing so the government could generate the necessary fiscal savings inthe health budget and re-allocate them within the sector towards quality-enhancing expenditures and investments (such as medicines, training, modem equipment, and other investments). A simple estimation suggests that reducing the number o f hospital beds (and with them physicians and nurses) to the EU-10or EUlevels would yield savings o f 0.25 and 0.34 percent o f GDP (per annum), respectively. The savings could reach up to 0.4-0.6 percent o f GDP (per annum) ifthe merging o f facilities takes place and with it a reduction in utilities expenses and in non-medical staff. Although this is a target, gradual adjustments towards this goal would nevertheless allow enough resources to gradually meet the investments needs estimated inthis report. 3.62 Inview o fthe problemsand issues raisedinthis chapter, the following selectedrecommendations are identified: 0 Revise and reformulate the norms dictated by the M o H for budget formation at the facility level. The present situation is a bottleneck for health financing and local government budgeting and also creates serious inefficienciesand perverse incentives inthe system. Its elimination would allow a more efficient use o f resources and more flexibility, which, inturn, would permit higher 75 budget allocations for investments and higher salaries withinthe health sector. This is inline with actions proposed by the government's program in relation to creating the regulatory conditions for budgetary (decision-making) decentralization. Local governments through hospitals and clinics shouldbe able to adjust their staffing, capital, and other input levels as needed to optimize performance and service delivery. The system must move away from such an input-based approach to an output-based one. The first step (and priority) i s an item-by-itemrevision of such M o H orders as Order # 33. The guidingprinciple of this revision should be the identification and elimination of all norms that dictate health spending and the allocation of resources such as medical and non-medical staff and materials based on existing inputs (such as number of beds or other facility physical characteristics). Then, if the government desires to replace the norms, the right vehicle is the establishment o f purely demand driven factors as a guideline. These factors could evolve by pricing unit costs of each service (in line with the H factor of the equalization formula), that is, a minimum standard per unit of service (in UAH terms). These units costs, in turn, would help to attempt more sophisticated financing mechanisms in the future such as contract agreements for service packages (with outcomes as measure of results) or other schemes that would allow the government to pay services and obtain better health outcomes for the populationas a return. 0 Move those norms that assure a minimum standard of quality (not related to financing rigidities or staffing) to the accreditationprocedure for facilities. The remaining normatives that do not impose any staffing rigidities based on inputsbut rather assure a basic level of quality based on existing demand could be transferred to a more efficient and quality-enhancing accreditation procedure for health facilities and health providers. The M o H should focus on strengthening the current accreditation procedure, as this would help regulate quality standards. Accreditation (not financing norms) shouldbe one o f the principal responsibilitieso f the MoH. Give localgovernmentsthe power to start the process of rationalizing the network of health facilities. Although the constitutional prohibition o f facility closure is an impediment that would be difficult to overcome, a revision should be attempted. Inthe meantime, the government could begin the process through the "merging" of facilities - an action which i s not prohibited and i s allowedby the Budget Code. 0 Ensure that factors within the formula of intergovernmentaltransfers do not contain any linkages to the network of facilities. On the contrary, transfers should fully account for the demand for services and, ideally, should reward efficiency gains and performance (see Chapter 2)* 0 Address the inequalitiesinthe system by offering poor householdssome financial protection fromthe costs of illness. The government shouldidentify mechanisms to make the healthbudget more pro-poor. Among the principles that shouldbe pursued inestablishing these mechanisms are the following: o Improve the targeting of expenditures at facilities that the poor prefer to use. More equitable spending for the benefit of the poor could be undertaken ina number of ways. The poor appear to use rayon level hospitals more often than other facilities. Targeting additional budget allocations to these facilities could help them purchase such necessary inputs as supplies so that households would not be forced to finance them through OOP. In addition, capital expenditure, which i s often quite low and i s indicative o f quality, could be directed from central government funds to polyclinics and rayon level hospitals. Better targeted public spending could also focus on public goods, such as health information programs regardingalcohol and tobacco. o Purchasedrugs especially for the poor, ifnondiscretionaryexpenditures can be reduced by reforming the norms. Drugs constitute the largest share o f OOP spending, and about 10 76 percent of Ukrainian households exhibit catastrophic OOP spending on drugs alone. A switch from an input-based to an output-based paymentheimbursement system between oblasts and facilities could help compensate facilities that target the poor. Facilitating the purchase of needed drugs, especially at polyclinics and rayon hospitals where utilization by the poor i s highest, would be an important step towards improving equity and reducing high OOP spending. Provide adequate incentivesto shift a larger proportion of doctorsinto primary care and to integrate the family medicine model into Ukraine's health system. The large number o f specialists and the absence o f gate-keeping translates into higher than necessary health costs. While the current health insurance reform focuses primarily on financing health services, it will be critical to address the issues on the delivery side as well. A key component o f any health reform would be better integration o f and coordination between primary, secondary, and tertiary care to allow for more cost-effective service provision. The establishment of a strong gate- keeping role at the primary care level could contribute to the fiscal sustainability o f additional financing inthe sector. Increase excise taxes on alcohol and tobacco to deter consumption.A highuse of alcohol and tobacco leads to poor health outcomes, especially to a low life expectancy among males. Experience shows that such taxes can have a positive impact on health status. Moreover, in Ukraine these excises are low when compared regionally, and there is considerable room for an increase. This i s inline with World Bank advice to Ukraine on tax policy. Increase the preventionprograms' spending in relation to HIV/AIDS and tuberculosis,but in parallel with improving their quality. The high rate of increase in HIV/AIDS and the reemergence of tuberculosis could compromise any future achievements in health status or advances in service delivery and financing. Tackling this problem would require the scaling up of treatment and care, particularly through enhancing coordination between the central government and local governments and other public and international organizations involved in service provision. Butthis task would also needto be combined with improving the quality of the prevention programs, otherwise resources could be inefficiently spent. On the latter, an updated and in-depth evaluation of prevention programs in Ukraine is necessary. Additionally, the emphasis should be placed on the availability of sustainable financing for the national A I D S program, particularly on the development of effective mechanisms for the integration of vertically oriented programs (HIV, tuberculosis, drug abuse, maternal and child health, etc.) and the decentralization of service provision inorder to reachvulnerable groups. a Arrange for additional financing frdm external donors through mechanisms geared to increasing the current fiscal envelope of selected activities in coordination with the government. To avoid the budget fungibility o f resources devoted to certain areas such as HIV AIDS, tuberculosis, and ARV treatment, donors should require the government to maintain the current level of spending as a percent of GDP inthese activities (as o f 2007) as a minimumfloor and then, on top o f that minimum floor level (if it i s met in the approved budget allocation), should contribute on a matching grant scheme basis. a Work with donors to allow financing of health care programs through the public financial system. This also requires a commitment of donors to fund programs using country systems (as they are strengthened). Over time, and through the restructuring of the system, reform the wage policy in the sector. Currently, salaries do not offer doctors the incentive to deliver efficient or quality care, and restrictions on salary levels may exacerbate the problem. Over time, and under the close control of the Treasury, local governments (at the oblast and rayon and city levels) should be able to determine the salaries o f different types of doctors to provide appropriate incentives for 77 improved service delivery. For example, some rural areas may lack a sufficient number o f primary care doctors, and raising the corresponding salary level might attract doctors from other areas. Provider payment reforms are complex, and may take time to devise, implement, monitor, and evaluate, but as economies grow and health systems expand, this i s likely to become a salient issue. 3.63 These steps could help correct some of the underlyingproblems facing Ukraine's health system and could help provide the necessary conditions for scaling up spending in an efficient, equitable, and sustainable manner. Restructuring the allocation of financing among oblasts and the budgeting o f resources at individual hospitals are critical steps to ensuring that the government's health expenditures reach the objectives of improving population health and financial protection. Improvements in health outcomes and better financial protection from catastrophic illness costs are unlikely to occur if measures are not taken to reform the existing inefficiencies. Inequities across oblasts and income levels generated by high OOP will continue and cannot be fixed by the channeling of more money into a health system hamperedby institutional arrangementsthat uphold misaligned incentives. 78 4. OVERCOMING FISCAL, EFFICIENCY,AND EQUITY CHALLENGES INPUBLIC EDUCATION SPENDING A. INTRODUCTION 4.1 In Ukraine, the state is the main provider of education services at all levels and private participationis negligible. Inrecent years public expenditures on education have grown rapidly inreal terms and absorb a high proportion of public expenditures. However, this growth i s coupled with worrisome (and increasing) levels o f inefficiency ineducation provision (at all levels). At the same time, the quality enhancing inputs and investmentsnecessaryto modernizing the system and to making it more responsive to market needs are under-provided. The inefficiency arises from a variety of sources, particularly in relation to financing, administration and intergovernmental fiscal relations (the bulk of public education spending takes place through local budgets). Inparallel, inequalities are surfacing owing to increasing out-of-pocket payments (OOP) and structural governance issues of the system. 4.2 This chapter is organized as follows: Section B provides a brief description of the education system structure in Ukraine. Section C examines selected education indicators and trends in public expenditures in the sector. Section D discusses key problematic issues facing Ukraine interms of supply side efficiency in education provision. Section E discusses financing and administrative issues that affect efficiency and deter improvements to the system. Section F discussesout-of pocket expenditures inpublic facilities and the equity and the governance issues linkedto them. Section Gprovides a briefdiscussion o f quality issues. Finally, Section H provides a discussion of and selected options for reform and recommendations to improve the value of public spending inthe sector. B. THEEDUCATIONSECTORSTRUCTURE 4.3 Ukraine has a national system of public education that dominates the provision of this service in the country. Private participation in education i s negligible, from pre-school to higher education, and i s limited to a few private schools and universities in some major cities and to a limited number of schools runby SOEs and largeprivate firms. All non-public schooling covers only a marginal fraction of enrollees. 4.4 The structureof Ukraine's education system is in transition from the former 10-year system o f primary and secondary education to a 12-year system, and from a 5-year program o f higher education qualification to a 4-year bachelor's degree with the option of a master's degree with an additional year o f study (Annex 4.1 presentsa graphic representation of the transition scheme).lo2Inaddition, the education lo*Ukraine's education system comprises a 4-year cycle o fprimary education, followed by 5 years o fpost-primary basic education (or lower secondary education), 3 years o f secondary education, short programs o fhigher education (called levels Iand 11), bachelor's levelundergraduate education (called levels I11and IV), and various programs o fpost-graduate education. (Annex 4.1 provides a description o frecent changes inthe structure o f the education system. ) Secondary education i s o ftwo kinds: general secondary education and vocational education. Children may start primary school before their eighth birthday, and may start at age seven. Attendance inthe entire 12-year program o f basic education and secondary education is compulsory. In Ukraine, the term "general secondary education" refers to the entire sequence o f primary, post-primary basic, and general secondary education. That terminology is used throughout this chapter, except indiscussions o f general and vocational 79 system maintains schools for special needs students at all levels before university education. Inparallel to the mainstream programs and institutions, Ukraine maintains a pool o f special schools for advanced or gifted students at each level: special pre-schools for early start education, "gymnasiums" and "lyceums" for advanced students inprimary and secondary education, and special 2-year colleges. 4.5 Ukraine has an extensive and oversized network of public education facilities, as discussed and examined indetail inthis report. The network encompasses more than 18,500 pre-schools (including those for special needs education) and close to 27,000 primary and secondary (general secondary) schools (including 400 boarding schools for disabled and special education institutions). There are around 1,700 special pre-school centers for early start education, 273 gymnasiums, 232 lyceums, and 25 (2-year) special colleges in the country,'03 and the number o f these special education institutions i s increasing annually. 4.6 Post-secondary (higher) education in Ukraine is made up of a diversity of types of institutions, comprising 345 universities, academies, and institutes, and including a few private institution^.'^^ The programs are intransition from a five-year "specialist" program to the new Bologna model o f four-year bachelor's programs, and graduate programs at the master's and doctorate levels. The responsibility for managing higher education institutions i s fragmented across many ministries-a legacy o f the Soviet period, when many universities were closely linked to sectoral ministries. The Ministryo f Education and Science (MoES) i s responsible for generic education policy, including the elaborate schedule o f input norms that are meant to control higher education resource allocation and quality. The Ministry of Finance is responsible for setting and enforcing recurrent expenditure standards. The MinistryofLabor and Social Policy setsthe standards for faculty compensation. 