25132 wVr.44L rIANK 'F F-tAN DECENTRALIZATION IN THE TRANSITION ECONOMIES: CHALLENGES AND THE ROAD AHEAD June 14, 2001 Poverty Reduction and Economic Management Unit Europe and Central Asia The World Bank TABLE OF CONTENTS EXECUTIVE SUMMARY ............................................................. IV WHY DECENTRALIZATION? ..............................................................1 THE DEGREE OF DECENTRALIZATION TO DATE ...............................................3 CHALLENGE 1: ADAPTING TO CHANGING GOVERNMENT STRUCTURES ... 7 Government Structures ..............................................................7 Country ...............................................................................8 The Role of the Regional Level ..............................................................9 The Appropriate Size of Local Government for Service Delivery .............................9 CHALLENGE 2: STRENGTHENING ACCOUNTABILITY ................................... 10 Accountability ............................................................. 10 Mobility ............................................................. 13 Information ............................................................. 13 Rule of law ............................................................. 14 State capture ............................................................. 15 CHALLENGE 3: MANAGING SHARED RESPONSIBILITIES ............................. 17 Expenditure Assignments ............................................................. 17 Aligning responsibility for outcomes with decision-making authority ..................... 19 Minimizing parallel decision making authorities ...................................................... 21 Strengthening checks and balances ............................................................. 21 Delineating boundaries for the public sector ............................................................. 22 CHALLENGE 4: AUGMENTING REVENUE AUTONOMY ................................... 23 Revenue assignments ............................................................. 23 User Charges ............................................................. 24 Own revenues ............................................................. 24 Shared Taxes ............................................................. 27 Strengthening the link between revenues and services provided . ............................. 28 CHALLENGE 5: CLARIFYING TRANSFERS .......................................................... 28 Design of transfer systems ............................................................. 29 Transfer systems in transition countries ........................... .................... 3 1 Further strengthening is needed ........................... 35 CHALLENGE 6: MANAGING AND MONITORING LOCAL BORROWING ....... 36 Methods of Local Borrowing .......................................................... 36 Regulatory frameworks for subnational borrowing .................................................. 38 CHALLENGE 7: BUILDING CAPACITY .......................................................... 47 Assessing Local Capacity .......................................................... 47 Building capacity .......................................................... 47 Community Driven Development as a Tool for Building Capacity ......................... 49 THE ROAD AHEAD .......................................................... 52 Annex Table 2: Basic Characteristics .......................................................... 61 Expendittre Functions ............................................................................. 61 TABLES Table 1: The Structure of Subnational Government in Transition Countries ............... 8 Table 2: Elections for Local Governments in Selected Transition Countries .............. 11 Table 3: Taxes of Subnational Government by Type of Tax Autonomy .................... 25 Table 4: Source of Financing in Transition Economies ....................................... 26 Table 5: Types of Transfers in Selected Transition Countries ................................ 32 Table 6: Regulatory Framework for Subnational Borrowing in Transition Countries .... 41 FIGURES Figure la: Local Expenditure as a Share of GDP (percent). 5 Figure lb: Local Expenditure as a Share of Total Expenditure (percent). 5 Figure 2: Local Expenditure as a Share of GDP and Press Freedom .......................... 14 Figure 3: Local Expenditure as a Share of GDP and Rule of Law ............................. ] 5 Figure 4: Local Expenditure as a Share of GDP and State Capture ........................... 16 Figure 5: Share of Subnational Expenditure in Total Government Expenditure ............. 19 Figure 6: Shared Taxes and Grants as a Share of Total Revenue (percent) ................... 29 Figure 7: Elements of a Fiscal Decentralization Strategy ........................................ 54 BOXES Box 1: Fiscal Federalism in Bosnia & Herzegovina ................................................. 6 Box 2: Decentralization of Education in Central and Eastern Europe .................... 20 Box 3: Community Driven Development and Capacity in Central Asia .......................... 51 Box 4: Decentralization and Poverty Reduction ............................................. 52 ANNEX TABLES Annex Table 1: Sources of Data . 59 Annex Table 2: Basic Characteristics .60 Annex Table 3: Expenditure Assignments for Selected Transition Countries .61 Annex Table 4: Subnational Shares of Total Government Spending by Functional Class.62 This study was prepared by Deborah Wetzel with the able research assistance of Lavanchawee Sujarittanonta. Jonathan Dunn (consultant) contnbuted significant inputs to an earlier version of the paper. Alison Panton assisted in formatting and preparation of the document. The errors remain the author's sole responsibility. EXECUTIVE SUMMARY Why Decentralization? ii. This raises a fundamental question that must be addressed for any i. Across the world countries are country in any region, but which is increasingly focusing on decentralization particularly important in the transition of decision-making authority to lower economies: Why decentralize? While in levels of government. The transition any given situation the reasons will be countries of Eastern Europe and the distinct, in transition countries there Former Soviet Union have not been appear to be three key imperatives excluded from this phenomenon. Indeed, driving decentralization: a political the very transition itself is a process of imperative, an economic imperative and moving away from a centrally a service delivery imperative. controlled, command-driven economy towards more politically, economically iii. The objective of this paper is to and administratively decentralized and highlight the key challenges to effective market based structures. No doubt that decentralization that have become some countries have gone further than apparent over the past decade; and to set others, but the nature of this type of out ideas about how to address these transition raises special issues for issues more effectively in the future. It decentralization in all of these countries. starts by giving a perspective on the Early steps in this process were degree of decentralization across the documented in Decentralization of the transition economies. Subsequent Socialist State: Intergovernmental sections then consider a number of Finance in Transition Countries.' critical issues and challenges. Finally the Several years on, we find that progress paper closes with suggestions has varied and that in most countries, the concerning policy steps and sequencing central government has delegated the that may help address these issues in the administration of many activities to future. lower levels of government. Only in a few cases, however, has real decision- The Degree ofDecentralization making authority been devolved. Unsurprisingly, it has taken longer than iv. The notion underlying theoretical expected for countries to work out new discussions of decentralization is that structures of government. While service local governments, which are closer to responsibilities have been delegated or citizens, are likely to have better devolved, central governments are often information on citizens needs and/or reluctant (and often for valid reasons) to demands and therefore will be able to give up control over revenues. Perhaps provide services more efficiently than most important, effective the central government. Underpinning decentralization presupposes a broad set this general idea are a range of of institutional structures that did not assumptions concernig the exist under the old system and these take accountability of local governments, the time to develop. degree of information flow, and the mobility of individuals. In the transition countries, the extent to which these See Bird, Ebel and Wallich. 1995 conditions do hold varies considerably iv and has an influence on the degree and monitoring and regulation of local effectiveness of decentralization efforts. govemment activities, especially local government borrowing. Finally it v. Even measuring decentralization requires capacity at both the central and poses difficulties. The most common especially the local level to manage measures of decentralization -- local public revenue and expenditure, govemment expenditure as a share of effectively implement policies and GDP and local expenditures as a share of deliver services. Each of these areas total general government expenditure -- raises particular challenges to effective indicate that these two measures vary decentralization in the transition considerably across the transition countries and sheds light on steps that countries, and even among sub-regions. may be desirable in the future. Russia, Ukraine, Kazakhstan, Hungary, Latvia and Estonia all have local Challenge 1: expenditure over 10 percent of GDP and Adapting to Changing Government a relatively high share of local Structures expenditure in total general government expenditure. Russia's shares are similar vii. All but three of the transition to those of other large federations countries (Russian Federation, Bosnia & (Canada and the US), though they are Herzegovina and the Federal Republic of lower than a number of northern Yugoslavia) are formally unitary. In European countries (Sweden, Denmark, each country there exist multiple tiers of the Netherlands). At the other end of the govemment, and most countries have spectrum, local expenditures are less two or three subnational levels. There then 5 percent of GDP in countries such are a number of issues that most as Armenia, Romania, and Slovakia, countries in transition are confronting, suggesting a less extensive role for lower with respect to govermment structures, level govemnments. two of which stand out. The first is the appropriate role for the regional level. vi. These standard measures, The second is the appropriate size of a however, fail to take into account the local government for effective service different degrees of decentralization and delivery. the true decision-making authority of localities over expenditures. They also viii. In recent years a number of do not take into account the degree to countries, especially (but not which the assumptions underlying the exclusively) in the EU Accession efficiency arguments for decentralization countries, have sought to reorganize and are in place. To fully assess the degree consolidate their intermediate levels of of decentralization it is necessary to govermment. While the role of this consider fiscal decentralization as a regional level is relatively clear in the comprehensive system. Such a system federal countries, it will continue to includes the basic structure of evolve in the unitary countries. For government and institutions that support implementation of good policy it is accountability. It includes a framework critical that as administrative and for expenditure and revenue territorial reform evolves, the specific assignments, as well as a system of functions and authorities, as well as the effective transfers. An important sources of finance of these intermediate component of such a system is also the levels be clearly specified. vii ix. The size of the lowest level of xi. Two points stand out in government varies significantly in considering the degree to which transition countries. Countries such as subnational governments are politically Czech Republic, Hungary and Latvia accountable. First, in most countries of that have very small local governments the region the intermediate levels of face the potential for service delivery government remain ones in which that is fragmented and inefficient. Small accountability is to higher levels of local governments often lack the government, rather than to the electorate. capacity to manage all the functions This suggests that activities undertaken assigned to them. At the same time at these levels remain deconcentrated many, smaller governments may allow and perhaps delegated, as opposed to for greater participation and truly being decentralized. This raises the accountability and may be politically issue of how to effectively meld desirable. The challenge for transition deconcentrated functions and countnes is to find stable structures of responsibilities with those that may be government that are politically devolved to lower levels. It also raises acceptable and at the same time help to the important question of how the flow promote efficient and effective local of funds through deconcentrated governments. At both the regional and intermediate levels (with accountability local level, administrative reform is to the center) affects the real autonomy likely to continue, and assignment of and decision-making power of the functions and finances will need to adapt lowest levels of government (with over time. Of critical importance is that accountability to the electorate). measures to reform government structures clearly take into account the xii. A second issue that arises relates appropriate scale for key services and to the degree that elections at the lowest that roles and responsibilities are set out level are free and fare, or are dominated in a clear legal framework. by national political parties, or other specific interests. Local bosses that gain Challenge 2: substantial influence exist in all Strengthening Accountability countries, including countries in transition. The concern exists as to x. The notion that decentralization whether the possibility of corruption can lead to more efficient delivery of may be even greater at local levels of govermnent services rests on the government than at national levels. possibility that local governments can be While much of the literature suggests held accountable. This section considers that citizens are likely to hold local the extent to which some of the governments more accountable than institutions that support local national ones, this presupposes the government accountability are in place. existence of institutions and a political Accountability for local government can culture that supports this accountability. take place in many fashions, but During Soviet times, incentives worked elections ("voice") and freedom to move against this type of accountability, and to another jurisdiction ("exit") are this legacy remains in many countries. probably among the most discussed in the literature on decentralization. xiii. Elections, mobility, information, rule of law and the extent to which the state can be influenced by outside viii interests all represent different given areas. Typically it does not go far mechanisms that affect the degree to enough in specifying the details of each which local governments can be held level's involvement. As structures of accountable. They also give some government and the intergovernmental indication of whether the institutional system evolve, it is important to refine underpinnings that are necessary for expenditure assignments so as to better successful decentralization are in place support better service delivery of and therefore the extent to which outcomes. Ideally