Report No. 23492-RO Romania Local Social Services Delivery Study (In Two Volumes) Volume 2: Main Report January 16, 2002 Human Development Sector Unit Europe and Central Asia Region Document of the World Bank CURRENCY EQUIVALENTS (Exchange Rate Effective January 15, 2002) Currency Unit Romanian LEI (ROL) LEI 1 US$0.000035 US$1 ROL 32,070 FISCAL YEAR January 1 - -- December 31 ABBREVIATIONS AND ACRONYMS CEB Budgetary Execution Center CPD Judet Child Protection Directorate EBRD Econonuc Bank for Reconstruction and Development EU European Union GDP Gross Domestic Product JCPC Judets Child Protection Commission JEI Judets Education Inspectorate JHIH Judets Health Insurance House JSIPH Judets State Inspectorates for Persons with Handicap LLPF Law on Local Public Finance LSSD Local Social Services Delivery MAP Ministry of Public Administration MIG Minimum Income Guarantee Program MMSS Ministry of Labor and Social Solidarity MOE Ministry of Education MOF Ministry of Finance MOH Ministry of Health MTSAB Means Tested Social Assistance Benefit NACPA : National Authority for Child Protection and Adoption NGO Non Government Organization PHD Primary Health Directorates RSDF Romanian Social Development Fund SSHP State Secretariat for Handicapped Persons TIMSS Third International Mathematics and Science Study Vice President: Johannes F. Linn, ECAVP Country Director: Andrew N. Vorkink, ECCO5 Sector Manager: Maureen Lewis, ECSHD Task Team Leader: Dena Ringold, ECSHD Table of Contents Table of Contents ............................................. iii Chapter 1: Introduction and Conte.xt ............................................ 1 1.1. Objectives and Background . . . 1 1.2. Local Governments and Decentralization in Romania . . . 1 1.3. Decentralization and Service Delivery: A Theoretical Introduction ...3 1.4 Methodology . . . 7 1.5. Scope of the Study . . . 9 1.6. Road Map of the Report . . .14 Chapter 2: Local Financing and Administration . . .......................................... 18 2.1. Local Government Finance .................................................... 18 2.2. Implementation at the Local Level ................................................... 29 2.3. Summary Issues ................................................... 33 Chapter 3: Institutional Arrangements for Service Delivery ............................... 34 3.1. Introduction ................................................... 34 3.2 Institutional Arrangements in the Social Sectors . ................................................. 35 3.3. Implications of Institutional Arrangements .................................................. 40 3.4. Summary Issues ................................................... 47 Chapter 4: Financing the Social Sectors .. . .. 52 4.1. Itroduction . . .52 4.2. Financing the Sectors . . .54 4.3. Education . . .59 4.4. Health . . .64 4.5. Summary Issues . . .66 Chapter 5: Service Delivery at the Local Level . . . .................. 71 5.1. Service Delivery . . .71 5.2. Access . . .72 5.3. Allocation of Resources and Services . . .79 5.4. Participation and Accountability . . .84 5.5. NGO Involvement in Social Services . . .86 5.6. Responsiveness of Local Social Services . . .89 References .... 93 iii Boxes Box 1.1 Defining Decentralization ........................................................4 Box 1.2 Decentralization and Government in Romania ........................................................7 Box 2.1 Shared Taxes: advantages and Disadvantages ....................................................... 23 Box 2.2 Allocation of Equalization Grants by the Judet Councils ...................................... 26 Box 2.3 Coping Strategies at the Judets Level ....................................................... 32 Box 3.1 Representative Schools ....................................................... 38 Box 4.1 Collection of the Social Solidarity Tax ....................................................... 55 Box 5.1 Poverty and the Roma in Romania ....................................................... 76 Box 5.2 Public Participation in Local Governm ent ....................................................... 85 Box 5.3 NGO Involvement in Child Protection Services .................................................... 89 Box 5.4 Community Based Services and the RSDF ....................................................... 90 Tables Table 1.1 Administrative Arrangements, 1996 ........................................................3 Table 1.2 Study Sites, Population 1999 ........................................................8 Table 1.3 Select Human Development Indicators for the Case Study Judets, 1998 ................8 Table 1.4 Enrollment Rates, 1990-98 ....................................................... 12 Table 1.5 Health Status in Comparison, 1998 ....................................................... 14 Table 2.1 Major Legislative Development in Local Public Finance ..................................... 19 Table 2.2 Macroeconomic Indicators ....................................................... 19 Table 2.3 Selected Fiscal Indicators ....................................................... 20 Table 2.4 Total Revenues, Judets and Local Councils ....................................................... 21 Table 2.5 Judets and Local Councils Revenues in real terms (1996=100) ............................ 21 Table 2.6 Shared Taxes (% of total local revenues) ....................................................... 23 Table 2.7 Criteria for Allocation of Transfers (Sume Defalcate), 1997-99 ........................... 25 Table 2.8 Total Expenditures of Judets and Local Councils .................................................. 27 Table 2.9 Coefficient of Variation, Per Capita Revenues of Local and Judets Councils ...... 30 Table 2.10 Transfers per capita on previous years own revenues and poverty ....................... 31 Table 2.11 Trends in Local Revenues in the Study Sites ....................................................... 31 Table 4.1 Expenditures by Source ....................................................... 52 Table 4.2 Distribution of Responsibilities for Financing Child Protection ........................... 54 Table 4.3 Total Social Assistance Spending ....................................................... 56 Table 4.4 Social Assistance Expenditures of Judets and Local Councils .............................. 57 Table 4.5 Trends in the Number of MTSAB Beneficiaries ................................................... 58 Table 4.6 Summary Data on Education Spending ....................................................... 61 Table 4.7 Summary Data on Health Spending ....................................................... 65 Table 4.A. 1 Trends in Education Spending ....................................................... 68 Table 4.A.2 Trends in Social Assistance Spending ....................................................... 69 Table 4.A.3 Local Education Spending ....................................................... 70 Table 5.1 Poverty Rates in Romania ....................................................... 72 Table 5.2 Socioeconomic Status, Rural and Urban Areas, 1998 ........................................... 72 Table 5.3 Romania, Poverty by Ethnicity, 1995 and 1997 .................................................... 76 Table 5.4 Education Services in the Study Municipalities and Communes (grades 1-8) 1996 ........................................ 80 Table 5.5 Health Services in the Study Sites (per 1,000 pop) 1996 ...................................... 81 iv Table 5.6 Capacity of Child Welfare Institutions, 1999 ............................................... 81 Table 5.7 Out-of-Pocket Expenditures on Education ............................................... 82 Table 5.8 Out-of-Pocket Health Expenditures (1994 & 98) ............................................... 84 Table 5.9 Trust in Public Institutions, 2000 ............................................... 85 Table 5.10 NGOs in the Study Judets (per 100,000 pop), 2000 .............................................. 87 Figures Figure 1.1 Responsibilities for Financing the Social Sectors,1999 ...........................................9 Figure 1.2 Poverty Rates in the Study Judets, 1998 ............................................... 11 Figure 2.1 Total Revenues in Real Terms ............................................... 30 Figure 4.1 Social Sector Expenditures ............................................... 53 Figure 4.2 Composition of Local Social Assistance Spending (1996-99) .............................. 54 Figure 4.3 Education Spending in Real Terms ............................................... 62 Figure 4.4 Evolution of Local Education Spending in the Sampled Towns ........................... 63 Figure 4.5 Evolution of Local Education Spending in the Sample Communa ....................... 63 Figure 4.6 Health Expenditures in Real Terms ............................................... 65 Annexes Annex IA Description of the Study Methodology ............................................... 16 Annex 3.1 Intergovernmental Responsibilities in the Social Sectors ..................................... 49 v Chapter 1: Introduction and Context 1.1. Objectives and Background Through the mid-nineties, Romania remained one of the most centralized states in Central and Eastern Europe (Wetzel and Dunn, 1998). However, more recent developments, beginning in 1998, including the adoption of new legislation on intergovernmental fiscal relations and property ownership have accelerated the decentralization process, and local governments are gaining increasing responsibility for financing and managing social services, including social assistance cash benefits and services, education and certain health care services. This transfer of responsibilities is proceeding in a dynamic context alongside major reforms in each of the sectors, and the capacity of local governments to adapt to the new processes varies significantly across the country - increasing the risk of regional disparities in social service provision and outcomes. This is the second volume of a study which was initiated to understand the implications of increasing local government autonomy for social service provision in Romania. It examines the changing institutional context for service delivery, fiscal trends and reforms in the social sectors from different perspectives: including local government administrations, central government agencies at the national and local level, social service institutions, NGOs and the beneficiaries themselves. Volume 1 of the report provides an overview of the issue, summarizes the major findings and presents a discussion of policy options and recommendations. This second Volume includes the detailed discussion, including theoretical background and empirical analysis which underlie the conclusions and policy recommendations in Volume 1.1 This first chapter of the study provides an overview of the decentralization process in Romania and discusses thematic issues which are explored throughout the report. It includes a brief discussion of the existing theory on decentralization and public services, focusing on potential benefits and pitfalls of increasing local government involvement in services, and the corresponding implications for equity, efficient and adequate financing, responsiveness and institutional incentives. Following a description of the methodology used for the study and an introduction the case study sites, the report concludes with an overview of the status of the three sector areas addressed: social assistance, education and health care. 1.2. Local Governments and Decentralization in Romania Like other countries in Central and Eastern Europe, Romania has a tradition of strong local governments. Romania's pre-communist system, modeled on the French system, included financially autonomous and democratically elected districts and municipalities. However, power was centralized following the communist takeover and local governments became limbs of the central government. In 1991, after four decades of central control of local affairs, Romania adopted a constitution mandating that responsibility for local affairs management be transferred to local governments. Shortly afterwards, the Law on Local Public Administration2 defined the institutional framework for local governments by re-establishing the two-tiered system that lAs much as possible, there is no duplication across the reports, however to ensure continuity for readers some material from Volume 1 is included here. 2 Law 69/91. existed before the advent of the communist regime. Nonetheless, little effort was made at the time to amend existing legislation and to make it compatible with the new administrative arrangements. In particular, financial management and expenditure decisions remained under the strict control of the central govemrnent and approval of the Minister of Finance was required for all major investments (Jakob, 1995). Since then, effective decentralization has been singled out as a priority for the Romanian government on a number of occasions.3 In October 1998, Romania's Parliament passed a new Law on Local Public Finance which provided for a significant restructuring of fiscal authority of public service provision. This development has had far-reaching consequences. The Law fundamentally changed local public administration in Romania by expanding local authority in the establishment of expenditure priorities,4 the approval of investments,5 long-term budgeting,6 access to credit from the internal or external capital markets,7 and assessing and collecting revenues from local taxes and fees. 8 The new Law transfers authority for the financing of aspects of many social services, including education, social assistance, and social services such as housing and community services to local governments. The Law, together with companion legislation on ownership of public property, strengthened the fiscal basis of the subnational governments, through increases in their shares of taxes and fees, and new provisions to increase their revenue raising and borrowing capacities. Local governments in Romania consist of judets (counties), and local councils which are comprised of: municipalities (municipiu), towns (oras) and communes (communa), consisting of one or more villages. In 1996, there were 41 judets, including the General Muncipal Council of Bucharest, 79 municipalities, 182 towns and 2,682 communes in Romania (Table 1.1). Municipalities, towns and communes all have the same legal authority and administrative structure, the distinction between these different types of local councils is historical and geographic.9 Members of local and judet councils are all locally elected for four year terms. Legislation passed in 1999 puts all forms of local government on an equal footing. There is no hierarchical relationship between judets and local councils, other than for the distribution of specified budget transfers which are made by judets to local councils. Throughout this report 'local governments' refers to all judets, municipalities, towns and communes, 'judet councils' refers to county administrations, and 'local councils' refers to the administrations of municipalities, towns and communes. 3 See for example the 1997 Country Assistance Strategy. 4Law no 189/98 - art. 16, paragraph 1 5 Law no. 189/98 - art 16, paragraphs 5 and 6 6 Law no. 189/98 - art 6, paragraphs 4 7Law no. 189/98 - art 48, 50, 55 s Law no. 189/98 - art 5, art. 26 - paragraphs l and 2 9 The only distinction is that municipalities may include elected bodies in sectors, as in the case of Bucharest, which is subdivided into 6 sectors. However, the implementation of this provision requires adoption of a special law. Other than Bucharest, none of the other municipalities are organized in sectors. 2 Table 1.1: Administrative Arrangements, 1996 Units Average Population Minimum Maximum Population Population Population Density (people/km) Judets 40 507,397 231,872 868,099 88 Bucharest 1 2,037,278 -- -- 8,929 Municipalities 79 100,613 18,554 346,830 1,354 Towns 182 13,327 1,856 35,861 253 Communes 2,682 3,796 119 27,006 62 Source: National Commissionfor Statistics 1.3. Decentralization and Service Delivery: A Theoretical Introduction Decentralization has become a world-wide phenomenon in the last two decades, mostly stemming from political momentum (Box 1.1). In particular, for the countries of Eastern Europe and the former Soviet Union, decentralization has represented an important component of the reorganization of the state in the transition from planned to market economies (Bird, Ebel and Wallich, 1995). Although mainly political in nature, the modalities of the transfer of responsibilities to local governments have important economic consequences, as they affect govermment organization and incentives. In turn, recent research has identified institutional design arrangements as important determinants of the quality of public service provision. The traditional debate on the benefits and costs of decentralization has crystallized along the trade-off between equity and efficiency. Theory predicts that a system of local elections accompanied by decision making and financial autonomy, will allow local policymakers to have direct information on the preferences of their constituencies, and therefore provide public services that are consistent with local preferences (so called allocative efficiency) (Tiebout, 1956). Similarly, locally elected policymakers will more be more responsive to local needs than centrally elected policymakers. Moreover, thanks to the beneficial effect of inter-jurisdictional competition, decentralization is expected to decrease the cost of public services and obtain cost or technical eff iciency. A recent strand of literature highlighted that the benefits of decentralization cannot be obtained unconditionally.'0 These studies highlight the importance of the incentives laid out by the institutional framework within which the different layers of govermment interact and of well functioning accountability mechanisms that effectively keep the actions of public officials in check. The benefits of decentralization should be weighed against its costs. As the provision of public services becomes more dependent on local resources, substantial inequalities can emerge across regions and threaten horizontal equity within the country and differentiate access to public services across different strata of the population, often affecting the most vulnerable. In addition to these concerns, recent contributions have stressed other possible pitfalls of decentralization, in particular that governance might substantially worsen at the local level because of capture by local elites (Box 1.2). 10 For a detailed discussion of theories of decentralization see, amongst others, Azfar et al. (1999), Gatti (2000), Litvack et al. (1999). 3 Disentangling the impact of the decentralization process on social service provision in Romania is an impossible task. Changes in public service delivery in Romania are proceeding in a dynamic context of political and economic change. Social services have been affected by the severe fiscal declines of the last decade, as well as substantive reform which has proceeded, or is underway, in all of the social sectors. Decentralization itself has been a dynamic process, with annual changes in legislation. The level of decentralization varies across sectors and programs. As such, this study is concerned with understanding the implications of changed fiscal and administrative arrangements on social service provision. It is especially concerned with equity, adequate financing, responsiveness and institutional incentives. These concepts are discussed in more detail below in the context of the delivery of social services in Romania. Box 1.1: Deflning Decentralization Decentralization is a complex phenomenon generally involving the shift of authority and/or financial resources from higher- to lower-tier governments and ranges from simple dispersion of central government responsibility to central government regional offices - deconcentration - to full privatization. Decentralization has many different aspects including allowing local government representatives to be selected through local elections (political decentralization), transferring authority and responsibility for service delivery to local governments (administrative decentralization) and shifting authority to raise adequate revenues and to make expenditure decisions to the local level (fiscal decentralization) (Litvack et al., 1999). Moreover, the literature has identified three main forms of decentralization based upon the location of decision making authority: Deconcentration occurs when lower levels of government (e.g. regional or municipal) execute functions on behalf of the central government without having decision making authority. The lower levels of government essentially function as "branch offices" of the central government. Delegation is the process in which the central government transfers decision making over certain functions to relatively autonomous local governments, which have limited autonomy and are ultimately responsible to the central government. Devolution involves the complete transfer of decision making over finance and management of public services to quasi-autonomous local government units. Devolution usually transfers responsibilities for services to municipalities where residents elect their own mayors and council members, revenues are raised locally and local governments have independent authority to make investments. Decentralization in Romania has involved aspects of each of these different types of institutional change. Equity A key concern of decentralization is the implication for equity - both across regions (horizontal equity), and among socioeconomic groups within society. Defining and measuring equity is a challenging task that can be separated into considerations of equity of opportunity and equity of outcomes. Equity of opportunity is concerned with providing a level playing field such that all individuals have the same chances, while equity of outcomes focuses on final welfare gains. Distinctions between the two can be easily blurred at the conceptual level, and are even more difficult to sort out in practice (Klugman, 1997). Analysis of horizontal equity and changes in patterns of service delivery across regions are a main focus of this analysis. Romania is one of the more geographically and ethnically diverse countries in Central and Eastern Europe. Rural and urban differences are notable. There are also significant differentials in levels of regional development, as indicated by variation in 4 poverty and unemployment rates, infrastructure and other variables. Recent poverty analysis for Romania using 1998 household survey data found that judets in the south, southwest and northeast have significantly higher poverty ratios (Tesliuc, Pop and Tesliuc, 2000), and across the country poverty is greater in rural areas. Emerging evidence indicates that regional disparities in development have been growing during the transition (Government of Romania and EU, 1997). These gaps are likely to widen as restructuring proceeds in Romania, as certain regions, which are home to a greater share of unproductive industries are likely to be more affected than others by unemployment and poverty. Fiscal decentralization has the potential to exacerbate these differentials in development even further, as poorer districts and municipalities will be under pressure to provide public services with a smaller resource base and weaker administrative capacity than more well-off areas. For example, decentralization of social assistance has had dramatically varied effects across localities, and availability of benefits and quality of service has come to differ significantly. In this regard, the ability of local governments to provide social services is critical for regional development, as regions which are less well-off in welfare terns -- such as low educational attainment and health status -- will be disadvantaged in terms of employment and investment and, consequently, growth over the longer term. This study examines the impact of the reform process on equity of social service provision across Romania to look at the extent to which disparities in available resources have shifted during the transition, and how they might be expected to evolve given the new legislative framework, focusing on social sector expenditures of local governments. The study also looks at the redistributive effects of social policies across different strata of the population. There is some evidence from the household surveys, that transition has led to increasing costs of social services, particularly health and education, to households, through user charges and under-the-table payments. In the context of growing poverty in some areas, these are serious developments which may effect the ability of families to send their children to school, or to make use of health services. Specific issues arise with regards to ethnic minorities, and particularly the Roma population, and this study will look at the impact of changes in social service delivery on Roma. Efficiency and Adequacy Like other countries in Central and Eastern Europe, the legacy of central planning has left Romania with inefficient and ineffective social services. In health, this appears through overstaffing and overspecialization of staff and facilities (e.g. excessive emphasis on secondary care at the expense of primary medicine), poor use of space, and low levels of capital investment and maintenance expenditures. Similar trends appear in education where overstaffing persists and investments have eroded. In social welfare, Romania's overemphasis on institutionalization of children, elderly and the disabled - trends which have continued into the transition period - come at the expense of more cost-effective and humane policies which encourage home and community based care and services. Analysis of technical efficiency raises a number of key issues which are critical from a policy perspective. For example, decentralization may eliminate benefits gained through economies of scale in education and health provision. Analysis can help to inform decisions about the appropriate size of local government areas from an efficiency perspective. In addition, changes in institutional responsibilities and accountabilities can have a direct impact on incentive 5 structures and it will be important to examine how institutional factors and processes facilitate or impede efficiency gains. At the same time, the rapid decentralization process coupled with continuing fiscal constraints has the potential to endanger the adequacy of public spending in social sectors. Issues of fiscal adequacy as well as of technical efficiency are an important focus of the analysis. Responsiveness Related to the above, the study examines the responsiveness of local governments and social services, and assesses whether and how increased local autonomy is affecting the ability of service institutions to meet the needs of local communities. The concept of local governments as service providers is a new one in Romania, requiring a reconsideration of the roles of institutions and individuals, and their approaches to service delivery. This issue takes on different characteristics across regions, as localities have a variety of special needs, including socioeconomic variation and ethnic diversity. An important issue is the extent to which local governments have the flexibility to meet local needs and have the capacity to identify and initiate innovative approaches to local service provision. The capacity of local governments to collaborate with civil society and to involve the community, NGOs and other actors in service delivery is another important ingredient in this regard. Incentives and accountability Incentives have a crucial role in the achievement of the benefits of decentralization. The study devotes substantial attention to the analysis of the set of institutional incentives and constraints placed on the many actors who interact in the financing and provision of social services. Borrowing from the theory of the firm, the recent literature has often characterized the tasks of local governments within a principal-agent framework where the central government or the local electorate (principal) endows the local government (agent) with resources and authority to carry over a specific set of tasks (Cremer et al., 1994). According to these models, the principal should design a contract where rewards and penalties are such that the agent is willing to carry on the tasks in the principal's best interest. Nonetheless, reality - especially concerning the provision of social services - is far more complex. Central and local governments are heterogeneous structures that might pursue conflicting objectives and where politicians and bureaucrats are often motivated by self-interest. Moreover, service providers are linked to both central and local governments by multiple arrangements that both regulate and constrain the delivery of services. For example, in Romania schools follow nation-wide education policies and are under the supervision of deconcentrated branches of the central government. At the same time local governments are involved in financing of current and capital spending of compulsory education and can therefore substantially influence the schools' ability to carry over their tasks. In this context, having decision making aligned with financial responsibility as well as minimizing the fractionalization of responsibilities across different institutional actors is particularly crucial. The study also identifies a number of instances in all sectors where simply improving transparency and clarifying responsibilities can prevent coordination problems and coverage gaps. Effective institutional design is directly related to the issue of accountability - i.e. to how and by whom control over public officials and service providers is enforced. Through local elections, citizens can discipline the actions of local politicians. Nonetheless, often no means are available to citizens to provide feedback to service providers on the quality of their services nor 6 to enforce change effectively if the performance of service providers falls short of expectations. This is particularly relevant for Romania, where bringing the government "closer to the people" represents an important chance for enhancing governance (see Box 1.2). Box 1.2: Decentralization and Governance in Romania The issue of whether governance improves or worsens with decentralization is highly debated in the literature. Some authors suggest that decentralization strengthens citizen's participation and local governance (Shah 1998) while others point at the lack of an arm's length relationships between local public officials and citizens as potential causes of weakened governance at the local level (Prud'homme, 1995) and (Tanzi, 1996). The empirical literature on the subject is still scant. A study by Fisman and Gatti (2000) finds that improved governance is associated with fiscal decentralization in a cross section of 68 countries. Similarly, a country study on the Philippines found that corruption at the local level was perceived to be less pronounced than at the national level after decentralization (Azfar et al., 2000). Nonetheless, while providing useful information on the association between the two variables, none of these works can actually measure the impact of decentralization on governance and more in-depth and country-specific analysis is needed. Governance in Romania has been assessed as weak in comparison with other countries in the region - although comparative measures are methodologically difficult. The Business Environment and Enterprise Performance Survey conducted by the EBRD and the World Bank surveyed entrepreneurs and managers about their interactions with the state and factors which influenced their ability to function. Categories of indicators included micro and macroeconomic governance, physical infrastructure and law and order. In the survey, Romania ranked 18' out of 20 countries included from Central and Eastern Europe, ahead of the Kyrgyz Republic and Moldova. Although it is not possible at this stage to assess the impact of the decentralization process on governance in Romania, it is important to note that the ongoing shift of resources and authority to local governments represents a considerable opportunity to deepen democratic institutions and strengthen mechanisms of accountability in the country. In particular, attention should be paid to the design of the new intergovernmental relations to gradually decrease the regulatory burden and improve transparency to consolidate better governance. 1.4 Methodology This study draws from a number of different methodologies, building upon a range of primary and secondary sources. It examines issues at both the national and local level, through in-depth case studies of local governments in five judets and Bucharest. It is important to note that these individual case studies are not representative and generalizations should not be made based upon these methods. However, the case studies do document actual experiences faced in localities across Romania and provide a context for understanding national level developments. Throughout this report findings from the fieldwork are identified. The fieldwork for the case studies was undertaken between June and November 1999 by three teams of Romanian social scientists which approached the issues from different perspectives including: (a) institutional and financial analysis of local public administration; (b) institutional analysis of social service providers; and (c) beneficiary assessment. A short description of each of these approaches is provided at the end of this chapter (Annex 1A). The study took place in a dynamic policy context, with some aspects of decentralization moving faster than others, and the entire administrative organization of the country under review. 7 This naturally complicated analysis and required a creative mix of quantitative and qualitative approaches. The research was spread over six months (June - November 1999) that coincided with the first year of the implementation of the Law on Local Public Finance. During the fieldwork, the judet and local councils included in the analysis were at different stages in implementing the new provisions of this legislation. In addition, analysis of the health sector was complicated by the fact that financing reforms following the introduction of health insurance were underway. A number of institutional arrangements and procedural factors had not been clarified at the time of analysis and this led to confusion, which was reflected in the interviews with key informants and beneficiaries. Selection of the Study Sites The sites for this study were selected to represent a range of different situations and circumstances. Criteria for selection included geographic diversity, socioeconomic characteristics and existing administrative capacity. The study covers five judets: Covasna, Hunedoara, Timis, Tulcea, Vaslui and Sector 5 of Bucharest. Within each judet, one municipality and one commune were selected for in depth analysis (Table 1.2). Table 1.2: Study Sites, Population 1999 Judet Municipality Population Commune Population Covasna Tirgu Secuiesc 21,493 Sinzieni 4,634 Hunedoara Lupeni 33,246 Vorta 1,231 Timis Lugoj 48,415 Nadrag 3,519 Tulcea Babadag 11,074 Sarachoi 7,969 Vaslui Vaslui 79,783 Zapodeni 3,848 Note: Population data are projections; Source: local councils The selected sites represent a diverse set of localities. Among the judets, Vaslui, located in the northeast in the Moldova region, and Tulcea, at the Danube delta on the Black Sea, are the poorest and least developed. Both judets have higher than average unemployment and poverty rates (Table 1.3). Hunedoara, in northwest Transylvania, is a center of the mining industry, and as a result has been strongly influenced by restructuring. The unemployment rate in Hunedoara in 1998 was over 21 percent, among the highest in the country. Table 1.3: Select Human Development Indicators for the Case Study Judets, 1998 Total Population % Rural Unemployment Infant Mortality Rate Population Rate (%) (per 1,000 births) Covasna 230,992 47.6 8.9 18.7 Hunedoara 531,238 24.0 21.3 17.1 Timis 682,409 38.5 6.2 21.6 Tulcea 265,173 51.1 11.5 26.9 Vaslui 432,375 58.9 15.3 17.2 Bucharest 2,016,131 - 6.2 13.6 Romania 22,502,803 45.1 10.4 20.5 Source: National Statistical Institute Covasna, in central Transylvania, ranks as an average judet in terms of socioeconomic development. It is unique in that over 90 percent of its population is Hungarian. Finally, 8 Bucharest, the capital city, is a region in itself It has a large population of 2.3 million (9 percent of the national population), and is divided into 6 'sectors' or administrative units. Sector 5 was chosen as the focus of this study as it is distinguished by high levels of poverty and has a significant Roma minority." The municipalities and communes included in the study also represent contrasting circumstances. Vaslui is a municipality and the judet capital, Lugoj is categorized as a municipality, and Babadag, Lupeni and Tg. Secuiesc are towns. Municipalities tend to be larger and more developed than towns. 