4.7 The education system is managed by the MoES in a delegated system: local government administrations at all levels (see Box 4.1) administer all levels of schooling prior to university education and public universities administer higher education on a contractual basis. Although local governments are an integral part o f today's education system structure, in practice the system i s highly centralized fiscally and administratively. The MoES coordinates the system o f delegated functions through the 24 education departments o f the oblast state administrations and the education departments o f the Autonomous Republic of Crimea (ARC) and the Kyiv and Sevastopol city state administrations. These regional departments, inturn, coordinate downwards witWthrough rayons and cities (second tier o f local government) and villages (thirdtier o f local government) under a fairly hierarchical system (see Box 4.1). Inaddition, several line ministries maintain specific field educational programs, mainly inthe areas o f agriculture and transport. secondaryeducationingrades 10through 12, inwhich it is clear fromthe context that the discussionrefers only to grades 10 through 12. lo3Ministry of Education. IO4This sectiondraws inparton the data andanalysisprovidedin Ukraine Tertiary Education Assessment, EU-TACISDraft FinalReport, 2006. 80 Box 4.1: EducationSector StructureandAdministrativeAccountabilities 4.8 The central MoES is responsible for the formation of the education policy and the control of i t s execution. Local governments at all three levels have specific expenditure responsibilities assigned in the education sector. In broad terms, small towns and villages are in charge o f pre-school and in some cases primary education, rayons and cities are in charge o f general secondary education (in many cases this includes primary education and even pre-schools), oblasts and cities o f special status are incharge o f technical and vocational schools, and the central government i s in charge o f higher education with contracts with the universities. Chapter 2 provides details o f overlapping of functions as well as financing arrangements betweenthe central and local government levels. 4.9 The overall education system has seen limited reform. Aside from the intergovernmental reform (through the Budget Code) which changed and clarified to some extent the responsibilitieso f local governments in the education sector (see Chapter 2), the rest o f the reforms have been reactive to short- term budgetary constraints in the late 1990s and thus have had marginal impact. These reforms include the following: (i) providing "contracted" higher education within public universities for students with entry scores below the threshold for budget-financed students admission (under this scheme, universities 81 can admit additional students against fees, which are retainedat the university level) (see Box 4.2)'05; (ii) requiring parents to purchase textbooks and other educational materials that were formerly provided for free by schools; and (iii) allowing schools to raise and retain funds through actions such as the rental or sale o f unneeded facilities and the provision o f paid extracurricular courses. But while most o f these reforms were well intentioned, they also contributed to emerging new problems, particularly problems related to equity and quality. Furthermore, none o f the reforms to date has resolved the key structural and fiscal bottlenecks inthe system. Box 4.2: Financing of Higher EducationinUkraine:Contract Education Within public higher education, there is a system o f "contract education" which is equivalent to private universities operating within ostensibly public universities. Currently, slightly under half o f the students in public higher education institutions are "contract students," and pay fees ranging from US$400 to US$1,500 per year. Budgetary finding o f higher education students is limited to "budget students" who score above a designated threshold on the university entrance examination. The number o f budget places for students in higher education is implied by three key parameters: (i) an overall enrollment target which is set by Parliament"; (ii)limitonthemaximumshare a (currently 49 percent ) o f contract students in total higher education enrollments), which is also set by Parliament; and (iii) the available budget for higher education. Budget students are especially numerous in fields such as agricultural and industrial engineering, which were developed and promoted to meet the needs o f the planned economy under the former system. Contract students are heavily concentrated in fields such as management, economics, and the humanities that have grown rapidly in recent years to meet the needs o f the new market economy. Budgets for individual higher education institutions are essentially incremental adaptations o f prior-year budgets, with adjustments designed to reflect impressions (rather than documentary evidence) o f evolving needs, and enrollment targets set by the Cabinet o fMinisters. Line-item budgeting leaves little room for expenditure discretion at the institutional level, amounting, o n average, to less than 5 percent o f the total institutional budget." Similarly, rectors and deans have little flexibility in staffing, salaries, and faculty workload, since these are all constrained by schedules and norms set by the Ministry o f Education and Science and the Ministry o f Labor and Social Policy. Higher education institutions are able to retain earnings from student fees, service provision, and the rental o f unused facilities, but they are not allowed to buy and sell premises. Capacity expansion i s thus dependent o n the donation or transfer o f new premises. Inthis highlyconstrained environment, a key revenue variable at the institutionallevel is the number o f fee-paying "contract" students. Student fees for contract students range between US$400 and US$1,500 per year. On average, these fees are considerably higher than the average budget financing provided for the budget-financed students. The receipts from these fees constitute an important source o f supplementary income for most universities. But even here the opportunity for individual institutions to increase revenues from student fees is constrained by legislation specifying the maximumpermissible percentage o f fee-paying "contract" students among newly admitted student^.^' Students with the highest entrance examination scores are admitted without fees. Contract students tend to be o f lower ability because o f the lower examination performance requirement for entran~e.~'In addition, budget- financed students who maintain an acceptable grade-point average receive stipends from the state budget. Under the 2002 Law on Higher Education, stipends were set at twice the pre-tax minimumlevel o f income established by the Ministryo f Labor and Social Affairs. Actual stipend levels are much less - currently about 30 percent o f the minimumlevel o fincome. l' The 2001 Education L a w establishes the target o f 180 students per 10,000 population for university degree rograms, and the target o f 100 students per 10,000 population for post-secondary non-degree programs. Ibid. 3'Parliament has just increasedthe maximum percentage o f contract students among newly admitted students inany individual university from 49 percent to 50 percent. 4` There are reports that some budget-financed students are actually o f lower ability but obtained their budget positions through corruption, rather than by scoring above the threshold score on the national examination. ~~ ~~ lo'Higher educationisnotthe mainfocus ofthis report; however, some issuesrelatedto publicspendingand governanceinthis areaare raised. 82 C. PUBLICEDUCATIONEXPENDITURESANDMAINEDUCATION INDICATORS: TRENDSANDANALYSIS PublicExpenditures 4.10 Public expenditures on education are high and have grown rapidly over the last few years. They grew steadily from 4.9 percent o f GDP in 2000 to 6.3 percent o f GDP in 2006 (Figure 4.1), outpacing many other functional categories in terms o f real growth over this period o f time. This high level o f budget commitment currently represents more than one-fifth o f the consolidated budget o f Ukraine.lo6 This level o f expenditures and their share inthe budget are very high in comparison to most other transitioncountries and OECD countries (Figure 4.1). Figure4.1: UkrainePublic EducationExpenditure as Percentof GDP (2001-06) andRegionalComparison" - Public Education Expendituresas % of GDP Public Expenditure in Educationas % of GDP 7,00r---- - 7 l--------- 6.00 -- 5.00 -- 4.00 -- 3.00 -. 2.00 -. 1.00 f 2002 2003 2004 2005 2006 1Datafor Ukraine are for 2006; data for other countries are for 2004. Sources: MoF; State Treasury o f Ukraine; State Statistics Committee; Eurostat; OECD; Bank staff calculations. 4.11 Public expendituresby levelof educationhavebeenrelatively stable. The share of each level as a percent o f total public expenditure in education has varied little since 2002 (Figure 4.2). Higher education expenditures were temporarily contained and even slightly reduced through the contracting mechanisms with public universitiesbutthey have returnedto pre-2003 levels. I O 6It should beborne inmindthat the consolidated budget plus the extrabudgetary social insurance funds together form the notion o f general government budget. 83 Figure4.2: PublicExpenditures on Educationby Level as Percentof GDP, 2002-06 7.00 BIOther education institutions and 6.00 other activities UllNon-school education 8, non- 5.00 school work with I @%&?Zduate 4.00 education S Higher education 3.00 2.00 0Vocational-technical education 1.oo 0General secondary education I 0.00 2002 2003 2004 2005 2006 Sources: MoF; State Treasury o f Ukraine; State Statistics Committee; Bank staff calculations. 4.12 Local governments spend more than 64 percent of the public education budget of the country (see Figure 4.3). The bulk of local government expenditures for pre-schools and general secondary education i s financed through the intergovernmental finance system (non-earmarked shared revenues and transfers, as detailed inChapter 2). Central budget resources directly finance the 448 higher education institutions and some vocational schools. 107 Central government budgets for the institutions which are directly managed by the MoES are negotiated by the Ministry directly with each institution, basedon budget "indications" providedby the institutions. Figure 4.3: Public Expenditures on Educationin2001-06, byLevelof Government and Levelof Education Shares of Public Education Spendina: Central vs Local 60% 50% 40% 30% I Local 20% 10% 0% 2002 2003 2004 2005 2006 Total Local C8ntraI Pre-school 100% 99-/- 1-/o Oeneral secondary 100% 9a90 1-/o Vocational-technical 100% 4-/o 96-/o Hiaher education 100% 7-/- 93-/o ofwhlch levels I & II 100% 30% 70% Ofwhich level8 Ill & IV 100% 1-/o 99=% Educatlon Total 100% 63% 37-/- Sources: MoF; State Treasury o f Ukraine; State Statistics Committee; World Bank staff calculations. 4.13 Per pupil public spending on education is not visibly correlated with regional income. Figure4.4 shows per pupil allocations across Oblasts with oblasts ordered from higher to lower regional lo' Additionally, many specialized vocational schools are managed and financed by various sectoral ministries, but these expenditures are line items o f other ministries' programs (for example, education and extension programs inthe Ministry o f Agriculture) and they are not counted for the purposes of this review. 84 income, from left to right. These trends, however, mask problems such as high pressures on recurrent spending on education on all local governments but more acutely on poorer oblasts andrayons, which in turntranslates into less capital spending and less spending on quality-enhancing inputs. Moreover, these amounts significantly understatethe full differencebetweenricher and poorer oblasts becausethey do not include parental informal contributions, which in the more affluent areas constitute an important additional source of financing. Figure4.4: Per PupilPublic ExpendituresonEducationAcross Oblasts (consolidatedbudgetinUAH),2005 3,000 2,000 1,000 500 wrces: MoF; State Treasury of Ukraine; State Statistics Committee; Bank staff calculations. Why Has Spendingon EducationIncreasedRapidly? 4.14 Given Ukraine's demographic profile, the increase in public spending on education bears further investigation. Ukraine's birthrate has fallen significantly since independence and consequently more recent student cohorts are smaller than in the past. This should have an effect on the total levels of spending as well as the composition of spending, as the demographic changes are first noticed at the primary level before translating into lower intakes at higher levels of education. This section investigates the sources of increased spending by looking in more detail at (i) enrollments and coverage, and (ii) theunit costs of education. Enrollments and Coverage 4.15 Total enrollments are falling. Enrollments have declined at all levels of education except higher education over the past five years (Figure 4.5). This pattern is similar to the enrollment changes that have occurred in other transition countries in the region, except that the onset o f the decline came much later in Ukraine.'os As described below, the enrollment declines in pre-school, primary, and secondary education result from falling enrollment ratios and from shrinking school-age cohorts. Enrollments are certain to decline further as the school-age population continues to shrink. `OsMichael Mertaugh, "Aging and Education," inFrom Red to Gray: The "Third Transition" of Aging Populations in Eastern Europe and the Former Soviet Union,The World Bank, 2007. 85 Figure4.5: EvolutionofEnrollmentsbyLevel, 1990-2006 -%-- 7,000.0 I- :: :: . :: _ _ _ I _ _ ~ _ _ _ _ ~______llll_l -general 6,000.0 secondary (full time) 5,000.0 -higher 4.000.0 3,000.0 -Preschool 2,000.0 vocational 1.ooo.o Source: Based on figures ofthe State Statistics Committee. 4.16 Higher education enrollments have expanded rapidly, with current coverage rates now higher than in many new EU member states, but there may be significant gaps in education coverage at the compulsory level. The prevailing view in Ukraine i s that compulsory education coverage i s virtually complete. The available data, however, provide a mixed picture of education coverage. Table 4.1 presents the publishedcoverage estimates for compulsory (primary plus secondary) education in Ukraine. Gross enrollment ratio (GER) estimates indicate the number of registered enrollments at a given level, dividedby the estimated population of the normal age group for that level of education. But GERs tend to overstate actual coverage because they include overage students in the numerator. There i s clearly a problem with the UNICEF estimate o f the GER for upper secondary education since an enrollment ratio as low as 62.6 i s not plausible by any basis o f comparison. The preferred indicator of education coverage at any level of education i s the net enrollment ratio (NER). NERs are usually calculated from registration data as the number of students enrolled at a given level of education who are within the normal age range for that level o f education, divided by the estimated population of that age group, but this estimation i s not without the potential for errors.1o9 The only available NER series for Ukraine i s that publishedby the UNESCO Institute o f Statistics, which i s calculated from data provided by the State Statistics Committee o f Ukraine. The UNESCO Institute of Statistics reports anet enrollment ratio o f 82 percent for basic education in2004, and 84 percent for upper secondary education (Table 4.1). If accurate, these estimates imply that there are some 1.4 million children o f compulsory school age who are not attending school. Verifying the extent o f the incomplete coverage of compulsory education and filling that gap should be an important goal o f public policy, includingpublic expenditure policy. logThese figures are subject to error from two sources: Errors can arise inthe numerator from errors in identifying underage and overage enrollments, and inthe denominator from errors in estimating the size o f the population for the relevant age group. 