all stakeholders would decentralization is likely to be be brought together to contribute to these meaningful. Supporting institutions in revisions. First, overall objectives the Baltics and Czech Republic, should be agreed upon. Second, the Hungary and Poland are furthest along, continued scope for government and are weakest in Russia, Ukraine, and involvement should be evaluated. Third, the Central Asian economies. The should be a detailed evaluation of who strength of these institutions varies in actually does what both formally and Caucasus, the Balkans and the South informally and what that implies for the Eastern European countries. incentive to meet the stated objectives. Fourth should be consideration of xiv. Institutions develop slowly and realignment of responsibilities to make are not easily changed. Nevertheless a sure that responsibilities are matched fundamental challenge to implementing with both decision-making authority and effective decentralization policies in accountability. Responsibilities for many transition countries is to regulating, financing and provisions strengthen the institutions that support should be clearly established. Efforts the accountability of lower levels of should also be made to develop systems government. Some aspects of this-- such of checks and balances, including as elections -- are political and therefore alternative sources of service provision out of the domain of an institution like (both private sector and by NGOs) and the World Bank. Other areas, however, in terms of beneficiary involvement. The such as improving labor market specifics will vary by sector, but mobility, supporting information flow, deliberate efforts to manage shared increasing transparency and responsibilities more effectively, can do strengthening civil society and the a great deal to strengthen the participation of local communities may effectiveness of decentralization efforts. help to build a solid foundation for more effective decentralization. Without Challenge 4: improvements in these institutions in the Augmenting Revenue Autonomy countries in which they are weak, the desirability of decentralization and the xvi. The limited amount of revenue extent to which it is meaningful will autonomy at the local level is perhaps continue to be ambiguous. the most significant signal that countries have not really decentralized, but rather Challenge 3: have deconcentrated or delegated Managing Shared Responsibilities responsibilities. Given the legacies of central planning and weak financial and xv. Transition countries usually have administrative capacities in many legislation that indicates the levels of countries, it will take many years before government that have responsibilities in localities will be given responsibility for ix raising their own revenues. Nevertheless, Challenge 5: if governments are serious about Clarifying Transfers improving service delivery through decentralization, at some point the link xx. While there is no doubt that a between services delivered at the local great number of actions and efforts have level and the funds necessary to provide been taken to transform the system of those services will need to be made. transfers in most transition countries, Several steps can be taken to pave the clearly more needs to be done to make way for greater revenue autonomy in the sure that these systems create the correct future. incentives and have the desired economic impact. First, what is fairly xvii. First, at the national level, basic, but often disregarded, is the need countries that are interested in to use the mechanism created as the decentralization can lay the groundwork basis for determining the resources by improving the overall tax framework. transferred. Countries will go to great Many transition countries have very high effort to create formula-based systems marginal tax rates and this discourages and then carry out actions which local governments from even making completely undermine these systems. use of owIn taxes that they may have For example, Bulgaria has created a now. formula for allocating resources, but typically, towards the end of the year it xviii. Second, there is a substantial provides and extra, unplanned transfer to degree of preparatory work that needs to localities to finance unpaid bills. This be undertaken to support localities in action i) indicates that the formula itself their ability to levy taxes. For example, must be inappropriate and ii) creates an development of a market value-based incentive for all jurisdictions to stop property tax requires the establishment paying their bills and to anticipate that of a cadastre and all the other requisite additional resources may become information. In many transition available towards the end of the year. countries it takes time to bring together The crucial point is that once an the various parties that need to approach for allocating resources is coordinate and implement such agreed upon, it is critical that it be used programs. The greater the degree of and there not be other ad hoc elements preparatory work, the greater the for providing resources, that undermine likelihood of effectiveness when systems the incentives of the system. are implemented. xxi. Most countries in the region could xix. Finally, revenue autonomy can benefit from stepping back and thinking probably best be increased by starting through the total effect of the transfer simply and modestly. A surcharge on a system that is evolving. Typically, grant pre-existing tax within an upper and systems are developed in a rather ad hoc lower band can draw on already existing way over time. Conditional grants are administration and collection capacity. created to serve one purpose, a range of As administrative capacity and capital grants mechanisms exist, accountability develop, localities can equalization adds yet another take on more sophisticated approaches to component. As was discussed above revenue autonomy. typically one grant mechanism is used to address a particular purpose or objective, x but it may also have indirect impacts on Challenge 6: other systems. As systems with multiple Managing and Monitoring Subnational transfer mechanisms develop, it is quite Borrowing important to assess their overall impact and whether those remain with the xxiii. In quite a few transition countries, overall objectives of the system. there have been some very dramatic developments in the area of subnational xxii. Additional efforts are also needed borrowing in recent years. Over a very to improve the quality of data underlying short time period, borrowing in the form specific transfer mechanism, the ability of loans from commercial banks and to monitor and the transparency of the bond issue has evolved from a negligible system as a whole. In countries that are factor into an important source of less far along in the development of autonomous financing for subnational transfer mechanisms, it can be quite a government, substituting for the challenge to collect the information traditional dependence on financing needed in a consistent and coherent way. provided by the central government. Agreement on a formula or capitation Even where the volume of funds is not grants require adequate information to large, borrowing on commercial terms determine expenditure needs in a manner represents a qualitatively different type that focuses on resources required to of financing than the traditional reliance provide the desired outcomes, as on highly subsidized loans from central opposed to being based on input authorities. As such, the development of requirements. Developing an ability to commercial borrowing imposes new monitor the use of funds is an important demands on both subnational and element of conditional grant systems, national governments. Subnational otherwise the "conditions" become authorities must adapt to a more ngorous meaningless. Finally, a key element of standard of accountability associated developing formula based system is to with financing their activities on provide local governments with the commercial terms. National authorities predictability and stability that using must adapt to the role of an effective such a system can provide. A regulator, moving away from the more prerequisite for this, however, is that the traditional role of directly controlling the system be transparent. It should be finances of subnational governments. absolutely clear how and why resources are allocated to different jurisdictions. xxiv. In addition a move towards Although many countrnes have worked increasing subnational borrowing hard in developing systematic transfer requires increasingly more effective systems, these are typically less information and monitoring systems, as transparent than desirable for having the well as reporting mechanisms. Even desired impact on incentives. Often lack some of the more advanced reformers in of transparency comes in the form of the region might be hard pressed to excessively complicated formulas, but it specify total local government also may result from the difficulties borrowing. Old systems make strong localities may have in obtaining the distinctions between taking loans and formulas and information used in issuing debt and monitoring mechanisms calculating them. do not always capture different types of local borrowing. Strengthening information and monitoring systems for xi subnational debt is a critical element for contracting for services and buying improved debt management overall. equipment. Ultimately there are four fundamental functions that organizations xxv. Generally, there is a close must be able to do in order to reach their correspondence between those countries objectives: decision-making, (which that have been more or less active and includes budgeting, planning and successful in supporting the evaluation); resource mobilization and development of subnational borrowing management; communication and and the countries that have demonstrated coordination; and, conflict resolution. greater commitment and success in reshaping other aspects of xxvii. The traditional approach to intergovernmental fiscal relations. decentralization has been to build Beginning with the significant share of capacity through training and technical transition countries in which subnational assistance before transferring borrowing has not emerged as a responsibilities or revenues. This method significant institution (Belarus, was based on concerns about Moldova, Central Asia and the irresponsible spending, local corruption, Caucasus), along a continuum through to regional inequities, and service collapse those countries that have been most as well as many central governments' active in establishing a stable and reluctance to devolve authority and effective regulatory framework for inability to monitor decentralization. subnational borrowing (Hungary, Czech However, it is difficult to expect local Republic, Poland and Estonia), the governments to change behaviors and survey helps to highlight which respond differently based solely on countries are most committed to training and technical assistance. supporting the development of Localities may understand better how to autonomous subnational governments. do things to suit their citizens' needs, but if the incentive to implement effective Challenge 7: decision-making, resource mobilization, Building Capacity and coordination are not in place, there will be little reason for localities to xxvi. Measuring local capacity is pursue them. difficult and there are no ready variables to apply across the transition countries. xxviii. Building capacity therefore Some of the basic components of requires two key elements: creating the planning, implementing and sustaining incentives for localities to undertake the basic services include: analyzing and actions that support effective capacity, solving local problems; determining and giving support and guidance as they community needs; organizing local and build their skills at decision-making, national political support for programs; planning, resource mobilization, mobilizing resources for programs; coordination and the like. An raising tax revenues or collecting user appropriate framework for fees; writing specifications for the decentralization combined with support technical elements of programs; to individual local governments can help maintaining and sustaining the services; give local governments the incentive to evaluating the impact of programs on the strengthen their abilities. Rather than local environment; providing for those plan and make large up-front affected adversely by the program; investments in local capacity building as xii a prerequisite for devolution of who provide services to the public responsibility, it is likely to be quicker sector. Finally, it is a mechanism for and more cost-effective to begin the strengthening voice and partnership. process of devolution, to permit learning by doing and to build up capacity xxxi. When done well, fiscal through practice. The key is to make decentralization can support hard budget sure that capacities that are built from constraints and macroeconomic stability, the bottom-up are consistent with the as well as reducing moral hazard and strategies of decentralization being generating responsive, effective and proposed from the top down. sustainable service delivery. This in turn can promote increased incomes and xxix. Community driven development productivity, improved literacy, better (CDD) efforts are likely to be a tool for health and strengthened civil society. supporting changes in institutions and The problem is when done badly, it can accountability, and in strengthening have the opposite results. The key social capital. By building capacity at question then becomes, how to "do the lowest levels, CDD can be an decentralization" well. Given the diverse important element of effective starting positions and idiosyncrasies that decentralization. However, as with other exist in the transition countries, the types of capacity building, it is critical answer is likely to differ, but all can be that CDD be developed in the context of considered in the context of what we a consistent and coherent strategy for have learned about the challenges that decentralization, ideally as part of that transition countries have faced so far. overall strategy. This assures that the learning and capacity developed will xxxii. Figure 7 presents a fiscal contribute to effective local decentralization strategy with three governments. layers. The first layer reflects aspects of the contextual environment ( the TheRoadAhead "overpinnings", so to speak ) that go beyond the system of intergovernmental xxx. The rationale for pursuing finance, but that have an important decentralization is not always influence on how effective straightforward, but there is reason to decentralization may be. These include believe that decentralization in an the structures of government and several accountable environment can help to dimensions of accountability, such as improve the performance of the public political accountability, freedom of sector, assist in combating corruption information, rule of law, the extent of and help countries in their efforts to state capture and even the extent to reduce poverty. In terms of public sector which fiscal management, both on the reform, decentralization is at the nexus expenditure and revenue side, have of key forces that help to strengthen advanced . The second layer sets out the public sector institutions overall. It can constituent parts of a system of be a vehicle for establishing the intergovernmental finance: expenditure appropriate rules and restraints for all assignments, revenue assignments, the levels of government. It is a mechanism transfer system and the legal and for creating competitive pressures, regulatory framework for subnational sometimes among governments, and borrowing. It is important that this sometimes among private contractors system be coherent and create the right xiii incentives for responsible behavior for size of many of the local governments. each part of the system. Finally the third However, information & monitonng layer is composed of elements such as systems and social capital are relatively capacity building, strengthening of civil well developed. While there is work to society, specific systems of information be done to improve elements of the and monitoring-- the "underpinnings" systems in these countries, they are well that need to be developed from the on their way. bottom up. xxxv. Therefore in the "keen xxxiii. Countries find themselves at decentralizers", a priority will be different starting points, with some providing support to countries in this elements of such a strategy in place and area in the run-up to EU accession. some not. Other countries may have few Specific issues that will likely need to be of these elements in place. While each addressed include: the role of regional country is different, the information set governments and implications for out in the preceding sections suggests expenditure assignments; developing that the transition economies can be "own revenues" through property taxes, grouped into four categories. Each personal income tax surcharges, motor requiring a slightly different approach to vehicle taxes; transfer mechamsms for the elements of a strategy. The four the allocation of funds, especially groups are as follows: "keen investment funds; assistance in helping decentralizers", "uncertain to contract out infrastructure and support decentralizers", "non-decentralizers" and to help build capacity in the smallest "decentralizers by necessity". communities. xxxiv. The "keen decentralizers" xxxvi. The "uncertain decentralizers" comprise Czech Republic, Croatia, include Albania, Bulgaria, Kazakhstan, Estonia, Hungary, Latvia, Lithuania, Kyrgyz Republic, Romania, the Russian Poland, Slovak Republic and Slovenia. Federation and Ukraine. Many of these While these countries have chosen countries are working wholeheartedly to different degrees of decentralization. improve their systems of Most have achieved some degree of intergovernmental finance, but do not consensus on what they are aiming for. always have a consensus on where they Structures --although evolving-are in are going, and have weaknesses in the place and there is open discussion about "overpinnings" and the "underpinnings" how and if they might be changed. for effective decentralization. Systems for accountability, such as Administrative reforms and the legal and elections, freedom of the press, etc, are regulatory frameworks are less far along largely in place. In most of these and the elements necessary to support countries the major elements of the accountability are in need of further system of intergovernmental finance are development. Typically in these in place - efforts need to be focused on countries the framework for developing own revenues and on intergovernmental finance requires more developing sources of financing for than just changes on the margin, but renewing and rebuilding infrastructure some major realignments to improve (to bring it up to EU standard). Capacity incentives. Significant efforts are needed in these countries is typically an area to build capacity at lower levels of that needs work because of the small government, to strengthen information xiv and monitoring, and to support the context, some of the "standard rules" of development of civil society. The intergovernmental finance are over- specific entry point and sequencing of ridden, so efforts at continued reform measures will vary in each country, need to focus on strengthening depending on the progress made so far in accountability, building capacity and each area and the degree of political will. developing civil society. An additional consideration is that many of these countries have not yet fully xxxix. In sum ,there are many different come to grips with macro stabilization roads to decentralization. To the present, and therefore efforts at decentralization most countries have made great strides, must be careful not to cause further particularly given the monumental deterioration on the macro front. changes underway in these societies. Some have come quite far, but only very xxxvii. The "non-decentralizers" include: few countries in the region can truly be Armenia, Azerbaijan, Belarus, Georgia, said to have decentralized. Real, Moldova, Tajikistan, Turkmenistan and effective fiscal decentralization will take Uzbekistan. While some of these more effort and more time on a variety countries such as Georgia and Moldova of fronts. have begun to pursue decentralization recently, others remain highly centralized. In these countries, effective decentralization is unlikely until the "overpinning" and "underpinnings" can be significantly strengthened. Structures may be developed, but accountability is particularly weak among a number of its dimensions. System of Intergovernmental finance inmost of these countries continue to largely reflect old modes of doing business. Local capacity is minimal and civil society and strengthening of social capital is in its infancy. In these countries, priority should be placed on strengthening accountability and supporting the development of civil society. Gradually, if the situation permits, countries may then consider devolving more authorities. xxxviii. The "decentralizers by necessity" include Bosnia & Herzegovina, the Former Republic of Yugoslavia and, potentially, Macedonia. These countries are not cohesive states that choose to decentralize, but rather countries in which decentralization pursues specifically political purposes. On this xv WHY DECENTRALIZATION? 1. Across the world countries are increasingly focusing on decentralization of decision-making authority to lower levels of government. Large countries of the OECD such as Australia, Canada and the United States have for many decades developed the systems by which different levels of government are provided with particular authorities and decision-making responsibilities. The countries of Europe, each in their individual way, have subscribed to the European Charter of Local for Self-Government and the underlying principle of "subsidiarity" by which public authonties are generally to be exercised by those governments closest to the people. In more recent years, there has been a wave of decentralization across developing countries. Latin America has seen rapid decentralization since the fall of many authoritarian regimes.2 China began a series of reforms of its system of intergovernmental finance in the mid-199Os.3 Countries as diverse as Indonesia and Pakistan are pursuing rapid ("big bang") decentralization. Other countries, such as South Africa, are taking a more gradual approach.4 2. The transition countrnes of Eastern Europe and the Former Soviet Union have not been excluded from this phenomenon. Indeed, the very transition itself is a process of moving away from a centrally controlled, command-driven economy towards more politically, economically and administratively decentralized and market based structures.5 No doubt that some countries have gone further than others, but the nature of this type of transition raises special issues for decentralization in all of these countries. Early steps in this process were documented in Decentralization of the Socialist State: Intergovernmental Finance in Transition Countries.6 Several years on, we find that progress has varied and that in most countries, the central government has delegated the administration of many activities to lower levels of government, but in only a few cases has real decision-making authority been devolved. Unsurprisingly, it has taken longer than expected for countries to work out new structures of government. While service responsibilities have been delegated or devolved, central governments are often reluctant (and often for valid reasons) to give up control over revenues. Perhaps most important, effective decentralization presupposes a broad set of institutional structures that did not exist under the old system and these take time to develop. 3. This raises a fundamental question that must be addressed for any country in any region, but which is particularly important in the transition economies: Why decentralize? While in any given situation the reasons will be distinct, in transition countries there appear to be three key imperatives driving decentralization: 2 See Burki, Perry, Dillinger and Webb. (2000) and Burki et al. (2000) 3See Bahl. (1999). 4 For a broad perspective see the World Development Report 2000; and Vaillancourt (1997) Ter- Minassian, Teresa. ed. (1997). 5 See the World Bank Development Report 1996 and the EBRD Transition Reports 6 See Bird, Ebel and Wallich (1995).. 1 * The political imperative. Throughout the world decentralization has been driven by a political process and not an economic one. In transition countries, the back-lash to years of central domination and control has manifested itself in many countries as a strong push for autonomous local self-governments. In some countries there is a clear vision and political consensus on the degree of decentralization that should exist. In others, such consensus does not exist. In yet other countries, such as Bosnia & Herzegovina, the system of decentralization acts in some sense as the glue that holds the country together. The critical point to make is that the system - no matter where it lies on the continuum from centralized to decentralized-- is the result of a political process. * The economic imperative. For transition economies that spent decades under state- run, planned economies, there has been virtually no choice but to reduce the role of the state in the economy and "decentralize" to markets and the private provision of goods. Nevertheless, state and local governments in some transition countries still have significant involvement in the regulation and provision of services that are not of a public good nature. Such involvement in many cases leads to an un-level playing field and works against both the development of new small- and medium-enterpnses and increasing new investment in the economy. There is thus a strong economic motive for redefining the government's role at all levels. * The service delivery imperative. Under the old system, central governments used regional and local branches to deliver services, but the bulk of decisions were made at the center and in many countries the sum total of mandated services and benefits were fiscally unsustainable. In the new systems, under increasing fiscal constraints, most countries have passed the responsibility for provision of key services on to lower levels of government. In some cases, local governments have the decision-making authority to prioritize expenditures and make the choices that meet the expectations and demands of citizens and the capacity to implement them. However, in many countries, the central government still has decision making authority over key aspects of service delivery, which can hamper the efficient and effective provision of the service. While there are some services that can and should be provided by the central govermnent, others can be provided more efficiently at the regional or local level in many countries. 4. In sum, efforts to decentralize are occurring in the transition economies for political, economic and fiscal reasons. The objective of this paper is to highlight the key challenges to effective decentralization that have become apparent over the past decade; and to set out ideas about how to address these issues more effectively in the future. The next section will start by giving a perspective on the degree of decentralization across the transition economies. Subsequent sections will then consider a number of critical issues and challenges. Finally the paper will close with suggestions concerning policy steps and sequencing that may help address these issues in the future. 2 THE DEGREE OF DECENTRALIZATION TO DATE 5. The notion underlying theoretical discussions of decentralization is that local governments, which are closer to citizens, are likely to have better information on their needs and/or demands and therefore will be able to provide services more efficiently than the central government. Underpinning this general idea are a range of assumptions concerning the overall context for decentralization. A principle assumption is that local governments can be held accountable by individuals (e.g. throwing local officials that don't meet demands out of office) or that individuals have the capacity to move to another jurisdiction if they are unhappy with the services provided (e.g. "voting with their feet"). Not only must individuals and local governments have access to information, but they also must have the capacity to use it. Finally, there is also an underlying assumption that there is a link between the services that a locality provides and its revenue base (at least at the margin) so that local individuals bear the cost of the services provided by their local government. Little is said in the literature about outcomes when these assumptions do not hold. In the transition countries, the extent to which these conditions do hold varies considerably and has an influence on the degree and effectiveness of decentralization efforts. 6. Even measuring decentralization poses difficulties. Generally speaking, decentralization refers to the transfer of authority and responsibility of public functions to subordinate governments, however, this can include different types of decentralization: political, administrative, fiscal and market decentralization. All of these different types of decentralization overlap. This paper focuses in particular on fiscal decentralization. There are also differences with respect to the degree of decentralization. Deconcentration has lower levels administering central government decisions; delegation transfers some responsibility for decision-making to lower levels, and devolution gives full responsibility for decision-making, financing and management to lower levels of government. While people referring to "fiscal decentralization" often mean devolution; in practice most countries have (and should have) a mix of deconcentrated, delegated and devolved functions. 