1.5. Scope of the Study Social services are affected by the new legislation and the decentralization process in general, as local governments have become responsible for financing nearly all social assistance benefit programs and services, including orphanages and homes for the elderly, as well as an increasing share of education expenditures. Local governments also have responsibility for delivering benefits, managing social assistance institutions and contracting services with NGOs. The level of local govenment involvement in financing social services varies across sector areas (Figure 1.1). Social assistance and education, have been most notably affected by the transfer of financing and administrative responsibility to local and judet councils. Social assistance cash benefits and services are financed out of local budgets and local governments retain responsibility for establishing and managing services. Local government financial involvement in the education sector is also increasing. At the time of the fieldwork for this study, local governments were responsible for most expenditures in the sector, with the main exception of personnel costs. Beginning in 2001, personnel costs for education were also transferred to local government budgets. Decentralization in education is also proceeding to the school level. A current experiment in 3,404 schools in Romania has transferred responsibility for financial and personnel decisions to schools. Based on the results of this program, the central government may delegate further responsibility to schools and local governments. In conjunction with these changes, responsibilities for financing personnel costs for education may be allocated to local councils. Figure 1.1: Responsibilities for Financing the Social Sectors, 1999 (% of total spending in the sector) 100% 90% 80% 70% 60% 50%F - VI- 40% Ceta 30% 20% 10% 0 % Health Education Soc ialI A ssis ta nc e Sources: LSSD database; IMF 1 I Qualitative work focused on Sector 5, while quantitative data were collected for the entire municipality of Bucharest. 9 Of the three areas of social service provision, the direct role of local councils in health care is the most limited, because of the introduction of national health insurance in 1997. Health insurance funds at the judet level now finance the majority of health care costs through contracting arrangements with physicians in local clinics. While local governments had been responsible for financing a small share of health expenditures - 13 percent in 1996 - this has since fallen to negligible amounts. The future role of local councils in health care is likely to be limited, but may emerge in certain areas including ownership of some facilities and providing incentives to recruit physicians into the locality. Given this situation, the most important role for local councils in impacting the health status of the community and the quality of health care services in a given municipality or commune is likely indirect - through provision of local infrastructure (e.g. roads, communications) and utilities (e.g. water supply). This study focuses on public social services where local governments play the greatest role. The emphasis throughout the report is greatest on social assistance benefits and services. The study addresses those cash benefits which are delivered and financed by local governments. These include national programs such as birth grants, emergency assistance, and most importantly, the main poverty alleviation program, the means-tested social assistance benefit (MTSAB).12 Social services and institutions for the elderly, disabled and children are also included. It does not address social insurance benefits such as pensions and unemployment benefits, or benefits paid centrally such as family allowances. In education, the study focuses on compulsory and secondary education, but does not address specialized secondary education which is under the purview of central ministries, or post-secondary and university education. Because of the limited role of local governments in health care, health receives more limited attention throughout the report. In health care, the focus is on those services which are delivered locally in dispensaries and polyclinics. Hospital care is outside the scope of this study. The following sections provide a general overview of the three social sector areas covered in the report. The discussion provides a brief introductory description of the social services provided and major sector issues and current reforms. A comprehensive discussion of the reform agenda facing each of the sectors is beyond the scope of this report, the intention here is to provide a snapshot of some of the main sector issues which frame the discussion of institutional change and fiscal developments in the later chapters.13 Social Assistance Social assistance is a complicated set of policies including a range of different cash and in-kind programs and services designed to address a variety of social needs (Andrews and Ringold, 1999; World Bank, 2000). Since 1990, Romania has faced the challenge of adapting its social assistance system to address the needs of a market economy. Explicit social assistance programs did not exist under the socialist regime, as the majority of the population was protected through the provision of full employment. Limited categorical benefits did exist for individuals 12 Throughout the paper the current version of the means-tested social assistance benefit program is referred to as MTSAB. Draft legislation approved by the cabinet in November will restructure the program and rename it as the Minimum Income Guarantee program (MIG). This term is used to describe the future program. 13 More detail on the sectors can be found in separate reports, including Tesliuc, Pop and Tesliuc, 2000 on social assistance; World Bank 1999 on health, and chapters on health and education in the forthcoming Country Economic Memorandum. 10 in particular circumstances, such as the disabled, or orphans. Reforms to date have entailed the introduction of means-tested social assistance to create a safety net for the poorest in society, reforms to family benefits, and restructuring of services for children at risk. The primary objective of social assistance in a market economy is to address the needs of the poor. Poverty in Romania has escalated during the transition period. Analysis of household data for 1998 indicated that 34 percent of the population was living below a relative poverty line set at 60 percent of mean per capita consumption (Tesliuc, Pop and Tesliuc, 2000). Larger families with three or more children, female-headed households and Roma have been most at risk of poverty. Poverty in Romania also has significant regional dimensions. In particular, rural poverty has consistently been twice as high as in urban areas. In 1998 the poverty rate was 41 percent in rural areas, in contrast with 28 in urban areas. Across Romania, poverty was found to be most concentrated in the northeast (including Vaslui in our study), south and southwest of the country (Figure 1.2). 14 Figure 1.2: Poverty Rates in the Study Judets, 1998 60.00 ts,t~ ~~~~~~~~~~~~~~~~~ Tota >0 1*00~ ~~~~~~~~~~~~~~~0 u~ 3UO 0. 20c00 0 10AD0 Covasna Huredoam TwiF Tulcea Vaslui BchaTest Rormnia Notes: Poverty line is 60 percent of mean consumnption per adult equivalent. Source: Rornania Integrated Household Survey 1998 (RIHS). Income transfers, including social insurance programs such as pensions and unemployment benefits and family benefits, play a significant role in lifting households out of poverty.15 In 1995 Romania introduced a means-tested social assistance program, designed to support the poorest households. However, because of the low level of resources allocated to the program, it has been weak in terms of coverage and mitigating poverty. In 1998 only 8 percent of poor households received the benefit and total expenditures on the program were less than 0.05 14 In particular, in 1998 the judets in our study ranked in the 2nd (Vaslui), 3rd (Tulcea), 4tD (Covasna) and 5th (Hunedoara and Tiniis) quintiles with respect to poverty. Ranks are from highest poverty ( ISt quintile) to lowest poverty (5th quintile). 5 Refer to Tesliuc, Pop and Tesliuc, 2000, for a full review of the effectiveness of transfer programs in poverty alleviation. percent of GDP in 1997. The Ministry of Labor and Social Solidarity (MMSS) is in the process of adopting new legislation which would strengthen the administration of the program and substantially increase the levels of the resources allocated to it. 16 Another critical item on Romania's social assistance reform agenda is the status of social services for vulnerable groups including children, the elderly and disabled. Child protection services in Romania, in particular, have received widespread international attention. Under socialism, governments placed an undue emphasis on institutional care at the expense of more humane, efficient and cost-effective community based services including home and day care (Tobis, 2000). In Romania the number of children in institutional care increased nearly 15 percent between 1990 and 1994 (UNICEF, 1997). However, since then Romania has taken important steps to restructure its system of child protection and has begun to promote the shift from institutions to community based care. Both the reform of social services and the steps toward strengthening cash benefits require substantial institutional reform and restructuring of the roles and responsibilities of the central and local governments, as well as communities and civil society. Education Like other countries in Central Europe, Romania's education system is fully developed and provides high levels of access to education for compulsory education. Official national statistics indicate that literacy is close to 100 percent. Enrollments in compulsory basic education (grades I-VIII) are high, and increased to 97 percent in 1999 (Table 1.4). In 1995/96 around 13,985 schools were devoted to primary and general secondary education with a total of 2,541,945 students and a teaching personnel of approximately 174,026. This translates into a relatively low student-teacher ratio (15), indicating the presence of many small schools, mostly in rural areas. Aggregate secondary education enrollments have declined during the transition period. This development reflects a decline in enrollments in vocational and technical education, which provide narrow technical training ill-suited to the needs of a market economy. In contrast, enrollments in general secondary schools, which provide more flexible academic education, and allow students to continue on to university, have more than doubled since 1989 (World Bank, 2000b). Table 1.4: Enrollment Rates, 1990-98 1990 1991 1992 1993 1994 1995 1996 1997 1998 Pre-primary' 54.3 51.9 53.3 50.2 55.2 58.4 55.1 -- -- BasiC2 89.5 89.4 89.6 90.3 91.4 92.6 93.9 95.0 97.0 Secondary3 90.8 76.2 67.0 63.4 65.7 68.3 68.1 67.3 -- Post-secondaryand 8.2 9.0 10.1 11.1 11.5 15.1 15.6 15.6 17.3 university4 Source: UNICEF, 2000; from official Romanian statistics. Notes: (I) net rates, % of 3-6 year olds; (2) % of 7-14 year olds; (3) academic and vocational education, % of 15-18 year olds; (4) gross rates, % of 19-24 year olds. Evidence on enrollments in education at the local level illustrates a more diverse picture across Romania, and highlights gaps in coverage, particularly in poorer rural areas and among 16 In January 2001, the name of the Ministry was changed to the Ministry of Labor and Social Solidarity from the Ministry of Labor and Social Policy. The new name, or abbreviation to 'Ministry of Labor' is used in this report. 12 Roma communities.17 In the five judets included in this study, enrollment rates for basic education in 1997 ranged from 91 percent in Covasna and Hunedoara to 98 percent in Bucharest. While official enrollment statistics were not available for local councils, interviews with parents and school officials in municipalities and communes suggested that children in some remote rural villages were not attending school, and that enrollment gaps were notable among some Roma communities. National household survey data found that in 1994, 41 percent of Roma children ages 7-14 were not attending school, in contrast with 7 percent of the population in general (Rughinis, 2000).18 The roots of these differences are discussed further in Chapter 5. Romania's education system shares many of the traits of systems in other countries in Central and Eastern Europe. In particular, education in transition countries is characterized by high levels of inefficiency in the allocation of resources, including education personnel and facilities.19 This is a result of the system of central planning, which determined the number and location of schools and teachers through centrally determined norms, rather than actual needs. This misalignment of resources and needs has been compounded by changes in the demographic composition of the country and changing prices of inputs, such as utilities (World Bank, 2000b). Addressing gaps in access across Romania, and the particular challenges of rural education provision is on the agenda of the Ministry of Education. Other priority reform issues for the education sector include raising learning outcomes through improved monitoring and evaluation and investments in teaching quality, as well as strengthened governance and management within the system.20 Results from the 1995 Third International Mathematics and Science Study (TIMSS), which compares learning outcomes for 8th grade students across countries suggested that Romania is lagging in education performance in these areas. Teaching generally focuses on rote memorization of facts, rather than creative thinking and functional literacy. Improvements in educational outcomes require a reorientation of the education system from a norm and input-based system, toward teaching and learning. This implies a rethinking of the roles and responsibilities of institutional actors within education to improve the flexibility of the system and to increase the accountability of schools for student outcomes. The following chapters address these issues as they relate to the increasing role of local governments in the education sector. Health The health status of the Romanian population is poor in comparison with neighboring countries and countries at similar income levels. In 1998, basic health status indicators, including male life expectancy, and infant and maternal mortality were significantly worse than those of other Central and Southern European countries (Table 1.5). Similar to other countries in the region, Romania's morbidity and mortality patterns are characterized by a disproportionate burden of disease from heart and circulatory diseases stemming from preventable lifestyle factors, including high levels of alcohol and tobacco consumption and poor diets (World Bank, 1999). Health status indicators vary across Romania. In the judets included in this study, 17 It is important to note that official enrollment rates do not reflect actual attendance. Gaps in coverage captured in the qualitative study may reflect students who are officially enrolled, but are not attending school. 18 Refer to Chapter 5 on the limitations of household surveys for analysis of Roma. 19 Refer to the World Bank's recent education sector strategy for Europe and Central Asia for more on this issue (World Bank, 2000). 20 For more on the status of the education sector see the education chapter of the forthcomning Romania Country Economic Memorandum, 2000. 13 according to administrative data, infant mortality in 1998 varied from 14 in Bucharest to 27 in Tulcea. Table 1.5: Health Status in Comparison, 1999 Male Life Infant Mortality Maternal Mortality Expectancy (per 1,000 live births) (per 100,000 live births) Romania 66.1 18.6 41.8 Czech Republic 71.4 4.6 6.7 Slovakia 69.0 8.3 10.7 Poland 68.8 8.9 4.8 Hungary 66.3 8.4 4.2 Bulgaria 67.9 14.6 23.5 Source: UNICEF, 2001 Health services, including hospitals and local health facilities such as dispensaries and polyclinics are widespread across the country. However, Romania's system is characterized by high levels of inefficiency in the allocation and use of services. Like other transition countries, health care in Romania was overly focused on specialized, tertiary care, at the expense of primary health services. The physical conditions of many health facilities in Romania are poor and deteriorating as a result of low levels of maintenance and a lack of capital investment in recent years. Romania's health system is undergoing a major systemic reform of financing and provision which aims to address these structural inefficiencies in the allocation of human and physical resources, and to realign the organization of the health system with the health needs of the population. The cornerstone of the reform was the passage of a Law on Health Insurance in 1997, which shifts financing, organizational and management responsibilities to a national health insurance system. This reform entails a significant change in management and provision arrangements for primary care. Physicians now provide health services in dispensaries and polyclinics under contract with the health insurance system. This has led to fundamental changes in incentives for providers and a restructuring of institutional roles and responsibilities. In particular, the role of the Ministry of Health has shifted from overall management of the system, to the lead agency for policy development and monitoring and evaluation. 1.6. Road Map of the Report All of the reforms discussed above have significant implications for intergovernmental institutional relations and the roles and the responsibilities of central and local government in service provision. Indeed, the success of the reform efforts will depend on the effectiveness of institutional arrangements, the presence of clearly defined roles and responsibilities, functioning incentive structures and smooth information flows across levels of government. The next three chapters describe the changes underway, in the context of the overall increase in local government autonomy, and assess the implications for service delivery. Chapter Two discusses developments in the financing and administration of local governments, and assesses the progress in implementation of these changes at the local level. Chapter Three describes the institutional context of the three social sectors covered in this study: education, health and social assistance, and the implications of institutional arrangements for the quality of service delivery. Chapter Four examines financing issues in the social sectors in more 14 detail, focussing on the changing roles of local governments and Chapter Five discusses changes in service delivery from the perspective of beneficiaries. Finally, Chapter Six presents the main findings and policy implications. 15 Annex JA: Description of the Study Methodology Institutional and financial analysis of local public administration. Analysis of local governments was conducted in the 5 judets, as well as 6 municipalities and 5 communes. The analysis consisted of two phases. First, budget data on local public expenditures and revenues were collected for all sites from the Ministry of Finance (MOF), sector ministries and their offices at the judet levels, as well as judets and local councils. These data were compiled into a database which is used for analysis throughout this report. Second, alongside collection of quantitative budget data, an assessment of the budgeting process and the roles and responsibilities of key players -- including budget officials at the MOF, judets, municipality, town and commune levels, and representatives of sectoral ministries at each level -- was conducted in each study site using key informant interviews. The study examined intergovernmental fiscal relations by identifying roles and responsibilities of the actors involved in the budget process on paper and in practice, and assessing the distribution of authority among state, regional and local officials to identify factors which impact the budgeting process and financing trends, including, inter alia, institutional and legislative controls over spending (e.g. constitutional, budget law, public administration controls over civil servant employment and pay), and existing capacity. Areas of emphasis in the interviews included budgeting processes and criteria for setting local expenditure priorities. Changes in the revenue structure of local government, especially the transfer system, and the impact on the distribution of resources across local governments. Interviews also examined the efforts of local officials to implement the new Law on Local Public Finance and the constraints to its effectiveness. Institutional analysis of social service providers. In parallel with the analysis of local governments, in depth analysis of institutions involved in social service provision was carried out. This included both collection of quantitative data, including detailed information on the financing of social services at all levels of government, and statistical infornation on social services themselves. Outcome indicators of social welfare are complicated by measurement problems (e.g. poverty and income) and timing issues (e.g. health indicators whose causes may be rooted in long term trends), and efficiency and effectiveness of resource use. As a result, quantitative data collection focused mostly focus on intermediate 'process' indicators to assess equity, such as expenditures and levels of service provision. The data collected from local governments, ministries and social service providers were compiled into a database which was analyzed for this report. In the tables presented here it is referred to as the "LSSD Database." In addition to data collection, in depth key informant interviews of social service providers in the areas of education, health and social assistance (cash benefits and social services) were carried out. These included discussions with officials in the central government ministries, at the level of deconcentrated institutions at the judet levels, as well as within social service institutions themselves (e.g. schools, dispensaries and child welfare institutions). The institutional analysis focused on defining the roles and responsibilities of different levels of government in providing social services, assessing factors which facilitate and/or impede effective functioning of the institutions, and understanding how key changes at the level of local governments influence service provision. Beneficiary assessment Assessment of the beneficiaries of social services was undertaken to gauge the population's actual experience with social services. The qualitative approach 16 allowed for the exploration of experiences and patterns of decision making. The beneficiary assessment was conducted in each municipality and commune in the study judets, as well as Bucharest Sector 5. Multiple methods of data collection were used including in-depth interviews with local key informants and focus groups with target beneficiaries of social services. In addition, nine case studies were undertaken of Roma communities, in order to document further the conditions and experiences of Roma with social services. Findings from these case studies are included here. 17 Chapter 2: Local Financing and Administration 2.1. Local Government Finance This chapter reviews the current situation of judets and local councils in terms of administration and financing, and reviews the main changes that have occurred with the adoption of the new Law on Local Public Finance (LLPF). The first section explains developments in local finance across Romania and the second section looks in depth at how these processes are unfolding at the local level, drawing from the fieldwork in the five judets under analysis. It examines how reforms in local finance have affected different types of local govenments, and the impact of changes in the intergovernmental transfer system on horizontal equity across local governments. The section also looks at coping strategies employed by different judets and local governments and assesses the existing capacity of local governments to perform their new functions. Prior to the adoption of the LLPF in 1998 local governments had limited fiscal autonomy.21 Before then the Law on Public Finances22 and the annual state budget laws governed the local budget process. Because local governments depended on intergovernmental transfers for over 75 percent of their revenues, the state budget law, adopted each year, served as the principal guiding legislation regarding local revenues and expenditures. Local governments did not learn the level of their revenues from national and local sources and the extent of their expenditure responsibilities for any given year until after the state budget was approved and published (often in May or June for the fiscal year beginning in January).23 Local governments had to wait for guidelines set by the state budget to approve their own budgets. If the state budget still had not been approved by the beginning of the fiscal year, the local govermments received authorization from the Ministry of Finance to spend one-twelfth of their total expenditures for the prior year each month. Sometimes, local governments also had to cover any expenditures at the beginning of the fiscal year which had been made on a temporary basis in the prior fiscal year. Key legislative reforms have fundamentally changed intergovernmental fiscal relations and the structure of the finances of communes, cities, municipalities and judets (Table 2.1). In particular, the LLPF, adopted in 1998, put local finances and the local budget process on an equal legislative basis with those of the national government. The law addresses basic principles of fiscal federalism including: (a) the alignment of expenditure responsibilities between the various levels of government; (b) taxation responsibilities (setting, collection, admninistration) of local governments; (c) transfers from the central government to local governments; and (d) the borrowing power of local governments. Amendments to the Law on Local Taxes and Fees 21 Law on Local Public Finances, No. 189 of 1998. 22 Law on Public Finance, No. 10 of 1991, as amended by Law No. 72 of 1996. This law should not be confused with the Law on Local Public Finances adopted in 1998. 23 The local budget process begins roughly in June or July of the preceding year with a request to the internal departments and local public institutions for a proposal for expenditures for the next year. Until fiscal year 1998, these estimates of future expenditures would then become the basis for a dialogue with the MOF on the level of transfers each local council would receive from the State budget. The individual budget of each local council was consolidated in a single county-level budget and forwarded to the MOF. The Ministry, in tum, gave each county and local council a spending limit that they used to finalize the draft budget for the next year. 18 authorized in 1997 and 1998 greatly expanded local control over their own revenues and authorized local councils to administer their own taxes. The Law on Public Patrimony, also adopted in 1998, created the basis for ownership by the local and judet councils of property associated with the functions they perforn. Table 2.1: Major Legislative Developments in Local Public Finance Year Law Focus 1991 Ratification of the Romanian - increased legal autonomy of local Constitution goveniments 1991 (amended 1996) Law on Local Public - defined the structure of local Admninistration, No. 69/91 governments 1991 (amended 1996) Law on Local Elections, No. - established election procedures for 70/91 local govenmments 1991 (amended 1996) Law on Public Finance, No - governed local budget process 10/91 1994 (amended 1997, Law on Local Taxes and Fees, - authorized local councils to administer 1998, 1999) No. 27/94 their own taxes 1998 Law on Local Public Finances - established framework for fiscal __(LLPF), No. 189/98 decentralization 1998 Law on Public Patrimony, No. - provided for division of public property 213/98 between central and local governments Fiscal decentralization in Romania has progressed for a number of important reasons, including the overall shift from centralized planning, and the commensurate transfer of increasing political and economic responsibility to lower levels of government. Fiscal decentralization has also been driven by the severe fiscal crisis associated with transition. Macroeconomic decline during the past decade in Romania has been dramatic. Aggregate GDP fell 18.9 percent between 1991 and 2000, among the deepest declines in Central and Eastern Europe. Following an initial drop in the early 1990s, GDP growth resumed between 1991 and 1996, and fell again in 1997 with delays in structural reforms (Table 2.2). These declines have been accompanied by a persistent fiscal deficit. Table 2.2: Macroeconomic Indicators 1995 1996 1997 1998 1999 2000 GDP growth (%) 7.1 3.9 -6.1 -5.4 -3.2 1.6 GNP per capita (US$-Atlas method) 1,400 1,450 1,420 1,430 1,520 1,640 Current Account Balance (US$ million) ' -1,732 -2,611 -2,360 -3,112 -1,302 -1,200 Current Account Balance (% of GDP) -4.9 -7.4 -6.7 -7.5 -3.8 -3.4 General Governnent Budget Balance (% of GDP) -3.4 -4.8 -5.2 -5.5 -3.8 -3.8 Inflation (CPI, y-o-y % change) 32.3 38.8 154.8 59.1 45.8 45.7 Unemployment rate (%) 2 9.5 6.6 8.9 10.4 11.5 10.5 1) Excluding non-convertible currencies. 2) As percent of total active civil population. Source: Romania CAS, 2001; Romanian National Conrnission for Statistics, 1999. The need to confront the budget shortfall provided an incentive for the government to shift increasing expenditure authority to lower levels of government. The increase in local government fiscal autonomy has taken place in a context of macroeconomic decline, and local 19 government revenues declined steadily between 1996 and 1998, both in absolute numbers as well as a share of GDP (Table 2.3). Central government transfers to local governments budget decreased in 1998 relative to 1996 and 1997. However, revenues from local taxes and fees increased in real terms in 1998 relative to 1997. This reflects the initial impact of the amendments to the Law on Local Taxes and Fees adopted in 1997 and 1998, that adjusted rates to inflation. Table 2.3: Selected Fiscal Indicators 1996 1997 1998 1999 2000 billion lei GDP 108,391 252,926 368,261 521,736 796,534 Total Govermnent Expenditure 36,810 85,639 131,123 193,567 283,566 Local Government Expenditure 4,955 10,370 13,382 21,599 33,217 Local Govemment Revenues 5,000 10,469 13,454 22,246 33,394 Transfers to Local Governments 3,821 8,436 10,052 4,308 4,308 Own Revenues of Local Governments (includes shared taxes) 1,130 1,984 3,328 17,935 17,935 % GDP Total Government Expenditure 34.0 33.9 35.6 37.1 35.6 Local Government Expenditure 4.6 4.1 3.6 4.1 4.2 Local Govermment Revenues 4.6 4.1 3.7 4.3 4.2 Transfers to Local Governments 3.5 3.3 2.7 0.8 0.5 Own Revenues of Local Governments 1.0 0.8 0.9 3.4 2.3 % Government Expenditures LocalGovernmentExpenditure 13.5 12.1 10.2 11.2 11.7 LocalGovernmentRevenues 13.6 12.2 10.3 11.5 11.8 Transfers to Local Governments 10.4 9.9 7.7 2.2 1.5 Own Revenues of Local Governments 3.1 2.3 2.5 9.3 6.3 1996=100 GDP 100 93 95 87 132 Total Government Expenditure 100 93 100 95 139 Local Government Expenditure 100 84 76 78 121 Local Government Revenues 100 84 75 80 120 Transfers to Local Governments 100 88 74 20 20 Own Revenues of Local Governments 100 70 82 286 286 Sources: IMF, LSSD Database This trend stabilized in 1999, when total local government expenditures rose to 4.1 percent of GDP from 3.6 percent in 1998. Local expenditures also increased as a share of total government expenditures. Revenues in 1999 were 20 percent lower than those in 1996, in real terms. The reduction in revenues occurred in all categories. In the case of "own" revenues the decline reflects the fact that the valuation of the local property tax base and the value-added tax (VAT) rates were not adjusted for inflation. These trends have been maintained in 2000. Revenues Local government budgets in Romania are comprised of three main sources: (i) own revenues, (ii) transfers and (iii) borrowing. The new LLPF has completely transformed the 20 revenue structure of local governments (Table 2.4). Most significantly, the main source of revenues for local governments has shifted from state budget transfers to own revenues. 24 The following section describes the main revenue sources for local governments and the changes since the introduction of the LLPF. Table 2.4: Total Revenues, Judets and Local Councils 1996 1997 1998 1999 Transfers (% of total) 76.4 80.6 74.7 19.4 earmarked (% of transfers) 48.4 44.9 50.3 41.4 general 51.6 55.1 49.7 58.6 Own revenues (% of total) 22.6 19.0 24.7 78.1 own taxes, fees and non-fiscal revenues 95.7 96.5 96.5 56.9 share of wage tax (from 99) -- -- -- 43.1 capital income 4 3 4 n/a Borrowing (% of total) 0.9 0.5 0.5 n/a Other (1) (% of total) 0.002 0.001 0.005 2.6 Total Revenues 100 100 100 100 Source: LSSD Database Notes: '--' = data not applicable; 'n/a' = data not available (1) data for 1999 includes borrowing, breakdown not available The LLPF along with the Law on Local Taxes and Fees greatly increased the autonomy of local governments over their own revenues. The revenues of judets and local councils have undergone significant adjustments, both in terms of their structure and levels. Local revenues have fallen since 1996, and revenues in 1999 were 80 percent of 1996 levels (Table 2.5). This shift from transfers to own revenues was particularly marked between 1998 and 1999, following the introduction of the LLPF. While transfers fell from 74 to 15 percent of 1996 levels, own revenues increased from 84 percent, to over 280 percent of 1996 levels with the introduction of shared taxes. Table 2.5: Judets and Local Council Revenues in real terms (1996=100) 1996 1997 1998 1999 Transfers 100.0 88.3 73.7 20.3 eannarked 100.0 82.0 76.6 17.4 general 100.0 94.2 70.9 23.0 Own revenues 100.0 70.2 82.4 276.5 Borrowing 100.0 42.1 43.5 n/a Other 100.0 65.5 220.2 n/a Total Revenues 100.0 83.8 75.4 80.1 Source: LSSD Database Notes: 'n/a' = data not available Own Revenues. Local government revenues include local taxes and fees and other financing sources, such as distribution of the profits of local utility companies (regii autonome), 24 In the literature on fiscal decentralization there is a debate whether shared taxes are transfers or own revenues. Here shared taxes are classified as own revenues to highlight the change in revenue structure. 