86 (Primary + Lower Basic(YO) Indicator/ DataSource Upper Secondary ImpliedGap inCompulsory Secondary) ("/.I EducationCoverage Gross enrollmentratio UNESCO Institute of Statistics 95.0 93.0 n.a. UNICEF InnocentiCenter 94.5 62.6 n.a. UNESCO Institute of Statistics I 82.0 84.0 1,384,900 4.17 Survey data indicate greater coverage but also show that coverage is lowest for low-income households and rural areas. Estimations based on the HBS Health and Education Module Sample Survey carried out in 2004 show overall enrollment rates at above 95 percent for both basic and upper secondary education but also reveal a strong income and locational gradient in enrollments - with significantly lower pre-school and higher education coverage in rural areas and low-income households.' lo Unit Costs 4.18 The increase in education expenditures is not justified by enrollment changes. Inorder to understand the sources o f the increase in education expenditure, it i s useful to start by decomposing the increase into its component parts. Table 4.3 does this by looking at the pattern o f expenditures by level o f education over the period 2002-05,'" and comparing the actual expenditure increases with the expenditure increases that would have occurred with constant unit costs. "* To abstract from the effect o f price changes, this comparison i s made using constant 2002 prices. Table 4.2 also distinguishes the costs that are directly related to enrollments (for which unit costs can be calculated) and the costs which are not directly related to enrollments (for which unit costs are less appropriate). As shown in column E, the increase in higher education enrollments was less than half the size of enrollment declines in pre-school, primary and secondary education. Thus, even though per student costs for university education are twice as highas those for primary and secondary education (or general secondary education), this difference i s more than offset by the rapid contractiono f enrollments inprimary and secondary education. 'loSee World Bankpovertyupdate2006. ''IThis comparison i s carriedout only for the period 2002 -05 becausedetailedexpenditurefigures are not availablefor earlier years, and enrollmentfigures are not available for pre-schooleducationfor 2006. 'I2That is, average cost per student. 87 r A B C D E F G H I Changes in 2005 2005 Projected Excess 2002 2002 2002 2005 Enrollment Actual Actual 2005 Expend Expend. Enrollment Unit Cost (in Enrollment (thousands) Expend Expend of 2002 (In millions (inmillions (thousands) 2002 UAH) (thousands) D-B (Inmillions (Inmillions = (C D)/ * of2002 of2005 of 2002 UAH); ofUAH A/B 1000 * UAH) UAH) 1000 G-H Preschool Generalsecondary Vocational/technical Higher, ofwhich Levels I& I1 Levels 111& IV Graduate SUBTOTAL Non-school Ed. materials Ed.research Other TOTAL I Source: State Statistics Committee; MoF; State Treasury; WB World Development Indicators (including deflators); Bank SI Iff I Calculations 4.19 Essentially all of the recent increases in education expenditures reflect the very large increase in unit costs for all levels of education except levels I11 and IV (bachelor level) higher education. Inaddition to these increased outlays for directly student-related categories, there were large real increases for expenditure categories not directly related to enrollments: non-school expenditures, educational materials, educational research, and, especially, the residual "other" category. Table 4.3 summarizes the growth inunit costs, based on data presented inTable 4.2. Expressed innominal terms, unit costs increased by well over 100 percent for all levels o f education except bachelor degree level higher education. After adjusting for price inflation, unit cost increases were still very high - ranging from 45 percent to over 81 percent injust three years (as shown inthe final column o f Table 4.2). Table 4.3: ChangesinUnit Costs of Education by Level, 2002-05 2002 2005 2002-2005 Unit Cost Unit Cost YOChange in YOChangein YOChangein (incurrent UAH) (inUAHof 2002) Enrollment Unit Cost Unit Cost Preschool 1,295 1,884 -0.40% 115.10% 45.50% General secondary 786 1,424 -15.50% 167.80% 81.20% Vocational/technical 1,46 1 2,388 -1.10% 141.70% 63.40% Higher 1,836 1,98 1 19.30% 59.50% 7.90% Levels I& I1 1,40 1 2,321 -13.30% 144.90% 65.70% Levels I11& IV 1,986 1,903 30.60% 41.60% -4.20% All levels 1,242 1,749 -0.80% 108.10% 40.80% Source: State Statistics Committee; MoF, State Treasury; WB World Development Indicator; Bank staff calculations 4.20 The decline inthe constant price unit cost of bachelor level higher education does not imply improved efficiency in delivery. The unit costs shown in Table 4.3 for higher education significantly understate the actual costs o f education provision because they reflect only budget-financed costs. A significant share o f the costs o f higher education are paid by students in the form of student fees which are retained by their institutions and used to meet routine costs (around 40 percent o f the total public budget at this level). Since about half o f the enrolled students in higher education pay fees (which on average cover the full costs o f their education), the recorded unit costs o f higher education significantly 88 understate the actual per student cost of higher education. Another source of incomparability inthe unit costs of higher education i s the stipends which are paid to most students to motivate an acceptable performance and to help meetliving costs. 4.21 The rise inunit costs is driven by wage increasesand the failure to adapt staffing needs and network facilities to the shift in demand for ,education reflected in changing enrollments. The increases in real wages in the public sector drove most of the overall increase in public spending on education in2005 and 2006. As explained inChapters 1and 2, wage increasesimpose a higherburden on local budgets, which manage pre-schools, and primary and secondary education where salary inputs are higherinthe cost function. 4.22 But despite recent increases, teacher salaries remain low (see Figure 4.6). Salary increases were by far the most important source of the recent major increase inpublic expenditures for education. Despite this fact, education sector salaries are low in comparison with other sectors (73.7 percent of average earnings) and with the relative salaries o f teachers in other countries. Average education sector earnings of 672.6 UAHper month in 2006 in Ukraine amounted to 75 percent o f per capita GDP, while teacher earnings inthe OECD countries average 130 percent of per capita GDP.*l3 Low teacher salaries have contributed to the solicitation of informal payments from students and parents, and may be contributing to deficiencies in attracting qualified candidates to the teaching profession (with long term quality implications). It i s therefore clear that Ukraine needs to contain unit cost increases in education primarily through efficiency improvements rather than by containing teacher wage increases. Figure4.6: Evolutionof AverageEarnings inthe EducationSector and Other Sectors (in UAH)2002-07 1 1400 -c Industry 1200 +All - Sectors 1000 800 Education 600 -Agriculture 400 200 I 2002 2003 2004 2005 2006 2007 I Sources: State Statistics Committee; World Bankstaff calculations. D. EFFICIENCYISSUES 4.23 At independence, the education sector, like other sectors, embodied considerable inefficiency, resultingfrom the absence o f hard budget constraints and the lack of efficiency incentives inthe planned economy. The limited reforms since independencehave not effectively addressed the root causes o f this inefficiency. Table D3.1, Education at a Glance, 2006, OECD. 89 4.24 Ukraine's average studentheacher ratio is extremely low (and falling) in comparison with other countries in the region, indicating room for major improvements in the efficiency of teacher use. Measured by the universally applicable metric o f studendteacher ratios, Ukraine's education system i s presently among the least efficient in the world (Figure 4.7). The studendteacher ratio in general secondary education (which comprises primary and secondary levels) inUkraine i s at 9.4, only 55 percent o f the OECD average and far less than halfo f the ratio ina number o f other high-achieving countries such as Japan and Korea. This deterioration in the number o f teachers per student was responsible for 12.5 percentage points o f the increase inthe unit costs o f general secondary education between 2002 and 2005. Inpre-school educationUhaine's studendteacher ratiois alsoextremely lowbyinternationalcomparison. Figure4.7: The Studentto Teacher Ratio inUkraine is Acutely Low ComparedInternationally (primary + secondary) 35 30 25 20 15 10 5 0 ources; OECD; Eurostat; SSC: Bank staff calculations. 4.25 Moreover, student teacher ratios have been falling dramatically over the last years (see Figure 4.8). This reflects an adjustment o f only 6 percent in the number o f teachers over the last eight years while net enrollees fell by more than 25 percent over the same period o f time. This negative trend will continue ifmeasures are not taken soon. 90 Figure4.8: DecliningStudent/Teacher Ratio inUkraine 1 - l2 10 8 6 studenwteacher Ratio in General Secondary Education (primary + secondary) 4 -- 0 1 I 1999 2000 2001 2002 2003 2004 2005 2006 source: State StatisticsCommittee; Bank staff Calculations 4.26 The positive effect of attrition, that is, the natural reduction of teaching staff due to retirement, is being offset by new hiring. Currently, around 15 percent of the teaching staff i s between the ages o f 50 and 55 years o f age, and there i s a large portion (close to 30 percent) between 41 and 50 years. According to this profile, aging staff, together with the rest o f the populationprofile o f the country, would in theory, contribute to improving this efficiency ratio through attrition over the next 10 to 15 years. But the reality i s that the system continues to re-hire teachers who have reached pension age (currently, more than 12 percent of the teaching staff are above the age of 55 and this figure has been growing) while keeping additional regular hiring. The continued re-hiring of retiree teachers (and new hiring) i s carried out to fulfill the current normative of staffing. 4.27 Demographic trends will lead to even greater inefficiency. This i s particularly concerning since falling fertility i s leading to shrinking school-age cohorts. The basic education cohort i s projected to shrink by almost 1.5 million by 2025 (Figure4.9). The cohorts for secondary and higher education are projected to shrink by 1million and 2.3 million, respectively (Figure 4.9). Unless the government takes decisive action to strengthen efficiency incentives in education management, inefficiency in the system will increase as the school-age populationand enrollments continue to shrinkand teacher numbersremain largely without adequate adjustments, leadingto further redundancy of teaching staff and facilities. Figure4.9: The SchoolAge PopulationWill Continueto DeclineOver the Next20 Years I School Aged Population Trends 2005-2025 6,000,000 5,000,000 4,000,000 15 to 17 3,000.000 . 18 to 23 2.000.000 1 .ooo.ooo o ! I 2005 2010 2015 2020 2025 Source: Basedon UnitedNations Population Division (medium variant projection) 91 4.28 The ratio of non-teaching staff to teaching staff is also a matter for concern. In general secondary education this ratio i s as high as 0.7 for regular schools, but increases to 1.5 for pre-schools and between 2 and 3 for special schools according to the World Bank facility survey. `14 The large number o f non-teaching personnel i s generated by the normative that dictates the number o f staff that have to serve based on the current network o f facilities, often without a valid rationale or notion o f efficiency (as discussed below). 4.29 Underlying these striking inefficiencies are the norms contained in the orders and regulations of the education sector. The MoES issues established orders with mandatory guidelines, also called "norms," that mandate how budgets are to be formed at the school (or education facility) level interms of inputs such as staffing (teaching and non-teaching) and materials required and that are based on the existing oversized network o f education facilities. Most o f these norms mandate budget formation at the school level irrespective o f demand (enrolled children or school age population in the jurisdiction). They have their roots in the Soviet system o f planned economy and they have not changed, despite financing changes inthe Budget Code reformin2001-02. Box 4.3 contains some examples. ~ Box 4.3: Normsfor Education Facilities 1 1deputy school director for 11class groups 1 1 managing director if the village has more than 600 people (regardless the total number of children in the school) 1 1pedagogue per each 8 class groups (regardless ofthe total number of children inthe school) 1 1extracurricular activities group leader per each 16class groups (regardless o fthe total number of children inthe school) 1 1cleaner per 500 square meters (0.5 per each 250 square meters) 1 1coat roomattendant per 200 coat spots inschool 1yard keeper per each 1.5 hectaresineach school 1 Norms that allow very low teaching hours loadper teachers compared internationally 4.30 The situation is worsened by a Verhovna Rada Resolution that does not allow the closing of schools. Consequently, local governments are bound to maintain a large network o f service facilities despite a decreasing demand in education and underutilization. Merging i s allowed by the budget code and this would be an option, but it i s not being pursued as a policy to date (there are isolated cases o f mergers initiated by local governments). While merging i s more viable in larger cities, given agglomeration effects, it still feasible even in rural settings as shown by the experiences o f neighboring countries (including New EUmember states and several regions o f the Russian Federation). 4.31 Largely because of these systemic issues, the increase in educational expenditures has not translated into increases in their capital component or in quality enhancing current expenditures. The large and increasing share o f the budget spent on wages and utility expenditures leaves little room for other education-enhancing inputs such as textbooks and other goods and services (some o f these may be part o f current spending) and for capital inputs such as laboratories, computers, broad band internet connection and the like. Recurrent spending represents about 95 percent o f total education expenditures from the state budget. Allocations for instructional materials and teacher training, which are essential to ensure the quality o f learning, are thus severely limited. Moreover, investments in the maintenance and repair o f educational facilities and other capital outlays i s low and i s declining further owing to the pressures o f increasing recurrent spending. `I4 Estimationsbasedon World Bank education andhealthfacility survey. 92 E. FINANCINGANDADMINISTRATIVEISSUES FinancingIssues 4.32 Expenditure assignments in the education sector between different levels of government requirefurther adjustment. First, there is still some overlapping of the functions delegated among the levels of local government (see Table 4.4) and this i s particularly evident in the responsibilities of pre- schools and primary education beingundertakenin some cases by towns and villages, or by rayons, or by both."5 Second, there i s a loss of economies of scale (pre-schools andprimary education) inthe activities undertaken below the rayon level, with townships and villages, on average, being too small to allocate resources effectively and efficiently. Third, having this responsibility at the village level would be an obstacle to restructuringand rationalizing the network of facilities and staffing. Table 4.4: Expenditure Responsibilities in Education, by Local Government Level Level of Education Oblasts, ARC, and Rayons and cities Townships, settlements, cities of special status and villages Pre-school X X General secondary (primaryand secondary) X x (Primary) Vocational /vocational secondary X X 4.33 The factors determining financing needsfor education inthe equalization transfer formula do not create appropriate incentivesfor efficiency improvements. Education is just one of six areas o f "expenditure needs" estimation in the equalization grant. Its overall weight in this estimation i s 0.3. The equalization grant is distributed separately to oblasts, the ARC and cities of special status with one formula, and to rayons and cities with a different one, but in both cases the transfers are made directly from the Ministry of Finance. Using a very similar formula, rayons and cities distribute equalization grants to townships and villages. Key issues related to the education component of the equalization transfer are discussedbelow: 0 The sub-factor for pre-school does not provide incentives for actual enrollments. The financing formula for preschool education i s driven by the number of children aged 0-6 in the district and the centrally set theoretical enrollment ratio (which i s based on budget availability). Because this calculation does not depend on actual enrollments, it provides no incentive for local education managers to expand or even to maintain existing pre-school enrollments. Actual enrollment ratios infact have been significantly lower than notional enrollment ratios at the pre- school level, directly reflecting the lack of incentives to expand pres-school coverage.'l6One consequence of the lack o f incentives in the overall system for dealing with this issue i s that the unitcosts ofpre-schooleducation inUkraine arehighinrelationto other levels of education, and inrelationto unitcosts inother countries (see Table4.5). `I5This initiallytook placebecausethe earlydecentralizationprocess implied adelegation ofresponsibilitiesaccordingto the facilities ownedby different levelsof government-that is, facilities were decentralizedandnot true functions.Reformsinthe BudgetCode in2001helpedto improvethis situation,but work remainsto bedone. Volodymyr Voytov, "Educational ServicesFinancingfrom Local Budgets," inReform Strategyfor Education in Ukraine: Educational Policy Recommendations, Ministry of Educationand Science of Ukraineand UNDP, Kyiv, 2003. 