7. The most common measures of decentralization -- local government expenditure as a share of GDP and local expenditures as a share of total general government expenditure (Figure 1 & 2) -- indicate that these two measures vary considerably across the transition countries, and even among sub-regions. Russia, Ukraine, Kazakhstan, Hungary, Latvia and Estonia all have local expenditure over 10 percent of GDP and a relatively high share of local expenditure in total general government expenditure 7. Russia's shares are similar to those of other large federations (Canada and the US), though they are lower than a number of northern European countries (Sweden, Denmark, the Netherlands). At the other end of the spectrum, local expenditures are less then 5 percent of GDP in countries 7The data available for Poland does not include the intermediate levels of government (vovoidships and powiats) and therefore underestimates total local expenditure relative to the data for countries such as Romania, Russia, Ukraine and Kazakhstan, which include the intermediate levels of government 3 Figure la. Local Expenditure as a Share Figure lb. Local Expenditure as a of GDP (percent) Share of Total General Government Expenditure (percent) Estonia' Estonia Latvia Latvia Lithuania Lithuania Croatia Croatia Czech Republic Czech Republic Hungary Hungary Poland (a) ;land Slovak Republic Slovak Republic Slovenia n Slovenia Albania Albania Bulgaria 1 Ronania B Russian Federation si Federain == = Ukraine _ ______ l__ _ .Ukraine _ _ _ Arnenia Arnenia Kazakhstan Kazakhstan 0 5 10 15 20 0 20 40 60 Sources: See Annex Table 1 5 such as Armenia, Romania, and Slovakia, suggesting a less extensive role for lower level 8 governments 8. By sub-region, the available data suggest that the central CIS countries, Russia & Ukraine, have the most significant share of local (including regional) expenditures, followed by Kazakhstan in Central Asia, and the Baltic countries. In the Central and Eastern European countries, Hungary, Poland and the Czech Republic9, local governments have a larger role than in the others. Relative to other sub-regions, South Eastern European countrnes (Albania, Bulgaria, Romania) and the Caucasus indicate a smaller local government share in GDP and total expenditure. Because of its own distinct character, measurement of the degree of decentralization in the Balkan countries is not strictly comparable with the data above . However, fiscal decentralization plays a key role in Bosnia & Herzegovina and is likely to do so m Macedonia and Yugoslavia (Box 1). Box 1: Fiscal Federalism in Bosnia & Herzegovina Under the Dayton Accords, Bosnia & Herzegovina (BiH)was set up as a single sovereign state composed of two entities: the Federation of Bosnia & Herzegovina (FBiH) comprised of two major ethnic groups and split between Croat majority areas and Bosniac majority areas) and the Republic Srpska (RS). Under the State Constitution most functions are assigned to the entities, including defense and the authority to tax. Two-third of the States revenues are to come from FbiH and one-third from RS. Overall, the State expenditure as a share of total consolidated expenditure was about 4 percent in 1998. The Entities themselves have very different structures and wide responsibilities. FBiH has two additional levels of subnational government, the cantons and the municipalities. Entity functions include defense, internal affairs, police, economic and social sector policies, reconstruction programs, displaced persons, justice, tax and customs. The Federation and cantons have joint responsibility for health, environmental policy, infrastructure, communication, transport and social welfare among others. The cantons are responsible for other matters including education, culture, housing. They are allowed to delegate functions to the municipalities. In 1998 some 83 percent of total consolidated expenditure was by FBiH RS has only one level of municipalities and is more centralized, although efforts have been made recently to move more responsibilities to the municipalities. RS accounted for 17 percent of total consolidate BiH expenditure in 1998. Referred to as "asymmetric federalism", this type of arrangement in makes it possible to find a politically acceptable solution in a highly polarized society. However this is not done without costs to effective decentralization and fiscal management. The State's limited role is difficult to carry out because of its reliance on the Entities for resources. Highly fragmented local governments make efficient provision of services quite difficult. With no transfer across entities, or even among communities, it has proven extremely difficult to address the significant disparities that exist within the Entities, much less across the CAlilntrv 8The data used in this report, unless otherwise specified, come from the country reports and documents specified in Annex Table I Every effort has been made to assure consistency in the classification across countries Nevertheless, data is not available for all transition countries and for those in which it is available, country circumstances are changing rapidly as is the data. Any corrections and updates are most welcome. 9 Although data is not available for all FSU countries, it appears that for most countries (with the exception of Azerbaijan and Armenia) local expenditure as a share of total expenditure is relatively high - about 3540% This is a result of most education and health expenditures being administered by lower levels. 6 9. These standard measures, however, fail to take into account the different degrees of decentralization discussed above (deconcentration vs. delegation vs. devolution) and the true decision-making authority of localities over expenditures. In Ukraine for example, local expenditures are considerable at 14 percent of GDP, but much of this expenditure is administered by oblast and rayon levels of government in a deconcentrated framework and is not really expenditure over which localities have control. They also do not take into account the degree to which the assumptions underlying the efficiency arguments for decentralization are in place. To fully assess the degree of decentralization it is necessary to consider fiscal decentralization as a comprehensive system. Such a system includes the basic structure of government and institutions that support accountability. It includes a framework for expenditure and revenue assignments, as well as a system of effective transfers. An important component of such a system is also the monitoring and regulation of local government activities, especially local government borrowing. Finally it requires capacity at both the central and especially the local level to manage public revenue and expenditure, effectively implement policies and deliver services. Each of these areas raises particular challenges to effective decentralization in the transition countries and sheds light on steps that may be desirable in the future. CHALLENGE 1: ADAPTING TO CHANGING GOVERNMENT STRUCTURES 10. The decline of the Soviet Union and the centrally planned economy brought with it fundamental changes in the structure of government across the transition economies and after a decade these structures continue to evolve. There are a number of issues that most countries in transition are confronting, two of which stand out. The first is the appropriate role for the regional level. The second is the appropriate size of a local government for effective service delivery. Government Structures 11. All but three of the transition countries (Russian Federation, Bosnia & Herzegovina and the Federal Republic of Yugoslavia) are formally unitary. In each country there exist multiple tiers of government, and most countries have two or three subnational levels (Table 1). Seventeen of the 27 countries have a regional/oblast level of government, which typically serve populations of above 500,000. 19 countries have county/rayon/district levels of government, which typically serve populations between 60,000 and 100,000. A level of government corresponding to towns and municipalities exists in virtually every country, but the size of the average population of this lowest level varies considerably, from only two to three thousand in the Czech Republic, Hungary, Latvia, Slovakia and Slovenia; to over 50,000 in Belarus, Lithuania, Tajikistan and Turkmenistan. 12. The number of tiers and the number of units at different levels depend on a range of country specific characteristics, such as total population, geographical size, urbanization and factors such as historical structure (Annex Table 1). For example, the Russian Federation, with a large and heavily urbanized population and vast territory, has the largest number of regional and local units. In contrast, Kyrgyz Republic, Tajikistan and Turkmenistan, which have smaller populations and territories and are less urbanized, have 7 a smaller number of regional and local units. Despite their differences, transition countries face some common issues with respect to government structure. Table 1: The Structure of Subnational Government in Transition Countries No. of No. of Average No. of Average No of Average Country Subnational Top tier Population Lower Tier I Population Lower Tier 2 Population Tiers (Region/pro (000s) (districts/ray (000s) (towns/municip (000s) vince/ ons/counties) alities,etc) oblast) Albania 2 12 97 374a 9 Armenia 2 11 346 930 4 Azerbaijan 2 71 107 3,500-4,000 Belarus 3 7 1454 222 133 58 Bosnia & Herz. 3 2 1836 lO(FBiH) 398.5 80(FBiH) -(RS) - 62 (RS) Bulgaria 2 9 921 255 33 Croatia 2 21 232 5957_ _ 9 Czech Republic 3 14 737 76c 119 6292 2 Estonia 2 15 96 253 6 Georgia 2 12 450 61 -1000 Hungary 2 19 3 170d 3 Kazakhstan 3 1 6 978 242' _? Kyrgyz Republic 3 6 645 619 Latvia 2 33 h 100 563' 3 Lithuania 2 10 371 56J 66 Macedonia, FYR 2 34 63 123 17 Moldova 2 11 446 911 4 Poland 3 16 2400 373 90 2483 12 Romania 2 41 557 1 2948 10 Russian Fed. 3 89 1652 29671 2337m Slovak Republic 3 8 674 79 68 2875 2 Slovenia 1 192 2 Tajikistan 2 3 1967 70 84 Turkmenistani 2 5 920 64 72 Ukraine 3 27n 2058 9370 105 30,000P Uzbekistan 3 14 1714 281 85 1749 13 Yugoslavia, FR 3 2 (Serb./Mont) I I_I a 65 communes + 309 municipalities. b 472 municipalities + 123 towns c Districts to be dissolved by end 2002 d Includes 22 cities with county nghts. e Includes two cities with oblast status. f. 163 rural rayons and 79 towns of rayon status g. 44 rayons and 21 cities. h. 26 districts and 7 republican cities i 70 town municipalities and 493 small urban and rural municipalities. j. 12 urban local governments and 44 rural local governments k. 2686 communes, 182 towns and 80 municipalities. 1. cities and rayon level with budgetary rights m sub-rayon level with budgetary rights - there are estimated some 27,500 urban and rural settlements and some 12,000 municipalities in the Russian Federation, but most without budgetary rights. n Includes 24 oblasts plus Autonomous Republic of Crimea and Cities of Kiev and Sevastopol. o 490 rayons and 447 municipalities (of which 167 under oblast admimstrative control). p. Villages and settlements. Sources: See Annex Table I 8 The Role of the Regional Level 13. Under central planning, administrative arms of the central government carned out functions and delivered services at the regional/county level (in Central and South Eastern European countries) and often at the district/rayon level (in the former Soviet Union countries). In response to the end of central planning, many countries created a new layer of autonomous local self-government at the lowest level. In some countries, as in Hungary, there was a backlash against the intermediate level and its roles and responsibilities were significantly curtailed. In other countries, such as in Ukraine and Kazakhstan, representatives of the central government at the oblast or rayon level continue to vie for authority with elected councils at those levels (Table 2). 14. In recent years a number of countries, especially (but not exclusively) in the EU Accession countries, have sought to reorganize and consolidate their intermediate levels of government. Poland consolidated from 49 vovoidships to 16. Albania has recently created 12 regional governments (to match the prefects) that will eventually become autonomous governments. Hungary has developed an administrative structure of seven regions to manage regional development, which over time may become the basis for elected regional governments. In most countries in which the intermediate levels were downplayed, governments are finding a need for them to fulfill certain responsibilities. 15. While the role of this regional level is relatively clear in the federal countries, it will continue to evolve in the unitary countries. For implementation of good policy it is critical that as administrative and territorial reform evolves, the specific functions and authorities, as well as the sources of finance of these intermediate levels be clearly specified. As the role of the intermediate level(s) change over time, the system of intergovernmental finance will have to adapt as well. The Appropriate Size of Local Government for Service Delivery 16. As pointed out above, the size of the lowest level of government varies significantly in transition countries. Countries such as Czech Republic, Hungary and Latvia that have very small local governments face the potential for service delivery that is fragmented and inefficient. Small local governments often lack the capacity to manage all the functions assigned to them. At the same time many, smaller governments may allow for greater participation and accountability and may be politically desirable. In some countries, such as in Lithuania, the solution pursued is to aggregate smaller localities, so they are a sufficient size to provide their assigned functions effectively. In countries where aggregation is not politically possible, municipal associations can be created and may perform an important role, but clear incentives are needed for the associations to be able to function effectively. 17. In other countries the size of what is called "local government" may be too big to effectively provide services that respond well to local needs. In Tajikistan and Turkmenistan, regional and local governments serve large populations. In Russia, current legislation refers to regions and to "local government" assuming that there is only one level 10 See Bird et al (1995) p. 15. 9 of local government. But in practice, there are a variety of different arrangements across the regions, with sub-rayon levels governments in some cases having budgetary responsibilities and in others not."1 Recently the Government has begun to try to address this mis-match between what is stated in legislation and what exists in practice. 18. The challenge for transition countries is to find stable structures of government that are politically acceptable and at the same time help to promote efficient and effective local governments. At both the regional and local level, administrative reform is likely to continue, and assignment of functions and finances will need to adapt over time. Of critical importance is that measures to reform government structures clearly take into account the appropriate scale for key services and that roles and responsibilities are set out in a clear legal framework. CHALLENGE 2: STRENGTHENING ACCOUNTABILITY 19. As discussed above, the notion that decentralization can lead to more efficient delivery of government services rests on the notion that local governments can be held accountable. This section considers the extent to which some of the institutions that support local government accountability are in place. The data suggest that there is a substantial degree of difference across the region in the extent to which supporting institutions are developed. In those countries where such institutions are weak, policies to pursue decentralization may not have the desired outcomes, and where possible, strengthening such institutions may help to support more effective systems of intergovernmental finance. Accountability 20. Accountability for local government can take place in many fashions, but elections ("voice") and freedom to move to another jurisdiction ("exit") are probably among the most discussed in the literature on decentralization. Just as there are quite a variety of structures of government in the transition countries, there are a wide variety of electoral arrangements as well (Table 2). At the lowest level of local government (towns and municipalities) of the 27 countries considered, 19 hold elections for both local councils and the local executives. Five countries have arrangements in which the local council is elected but the local executive is appointed by either the President or the Governor of the next highest level. Three countrnes (Belarus, Azerbaijan and Turkmenistan have no elections at all for local governments at any level.) At the county/district/rayon level, of the 19 countries that have such a level, seven countries have elections, 7 countries do not, and in 5 the county/district/ rayon council is elected but the executive is appointed by a higher level - sometimes the President, sometimes the governor at the next highest level. Of the 17 countries that have regional/oblast levels, " See Kurlyandskaya (2001). 