21 and revenues from property (capital income). Until 1998 the most important local taxes were the personal and corporate property tax and the vehicle tax. Prior to the amendments to the Law on Local Taxes and Fees, local councils had no control over local tax rates or the valuation of the tax base. This was solely the prerogative of the national government which made few adjustments. Through 1998, own local revenues of the local councils never exceeded 25 percent of total revenues. Following the introduction of the LLPF, local councils were granted the ability to adjust the rate of the corporate and personal property tax within a range. This combined to increase the amount of local taxes and fees received by the local councils between 1998 and 1999 by over 70 percent in real terms. This source now accounts for over 50 percent of total local revenues, one of the highest percentages in Central and Eastern Europe. Both the amended Law on Local Taxes and Fees and the LLPF shifted responsibility for the collection of local taxes and fees to the local councils. This was to come into play on January 1, 1999. However, because neither the local councils nor the MOF were fully prepared, the laws were amended to delay implementation until January 1, 2000. The amendment also allowed individual local councils to request the authority to administer their own taxes at any point in 1999. Several local councils did so. At the same time, the LLPF introduced tax revenue sharing in Romania for the first time (Box 2.1). Under the new arrangements, the judets and local councils receive a share of the wage tax. According to the LLPF, the judets and local councils receive a share of the wage taxes collected from taxpayers in their geographic area of jurisdiction. The actual shares received by the judets and local councils can be modified in the annual state budget law. In 1999, 35 percent of wage tax revenues went to the local councils, 50 percent to the state budget and 15 percent to the judet councils. The remainder is allocated to the state budget. When the State Treasury receives payments from the taxpayers, it deposits a share into the account of the national government and the share of the judets and local councils into their respective accounts. In this sense, from a budget perspective, they are equivalent to the revenues that judets and local councils receive from their own taxes and fees. That is, the funds flow directly to the local treasury accounts from the taxpayers and there is no redistribution of taxes across judets. The important distinction between shared taxes and local taxes and fees is that local governments do not have decision making authority over taxation rates. There are several other important differences with the new system. Under the previous system, the state budget transfers had to be authorized annually. The new system puts an end to the effective subordination of local councils to the central government as the authority for revenue sharing stems from the LLPF and the revenue sharing provisions of the law are in effect each year, whether or not the state budget law has been adopted. Finally, the rules for revenue sharing between levels of government are clearly defined in the legislation. There is no discretion at the national level in the allocation of revenues, and to do so would require an amendment to the LLPF. Transfers. The LLPF greatly simplified the transfer system between the state and local governments. The introduction of tax sharing arrangements in that law reduced the overall importance of transfers in local government budgets. Most significantly, the LLPF scaled back the amount of earmarked, or dedicated, transfers provided to local governments, increasing the autonomy of local and judets councils over their expenditure allocation process. Between 1996 22 and 1998, transfers accounted for between 75 and 81 percent of total local government revenues, this dropped off dramatically to 17 percent in 1999 under the new law with the shift of the wage tax to shared taxes. Box 2.1: Shared Taxes: Advantages and Disadvantages With the reforms, own revenues, composed mostly of shared taxes, have become the most important revenue source for local governments. In Romania, shared taxes comprise a substantial share of total local revenues, higher than some neighboring countries in the region (Table 2.6). Table 2.6: Shared Taxes (% of total local revenues) % Year Bulgaria 52 1998 Czech Republic 42 1999 Hungary 15 1998 Poland 24 1997 Slovak Republic 24 1999 Romania 41 1998 Source: Wetzel, 2001 Tax sharing has distinct benefits including: administrative simplicity and continued central government control over the macroeconomic balance. Tax sharing may also provide incentives for local governments to increase revenue collection. However, there are substantial drawbacks, including: (i) Lack of capacity: local governments may lack administrative capacity and financial incentives to mobilize their own revenues - especially if transfers are designed to address fiscal disparities across regions. This has implications for the design of the transfer formula, as well as the need to incorporate incentives into tax sharing arrangements (e.g. related to collection rates). (ii) Unpredictability: shared taxes may be more susceptible to changes than other types of revenues. Rates for tax sharing can change annually, limiting the ability of local governments to plan. This is the case in Romania, similar to other countries in the region (Bulgaria, Ukraine, Georgia). (iii) Weak distributional impact: Because shared taxes are allocated based upon the local tax base where they were collected, regional disparities are reinforced in revenue collection. The poorest localities with the weakest revenue base will have the lowest amount of resources. (iv) Discretionary nature: Finally, shared taxes can increase discretion within intergovernmental fiscal relations, as they encourage lobbying and negotiation over tax rates. These factors underscore the risks of increasing reliance on shared taxes. As has been the case in Romania, shared taxes can increase regional disparities and reduce incentives for local tax collection. Source: Wetzel, 2001. 23 Transfers under the old system. Transfers were nearly equally divided between general (untargeted) and earmarked transfers, prior to the introduction of the LLPF. In theory, the total amount of general transfers allocated to local governments was based upon the total amount of revenues received from the wage tax. However, in practice the amount of transfers distributed were based upon overall budget constraints, rather than the amount of revenues collected from the wage tax, and transfers worked exactly as if they were funded from the general revenues of the national government. Econometric analysis of transfer data for 1993 and 1994 shows that transfers were not targeted to poorer Judets, as a positive correlation between per capita transfers capita and per capita revenues emerges (Jakob 1995). Until 1996, the local and judet councils were obliged to maintain minimum spending levels for certain sectors (including education, health and social assistance) to be financed from the transfers. This was discontinued in fiscal year 1997. The earmarked transfers from the state budget to the judet and local councils until 1998 were of two types. The first were subsidies paid by the judet and local councils to local heating companies to compensate for national price caps and to the public transport companies to compensate for free transportation benefits for certain categories of persons (disabled, veterans). The second type of earmarked transfers were subsidies to finance local investments. In the case of the operating subsidies, the national government retained the right to set prices and provide specific benefits related to services which were nominally under local control. In the case of investments, the national government retained the right to approve each investment project of the judet and local councils, and to decide the volume of investments at the local level.2 Both the general and earmarked transfers provided considerable discretionary authority to the judet councils. The process for allocating the transfers had three phases. In the year preceding each fiscal year, the MOF reviewed the consolidated draft budgets of all judets, and used the deficit in the budget as the basis for determining the total transfer for each judet. The final amount by judet was set in the annual state budget law. Then each judet council would allocate the total to each local council within its geographic boundaries and to itself. There were no rules to guide this process, and, as a result, from the perspective of local councils, the general transfers were arbitrary and negotiable. At the beginning of the process, each judet and local council would prepare a "deficit" budget showing the difference between revenues from their own sources and planned expenditures. Their expectation was that the MOF would agree to cover the "deficit" through transfers from the state budget. The MOF would review these figures and provide back to each council its "spending ceiling," usually some amount smaller than the "deficit." Transfers under the new system. The new legislation greatly simplified the transfer system, and reduced the amount of discretion applied in allocating revenues across local governments. All dedicated transfers for operating subsidies of public service companies and investment subsidies to the judet and local councils were eliminated.26 In this context, the LLPF 25 In theory, each judet or local council was free to provide additional funding for investments from their own revenues. Many took advantage of this option. However, the investments subsidies from the State budget still accounted for the vast majority of local investments. 26 Only the funding for local investments obtained from foreign donor loans made directly to the national govermnent remain as dedicated investment subsidies in the new system of local finances. Heating and transport subsidies that had been funded through transfers are now funded by local and judet councils from their general revenues. 24 also reduced the authority of the national government to approve local investment projects to larger projects above certain minimum cost levels, and to those funded partially or fully by the state budget. The general transfer was replaced with an 'equalization grant' which aims to correct for differences in expenditure responsibilities and fiscal capacity among the judet and local councils. The equalization grant is funded through a transfer from the state budget that is allocated directly to the judet councils for redistribution by them to the local councils in their area of jurisdiction. The state budget also specifies the amount of these equalization grants that the judet councils may retain for their own use, within a limit of 25 percent set in the LLPF. The total amount of funds that was distributed to the judets as the equalization grant in 1999 was ostensibly based upon projections of the shortfall between estimated local government revenues and expenditures. However, in practice it was based upon available resources. The grant is allocated among the judets based upon a formula defined in the state budget (Table 2.7). The formula can be changed annually with the state budget law, limiting the transparency and predictability of the system. The formula was similar to that which is used for allocating transfers under the old system, with the addition of a criteria for local fiscal capacity. In the 1999 formula the fiscal capacity criteria had a weight of 30 percent. The remainder of the formula was based upon indicators of demographic and physical infrastructure (Romanik, et al., 1999). The allocation of the grant from judets to local councils is based upon the same formula, however this was not implemented. Table 2.7: Criteria for Allocation of Transfers (Sume Defalcate), 1997-99 1997 1998 1999 Population 12 8.5 5 Length of city streets 4 8.2 5 Length of roads 2 15 5 Number of housing units 2 5 5 Length of water and sewage network 3 10 8 Number of pre-university students 34 45 25 Number of children in orphanages -- 3.3 15 Number of territorial-administrative units 3 5 2 Fiscal capacity -- -- 30 Number of hospital beds 34 -- -- Number of exemptions 6 -- -- Total 100 100 100 Source: Romanik, et al., 1999 Processes for determining the allocation of the equalization grant differed greatly across judets. As highlighted in Box 2.2, none of the judet councils included in this study followed the formula outlined in the state budget law. Most judets councils devised their own criteria, some based upon norms (e.g. population size), others on the revenue raising capacity of the municipality or commune. The most common criteria across the five judets in the study were the financing needs for own administration of the local councils and social service institutions for 27 elderly and children. Analysis presented in the next section of this chapter illustrates that the change in the transfer mechanism led to increasing horizontal inequality across judets. As stated in the LLPF, the objective of the equalization grant is "to create a budgetary balance among certain territorial-administrative units." However, in practice, the effectiveness of 27 The latter is not surprising, as social services are decentralized to the county and muncipal level. 25 the transfer in reaching its equalization objectives was limited by (i) the unpredictability of the funds available for the transfer; (ii) the limited targeting effectiveness of the formula; and (iii) the lack of accountability at the judets level for applying the formula (Romanik, et al., 1999). Box 2.2: Allocation of Equalization Grants by the Judet Councils In 1999 judet councils adopted varying strategies for allocating the equalization grant among local councils and setting guidelines for the use of funds. In Timis, the judet council adopted explicit norms for allocating the resources, including a measure of fiscal capacity calculated based upon the local council's revenues and expenditures during the prior year and administrative expenditures, as well as 8 other indicators such as the number of pre-school age students, number of children in foster care, total population and length of the water, sewage and road networks in the locality. Other judets used broader guidelines, rather than a specific formula. For example, in Tulcea allocation was based on the level of own administrative expenses of the local council, needs in social protection services and on-going investment projects. Other factors included operating expenses for homes for the elderly (Hunedoara and Vaslui), and expenditure responsibilities issued after the passage of the LLPF (Hunedoara and Covasna). All of the judets in the study used the level of own administrative expenses as a prionty criteria. The judets also employed differing approaches to the use of the grant funds. In most cases the objective of the funds was to ensure minimum operating expenses of the local councils. Timis set more specific parameters on the use of funds allocating 58 percent for social assistance; 28 percent for fiscal capacity; 7 percent for education and 7 percent for covering prior-year arrears. Hunedoara opted to set aside 30 percent of the equalization grant for emergency circumstances. Source: Interviews with judet officials. Borrowing. One of the most significant changes included in the LLPF was the provision for local governments to obtain credits from domestic and international credit markets, through loans and issuance of bonds. The LLPF sets a ceiling for borrowing such that total debt service may not exceed 20 percent of own revenues, including the shared wage tax. Local governments do not require a governmental guarantee for domestic loans. However they do have to send prior notice to the Ministry of Finance. For external credits, local governments require the approval of a newly established "Commission for Authorization," whose members are nominated through a Governmental Decision and include representatives of local public administration, the central government and the National Bank of Romania. Experience to date with local government borrowing has been extremely limited. In 1999 budgeted revenues from borrowing of judets and local councils were 0.7 percent of total local government revenues. Very few local governments have taken advantage of the new provisions for borrowing, largely because of current weaknesses in the banking and financial markets, high borrowing costs and unclear ownership rights on property which could be used as collateral. Only one of the five towns, and one of the five communes included in this study took out loans in 1999, both from the state treasury. Lack of experience and administrative capacity was also identified by local officials as a constraint. Expenditures The LLPF provided clear authority to local governments to adopt and amend their own budgets independently of the state budget and clarified expenditure responsibility across levels of government for the first time. Although the new legislation expanded the scope of local 26 government autonomy over public spending in certain areas, expenditure management remains tightly proscribed. According to the LLPF authority of the MOF to review local budgets is limited by law to confirming the estimates made by the local goverments of equalization grants they might receive from the state budget. However in practice the central government has retained control over the budget process in a number of ways. The central government monitors local budgets through prior MOF review of the annual budgets of judets and local councils, a quarterly report on budget execution to the MOF, and prior review by the MOF for the majority of investments of local and judets councils. In addition, if the state budget is approved until after the beginning of the fiscal year, local governments - even those that do not receive transfers from the state budget - are subject to controlled spending at a rate of one-twelfth of prior year expenditures. Under the new system, the judet and local councils should be able to adopt their budget effective January 1 of each year, using their own estimates of local revenues and the estimates approved by the MOF of transfers from the state budget, regardless of whether the state budget has been approved. However, this has not yet occurred. Finally, expenditure decisions are controlled within sectors by central government norms on service provision. These issues are discussed in depth for social services in chapter 4. Notable developments in the area of local expenditures have included the increase in expenditures on local administration, brought about by the new legislation and the need for new personnel at the local level to manage the implementation of the LLPF. Other developments have included the shift of responsibility for financing health care from the state and local budgets to the health insurance system in 1998, and the increase in responsibility for financing education, with the transfer of responsibility for educational expenditures including materials, capital expenditures, scholarships, transportation, and equipment for the educational process to the local governments. Table 2.8: Total Expenditures of Judets and Local Councils 1996 1997 1998 1999 budget Local administration 1) 6.9 7.9 13.4 15.4 Education 11.9 10.0 9.0 9.3 Health 12.4 12.8 0.5 0.5 Culture, religion and sports 3.2 4.7 4.9 4.3 Social assistance 6.3 4.6 9.3 7.4 Public works and housing 37.6 39.0 41.3 38.4 Transport and communication 19.9 19.6 20.3 12.6 Other economic activities 0.0 0.1 0.0 0.1 Other activities 2) 1.5 0.9 1.0 1.0 Agriculture and forestry -- 1.1 Special destination expenditures 3) 9.0 Repayment of loans and interest payments 0.2 0.4 0.2 0.4 Reserves -- -- -- 0.6 Total Expenditures 100 100 100 100 Notes: 1) Own expenditures of the local government. 2) Includes military commnands, fire brigade, civil guards. 3) Expenditures from targeted transfers. Source: LSSD Database The predictability of expenditure responsibility is an important issue. Like other countries in the region, in a constrained fiscal enviromnent, decentralization in Romania has led to the off- 27 loading of central government fiscal responsibilities to local governments without ensuring that local governments have the necessary revenue raising authority. In Romania, this issue has been compounded by delays in the passage of the state budget law which defines expenditure responsibilities. For example, the 1999 state budget law was not passed until March 1999. This led to the transfer of financing responsibilities late in the fiscal year after budgets have been allocated. An example in 1999 was the shift of student scholarships, which were then transferred back to the central government in 2000. The lack of fiscal autonomy in expenditures also increased due to the fact that the 1998 Law on Local Public Administration is largely silent on the issue of the functions and responsibilities of local governments at all levels. Each year since 1991 the state budget law has added, modified, or removed local responsibility for specific services or for certain aspects of these services. Responsibilities for expenditures of both judet and local councils have changed frequently. Even the direction of the change was difficult to predict. Responsibility for expenditures shifted back and forth between the national and local governments and between judet and local councils. Local governments also moved certain expenditures off budget by shifting them to local entities under their authority. The government has prepared a major amendment to the Law on Local Public Administration that is currently before the Parliament in Romania. This new version of the basic local government law still remains largely silent on the definition of local fimctions and responsibilities. The LLPF does include such a definition. In an annex to the law there is a complete list of local responsibilities.28 However, the law also states that the annual state budget law can modify this list, which happened in both 1999 and 2000. Until 1998 there were no laws dealing with public property, at both the national and local levels. Effectively the national government was in control of all public property, including the property directly associated with functions performed by the local governments. This made it easier for the national government to modify local responsibilities by shifting responsibility for property ownership. For example, a specific health facility might become a municipal responsibility at one point, that of a judet the next, and finally a national responsibility. Shifts of responsibility have occurred in the middle of the fiscal year. Lack of clarity on property ownership has created uncertainty and reduced incentives for local governments to invest in maintenance of facilities and capital investments. The implementation of the Law on Public Patrimony passed in 1998 will clarify property ownership. Once a specific property, a school building for example, has been registered in the name of a local government, it will not possible to shift responsibility for maintaining and improving this property to another level of government through the state budget law. In effect, the process of registering the title of specific properties will help clarify and stabilize local functions and responsibilities. This will strengthen the value and usefulness of the local budget as a financial management tool. However, more than one year after Parliament adopted the Law on Public Patrimony, no transfer of property has yet occurred. Reviews of the ownership of health and education facilities have recently been completed, and a similar analysis for social assistance offices and institutions is underway. 28 LLPF, Chapter 2. 28 Changes in expenditure assignment are unpredictable and have a differential impact across judets. For example, as a result of new rules adopted for budget year 2000, local councils will have to pay a fixed amount per child in a child-care institution to the corresponding judet council. This promises to be a major new expenditure for the local councils. Yet, when the state budget was adopted, the amount that each local council would have to pay had not yet been determined. This is the responsibility of each judet council. It is only in the second half of the budget year that they have begun fixing the amount. The impact on individual local councils will be a function of the number of children from each community who are in child-care institutions. The same is true for the heating price subsidies paid mostly by the larger municipalities that have a district heating plant in their area of jurisdiction. The amount of the subsidy is the difference between the approved price for heat delivered to residential consumers in each city and a national reference price established by the national government. The heating subsidy has become the largest individual expenditure for most municipalities, representing anywhere from 30 to 50 percent of recurrent revenues. The national reference price often changes after the beginning of the budget year. The problem is that the approved price for heat delivered to residential consumers is different in each city. Therefore, the amount of the subsidy (the difference between the approved price and the national reference price) varies from city to city. The transfer to local funding of all ongoing investments of the Ministry of Education (MOE) starting with budget year 1999 had a similar effect. The impact on individual local councils varied according the what investment projects that the MOE might have in a given community. Some local councils had to absorb significant new investment expenditures, while others had almost none. The MOE did not inform each local council of the projects they would have to fund out of their local budgets until almost halfway through the 1999 budget year. The unpredictable effect of the changes in expenditure assignment has detracted from the positive impact of the changes in the structure of local revenues produced by the, amendments of the Law on Taxes and Fees and the adoption of the LLPF. For example, it might appear that the failure to implement the equalization provisions of the LLPF is a function of the revenue structure of local governments. In fact, it is also function of the instability in the expenditures of local governments. 2.2. Implementation at the Local Level In 1999 total local revenues increased as a share of GDP and in real terms. However this recovery did not keep pace with trends in GDP and government expenditures. At the local level the experience across judets and local councils has been diverse. While some local governments were able to take advantage of new opportunities for raising revenues, others have not (Figure 2.1). This section discusses the impact of the LLPF on local governments drawing from judet level data on financing and the individual case studies. 29 Figure 2.1: Total Revenues in Real Terms (1996=100) 120 100 80 01996 M 1997 60 0 1998 01999 40- 20- Bucharest Covasna Tulcea Timis Vaslui Hunedoara Romania The variation in revenues of all local governments in Romania increased with the LLPF (Table 2.9). This has largely been driven by the increase in variation of transfers received by local governments. With the introduction of shared taxes and the equalization grant, the coefficient of variation, which measures the variance of transfers received across the judets, grew from 0.19 to 0.35 between 1998 and 1999. In contrast, own revenues became relatively more equal during the same period, although the level of own revenues varies more significantly than transfers across local governments. It is important to note that growing variation is not necessarily a negative development, and may reflect increasing efficiencies through adjustments to local prices and demands for services. However, as the in-depth discussion of social sector financing in Chapter 4 illustrates, in a number of judets, declining local revenues have led to a collapse in financing of basic social services and cessation of services in some cases. Table 2.9: Coefficient of Variation, Per Capita Revenues of Local and Judets Councils (1) 1996 1997 1998 1999 Total Revenues 0.30 0.34 0.34 0.36 Transfers (2) 0.23 0.19 0.19 0.35 Own Revenues (3) 0.42 0.53 0.52 0.44 Notes: (1) coefficient of variation is the standard deviation/mean (2) transfers in 99 include the equalization grant and targeted transfers (3) own revenues in 99 include shared taxes Further analysis of the data on per capita revenues in each judet illustrates that the increased variation in the allocation of transfers across judets does not reflect improved targeting of transfers to poorer judets. In fact, basic regression analysis found that judets with higher levels of own revenues per capita received greater amounts of transfers per capita (Table 2.10). Rather than having an equalizing effect, transfers are allocated to those judets which have had greater success in raising own revenues. The change in the transfer mechanism following the introduction of the LLPF strengthened the relationship between local fiscal capacity and the 30 amount of transfers received by the judet. In other words, judets which were more successful at raising own revenues (taxes and fees) in 1998 received more transfers per capita in 1999 than they had in 1998. Analysis also found that the level of poverty in the judet had no influence on the level of transfers allocated to the judets, further indicating the limited targeting of the system. Table 2.10: Transfers per capita on previous year own revenues and poverty 1997 1998 1999 Own revenues per capita (previous year) 0.22** 0.16* 0.26** (-3.55) (2.10) (3.19) Poverty Rate (previous year) -83.73 231.44* 365.79 (-0.40) (2.08) (1.65) Constant 57764.63** 41101.53** 848.12 (5.83) (5.29) (0.07) Source: LSSD Database Notes: OLS regressions; t-statistics in parentheses;** indicates significance at 5% or more; *indicates significance at 1 0% or more; regressions are for 40 judets, excluding Bucharest and Ilfov which are outliers This analysis highlights the need for a reevaluation of the objectives of the transfer system and its equalizing role across judets. Due to the lack of data on the allocation of transfers from judets to local councils, it was not possible to assess the targeting of transfers to municipalities and communes. However, analysis of budget data for the case study sites illustrates the vulnerability of local councils to fluctuations in resource levels. Total revenues in some of the sites have declined dramatically since 1996. Most notably, budgeted revenues in 1999 in Sarachioi, Tirgu Secuiesc, Babadag and Lugoj remain between 29 and 43 percent of 1996 levels. Dramatic changes in revenue levels can, in some cases, be explained by local developments. For example, Zapodeni received a targeted transfer for telecommunications investments in 1998, explaining the significant increase during that year. Wide variations in resource availability from year to year highlight the uncertain fiscal situation faced by local councils and underscore the need for mechanisms to improve the predictability of revenues across years. Table 2.11: Trends in Local Revenues in the Study Sites (1996=100) 1996 1997 1998 1999 (budget) Covasna Sinzieni 100 86 65 51 Tirgu Secuiesc 100 75 49 33 Hunedoara Vorta 100 89 99 122 Lupeni 100 83 51 56 Timis Nadrag 100 88 106 75 Lugoj 100 86 61 43 Tulcea Sarichioi 100 56 43 29 Babadag 100 73 63 42 Vaslui Zapodeni 100 84 163 60 Vaslui 100 74 53 78 31 Coping Strategies Local governments have addressed resource constraints in different ways, including cutting current expenditures, off-loading responsibility to other local governments (e.g. from judets to municipalities), and raising additional resources as allowed by the LLPF (Box 2.3). In 1999 a few local councils began to take advantage of new opportunities for revenue raising. Most comrnon was property rental. All of the towns, and all communes, with the exceptions of Zapodeni and Nadrag, rented out public land. Despite the new legislation, borrowing remains relatively rare. Only two local governments, Vaslui and Nadrag, took out loans from the state treasury. Introduction of new taxes and fees was also rare. Lugoj adjusted the rate of the turnover tax. Zapodeni introduced fees for dog and cart ownership, while Lupeni reduced levels of local taxes by 50 percent. Other local governments raised revenues through property sales and the establishment of joint ventures. Box 2.3: Coping Strategies at the Judets Level Judets councils have dealt with the tight fiscal situation in differing ways. In most of the study cases, judets undertook measures to reduce overall expenditures. Vaslui, Tulcea and Covasna decreased operating expenses by reducing staff numbers in the judets administration. Vaslui rotated some staff on unpaid leave to reduce the wage bill. Vaslui and Tulcea also undertook steps to limit operating expenses for fuel, telecommunications and materials. Another strategy has been to devolve expenditure responsibility further to local councils. Covasna transferred expenditure authority for some cultural activities to a municipality. Hunedoara was the only judet in the sample which identified a explicit strategy for dealing with fiscal constraints. The judet council maintained a reserve for its own operating expenses, equal to two months own administrative expenditures, as well as a reserve for local councils, by retaining 30 percent of the equalization grant to allocate for emergency situations. Most judet council officials expressed frustration with the budgetary situation and felt that fiscal control was out of their hands due to rising inflation, the unpredictability and uncertainty of wage tax revenues and own revenue sources. Source: Interviews with judet council officials Local officials explained that they had been reluctant to make more extensive use of new resources. In Vaslui, the mayor was concerned about the difficulty of imposing additional taxes and fees given the difficult economic conditions in the judet. Other issues which were raised included the high costs of borrowing, and a lack of clarity in property ownership which also limited borrowing because of the lack of collateral, as well as revenues from renting land. Administrative Capacity Increased local government fiscal responsibility implies a significant increase in the need for local capacity in fiscal management, tax collection, and budgeting. Expenditures on local administration did increase as a share of total local expenditures in 1998 and 1999 (Table 2.5) and were protected in real terms. Limited training and guidance were provided to local governments to cope with the changes. As mentioned above, because of the lack of preparedness, the full introduction of the LLPF was delayed until 2000 in most areas. 