93 Table 4.5: Unit Costs of Educationby Level,Ukraine and OECD Average, 2005 Sources: OECD data fromEducation at a Glance, 2006; State Statistics Committee; MoF, State Treasury; WB World Development Indicators; Bank staffcalculations 0 The sub-formula for general secondary education provides the wrong incentives in terms of class sizes. For example, it provides two-thirds more for students in smaller classes than the average. With such an arrangement, there are incentives to divide classes into smaller class sizes inorder to increasefinancing. Indeed, averageclass sizes are declining. 0 The adjustment coefficients for secondary schools for students with special needs may generate perverse incentives. The sub-formulas for special needs schools distinguish three categories with very different missions. The first category i s schools whichprovide housing and board for orphans; the second i s schools for students with physical or mental learning disabilities and for delinquent children who require specialized attention; the third category i s elite schools for athletes and military aspirants. An anomaly in the adjustment coefficients for the first two categories i s that the coefficient for boarding schools for orphans (who presumably do not otherwise require special care) i s higher than that for boarding schools for students with learning disabilities (who do require specialized care). A more basic concern is that the very large coefficients for institutionalized children provide an incentive to maintain children in institutions even when their condition would allow a transfer to mainstream programs. In support of this argument the number o f institutionalized children has declined over the past five years at 2.0 percent per year - less than halfthe rate of decline of that population age group (5.3 percent) over the same period. This imposes unnecessarybudgetary costs and arguably also hurts the prospects of these children for a smooth integration into the labor market later on. 0 The large adjustment coefficients for specialized schools for elite students and for athletes and military aspirants raise a question about the appropriateness of these programs. In most OECD countries, specialized programs of this nature are managed and financed separately fkom secondary education, and typically do not start until after compulsory education has been completed. 0 The unit costs of secondary vocational schooling in Ukraine are 70 percent higher than the costs of general secondary schooling. This difference i s much greater than the 14 percent average vocationallgeneral education differential inthe OECD countries (Table 4.5). The higher level o f budget financing for vocational education provides an incentive for the excessive vocationalization o f secondary schooling. Indeed, data from the State Statistics Committee recordthat vocational secondary enrollments contractedby only 1.6 percent per year over the past five years, while general secondary enrollments declined by 5.1 percent per year. The rising proportion of vocational education in total enrollments i s opposite to the trend in the EU-8 94 countries and i s all the more costly given concerns about the labor market relevance o f this education."' (Box 4.2 discussescontract education as ameans of financing highereducation.) AdministrationandAccountabilityIssues 4.34 In theory, local governments can hire and fire personnel, but in practice the process of firing teachers and non-teaching staff i s extremely difficult. As explained in Chapter 2 the process of firing i s cumbersome and follows a hierarchical path o f sectoral accountabilities. The Ukraine Social Facility Survey of Public EducationalInstitutions of General Education shows that, on questions of hiring and firing teachers, 85 percent of school principals reported that an upper line o f ministerial authorities (i-e., either the oblast or rayon education department or the Ministry of Education and Science) played an important or very important role inthis area. 4.35 Similar attributions in the management of schools are held by sectoral departments. These include setting fees for extra educational services, and admitting students. Salary levels are set at the central Ministry level. 4.36 Moreover, there is widespread uncertainty at the school level about the discretion of local school managers to actually manage schools. For example, according to the World Bank's Ukraine Social Facility Survey o f Public Educational Institutions of General Education, 17.5 percent of school principals reported that they had no input into the preparation o f their school budget, 48.9 percent reported that their budget was prepared based on the school's reported needs (that is, "norms"), 25 percent reported that they prepared a draft budget which was sent to the rayon for review, and 7.9 percent reported that they prepared a draft budget which was sent to the local government for review.'" Given the limitations imposed on school managers' autonomy and the confusionabout what decisions are within their authority, the incentives and accountability mechanisms for improved school management and administration are highly inefficient. F. OUT OFPOCKETEXPENDITURES,EQUITY,AND GOVERNANCEISSUES Profileof Out of PocketExpenditureson Education 4.37 Out of pocket payments (OOP) in education are prevalent in Ukraine. According to rough (and conservative) estimations based on the household surveys, OOP without counting higher edu~ation,"~may be around 0.7 percent of GDP. Based on facility surveys for general secondary education, this rough estimate i s around 0.8 percent of GDP, although both estimations are on the conservative side and may be underestimating the reality. While intheory, OOP are tightly circumscribed by the legislation, inpractice they take place at all levels and with different degrees of transparency. The first broad distinction that needs to be made i s between the OOP taking place from pre-school to general secondary education and the OOP taking place in higher education. The second distinction i s between formal or largely legalized fees and "informal" payments, although there are "in between" grey areas. Box 4.4 presents a broad brush treatment of formal and informal payments in the education sector in Ukraine. 'I7 "Fiscal Efficiency and VocationalEducationinthe EU8 Countries" the World Bank September 21,2006. `I8 Ukraine SocialFacility Survey ofPublicEducationalInstitutionsof GeneralEducation, Analytical Report,Annex 2; report commissionedby the World Bank andcarriedout by the Kyiv InternationalInstituteof SociologyJuly, 2006. ' I 9Higher OOP are roughlyestimatedat 0.75 percent of GDP. 95 I Box 4.4: A BroadBrushTreatment of FormalandInformalOut- of- Pocket-PaymentsinEducation I Inthe case ofpre-schoolto general secondary education, afew years ago the government undertook the positive of step o f formalizing certainpayments that were already required o f all parents. This formalization impliedthe creation o fthe so-called "charitable finds" managed by schools. These funds collect formal fees for certain textbooks and extracurricular activities and the materials they require (music lessons, language lessons andthe like). This modelprovides the potential for higher transparency but the experience hasbeenmixed. Some charitable h d s publicize and monitor the implementation o f the activities they implement,while others do not. Local governments currently lack a framework for the oversight o f these bodies. I'Onthe informal side, a highly prevalent means o fpayment i s through parents' committees which collect uniformpayments from students and provide themto the school principal for a variety o f items such as uniforms, donations for school renovations, and learning materials. Infact, this form o fpayment is largely unspecified and goes unrecordedinmost cases, providing little transparency and accountability. Another highly common mode o f informal payments (out o f several) i s one that takes place to obtain extra tutoring at schools. Inthe case ofhigher education, close to halfofthe studentsincolleges anduniversities pay their tuitionina "formal" manner and the other half o f the tuition i s financed through the budget. Inaddition, there are private expenditures for roomandboard. Onthe informal side, a variety o fpayments exist, the most substantial one being at the stage of admission, andthey are geared to several purposes: for example, to getting the "right" test interms of difficulty, to receiving a more positive revision, to entering the desired department. Most o f themare associated with cormpt behavior and with equity implications. There are also payments made to colleagues (levels Iand I1o fhigher education) to obtain additional tutoring andpreferential treatment intesting. The payments related to tutoring at all levels are a direct result o f low teacher salaries. " Kyivmunicipalityhasprepareddraftlegislationwhichwouldallowthe supervisionofthe activities ofcharitable trusts to prevent such abuses. 4.38 Unbundling and estimating the size of formal and informal OOP presents significant challenges. It is particularly difficult to come up with an accurate estimate for the following reasons. First,householdexpenditure surveys, the main source of data on OOP, donot make a clear-cut distinction between formal and informal payments in education; in fact, most people cannot tell the difference (as revealed by interviews). Second, as explained in Box 4.4, roughly half of the higher education students pay formal tuition and fees, and parents in general secondary education pay formal payments to the established charitable funds ineach school, but inthe household survey figures they are bundledtogether with informal payments for tutoring, admission to schools, and the like. Moreover, they are bundled as well with the contributions to the parents' committees that purchase uniforms or donate the resources to the school for repairs, often without oversight on how the resources are spent. Table 4.6 presents some figures basedon household budgetary surveys. 96 Table 4.6: A Profileof OOP Spending on Educationduringthe 2003/2004 Academic Year Teachingservices Equipment & overheads Living costs Uniform TotalI Extra Giftsto Textbooks Facility & Other Food& Overail Tuition Training staff & Materials Needs Clothing Lodging Primary andsecondaryschools 123 243 24 50 23 180 262 267 Specializedboarding schools 133 98 53 52 33 154 74 191 3, Gymnasiumsorlyceums 416 270 43 62 45 197 311 489 Boarding schools for children w. specialr 0 0 20 49 21 365 352 S a 33 SubTotalGENERAL EDUCATION 248 244 25 50 24 181 252 277 938 145 51 63 37 219 568 670 B .aM Universities (accred.leve1I11andIV) =g Technical schools (accred.leve1 I) Colleges (accred.levelI1) 1,476 243 162 95 59 265 900 1,429 2,116 326 132 91 as 295 883 2 'g SubTotalHIGHEREDUCATION 1,702 1735 269 103 81 53 258 784 1267 Primary andsecondaryschools 1% 4% 5% 19% 13% 57% 0% 100% Specializedboarding schools 1% 2% 9% 25% 9% 45% 9% 100% Gymnasiumsor lyceums 13% 6% 8% 16% 12% 42% 3% 100% Boarding schools for children w. specialr 0% 0% 1% 28% 10% 61% 0% 100% SubTotalGENERAL EDUCATION 1% 4% 6% 19% 13% 57% 1% 100% Technical schools(accred.levelI) 16% 1% 7% 15% 9% 36% 15% 100% Colleges (accred.levelI1) 28% 1% 10% 14% 4% 20% 23% 100% Universities (accred.level I11and IV) 36% 3% 6% 16% 2% 18% 19% 100% SubTotalHIGHEREDUCATION 27% 2% 7% 16% 5yo 26% 18% 100% Specializedboardingschools 207 1129 48.82 81.50 88.65 86.08 1.17 97.52 ._ Gymnasiumsor lyceums 4.26 6.38 31.91 51.06 3830 59.57 1489 89.36 Boarding schools for children w. specialr 28.57 17.99 64.02 88.89 88.36 84.66 7.94 98.41 SubTotalGENERAL EDUCATION 3.46 11-56 49.33 81.47 87.95 85.64 1.69 97.44 Technical schools (accred level I) 23.68 5.53 45 53 72.11 52.89 71.84 28.68 95.26 Colleges (accred.levelI1) 39.42 6.73 59 62 84.62 27.88 72.12 48.08 99.04 5 8 45.76 10.38 38.98 82.42 20.76 57.20 37.71 5 Universities (accred level I11and IV) 96.82 SubTotalHIGHER EDUCATION 36.30 8.05 43.83 78.56 34.31 64.64 35.25 96.44 Source: KatyaMaynzyuk,"Out-of-Pocket SpendingandGovernanceIssues inEducationFinancing," February, 2007. 4.39 The absolute size of parental contributionsrises considerablywith the levelof specialization in education services. Overall revenues from parents ingeneral secondary schools tend to be 4.87 times smaller than in gymnasiums and 6.46 times smaller than in lyceums. The size o f contributions also rises sharply with the number o f grades coveredby the schools. Issues of Concern 4.40 Householdexpenditures on education are sizable, and in some cases they show significant inequities. For example, in cities, household outlays for education in pre-school and general secondary education are about as high in the lowest quintile as in the highest quintile o f income (Table 4.7). As reported by households,'*' expenditures on education are for formal fees (private tuition, school uniforms, extracurricular activities, textbooks and other educational materials)l2l and also for informal payments relatedto admission to schools, tutoring, and other school activities. I 2 OInthe2005 HouseholdSurveyon HealthandEducation. 12' Although the government's intention is to providefree textbooks throughoutsecondaryeducation, budget financingis currentlyavailableto providefree textbooksonly ingrades 1through6. Inthe higher grades, studentsmust buy textbooks. 97 Table 4.7: TotalHouseholdExpendituresonEducationper Student,by IncomeQuintilesandArea of Residence( inUAH) Type of Settlement Q1 Q2 43 Q4 Q5 Total City 801.3 381.2 448 593 730.3 604.3 Town 314.8 378.6 350.3 555.2 730.5 430.3 Rural 293.6 438.3 514.7 657.4 1,089.00 399.5 Total 402.2 404.4 435.9 591.8 746.9 492.9 Source: DaryaNachinkina, "Education inUkraine:Illustrating SocialDimensionsofPovertyandInequality," backgroundpaper preparedfor the 2005/2006 World BankPovertyAssessment, 2006. Datafrom 2005 of the HBSHealthandEducationModule Sample Survey. 4.41 With little oversight and accountability, quality is not secured by the substantial informal payments that parents make. Lack of clarity and of institutionalized checks over performance and quality, particularly at the general secondary level, remain a window for rent-seeking. The resulting informal payments differ and are often masked under alleged "additional training" or "facility needs" or plain payments to obtain admission to desiredpublic schools (this i s more common inlarge urban areas), or to obtain the needed grades to move forward to other education levels (particularly for students that would like to pursuehigher education). Survey results show that informal payments ineducation are more likely to take place inUkraine than inmost countries inthe region. (See Figure4.10) Figure4.10: PercentofRespondentsThat ThinkUnofficialPaymentsAre Neededfor Education:Ukraine andComparator Groups 20 . 7 mAlways 18 I Sometimes 0Usually 16 g ::8 -I4 3 6 4 2 0 CIS-low Income CIS-middle South-Eastern EUstates member Ukraine Income Europe Source: EBRD-World Bank Life inTransitionSurvey (LITS)2007. 4.42 Access to pre-school for children from poor households is limited. Currently, pre-school education covers relatively few children from low-income households: the pre-school enrollment ratio for children from households in the lowest income quintile i s 30 percent versus enrollment ratios well over 50 percent for children from the highest three income quintiles (Table 4.8).'" There i s evidence in the education literature that pre-school education can play a very important remedial role in preparing children of high-risk groups (especially children from low-income households) for a more successful performance once they enter primary and secondary schooling. Budget financing of pre-school education for this purpose would be well justified, but the provision o f pre-school education would need to be fundamentally reorganized to make it more accessible for children from poorer households. Inthe case o f urban areas, the admission payments may be one of the significant barriers to access. "'See "Ukraine PovertyAssessment:Povertyand Inequality in SocialDimensions," World Bank, June, 2006.) 