10 Table 2: Elections for Local Governments in Selected Transition Countries Country Representatives Elections for Elections for Elections for Comments of Central Gov't regional units? district level? Local level? at lower levels? Albania Yes- 12 7 Yes New law (July 31, 2000) creates 12 regions and prefectures disbands districts as administrative units Armenia No Yes Central govemment appoints and dismisses heads of rcgional govemment Azerbaijan Yes No Yes Local executive power lies with regional govemors appointed by president Belarus No No No Prcsident appoints govemors, who then appoint subordinatc officials Bosnia & Herzeg N Y (entities) Y/N Y Each entity has separate Law on Local Self- Govemment FbiH has autonomous cantons and elected municipalities The RS more centralized Bulgana N Y Council of Ministers appoints rcgional govemors Mayors and municipal councils arc directly elected for 4-year terms Croatia Y Y Y Local prefects are appointed and approved by president to represent state authorities Local executives (mayors) chosen by councils Czech Republic Y V V Y Distncts to be dissolved at the end of 2002 Estonia Y N Y County govemor appointed by central govemment on proposal of Pnme Minister Georgia Y N Y/N Y/N Top tier govemed by appointed commissioners except in 2 Autonomus Republics and Tbilisi Rayons/cities and lowest tier have elected councils but appointed executives. Hungary Y Y Y Kazakhstan Y/N Y/N Y/N Oblast akims appointed by president Local level akims appointed and removed by oblast-level Regional/Local councils elected Kyrgyz Republic Y/N Y/N Y/N President appoints oblast, rayon and town executives Subnational councils are elected. Latvia N N Y District councils consist of chairs of murnueipal councils. Local councils elected, local councils appoint mayor Lithuania N V Central govemment appoints govemors of regional administrations. Local councils are elccted. Mayors appointed by local councils Macedonia, FYR Y N Y Status of 34 administrative units is unclear, but officials centrally appointed Municipal councils and mayors are elected Moldova N Y Local councils are elected Executive selected locally, but approved by the president Poland Y/N V V Vovoidship (regional)council elected, but head appointed by prime minister. Romania Y Y Y Russian Federation Y for super- Y Y Y Heavy influence from layer above. districts Slovak Republic Y N N Y Administrators of regions and districts are appointed by central govemment Mayors and municipal assemblies are directly elected. Slovenia Y Y Municipal councils and mayors are directly clected to four-year terms. Tajikistan N Y/N Executive at regional level appointed by president. President can also remove local officials, mcluding locally elected mayors Turkmenistan N N Regional and municipal hakim (executives) are appointed and dismissed by president Ukraine Y Y/N Y Y Rcgional Govemors appointed Rayon and local level elected (but have in ast becn removed by President) Uzbekistan N Y/N Y/N Regional executives (hokims) are appointed and dismissed by President Mahalla leaders (local) selected from "village elders" Wilayat (regional) and local councils are clected but elections influenced Yugoslavia,____FYR _________________ by higher levels. Yugoslavia, FYR ________ Y Y only four have elections (the Federal governments and the Czech Republic). Four have elections of regional councils, but appointment of regional/oblast governors and ten have no elections at the regional/oblast level. 21. There do seem to be some tendencies exhibited by sub-regions. In the Baltics, the intermediate level is appointed (though in Latvia district councils consist of the chairs of municipal councils) and local levels are directly elected. In the Central and Eastern European Countries, in all but Poland and the Slovak Republic, elections are held at all levels. In the South Eastern European countries, elections are held for all levels in Romania, but intermediate governments are appointed in Bulgaria and Moldova. In Georgia, executives of lower levels of government are appointed, while the councils are elected, and in Armenia regional governments are appointed and local governments elected. Of the central CIS countries, Russia has elections at all levels, Ukraine appoints executives and elects councils at the oblast and rayon level and elects both mayors and councils at the lowest level, and Belarus has no elections. Finally, in the Central Asian countries there is a much greater tendency to appoint the executives at all levels of Government, although in some countries, such as the Kyrgyz Republic, councils may be elected. 22. Two points stand out in considering the degree to which subnational governments are politically accountable. First, in most countries of the region the intermediate levels of government remain ones in which accountability is to higher levels of government, rather than to the electorate. This suggests that activities undertaken at these levels remain deconcentrated and perhaps delegated, as opposed to truly being decentralized. This raises the issue of how to effectively meld deconcentrated functions and responsibilities with those that may be devolved to lower levels. It also raises the important question of how the flow of funds through deconcentrated intermediate levels (with accountability to the center) affects the real autonomy and decision-making power of the lowest levels of government (with accountability to the electorate). 23. A second issue that arises relates to the degree that elections at the lowest level are free and fare, or are dominated by national political parties, or other specific interests. Local bosses that gain substantial influence exist in all countries, including countries in transition. The question arises as to whether the possibility of corruption may be even greater at local levels of government than at national levels. l While much of the literature suggests that citizens are likely to hold local governments more accountable than national ones, this presupposes the existence of institutions and a political culture that supports this accountability. During Soviet times, incentives worked against this type of accountability, and this legacy remains in many countries. 12 The recent empirical literature on the relationship between decentralization and corruption provides mixed results Treisman (2000) finds that large federations with greater decentralization are associated with higher corruption, whereas Fisman and Gatti (2000 ) finds the opposite. Of course there are substantial issues of how to effectively measure both decentralization and corruption. 12 Mobility 24. The mobility of a population provides a different mechanism for accountability. In theory, those who are dissatisfied with the services delivered by one local government can "vote with their feet" by moving to another jurisdiction that better meets their preferences. In practice, there are a number of constraints on the mobility of the population in transition countries. In the past, mobility was restricted by resident permit systems and by restrictions on housing such as limited private ownership and lifetime occupancy. The relaxation of formal controls on mobility has led to major urban-to-rural migrations in Macedonia, Romania, and Poland, and rural-to-urban migration in Albania.1f While restrictions on housing may have been reduced in some countries, shortages of housing appear to have had some affect on mobility in countries like the Czech Republic. Housing is also a key constraint to mobility in Russia. In many of the less reform minded countries of the region, mobility is limited by housing and benefits that are associated with the insolvent firms. Workers may no longer receive wages, but they are unwilling to lose rights to suchI benefits. 25. Although mobility has improved significantly in the Baltics and CEE countries, (less so in other countries of the region), it is unlikely to play a significant role in local government accountability in the near future. This is partly because mobility remains limited, but also because factors beyond local service provision often influence individual decisions about where to locate. While efforts to improve labor mobility can go some way to strengthening "exit" as a form of local accountability, it will likely be some time, if ever, before it plays a significant role . Information 26. An important aspect of the ability to hold any level of government accountable is the degree of information available to citizens. It is difficult to participate effectively and exercise "voice" if there is no information both on decisions to be made and the basic information necessary to evaluate different points of view. There is no data available differentiating access to information at different levels of government. However, rankings of press freedom at the national level may be considered a proxy (albeit an imperfect one) for the degree of information flow more generally. Data available at the national level ranking press freedom suggest substantial variation across the transition countries (Figure 2). The data do not suggest that there is a straightforward association between local expenditures as a share of GDP and a ranking of press freedom, but groupings of countries do appear. In the northeast quadrant of the figure, three countries have a high share of local expenditures in GDP and a scoring in the press freedom ratings, that suggest that there is no freedom or limited freedom of the press (in the ranking higher scores are associated with lower press freedom). These countries include Kazakhstan, Russia and Ukraine. In the northwest quadrant of the figure, a second group of countries is scored with partial press freedom or a free press, but has a lower share of local government expenditure in GDP. These include Albania, Bulgaria, Croatia, Romania, Slovak Republic and Slovenia. A third group of countries have press freedom, and the share of local government in total GDP 1' See World Bank 2000. Balancing Protection and Opportunity: A Strategyfor Social Protection in Transition Economies. Washington, DC: The World Bank. P. 15-15. 13 ranges from 8 to 13 percent of GDP. This group includes Czech Republic, Estonia, Hungary, Latvia, Lithuania and Poland. Figure 2: Local Expenditure as a Share of GDP and Press Freedom 14 6- E 0 5 2a 4 -1 0 0 3 -l %.2 2 - 0 5 10 15 20 Local Exp./GDP (%) 27. Better information flow is likely to be found in those countries with more freedom of the press. Such information flow can play a role in improving transparency and supporting accountability, both generally and at the local level more specifically. Other forms of information flow such as public availability of programs and accounts, open budget hearings, provision of information over the internet are tools that can also help to hold localities accountable. It is in the third group of countries discussed above, where information flow appears to be strongest, that local governments are more likely to be accountable. In other countries, measures to improve both information and transparency can be an important supplement to decentralization policies. Rule of law 28. The ability of a country to make, implement and enforce laws is also an important factor in its ability to hold elected governments accountable. Without rule of law, the notion that local government (or any level of government for that matter) can be held responsible for its actions becomes more tenuous. Again data on rule of law are available only for the national level, but are likely to be relevant at the subnational level as well. Figure 3 shows no direct association between local expenditure as a share of GDP and an index ranking rule of law in the transition countries.' It does, however, show a similar 14 The Press Freedom Rating comes from Freedom House, Nations in Transit 1998, Press Freedom Survey 2000 and Freedom in the world 1998-99. I The rule of law index is that composed by Freedom House and reflects the degree to which the citizens of a country are willing to accept the established institutions to make and implement laws and adjudicate disputes. 14 differentiation among countries as seen in Figure 2. In this case Slovenia is part of the group of countries with a relatively low "rule of law" ranking. 16 Figure 3: Local Expenditure as a Share of GDP and Rule of Law 17 6- 5 - 3 41 1 -11 0| i 3 - x 0 0 5 10 1 5 20 Local Exp.IGDP (%/) State capture 29. Recent work in the World Bank has highlighted the role of "state capture" as a response on the part of firms or private individuals that compensates for weaknesses in the legal and regulatory framework . State capture reflects the ability of firms to shape and influence the formulation of the rules of the game through private payments to public officials (or the ability to influence without payments). High levels of state capture are likely to signify an institutional environ-ment that will pose difficulties for effective decentralization. It basically implies that governments become accountable to a subset of the population. The survey data available on state capture, largely focus on the central govermnment, but as with the measures discussed above, information on the degree of state capture generally is helpful in considering the extent to which the institutional environmnent may be a conducive one to decentralization policies. Low ratings indicate sound political institutions, a strong court system and provisions for an orderly succession of power. High ratings indicate a tradition of depending on physical force or illegal means to settle claims. 16 The correlation coefficient between the index of press freedom and the index of rule of law is 0.885 17 The Rule of Law index comes from Freedom House, Nations in Transit 1998, Press Freedom Survey 2000 and Freedom in the world 1998-99 18 See Pradhan et al. 2000. Anticorruption in Transition, a Contribution to the Policy Debate. Washington, DC: The World Bank. 1 5 30. The share of local expenditure in GDP does not demonstrate a straightforward relationship with the degree of state capture (Figure 4).19 Indeed, there is a wide degree of variation in the importance of local expenditures at both low and high levels of state capture. The countnes with a relatively low level of state capture (Armenia, Czech Republic, Estonia, Hungary, Kazakhstan, Lithuania, Poland and Slovenia) have shares of local expenditure in GDP that range from 1.2 percent to about 13 percent. The countries with relatively high shares of state capture (Bulgaria, Croatia, Latvia, Romania, Russia, Slovak Republic and Ukraine) have shares of local expenditure in GDP ranging from 3.8 to fifteen percent. There does, however, appear to be a demarcation between countries with low and high levels of state capture and it is certainly an important consideration to take into account when assessing the effectiveness of decentralization more broadly. In those countries with high state capture, the likelihood that local governments will be subject to more traditional types of accountability may be severely diminished. Figure 4: Local Expenditure as a Share of GDP and State Capture 35 - 3 25- m 20- 5 _ 0 0 5 10 15 20 Local Expenditure/GDP (%/6) 31. Elections, mobility, information, rule of law and the extent to which the state can be influenced by outside interests all represent different mechanisms that affect the degree to which local governments can be held accountable. They also give some indication of whether the institutional underpinnings that are necessary for successful decentralization are in place and therefore the extent to which decentralization is likely to be meaningful. This overview suggests that the supporting institutions in the Baltics and Czech Republic, 19 The measure of state capture used was developed by Hellman, Jones and Kaufmann.2000 It is a composite index comprising the share of firms that report a significant impact on their business of the "sale" to private interests of parliamentary legislation and presidential decrees, corrupt mishandling of central bank funds, "sale" to private interests of civil and criminal court decisions, and illegal contnbutions by private actors to political parties. 16 Hungary and Poland are furthest along, and that they are weakest in Russia, Ukraine, and the Central Asian economies. The strength of these institutions varies in Caucasus, the Balkans and the South Eastern European countries. 