32 At the time of the study most judets and local councils had not significantly altered their organizational arrangements to adopt to the provisions of the LLPF, although all had set up revenue departments. The deconcentrated judets-level offices of the MOF, the General Public Finance Directorates (GPFDs), continued to assist with tax collection responsibilities in 1999. As was apparent in the discussion of financing trends, during the first year of implementation of the LLPF most judets focused on ensuring compliance in the payment of the wage tax, the most important revenue source, rather than other local taxes and fees. Judets and local councils which had the most to gain from taking on tax collection responsibility were best positioned and most proactive in setting up their new revenue departments and bringing in qualified staff. Municipalities such as Vaslui and Lugoj have larger local administrations than towns and communes. This a function of both the size of the locality and the availability of qualified staff. In Vaslui and Lugoj at the time of the interviews half of the personnel in the new revenue departments held academic degrees, while in the communes very few employees in the local government administration had higher education credentials. In both judets and local councils, the main obstacles identified by local officials were the lack of training and guidance from the GPFDs and the Department of Local Public Administration in the central government and the lack of monitoring both at the judet level and of local councils. Not surprisingly, rural communes have the least capacity for revenue raising. Local officials in judets and local councils lamented the lack of a strategy at both the level of GPFD and Department on Local Public Administration on how to handle the transfer of this activity. The only assistance provided was training of local council officials organized jointly by the GPFD and the judet councils and the transfer of data on taxpayers and delinquent accounts from the GPFD to the local councils. There was no guidance or assistance on the way this new activity should be organized and operated by the local councils. Monitoring and support to local governments was also missing. Contacts between GPFD and mayors regarding collection activities were reportedly sporadic. Generally, the contacts originate with the mayors in those localities where the new tax personnel do not have prior experience working for the GPFD. 2.3. Summary Issues The legal framework established in the LLPF and other recent legal reforms has expanded the financial autonomy and authority of judet and local councils in Romania. Positive developments in this respect include increasing clarity in intergovernmental fiscal roles and responsibilities, as well as strengthening local governmental fiscal accountability. Although it is still too early to gauge the overall impact of the new legislative framework, the experience of the first year of implementation highlighted key issues which require monitoring and policy attention. While the establishment of the new legal framework for intergovernmental finance represents a significant achievement, there is a need for clear legislative language describing not just the functions of local governments, but also the authority they have to carry them out. 33 Chapter 3: Institutional Arrangements for Service Delivery 3.1. Introduction As far as institutional arrangements are concerned, the ability of local governments to succeed in their roles and to ensure the effectiveness, efficiency and equity of social service delivery - both through direct provision and in partnership with other institutions - rests on three broad sets of factors (i) local capacity, including human and financial resources; (ii) the quality of intergovernmental relations, including clearly defined roles and responsibilities which ensure transparency, accountability and correctly aligned incentives; and (iii) information to identify local needs and preferences, target resources and design effective programs. This chapter focuses on the institutional setting for service delivery. Issues of financing are addressed in detail in the next chapter. The chapter first provides an overview of the current role of local governments in service delivery, and then discusses institutional arrangements and changes with recent sectoral reforms. Tables included at the end of this chapter illustrate the distribution of responsibilities for service provision across levels of government. Finally, it assesses issues related to intergovernmental relations and their impact on the delivery of services. Local Governments and Social Services Social assistance cash benefits and services are the most decentralized policy areas among those covered in this report. The discussion in this chapter focuses on the social assistance programs that entail a substantial involvement of local governments: cash benefits, particularly the means tested social assistance benefit (MTSAB), emergency aid, and birth grants, child protection, services for the disabled, and a number of in-kind services. Local councils are direct providers of cash benefits for low income households. They oversee the implementation of benefit programs including budgeting, screening eligibility of beneficiaries and actual payment of benefits. Local governments also have important responsibility for in-kind social assistance including programs and institutions for children and the disabled. Child protection is the only social service which has been completely decentralized to the judet level.29 All decisions regarding provision and financing of programs are undertaken by judet level institutions. Local governments are also responsible for coordinating policies and projects with NGOs and donors, including developing and implementing projects, and contracting out services. Local governments also have critical responsibilities for determining eligibility of beneficiaries and allocating resources for social assistance programs and institutions. In contrast, the role of local governments in education is mostly confined to financing. With the new LLPF, responsibility for financing all capital expenditures and some current expenditures, with the notable exception of personnel expenditures, was transferred to local governments. In the health sector the role of local governments has decreased with the reform process. However local councils do have involvement in areas such as property ownership and aspects of personnel policy, such as setting incentives for providers. 29 Child protection services were recentralized beginning in 2001. This discussion focuses on the system prior to the change. 34 3.2 Institutional Arrangements in the Social Sectors Social Assistance Social Assistance Cash Benefits and Services. Local councils have responsibility for delivering social assistance cash benefits for low income families including the means tested social assistance benefit (MTSAB) and emergency assistance. Local councils also provide birth grants which are given for all children. Delivery of benefits is undertaken by local government employees who are hired, fired and trained by local councils. The central Ministry of Labor and Social Solidarity (MMSS) initiates new programs and establishes broad principles of benefit eligibility at the national level and has a coordinating role, through its deconcentrated offices at the judet level, to evaluate staff quality and collect data on the number of beneficiaries and expenditures of different programs. However, because staff who implement the programs are under the jurisdiction of the local councils, the MMSS has limited authority to influence implementation of its programs. Social assistance services are provided by a diverse set of institutional actors. Responsibility for provision of services for poor households, the elderly, children at risk and the disabled fall under central and local institutions, including: (i) the Department for Child Protection and the decentralized Child Protection Directorates at the judet level; (ii) the State Secretariat for Handicapped Persons (SSHP) and its deconcentrated offices at the judet level; and (iii) judets and local councils. While decision making for child protection is decentralized to the judets level, decisions regarding services for the disabled are made centrally through deconcentrated organizations. Still, other institutions, including soup kitchens, fall under the responsibility of local councils. Child Protection. Beginning in 1998, responsibility for managing and financing child protection services has rested with two institutions at the judet level, the Judet Child Protection Directorate (CPD), the implementing agency, and the Judets Child Protection Commission (JCPC), a decision making body comprised of representatives of all deconcentrated institutions in the judets. CPDs have the main responsibility for child protection activities, and have authority over personnel policy, establishing eligibility of beneficiaries and monitoring. Although Judets Child Protection Commissions are responsible for decisions related to the types of services and programs provided in a judet, proposals and resolutions which are adopted by the Commission are generally proposed by CPDs. In some cases the judets councils play an important role in setting policies, but this varies across judets, as it is not regulated by specific provisions. Interviews with local officials indicated that the level of involvement of the judet council depends mainly on the commitment and initiative of the secretary of the judet council, who is also the president of the Commission. At the national level, child protection activities are overseen by the central National Authority for Child Protection and Adoption (NACPA).30 The central Department does not implement programs, rather, it is responsible for developing national strategies, initiating legislation in the field of child protection, and coordinating international assistance programs. Local councils have a role in implementing some child protection activities, particularly screening beneficiaries for eligibility for services. 30 The Department of Child Protection was founded in 1998, it's name was changed to the NACPA in 2001. 35 Other local institutions share a role in child protection, including most notably the police, which oversees initiatives related to juvenile crime and child abuse. Collaboration between the police and local authorities is close in many cases, as CPDs do not have hotlines or other mechanisms to react directly to emergency situations. In cases of violence or abuse, the police refers cases to the local councils, which in turn request assistance from the Child Protection Directorates in the judets. Services for the Disabled. Social services for the disabled fall under the responsibility of the State Secretariat for Handicapped Persons (SSHP) and its deconcentrated institutions at the judet level, the Judet State Inspectorates for Persons with Handicap (JSIPH). The Judet Inspectorates implement policies set centrally. The State Secretariat is responsible for the full range of decisions regarding the type and number of facilities provided in each judet, personnel (hiring, firing, setting and paying salaries), and establishing eligibility for services and monitoring. The Judet Inspectorates play an advisory role to the State Secretariat by making recommendations regarding the number of facilities and the number of staff needed. Local councils have a significant role in financing and administering services for the disabled. They are responsible for personal care services, benefits for transportation and telecommunications, and institutions for the disabled. Local councils are also responsible for assessing eligibility for services based on central guidelines. The level of involvement of individual local councils depends on the number and type of services provided in the area. Some local councils are also involved in supervising the activities of social workers who work with the disabled. The amount of contact between local councils and JSIPHs varies significantly. Local Governments. In addition to the responsibilities for delivering social assistance cash benefits described so far, local councils screen beneficiaries for entry into institutions for child protection and the disabled, and monitor the welfare of foster children. Some local councils also receive applications for benefits paid out of the state budget, such as child and family allowances, although these benefits are delivered by the deconcentrated offices of the Ministry of Labor. Finally local councils operate social services, most commonly soup kitchens. In principle, judet councils do not have direct responsibility for social assistance provision beyond their involvement in child protection, however in practice they play varied roles. In some cases they attempt to assist local councils by promoting investment and activity from donors and NGOs, or by channeling information on specific issues to the central government. Judet councils also have the capacity to address urgent financial constraints of local councils through the allocation of the equalization grant, as discussed previously. However, because the funds are not earmarked, judet councils cannot ensure that local councils spend the funds for their intended purpose. Monitoring and Evaluation. Local governments are also responsible for communicating information on social assistance programs, including data on financing and numbers of beneficiaries, to the deconcentrated offices of the Ministry of Labor at the judet level. However, under the current system, no institution is responsible for monitoring the effectiveness of the social assistance programs, and whether they address the poverty needs within the locality. Because the means tested benefit program is implemented by officials in local councils, the Ministry of Labor is unable to address gaps in financing or other issues with implementation at the local level. As discussed further below, new legislation proposed for social assistance will 36 strengthen the role of the Ministry in monitoring the delivery of social assistance at the judet level. Education Provision of education in Romania remains substantially more centralized than social assistance, but is involving an increasing number of institutional actors. Local councils have growing fiscal responsibility and pilot reforms are underway to delegate decision making in some policy areas, particularly personnel, to individual schools. The Ministry of Education (MOE) is in charge of nation-wide policy making, teacher training and evaluation, personnel, research, libraries, reform strategies, quality assessment, and development of the methodology for the national examination at the end of compulsory education cycle.31 The Ministry of Education is also responsible for general education administration.32 The Ministry's deconcentrated bodies - the Judet Education Inspectorates (JEIs) - are responsible for carrying out the administrative side of these duties, including inspections, personnel policy, organization of training courses for teachers and principals, and coordinating national examinations. Judet Inspectorates are relatively large structures and their staff includes inspectors and financial, and human resource experts. Judet Inspectorates are presently undergoing a substantial restructuring which involves, amongst other issues, a reform of the school inspection system. The role of local governments in education is essentially limited to financing. Beginning in 1995, and especially after 1998, local councils have been delegated growing fiscal responsibility for education and now fund a wide range of expenditures including spending for infrastructure, maintenance, and school equipment. In towns, funds are first channeled through Centre de Executie Bugetara (CEBs), literally 'budgetary execution centers.' These are usually large schools that act as budget intermediaries between local councils and Judet Inspectorates on one side and schools on the other.33 These issues will be discussed further in the next chapter. Personnel. Decisions regarding personnel have been traditionally centralized, as Judet Inspectorates have been charged with hiring and firing of teachers and appointing principals. Judet Inspectorates also organize national hiring competitions and assign the teachers to schools in the judet. Individual schools do not have any input into the selection of teachers for their schools. These hiring procedures have led to frustration among some schools, as documented in interviews with school officials in Covasna and officials in the Judet Inspectorates. Starting in the 1998-1999 school year, personnel policy was directly delegated to some "representative" schools (see below). Infrastructure. Starting with 1998, local councils have financial responsibility for capital investments, while decision making regarding the number and size of schools remains with the central Ministry of Education. A recent ordinance transferred the property rights of schools to local councils. This measure aligned ownership of schools with the responsibility for 31 The case of "representative schools," which have autonomous management of human resource policies is discussed further below. 32 Until recently, the Ministry of Education was also in charge of all special education schools. These responsibilities are in the process of being transferred to the newly established National Agency for Child Rights Protection. 33 JEIs channel resources for salaries and educational materials through CEBs. 37 repair and maintenance spending - which had been transferred to local councils earlier in 1995 - and corrected incentives for local councils to maintain and invest in school property. Education Policy. The last decade has witnessed wide-ranging reforms aimed at bringing the Romanian education system in line with the needs of a market economy. The education reform has entailed the design of a national curriculum framework including school-based curricula, as well as decentralization of textbook choice. Presently, school curricula consist of common core classes and additional modules based on school-specific demand. Schools are responsible for securing the necessary resources for the optional courses they offer. Although the public provision of education services is at present still best characterized as a deconcentrated system, recent reforms have brought about an increasing degree of decentralization. For example, in the 1998 school year, the MOE began a large-scale initiative which further decentralized autonomy in important areas to a number of selected schools known as "representative schools" (Box 3.1). Box 3.1: Representative Schools In the 1998/1999 school year, the MOE undertook a pilot initiative to delegate decision making authority to the school level. Under the experiment, 3,404 primary and secondary schools (approximately 22 percent of total schools) were designated as "representative" and were granted autonomy over budget making, resource allocation and personnel decisions. In particular, representative schools may organize their own competitions to select teachers whenever a position opens, and may fire teachers. Moreover, they oversee their own annual budgets. This type of autonomy may be extended to all schools in the future. Representative schools were not selected randomly. Rather, high performing schools were chosen on the basis of a number of indicators, including student achievement and the share of qualified teachers, and involvement in the local community. Preliminary interviews with school principals suggest that schools welcome such autonomy, although they recognize the increased administrative costs associated with it. Although increasing school autonomy in personnel activities would further the decentralization process in the education sector, extending the experiment to all schools should be approached with caution. In particular, taking the feedback from the current representative schools as indicative for the whole country might be misleading, because of their higher capacity and better initial performance. Unless a systematic effort in training principals and personnel is undertaken, poorer schools with little accounting and managerial capacity could be overwhelmed by the administrative burden. Monitoring and evaluation. The Ministry of Education is responsible for student assessment standards through the National Assessment and Evaluation Service. A standardized exam at the end of grade VIII was introduced in 1998. Although presently reliable data is still lacking, a careful record of student's achievements in the standardized exams will provide policymakers with an important tool for formulating policy.34 Health Roles and responsibilities in health care are in flux following the introduction of the health insurance system. The reform has reduced the role of local governments in the provision 34 See Voicu and Bega (1999). 38 of health care, while at the same time shifting many aspects of decision-making to individual physicians. The Law on Health Insurance was passed in 1997 and the Health Insurance Fund began collecting contributions in 1998. The institutional structures of the insurance system - a National Health Insurance Fund (NHIF), and 41 deconcentrated Judet Health Insurance Funds (JHIFs) at the judet level - were established in January 1999. The public health system consists of the central Ministry of Health (MOH) and 42 judet level Primary Health Directorates (PHDs) including Bucharest. Prior to the introduction of health insurance, responsibility for service provision was based in the Ministry of Health and administered through its deconcentrated branch offices. Beginning in 1998, the Judet Health Insurance Funds in each judet assumed responsibility for providing primary health services at the local level by contracting for services with primary care physicians. The role of the Ministry of Health, and its judet level offices, the Primary Health Directorates has been narrowed to focus on public health activities and coordinating functions including monitoring public health and system quality, and collecting and processing data on health status. Under the previous arrangements, local councils had responsibility for financing some health expenditures, including maintenance and supplies for dispensaries. From 1998, this was changed and local councils only finance some materials and services for creches and blood donation activities. The role of local councils also includes ownership of dispensaries in some cases. According to the Law on Health Insurance, local councils can also provide incentives - including salary supplements and facilities - to attract physicians to their locality. The case studies also revealed exceptional examples where the local councils stepped in to provide support to health facilities faced with serious budget crises, such as in-kind and financial support for financing utilities in some rural communes. Personnel. Under the new system, physicians contract with the Judet Health Insurance Funds to provide health services. In turn they receive a budget for personnel (e.g. nurses), office administration, maintenance and consumables and medical equipment. The physician decides how many staff to hire based upon guidelines set by the Ministry of Health. Physician's salaries are set based upon the number and age of patients enrolled on their list, and the number of consultations, within parameters which are negotiated annually between the National Health Insurance House and the College of Physicians. Infrastructure. Issues related to the ownership and management of health facilities under the new system are currently being clarified. Dispensaries and other primary health units are owned by different institutions and, in some cases, property ownership has not been clear. Some are owned by the Ministry of Health, and others by judets and local councils. The Ministry has recently completed an inventory of the ownership of facilities. Under the new system, local councils which own health facilities may lease them to physicians who then enter into a contract with the Judet Health Insurance Fund to provide services in that locality. Public Health Activities. The Ministry of Health runs national public health programs which are implemented by the Primary Health Directorates including health promotion campaigns through public information, immunizations and HIV prevention activities. The range of activities which are undertaken in each judet differ significantly and not all judets run all activities. The introduction of health insurance has not affected the financing of public health programs. None of the staff in Primary Health Directorates interviewed in this study were aware 39 why the Ministry was supporting the programs it was supporting in their judets, and it was unclear whether the particular health needs of the judets were taken into consideration in programming public health. Family planning activities are included in the national programs, but the level of activity varies across judets. Some PHDs organize special radio courses (Hunedoara, Covasna, and Timis) and others disseminate information through local newspapers. In some cases (Hunedoara, Covasna) physicians working in family planning clinics conduct lectures in schools (especially at 6th and 7th grade). Interviews during the case studies indicated that family planning activities had limited coverage in rural areas. Monitoring and Evaluation. The fieldwork indicated confusion regarding responsibilities for monitoring and evaluation in health. While most directors of Primary Health Directorates perceived monitoring as their responsibility, as the Ministry of Health is formally responsible for accreditation of physicians, others believed it to be under the purview of the Judet Health Insurance Funds. One Primary Health Directorate director noted: "the JHIFs represent the interest of the insured, therefore they should care about the quality of the services being provided, as well as coverage problems. " Primary Health Directorates are responsible for collecting data on health status. All PHDs included in the study have a departnent for statistics which gathers standard data requested by the Ministry of Health. The majority of the departments do not undertake monitoring activities beyond those required by the Ministry. An exception was the Primary Health Directorate in Hunedoara, which undertook independent evaluations on health status topics (e.g. reproductive health and communicable diseases) to determine health needs in the judets. The Primary Health Directorate was able to secure funds from the state budget through a public health program to support this initiative. 3.3. Implications of Institutional Arrangements Roles and Responsibilities During the ongoing reform processes in the different sectors, roles and responsibilities among institutional actors have frequently been unclear due to changes that have introduced new institutions and altered the mandates of others. These developments have the potential to influence service delivery by diluting accountability and distorting incentives. In some cases confusion regarding roles and responsibilities can lead to gaps in the coverage of services. Moreover, even when well identified, in some areas decision-making and financing responsibilities do not appear to match, and, as a result, incentives are weakened and transparency jeopardized. The review of institutional arrangements across sectors, and interviews with local officials and social service providers identified a number of examples. In social assistance, responsibilities for delivering services are fragmented across institutions and levels of government, limiting coordination and a comprehensive approach to the sector. In education, the lack of clear guidelines for budget allocation and information sharing at the level of budget allocating institutions, the budgetary execution centers (CEBs), limits transparency and creates disincentives for schools to multi-year financial planning. In health, roles and responsibilities are uncertain following the recent reform process. 40 Social Assistance. Provision of social assistance benefits and services has been increasingly diversified and is fragmented across numerous government agencies at all levels of government. Despite this situation, significant improvements have been made in recent years to consolidate functions and pending legislation promises to further reform the system. For example, institutional arrangements for the provision of child protection services were streamlined with the creation of the Department of Child Protection (now the National Authority for Child Protection and Adoption) in 1998. Until mid-1997 responsibility for child protection fell under the jurisdiction of numerous agencies including the Ministries of Health, Education, Labor and Social Protection and the State Secretary for the Handicapped.35 Fragmentation of roles within social assistance persists within the current system, which involves many different agencies. Responsibility for managing social care institutions falls under the Department for Child Protection, the State Secretariat for the Handicapped, as well as local councils. This fragmentation limits coordination of policies and likely increases administrative costs. The division of responsibilities also leads to gaps in coverage, as no single agency is responsible for policy on social assistance as a whole. For example, none of the institutions is responsible for children leaving institutions at age 18, as they no longer fall under the jurisdiction of the child protection, or education system. A draft Law on Social Assistance, approved by the Cabinet in November 2000, is expected to consolidate social assistance functions and existing structures under the Ministry of Labor and Social Solidarity. The MMSS would assume responsibility for coordinating policies for all social assistance benefits and services, possibly including institutions and programs for children, the elderly and disabled. To date it is unclear whether the functions of the State Secretariat for Handicapped will be under the MMSS or the Ministry of Health. The new Law also envisages substantial structural changes in institutional arrangements for social assistance provision in local govermments. At the judet level, judets would establish decentralized social assistance departments within their administrations, to oversee the implementation of social assistance programs and to monitor the needs of residents in the judets. Local councils would also establish social assistance divisions within their administrative structure to identify needs, develop and administer services and to coordinate programs with NGOs. Actual provisions for the implementation of the new Law have not yet been defined and significant pending issues remain. Specifically, the future role of the Ministry of Labor is not clear, and its organization at the local level has not been identified. It is unknown whether the Ministry of Labor will have offices in each judet, what their relationship with local governments will be, and how they will operate. Related to these issues, the process for consolidating institutions into the Ministry of Labor will require careful planning, coordination and political will. Education. In education, unclear rules and overlapping institutions blur responsibilities and increase uncertainty in the sector. As mentioned above, CEBs act as budget intermediaries between Judet Education Inspectorates, local councils and schools, and also provide schools with administrative and accounting services. Interviews conducted for this study suggest that a set of agreed upon guidelines on the mechanism of budget allocation and on the sharing of information 35 Distribution of responsibility for child welfare institutions was as follows: institutions for children 0-3 years (MOH); institutions for children 3-6 and 6-18 (MOE); institutions for disabled children (SSHP); temporary shelters (local councils). 41 with schools is lacking. As a result, CEBs often operate in a discretionary and non-transparent manner that weakens schools' incentives to develop much needed financial planning. Moreover, the presence of this additional administrative layer between the sources of funds and the service provider reduces accountability substantially. As these issues directly concern the effectiveness of financial arrangements, the problems linked to the role of CEBs will be discussed further in the next chapter. The ongoing changes and shifts of responsibilities within the education sector have brought about other distortions in incentives and gaps in coverage. For example, the division of responsibilities between the Ministry of Education, that was in charge of the education of students with special needs, and the local councils, that have been recently assigned the responsibility for the accompanying personnel might give rise to an excessive fragmentation in the care of disabled students.36 In other cases decision-making is not aligned with financial responsibilities. This is the case of school infrastructure, where the decision to build a new school lies with the Ministry of Education while local councils have responsibility for financing. Although this might be motivated by the need to frame the construction of new schools in a nation-wide strategy, local councils may feel both excessively burdened by financial obligations, while the lack of decision- making authority might weaken their ownership. Finally, in some cases the substantial degree of centralization does not foster accountability at the level of service provision. This has been traditionally the case with personnel policy. Here centralized decision making regarding personnel in education has limited the scope for local quality control, as there have been few built-in incentives for teachers to perform better, or to enhance their qualifications. Nonetheless, steps have recently been taken to address these issues with the restructuring of Judet Education Inspectorates and of the inspection system. Health. In the health sector, because of the newness of the insurance system and the recent transfer of responsibilities for financing services from Primary Health Directorates to Judet Health Insurance Funds there is understandable confusion over roles and responsibilities. The dynamics between the public health and insurance systems are beginning to take shape. In a number of the judets in the study tensions were notable - particularly because the Judet Insurance Funds are responsible for a larger share of the budget and because salaries disparities across the Judet Insurance Funds. Lack of clarity regarding the roles of the Primary Health Directorates and Judet Insurance Funds - particularly over shared activities - leads to blurred accountability and the risk of gaps in coverage. Interviews with officials in both types of institutions indicated a high degree of uncertainty regarding responsibilities. Tensions also emerged over perceived overlapping responsibilities, or activities which each institution thought to be under the jurisdiction of the other. An example is responsibility for immunization campaigns. While Primary Health Directorates are responsible for providing vaccinations, Judet Insurance Funds have authority over physicians as the contracting body. As a result neither institution is explicitly responsible for ensuring coverage. According to a Primary Health Directorate director: "We only provide the 36 These arrangements might be subject to change as responsibility for special schools was transferred to the National Agency for Child's Rights Protection in 2000. 42 injections and ensure technical assistance.' All other problems related to these campaigns are the responsibility of the JHIFs, because they signed contracts with the family physicians. " Ambiguity over the ownership of health facilities affects incentives for maintenance and investment. Property rights and the accompanying responsibilities for managing health facilities under the new system are currently being clarified. While some dispensaries and primary health units are owned by local governments, others are under the jurisdiction of the Ministry of Health. The Ministry has recently completed an inventory of facilities to determine ownership. The decision over ownership can affect local councils' budgets in an important ways, as under the new system, local councils which own health facilities may lease them to physicians. Without clarity of ownership, neither actor has any incentive to maintain the facilities and make necessary capital investments and as a result the status of physical infrastructure in the sector is threatened. Cross-Cutting Policies. Equally important, close collaboration across institutional actors in the different sectors is needed to ensure coordination on cross-cutting issues and to prevent gaps in service provision. An example is health insurance coverage for poor households, where the legislation requires coordination between local councils, the Ministry of Health and the Ministry of Labor. Under the new insurance legislation, contributions for beneficiaries of social assistance are to be covered by the Ministry of Labor. In order for this to occur, local councils must provide a list of beneficiaries eligible for social assistance to the branch office of the Ministry of Labor at the judets level. Because of the novelty of the health insurance system, a number of judets and local councils were unaware of these arrangements and did not provide the information to the Ministry of Labor. In the cases where the local council was no longer funding social assistance, local councils frequently did not screen benefits for eligibility, leaving poor households without health coverage. Local Government Capacity for Service Delivery While local governments are assuming growing responsibility in the social sectors, to date there have been few changes in the organization of local administrations to deliver these services. As discussed in the previous chapter, judet and local council administrations are in the process of forming departments for revenue collection, and some limited training has been provided in some local governments for the implementation of the Law on Local Public Finance. However none of this relates directly to the social sectors. Central government institutions have an important role to play in ensuring local government capacity to implement social services, both by monitoring delivery of services across localities and building capacity through training at the local level. The following discusses local capacity for service delivery, focussing on social assistance where local councils are most involved. Social Assistance. Although judet councils do not have direct responsibility for social assistance beyond their involvement in child protection, some judets (Timis in the study) have established departments for social assistance. The current draft Law on Social Assistance calls for the establishment of such departments in all judets which would be involved in a range of tasks, including monitoring of social assistance needs, gathering information on the implementation of programs, and managing social assistance institutions. The Law states that these departments will be decentralized offices, suggesting that staff in the offices will be employees of the local administration. Their relationship with the Ministry of Labor will require close coordination, particularly on training, as they will be responsible for implementing programs and policies set by the central ministry. 