98 Table 4.8: EnrollmentRatio Estimates(percentof age group) basedonthe HBSHealthand Education ModuleSample Survey, byAge Group,Region,Income Quintile, andGender Age City Town Rural Q1 Q2 43 4 4 Q5 Male Female Total 3-5 62.6 59.5 26.3 30.6 38 58.2 67.1 78.4 49.4 44.8 47.3 6-9 96 94.8 91.8 90.9 94.4 96.6 96.4 95.6 92.9 95.5 94.1 10-14 99.9 99 98.7 98.3 99.4 99.8 99.4 100 99.1 99.3 99.2 15-17 95 90.9 88.8 89.8 90.4 94.1 94.4 93.6 91.3 92.7 91.9 18-25 36.5 20.9 19.1 28.3 23.2 28.3 29.7 28.2 26.1 29.4 27.7 Source: Background paper prepared for the 2005/2006 World BankPoverty Assessment, 2006. Datafrom 2005; HBS Health and Education Module Sample Survey, as reported inDaryaNachinkina, "Education inUkraine: Illustrating Social Dimensions of Poverty and Inequality," 4.43 Special schools such as lyceums and gymnasiums (for advanced students) generate more OOP spending and seemingly provide a better quality of education, but also create some equity concerns. These elite schools have developed rapidly during the past decade and will continue to do so as expressed by the authorities. They are popular with parents because they offer a more intensive preparation for university than conventional general secondary schools. Moreover, testimonials from teachers and school directors, and evidence collected duringstaff visits, have confirmed that the quality of inputs,including facilities, laboratories, teaching aids, teacher training, and the like, are far better inthese institutions than in the mainstream system. This in itself is positive, as it provides a structure for advancing more gifted students. At the same time, there are questions about who can access these schools. Household survey evidence shows that these schools require more OOP spending than do the mainstream general secondary education schools. In larger cities this seems even more acute: for example, anecdotal evidence indicates that access to some Kyiv lyceums requires an initial payment o f more than US$l,OOO, which would be equal to six average monthly wages in the country. Higher OOP in itself could be a barrier for gifted students from poorer households, even assuming that the rest o f the process of selection i s completely fair. 4.44 The increased reliance on financing from students and parents has contributed to the emergence of sizable differences in education quality. Reliance on parental contributions and other non-budget revenues often creates perverse incentives, such as the incentive for teachers not to cover the complete curriculum in class in order to create a demand for paid, extra tutorial instruction, and the incentive for production activities in vocational schools and service provision in general secondary schools (such as the offering o f computer classes to the community) to displace educational activities. Although it i s difficult to document, corruption in the form o f the solicitation o f informal payments for better examination scores and for admission to university programs also has become (and remains) a serious concern inUkraine, as it has inother countries inthe region. G. QUALITYISSUES 4.45 Quality control in education is hampered by the lack of an acceptable and objective performance standard. Measuring learning achievement was a priority for the educational authorities untilrecently. The existing universityentry examination varies significantly across universities and even within a university in the degree of difficulty o f the tested subjects. More recently, the Ministry of Education has made a laudable effort to implement independent student assessments, including through the full roll-out in 2007 of the Trends in International Mathematics and Science Study (TIMSS), for which Ukraine signed up in 2005. This represents a critical milestone in establishing a baseline for performance and quality measurement in education in the country. An objective performance measure serves two key purposes: (i) will raise awareness o f the current quality gap existing between Ukraine's it education performance and the highest world standards; and (ii) it will provide a baseline against which 99 the effectiveness o f future reforms can be evaluated and thus will be a critical component o f a move from an input-orientedto a more output-oriented financing and management system inthe sector. 4.46 Nevertheless, at present there are signals (although there i s no firm evidence) that education quality is declining. There i s indirect evidence o f increasing dispersion in the quality o f the learning environment as a result o f increasedreliance on non-budget financing: 0 Many important educational inputs, such as educational equipment and materials, programs for poor students and at-risk students, and school maintenance, are usually locally financed, which leads to significant disparities in the quality and coverage o f education. This i s particularly so because these needs are greatest in those localities with the lowest incomes, revenues, and educational performance. It i s important to ensure adequate financing for these expenditures, preferably through a central financing instrument that targets the schools and localities with the greatest needs. 0 Teacher training, which i s inprinciple included in central transfers, i s inreality not provided to any significant degree. Failure to fund teacher training i s a high-risk and low-efficiency option because it means that an educational system's most important assets - its teachers - are not equipped to perform their work effectively, especially in introducing the important innovationsplanned under ongoing reforms. 0 There are no incentives for teachers to improve or even maintain education quality in the form that matters most: student learning achievement. 4.47 Moreover, in comparison with other countries in the region there appears to be an unusually high level of dissatisfaction with the quality of the provision of education in Ukraine. While Figure 4.11 does not provide any solid evidence regarding the quality o f education or the changes in this quality over time, it shows that the level o f dissatisfaction in Ukraine is higher than in other regional groups o f countries, including low-income CIS c~untries.'~~ This situation should be o f concern, since other countries inthe regions face similar or higher constraints interms o f resources andbudget. '23Ukraine is also part of the average numbers that appear for middleincome CIS countries. 100 Figure4.11:PercentofRespondentsDissatisfiedwith PublicEducation:Ukraineand Comparator Groups 30 ElVery unsatisfied .IUnsatisfied i 3 2o 10 0 E 'Z ource: EBRD-WorldBank Life in TransitionSurvey (LITS) 2007 H. CONCLUSIONSAND RECOMMENDEDACTIONS 4.48 Ukrainedoes not needto spend more on public educationbutit needsto spendbetter.Major improvements in efficiency in the provision of education are needed, and they have the potential of creating the necessaryfiscal space, withinthe same education budget envelope, for re-allocating resources towards quality-enhancing inputs and investments. These inputs and investments include better teaching aids, the training and re-training o f fewer, but better paid, teachers and professors at all levels, needed equipment (laboratories, computers, and the like), andinternet access for all schools. 4.49 The existinginefficiency and excess capacity resultfrom an inheritedoversized staffing and network system that is perpetuated by regulations and incentives present in the budgeting, financing, and administrative mechanisms of local governments and the education system. According to the Ukrainian legal framework, local governments are meant to be accountable to the local constituents for managing public education (before higher education) effectively. But because the Ministry o f Education and Science (MoES) retains control over most o f the decision-making, local governments are incapacitated to perform the needed adjustments. The MoES continues to retain full control over the norms governing staffing arrangements, teaching hours, non-teaching staffing ratios, and class sizes-all basedon the oversized school network all dictating the way inwhich budget are formed at the school level and, consequently, at the local government level. The current arrangements undermine local accountability, hamper the opportunity of allocating resources to improve quality, and come at a highcost (low value) for the country. The constraints embedded inthe systemmake it difficult for local governments to take actions to improve efficiency, such as school consolidation and the re-allocation and restructuringo f staffing. 4.50 A simple estimation suggests that if, over the medium term, Ukraine could attain a level of studendteacher ratio for pre-school, primary and secondary education comparable to the OECD country average, it could save approximately 0.7 percent of GDP per annum.'24 The savings could reach up to 1 percent of GDP per annum if the merging of facilities also took place and with it a reduction in utility lZ4 Estimationsare based on unit costsandaggregate cost structure ofpre-school,primary, and secondarylevels. 101 expenses and in non-teaching staff. In addition, significant savings could bemade in higher education with proper targeting. While this is merely a benchmark, gradual adjustments towards that objective would allow enough additional resources to gradually meet the education investmentneeds estimated in thisreport. 4.5 1 The following selected recommendations are geared to achieving the goals stated above through greater financial and administrative efficiency inthe provision of education. 0 Revise and reformulate the norms dictated by the Ministry of Education for budget formation at the facility level. The elimination o f staffing norms for teaching staff (based on inefficient levels of teacher hours provided by each teacher) and for non-teaching staff (mainly based on the physicalmeasures of each school) would provide the neededflexibility at the facility and the local government level. There are a number of MoES orders and resolutions that need to be revised, including Order N!?1/9-234 dated 19.06.2001, and the related VR resolution N255 1- 111. The guiding principles in the revision of these norms should be the elimination or reformulation of norms that impose staffing rigidities based on physical measures o f schools (for example for managerial and non-teaching positions, as well as teaching positions) and the reformulation of norms basedon inefficient levels o f teaching hours renderedby each teacher and inefficient class sizes. Ifthe government desires to replace the norms, then the right vehicle is the establishment of purely demand driven factors as a guideline. These factors could evolve by pricing unit costs of each service in line with the H factor of the equalization formula, that is, a minimumstandardper unitof service (inUAHterms). Give local governments discretion in the closure of under-utilized education facilities. In the meantime, while legislative obstacles are lifted, local governments should proceed with the merging of schools. In parallel, the central government and local governments would need to continue investing in buses to transport children to and from schools. While there i s some experience in Ukraine in this area, this policy should be pursued comprehensively. Immediate efficiency gains can be expected from a reduction in non-teaching staff (and in some cases teaching staff), in current spending in heating, and in expenditures on inputs to maintain the regular operation o f schools. Allocate decisions on the hiring and firing of teaching and non-teaching staff to the corresponding level of local government handling the function. This recommendation implies the lifting of all of the current central Ministry hurdles for firing personnel. Rayon and city governments should also be allowed (without unnecessary obstacles) to re-allocate personnel across their jurisdictions. Use the naturalprocess of attrition to help the rationalization process, but without re-hiring retired personnel and minimizing new hiring.New hiring should be undertaken exclusively to meet urgentknowledge gaps (for example, inthe Information Technology field). 0 Streamline the education component of the equalizationtransfer around the basic principle of financing pure demand for education services while encouraging enrollments. The following more specific recommendations complement those provided inChapter 2: o Remove from the formula all coefficients that adjust per student allocations linked to the existing network of schools. A fundamental problem i s that the current financing formula for general secondary education contains a large number o f adjustment coefficients for schools depending on the number and type o f school facilities, favoring higherfundingfor larger networks of facilities at the local level. This sends the wrong signal to local governments, discourages reforms in service provision, and perpetuates the inefficiencies o f the system. 102 o Remove from the formula all adjustment coefficients related to class sizes. The formula provides two-thirds more financing per student to local governments with schools that have lower minimum average class sizes than to local governments with schools with larger average class sizes. This clearly creates negative incentives. In a few cases, this provision reflects the higher costs faced by schools in small, rural communities where higher per student costs are likely-but the formula already contains coefficients to account for rural and mountainous areas. Even schools with only a few children at each grade can reduce costs without compromising learning achievement by consolidating two or more grades under a single teacher. 125 Similarly, regions with several small schools can consolidate schools, and bus students to central schools. The need for school and class consolidation will grow as school age cohorts continue to shrink while per student real unit costs will continue to grow.126 o Allow the formula to provide the same per student financing regardless of the mix of students across schools, includingspecial types of schools and vocational schools (except for schools for disabled or disadvantaged children). This will help to remove the distortions, inefficiencies, and inequities that arise from the current formula and will create powerful incentives for cost-effective management at the local level. o Introduce actual enrollment as a factor in the formula for pre-school education(in the same way as it is for general secondary education) in order to expand coverage. Moreover, given the acute problem o f pre-schools with regard to highunit costs, the formula could also include incentives for consolidation and even for the targeting o f children from vulnerable households. o Apply a pure per student allocation below the rayon level. Applying this funding in the formula from rayons to villages and settlements would leadto the consolidation of small rural schools which do not have the option of reducing costs by consolidating classes. For larger schools (incities) it wouldprovide an incentive for school consolidation. 0 Start efficiency enhancing reforms immediately: If the problem of inefficiency and overstaffing continues to grow at the present pace and no action is taken (particularly in view of the smaller lower cohorts of student-age population expected over the next years), a gradual response (as recommended inthis report) would not bepossible inthe future, and larger scale and harder changeswould beneeded at a highcost for the government and the population affected. 4.52 The following selected recommendations are geared to meetingthe challenges posed by equity, governance and quality problems (through expenditure targeting and other mechanisms): 0 Provide, over time, performance incentives. Under close control of Treasury, local governments at the oblast and rayon and city levels should be able to pay in addition to wages, bonuses or location or subject premiums as incentives for perf0rman~e.l~~ lZ5These actions would not necessarily compromise quality. There i s abundant international evidence that multigrade teaching with the right teacher preparation and support and the right learningmaterials for students andteachers can produce student learning achievement as least as highas conventional single teacherkingle grade teaching. And class sizes well above the maximum class sizes inUkrainehave been found not tojeopardize learning achievement. See, for example, Eric Hanushek and Javier Luque, "Efficiency and Equityin Schools Around the World, " in TheEconomics of Education Review, Volume 22,2003. lZ6Itshould be noted that higher class sizes as proposed in this chapter do not runcounter to quality within reasonable thresholds. For example, Korea, with average class sizes close to 40, that is four times larger thanthose ofUkraine, consistently performs at the top o f the PISA and TIMSS internationalstudent assessments (see OECD 2006) '*'The current salary scale for teachers rewards teachers primarily according to formal qualifications and length o f service. To the extent that it rewards teaching excellence, it does so in recognition o f the few highest-performing students in national competitions. This formula needs to be adjusted to also provide incentives for teachers to focus on what all their students are 103 0 Improve the targeting of budgetary expenditure for tuition costs and stipends in higher education by limitingthem to students from poor households (who also attain high scores in standardized entry tests-see below). 