32. Institutions develop slowly and are not easily changed. Nevertheless a fundamental challenge to implementing effective decentralization policies in many transition countries is to strengthen the institutions that support the accountability of lower levels of government. Some aspects of this-- such as elections -- are political and therefore out of the domain of an institution like the World Bank. Other areas, however, such as improving labor market mobility, supporting information flow, increasing transparency and strengthening civil society and the participation of local communities may help to build a solid foundation for more effective decentralization. Without improvements in these institutions in the countries in which they are weak, the desirability of decentralization and the extent to which it is meaningful will continue to be ambiguous. CHALLENGE 3: MANAGING SHARED RESPONSIBILITIES 33. A clear, stable and comprehensive assignment of expenditure responsibilities is often identified as the logical and essential starting point in the design of an effective system of intergovernmental fiscal relations. Standard prescriptions for expenditure assignment effectively call for setting out "who is responsible for what" in a clear and straightforward way. Often implicit in recommendations is the notion that one level of government is entirely responsible for all aspects for service provision. However, in many of the key areas of service provision multiple levels of government have a role to play. The challenge in transition economies is making the roles of each level clear with respect to regulation, financing and provision of service, and providing both the incentives and the checks and balances necessary to support such assignments. Expenditure Assignments 34. The first principle used to determine an appropriate set of expenditure assignments for a decentralized system of government emphasizes the objective of economic efficiency and is referred to as "subsidiarity". According to this principle, the provision for any given public service should be assigned to the lowest level of government (or jurisdiction) that allows for full internalization of the benefits (and costs) associated with that service. Assigning expenditures according to this principle allows for local variation in the service levels to best suit local variation in preferences. Furthermore, the efficient assignment of public expenditures must also take into account possible positive or negative externalities across more decentralized jurisdictions. In line with the principle of "subsidiarity", typically garbage collection, preschool education and city streets are allocated to local jurisdictions, while national defense, the conduct of foreign affairs, defining an international trade regime, regulation of interregional commerce, and immigration are examples of functions allocated to the national government. Between these examples are a wide range of specific expenditure responsibilities which can lie anywhere along a spectrum between being solely a local government responsibility and being solely a national govemment responsibility. 17 35. Studies addressing the issue of expenditure assignments in the transition economies, particularly those that were a part of the former Soviet Union, commonly note "a broad correspondence between the geographical dimension of benefits from a particular service and the level of government responsible for its provision."20 Annex Table 3 shows recent expenditure assignments in selected transition countries, based on data assembled from country-specific studies and other cross-country analyses21. In all but a very few functional areas, responsibilities are shared. University education and areas of transportation such as airports and railroads are assigned to the national level in all but one of the countries for which there is data. (In Croatia, railroads are assigned to the intermediate level.) Functions such as garbage and waste collection , water and sewerage are assigned only to local government in most countries. Local governments are assigned important, but shared, responsibilities in education, housing and community services, and to a lesser extent in health. The most typical way in which assignments in transition countries contradict the commonly accepted assignments, is through reliance on local authorities for provision of redistributive social assistance policies. Redistributive expenditures at the subnational level can take many forms, including price subsidies for locally provided services such as utilities, transportation and housing, subsidies on foodstuffs, direct income supplements, family allowances and the like. Typically, such redistribution is considered to be a central government function. Assigning responsibility for financing social assistance to lower levels of governments, creates that problem that those communities most likely to need social assistance will be least able to afford it. A more effective approach is to have the central government finance social assistance and local levels of government administer programs. 36. Data on subnational expenditure as a share of total government expenditure by functional category bear out these assignments (Figure 5 22). In education, housing and community service and recreation and culture, localities incur a substantial portion of expenditure. This is less the case in the area of health. There are also notable differences across sub-regions. Subnational governments in the Baltics and Central CIS countries spend appreciably more than other sub-regions in both education and health. However, as was the case with the measures used above, the share of local expenditure in total government expenditure may not necessarily reflect the degree to which local governments have actual decision-making authority over expenditures. Often local governments are assigned responsibility for certam services but are typically given little authority over the key decisions that determine effective service delivery. Some of the key issues that relate t expenditure assignments are as follows. 20 See Martinez-Vazquez (1998); Bird et al (1995), and Ter-Minassian (1997). 21 Note that the degree of specificity in which expenditure responsibilities are set out vary considerably across countries. In some cases the legislation simply specifies local functions, in others it specifies the roles of different levels. Annex Table 3 also reflects these reporting differences. 22 Figure 5 is based on data in Annex Table 4 from the IMF's Intergovernmental Finance Statistics. Data were not available for all countries. Baltic countries include Estonia, Latvia and Lithuania. Central and Eastern Europe countries include Croatia, Czech Republic, Hungary and Poland South Easter European countries include Albania, Bulgaria, and Romama. Central CIS includes Belarus and the Russian Federation Data for the Caucasus was limited to Azerbaijan and there was no data available for the CIS countnes. 18 Figure 5 Share of Subnational Expenditure in Total Government Expenditure 120 100 8l Baltics 80 - l CEE 60 -o SEE 40 - [I CCIS 2 OECD 20- 0 -~~~~~~ , .(Pt0) Aligning responsibility for outcomes with decision-making authority 37. There are numerous actors that participate in the delivery of services in most functional areas. These include the central government, in some cases regional administrations, local governments, units that directly provide services (such as schools, clinics, social assistance organizations), boards that govern these institutions (school boards, for example), associations, inspectorates, unions, professional organizations (teachers, health workers) and potentially other organizations. A first clear tension that arises with respect to expenditure assignments in the region is that between responsibility and decision-making authority. While lower levels of government are typically responsible for efficient delivery of many services, in many cases they do not have control over key decisions that affect service delivery. For example, in education, the curriculum to be taught may be fully determined at the central level in a manner that is far removed from experience in the classroom. Decisions on the number of teachers and the wages they will be paid are also often taken at higher levels, leaving little room for maneuver in terms of resource management (see Box 2). At the same time, central ministries and agencies often are so focused on day-to-day operational issues and administration that they are unable to focus on important functions such as strategic development, setting standards, monitoring and evaluation -functions which require a national perspective. 19 38. In the most severe cases, as in most of the Central CIS and Central Asian Countries, the mis-match between responsibilities and decision-making authority is apparent in significant unfunded mandates. In many countries, the Constitution requires that education and health be provided free of charge and that a whole host of benefits be provided, and typically these responsibilities have been assigned to lower levels of Government. Conservative estimates of such unfunded mandates in the Russian Federation amounted to 8-10 percent of GDP in 1999.23 Passing responsibility of delivery of such mandated (and fiscally unsustainable) functions to lower levels of government, in effect amounts to pushing the deficit down to lower levels of government. Without sufficient financing from the center, such mandates lead to a host of budgetary management problems (arrears, barter, mutual settlements). Box 2 Decentralization of Education in Central and South Eastern Europe A recent World Bank Institute Program (see Fiszbein 2001) focused on building analytical capacity for understanding the ways in which evolving intergovernmental structures influence the effectiveness of service delivery in education in six transition countries (Albania, Bulgaria, Czech Republic, Hungary, Poland and Romania). Groups of stakeholders from each country were brought together to: i) assess who was carrying out what functions in education (formally and informally), ii) identify inadequate assignments of responsibilities, mis-matches between responsibilities and authorities, and mismatches between authorities and accountabilities; and iii) identify how to resolve the principal contradictions through redefining responsibilities, providing new authonties or creating new forms of accountability. The purpose of the program was to help countries strengthen their ability to conduct independent analysis of intergovernmental arrangements in order to enhance policy fornulation, but the lessons learned also shed light on some common issues across these countries. In all cases, several levels are involved in the delivery process and no one actor has full decision-making power over key decisions. However, recognizing the principle of shared responsibilities does not mean that there are not issues related to the balance of power among different actors. Many countries are still struggling to find the right balance between the center, the localities and the schools. While there has been a clear shift in the direction of local governments, it is much less clear whether the decentralizing trends will go as far as making schools the "center of gravity " m the education system. Countries differed considerably on whether this is desirable, but most recognize that the level of institutional development (both of the school sand networks that could link them) is insufficient to undertake this vision without a considerable disruption to the overall effectiveness in service delivery. An effective system of shared responsibilities requires the presence of checks and balances. In most countries in the study, these remain weak. Lack of transparent, accurate and timely information is also a barrier. Information is a prerequisite for accountability. In addition to better information, what is needed is development of a more performance-oriented management culture. Countries still rely too much on measurement of inputs as opposed to outcomes. Without exception, the reports identify participation as an essential aspect of inproving the effectiveness of education services. 23 See Novikov (2000) 20 Minimizing parallel decision making authorities 39. A second tension that arises in sharing responsibilities is that between the "professionals" versus the "politicians" and the degree to which service delivery is managed through a parallel system as opposed to local governments. Those who make their careers in particular sectors, such as education and health, and who have specific views on the needed substantive sectoral reforms have understandable concerns both about the capacity of local governments to pursue sound sectoral policies, and the extent to which sectoral goals may be compromised as a result of political forces at the local level. Studies in Fiszbein (2001) found that the degree of controversy in this respect is broadly related to the specifics of how decentralization has evolved. Resistance in some countries (Romania, for example) to local elected authorities playing a an active role in management of education is in part a reaction to a perceived lack of political reform. "It is hard not to sympathize with concerned education reformers resisting a movement that may make schools hostage to non-reformed political bosses. At the same time it is hard to see how the stated goals (particularly efficiency and quality) can be achieved in the long run if education management works in a parallel but unengaged way with local governments."24 40. Addressing these tensions requires careful consideration of the particular roles for each level. Central government has a particularly important role to play in setting realistic goals for the sector and a strategy for achieving them. It also has important role in setting standards and ensuring quality, monitoring performance and outcomes and ensuring fairness. The central government must make basic decisions concerning national level institutions Finally it may have a role to play in providing financing for financing public goods that are of a national priority. When this is the case it is important for the method of finance to encourage efficiency at lower levels. Regional level governments have a role to play in determining "economies of scale" decisions for regional level institutions (whether to open, close or consolidate institutions across small jurisdictions and addressing spillover effects at the lowest levels. At the local or service delivery level (school, clinic) responsibilities should focus on effectively managing resources to deliver services to the specified standards in the most efficient and equitable manner possible. Beyond specifying which level is responsible for what function, it is necessary to make the responsibilities for regulation, financing and provision very clear. This needs to be done on a sector by sector basis. Strengthening checks and balances 41. In addition to specifying the powers and responsibilities among the different actors, a systems of checks and balances also needs to be considered. Such checks and balances draw upon different stakeholders. These would include: the government and professional groups, who provide rules and standards; the private sector, that potentially provides competition; and taxpayers, users and beneficiaries, who press for better service and effective use of resources. In the Baltics and CEE countries, mechanisms for checks and 24 Fiszbein (2001) page 8. 