43 Local councils will continue to have substantial responsibility for social assistance under the new legislation. Despite their significant role, local councils have limited human resource capacity in social assistance. Social assistance benefits are delivered by employees of local councils, called 'referents,' who do not necessarily have expertise or experience in social work and social issues. This reflects the fact that social assistance in Romania has been traditionally considered an administrative function. The current system also does not differentiate between delivery of cash benefits and social services. The same personnel who provide birth grants and means tested benefits are responsible for referrals to social service institutions. As social assistance in Romania becomes increasingly diversified, there is a greater need for qualified and trained staff at the local level who are equipped to address the needs of a wide range of beneficiaries. The organization of social assistance services within local administrations varies notably across local councils, depending on the range of services provided and the size of the locality. In the smallest communes, Vorta and Zapodeni, there were no designated employees for social assistance. Instead the mayor, city hall secretary and accountant implement the programs, including registering and screening beneficiaries for means tested benefits. In the study sites the number of persons involved in social assistance varied from one, in Tirgu Secuiesc, to five in Vaslui. None of the social referents in the case studies had specific training in social work. Capacity for screening beneficiaries for in-kind services is similarly varied. In localities which have a social assistance institution or have many children in foster care, local councils frequently have a staff person assigned to oversee these activities. This was the case in three of the municipalities in the study, Vaslui, Lupeni and Lugoj. In the other cases, Babadag and Tirgu Secuiesc the social assistance referents handle referrals to institutions. Training is rare and depends on the initiative of local officials. In 1998 in Hunedoara, the Child Protection Directorate organized two courses for the social assistance admninistrators in local councils who are involved in child protection activities. Education and Health. Because the role of local governments in education and health is limited to financing services, local administrations do not have specialized expertise in either area. In both sectors decentralization is increasing the responsibility of individual service providers. In health, primary care physicians are responsible for managing the budgets as contracted with the Judet Health Insurance Funds, and take decisions regarding the allocation of resources for personnel in the clinics (e.g. nurses), pharmaceuticals and other materials. In education, individual schools are gaining greater responsibility for resource allocation under the program of 'representative schools' discussed above. Both of these developments imply a need for capacity in accounting and budgeting within the local administrations, as well as social service institutions. The general impression from the fieldwork was that capacity in the education sector, particularly as far as accounting and managerial skills are concerned, is limited, both in most local councils - where no specialized personnel are assigned to manage education spending, and in schools - where principals and administrative personnel have little or no training in management. This weak capacity will be particularly problematic as decentralization moves forward and increasing administrative and managerial responsibilities are shifted to local councils and schools. 44 Local Decision Making on Social Services Local govermment decision making regarding social services is largely focused on budgeting, the allocation of resources to the different sectors, and the provision of social assistance benefits and services. In the local governments included in this study, mayors played a pivotal role in decision making. In other areas the mayor delegates to the staff working on specific sectors (e.g. social assistance), however in all of the sites included in this study, the mayor retained authority over decisions regarding the amount of resources available for various programs and the prioritization of programs. In some cases the local councils do play a role in taking strategic decisions, sometimes contrary to the views of the mayor. Because of the centralized nature of decision-making, personality and the individual initiative of mayors and other key local government officials has a strong influence on the types of decisions which are taken. The lack of clear and transparent guidelines for operations and for sharing information is likely to make personal initiative and connections even more crucial. Social service providers rarely have a role in the budget process. However, the case studies documented some exceptions. In Tg. Secuiesc in 1999 the local council included principals from all schools in the area in the discussion of the education budget - not just the CEBs which manage budgets - in the budget discussions. In Lugoj the local council's technical expert collaborated with the CEBs to make sure that budget allocation covered schools needs uniformly. Nonetheless, in the education sector, most decision making lies with the central government. The constraints on local governments' involvement with education (mainly the financing of infrastructure and maintenance) seem to derive more from the continued lack of funds than from pitfalls in the decision-making process. The approach of local governments toward social services has been shaped by the nature of the decentralization process itself. For the local officials interviewed for this study, decentralization in social assistance and education has implied an increased fiscal burden. The transfer of additional financing responsibilities, often following the start of the fiscal year - and in an environment of tight fiscal constraints - has led local government officials to perceive decentralization as an increase in imposed obligations, rather than as an opportunity to develop locally responsive policies. As a result, local government decision making in social assistance and education has been largely focused on developing coping strategies to address the tight fiscal enviromnent. The nature of these strategies is discussed further in the next chapter. Intergovernmental Relations. The relationship between central government institutions, particularly the deconcentrated institutions - in social assistance the Child Protection Directorates and Judet State Inspectorates for Persons with Handicap (JSIPHs) and in education the Judets Education Inspectorates - and local governments is an important one which affects the level and quality of service delivery in each locality. Throughout the case studies, the nature of intergovernmental relationships varied significantly and depended greatly on the personalities of the local officials involved. The transfer of financing responsibility to local governments led to tense relationships with central government institutions, particularly where local govermments lacked involvement in decision-making. For example, the shift of responsibility for financing institutions for the elderly and disabled to local councils led to difficult relations between local councils and the JSIPHs in Lugoj and Hunedoara, when local councils did not allocate sufficient resources to the institutions. In Lugoj, in 1999, the local council did not budget for the two social service 45 institutions in the locality, one for disabled children, with 120 residents, and a day care center for disabled children and adults, with 44 clients. At the time of the interview, the general inspector of the JSIPH was considering legal action against the local council to address the impasse. Similarly, in Tg. Secuiesc, the construction works for a school building came to a halt when, in the midst of the budget year, the central government transferred the financial responsibility for the new investment to the local council without a clear definition of property rights. Moreover, mayors and local officials often lamented that financial and administrative responsibilities were transferred from the central government to local government without any explicit consultation and, in some cases, after the local budget had already been approved, for example in the case of student scholarships. This lack of collaboration across levels of government has the potential to create discontent and mistrust. Incentives for Local Officials. Local government officials have limited formal authority and capacity to address gaps in service provision. As a result their incentives to innovate, initiate new social service programs, or to tailor existing programs and services to local needs are weak. The case studies found that local initiative in service provision was the exception, rather than the rule. For example, the open approach of the budgetary execution center (CEB) in Tirgu Secuiesc, where transparent information about allocation rules encouraged trust and promoted responsible budgeting at the school level. In the same town, substantial collaboration and openness was also promoted by meetings with school principals organized by the local councils where yearly school spending was discussed. Greater proactivity and innovation was also found in the area of child protection, as the decentralized Child Protection Directorates at the judet level had flexibility to form partnerships with NGOs, donors and other civil society organizations. Some judets (Hunedoara and Timis) have undertaken organizational reform of the system and others have adopted strategies for children leaving institutions, who do not fall under the jurisdiction of any of the social service institutions, including child protection or education. Despite the exceptions, the interviews confirmed that local initiative is mainly dependent upon the personality and proactivity of individual officials. Moreover, where local governments did get involved in adapting services to local conditions, this was most frequently to address financing constraints. For example, to introduce new eligibility criteria to reduce social assistance payments, or to stop paying student scholarships or transportation subsidies. Once local governments have adapted more fully to the new decentralized environment and with greater fiscal capacity at the local level, establishing rules for information sharing and monetary and non-monetary rewards for constructive innovation at the local level could improve the quality of local public management. Built in financial incentives also impact the proactivity ofjudets. The budget of the CPD is still input based and is largely determined by the number of children in institutions. On the other hand, funds are available from extrabudgetary sources for deinstitutionalization activities, including the maternal assistant system and mother child centers. As a result, CPDs generally maintain existing institutions and implement deinistutionalization activities financed from extra- budgetary sources. Additional activities depend on the personal commitment of management. Monitoring and evaluation. At present, monitoring and evaluation practices and effectiveness vary widely across sectors, but in general are limited. Monitoring of outcomes of policies has been particularly limited. This has changed recently in education, where the 46 introduction of a standardized tests in the eighth grade will allow for country-wide monitoring of lower-secondary education outcomes. More importantly, it will provide a useful benchmark for impact analysis of the ongoing process of financial decentralization of education spending to local councils, as well as of the transfer of administrative and personnel responsibility to the representative schools. In social assistance, because of the multiplicity of programs and the current fractionalization of responsibilities among many ministries, a systematic evaluation of program impact, coverage, and targeting is crucial. Nonetheless, at present monitoring is limited and is not linked to feedback mechanisms that convey to policymakers at the local and central government the necessary information in a timely and effective way. The draft Law on Social Assistance addresses these issues and envisages a strengthening of monitoring arrangements by the Ministry of Labor through its judets level offices. The Ministry envisages that this will be supported with improved information technology, as well as enhanced capacity for analysis and policy making. In the health sector, more transparent guidelines should clarify the distribution of responsibilities for monitoring and evaluation between the public health system and the health insurance fund. The need for systematic monitoring and evaluation is increasingly pressing. As decentralization moves forward, the supply of social services will be increasingly diversified across localities and, quite likely, will depend more heavily on local financial resources and capacity. Moreover, as the practice of contracting out services to providers becomes more common, the quality of services will also vary considerably. These processes can substantially jeopardize equity across poor and wealthy localities. Furthermore, as we already stressed, the increasing fragmentation of responsibilities across different institutional actors is likely to produce coordination problems that will result in significant coverage gaps. In this context, local governments can take advantage of their proximity to the citizens and to service providers to obtain direct feedback on service quality and adequacy, to be used as an input in policymaking at the local level. The central government in turn can make use of this and other nationally collected information to monitor equity of service access and provision across the country as well as to evaluate the effectiveness and coverage of programs across the different social sectors. For monitoring and evaluation to be effective, an institutional mechanism that links the results and findings of the evaluation process directly to policymaking is essential. 3.4. Summary Issues In recent years, local governments' involvement in social services has increased substantially, although the role of local governments varies considerably across sectors and, within sectors, across different functions. Social assistance policies are the most decentralized. Local governments not only provide around fifty percent of the total funding but are also directly involved in policy making, delivery and administration of social assistance. Starting in 1995, and more substantially after 1998, local councils have played an increasingly important role in financing all capital and some current spending for the provision of education services. This responsibility is likely to expand, as the current government is planning to allocate responsibility for personnel expenditures in education to local governments. In the health sector, on the other hand, the reformns establishing the Health Insurance Fund have reduced the role of local councils to financing a limited range of activities (services and materials for creches and blood donations) and to the management of the buildings that are local council property. 47 The many recent institutional changes have configured a complex system of intergovernmental relations where financing, decision-making, and delivery are often fragmented across different actors and levels of government. In the absence of clearly defined roles and agreed upon criteria for coordination of these activities, such fragmentation has the potential to bring about gaps in service coverage (as in the example of the immunization campaigns) and make service provision excessively dependent on such factors as personal relations between key personnel in the involved institutions. Moreover, whenever responsibilities are not appropriately matched with decision-making power, incentives for quality service are weakened and accountability blurred. Finally, the frequent changes in the number and types of responsibilities that are assigned to local governments (for example financing of student scholarships was decentralized in 1998 and re-centralized in 2000) increases the uncertainty in the operations of local governments, decreases incentives for responsible multi-year planning and fosters "wait and see" attitudes. In many instances local governments appeared unprepared to assume the many financial and administrative responsibilities progressively transferred from the central government. This was particularly true as decentralization has progressed in a period of economic decline and tight fiscal constraints and was not accompanied by parallel training and information programs aimed at increasing capacity at the local level. Moreover, as highlighted by many interviews with local officials, the central government shifted new responsibilities and tasks to local councils without a comprehensive consultation process with local governments, thus contributing to a lack of ownership and diffused discontent at the local level. The increased involvement of local governnents also creates important opportunities for exploiting their natural information advantage and ability to tailor service delivery to local needs. For example, the wider scope in local government activities can be particularly beneficial to local communities as local governments can be more flexible than the central government in collaborating with NGOs and can more easily integrate innovation in their activities. Although the novelty of many institutional arrangements makes it difficult to disentangle transitional issues from structural pitfalls, the importance of clearly defined intergovernmental roles, transparent procedures, correctly aligned financial and decision-making responsibilities, and capacity at the local level cannot be overstressed in the present Romanian context. 48 Annex 3.1: Intergovernmental Responsibilities in the Social Sectors A. Social assistance: Cash Benefits Who Decides? Ministry of Labor Judet Directorate Judet Local (municipal, and Social Solidarity for Labor and Administration commune) Social Protection administration Service organization Initiating new programs X X Personnel X Number of positions X Minimum credentials X Hiring X Firing X Salary settings x Paying salaries x Training X Evaluating staff quality X Access Rules Defining eligibility criteria X x Applying eligibility criteria x Monitoring and Evaluation Monitoring number of X X beneficiaries Monitoring quality X x (responsiveness) B. Social services for the disabled Who Decides? State Secretariat Judet State Inspectorates Judet Local Special for Persons with for Persons with Admin. (municipal, centers Handicap handicap commune) Admin. Service organization Number of facilities X X" Capacity (# of beds) X Revenue raising X X '° x Required Programs/Norms X Personnel Number of positions X X__ Minimum credentials X Hiring X _ X Firing X Salary settings X Paying salaries x Promotion/evaluating quality X X Access Rules Eligibility for access to care X Monitoring and Evaluation Setting standards X Monitoring quality X (achievements) 37 JSIPHs propose and SSPH approves. 38 Judet council budgets are the main source of financing. 39 JSIPHs propose and SSPH approves. 49 C. Social assistance services for child protection Who Decides? Central Judet Council Local (municipal, Government Directorate for Child Protection commune) administration Service organization Initiating new programs X X Personnel X Number of positions x Minimum credentials X Hiring X Firing X Salary settings X Paying salaries X Training X X Evaluating staff quality ( on the job) X Access rules Defining eligibility criteria X X Applying eligibility criteria x x Monitoring and Evaluation Monitoring number of beneficiaries X x Monitoring quality (responsiveness) X x D. Compulsory edu cation preschool, primary and gymnasium) Who Decides? Law Ministry of Judet Education Judet Local (municipal, Schools Education Inspectorate Admin. commune administration) Service Organization Number of schools x Capacity (class size) X x Number of places in X42 education units per grades and education forms Revenue raising (of schools r x Number of positions X44 _ Minimum credentials X4 Hiring X46 _ Firing x X 4 7 Salary setting- X4 x__ _ _ _ _ _ __ _ _ _ _ Paying salaries x Teacher training X49_ x _ 40 Art. 10, Law 84/1995 "Local authorities and economic agents may establish and finance school units according to the Law provisions." 41 Minimum ceiling of children per class is established by Law 84/95, art 158 42 Annual Order of the MOE 43 "Identification and use of extrabudgetary resources by schools ( Order of MOE 3378/1998) and "completely autonomous use of extrabudgetary resources by schools (Order of MOE no. 3176/1998). 44 Annual Order of the Ministry (See Order of the MOE No. 3589/1998 on allocation of positions in schools, lyceums and universities within the Budget 1998. 45 Established by Law on the status of education personnel. 46 Art. 30 Law on status of education personnel. 47 Order of MOE No 3207/1998 establishing the organization and functioning of the Discipline Collegium attached to the Judet Education Inspectorate. 48 Law on the status of education personnel and the Education Law set conditions and ceilings for levels of salaries and incentives. 50 Promotion/evaluating teacher X x x quality Access Rules Who goes to which school x Quality monitoring _ Setting standards Determining curriculum X Monitoring quality X2 XX53 (achievements) E. Health care services Health care services Central Territorial Local Admin. Who Decides? Ministry of National Judet Public Judet Judet Local Health Health Health Health Admin. (municipal, (MOH) Insurance Directorate Insurance commune House (PHD) House admin.) Service Organization Number of facilities X Capacity (# of beds) XX55 Revenue raising X X x.6 Contracts with Private X Providers Number of positions X X57 Minimum credentials X X Hiring X X X X Firing X X Salary settings X X X X Paying salaries Promotion/evaluating quality Physicians Collegium Access Rules Eligibility for access to care X Monitoring and Evaluation Setting standards Physicians Collegium Monitoring quality x (achievements) I I I 49 Art.32 para.2 Law on the Status of Education Personnel states that the overall training of didactic personnel is coordinated by the MOE and the Education Inspectorate together with Houses of Education staff organize teacher training according to MOE methodology. 50 Procedures mentioned by the Law on the status of education personnel; criteria established by the MOE; self assessment judged by the school administrative board (the school headmaster is the president of the Administrative school board). l MOE sets quality standards. 52 Institutional evaluation is ensured by the Evaluation Department; Students performance evaluation is coordinated by the National Evaluation Service. 5 Schools may decide upon providing certain optional and alternative teaching classes. Curriculum is set by the MOE. 54 PHDs propose and MOH approves. PHDs propose and MOH approves. 56 Health units may raise their own revenues. 57 PHDs propose and MOH approves 51 Chapter 4: Financing the Social Sectors 4.1. Introduction The role of local governments in financing social services has changed, both with on- going sectoral reforms, and as a result of the changes in intergovernmental finance discussed previously. Relative to the other policy areas, the health sector has been unaffected by changes in local finance, as responsibility for nearly all health financing was shifted to the health insurance system in 1998. In contrast, local finance reforms have had a direct impact on financing of social assistance and education. Both are now dependent upon local budgets for a substantial share of their resources. In addition to increasing the role of local governments in revenue raising and budgeting, the Law on Local Public Finance transferred new expenditure responsibilities for capital expenditures in education and social assistance institutions to local budgets. As a result, local governments are now responsible for financing all social assistance cash benefits and services for poor households and the elderly, and an increasing share of services for children at risk and the elderly. In education, local governments are now responsible for all capital and for some current education expenditure, with the exception of personnel costs, (Table 4.1). Table 4.1: Expenditures by Source (% of total expenditures in the sector) 1996 1997 1998 1999 Health 100 100 100 100 central 81 81 99 100 local 19 19 0.6 0.5 Education 100 100 100 100 central 84 87 90 85 local 16 13 10 15 Social Assistance 100 100 100 100 central 62 69 49 53 local 38 31 51 47 Source: LSSD Database, National Commission for Statistics Notes: 1999 data are budget. In both sectors, local governments have become responsible for an increasing share of expenditures, but have faced growing fiscal constraints due to the overall decline in local revenues. Expenditures on both education and social assistance expenditures remain below 1996 levels in real terms while health expenditures increased (Figure 4.1). Local governments have dealt with the tight budget constraints and new expenditure responsibilities in different ways. The case studies documented responses and coping strategies across local governments. Some local councils stopped providing services altogether. The services which were most commonly threatened were poverty alleviation programs, leading to a situation in which some of the poorest local governments with the greatest needs have been unable to fund redistributive programs, such as means-tested social assistance benefits. Other benefits which were ceased include scholarships for students and transportation for school children. In other cases, local govermments restricted access to benefits by introducing new eligibility criteria. 52 Figure 4.1: Social Sector Expenditures (1996=100) 120.00 110.00 - -Health 100.00 I00 00 00 i--Education 90.00 80.00 -Social Assistance 70.00- - Total Local 60.00- Expenditures 50.00 1996 1997 1998 1999 Note: 1999 data for social assistance are budget. Sources: IMF data and LSSD database. Given the limited public resources, local governments and service providers, including schools, have increasingly turned to households and civil society organizations for funds. Donor organizations and NGOs have become involved in financing child protection activities and services for the poor and disabled. In education, schools have relied increasingly on parents for support. In many of the local governments in our study, parents have become involved in financing school expenditures, or providing in-kind labor to support schools. These developments have substantial implications for equity. Budget constraints have also led to shifts in types of expenditures. In education, fiscal constraints at the local level have led to a collapse in capital expenditures and spending on maintenance and materials. In contrast, personnel expenditures have been protected. In social assistance, local governments have prioritized spending on institutions and services, over cash benefits. Changes in financing arrangements have altered the roles and responsibilities of the institutions involved in service delivery - including the deconcentrated representatives of the central government ministries, local governments and service providers - and have modified the incentives they face. These dynamics are complex and difficult to disentangle, because of the wide range of institutions involved in both financing and provision. This chapter discusses financing responsibilities for each of the three sectors, focusing in depth on social assistance and education where fiscal decentralization to local governments has taken place. It discusses the impact of these arrangements on the adequacy of resources channeled to the sectors and the equity of resource allocation across judets and local governments. In particular, it focuses on (i) arrangements for financing each of the sectors and the major changes which have taken place since 1996; (ii) trends in financing within each of the sectors; and (iii) the impact of increasing local government involvement in the sectors, including some of the coping strategies they have employed. 53 4.2. Financing the Sectors Social Assistance Social assistance benefits and services have been more dependent upon local resources than other sectors and have therefore been more vulnerable to local budget fluctuations. The involvement of local councils in financing social assistance services has changed significantly during the last four years (Figure 4.2). A notable change took place in 1998, when local governments assumed responsibility for financing child protection. The following section details the financing responsibilities and sources of funds for child protection; services for the disabled; and social assistance benefits and services for low income households. It then reviews trends in spending levels and the allocation of resources within social assistance, and determinants of spending and implications for equity across local governments. Figure 4.2: Composition of Local Social Assistance Spending (1996-99) 100% 90%- _ Other actions 80%- E Public services for child protection 70% - Birth allowances 60% 0 Cash benefits U~50% 12 Social canteens 40%- E Institutions for the handicapped 30%- U Institutions for the elderly 20%- 10% 0% 1996 1997 1998 1999 budget Financing Arrangements Child Protection. Child protection services are funded from both the state and local budgets. Until 1997, over 85 percent of resources were provided from state budget sources through the numerous ministries engaged in child protection. This changed with the creation of the Department of Child Protection in 1998, and in 1999 judet councils provided 85 percent of expenditures on child protection (Table 4.2). In addition to funds from the Child Protection Directorates at the judet level, until 1999 the Ministry of Education has funded special education for the disabled through the Judets Education Inspectorates. The judets now finance child protection from their own budgets and are also responsible for managing funds from NGOs and donors which have become active in financing projects for children at risk. Table 4.2: Distribution of Responsibilities for Financing Child Protection (% of total budget) 1996 1997 1998 1999 State 86 87 16 15 Judets 14 13 84 85 54 The Child Protection Directorates (CPDs) prepare budget projections which they present for approval to their board of directors and ultimately to the judet councils. In most cases CPDs involve the institutions under their jurisdiction in the budget preparation process. Child protection institutions receive their funds directly from the CPDs. Benefits and Services for the Disabled. When compared to child protection, financing of services for the disabled have been quite centralized. A major change took place in 1999 when full responsibility for financing institutions for the disabled was transferred to local governments. Funds from benefits and services came from three main sources in 1999: (i) the Social Solidarity Fund, an extrabudgetary fund which finances institutions, services for children, social pensions and transportation subsidies; (ii) the Ministry of Labor which supports home care services; and (iii) local budgets which fund institutions, home care services, and transportation and telecommunication subsidies. Despite the transfer of financial responsibility for the institutions to local budgets, the share of expenditures coming from local budgets declined steadily, reaching 9 percent of total expenditures on the disabled in 1999. Box 4.1: Collection of the Social Solidarity Tax Deconcentrated offices of the State Secretariat for the Handicapped, the Judet State Inspectorates for the Handicapped (JSIPHs), are responsible for collecting the main source of funding for services for the disabled, the Social Solidarity tax, a 3 percent payroll tax collected from all employers in the judets. The tax rate is set centrally. This represents a significant task for the JSIPHs, which generally do not have the capacity for tax collection. The challenge is particularly severe in depressed areas where enterprises have significant debts, and which are undergoing substantial restructuring. The process for collecting the tax is complex and time consuming and involves substantial administrative costs. The JSIPHs send the collected revenues to the SSHP which returns the funds back to the JSIPHs. It is not clear whether any redistribution of resources among judets takes place at the level of the SSHP. The process of transferring revenues leads to delays in financing of 1-2 weeks, and in some cases longer. In some judets this led to delays in wage payments for social workers of between 2-4 months. Cash Benefits and Services for the Poor. Social assistance benefits and other services for the poor and elderly, including soup kitchens and home care are financed entirely out of the budgets of local councils. Decisions regarding the allocation of resources to social assistance are undertaken differently across local councils. In some cases, the mayor undertakes most decisions, in other cases administrative staff in the local government and elected local councilors negotiate different arrangements. All of the local officials interviewed agreed that within social assistance expenditures, payment of birth allowances was the first priority, followed by scholarships for students and emergency assistance. Means tested social assistance was identified by all as the last priority. Trends in Social Assistance Spending National Level Developments. Total social assistance expenditures represent a relatively small share of public expenditures, amounting to 0.8 percent of GDP and less than two percent of government expenditures in 1999. However, they consumed over seven percent of total local government expenditures in the same year. Real trends in local social assistance expenditures since 1996 illustrate dynamics in the sector (Table 4.4). Cash benefits, including the Means Tested Social Assistance Benefit (MTSAB) program, fell to 16 percent of 1996 levels by 1999. 55 In contrast, expenditures on institutions for the elderly and disabled have decreased less. The notable increase in resources for child protection in 1998 in part reflects the influx of donor and NGO support to that sector. Table 4.3: Total Social Assistance Spending (1996=100) 1996 1997 1998 1999 budget Cash Benefits 100 36 24 16 Child Protection 100 88 121 80 Benefits and Services for the Disabled 100 83 80 69 Benefits and Services for the Elderly 100 83 90 78 Other 100 103 158 165 Total 100 75 84 66 Trends in Local Spending. Empirical analysis provides insight into the determinants of spending on social assistance across judets (the results are reported in table 4.A.2 at the end of this chapter).58 The data reveal a strong trend over time in social assistance expenditures as a share of total local spending, with spending substantially lower in 1997. Poorer judets appear to devote a higher share of their total expenditures to social assistance benefits and services. In particular, the association between the share of local social assistance spending in the judet and the poverty rate for the previous year holds strongly for 1996 and 1998.59 Moreover, relatively poorer urban judets seem to spend a larger share of their budgets on social assistance - quite likely because most institutions for the poor are concentrated in urban areas. For example, urbanized judets have, on average, almost twice as many social assistance institutions for the elderly in comparison with rural areas. Although poorer judets spend a higher share of their local budget on social assistance, their per capita spending is not necessarily higher. A pooled regression of local per capita spending on social assistance on poverty rates in the previous year returns no correlation between the two variables.60 Repeated cross sections highlight that this result is due to the combination over time of different effects - per capita spending is positively associated with poverty in 1998 and negatively associated with poverty in 1999,61 indicating the critical need to improve steady targeting of social assistance spending to poor areas. Further analysis illustrates that the fiscal capacity of the judet affects social assistance spending. The pooled data reveal that, for a given level of poverty, judets with higher revenues per capita (analyzing both data for the same year and for the previous year) spend more in per capita terms than judets with lower fiscal capacity but devote a lower share of their budget to social assistance spending than poorer judets, while they.62 58 Econometric analysis of spending in the different sectors was conducted on judet level data. Data on spending at the level of local councils were not available. 