0 In parallel, accelerate the introduction of standardized tests across the country to be used for application to higher education and to measure overall performance after secondary education. And consider letting the standardized test be administered externally by a thirdparty entity incoordination with the MoES. This would bringfurther transparency to the system. Inany case, the MoES wouldhave the role o f overseeing and guidingthe process. 0 Consider, on a pilotbasis, fully attaching part of the budget resources for higher education to the studentsthemselves(for selected "budget financed" students). This mighthelp create a market-like environment where universities could compete intensively to attract these students (and their funding) though better programs and better skills fitting with labor market demand. 0 Strengthen the monitoring over admissions to gymnasiums and lyceums. This is essential to help curving governance concerns and ensure fairness for the children o f low income households. Strengthen the regulations regarding "extracurricular" classes and activities in schools. Children from poorer families are not able to afford the payments solicited for language training, special tutoring, and other school activities. More stringent and clear-cut regulations regarding what i s and what i s not admitted as a formal fee are neededto improve transparency. Monitor outcomes and outputs rather than inputs. The performance o f general secondary education should be monitored by the assessment o f student learning achievement through standardized tests. The performance of vocational secondary education should be monitored by both learning achievement and labor-market outcomes (i.e., employment, earnings, and job mobility). learning(in additionto preparingtop students only). Revision of the evaluation criteria and compensation formula needs to be carried out within a comprehensive professional development program for the teaching profession. A number or OECD countries, includingEnglandandScotland, have suchsystems. 104 5. STRENGTHENINGCAPITALBUDGETINGINLOCAL GOVERNMENTS A. INTRODUCTION 5.1 Local governments have a large role in infrastructure investments in Ukraine. One-third o f the total capital expenditures are made through local budgets, and more than half o f the resources intended exclusively for (fixed) capital investments are spent at the local level. Investments in the housing and communal sector (which includes heating, water, and sanitation) and in the health care and education sectors are made mostly through local budgets. In addition, other important sectors such as transport (mainly roads), social assistance, and agriculture also depend significantly on local government budgets. Moreover, investment needs at the local level are estimated to be large over the next 10 years (on the order o f US$25 billion) (See section B). 5.2 Creating fiscal space for national and local investments across sectors, as recommended in the first installment o f the Public Finance Review (PFR I) in this report, i s a necessity, but such efforts and would be compromised in the absence o f a strong framework o f capital budgeting: that is, a framework that would make it possible to plan, evaluate, select, allocate, implement and maintain investments efficiently and coherently with the growth strategy o f the government. Indeed, if such a framework i s not in place, the resources saved and re-allocated might be wasted in less productive capital expenditures. Building upon the work undertaken at the national level (in PFR I), report examines the existing this mechanisms and processes o f capitalbudgeting at the local level. 5.3 Inthe present chapter, section B provides the context for the capital budgeting system, while sections C and D discuss general trends and also issues inlocal capital budgeting. Recommendations and options for reformthat are consistent withreforms at the national level are offered inthe final section. B. CONTEXT 5.4 Ukraine has a capital budgeting system at the national level that presents some good features but also has significant weaknesses. Among the positive features is the fact that the line ministries and their programs are the unit o f analysis for capital spending and select their projects within their annual budget ceilings. The core weaknesses include the following: (i) a short-term perspective on budget ceilings and capital spending strategy (1 year-budget cycle), which affects planning significantly; (ii) poor economic evaluation, screening, and selection of projects, without a strong cost-benefit analysis framework; (iii) a problematic (and sometimes non-transparent) project selection process, whereby new projects, sometimes without an economic analysis, are placed directly at the end o f the pipeline, taking preference in funding over ongoinghnfinished projects; (iv) weak coordination betweenthe Ministry o f Finance (MoF) and the Ministryo f Economy (MoE), with unclear boundaries and functions inthe capital budgeting process; (v) weak control and monitoring o f current assets and their conditions, which hampers the planning o f proper maintenance and asset replacement spending; (vi) a poor record o f capital projects completion (low execution rates) which stems from several o f the weaknesses pointed out earlier; and (vii) lack o f a 105 coherent strategy for public investments that i s linked to an economic growth strategy for the country. Recommendations of the PFR-Ifor capitalbudgetingat the national level are giveninBox 5.1. Box 5.1: Key Recommendationsof the PFR-Ion CapitalBudgetingat the NationalLevel Introduce multi-year ceilings for total spending (capital andrecurrent). Introduce multi-year sector strategiedplanning within resource ceilings inall line ministries and agencies. Integrate the capital budgeting process into the overall budget process. Integrate donor financed investments (and assistance) into the core capital budgeting process (and into the overall budget process- on a multi-year basis) by streamlining the complex procedures o f project processing and approval inthe government, andby synchronizing project preparationwith the budget cycle. Establish a formal structure o f collaboration between the MoF and the MoE to review all aspects o f capital budgeting, fiom identification and evaluation through completion and monitoring. Strengthen project evaluation capabilities inthe MoF and the MoE, andparticularly inline ministries. The MoE should take the lead inthis process. , Apply cost-benefit analysis consistently as part of the formal process o f project evaluation and selection, and implement transparent and standard methodologies o f evaluation. Strengthen the analytical capabilities for evaluating potential fiscal risks in all types o f public private partnerships andprivatization schemes. Establish a databaseo f current assets and their condition. Strengthen coordination between the MoE and the MoF to closely monitor capital project implementation during the year with line ministries and to identify and resolve any bottlenecks, and assure enough funding to complete projects. Review the current budget execution procedures for capital projects to assess their adequacy and to ensure that they are not contributing to the under-executiono f capital spending. Improve the monitoringand assessmento f operations and maintenance. The MoE and MoF incoordination (and together with line ministries) should monitor information on asset conditions and should feed their findings back into the recommended assets database. Perform annual audits o f internal control in line ministries to assure that they support capital project management and completion, particularly for high risk ministries. Sources: World Bank (2006), "Ukraine Fiscal Space for Growth: Public Finance Review" (PFR I). 5.5 The Budget Code 2001-02 reforms, and further legislation changes in 2002, established a new framework for budgetary practices, including those for the capital budgeting system. However, this legislation left a number of gaps regarding capital budgeting at the local level, not only in terms o f the allocation o f capital expenditures from the central government (through capital subventions/transfers, discussed in detail in Chapter 2) but also in terms of the processes for the planning, evaluation and selection, implementation, and monitoring and maintenance of capital projects. Moreover, despite these reforms and some shifts inpower over capital projects between the different levels o f local government, the capital budgetingsystemremainedhijackedbypoliticalneeds. 5.6 The main financing sources for capital expenditures at the local level include: (i) subventions from the central budget to local budgets for investment projects (that is, capital subventions/transfers to local governments as discussed in Chapter 2); (ii)subventions for agencies and enterprises, including those for "targeted and (iii) other local government revenue sources, including "own" local 12'"Targeted programs" are discussedinsection D ofthis chapter. 106 taxes and fees, receipts from the sale o f land, environmental fees, and other minor revenue sources. The first two sources dominate the overall structure o f capital investments at the local level. C. GENERALTRENDSINLOCALCAPITALSPENDING 5.7 In 2006, local budget capital expenditures recovered after a dip in 2005. Total local capital expenditures, as classified in the Ukrainian budget, represented just above 2.6 percent o f GDP in 2006 (see Figure 5.1). However, as in the case o f the national consolidated budget, capital expenditures are comprised o f (fixed) capital investments and "transfers to enterprises and government agencies." Under the latter category, close to two-thirds are subsidies to state owned enterprises and specific transfers to spending units o f the government (i.e., Ministry o f Agriculture, Ministry o f Energy) to support programs that provide subsidies to various sectors. This leaves public investments in fixed capital at the local level at close to 1.4 percent o f GDP. Figure 5.1: Evolution of LocalBudget Capital Expenditures as Percent of GDP (breakdown by (fixed) capital investments and "transfers to enterprises and government agencies," 2001-06 - - 3.0% 1 2.5% I to Transfers agencies 2.0% and SOEs 1.5% 1.O% 0.5% Investment 0.0% 2001 2002 2003 2004 2005 2006 ources: Ministry o f Finance; State Treasury o f Ukraine; Bank staff calculations. 5.8 The category "transfers to enterprises and agencies" at the local level, inparallel with the trends at the national level, has increased significantly inreal terms and as a share o f local capital expenditures. As a percent o f GDP it went from 0.24 in 2002 to 1.3 in 2006 (see Figure 5.1). Currently, "transfers to enterprises and agencies" at the local level represents close to 52 percent o f the total spending in this category in the consolidated budget. Within local budgets, this item went from being less than 20 percent o f the share o f total local capital expenditure in2002 to close to 52 percent in2006 (see Figure 5.2). Figure 5.2: Share of "Transfers to Enterprises and Agencies" inLocal Capital Expenditures 2002-06 I to Transfers aC)enGlesend (Fixed) C-pitel lnveatment Sources: Ministry of Finance; State Treasury of Ukraine; Bank staff calculations 107 5.9 The share o f local capital expenditures in total capital expenditures in the consolidated budget (net o f "transfers to enterprises and agencies") increased considerably over the last five years (see Figure 5.3). This was due mainly to a significant overall reduction in (fixed) capital investment spending for national programs in 2005, and to a recovery in total capital expenditures in 2006, which, while largely drivenby "capital transfers to enterprises and agencies," also experienced growth inthe "fixed" portion. Figure 5.3: The Growing Share of Local Budgets in Total (fixed) Capital Investments, 2002-06 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0Yo 2002 2003 2004 2005 2006 ources: Ministryo f Finance; State Treasury o f Ukraine; Bank staff calculations. 5.10 Oblasts spend close to 73 percent o f the total resources allocated for fixed capital investments at the local level, although their share has been slowly diminishing(Figure 5.4). The shares o f rayon and city governments have remained relatively stable since 2002, but township and village budgets for capital investments have grown marginally since 2002. Nevertheless, it i s important to note that oblast budgets often finance investment projects at the rayon and city levels, and sometimes even below rayon level. Figure 5.4: (Fixed) Capital Investments by Level of LocalGovernment, 2002-05 loo-?& =Townships. SO-/- setllements. and villages BO?& 70% 0Rayon 60% 50e/- IOblast 40-/- 30-h 20% lo-?& 00% 2002 2003 2004 2005 Sources: Ministry o f Finance; State Treasury o f Ukraine; Bank staff calculations. 5.11 Moving forward, this report roughly estimates that the investment needs o f local governments over the next 10 years (2008-2017) may reach US$29 billion (see Table 5.1).'29 District heating i s probably the service with the most pressing needs interms o f physical investments, particularly to update its equipment and distribution network, which i s not only outdated but i s also energy inefficient. Water 12'These figures are merely a rough estimate o f the potential public investment needed at the local level and are expected to be revisedjointly with the government through the process o f discussion and dissemination o f this report. 108 and sanitation services and regional and rural roads (and other transport infrastructure) also require significant refurbishing to avoid further deterioration. But investment needs are also significant in social services. For example, in education they include the refurbishing o f deteriorated schools, the purchasing o f new equipment, laboratories, and the like. The investment needs in the health sector include the refurbishing o f deteriorated health facilities at all levels o f government and, mainly, the replacement o f outdated medical equipment. Table 5.1: Investment Needs at the Local Level In Sector Billions Description 1 of US$ Housing and 14 Investment and rehabilitation needs of water supply and sanitation are USS7 billion Communal through 2012. Water supply investments require 60% o fthat amount, and the Services (includes remaining 40% is needed for the rehabilitation o f sanitation systems. An additional housing, heating, US$7billion is needed for district heatingcompanies andretrofittingactivities on the water, and demand side. sanitation The estimates for health and education are based on public capital investment in Social Protection these sectors as a share o f GDP made by new EUmember countries and then multiplied by the percent o f investment spending inthe sector at the local level. This estimate mostly consists of road construction and repair over the next 10years and rural roads based on PFR Iestimates (assuming that about one-third o fpublic capital investment i s made locally). Other 7 Including investment in agriculture and land reform, local energy projects, local transport projects (including regional airports), waste management, and other local infrastructure 5.12 At the local level, the private sector is likely to invest only a small portion of the total needs, perhaps in some areas o f housing and communal services such as district heating, and that only if the regulatory framework for sub-national public-private partnerships (PPP) improves significantly. Thus, even considering some private investments, local government spending in (fixed) capital investments would needto grow inreal terms by at least 70-90 percent from its current level. D. ISSUESINLOCALCAPITALBUDGETING InvestmentPlanning 5.13 Planning is undertaken on an annual basis and with little medium-term perspective. The major sources o f capital investments (that is, intergovernmental capital subventions/transfers for local investments, allocations for targeted programs and the oblast own funded "development budget") are decided inthe annual budget process. Thus, local governments have a very short time frame for planning, which i s coupled with significant uncertainties about funding levels (as discussed later). Staff visits to local governments (particularly city governments) evidenced that some basic attempts have been made to create a framework for investmentplanning. Butwithout a well-established medium-term framework, this task limits itself to setting lists o f projects with certain urgency/priority levels, which may be changed later because o f annual allocations or political considerations. 