21 balances are starting to appear through both the private provision of public services and the input and feedback of those who receive services. In other sub-regions, governments tend to rely almost exclusively on hierarchical rules as incentives. This often limits the desire and the ability of those closest to beneficiaries to exercise any authonty and effectively removes any scope for initiative. Regardless of the final distribution of power between levels of and agents of government, an effective system of shared responsibilities requires the presence of checks and balances. Delineating boundaries for the public sector 42. Subnational governments also continue to provide, in some cases, direct subsidies to local agricultural and industrial enterprises. This subsidization of the local economy is one aspect of a much broader issue affecting subnational expenditures and finances, i.e. the relationship between subnational governments and the enterprise sector. This relationship has changed dramatically over the past decade with developments such as privatization, divestiture of enterprises' social assets and growth in the number of new "public enterprises" (often joint ventures of some kind) involving subnational governments. Unfortunately, the impact these developments have on subnational roles is not very transparent. The divestiture of social assets (more appropriately labeled liabilities) and reductions in recurrent social spending by enterprises have had a substantial impact on subnational expenditures. With respect to some areas of social spending inherited from the enterprise sector (e.g. housing, recreation and culture), the issue may be less one of determining an appropriate level of government for those expenditures than it is establishing an appropriate boundary on public sector responsibility per se. However, to the extent financially strained subnational governments do take on additional expenditure obligations transferred from the enterprise sector, this complicates subnational governments' efforts to finance other activities accepted as legitimate local public services. In addition, local government involvement in areas that are not strictly related to the provision of public goods, may hinder the degree to which private sector entities, especially small and medium scale ones can compete, and therefore may hinder not only the potential for contracting out the provision of public services, but also the development of private sector entities. 43. Transition countries typically have legislation that generally specifies levels of government that have responsibilities in given areas. Typically it does not go far enough in specifying the details of each level's involvement. As structures of government and the intergovernmental system evolve, it is important to refine expenditure assignments so as to better support better service delivery of outcomes. Ideally all stakeholders would be brought together to contribute to these revisions. First, overall objectives should be agreed upon. Second, the continued scope for government involvement should be evaluated. Third, should be a detailed evaluation of who actually does what both formally and informally and what that implies for the incentive to meet the stated objectives. Fourth should be consideration of realignment of responsibilities to make sure that responsibilities are matched with both decision-making authority and accountability. Responsibilities for regulating, financing and provisions should be clearly established. Efforts should also be made to develop systems of checks and balances, including alternative sources of service provision (both private sector and by NGOs) and in terms of beneficiary involvement. The 22 specifics will vary by sector, but deliberate efforts to manage shared responsibilities more effectively, can do a great deal to strengthen the effectiveness of decentralization efforts. CHALLENGE 4: AUGMENTING REVENUE AUTONOMY 44. An important notion underlying the effectiveness of decentralization is the link between revenues raised at the local level and the services provided at that level. Local citizens express their preferences and (in theory) are willing to pay for these services with their taxes. Because local citizens pay for these services directly, they have a much greater interest in making sure they are used well and in holding the local government accountable. However, while it is easy to determine services demanded, it is much more difficult to define revenue sources that are appropriate for lower levels of government. In the transition countries, even those taxes that are considered to be appropriate sources of revenue for local governments are difficult to implement. As a result, one of the underlying premises of decentralization - that local governments have control over their revenues at the margin - does not generally hold. If a decentralized approach to service delivery is to be effective, revenue autonomy of local governments needs to be augmented. Revenue assignments 45. A large literature on revenue assignment highlights some basic principals of tax assignment for local governments and the limitations that these imply for sound local sources of finance.25 Wherever possible services should be paid for by those who benefit from the service, which implies that user charges should be high on the list of potential sources of finance. Where such charges are not possible because of the nature of the service, local governments will need to finance service provision through their general revenue base. However, for efficiency, and equity reasons the number of tax instruments available to local governments may be limited. Ideally, subnational governments should have access to relatively stable sources of revenue and whose burden is not easily exportable beyond local boundaries. To minimize potential tax distortions a high degree of national uniformity seems desirable with respect to the corporate income tax and the value added tax, makmg these most appropriately national taxes. It may be desirable to maintam personal income taxes at a national level because of their buoyant nature and usefulness in affecting national fiscal policy, however, in may countries, income tax surcharges are a useful mechanism for giving discretion to local governments over part of this revenue source. Excise taxes and single-stage retail taxes are frequently thought to be appropriate taxes for local governments, although frequently national governments are reluctant to give up these taxes when they are a significant revenue source. One exception to this is the motor vehicles tax, a portion of which is frequently assigned to local governments. The property tax is recognized in many countries as an important source of finance for local governments, however, major problems associated with implementation and valuation of property values, imply that it may be a difficult tax to administer effectively particularly in countries with only nascent property markets. Finally, most localities are assigned a number of specific local taxes (such as dog taxes, gambling taxes) over which they have 25 See among others Bahl and Linn (1992), Shah (1994), and Spahn (1995). 23 discretion, but whose revenue take is relatively low and may not even outweigh the administrative cost to collect them. User Charges 46. Although there is little systematic data available on the prevalence of user charges and degree of cost recovery across the region, there are a few general points that can be made. In the countries furthest along in intergovernmental reforms (the Baltics, Czech Republic, Hungary, Poland), local governments are increasingly provided with the autonomy to contract out services and to determine the required fees for managing these services. They still may be confronted with difficulties concerning the appropriate method for price determination and the ability to pay for fees which may require special assistance. In these countries an additional issue is setting the legal and regulatory framework and providing the necessary information base for effective pricing of such services. In the other transition countries, the pricing of such central services continues to be controlled by the central government and is the focus of important political debates. In Ukraine, the ability to impose user charges at a cost recovery level was a significant point of conflict between the Executive branch of the government and the Parliament and decrees limiting the ability of utilities to raise prices above a certain level have been brought before the Constitutional Court. Restrictions on the right to employ and determine user charges have a broad impact on the possibility of local governments and municipal enterprises to provide services effectively. In effect little progress has been made in these countries (for all levels of government) in adjusting prices for public services in traditional public service areas such as housing, water, heating and transport. Own revenues 47. In the strict sense of the term, revenue assignments barely exist in the transition countries, for the simple reason that subnational governments have little or no discretionary control over either the definition of tax bases or the setting of tax rates. Autonomy is determined by the extent to which localities have any influence over the taxes they receive. The OECD has recently developed a classification for identifying the tax autonomy of lower levels of government based on the extent to which a local government has authority to set the tax rate and base and the degree to which it influences arrangements for tax sharing. They have applied this classification to selected OECD countries and to six transition economies (Table 3). Categories A-C in the table are considered to be "own revenues". Hungary and Poland control the rates of between 40 and 50 percent of their taxes, whereas the Czech Republic, Estonia, Latvia. and Lithuania have little or no autonomy over their revenues. 48. For comparison, it is interesting to consider the degree of "own revenues" classified in this way in other OECD countries. In all countries but Austria, the local governments have some degree of autonomy (setting the rate or the base) of at least 50 percent of its revenues and in most cases considerably more. For intermediate levels of government the situation varies considerably. Some have authority over their revenues (regions in Belgium, 24 Finland, Sweden and counties in Denmark). Some have a role in agreeing the sharing rate (Lander in Germany) and some have no autonomy at all (regions in Portugal). Table 3 Taxes of Subnational Government by Type of Tax Autonomy Country Subnat. Taxes Sub-central government tax revenues by type of "tax autonomy" as percent of total taxes A B C Dl D2 D3 D4 E Czech Republic 11.1 2.7 5.6 91.8 Hungary 10.4 49.2 50 8 Poland 8.3 41 9 0.6 57.6 Estonia 16 2 9 2 90.8 Latvia 17.1 100 Lithuania 22.0 100 Austria 19 Local 8 9 11 81 Lander 10 2 98 Belgium 28 Local 6 13 84 2 1 Communities 13 3 97 Regional 10 8 92 Denmark 31 Municipalities 22 96 4 Counties 9 93 7 Finland 22 Local 22 89 Regions 0 100 Germany 29 Local 7 1 52 47 Lander 22 100 Portugal 6 Local 3 49 14 37 Regions 2 100 Spain 1 3 Local 9 33 51 16 Regions 5 15 7 78 Sweden 32 Municipalities 22 4 96 Parishes 0 2 98 County councils 11 100 Switzerland 38 Communities 16 97 3 Cantons 22 89 1 6 5 A= sub-central govemment (SCG) sets basc & rate B SCG sets tax ratc only. C SCG sets tax base only. Dl. SCG determines revenue split. D2 Revenuc split only changed with consent of SNG. D3 revenues split fixed in legislation, may be unilaterally changed by central govemment D4 revenue split detemined by central govemment as part of annual budget process E Central Govemment sets rate and base of SCG tax Transition country data from 1999. OECD data from 1995. Source OECD (1999) and OECD (2001) 49. Ideally, survey data along these lines would exist for all transition countries, but unfortunately this is not the case. However, the data available from country reports give some indication of the degree of own revenues (again meaning the taxes over which localities have the authority to control the tax rate or base) for a broader selection of 25 transition countries (Table 4). In this table, percentages reflect the share of own revenues in total revenue and grants (not as a percentage of total taxes as in Table 3) There is broad correspondence with the OECD results. In addition, the Kyrgyz Republic, Russia, and the Slovak Republic have autonomy over 15 percent of their revenues. Not a significant portion, but better than the negligible amounts in other countries. Many countries also receive a significant portion of "own" revenues from non-tax sources as seen in colunn 3. In Croatia, Czech Republic, Poland and Romania over a third of total revenues and grants comes from "own" tax and non-tax sources. Other transition countries rely almost entirely on shared taxes and grants. 50. The taxes assigned as subnational taxes in countries with own revenues vary. Both Russia and the Kyrgyz Republic have introduced subnational sales taxes with optional maximum rates, (however, such sales taxes can conflict with national VAT s). Russia also has an optional surtax on the corporate income tax up to a maximum rate. Those countries in the CEE with higher own revenues often have business or communal taxes (Hungary) and most are in the process of developing market-value based property tax systems, although the development of these systems takes quite some time. Croatia allows a subnational surcharge on the personal income tax in cities greater than 40,000. Table 4 Revenue Autonomy in Transition Economies "Own" tax Shared Taxes Nontax revenues Grants and Year revenues Transfers "Own" nontax revenue Other nontax revenues Albania* 3 2 95 99 Armenia 3 55 24 18 99 Bulgaria 0 52 11 37 98 Croatia 0 1(a) 52 38 10 99 Czech Republic 4 42 32 22 99 Estonia 1 83 5 II 98 Hungary 22 15 3 60 98 Kazakhstan 3 86 14 -3 c 99 Kyrgyz Republic 14 23 13 50 99 Latvia 0 2(b) 60 10 30 99 Lithuania 3 63 5 29 97 Moldova* 5 91 3 99 Poland 22 24 14 2 38 97 Romama* 42 41 i7 99 Russia 14 50 6 30 98 Slovak Republic I 25 13 47 98 Slovenia 0 4 41 37 22 98 Ukraine* 3 83 14 98 Sources- See Annex Table I* Bold indicates sum of "own taxes and own fees". a. Cities with populations above 40,000 are permitted to levy a personal income tax surcharge of up to 30 percent on the 25 percent share they receive from the national tax. b. In 2001, the property tax will become an "own tax" (localities will have rate setting authority. C. The negative number reflects the net transfer from the center to the regions. 26 Shared Taxes 51. In many countries shared taxes consist of over 50 percent of local government revenue. The main taxes shared by the center with subnational governments, usually on a derivation basis are the personal income tax, corporate income taxes, excises and in some cases VAT. A key issue for tax assignment is its role in encouraging accountability at lower levels of government and, even with appropriate assignments, tax sharing arrangements serve to mitigate the degree to which local taxation has this effect. Tax sharing has advantages in its administrative simplicity and ability to direct a significant quantity of financing to subnational levels of government, as well as giving central government the macroeconomic comfort of controls over revenue. In addition, many governments view tax sharing as a way to provide an incentives for increased revenue collections overall. However, there are some very serious shortcomings. 52. The mechanics of revenue sharing are straightforward enough, but they encourage lobbying and negotiation. Subnational governments are allocated a specific share of collections from the major tax instruments. In some countries, local agents of the central government collect taxes, give local authorities their approved shares and pass the rest up the administrative chain (e.g. Ukraine, Russia). In other cases, tax collection is centralized, and subnational tax shares are passed down from central authorities (Albania, Croatia, Hungary, Poland). Bases and rates for these shared taxes are in all cases determined by the central government. Subnational governments have very few formal channels or opportunities for increasing official budget revenues. The main channel is by sending representatives of subnational governments to lobby long and hard for a more generous allocation as part of the annual budget cycle. 53. Even where the legal and administrative capacity exists, local governments lack the proper financial incentives to mobilize their own revenues. This may well be the most critical and fundamental shortcoming of revenue sharing as practiced in many transition countries. Because revenues generated by subnational taxes and charges are often factored directly into the calculation of any "fiscal gap" covering subvention and setting of sharing rates for national taxes, subnational governments derive no net financial benefit from mobilizing their own revenues. Shared revenues can also create powerful incentives for local and regional officials to pursue informal and nontransparent methods for generating additional discretionary sources of revenue. 