59 Throughout this paper, poverty is defined as the percentage of individuals below as poverty line set at 60% of adult equivalent consumption. 60 Table 4.A.2, columns 1-5 report results for Ordinary Least Squares regressions (OLS) on pooled data only for 40 Judets and 4 years (1996-1999). 61 No notable correlation emerges for 1996 and 1997. Results are not reported here but available on request. 62 In these regressions we also controlled for the poverty rate to avoid that fiscal capacity acting simply as a proxy for poverty in the regression. The results are robust to controlling for income inequality (as proxied by the Gini coefficient)as well as for the percentage of MTSAB beneficiaries. 56 Composition of Local Spending. Fiscal constraints and the transfer of responsibility for financing child protection and services for the disabled has led to a general shift in expenditures from cash benefits to social services. Most strikingly, local expenditures on means tested benefits fell to 13 percent of 1996 levels in 1999. In contrast expenditures on in-kind benefits and institutions have been relatively protected. The dramatic real increase in child protection expenditures illustrated in Table 4.5 reflects the transfer of expenditure authority to local government budgets. Table 4.4: Social Assistance Expenditures of Judets and Local Councils (1996=100) 1996 1997 1998 1999 (budget) Total social assistance 100 61.0 110.7 81.1 Institutions for the elderly 100 83.3 90.4 77.8 Institutions for the disabled 100 84.9 81.9 58.0 Soup kitchens 100 93.8 128.6 106.9 Cash benefits (MTSAB) 100 33.2 20.7 13.0 Birth allowances 100 72.9 72.1 66.8 Child protection 100 77.9 2420.5 1611.6 Other benefits and services 100 182.2 555.4 639.7 Notes: Data differ from Table 4.4, due to diferences in the aggregation of programs. In addition to the changes in expenditure assignment, the shift from benefits to services may be a result of two factors. First, local governments do not have the authority to close institutions, and would be reluctant to do so, as closing institutions would lead to job losses for local employees. This would increase local unemployment and could ultimately increase demand for local social benefits. Local councils may not be able to provide alternatives for residents of social care institutions. Second, interviews with local government officials revealed a substantial stigma attached to social assistance cash benefits. Many local officials characterized beneficiaries of means-tested benefits as lazy and undeserving of support and were concerned about creating dependency on passive assistance. In response, local officials preferred to provide beneficiaries with one-time 'emergency' support, rather that regular benefits and, in some cases, conditioned receipt of benefits on participation on public works. Coping Strategies In some cases local councils have not allocated sufficient resources to maintain social assistance institutions. This has particularly been the case for institutions for the elderly which are completely financed from the local budgets. To address this, some Judet Inspectorates for the Handicapped (JSIPHs) have attempted to sign contracts with the local councils to ensure financial resources (Hunedoara). In other cases (Lugoj, Timis judet and Geoagiu, Hunedoara judet) the local council has not allocated funds for the institution. In localities where funds are paid, budget cuts are most frequently made for current expenditures such as maintenance, medicines and salaries. Although institutions are a local responsibility, not all residents of the institution may be from the locality where the institution is situated and no system of cross- subsidization is in place. The mayor of Babadag attempted to solicit funds from other local authorities where the beneficiaries of social assistance institutions in Babadag were from, but received no response. Cash benefits have been most adversely affected by budget constraints. Local councils have stopped funding the means-tested social assistance benefit (MTSAB) program in many 57 cases, or have adopted measures to limit the number of beneficiaries. According to household survey data only 8 percent of poor households received MTSAB benefits in 1997.63 On the contrary, the case studies did not point to any delays in the payment of birth allowances. This benefit can be targeted with ease and no stigma is attached to the beneficiaries. The MTSAB program, which was introduced in 1995, has proven ill-suited to meeting its poverty alleviation goals (Tesliuc, et. al, 2000). Coverage of the program is exceptionally low, because of budget constraints at the local level and weak incentives for local councils to pay benefits. Decisions regarding whether benefits are paid or not, their level and eligibility criteria are made entirely at the local level. This has fueled the decline in coverage, and in many municipalities the MTSAB has gone unpaid. Expenditures on the program are extremely low, amounting to 0.05% of GDP in 1997. Recently approved legislation (December 2000), established that not less than 0.4 percent of GDP, an eight-fold increase, should be allocated to the program. Notwithstanding a substantial increase in poverty in the country as a whole, the number of MTSAB beneficiaries dropped by 31% between 1997 and 1998 (Table 4.6). A pooled regression for 1998 and 1997 confirms the substantial decrease in MTSAB beneficiaries (as percent of total population) between the two years. Overall, poorer judets appear to have a higher share of MTSAB beneficiaries Nonetheless, when the relationship between the share of MTSAB beneficiaries and poverty is investigated in separate cross sections for 1997 and 1998, the association between the percentage of MTSAB beneficiaries in the judet and poverty turns out to be significant only for 1997 while no association emerges for 1998. This indicates a substantial weakening of targeting of resources in 1998. Moreover, the variation across judets in the share of the population receiving the MTSAB cannot be linked on average to variation in rural and urban poverty rates, nor to variation fiscal capacity, as measured by contemporaneous or lagged per capita local revenues, further highlighting the scope for improvements in targeting. Regression results are reported in Table 4.A.2, columns 6-8. Table 4.5 Trends in the Number of MTSAB Beneficiaries 1997 1998 % change Average number of MTSAB beneficiaries per Judet 4225 2956 -31% Poverty (lagged poverty rate) 21 32 +52% Three of the five communes in this study had stopped paying the MTSAB completely in 1999, while the other two paid the benefit irregularly. Babadag stopped paying altogether in 1997. In the cases where benefits were paid, delays were reported to be frequent because of the lack of funds and absence of personnel to deliver benefits. Some local councils have attempted to compensate for the lack of benefits by providing people with meals from soup kitchens. In Tg. Secuiesc the local council paid the benefit only during the winter. Bucharest was the exception in the case studies, as means tested social assistance was paid regularly in Bucharest Sector 5. 63 See Dhanji, Pop, Tesliuc and Tesliuc, 2000,"Romania: Social Protection and the Poor," World Bank, draft. Calculations refer to a relative poverty line of 60 percent of adult equivalent consumption. 58 Local councils have adopted various strategies to limit the number of beneficiaries of the MTSAB. Local councils in Babadag and Vaslui opted to grant to the benefit only to the ill or elderly, while others added additional eligibility criteria to exclude certain groups. In Lugoj the local council stopped paying benefits to individuals who had never worked, entrepreneurs (regardless of their income), and families with children who are not attending school. In Tg. Secuiesc, the local council attempts to screen applicants for informal economic activity and requires that potential beneficiaries present their passports, to determine whether they are involved in cross-border trade. Some local councils conditioned benefit receipt on participation in public works. In Vaslui, Lugoj, Nadrag and Tg. Secuiesc, beneficiaries are required to work from four hours per day to five days per month. An administrator in Vaslui described an extreme strategy for transferring responsibility for paying the MTSAB benefit elsewhere, explaining that "[w/e try to encourage those who askfor assistance to move to the villages." 4.3. Education Financing Arrangements Responsibilities for financing education are divided between central and local governments. The education sector has witnessed a progressive shift of financial responsibilities towards local councils. While prior to 1998 most education spending was the responsibility of the central government, by 1998 nearly all education expenditures, with the notable exception of personnel costs, had been shifted to local councils. Prior to 1998, the level and allocation of resources were determined based upon discussions between the Ministry of Education and the Ministry of Finance, which would negotiate the budget allocation for different items on the basis of central norms including the number of students, average salary of teaching staff, and equipment needs. Once the budget was approved by the Parliament, the expenditures were allocated between current (salaries, textbooks) and capital expenditures. Resources were then transferred to the Judet Education Inspectorates (JEIs), which, in turn, would transfer the funds to Centru de Executie Bugetara (CEBs) - schools which serve as intermediate budget centers - and finally to the schools. 4 Local councils assumed responsibility for school maintenance and repairs in 1995. The most substantial shift in financial responsibility from the center to local councils occurred in 1998 when spending for investment in infrastructure, student scholarships,65 as well as equipment was transferred to local councils. After this rearrangement, salaries remained the main responsibility of the Ministry of Education. This reallocation of financial responsibilities redefined the role of the Judets Education Inspectorates (JEIs). Until 1998, local councils transferred funds for maintenance, repairs and educational materials to JEIs which then redistributed them to CEBs and subsequently to schools. Now local councils bypass JEIs. Towns transfer funds to CEBs while communa transfer funds directly to schools. In 1999, JEIs were still the sole administrator of expenditures for special education. Figure 4.B at the end of this chapter depicts the flow of funds among the different institutional actors. CEBs channel funds from the JEIs and, after 1998, from local councils to local schools. CEBs are usually large schools in charge of allocating funds and managing budgets for 10-25 64 Although by law JEIs were required to channel funds for equipment to CEBs, JEIs sent funds directly to schools and took care of bids, and collected information about suppliers, etc. 65 Financing of student scholarships was re-centralized in 2000. 59 schools in the area. CEBs are intended to achieve economies of scale within the education system by pooling administrative costs and sharing accounting and planning services. Nonetheless, the intermediation of CEBs is not without problems. Although the provision of accounting services is particularly important in a situation where managerial capacity at the school level is limited and costly, CEBs appear to operate without clear guidelines on the transfer of information to individual schools. For example, many schools complain that they do not know how much their budget is and what criteria are used by the CEB for allocating resources across schools. Such lack of clarity and transparency creates discontent in schools and makes the relationship between local councils and schools difficult. On the contrary, in the cases where the CEBs did develop clear formulas for the allocation of spending across schools and shared them with the management of individual schools, the relationship between schools, CEBs and local councils were reported to be smooth and collaborative. In Tg. Secuiesc the process for the formation for the school budget seems to be particularly transparent. First the local council requests the schools to estimate their spending needs. Later on, the CEB convenes a meeting with all school principals to debate budget allocations with particular attention to capital investments. All CEBs allocate funds on a capitation basis and according to school surface (for heating and current repairs). The lack of discretion and the transparency of the process have enhanced ownership at the school level. Trends in Education Financing National Level Developments. The Education Law establishes that at least 4% of GDP should be allocated to education spending. Nonetheless, resource allocation to education fell short of the government's objective and was only 3.3% of GDP on average between 1996 and 1998. In comparison with other transition economies, Romania's public allocation for education is lower than average (in 1998 transition economies allocated on average 4.4% of GDP to education spending) and falls short of the 4% average that is typical of middle income countries.66 Although high levels of public spending for education might reflect inefficiencies in the provision of the service, both the data and interviews suggest that adequacy of education financing is the main concern. Total education spending decreased steadily in real terms, dropping in 1999 to 78 percent of 1996 levels. Such a marked decrease reflects a diminished emphasis on education vis-a-vis other sectors - in 1996, around 10% of general government expenditure was allocated to education, while in 1999 education absorbed only 8.5% of general expenditure (Table 4.7). 66 See Making Transition Work For Everyone, World Bank, Washington,2000, p. 242 and Human Capital Investment: An International Comparison, OECD, Paris, 1998. 60 Table 4.6: Summary Data on Education Spending 1996 1997 1998 1999 Education spending in real terms Total education spending (1996=100) 100 84 86 78 Local education spending (1996=100) 100 71 57 59 Education spending (°/) Total education spending as % of GDP 3.4 3.2 3.4 3.1 Total education spending as % of govt. spending 10.2 9.2 8.9 8.5 Local education as % of total education spending 15.7 13.2 10.4 11.7 Local education as % of total local spending 11.9 10.0 9.0 8.9 Coefficient of variation of judet spending per capita on 0.21 0.19 0.19 0.19 education (standard deviation/mean) Sources: LSSD database. 1999 data are from Romania: Selected issues and statistical appendix, IMF, 2000. Note: * 1999 data are budget. The progressive shift in financial responsibilities to local governments had an important impact on the composition of total education spending. Figure 4.3 depicts the evolution of real personnel, capital, and total education spending as percentages of 1996 values. The figure highlights that, throughout the period, the evolution of personnel spending (financed by the central government) closely patterned the evolution of total education spending, indicating that salaries have been protected relative to other expenditure categories. On the other hand, spending for infrastructure dropped dramatically after 1998, when the responsibility for its financing was transferred to local councils, reflecting the inability of local governments to cope with these new 67 obligations. Such a drop in capital spending is particularly problematic in Romania where a large part of the school infrastructure is in poor condition. Voicu and Begu (1999, p.17) suggest that this spending pattern is a clear sign of under-funding of education in the country. Given the wide diversity of conditions across the country - and in the absence of a functioning system of intergovernmental transfers - the shift of financial responsibilities for infrastructure to local councils is likely to increase disparities across rich and poor municipalities. 67 Spending for infrastructure was 7.7% of total spending in education in 1996 but was budgeted to be only 3.3% of total education spending in 1999. 61 Figure 4.3: Evolution of Total Education Spending in Real Terms (1996=100) 120 - 100 80 - -u--Personnel 60 - Capital - Total 40 - 20 0 1996 1997 1998 1999 Note: 1999 data are budget. Local Education Spending. The decrease in education spending was substantially more marked at the local than at the central government level. The data show that spending on education by local governments fell by 41 percent in real terms between 1996 and 1999, despite the increased financial responsibilities that were progressively transferred to local councils. Analysis of education expenditures as a share of total local spending provides some insights on how local councils prioritize spending. Table 4.1 highlights that, in parallel with the decrease in the local education spending as a share of total education spending, the share of resources devoted to education within local budgets declined from 13 percent in 1996 to 10 percent in 1998. Multivariate analysis (results are reported in Table 4.A.3) confirms a statistically significant time trend, which appears to follow the substantial increase of poverty nation-wide between 1996 and 1998. It also reveals that across judets the variations in the priority given to education spending at the local level varies positively with poverty (column 4) - i.e. poorer judets were likely to devote a relatively higher share of their budget to education. Once the share of rural population is included in the analysis, the significance of the coefficient on poverty drops while the share of rural population becomes highly significant. This result is not surprising, as the correlation between the share of rural population and the head- count ratio is substantial (0.30). In particular, the results suggest that priority to education spending within local budget is given at moderately rural judets, while councils in extremely rural areas tend to allocate relatively fewer resources to education.68 The association between education spending and the share of rural population in the judet suggests that education expenditure in rural areas is higher because of higher fixed costs in the production of education services in rural areas (rural areas have a higher ratio of schools per 68 The point estimate in table 4.A.3, column 5 suggest that the share of local education spending peaks when the share of rural population is approximnately 54%. 62 capita than urban areas as well as smaller schools). These results are notable and are confirmed by the regression analysis of per capita local education spending (table 4.A.3, columns 1-3). Nonetheless, the level of aggregation of the data (the judet) does not allow us to fully capture the diversity in the cost structure of the Romanian system where overcrowded urban schools co-exist with small multi-grade rural schools with fewer than 10 children per class.69 Quite interestingly, the share of local budgets devoted to education spending is significantly increasing in the number of teachers per capita in the compulsory cycle (primary and gymnasium). This association holds despite (or because of) the fact that the main expenditure associated with personnel - salaries - is paid out of the central government budget. Moreover, the association is virtually unaffected by controlling for the number of schools per capita. This result is consistent with a political economy-type explanation - in councils with a relatively higher number of teachers, teachers (who often participate as elected officials in local government councils) may be able to influence the share of the budget which is allocated to education. Figures 4.4 and 4.5 chart the evolution of local education spending in real terms in the communa and towns of the sampled Judets. Both graphs return an impression of substantial diversity and variation across communa and town as well as between them. With the sole exception of the town of Lupeni, all local councils faced an important drop in education spending when compared with 1996. In 1999, the correlation between the real revenues and real education spending as percentages of the 1996 levels is 0.27 across the 5 communa and 0.31 across the 5 towns, suggesting that the drop in real education spending is in part due to the decrease in real revenues but also reflects a shift in local spending priorities away from education. Figure 4.4 Figure 4.5 Bdacan of local edccUcn q3edrg inthe snpled caurau Evluton of local educaton qgndng in the snipled tns (%of 1996 value% real tenrw ('of 1996 value% real ternv 160.0%-_________ 140.0/0% 16-0/ Trg SRmim 12.0 _SaictHd 140.00%o - 1 20.0/o - 100.00/0 ~ ~~~~~~~~--B±abadg 8-*--- -Sawiai 60.0ok VasiLi 40.0% --VMXI 20.0%- oX QO%00/0 1996 1997 199 19R 1996 1997 1998 199b 69 See Voicu and Begu (1999). 63 Coping Strategies The attribution of new financial responsibilities for education spending to local councils in 1998 coincided with the local public finance reform which increased the dependency of local councils on own revenues. Because of this, some of the poorer local councils have found it difficult to cope with basic financial obligations for education. For example, the local councils of Sanzieni (Covasna), Vorta (Hunedoara), Lugoj and Nadrag (Timis), did not pay student scholarships for low income students. Scholarships were paid only partially in Sarichioi (Tulcea). Moreover, transportation subsidies were not paid in Vorta, and were paid only partially in Nadrag. In some cases, the decentralization of education spending created conflict between local and central administration. For example, in Tirgu Secuiesc the construction works of a new building, that had been initiated by the Ministry of Education and with Ministry funds, came to a halt after that fiscal responsibility for investment spending was transferred to the local council. As a result of the strain put on local council budgets, schools have been encouraged to diversify and increase their financial resources through specific activities such as consultancies, renting school halls, school festivals, or collection of user fees. For example, a school in Lugoj imposed a quarterly fee of ROL 10,000 per child, although no money was collected from poor families. Another school in Lupeni, asked for a contribution of ROL 5,000 per child, but only half of the families were able to pay. In many cases parents were more willing to offer support in kind than in cash, and offer services such as helping with maintenance or painting classrooms. Although principals have autonomy to manage these extra-budgetary funds, the financial autonomy of schools remains limited, as funds can cover only a limited range of school needs. Moreover, the lack of personnel with accounting and managerial skills and the lack of infornation about school budget restrict the ability of schools to manage funds optimally. 4.4. Health Financing of primary health care has been transformed significantly since 1998, as responsibility for funding health has shifted from the state budget to the health insurance system. Prior to 1998, the majority of funds for health (81 percent in 1997) were executed through the Ministry of Health (MOH) and its branch offices at the county level, the Judet Public Health Directorates. These institutions managed both the state budget allocation for health and an extrabudgetary fund for health, financed through a two percent payroll tax and revenues from taxes on tobacco, alcohol and advertising. Beginning in 1998, the government started collecting funds for the new Health Insurance Fund. Health insurance is financed through payroll contributions for the insured population and state budget transfers for non-insured, who include children, the unemployed, and pensioners. As the institutional framework of the Health Insurance Fund had not been fully set up, 1998 was a transitional year with funding coming from different sources and implemented by county offices of the MOH - the Judet Public Health Directorates and newly established Public Health Inspectorate). The judets offices of the national Health Insurance House were set up in 1999 and financing responsibility was then shared with the MOH institutions. National Level Trends. The introduction of health insurance in 1998 led to a increase in aggregate public resources to the health sector, both as a share of GDP and total government spending, as well as in real terms (Table 4.8). This increase was due to the influx of additional revenues from insurance contributions. In 1998 the main source of revenues for health became a 64 10 percent payroll tax, divided evenly between employers and employees. In 1999 the total taxation rate was raised to 14 percent. Local health spending declined in real terms, and as a share of total local spending, as responsibility for financing health was shifted out of local government budgets. Table 4.7: Summary data on Health Spending 1996 1997 1998 1999 Health spending in real terms Total health spending (1996=100) 100 87.5 100.6 113.4 Localhealthspending(1996=100) 100 86.1 3.1 3.0 Health spending (Y) Total health spending as a share of GDP 3.0 2.8 3.2 3.9 Total health spending as a share of govt. spending 8.8 8.2 8.8 10.5 Local health as a share of total health spending 19.1 18.8 0.6 0.5 Local health as a share of total local spending 12.4 12.8 0.5 0.4 Coefficient of variation of judet spending per capita on 0.23 0.32 0.33 0.28* health (standard deviation/mean) Source: National Commission for Statistics, IMF; * based upon budget data Collection of insurance contributions for health is undertaken locally by the Judet Health Insurance Funds. As a result, increases in revenues have depended upon the fiscal capacity of the judet and the capacity of the judet insurance fund to collect taxes. As a result, the increase in health expenditures has not been uniforn across judets. As shown in Figure 4.6, real expenditures were below the national average in Tulcea, Vaslui and Hunedoara, and higher in Bucharest, Covasna and Timis. This reflects the distribution of health facilities across Romania, as Bucharest has a higher concentration of hospitals than the other judets and serves patients from other areas of the country. Figure 4.6: Health Expenditures in Real Terms (1996=100) 160 - 140- 120 - - + Covasna g 100 - _-- Hunedoara 1080 - ejA Timis 80- ON 60 -- Tulcea 40 - WN Vaslui 20 - - Bucharest 0 - - *-Romania Total 1996 1997 1998 1999 Between 1996 and 1997 local councils (including municipalities and communa) were responsible for financing part of the expenditures for dispensaries (local health care units) including maintenance and medical supplies. This amounted to approximately 13 percent of health care costs in 1996 and 1997 dropping to 0.05 percent in 1999. The responsibilities of local councils for health were reduced to funding health services in creches and blood donation activities. 65 Spending on health across judets was analyzed to understand the impact of changes in financing levels at the local level.70 As shown in Table 4.8 above, the variation of spending across judets, as measured by the coefficient of variation, has increased between 1996 and 1999. This is despite mechanisms built into the health insurance system which are intended to redistribute resources across judets. According to the legislation, a share of revenues from insurance contributions are to be redistributed across judets to equalize revenues. Analysis conducted prior to the reform suggested that redistribution of 22 percent of revenues would ensure that health care revenues in all districts would be at least 95 percent of the national average. In practice, 25 percent of resources have been redistributed, however the effect has not been as intended. In 1998 spending in over half of the judets was below 95 percent of the national average. This has been due to the lack of adequate criteria for reallocating revenues according to needs. 4.5. Summary Issues In the last few years, wide-ranging sectoral reforms as well as the enactment of the Law on Local Public Finance (LLPF) have substantially changed the role of local governments in financing social sectors. Local governments' financial responsibilities have increased in education and social assistance, while the introduction of the health insurance system has minimized local governments' involvement in health spending. In 1998 spending for school infrastructure and equipment was transferred to local councils, leaving salaries as the main central government financial responsibility in the sector. Between 1998 and 1999, financing for child protection and for services for the disabled were also shifted to local governments. This was in addition to the fact that spending for a number of important cash benefit programs, including the Means Tested Social Assistance Benefit (MTSAB), and spending for other services for the poor (soup kitchens and home care), were already the sole responsibility of local councils. As a result, local governments finance almost 50% of all social assistance spending. Local governments were faced with these new financial responsibilities in a time of country-wide fiscal tightening. In 1999 the Law on Local Public Finance changed the structure of local revenues, by making local revenues much more sensitive to local resources and producing relevant disparities in local capacity across the country. When confronted with these constraints, local governments resorted to various coping strategies. Local governments may have tended to cut spending in areas of least resistance. For example, local councils cut spending for school infrastructure and for social scholarships. Overall local spending in education as a share of total spending in education fell notwithstanding the new obligations. Expenditures for maintenance of education facilities and capital investments were particularly eroded, threatening the overall sustainability of the system. In social assistance, overall spending fell in real terms and local governments shifted priority away from cash transfers towards in-kind benefits. In particular, eligibility for MTSAB was substantially restricted and often the benefit was not paid at all. The shift of financial responsibilities to local governments and the higher sensitivity of local revenues to local resources have important implications for efficiency and equity, all the more so in a country where differences in fiscal capacity are very accentuated. The reduction in 70 Judet level data were collected directly from the MOH, MOF and Health Insurance Funds. Trends in this data differ from the national level data presented earlier, the reasons for these discrepancies are unclear. 66 capital spending for schools is particularly problematic and is tantamount to a substantial under funding of education, since the stock of existing infrastructure is in generally poor condition. Moreover, due to the new revenue arrangements, poorer local councils, although more likely to have schools in need of repairs and investment, have the least resources to face these needs. Similarly, local councils with a higher percentage of people below the poverty line have fewer resources to support the poor and are also more likely to resort to cutting social assistance programs in favor of more pressing or politically rewarding spending. In the next chapter, we will discuss in detail these and other implications of the new institutional and financial arrangements for social service delivery at the local level. 67 Table 4.A.1 Trends in Education Spending by Local Governments Current million ROL Evolution over time Percentage of total (% of 1996 values) 1996 1997 1998 1999 1996 1997 1998 1999 1996 1997 1998 1999 Total Spending 590,358 1,040,998 1,209,235 1,780,420 100% 71% 57% 60% 100% 100% 100% 100% of which: Kindergartens 76,787 144,053 175,747 266,316 100% 75% 64% 69% 13% 14% 15% 15% Compulsory education (1-8) 241,725 442,185 527,999 724,480 100% 73% 61% 60% 41% 42% 44% 41% Secondary education (gymnasium, vocational, professional, special) 242,389 398,415 493,475 774,039 100% 66% 57% 64% 41% 38% 41% 43% University - - - - - - - - - - - - Orphanages (3-18 years) 25,708 47,801 - - 100% 74% - - 4% 5% - - Other 3,749 8,544 12,014 15,585 100% 91% 90% 83% 1% 1% 1% 1% Source: LSSD database. Budget data for 1999. 68 Table 4.A.2 Trends in Local Social Assistance Spending MTSAB beneficiaries Per capita spending' Share of total local spending (as % of totalpopulation) # (1) (2) (3) (4) (5) (6) (7) (8) 1996-1999 1996-1999 1996-1999 1996-1999 1996-1999 1997 and only 1997 only 1998 1998 Year: 1997 -5.9** -4.18** -0.015** -0.017** -0.028** (-13.72) (-6.18) (-6.49) (-6.73) (-6.73) Year: 1998 0.88 2.69** 0.03** 0.03** 0.018** -0.0045** (1.29) (2.75) (9.67) (10.25) (3.41) (-4.08) Year: 1999 -3.9** -1.42 0.008** 0.009** -0.003 (-5.73) (-1.23) (1.92) (1.85) (1.32) Poverty 0.03 0.07* 0.7** 0.005 0.13** 0.25** 0.03 (lagged head-count ratio*l000) (0.81) (1.77) (2.27) (1.78) (2.05) (3.09) (0.50) Rural poverty (lagged) -0.1 (-0.37) Urban poverty (lagged) 0.8** (2.26) Total local revenues per capita 39.22* -0.28** 0.02 0.009 (3.03) (3.71) (0.76) (0.98) Constant 13.79** -4.98* 0.05** 0.06** 0.12** 0.0048** 0.001 0.003 (10.82) (-1.73) (6.43) (6.8) (7.11) (2.32) (0.03) (1.62) R2 0.45 0.48 0.49 0.51 0.54 0.19 0.13 0.02 N (judets per year) 40,40,40,37 40,40,40,37 40,40,40,37 40,40,40,37 40,40,40, 37 38, 37 38 37 Coefficients are calculate with OLS. All regressions exclude Bucharest and Ilfov. t- statistics are reported in parentheses and standard errors are calculated correcting for heteroskedasticity and correlation within Judet clusters. IAII coefficients are multiplied by 1000. ttRegressions for the share of MTSAB beneficiaries also exclude Brasov for 1997 due to likely measurement error and Dolj (outlier). ** indicates significance at 5% or more; * indicates significance at 10% or more. 69 Table 4.A.3 Local Education Spending Per capita local spending Share of total local spending (1) (2) (3) (4) (5) (6) Year: 1997 -0.008** -0.008** 0.008** -0.01** -0.01** -0.02** (-10.22) (-8.18) (-8.7) (-5.08) (-5.18) (-5.00) Year: 1998 -0.011** -0.011** -0.011** -0.3** -0.03** -0.03** (-12.7) (-12.8) (-13.44) (-8.56) (-8.3) (-7.95) Year: 1999 -0.01** 0.02** (-10.8) (-5.7) Lagged poverty -0.01 0.03 0.03 0.06* 0.1 0.1 (head-count ratio*l000) (-0.17) (0.34) (0.31) (1.89) (0.45) (0.34) Share of rural population 0.08** 0.03 0.37** 0.21* (2.8) (1.43) (2.96) (1.76) Share of rural population -0.09** -0.04 -0.34** -0.18 squared (-2.19) (-1.7) (-2.48) (-1.38) Schools per Capita 0.49 -1.56 (0.17) (-0.13) Teachers per capita 2.52** 9.39** (3.11) (3.11) Constant 0.02** 0.007 -0.001 0.02 0.02 -0.007 (12.4) (1.5) (-0.19) (12.4) (1.02) (-0.23) R2 0.53 0.54 0.62 0.23 0.40 0.49 N 40,40,40,37 40,40,40 40,40,40 40,40,40,37 40,40,40 40,40,40 Estimation is obtained with OLS. t-statistics are reported in parentheses and standard errors are corrected for heteroskedasticity and correlation within judet clusters. Regressions are for 3 or 4 time periods (1996, 1997, 1998, and 1999) and 40 Judets. Expenditure data for 1999 are budget. Bucharest and Ilfov appear to be outliers and therefore excluded from the sample. * indicates significance at 10%; ** indicates significance at 5% or more. 70 Chapter 5: Service Delivery at the Local Level 5.1. Service Delivery During the transition period, poverty in Romania has increased notably (Table 5.1). This has occurred in a country where differences in poverty across regions and between urban and rural areas are extremely marked. Poverty trends mirror patterns in human development indicators and, to some extent, in the supply of social services. This chapter draws upon the beneficiary assessment, as well as quantitative data, to examine the situation of service delivery across the country. It first looks at access to social services and implications for equity across local governments. The section focuses on critical determinants of access, geography and the contrasting circumstances in rural and urban areas, and ethnicity and the exclusion of Roma. The chapter then looks at ways in which resource allocation effects the distribution of services across the country and among households. Finally, the chapter concludes with a discussion of public participation and the responsiveness of local services. In order to understand how the changing institutional arrangements and financing trends have affected actual delivery of services at the local level, in depth interviews and focus groups were held in all of the case study sites with current and potential beneficiaries of social services, as well as local service professionals including teachers, doctors and community leaders such as priests and Roma leaders. The results of these discussions were mixed, sometimes contradictory and difficult to generalize. However, they did provide valuable insights into the concrete experiences Romanians in different areas and circumstances are having with social services, and supplement the quantitative data on service delivery and outcomes from other sources. Unsurprisingly, beneficiaries and providers alike frequently identified fiscal constraints as the primary obstacle to service provision. In some cases, people pointed to services which were no longer delivered, such as means-tested social assistance, student scholarships or transportation to schools. In others, people noted the deterioration of schools and health facilities due to the lack of funds for maintenance, and operating costs, such as fuel. Some beneficiaries who experienced a withdrawal or decline in services expressed nostalgia for the Ceaucescu regime. This is consistent with findings of studies of poverty in other post communist- countries.71 The interviews also highlighted the low level of information flows between beneficiaries of services and providers. With few exceptions, social service providers do not generally solicit feedback from recipients through customer satisfaction surveys, or other channels. Beneficiaries were rarely asked for their views on the level, types and quality of services provided. The discussions pointed to the lack of a 'service culture' in Romania. Neither beneficiaries nor providers expressed the sentiment that public education, health and social services were public services, paid by taxpayer funds and delivered by public servants. Rather people seemed to have low expectations. A woman client to a soup kitchen from Lugoj noted: We can't complain. For us it is important to have this meal, good or bad. It is more than nothing. It keeps us alive. 