5.14 There is a lack of predictability in the allocation of capital investments to local governments,which severely hampers any planning attempts. As Figure 5.5 shows, there are abrupt 109 changes from year to year in capital investment allocations across oblasts (calculated on a per capita basis). This situation handicaps any planning efforts of local governments. This pervasive feature arises from the lack of a proper framework for intergovernmental capital subventiondtransfers for local governments that is formula-based and stable through time (Chapter 2 provides more details on this issue, as well as specific recommendations). Figure5.5: Year-to-Year Changeinper CapitaCapitalAcquisitionby Oblasts (consolidated), 2002-05 1 --- - 2002-2003 -C 2003-3004 2004-2005 1 I I Project Evaluation, Selection, and Allocation through Co-financing Schemes Project Evaluation and Selection 5.15 The project evaluation process lacks any consistent and solid economic framework. This process reflects similar weaknesses at the national level. Local branches of the line ministries commonly prepare a list of projects and their descriptions which are reviewed by higher levels and consolidated by the oblast branches of the line ministries. Then, each line branch at the oblast level, incoordination with the executive (typically, the Department of Economy) screens the consolidated lists and reassesses the priorities, but does this without having a solid economic basis for decidingbetweenprojects. A large part o f the problem comes from the lack of a proper framework for project evaluation on which local governments could base their work. 5.16 The selection of capital projects i s highly discretionary. Once the lists o f projects are consolidated and reassessedinterms o f priority by the oblast's Department of Economy, the lists are sent for approval to the oblast council. The only evaluation and selection principle, used first by the executive (and at all levels o f local government) and then by the local council, i s the loose criterion that ``new projects should improve or not decrease access of thepopulation to social services. Inthe absence of a 'I cost-benefit analysis framework, comparison among projects i s inpractice fully discretionary. 5.17 In theory, resolutions from the Cabinet of Ministers on capital allocations from the state budget should give preference to funding the completion of unfinished projects over new ones; in practice, this i s not the case. In fact, while there i s a long-standing Cabinet resolution that supports the principle o f favoring ongoing projects, every year, during the budgetprocess, there i s a Cabinet resolution 110 dealing with those capital allocations (through subventions/transfers or other mechanisms) that, in practice, supersedes that principle. With each new resolution the rules change, and often the selection of projects for subventions (capital transfers) i s based either on lists sent by oblasts or on suggestions made by membersof Parliament that do not provide preferentialtreatment to ongoing projects (see Chapter 2). KRU(the General Auditor Office), inthe report Use of Subventionsfrom StateBudget toLocal Budgets, points out that in 2006 "close to 20 percent of the funds were allocated without complying with the principle of favoring unfinished construction." (Box 5.2 highlightsthe opportunity the Euro Cup 2012 to be held in Ukraine and Poland offers strengthening the capital budget framework and the national and local levels.) Box 5.2: The Euro Cup 2012: A Challengeand an Opportunityto Strengthenthe Frameworkof Capital Budgeting The organization o f the Euro Cup 2012 inUkraine presents a number o f challenges, some o f them related to the investment side of the budget. First, Ukraine's budget will face the pressure o f increases in infrastructure spending in order to finance (basic) needed transport related investments, local services infrastructure, and facilities for this event. On the basis o f the figures the government o f Ukraine i s handling (on the conservative side), and aside from the investments that are expected to be undertaken by the private sector, the budget would need to increase fixed capital investments by about 0.9-1 percent o f GDP (or 35 to 40 percent above the current level) each year from 2008 to 2011. The handling o f the financing part coming from the budget will also imply trade-offs in terms o f reducing non-productive spending elsewhere in the budget if a higher tax burden i s to be avoided (good candidates for curtailment are the existing inefficient subsidies, poorly targeted social spending, and other government consumption and current transfers-as discussed in PFR I). Second, managing this process under time pressure and conflicting demands requires careful strategic prioritization and discipline in allocating public resources effectively and efficiently This situation also presents an opportunity. Giventhe pressures to allocate the limited resources inthe most productive way, the government could use the momentum to: (1) strengthen the capital budgeting process at the national and local levels (PFR Iand the present report provide specific recommendations); (ii) create a criteria based, stable mechanism to allocate capital transfers within the intergovernmental framework (see Chapter 2); and (iii) reform the service delivery platform across sectors and build capacity for strategic planning and monitoring and evaluation (M&E) at the local level. Indoing this the government should avoid another discretionary "targeted program" as well as ad hoc capital subventions. I I Allocations through Co-Financing Schemes 5.18 A significant portion of the capital investment allocations (mainly made through subventions/capital transfers to local governments) are made on the basis of co-financing, which has unintended consequencesfor regional equity. Generally, schemes such as co-financing are quite positive as they encourage ownership and project sustainability at the local level. In Ukraine, however, given that local budgets are heavily hijacked by recurrent spending, they have little left to contribute to their share of the co-financing arrangements. Wealthier regions tend to have greater access to these resources, which helps them to capture the co-financing resources, which in tum creates regional inequalities. Moreover, local governments in the second tier, such as rayons and smaller cities,13' and third tier local governments (townships and villages) lack the fiscal capacity to co-finance capital Except larger and wealthier city governments. 111 investments as required by the Cabinet o f Ministers resolutions. In the end, local governments in the second and third tiers receive some o f those resources via oblast allocations but they are fully discretionary and are subject to political bargaining. 5.19 Similarly, transfers to enterprises and agencies are higher in per capita terms in more industrialized oblasts (see Figure 5.6). This is to some extent expected, since these transfers are mainly geared to support state-owned enterprises and government agencies that support subsidy programs, and the larger state-owned enterprises are locatedmainly inindustrialized oblasts. Figure5.6 Tranfersto Enterprisesand Agencies by Oblasts, 2005 80 70 60 50 40 30 20 10 ote: Oblasts, ARC and cities o f special status are ordered fromhigher to lower per capita regional (oblast) value added (left to right). Sources: MoF; State Statistics Committee o f Ukraine; Bank staff calculations. ProjectImplementation 5.20 Annual under-executionof capital spending in consolidatedoblast budgets is not as low as at the national level, but it remains a problem in local budgets(see Figure 5.7). However, at the city and rayon level under-execution i s in some cases extremely high (see Figure 5.8). This may indicate middle o f the year re-allocations within the oblast budgets. A look across rayons provides no evidence that wealthier jurisdictions have lower (or higher) under-execution rates than poorer ones, which indicates that this i s a widespread problem that i s not necessarily linked to the level o f funding from the central government or to a below average fiscal capacity incertain local government^.'^^ '" The same could be said about consolidated budgets for 2006, except for the case o f Donetsk, a region that executed 400 percent o f its planned construction plannedbudget in 2006. 112 igure 5.7: ExecutedCapitalConstructionProjectsby Oblast as Percentof Allocated Funding 1000% I Construction BRenovation 100% 10% Sources: MoF, State Statistics Committee o f Ukraine; World Bank staffcalculations Figure5.8: ExecutedCapitalConstructionProjectsby Rayons(and Cities) as Percentof Allocated 'unding,2005. 100 90 80 70 60 50 40 30 / Rayons and cities are ordered from higher income to lower income (from left to right). This study used PIT collections per capita as proxy for income level by rayon. Sources: MoF, State Statistics Committee o f Ukraine; World Bank staffcalculations. 113 5.21 According to KRU audit reports under-execution is more acute in social infrastructure and communal services. Thus, while roads and large transport infrastructure projects show higher execution rates (commonly above 80 and 90 percent), social services and communal infrastructure show considerably lower-and alarming-spending execution rates o f between 7 and 50 percent. Other infrastructure projects are above 50 percent. 5.22 Under-execution of capital projects i s likely to stem from a number of issues (several of which are discussedinthis chapter), includingthe following: 0 The absence of multi-year budgeting and multi-year strategy setting for capital investments at the local (and national) level, which hampers proper planningbeyond each annual cycle. 0 A lack o fpredictabilityinthe level o ffundinginany given fiscal year. 0 New projects taking preference each year over ongoing projects. This stems mainly from the lack o f a consistent framework for project selection and from political pressures to introduce new projects each year. For example, during interviews local government officials at the oblast level argued that up to 50 percent o f the projects might clearly be under-funded, but that they are included for political reasons. 0 Delays in capital project funding during the fiscal year. Based on interviews with local government officials and also according to the KRU's report o f 2006 On Use of Subventionsfrom State Budget to Local Budgets,13* funds from the state budget usually arrive towards the second half o f the fiscal year, which does not allow enough time to spend them within the fiscal year. 0 Different sources o f funding for a single project. For example, this funding could come from subventionshapital transfers, local own funding, and even targeted programs (the latter are discussed in detail later inthis chapter), and often one o f the sources may face under-funding for the fiscal year or face a delay inreceiving the fundingavailable. MaintenanceandAsset Condition 5.23 Planning for maintenance purposes is not undertaken through a structured process. Since there i s no specific budget item (account number) that accommodates maintenance and routine repairs as current spending, these activities are not formally funded and are performed on an ad hoc basis. Moreover, routine repairs involving the purchase o f spare parts (that are periodically changed) are often charged to the capital expenditure account o f renovations (see Box 5.3 for a guide to distinguishing maintenance from renovations). This creates an unstructured process for dealing with this critical type o f spending and may lead to higher renovation spending (and replacement investments) over time due to poor maintenance spending and planning. 5.24 There are no data on current assets and their physical condition at the regional level. This situation makes the process o f allocating funding for capital projects much less efficient, as it may generate under-spending on maintenance. The planning consequences o f this lack o f information can be indeed damaging for the processes o f planning and allocating resources, and can increase waste from unproductivereplacement investments. 132KRUweb-site, http://www.dkrs,gov,ua/kru. 114 Box 5.3: Distinguishing Operational Maintenance and RoutineRepairs from Capital Renovations The distinction between maintenance and routine repairs (which constitute a recurrent expenditure) and major renovations and improvements to capital assets (which constitute a capital expenditure) is often confused. According to the Government Finance Statistics Manual 2001, major renovations and improvements has the following features: (i) the decision to renovate, reconstruct, or enlarge an asset is a deliberate investment decision that may be undertaken at any time and is not dictated by the condition o f the assets; and (ii) renovations or enlargement increase the perfomnce or capacity o f existing assets or significantly extend their previously expected service lives. On the other hand, maintenance and routine repairs have the following features: (i) they are activities that owners or users o f assets are obliged to undertake periodically in order to be able to utilize such assets over their expected service lives; and (ii)they do not change the asset or its performance, but simply maintain it ingood working order or restore it to its previous conditions in the event o f a breakdown. Finally, purchases o f military weapons and equipment are treated as use o f goods and services rather than acquisition o f fixed assets. Source: World Bank 2007 Belarus policy note based on IMF, GFSManual 2001. OtherIssues TargetedPrograms 5.25 The state "targeted programs" have multipliedin number to the point of being almost out of control. The targeted programs establishedjust after Ukraine's independence had the role of financing specific social and development programs that contained a strong component o f capital investment. Their objectives ranged widely from social programs to law enforcement (see Box 5.4). Inthe early 1990sthey fulfilled the task o f securing funds for certain favored programs in the absence of a more structured budgetary process and inthe midst of a large number of unorganized off-budget funds that did not allow any planning. But with the reforms o f 2001-02, which brought much more structure to the budgetary process, these targeted programs lost their initial rationale. Nonetheless, and given the political sponsorship of members o f Parliament, they remained in the system and grew almost out of control- from 7 programs in 1993 to around 300 by 2006. Box 5.4: State Targeted Programs inUkraine The state targeted programs, according to legislature, are a set o f tasks and activities within a program aimed at resolving social and economic development needs. The introduction o f the state targeted programs was an attempt to establish a mechanism o f budget financing that would protect certain types o f programs inthe midst o f a chaotic budgeting process. Before 2004, state targeted programs were not regulatedby any umbrella legislation. The Law o f 2004 systematized the targeted programs, defining the following types: (i) economic programs aimed at solving industrial sectoral and intersectoral problems, promoting energy saving, promoting the creation o f new industries; (ii)scientific programs, financing fkndamental research; (iii) and design programs, aimed at the research creation o f new technologies, and materials, and increasing the competitiveness o f production; (iv) social programs, aimed at improving the quality o f life, reducing unemployment, improving health care and education; (v) cultural programs, aimed at preserving the cultural inheritance and satisfying the intellectual and spiritual needs o f people; (vi) environmental programs; (vii) defense programs; (viii) law enforcement programs; and (ix) other programs, including regional development programs. Source: Law o fUkraine on State Targeted Programs, #1621-IV. 115 5.26 The fundingfor targetedprogramsis mainly capturedby the category "economic activities" (mostly for coal and agriculture subsidies, for example), while other programs are kept alive but are under-funded. In 2006, programs under "economic activities" (which are dominated by the coal, agricultural and transport sectors) took 44 percent of the hnding for these targeted programs (see Table 5.2). Given that the coal and agriculture programs are geared mostly to finance input and production subsidies rather than actual investments, they pose concerns about their adequacy compared to potential productive infrastructure investments that are urgently neededelsewhere. The second largest beneficiary (also close to 44 percent) was the program for the social and economic development o f the city o f Kyiv. The remaining 12 percent i s spread out among (literally) hundreds of programs under the umbrellas of social, ecological, law enforcement, research, and other minor categories. Type of State Targeted Programs ExecutedIPlanned Total Economic 126% 43.9% Scientific 3yo 0.4% Research and Design 32% 0.3% Social 49% 5.1% National Culture 40% 1.2% Ecological 139% 4.3% Defense 42% 0.7% Law Enforcement 53% 0.4% Other 1/ 1385% 43.8% Total I 138% 100.0% 5.27 No comprehensiveevaluation of the economic feasibility and achievements of these targeted programs has been undertaken in recent years. These programs tend to stick, even though many may no longer offer a strong economic rationale for their existence or any other reason for their remaining under the new budgetary environment. In their current structure and numbers, these programs are becoming a burden for the (national and) local capital budgetingprocess rather than an efficient tool for channeling funding to needed investments. Moreover, according to JSRU reports, a large number of targeted programs are close to non-operational since they do not disburse funds regularly or do so in marginal amounts. This i s also made clear by an examination in more detail o f the funding for 300+ programs among the 9 categories shown inTable 5.2. ReportingIssues 5.28 There are also some problematic reporting issues. Treasury reports on the functional distribution (i.e., across sectors) o f capital expenditures are unclear. For example, fixed capital expenditures are classified functionally by sector (education, health, housing and communal services, transport, etc.) but only for capital acquisition o f equipment, and all expenditures for construction (for all sectors) are lumped into a category called "construction" which does not distinguishthe sector for which construction was undertaken. The category "capital transfers to enterprises and agencies" suffers from lack o f clarity in the reporting as well: first, it incorporates plain subsidies to spending units and enterprises, which should not be part o f the capital expenditure category; second, it does not break down this spending functionally (but rather by spendingunit). 