71 See for example, "Bulgaria: Consultations with the Poor," 2000. The World Bank, Washington, D.C. 71 5.2. Access Access to social services varies across local governments as a result of a range of interrelated socioeconomic, geographic, institutional and financial factors. In particular, the case studies highlighted significant differences in conditions in rural and urban areas. While few of those interviewed in the urban sites noted difficulties in accessing health and education facilities, rural residents in a number of cornmunes noted problems in the availability of and access to these services. Rural areas face special challenges in designing effective arrangements for service delivery. Similarly, the case studies illustrated that ethnic diversity, and especially the presence of Roma communities influence service delivery. Geography There is a strong regional dimension to poverty. Contrasts between rural and urban areas are stark in Romania, as poverty is significantly higher in rural areas. For example, in 1999, out of 10 million people living in rural areas, 4.2 million were living below the poverty line (Table 5.1). Many rural residents rely on subsistence agriculture. Table 5.1 Poverty Rates in Romania 1995 1996 1997 1998 Poverty rate' (total) 26.8 21.2 32.5 35.7 Rural poverty 32.6 25.4 37.9 40.9 Urban poverty 22.0 17.6 27.7 32.8 1 The poverty rate is defined as the percent of population with consumption below 0.6 of per adult equivalent (Headcount ratio). Source: Romania LSMS 1995-1998. Aggregate indicators also illustrate lower coverage of social services in rural areas and gaps in health and education indicators (Table 5.2). In particular, survey data indicate that enrollments are lower in rural areas and that, between 1994 and 1998, rural enrollments for the poorest quintile of the population dropped from 93 percent to 90 percent.72 Table 5.2: Socioeconomic Status, Rural and Urban Areas, 1998 % of population Life Expectancy IMR Primary School Enrollment Urban 54.9 69.8 18.5 95.5% Rural 46.1 67.8 25.0 93.2% Source: Romania Integrated Household Survey, NSI Coverage of social services is lower in rural areas for a number of reasons. Rural settlements can be geographically isolated, making it more difficult for facilities to be maintained and for personnel to travel to provide services. Access to services is particularly challenging in larger communes with disparate and remote villages. Physical infrastructure is poorer in many communes, especially roads and telecommunications. Related to this, the unit costs of service provision are higher in rural areas. The case studies illustrated that these problems were 72 Although average urban and rural enrollment appear quite similar, a t-test of difference of means confirms that rural enrollment is lower than urban enrollment in a statistically significant sense. 72 particularly salient for health and education, although geography also affects access to social assistance benefits and services. Some beneficiaries in remote areas with limited access to public transportation noted the difficulty of getting to the local council office to apply for benefits. Although an assessment of the direction of causality in the association among poverty, poor social sector infrastructure, and gaps in human development indicators is beyond the scope of this work, the evidence suggests that these mechanisms might be self-reinforcing. The following sections draw together findings from the beneficiary assessment of access to health and education. Education. Residents of the communes Vorta and Zapodeni noted that some children had to travel for nearly one hour to get to school. Long distances discourage children from attending school and reduce their capacity to learn when they arrive. The problem is particularly pronounced for lower secondary schools (grades V to VIII), as these tend naturally to be more sparse over the territory. Teachers noted that some children arrived at school too exhausted to participate. The erosion of financial resources for school transportation has made access to school even more difficult. In 1999 the situation worsened as financial responsibility for transportation subsidies was transferred to local councils, many of which stopped paying the subsidy. This was the case in Vorta. However, despite the obstacles, teachers noted that families did make an effort to send their children to school. A school principal in Vorta noted: "The child who lives the farthest away is coming from about 2 km. Only 4 kids are from around here, the others come from far away. Even when the weather is bad they come to school. " Higher poverty levels in rural areas also threaten the ability of families to send their children to school. For poorer households the opportunity costs of sending children to school can be quite high, particularly in rural areas where agricultural activity is the main source of household income. Parents and teachers in communes noted that in some cases children missed school to work in the fields, or to help with household activities, such as child care for younger family members. The principal in Vorta noted: "Parents are happy when their kids have good grades. But parents do not help them at all. They have the children working in the household caring for animals or cultivating the land. That is why the children are tired when they come to school and because of this they cannot learn. " In this context, the fact that child allowances are linked to school participation may have helped to encourage attendance. Local governments and school principals also have an incentive to promote attendance, as there is a lower limit on the number of students schools need to have in order to stay open. Health. Similar issues were noted for health care. Coverage of facilities was lower in larger communes with more dispersed populations, such as Sarichioi, Zapodeni and Vorta, while fewer problems were noted in smaller communes with concentrated populations such as Sinzeni, which has three villages and Nadrag, which has two villages. Proximity to urban areas also facilitates coverage, as is the case with Sinzeni which is five kilometers from Tirgu Secuiesc. In the dispersed areas, coverage is more difficult. Inhabitants of Zapodeni and Vorta noted that ambulances usually come to the center of the commune, but sick people often have to be transported there by cart. 73 Roads are so poor in some areas that it is difficult for residents to use ordinary cars for transportation, particularly in the winter. In some cases there are few (or no) telephone lines available in the commune, which makes it difficult to report health emergencies. A resident of Macresti-Zapodeni, a village in Zapodeni commune, noted the challenge of getting to a physician: "I have not seen a physician for years. You have to pay 50,000 lei for transportation to Zapodeni (15 km away). The physician only comes here on Wednesday and Friday... We just walk and die walking. From here to Zapodeni it is 2 hours to go with the cart. After this traveling the sick get sicker. " The case studies noted additional problems in rural areas, including the lack of utilities for health facilities such as water and gas supply and heating. The level of development of the commune influences the availability of services. In this regard Nadrag represents a unique case. Although it is a commune, prior to restructuring Nadrag was an industrial center. It was home to a large enterprise which was the major employer in the area. As a result, Nadrag had more extensive health facilities than other communes. Although the enterprise has been closed, the health services were transferred to the Ministry of Health and continue to operate. The reverse could also occur, if the closure of the enterprise led to the collapse of health facilities, and if local facilities were not able to assume the responsibilities. Residents of the communes cope with the low level of services in different ways. In some cases people consult with other health workers such as midwives, pharmacists in neighboring cities, or opt for traditional medical treatment. For more serious cases beneficiaries reported checking into the judet hospital. However because of the prevalence of informal payments in hospitals, some poorer individuals were unable to afford care. Incentives for Service Providers. An issue frequently raised by beneficiaries, particularly in communes, was lack of qualified personnel in remote areas. All of the beneficiaries and officials interviewed acknowledged the difficulty of attracting physicians to work in local dispensaries when adequate living conditions were not guaranteed. In a number of the communes, residents recalled that prior to 1990 there had been a physician in every village. This was a result of communist era policies which required doctors and teachers to serve in remote areas. Following the transition, many professionals left for urban areas. A nurse in Zapodeni noted: "It is a serious problem with physicians. Nowadays everyone wants to go to the city. That's why we do not have a local physician. Can you imagine someone coming to such a place? We should probably insure them better conditions of living and they would come. " Similar issues were noted for education. In a number of cases, communes reported difficulty in attracting qualified teachers to their areas. As with health, communist policies forced teachers to settle in rural areas, however with the lifting of this regulation, many have left. The mayor of Zapodeni remarked: "How can we attract good teachers in these conditions? We don't have running water, we don't have a lot of things that make urban sites more 74 attractive. Even if you would give them a piece of land you couldn 't find anyone to move to this poor village. " In three of the communes, Sarichioi, Zapodeni and Vorta, teachers and principals reported that they were attempting to encourage talented students to continue their education in nearby pedagogical high schools and universities, with the hope that these students would come back to the commune after graduation to teach. Although the interviews indicated that so far few of these students were actually coming back to their villages. At present, a premium ranging between 5 and 80% of the salary can be paid to teachers in rural areas, however the effectiveness of this measure is not known. Salary premia for teachers in rural areas are paid out of the central budget and are awarded through a governmental decision on the basis of several criteria, including availability of water and electric supply, geographic characteristics of the area, and the need for teaching personnel in the area. In 1999-2000, 386 billion Lei were spent on salary premia, corresponding to approximately 5% of the total salary bill. While the urban sites reported notably fewer problems in staffing schools, there were a number of exceptions. Some schools in less desirable peripheral areas outside of urban centers, and schools with large shares of Roma students had difficulties locating qualified teachers. In the Romanian educational system 'qualified' teachers are defined as teachers who received a degree in the subject they teach. While many beneficiaries and school personnel pointed to the lack of qualified teachers as a serious issue, the impact of 'qualified' teachers on educational outcomes needs to be rigorously assessed. In their study of the deterrninants of test performance for primary school students, Voicu and Begu (1999) find that pupils of teachers who hold a teaching degree outperform pupils of new or unskilled teachers. More, importantly, they find that school location in a rural area is a substantially more relevant determinant of low student performance than the qualifications of teachers. This would suggest that targeting policies on structural characteristics that are highly correlated with rurality - such as poverty or poor infrastructure conditions - might be an effective way of improving students' performance. Lack of personnel also effects social assistance services. In some judets up to 50 percent of positions for physicians in institutions were unfilled. JSIPHs and local councils do not have mechanisms for introducing incentives to recruit personnel. Institutions are frequently isolated and located in rural areas where people are reluctant to live. In addition, salaries for staff working in institutions are among the lowest on the civil service pay scale. The level of employment varied across areas. In Timis, 76 percent of positions for staff in institutions for the disabled were filled, the figure was higher in Vaslui and Bucharest. JSIPHs also have difficulty recruiting staff and, as a result, the capacity of these institutions to function and to initiate new programs is limited in some cases. Lack of personnel limits the capacity of local administrations for monitoring and evaluation. In particular, the activities of assistants for disabled go largely unsupervised and monitored. The assistants are employees of the JSIPHs, but because of the large number of assistants and the small number of personnel in the JSIPHs it is not possible for the JSIPHs to monitor the activities. A general inspector commented, "Usually because of the great work-load the field inspectors conduct social inquiries only when hiring a personal assistant, what happens after, nobody knows." 75 Social Exclusion and the Roma Another theme which carne out strongly in the quantitative and qualitative analysis was that Roma are more vulnerable to exclusion and gaps in service provision than other groups in society. This is a complex issue which is closely linked with the overall welfare and poverty situation of Roma in the country. The available data suggests that Roma are among the poorest in society (Box 5.1). Roma face challenges in accessing social services in large part because of their high poverty levels, but also as a result of unique circumstances including cultural and sociological factors and discrimination, leading to a vicious cycle of poverty and exclusion.73 Box 5.1: Poverty and the Roma in Romania Measuring poverty among the Roma in Romania and other countries is an extremely difficult task. Seemingly straightforward questions, such as estimating the size of the population within a country, prove extremely challenging. Household surveys and censuses often do not distinguish respondents by ethnicity74, and even when ethnicity is included, a range of issues arise, including undersampling of areas where Roma are likely to reside, difficulties in locating and identifying populations which may not be officially registered, and problems with self-reporting. Roma may opt not to self-identify for various reasons, such as fear of discrimination, or political motivations (Druker, 1997; Liebich, 1992). Analysis of the Roma is also complicated by the diversity of the populations. Romania's Roma population is heterogeneous and contains various groups and subgroups based on historical, religious, occupational and other divisions. The Romania Integrated Household Survey (RIHS) is one of the few household surveys in the region which distinguish respondents by ethnicity. While the results are informative they do not provide a complete picture, because of the small share of the sample identifying as Roma. The survey, which encompasses about 40,000 households, has consistently identified about 2 percent of the survey population as Roma between 1994 and 1998. This is in contrast with estimates by local officials and commnunity leaders which indicate that the share of the Roma population in Romania could be as high as 9 percent (Ringold, 2000). Results from the RIHS indicate striking levels of poverty among the Roma, in comparison with the rest of the population. The poverty line used in Romania was set at 60 percent of mean consumption, per adult equivalent. For three survey years (1995-97) the poverty rate for Roma was significantly higher than for the other population groups, reaching 79 percent in 1997, in comparison with the national poverty rate of 31 percent. Between 1995 and 1997, poverty increased for the entire population and the gap between Roma and non-Roma poverty rates narrowed slightly. Table 5.3: Romania, Poverty by Ethnicity, 1995 and 1997 % of Poverty Rate Poverty Depth Poverty Severity population (97) 1995 (%) Share 1997 (%) Share 1995 1997 1995 1997 Romanian 89.8 24.5 87.4 29.7 86.6 6.1 7.3 2.2 2.7 Hungarian 6.8 22.8 6.1 28.4 6.0 5.7 6.7 2.2 2.3 Roma 2.3 76.4 5.4 78.8 6.3 28.2 33.2 13.1 17.1 Other 0.3 23.5 1.1 32.6 1.1 6.4 8.9 2.6 3.6 Romania 100 25.3 100 30.8 100 6.4 7.9 2.4 3.0 Source: Romania Integrated Household Survey, 1995 and 1997, from Tesliuc and Pop, et at, 2000. Notes: The poverty line is 60 percent of adult equivalent consumption. 73 For more information on these issues refer to Ringold, 2000. 76 In order to gain a more complete understanding of poverty and expenence with social services among different types of Roma communities, additional qualitative fieldwork in Roma settlements was undertaken alongside the case studies in the five judets. Interviews were conducted in nine Roma communities with different geographic characteristics (e.g. rural/urban), levels of ethnic and religious diversity and socioeconomic conditions. Results from this work are included in this report. Because of frequent geographic isolation at the periphery of villages and towns, or in urban ghettos, Roma communities may be more affected by gaps in physical coverage than other groups. Roma communities may lack access to services because of the absence of facilities or personnel, or because of lack of transportation to a neighboring town. In Bucharest Sector 5 the dispensary had reportedly moved to a different part of the city because of tensions between Roma and health care personnel. In other areas Roma reported that ambulances refused to travel into Roma neighborhoods. Issues related to Roma culture and the organization of Roma society impact access and relations with social service providers, although because of the diversity of Roma culture and communities these factors are difficult to define. Roma children may have difficulty starting school because of limited proficiency in Romanian. Similarly, parents may have difficulty communicating with teachers, understanding health personnel and maneuvering through local government offices to access social assistance. Poor communication between Roma patients and doctors perpetuates stereotypes and breeds mistrust on both sides. The case studies found that some doctors and nurses have resorted to innovative, and frequently underhanded, strategies to encourage immunizations, including making threats that Roma would have to pay for the services, enlisting the police, and linking the distribution of in-kind social assistance with immunizations. A physician in Babadag explained: "They have never been willing to let their children be vaccinated. We have to motivate them with methods adequate to their values. For instance we told them that a vaccine is very expensive, 60, 000 lei, and we administer itforfree now, later they will have to buy it. We threatened some illiterate parents with false papers, telling them that they will have to pay penalties if they do not let their children to come for vaccination. (She shows us such a false penalty certificate, smiling) We put a lot of stamps on it... Sometimes we brought policemen with us in the area to be more convincing. And it worked many times. " While possibly effective, these tactics likely perpetuated the mistrust between health personnel and Roma, and fail to address the underlying lack of information and understanding about the role of immunizations. Discrimination against Roma is difficult to measure, however examples did arise in the qualitative analysis. It is important to note that discrimination exists on both sides - Roma have negative stereotypes of non-Roma, and vice-versa, reinforcing social exclusion and the separation of populations. Discrimination against Roma by non-Roma parents, children and teachers contribute to low attendance and can both discourage children from attending school and affect the quality of education children receive in the classroom. 74 A recent review of World Bank Living Standards Measurement Surveys (LSMS) datasets for countries in Europe and Central Asia found that only two out of nine surveys distinguished respondents by ethnicity. 77 In some sites, both non-Roma parents and teachers expressed a preference for ethnically homogenous classes and especially in urban areas - where the density of Roma population is higher - the distribution of Roma children in classes and across schools is subject to negotiation among principals. Stereotypes about Roma and their attitudes toward education lower teachers' expectations about the potential of their students. Discrimination can be both explicit, as in the case of schools creating separate classes, or more subtle, for example if parents discourage their children from interacting with Roma classmates. A parent in Bucharest noted: "I cannot say that school no. *** is professionally worse or better than others. All I know is that the children are worse. I saw their behavior. That 's why I run away from * * *, because all gypsies are there. " Roma respondents pointed to fear of discrimination as an obstacle to school attendance. Families said they feared that their children would not be really accepted into the regular schools. The fear of discrimination, from both teachers and fellow students, is often a legitimate one. One father of a Roma student from Zabrauti, a Roma neighborhood in Bucharest, described the kind of stereotypes directed against his son by other non-Roma students. "The other children look at my children differently: 'Look at the Gypsies from Zabrauti!' Others are calling them 'Ghetto Boys!' ... So I sent my boy to the school with special classes, because it is closer and I have heard there are some assistance, somefree notebooks... " Other parents complained of discrimination against Roma and favoritism towards non- Roma students. Discrimination ranged from ignoring their needs, pejoratively calling them "Gypsies," and even violent treatment. In Valcele, a Roma settlement in Covasna, some students complained that teachers either ignored them or physically abused them. During the winter, for example, one parent told of Roma children being forced to stay outside during breaks since, as one teacher argued, if Gypsy children were along to congregate around the stove, their "unpleasant smell" would spread throughout the classroom. Students also complained of favoritism towards non-Roma students in which the teacher only offered help to non-Roma students on assignments. Discrimination was also found to influence access to health. In Babadag, near the Black Sea, a doctor reportedly stopped providing subsidized medications to Roma patients, because she had received written instructions to provide them only to those who were insured. The doctor noted that only one Roma household, a relatively well-off family engaged in cross-border informal trade, had paid the contribution for health insurance (approximately US $20 per month). Another serious issue with equity consequences is the ability of physicians to select only low risk patients for their primary care lists. As a result, discrimination and screening of patients has left some Roma uninsured. A physician in Tirgu Secuiese, a town in Covasna County in the Transylvania region explained: "I do not register gypsies as a family physician, I do not accept gypsies ... They come here and ask for money, ask for medicines ... They have a lot of nerve. You have to keep an eye on them when they enter here. I do not think they are poorer than other people. They go to Hungary with business, they probably have more money than we have. They go by car to askfor social aid." 78 Gaps in service delivery for Roma communities highlight the need for monitoring of social welfare and services across the country and tracking the impact of policies and programs on vulnerable groups. National level programs and services may not necessarily be effective at reaching marginalized communities, such as Roma, when implemented at the local level. Similarly, local governments may not face clear incentives to reach out to Roma communities. 5.3. Allocation of Resources and Services The difficult fiscal situation of the last few years, the strong dependence of local finances on local resources following the introduction of the LLPF, and the shift of spending responsibilities to the local level have combined to affect both the adequacy of social service financing as well as how equitably public resources are allocated across the country. Both these effects were strongly identified by beneficiaries and providers alike in the interviews. As discussed in chapter 4, budget constraints have led to the cessation or scaling back of services in many areas. Decentralization has the potential to improve the cost-efficiency of services. This is particularly relevant for Romania, where inefficiencies in the distribution of health and education services, as well as a strong emphasis on institutionalization of children and elderly have, amongst other issues, led to low cost-effectiveness. The level of aggregation of our data does not allow us to perform an in-depth efficiency analysis in the different sectors. Therefore, the following discussion focuses on a qualitative discussion of the issue, as well as data from the case study sites. The section discusses issues arising from developments in the allocation of resources, focusing on the adequacy of financing, as well as efficiency and equity concerns. Education. School principals and teachers almost uniformly stressed the difficulty of obtaining funds for current expenditures, particularly for heating and maintenance of school facilities. Schools which were not serving as budgetary execution centers (CEBs) reported having the most difficulty. The cessation of spending on maintenance, compounded with the poor state of existing facilities, is particularly problematic. Data for the municipalities and communes in this study highlight notable differences in the allocation of education resources across the country, and particularly between urban municipalities and rural communes (Table 5.4) Because of higher population density, schools serve larger population groups in municipalities than communes. In small communes, the number of students per school is extremely low, for example in Vorta, the least populated commune in the study, there were 13 students per school, and in Zapodeni, 24.75 Despite this high level of facilities to students, as discussed above, some children in both communes had difficulty getting to school because of long distances, poor roads and transportation. Similarly, ratios of students to teachers are lower in most communes, 7 in Vorta and 9 in Zapodeni, the ratio of students to teachers was low, reaching 7 and 9 respectively. Parents in both of these communes noted the lack of qualified teachers in their area. 75 Note that data are from 1996 and may not precisely reflect the circumstances at the time of the interviews in 1999. 79 Table 5A: Education Services in the Study Municipalities and Communes (grades 1-8),1996 Judet Local Council Population Students per Students per Exp. per per school school teacher student ('000 lei) Covasna Tirgu Secuiesc 2,848.6 106.9 12.0 94.2 Sanzieni 1,184.5 62.5 11.6 148.0 Hunedoara Lupeni 5,474.0 233.3 22.9 47.0 Vorta 209.8 13.3 7.2 243.3 Timis Lugoj 3,310.3 119.9 10.8 144.8 Nadrag 1,799.0 72.0 12.9 103.7 Tulcea Babadag 2,713.0 142.2 17.3 181.4 Sarchioi 1,358.0 88.3 8.3 169.0 Vaslui Vaslui 4,212.7 233.8 18.2 55.1 Zapodeni 491.8 24.3 9.3 178.6 Sources: National Commission for Statistics, local budgets. With the exception of Timis and Tuicea, expenditures per student are significantly higher in rural communes than municipalities. Although an imperfect proxy, expenditures per student can be a useful indicator for cost differentials. Average unit costs differ substantially across judets. According to a recent study, large school size, the presence of multi-grade education and a high number of shifts are correlated with lower unit costs (Voicu and Begu, 1999). Despite the issues related to access to schools in rural areas, the data on expenditures, and the low student per school ratios suggest that increasing the level of education services in rural areas would not necessarily be an efficient mechanism for addressing gaps in access to schools. This issue is of particular concern for primary schools. There are about 1,370 very small schools in Romania. A share of these have fewer than 10 students enrolled in rural areas and the remainder, in urban areas, have less than 13 students. If some of these schools could be closed or combined with gymnasia, alternative solutions to getting children to schools, such as improving transportation within communes and to schools in neighboring communes and municipalities, could be implemented. Health. Data from the local governments included in this study indicate substantial variation in the distribution of services across the country (Table 5.5). Because of higher population density, dispensaries serve significantly larger populations in urban than in rural areas. However, this does not necessarily reflect lower access to services in urban areas, because of the prevalence of hospitals and other tertiary health facilities in municipalities and towns. Residents of rural areas may also seek care in neighboring urban areas. More indicative of actual service levels are the numbers of population per physician, which vary significantly between the municipalities and communes. In the least populated communes, Vorta and Zapodeni, one physician serves the entire locality. 