116 E. RECOMMENDATIONS 5.29 The local capital budgetingprocess mirrors the problems o f the capital budgetingprocess at the national level. While most of the recommendations for improving the national budgeting process also fit the local realities, there are a number of specific recommendations put forward in this report that are geared to improving local capital budgeting, making it compatible with a strengthened system at the national level. These recommendations are as follows: Link the budgeting process to regionalinvestment priorities and nationalstrategic goals. Currently, the planning, evaluation, and selection processes are geared to be adapted to what i s likely to be funded by politicians, to the detriment of economic rationales and local (and national) goals. In coordination with line ministries and agencies, introduce multi-year strategies and resource envelopes. This is critical to improving planning and providing more security for investmentsat the local level. 0 Create a relativelystandard framework (including a template) for project evaluation for local governments. In fact, this framework and the criteria for assessing projects should not differ from those recommended in PFR Iat the national level (and for all line ministries). This framework should be comprised of a strengthened cost-benefit analysis methodology and an evaluation settingmethod with criteria clearly established, and with a rating scheme accordingly, to allow comparability among the investments examined. The framework could have different levels of detaildepending on the level of local government, the size of the investment (inUAH), and the type o f capital expenditure (renovation or new project). The MoE should take the lead in creating this framework and an appropriate and uniformtemplate and should also take the lead in training local authorities inits implementation and then enforcing its application. 0 Establisha consistentselectionprocess across sectors and regionsunderwhich projectsare moved up in the pipeline only if they rank high in the evaluation process. Moreover, consideration should be given to the creation of a different budget sub-item for projects put forward at the end of the pipeline without going through process o f evaluation and selection proposal. This would bring more transparency to the budgetary documents and discussions and couldhelp foster an environment inwhich productivepublic investments are privileged. 0 Enact a regulation (preferably within the context of amendments to the budget code) to grant preference in funding to ongoing projects superseding any annual Cabinet of Ministersresolution. Additional annual budget documents should be specific about ongoing projects and the funding they receive in a specific year attached to the timeline (and the remainingfunding) to completion. 0 Establish a formal criteria-based system of capital subventions/transfers to local governments. This would reduce political bargaining in the allocation of resources to local governments, improve the predictability of capital investments, and reduce somewhat the existingregional inequalities, Specific options for establishing a system are provided inChapter 2, Section E. 0 Reserve the co-financing scheme as a secondary mechanism for allocationthat could run parallelto the criteria-basedsubventions/transfers.Given the low level of fiscal capacity at the local level, this scheme would work only as the fiscal capacity increases (through the 117 recommendation in Chapter 2). In the meantime, the recommended scheme (see Chapter 2) for criteria-based capital subventions/transfers could provide a resource cushion for local government in terms o f their share o f co-financing arrangements. The use o f the co-financing scheme shouldbe kept but on more selective basis and only to create positive incentives for local government to invest innational priorities. Begin the collection of data on current assets and their condition at the regional and local levels and start feeding a national system for this purpose. Local governments, under their own initiative, could start to buildthese databases at the village and rayon (and city) levels, and oblast governments could consolidate the information under the tutelage o f the Department o f Economy. This initiative and information, by itself, would allow local governments to better plan the necessary spending for maintenance, repairs, and renovations. Collect and report data on maintenance spending separately from data on capital renovations. This would make it possible to more accurately plan the real spending needs for each item. Closely monitor the implementation process for capital projects at the local level, particularly in sectors where under-execution i s more acute. While the recommendations presented in this section would help mitigate some problematic issues that trigger under- execution, the regional and local Departments o f Economy and Finance would need to work together closely to identify bottlenecks at the early stages. Moreover, they should establish working groups with local authorities from line ministries such as Health, Education, Housing and Communal Services, where project implementation problems and under-execution are more severe. Re-evaluate the existing targeted programs and curtail those that no longer retain a strong economic rationale or are redundant in the current budgetary environment. As discussed earlier, many o f these programs are outdated, are close to non-operational, and constitute an extra rigidity inthe capitalbudgetprocess that affects the fundingand flexibility o f the allocation process o f local governments. Our examination shows that a large number o f programs could be curtailed on the basis o f the above-mentioned criteria. Moreover, a reducednumber o f programs are financing a large portion o fthe resources for input and production subsidy programs, creating inefficiencies in public spending and crowding out the fiscal space for needed investments elsewhere (more recommendations regarding these subsidies can be found inPFRI). 118 ANNEX 1.1:OrganizationalStructure ofthe General Government Budget of Ukraine 1_1 Consolidated Budget ~) [ LocalBudgets ~ ] General Fund ~ SpecialFund ~ Note: It i s important to clarify that Ukraine's legislation does not have a "General Government Budget" but two separatebranches: the "Consolidated Budget" and the off-budget Social Insurance Funds. For the purposes of this analysis (and to enable international comparisons) we consolidate both branches under the broad definition of General Government Budget. Source. World Bank Public Finance Review I(2006). 119 ANNEX 2.1: LocalGovernment Structure Figure A-2.1.l: Graphic Representation of Intergovernmental Structure and Intergovernmental Transfers 120 Table A-2.1.1 :Number of Local Governments in Each Administrative-Territorial Structure inUkraine, 2007 'Rayons City Rural Settlements kada City- Rayon Zity AR Crimea, Oblasts, City of -p5 % , = t y p e * -5 -4 b ) Q * * Special Status 4g $.r.ges villa S t 3 A 8 - 3 C b) 1 c 3 j8 .- 3 3 2 b 0 d 3 ' g- gm 9 g 5 i s * d . b g s $ gm O d m E PotalinUkraine I 490 I 118 I458 177 I279 I 2 6 I886 128,541 11,364127,1771112,0871 488 I I80 I456 117612781 784 110,2791 Source: State Statistics Committee. 121 ANNEX 2.2: PrinciplesintheDesignofTransfer(andEqualization)Systems'33 Objectivity An equalizationtransfer formula should have clear objectives and thus its designshould reflectthe objectivepursued.Within this specific type of intergovernmental transfers the objectives of the transfer could include the equalization of expenditure needs or the equalization o f fiscal capacity across the sub- national spending units, or both. The eventual decision on the objective o f equalization would depend upon such specific characteristics as income disparities across regions and differences inpopulation and demographic breakdown across regions; it would also depend on the determination of the size of the pool of resources, identification of which resources would go into the pool, and a decision on what portion of such resources would be retainedby the sub-national units. For example, if a larger amount of the total resources i s collected at the central level, a redistribution based on expenditure need may be more appealing. Likewise, ifmost of the revenue i s retained by sub-national spending units, then equalizing for fiscal capacity may be a more desirable objective. In other words, there are no "right" or "wrong" objectives, or less desirable or more desirable objectives. What i s necessary i s that the objectives be clearly enunciated and described. Adequacy Any equalization formula should provide the recipients with adequate funding to fulfill their mandates.This is particularly important when fundingis gearedto public services such as education and health, which have widespread externalities. Redirecting more resources to sub-national spending to recipients that have lower fiscal capacity and high expenditure needs could reduce the negative externalities (for example, from poor health care or a low quality of education provision) that would otherwise prevail. The criterion of adequacy also presents a convincing foil for the inevitable political considerations that often govern equalization transfers, since politically powerful local governments (which commonly are also the wealthier ones) tend to influence the design o f the formula in their favor. Providing inadequate resources to the poorer sub-national units defeats the very purpose o f the equalization formula, since the resultingunmet needs would eventually have to be met from the national pool of resources-but the delay would only increase the financial requirements. Simplicity The virtues of simplicity can hardly be overstated. Simple formulas are more easily and better understood by all stakeholders and have a much better chance of being sustained over time, since formulas that are not well understood are more likely to generate disagreements that would lead to the eventual abandonment o f the option. Moreover, the simpler the formula the easier it i s to collect and analyze the data needed to operationalize it. Timely and accurate data are essential inputs for any distribution formula, and mobilizing such inputscan be a challenge ifthe formula i s complex and the data demands are huge. Transparency The formula should be transparent in its design and in the rationalebehind the inclusion of the factors on which the equalization is predicated.Closely relatedto the principle of simplicity (lack of simplicity often translates into lack o f transparency) transparency increases acceptance and adherence, which improves the sustainability of the formula. In other words, transparency also ensures that the 133This annex extractsdirectly from Saavedraand Chawla(2005), and, overall, draws on the work of Jorge Martinez-Vazquez, (2003; 2004); and RoyBahl(2000). 122 factors, weights and coefficients used in a formula are not subject to sudden changes and political manipulation. Predictability The output of the formula needs to be easily predicted by all stakeholders. In particular, local government officials should b o w fairly accurately the amount o f resources they can expect to receive during the following years. Predictability i s a critical ingredient of financial planning, which in turn is essential for the efficient administration of service provision. A multiyear approach to planning resources i s always appealing for that purpose, though constant changes in the factors, weights, and coefficients of the formula can make that task very difficult. Planning and financial certainty in the medium and long terms are essentialto the enhancement of any service provision system.'34 Incentives Equalization transfers should provide the "right" incentives for local government. The goal of equalization transfers i s to provide the means for local governments to afford a level o f service provision that they would be unable to reach otherwise. However, the way in which the equalization formulas are set up should not provide incentives for spending sub-units to relax their own fiscal efforts to raise revenues. It i s important, therefore, for the equalization formula to be incentive compatible in the sense that while it should provide relief to the spending sub-units that qualify, it should not do so at the cost of diluting revenuecollectionefforts at the level ofthe spendingsub-units. `34 This principle also applies to the first stage of implementation o f a new funding formula. That is, the design o f the equalization mechanism should yield enough resources to local governments so that it does not affect its former level o f funding. 123 ANNEX 3.1:Selected Health Status Indicators and Cross-Country Comparisons Figure A-3.1.1: LifeExpectancyvs Income for Males and FemalesAcross Countries, 2004 LIFE EXPECTANCY VS INCOME, 2004 LIFE EXPECTANCY VS INCOME, 2004 MALES FEMALES 53 8 0 0 o n / P 53 $ iR f! 3 8 + e ip // d 8 0 0 / / 0 8 160 250 GDP per capita 1000 2500 1000025000 GDP per capita 1000 2500 1000025000 100 250 Note: GDP per ceplta Source: World Develo ment Indicators constant 2000 US$: Log scale Source World Development Indicators Nota GDP per capita in mnstant 2000 US$. Log scale Figure A-3.1.2: Infant and ChildMortality Trends, 1960-2004: Ukraineversus Comparator CountriesDtegions tENDS IN INFANTMORTALITY, 1960-2( RENDS IN CHILD MORTALITY, 1960-20C 1 -7 1960 1970 1980 Year 1990 2000 1960 1970 1980 Year 1990 2000 Souce: World Development Indicators 124 C i l i t y Groups as Percentages of Total DALYsinUkraine, 2002 3 Males Females Disabili ou Total DALYs("h Cardiovascular diseases 27.2 Cardiovascular diseases 33.6 Unintentional injuries 15.5 Neuropsychiatric 16.5 conditions Neuropsychiatric 11.3 Malignant neoplasms 10.4 conditions Infectious and parasitic 9.3 Sense organ diseases 5.4 diseases Malignant neoplasms 9.0 Unintentional injuries 5.2 Intentional injuries 7.2 Musculoskeletal 5.2 diseases Digestive diseases 4.9 Digestive diseases 4.1 Respiratory diseases 3.8 Respiratory diseases 3.8 Sense organ diseases 3.1 Infectious and parasitic 3.4 diseases Musculoskeletal 2.2 Intentional injuries 2.6 diseases Source: WHO. 125 ANNEX 3.2: Average Lengthof Stay RegionalAnalysis Table A-3.2.1:FindingRegionswith Higherthan Averape Lenpthof Stav-Probit Repression Variable Coefficient Z-statistic EconomicRegion Eastern 0.252 (1.94) Donetsk 0.241 (1.79) Prechornomorsk 0.291 (2.36)* Podi1ia 0.109 (0.87) Central 0.074 (0.55) Predniprovsk 0.235 (1.71) Carpathian 0.014 (0.11) Socioeconomic and demographic characteristics 19