80 Table 5.5: Health Services in the Study Sites (per 1,000 pop.), 1996 Dispensaries Doctors Total Population Covasna 3,998 768 231,872 Tg. Secuiesc 4,558 495 22,789 Sanzieni 4,738 2,369 4,738 Hunedoara 3,378 593 543,848 Lupeni 6,569 557 32,844 Vorta 1,259 1,259 1,259 Timis 3,395 444 692,645 Lugoj 2,613 355 49,654 Nadrag 1,799 1,799 3,598 Tulcea 3,033 1,003 266,897 Babadag 2,713 543 10,852 Sarichioi 8,148 4,074 8,148 Vaslui 4,095 1,142 462,703 Vaslui 5,003 560 80,041 Zapodeni 3,934 3,934 3,934 Source: National Commission for Statistics Child Protection. In the area of social services for children, elderly and the disabled, cost- effectiveness and efficiency are closely related to the legacy of institutionalization. Social services in Romania, like other countries in the region, remain overly dependent upon institutions, rather than more effective, humane and less costly community based services. In the area of child protection, access to institutional care is generally not an issue. There is excess capacity in child welfare institutions in nearly all of the judets included in this study, except for Vaslui (Table 5.6). Table 5.6: Capacity of Child Welfare Institutions,1999 # of institutions # of places # of children Occupancy rate Covasna 7 410 301 73% Hunedoara 9 1086 783 72% Timis 9 1362 1324 97% Tulcea 4 666 410 62% Vaslui 5 505 514 102% On the other hand, there is a lack of availability of alternative forms of care, including those targeted at prevention of institutionalization, and services such as foster care. Mostly such services cover only a small segment of the children and families at risk. Informal Payments for Services Tight budget constraints have led to a high prevalence of out-of-pocket charges for services. Both formal and informal charges are widespread in each of the three sectors, particularly education and health. In education, in particular, the growing importance of charges can be linked to increasing budget constraints at the local level. As schools have faced declining budgets for recurrent costs, schools have increasingly turned to parents, frequently through parent-teacher associations, to fill the gap. The presence of out-of-pocket costs has significant 81 equity implications for access to services and circumvent the efficiency of the system by distorting incentives for providers and draining resources from the system. Education. Budget constraints in education have increased costs to households. Although compulsory public education is legally free in Romania, households face a range of various formal and informal out-of-pocket expenses, ranging from the costs of school lunches and materials, to charges for tutoring and extra classes. Household data suggest that out-of-pocket education expenses are widespread and make up for a notable of total household expenditure. Wealthier households spend more and are more likely to spend than poorer households. Out-of- pocket spending as a percentage of total household spending, as well as the share of households that make expenditures on education increase with affluence (here proxied by household expenditures). The nature of the data makes it impossible to distinguish private tutoring spending from other education related expenses, such as school materials. Individuals were frequently reluctant to discuss informal payments during the fieldwork. Nonetheless, anecdotal evidence suggests that private tutoring might make up for a substantial share of out-of-pocket expenses. Table 5.7 Out-of-Pocket Expenditures on Education (% of households with school age children who spend money on education) Quintile' Secondary Primary % of total household expenditure Education Education 1 1.1 0.6 4.8 2 5.0 3.4 4.9 3 11.6 7.5 5.6 4 15.0 10.7 6.1 5 22.5 15.1 7.5 Total 10.5 6.7 6.9 Note: (1) Indicates household expenditure quintile. Source: 1998 Romania RIHS. Private tutoring has historically been widespread in Romania, particularly for entry into university and secondary education. Tutoring has been common in mathematics, physics, philosophy, history and, more recently in foreign languages and information technology. Parents may opt for tutoring to ensure that their child can continue to the next grade or level of education, or to improve their labor market potential. To the extent that private tutoring is a matter of individual choice, it does not necessarily have adverse implications for equity. In a system where the supply of education is mostly public, private tutoring could simply reflect parents' willingness to pay for additional education for their children. However, discussions with education personnel and parents indicated that tutoring is increasingly substituting for, rather than complementing, education received in the public sector. In some instances in the study sites, students received tutoring from their regular classroom teachers, or from close colleagues of their teachers. A school principal noted that teachers have incentives to encourage tutoring, but recognize its quasi-legality: "Those with money can afford private tutoring. But the teachers are afraid of getting in trouble. Two hours of private tutoring are about 50,000 lei. I have suspected that one of my teachers did not pass the children in the exams in order to force them to buy private tutoring (in 1996) - I know that at the level of high 82 schools this is common practice. There is a mafia of teachers from different schools that recommend children to each other. " Private tutoring was noted to be more prevalent in wealthier urban sites, than rural areas. However there were exceptions to this, a well-to-do family in Zapodeni provided their children with tutoring with teachers from Vaslui, the closest city. Although private tutoring for money is common, discussions with education providers and parents also noted that many teachers spend extra hours, beyond the regular school schedule with their students. Particularly in urban areas teachers provide additional lessons, and/or organize extra-curricular activities. In order to cope with budget crises, schools turn to parents, teachers and the community at large to solicit funds. Contributions to schools are made in cash, for example through fees for participation in parent-teacher associations, or payments for specific needs (e.g. for heating), or in-kind through volunteer labor. The case studies found that both parents and teachers frequently contribute to schools by painting, chopping wood for fuel, or performing other maintenance tasks. Social assistance. Administrative charges and opportunity costs of receiving social assistance, particularly the means-tested social assistance benefit, can exceed the actual amount of the benefit. The case studies illustrated that transportation and administrative costs can reduce the value of benefits and deter eligible beneficiaries from applying. Transportation costs are especially high for households in remote areas which need to travel to the local council to apply for benefits. Application charges, such as documentation certifying income, can also increase the costs. A local council official in Nadrag noted: "A file for a person costs: 34, 000 lei transportation to Lugoj, certificate of income, 5,000 lei, copies - 20,000 lei. Ifpeople have to get the difference of about 10, 000 or 11, 000 they give up. The benefit for one person is about 160, 000 lei. " Similar difficulties were reported for applicants of school scholarships for low income students. Households are required to present documentation and approvals from different offices in the local administrations and the schools every year to become eligible. Beneficiaries in some of the poorer areas: Babadag, Sarichioi, Vaslui and Zapodeni, reported that even when they had succeeded in securing the necessary certifications, they did not receive benefits because the local council was unable to cover the costs. In Babadag and Sarichioi, the local councils would not certify households for eligibility if they had any outstanding financial obligations to the local government. Health. Evidence from household data indicates that more than 50% of families incurred health related out-of-pocket expenses (Table 5.8). Out-of-pocket expenses have increased over time - possibly following the structural changes occurred in the health sector - and have represented an increasing share of monthly expenditures for the poorest households. Although, as for education, the data does not allow us to distinguish the nature of such expenses, anecdotal evidence as well as interviews conducted in the field suggest that informal payments in health are common in Romania. For example, a study by the Public Opinion Barometer found that 49 percent of adult Romanians have made informal payments for hospital care. 83 Table 5.8 Out-of-Pocket Health Expenditures, (1994 & 98) % of households reporting payments % of monthly expenditure per capita (for acute care) 1994 1998 1994 1998 Quintile' 1 40.7 40.0 0.16 0.48 2 49.9 49.5 0.17 0.53 3 55.0 55.2 0.15 0.61 4 56.6 56.5 0.18 0.65 5 58.0 61.0 0.17 0.65 Total 53.2 54.2 0.17 0.61 Note: (1) Indicates household expenditure quintile. Source: 1994 and 1998 Romania LSMS. In the fieldwork, the issue of informal payments came up frequently in the discussions with beneficiaries of health care. Physicians from hospital do not even look at you if you don't pay. Most of them make a fortune from surgery. Woman, Potari. Ifyou don 't give to the doctors they don 't even notice you. They don 't ask, but ifyou don 't give them they do not care for your children. I gave cheese, eggs, cream. Woman, Macresti. When I went to the hospital in Vaslui it was 100,000 lei for one day. I have stayed there for a week and they didn 't do anything to improve my condition. Then I went to Iasi and I stayed there for 3 weeks and gave 100,000 bribe. You have to give money to the doctors too. Man, Macresti. The prevalence of informal payments has important consequences for the efficiency of medical services, as to doctors evade the tax system. Official revenues from taxation are spent on salaries and modernization of services, the value of the payments is transferred only to the doctors. Moreover, to the extent that informal payments are a pre-requisite for obtaining a service, this practice has serious implications for the equity of the health system as well (Lewis 2000). 5.4. Participation and Accountability In order for public services to be responsive to the needs and preferences of the communities they serve, local governments and service providers require sufficient information about the needs of the community, as well as feedback about the effectiveness of the services themselves. As a result, effective communication channels and information systems are important for service delivery, as is a sufficiently informed and active civil society. Both are still under development in Romania. While some mechanisms for participation exist at the level of local governments and in education, in particular, active involvement of citizens remains limited. The status of civil society is related to many factors, including public trust in institutions. Trust indicates the extent to which people believe institutions are reliable and accountable to their interests. Recent research suggests that the level of trust among society is notably lower than in other transition countries in the region (Romania Academy of Sciences, et al., 2000). In 84 particular, measures of public trust in institutions, as well as interpersonal trust, were lower than other countries in the region including Bulgaria, Russia and Ukraine. A Freedom House survey of public perceptions of the accountability of public institutions indicated low confidence in central govermment institutions, and slightly higher confidence in local govermments and social services including health and education (Table 5.9). Increasing popular confidence in public institutions will take time, but can be facilitated through greater public participation in local affairs and service delivery, as well as greater openness and transparency. Table 5.9: Trust in Public Institutions, 2000 (% who partially or fully agree that: "The following institutions work in the public interest.") Romania Bulgaria Slovakia Government 13.1 23.2 27.7 Parliament 8.4 11.6 20.9 President 20.7 67.2 64.8 Courts 24.0 13.2 28.3 Local Govermment 39.3 23.2 51.0 Health System 47.5 26.2 34.5 Schools 60.9 52.3 51.2 Police 44.2 39.2 36.0 Source: Freedom House Survey, cited in Romanian Academy of Sciences, et al., 2000. Based on random samples of adults in each country (Romania n=1237; Bulgaria n=1 161; Slovakia n=101 1). Local Governments. All of the judet councils included in the study hold open council meetings (Box 5.2). Despite these provisions, citizen participation is low. Those attending council meetings tend to be representatives of organizations funded from the council budget, or representatives of local councils. At the level of local councils, the link between the citizens and the elected representatives, mayor and local councilors, is based on existing legal provisions76 which provide, among other measures, that the mayor must present a report on the status of the locality at least once a year. The Law also provides that in certain cases, the mayor can recommend to the local council that a referendum be held to consult the population directly. However, a second law providing for the procedures for such referendums has not yet been voted by the Parliament. Box 5.2: Public Participation in Local Government In the majority of cases, citizen participation in local council meetings is sporadic. Babadag reported the highest level of citizen participation in local council meetings, with 10-15 citizens attending on average. All judet and local councils in the study had mechanisms for consultations between the mayor and local administration. These ranged from occasional meetings with individuals and interest groups, to radio interviews (Lugoj), and classes on topics of local interest (Tirgu Seciuesc).Tirgu Seciuesc also broadcasts local council meetings live on television. In Vaslui, topics raised in local meetings included school financing, issues faced by apartment owner associations and local public works. In Zapodeni, concems raised from agricultural issues, to organization of a sport club and local taxes and fees. Sources: Interviews with local administration officials and citizens. 76 Law no 69/91 - art. 25 85 Education. Community participation in education has been limited. As noted above, participation of parents and others in education was most frequently in response to severe fiscal conditions. The beneficiary assessment found contrasting types of parental involvement in communes and municipalities. In rural areas, parents were more likely contribute to schools through in-kind labor such as cleaning and renovation. In urban areas, including Vaslui, Babadag, Lugoj and Lupeni, parents were more likely to make monetary contributions through parent teacher organizations. A principal from Vorta noted: "Parents helped us. They painted the school, they cut wood. We asked for help not for money or materials. When we needed to repair the roof, parents helped us. It is a tradition for parents to help the school. " Feedback from parents on the content and process of education is also becoming more frequent. In schools in Vaslui and Lugoj parents' committees meet monthly with school personnel to address the concerns and questions of parents. In schools in Babadag and Vaslui the school administration actively seeks feedback from parents. The principal from Babadag noted: "We had a questionnaire. Of course it was not a very professional evaluation but we tried to see what parents think, what are their complaints...Mainly they complained about the delay in scholarship payments, about the general schedule, there were minor complaints and some of them cannot be solved very easily by the administration. " Some of the interviews with parents suggested that interactions between parents and teachers had increased in the transition period and that education was becoming more responsive. Parents indicated that teachers were involving them in exams and disciplinary actions. Teachers mentioned that they had more difficulty engaging parents in rural areas, and attributed this to higher opportunity costs of participating in schools for poorer households. Parents in poorer areas, such as Vorta and Zapodeni were mainly concerned with the payment of school scholarships. This was also the case in towns, as described by a teacher in Lupeni: "Parents are not concerned about the school, only about scholarships. It is hard to work together with them. We try to organize parents committees, parents meetings, but only the parents of good children are coming. Parents from poor families do not come. " Parents in the more developed areas such as Bucharest, expressed more complex expectations for their children's education, including requests for civic education, sex education and counseling services. 5.5. NGO Involvement in Social Services77 Delivery of social services in Romania is becoming increasingly diversified with the involvement of non-governmental organizations (NGOs) in the provision and financing of social services. Local governments have an important role to play in working with non-governmental organizations, as well as other entities within civil society such as churches, community organizations, donors and other partners in the delivery of services. This role can have a number 77 This section draws from Dankova (2000) and Petrescu (1998). 86 of dimensions ranging from attracting resources, including donor support and project resources, to the locality, to contracting, monitoring and evaluating services, and referring beneficiaries to the appropriate service providers. In 1999 there were 27,000 registered NGOs in Romania, however a recent report on NGO activity noted that only 2,000 to 3,000 are estimated to be active (Dankova, et al., 2000). Over forty percent of registered NGOs are focussed on the social sectors. The 1999 Directory of Foundations and Associations indicated that 18 percent of NGOs had activities in the social sectors, 16 percent of these were classified as education, teaching and research and 8 percent in health. The level of NGO activity varies notably across regions, as illustrated by data for the judets in this study. NGOs (of all types) have a greater concentration in Covasna, Timis and Bucharest, while activity is notably lower in Vaslui and Hunedoara. Table 5.10: NGOs in the Study Judets (per 100,000 pop.), April 2000 Covasna 46.3 Hunedoara 19.5 Timis 36.1 Tulcea 15.3 Vaslui 5.9 Bucharest 52.7 Source: Foundation for the Development of Civil Society, cited in Mungiu-Pippidi and Mandruta, 2000 The implications of increasing involvement of NGOs in service provision are complex and influence the level, quality and effectiveness of social service provision at the local level in many ways. NGOs have the potential to complement or substitute for public services. Especially in times of severe budget constraints, the presence of NGOs might provide an easy way for local and central governments to off-load some of their responsibilities and effectively give rise to a "crowding out" of public services. NGO provision of services also increases the risks of inequitable service provision across local governments if NGO or donor activity, is more prevalent in certain areas than others, if coverage of services is limited, or if quality and effectiveness are not monitored. Legislative Framework. There have been important recent developments related to the legal status and financing of NGOs in Romania. Most importantly, in May 2000, following the completion of fieldwork for this report, a new NGO law was passed which eases registration processes and clarifies the relationship between local governments and NGOs and makes provisions for information sharing and the provision of resources to NGOs, such as office space. Legislation regarding the financing of NGOs has also encouraged activity and involvement in service provision. The Law on Local Public Finance makes provisions for collaboration between local governments and NGOs by laying out the framework for establishing partnerships with NGOs and transferring resources. Taxation legislation exempts NGOs which provide social services from certain taxes including the value-added tax and the profit tax. A law on sponsorship passed in 1994 and amended in 1998 allows business to make tax deductible contributions to NGOs, although critics note that the procedures for requesting the deduction are onerous and discourage businesses from contributing. A number of other laws have specific provisions for the involvement of NGOs in social services. Among these, Law 208/1997 which local authorities to contract with NGOs to run soup 87 kitchens and Law 34/1998 which provides central government support for NGOs which run social welfare institutions. Finally, the Law on Social Insurance allows insurance funds to contract NGOs to provide health and social assistance services. While the legislative framework for NGOs appears quite advanced, further analysis is needed to assess its effectiveness and impact. Interviews with NGO staff and officials from the central as well as local governments conducted by a team from the Mott Foundation indicated that NGOs and local governments were not yet sufficiently informed about the new legislation to take full advantage of its provisions. In addition, there are areas where additional legislation will be needed, including establishment of quality standards for service provision by NGOs. Relationships with Local Governments. Collaboration between local governments and NGOs is increasing. NGOs were active in all of the judets included in this study and in many cases were working directly with local governments. The relationship between local governments and NGOs varies significantly across Romania. A 1998 study by the MOLSP and the Civil Society Foundation found that local officials and NGO staff frequently collaborated in a wide range of areas including financing activities, sharing information, providing legal advice, distributing aid, screening beneficiaries, training personnel and preparing joint projects for donors. The study by the Ministry of Labor and interviews conducted for the Mott Foundation study pointed to similar issues constraining the effectiveness of partnerships between local governments and NGOs. Both studies indicated that communication between NGOs and local governments required attention and that both parties lacked information on legal arrangements and provisions for joint activities. The lack of funds at the local level also affected the level of NGO involvement, as local governments are unable to contract for services and locating support for recurrent costs is a challenge. Finally, local government capacity to work effectively with NGOs is low. Pilot training has been undertaken in some judets including topics such as project management, communications and public relations. There are increasing examples of successful collaboration between local governments and NGOs (see Box 5.3 in the final chapter of this report). NGO Involvement. NGOs have been most active in the provision of social assistance services, particularly in child protection and in services for the disabled. International donors and NGOs have been most active, while the relative involvement of local NGOs is low. Local NGOs identified during the fieldwork were generally small with limited fiscal and administrative capacity. NGO involvement in child protection has led to a significant increase in available resources and have allowed Child Protection Departments to experiment with prevention activities and alternatives to institutionalization. The decentralized nature of the CPDs allows for flexibility and innovation in partnerships at the judet level. NGOs are also active in education, where they address a diverse set of issues, such as learning programs for Roma children and fostering community involvement in children's education. 88 Box 5.3: NGO Involvement in Child Protection Services NGO and international donors are involved in child protection activities in all of the judets in the study. In some areas involvement focused on activities which prevent the entry of children into institutions in the first place. For example, in Timis activities include training social workers to assist children at risk, and the establishment of a center to work with mothers and children together. The Child Protection Directorate in Tulcea is receiving support through a World Bank Child Welfare Reform project for the organization of foster care and counseling centers. In Hunedoara funds from the same source are used to develop counseling services for at risk families. Other activities aim to provide children with an alternative to institutionalization. In Covasna a project supported by a Swiss foundation aims to build 12 family homes for families who will provide foster care to children at risk and possible adoption. In Hunedoara the Soros Foundation is supporting fostering centers to reintegrate young people leaving institutions. Local churches and community organizations have become involved in initiatives to address the needs of poor households. The level of involvement varies across areas and can depend on specific social and cultural characteristics. For example, in Sarachoi, the commune in Tulcea, there is a traditional Russian orthodox community, the Lippovans. Associated with the Lippovan orthodox church are hermitages, traditional institutions which provide shelter for elderly with no other source of support. Similarly, a church in Vaslui has established a small soup kitchen with support from its congregation. NGOs have the potential to strengthen local government provision of social services by establishing new service initiatives. In Vaslui the Elena Cuza Foundation established a day care center for the elderly which provided for a broad range of needs including medical services, physical therapy, home care and a soup kitchen. After two years, the center was passed on to the local administration. The center continues to function and is popular among beneficiaries. Community-based social services. In the areas of social assistance, there has been some development of community-based services involving NGOs, churches and other civil society organizations. Donor activity has been particularly intense in this area in recent years. The example of the Romania Social Development Fund (RSDF) provides useful lessons and examples for further work in this area (Box 5.4). 5.6. Responsiveness of Local Social Services In principle, decentralization should lead to an improvement in local governments' ability to tailor service delivery to local communities' preferences. However, the many changes in local finances that occurred in the last years coupled with serious budget constraints have often restricted local governments' and service providers' actions to implementing coping strategies, and have not improved the responsiveness of services to citizens. The lack of a 'service mentality' among civil servants also limits the responsiveness of services. However, in a number of ways, the potential for increasing responsiveness has increased. 89 Box 5.4: Community Based Services and the RSDF The Romania Social Development Fund (RSDF) was established in 1998 by the Romanian government in partnership with the World Bank, to support community driven development initiatives in disadvantaged areas of the country. A central component of the RSDF is support for community-based services, which provide useful examples of community participation in service delivery and of the significant potential for collaboration between NGOs and local governments. Since the inception of the RSDF, 39 such projects have been supported, representing a diverse range of activities. Projects include home care for the elderly, shelters for homeless adults, and housing and counseling services to assist teenagers who have left state institutions to reintegrate into society. In nearly all cases, projects are run by NGOs with support from local govemments. The contributions of local governments are sometimes in cash, but most frequently are in-kind, through the provision of staff time, consumables or facilities. In Galati, a local council pays the salaries for social workers in a home care program for the elderly. In another project in Braila, the local council contributed a building for a counseling center. The greatest challenges for these projects is to identify on-going sources of support. Many projects are not selected for RSDF funding because they are not able to demonstrate sustainability. In this regard, the ability of local governments to identify and allocate resources to programs and services which reflect the needs of the community is extremely important. The judets offices of the MMSS serve as another important source of support for programs. Based upon a law on NGOs passed in 1998, the MMSS is able to support consumables and equipment for community services. In addition to fiscal sustainability, the success of the projects and the collaboration with local governments also depends on local participation and support, and the ability of projects to demonstrate visible results. The personal initiative and involvement of local government officials can greatly influence the success and sustainability of projects. In Braila, on-going support from a mayor has led to the expansion of a number of initiatives. Experience from this first set of projects has demonstrated the capacity for local governments to work with NGOs in service provision. The lessons from these projects are valuable examples for other local governments which are unaware of the potential opportunities for increasing civil society involvement in services. The RSDF staff is considering ways in which to disseminate the experience across the countries, including study visits for local officials to see the results of projects. Education. A number of policy initiatives have made schools better equipped to respond to local preferences. For example, schools can now develop school-based curricula on the basis of students' demand. The offering of optional courses (often computer courses) is financed by the schools' own resources. As it is the case with most issues with decentralization, these financing arrangements provide the schools with the correct incentives - to be able to run a new optional course, schools must be able to secure funds - but interferes with equity - as schools in poor areas may lack the capacity and the resources to offer new courses that might complement the more traditional teaching covered under the common core curriculum. The institution of "representative schools" offers a different opportunity to schools to be more in tune with communities, as schools can independently hire the teachers that they believe fit best the school's needs. Nonetheless, as this policy initiative is very new, feedback from 90 students or parents is not available to indicate whether school-based hiring has effectively improved the quality of teaching. 78 Social assistance. In social assistance, greater involvement of local councils in the delivery of services has the potential to improve the targeting of benefits and programs to local needs. However, current financing arrangements weaken incentives for local officials to provide benefits. In the context of budget constraints, local officials face pressures to allocate resources to other areas. Social assistance is not a priority for local governments. The pending reform of social assistance, included in the draft framework Law on Social Assistance, is designed to address the financing constraint, by centralizing financing in the state budget and allocating it to local councils through transfers earmarked for social assistance. In principle, this reform should take advantage of the greater ability of local governments in targeting benefits, while ensuring adequate resources. However, the program will require careful monitoring to make sure that targeting is effective, as local governments will have new incentives to increase coverage. There is clearly greater scope for community and local government involvement in social services, which has the potential to enhance the responsiveness, allocative efficiency and accountability of public services. 5.7 Conclusions Romania's new government has identified increasing the role of local governments in service delivery as a priority and has rightly linked the process to the overall reform of public administration and intergovernmental finance in the country. Experience to date with decentralization of social services in Romania provides some important policy lessons which have already begun to shape future policies, and indicate potential benefits and pitfalls of decentralization. In some areas, decentralization has been a failure. Fiscal decentralization of poverty alleviation benefits has undermined their effectiveness and led to an erosion of the social safety net. Financing of the means tested social assistance benefits out of local government budgets has meant that many of the poorest local councils, which have the greatest need, have been unable to pay benefits. As a result, there is currently no national poverty alleviation program for the poorest households in Romania. In order to address this, the government is undertaking measures to recentralize financing and aspects of its administration. The MLSS has prepared the MIG Law which proposes to centralize financing for social assistance cash benefits by financing the program through earmarked cash transfers to local councils. Targeting of benefits will remain the responsibility of local governments. In other areas, decentralization appears to have been more effective. Decentralization of child protection to the judets level has diversified arrangements for financing and provision of child protection services by increasing opportunities for judets to work in partnership with NGOs, donors and other institutions within civil society. Other examples of collaboration between local governments and NGOs, such as those implemented under the Romania Social 78 Some inspectors are skeptical that school-based hiring will be superior to centrally run hiring unless schools base their hiring decision on an in-depth assessment of the candidate pedagogical abilities (as opposed to assessing the candidate's accumulated knowledge though a written exam). 91 Development Fund, also illustrate the potential for increased responsiveness, through greater community involvement in services. Education is the policy area where there is perhaps greatest potential for further decentralization. The new government has already transferred responsibility for financing personnel to local governments. In addition, the experiment with 'representative schools' indicates that additional responsibilities including managing personnel and curriculum may be decentralized to the local level. Lessons from the experience of local government involvement in education to date are essential for the success of future policy reforms. A careful evaluation of capacity at the school level and of the benefits that could accrue in terms of increased parents' participation and ownership could inform the decision of whether Romania should move towards more complete autonomy. 92 References Andrews, S. and Ringold, D. 1999. "Safety Nets in Transition Economies: Toward a Reform Strategy," Social Protection Discussion Paper No. 9914. The World Bank: Washington, D.C. Azfar, O., Kahkonen, S., Lanyi, A., Meagher, P. and Rutherford, D. 1999. "Decentralization, Governance and Public Services, The Impact of Institutional Arrangements: A Review of the Literature," IRIS Center, University of Maryland, College Park. Azfar, O., Gurgur, T. Kahkonen, S., Lanyi, A., Meagher, P. 2000. "Decentralization and Governance: An empirical Investigation of Public Service Delivery in the Philippines," IRIS Center, University of Maryland. Belcher, M. et al., 1997. "Romania: Municipal Creditworthiness and Local Government Decentralization," Urban Institute Report. Bird, Ebel, Wallich, 1995. Decentralization of the Socialist State, Washington, D.C.: The World Bank. Caluseru, G. 1999. "Romania: Municipal Development and Finance Policy Assessment Paper," The World Bank: Bucharest, Romania. Draft. Conway, F., Pop, L, and Zamfirescu, V., 2000. "Romania: Analysis of Local Financing for Child Protection Services," Urban Institute, prepared for USAID. Cremer, Jacques, Antonio Estache, and Paul Seabright, 1994. "The Decentralization of Public Services," World Bank Policy Research Working Paper no. 1345.' Dakova, V., Dreossi, B., Hyatt, J., Socolovschi, A. 2000. "Review of the Romanian NGO Sector: Strengthening Donor Strategies," Mott Foundation and Charity Know How, Bucharest, Romania. Draft. Dhanji, F. 2000. "Discussion Piece: Public Administration Reform, Judicial Reform and Anti- Corruption," The World Bank, Bucharest, Romania. Draft. Tesliuc, E.D., Pop, L., and Tesliuc, C. M., 2000, "Romania: Social Protection and the Poor." Forthcoming. The World Bank: Bucharest, Romania. Duthilleul, Y., 2000. "Education and Economic Growth: Next Phase of Education Reforms in Romania," Manuscript, The World Bank. EDI, 1998. "Intergovernmental Roles in the Delivery of Education Services in Central and Eastern Europe," Background papers. Fisman, R. and Gatti, R., 2000. "Decentralization and Corruption: Evidence Across Countries" forthcoming, Journal of Public Economics. Fiszbein, A., 2001. "Decentralizing Education in Transition Societies: Case Studies from Central and Eastern Europe." The World Bank: Washington, D.C. Gatti, R., 2000. "The Economics of Decentralization: a Review of the Literature and the Case of Education Provision", Working Paper no.109, Italian Society of Public Economics. Jakob, Felix A., 1995. "Decentralization and local government finance in Romania," in Bird, Ebel, Wallich, Decentralization of the Socialist State, Washington, D.C.: The World Bank. 93 Government of Romania, 2000. "The Governing Programme: 2001-2004." Bucharest, December, 28, 2000. Govemrnment of Romania and the European Commission, 1997. "Green Paper: Regional Development Policy in Romania." Kaufman, Daniel, Aart Kraay, and Pablo Zoido-Lobaton, 1999, "Governance Matters," World Bank Policy Research Working Paper, no. 2196. Ministry of Labor and Social Protection, 1997. "Evaluation of the Income Support Benefit: Executive Summary." Mimeo. Petrescu, D. 1998. "Report: NGO Stock-Taking in Romania," The World Bank, Bucharest, Romania. Prud'homme, Remy, 1995, "The Dangers of Decentralization," The World Bank Observer, vol.10, no.2. Romanik, C., Coman, P., Daianu, D. and Garzon, G. 1999, "Romania, Winners and Losers: the Impact of Reform of Intergovernmental Transfers, The Urban Institute. Rondinelli, Dennis, 1981, "Government Decentralization in Comparative Perspective. Theory and Practice in Developing Countries," International Review of Administrative Science no.47. Shah, Anwar, 1998. "Balance, Accountability, and Responsiveness: Lessons about Decentralization," Policy Research Working Paper no. 2021, The World Bank. Tanzi, Vito, 1996 "Fiscal Federalism and Efficiency: A Review of Some Efficiency and Macroeconomic Aspects," in M. Bruno and B. Pleskovic, eds., Annual World Bank Conference on Development Economics 1995, World Bank, Washington, D.C. Tesliuc, C. and Pop L., 2000. "Poverty, Inequality and Social Protection," Manuscript. Tiebout, Charles, 1956. "A Pure Theory of Local Expenditures," Journal of Political Economy, vol. 64, pp. 416-24. Tobis, D. 2000. "Moving from Residential Institutions to Community-Based Social Services in Central and Eastern Europe and the Former Soviet Union," The World Bank: Washington, D.C. Voicu, B. and Begu. L. 1999. "An Analysis of the Structure, Costs and Efficiency of Romania's Pre-university Education System," The World Bank. UNICEF, 2000. "Young People in Changing Societies," Regional Monitoring Reports, o. 7. Florence: UNICEF Innocenti Research Centre. UNICEF, 1997. "Children at Risk in Central and Eastern Europe: Perils and Promises," Regional Monitoring Report, No. 4. Florence: UNICEF Innocenti Research Centre. Wetzel, D., 2001. "Decentralization in the Transition Economies: Challenges and the Road Ahead." Forthcoming. World Bank, 2001a. "Country Assistance Strategy for Romania, June 19, 2001" World Bank, 2001b. "Romania: Country Economic Memorandum." Forthcoming. 94 World Bank, 2000. "Balancing Protection and Opportunity: A Strategy for Social Protection in Transition Economies," The World Bank: Washington, D.C. World Bank. 2000b. "Hidden Challenges to Education Systems in Transition Economies," The World Bank: Washington D.C. World Bank, 1999. "Romania: Health Sector Support Strategy," Report No. 184 10-RO. World Bank, and National Commission for Statistics, 1999. From Rural Poverty to Rural Development, Coordinated by Constantin Chirca and Emil Tesliuc. World Bank, 1992. "Romania: Decentralization and Local Govermment Reform," Working Paper. Zymelman, M. and DeStefano J., 1989, Primary School Teachers 'Salaries in Sub-Saharan Africa, the World Bank. Legislation Law on Local Public Finance, September 14, 1998 version. Law on Patrimony, 1998. 95