54. Instability and uncertainty are also an element of tax sharing arrangements. One source of uncertainty is the unpredictable annual revision of sharing rates to suit either central authorities' revenue needs or possibly the shifting political influence of some particularly powerful region or set of regions. Evidence from Kazakhstan, Georgia, Ukraine, Bulgaria and shows that sharing rates do change from year to year, though some countries have tried in recent years to provide more stability (Hungary, Kyrgyz Republic, Lithuania). Effective budget planning, formulation and execution require predictability and stability of revenues. Subnational governments in countries relying heavily on revenue sharing have largely been deprived of these two essential features. 55. A final shortcoming of a heavy reliance on tax sharing is its distributional impact. Because most shared taxes are allocated based on where the taxes were collected - systems 27 that employ significant shared taxes can find that revenues dispanties can be substantial. Those localities that have the weakest revenue base, will be the worst off and other forms of redistribution will need to come into play. Shared taxes are effectively transfers, but they are likely to need to be complemented by equalization schemes. Strengthening the link between revenues and services provided. 56. The limited amount of revenue autonomy at the local level is perhaps the most significant signal that countries have not really decentralized, but rather have deconcentrated or delegated responsibilities. Given the legacies of central planning and weak financial and administrative capacities in many countries, it will take many years before localities will be given responsibility for raising their own revenues. Nevertheless, if governments are serious about improving service delivery through decentralization, at some point the link between services delivered at the local level and the funds necessary to provide those services will need to be made. In the meantime, what steps can be made to pave the way for greater revenue autonomy in the future? 57. First, at the national level, countries that are interested in decentralization can lay the groundwork by improving the overall tax framework. Many transition countries have very high marginal tax rates and this discourages local governments from even making use of own taxes that they may have now. High overall tax rates have been an key factor in the unwillingness of mayors in Hungary to use the full complement of local taxes that they have at their disposal. In effect if governments hope to encourage local accountability, then they will need to make the tax room available for them to do so. Also, given the importance of maintaining overall macro economic stability, it is fundamental for changes in subnational tax assignments to be considered as part of an overall tax package. 58. Second, there is a substantial degree of preparatory work that needs to be undertaken to support localities in their ability to levy taxes. While it is not necessary for local governments to necessarily administer a tax for it to be local, there is a great deal of work to lay the groundwork. For example, development of a market value-based property tax requires the establishment of a cadastre and all the other requisition information. In many transition countries it takes time to bring together the various parties that need to coordinate and implement such programs. The greater the degree of preparatory work, the greater the likelihood of effectiveness when systems are implemented. 59. Finally, revenue autonomy can probably best be increased by starting simply and modestly. A surcharge on a pre-existing tax within an upper and lower band can draw on already existing administration and collection capacity. As administrative capacity and accountability develop, localities can take on more sophisticated approaches to revenue autonomy. CHALLENGE 5: CLARIFYING TRANSFERS 60. Because of limitations on effective local tax handles, it is rarely possible for local levels of government to raise sufficient revenues even when revenue assignments are 28 sound. As a result, a well designed system of intergovernmental transfers is a critical component of a decentralized system. The reliance of grants in the transition countries is highlighted in Figure 6, which shows both shared taxes ( a form of transfer as discussed above) and grants as a share of total revenues. While many countries rely on shared taxes for the bulk of their revenue, transfers also provide a considerable source of financing. The legacy of the period of central planning is one of negotiated, untransparent methods for determining how transfers were allocated. While some transition countries have made good progress, improving the transparency, predictability and incentives created by the existing systems remains one of the principle challenges of fiscal decentralization. Figure 6 Shared Taxes and Grants as a Share of Total Revenue (percent) Albanma* __i Armeniailni; t_ - BulgaraI _ Croatia- a Czech Repubi c Estonia Hungary I_! Kazakhstan Kyrgyz Republic = Shared Taxes Latvl____ _ -, * "= il=" i _ _ IR~~~~~~UGrants L.t u. . .... Z Moldova* Poland __ Romanma- Ru s sila _; Slovak Repubic SloveniaP _ Ukraine* 0 20 40 60 80 100 Design of transfer systems 61. Intergovernmental grants are widely used to address a number of important fiscal policy objectives. Furthermore, there are a variety of different types of grants, some of which are better suited than others for achieving specific objectives. Unconditional grants 29 from the central government are the preferred mechanism for addressing vertical imbalance. (Note that revenue sharing is effectively an unconditional grant). Unconditional grants are also viewed an option for equalizing or reducing horizontal disparities in net fiscal benefits across jurisdictions. 62. In addition to addressing fiscal gaps, intergovernmental transfers can also be employed to promote economic efficiency by encouraging subnational governments to mcrease their supply of services with positive "spillover" effects. For this purpose, open- ended matching grants are recommended. Grants can also support a decentralized approach to providing some minimum national standard for a given service. In this case, the standard taxonomy of grants offers the choice of a conditional non-matching grant. hi fact, grants basically of this type are in use in certain transition countries. Conditional matching grants are recommended for encouraging subnational governments to devote some of their own financial resources to a service area of especially high pnority to the central government, but for which local governments might assign a lower degree of prioritization. In addition to the objectives noted thus far, intergovernmental grants can also serve important political functions. Sharing some part of revenues on a derivation basis can help satisfy regional claims to funds raised in their jurisdiction. Meanwhile, specific grants can be (and are) used to quell separatist inclinations. Overall, the range of options can be confusing. As one author puts it, "The design of an effective transfer system is a complicated task. Part of the difficulty comes from the fact that there are a variety of objectives that a transfer system generally addresses: regional equity, achievement of national goals under decentralized provision of services, correction of interjurisdictional spillovers, or simply to fill the gap resulting from the asymmetric assignment of expenditure responsibilities and revenue powers among levels of government. Transfers that are appropriate to further one objective may complicate achievement of another. For these reasons, transfers systems in most countries include a variety of transfers, rather than just one." (Stein, 1997) 63. Whatever combination of grant types a country chooses to use, there are some key criteria against which systems should be measured.26 These criteria include: Autonomy - Provide some sources of finance with little or "no strings attached" through formula-based revenue sharing and block grants. Also, grant amounts should not be reduced as the result of additional effort by subnational governments aimed at mobilizing local revenues. Revenue Adequacy - The volume of financing from grants should be significant enough to provide meaningful assistance for the recipient to carry out expected functions. Equity - Transfers should vary positively with subnational fiscal needs and negatively with subnational fiscal capacity. 26 See Bahl and Linn (1992), Ma (1997), Shah (1994). 30 Predictability/Stability - Parameters determining size of grants for individual recipients should be fixed or stable for multi-year period to promote sound budgeting practices. Efficiency - The design of grants should be such that they are neutral with respect to a subnational government's choices regarding of resource allocation to different sectors or different types of activity. Simplicity/Transparency - Grant distributions should be based on objective and easily observed factors. Proper Incentives - Grants should encourage sound fiscal management and avoid financing deficits. In other words, grants should not create a disincentive for local tax effort or expenditure control. Safeguarding grantor's objectives - The system in place should permit for effective monitoring of actions by all levels of government, progress reviews, technical assistance as necessary, and the use of selective matching transfers. Transfer systems in transition countries 64. Over the decade efforts to reform the system of grants have been under way m most transition countries. Although individual systems are highly idiosyncratic and therefore difficult to compare, Table 5 summarizes some basic characteristics of transfers in selected transition economies. 65. Conditional current grants. Most countries for which data are available have some type of conditional current grant, in which resources are provided for a specific use. These grants usually take the form of an earmarked transfer for a particular, such as education, health or social assistance.. In the past, allocations for these current expenditures were focused on inputs and the funding required to keep existing institutions functional. Many countries are now moving toward capitation grants which provide a certain amount of funding per beneficiary, but these are not typically done on open-ended reimbursement basis. In 2000, the Russia Federation created the Compensation Fund in order to provide transfers to finance federal mandates such as child allowances and disabilities. There is also a proposal to create a conditional matching grant scheme for the social sectors in which conditional financing would be provided, but regional governments would need to match these expenditures. The idea behind this is to assure that regions use the transferred resources in priority areas, but putting additional resources in to support these areas as well. In Czech Republic, conditional current grants are provided on a per capita basis to fund delegated responsibilities in education, health 31 Table 5 Types of Transfers in Selected Transition Countries Country Conditional Conditional Unconditional Equalization Comments Current Capital Transfers (block Transfers Transfers grants) Albania Yes Yes Yes No Conditional grants not matching, based on use Formula for block grants, which might have equalizing component under consideration as of 1999 Armenia No No Yes Yes Equalization formula consists of item such as per person income, border region, earthquake zone, mountainous region, etc Bulgaria Yes Yes Yes Different mechanisms exist, but Ministry of Finance typically either pays off all unpaid bills or provides general subsidies to do so toward the end of the year Croatia Yes (ad hoc) Yes(ad hoc) Yes Yes For counties on whose consolidated revenues are below the per capita average a grant from the state budget is provided at the level of the difference between the actual per capita revenue and 75% of the state average revenue per capita The grant cannot be given to a county on whose territory the rate of surcharge on income is below 1.0 and tax rates and the amount of tax are lower than the highest stipulated rates A similar formula applies within counties. Czech Yes Yes No No Conditional current include earmarked subsidies on Republic per capita basis for delegated responsibilities (education, social assistance) and grants that are received through application and often require matching component Conditional grants for investment grants typically require matching funds The Environment Fund also provides matching investment grants Estonia Yes Yes Yes Yes Formula: Tn = (m*ak-an) 0 9*cn, where Tn = Subsidy appointed to local govt, m = Subsidy level coefficient denved from size of subsidy fund, ak = Average level of appropriations and receipts to the budgets of local govt units in Estonia allotted out of state taxes in the budget year(per capita), an = level of appropnations and receipts to budget of local govt unit, allotted out of state taxes in the budget year (99) in Kroons per capita, 0 9 = revenues are supplemented at amount of 90%, cn = population of local govt unit Hungary No Yes Yes Not specifically Normative grant based on extensive formula with some equalizing elements Targeted investments require process of application and usually matching funds Deficit grants for those who suffer shocks or deficits "through no fault of their own." Kazakhstan No No Yes Yes "Fraternal" system introduced in 1999 budget formulations- simultaneously determines I subventions to poorest regions and wthdrawals from nchest regions Transfers are general purpose with minimums for education and health The methodology used in resolution 529/99 takes into account Regional expenditure differentials which arise from natural and climatic conditions The ggal is to bring range of per capita expenditure amonq oblasts within range of -10% to +15% of the I average in 3 years' time. Not fully implemented in the 2000 budget System under revision. Kyrgyz Yes No Yes Yes Conditional current grants for education and health, Republic calculated on per capita basis with adjustment ' coefficients Equalization grant has no formula, but is gap filling Negotiated with Ministry of Finance 32 I Table 5 Types of Transfers in Selected Transition Countries-Contd. Country Conditional Conditional Unconditional Equalization Comments Current Capital Transfers (block Transfers Transfers grants) Latvla No Yes Yes Yes Formula implements equalization of both revenue capacity and expenditure need Also equalizes across all levels simultaneously Fraternal system with funding from wealthier local governments and from central govemment Negative transfers mostly from Republican cities Distributions are lump-sum and unconditional Lithuania No Yes Yes Yes Unconditional grants exist for i) reserves for extraordinary events, ii) equalization of tax revenues and ia) equalization of expenditure needs 1) allocated based on claims at end of year, ii) allocated based in order to bnng forecast tax revenues below national average closer to average, and iii) allocated on "fratemal system" based on relative measure of need adjusted with demographic coefficient Poland Yes Yes Yes Yes Equalization grant brings localities up to 85 percent of national average per capita revenues Romania Yes Yes Yes Yes Formula for equalization in place based on expenditure needs & revenue capacity Russia Yes Yes but not Yes Yes Formula has weighted index of expenditure needs systematic and fiscal capacity Ukraine No No Yes Yes Equalization based on measure of expenditure needs minus revenue capacity Formula includes extensive set of sectoral coefficients To be applied to oblasts and within oblasts Sources. See Annex Table I social protection. In Kyrgyz Republic, conditional grants for education and health are provided on a per capita basis, taking into account adjustments for geographical area, mountainous regions, etc. Countries which do not have conditional current grants include: Armenia, Hungary, Kazakhstan, Latvia, Lithuania and Ukraine. In Hungary a formula for determining normative grants includes an extensive list of expenditure needs, but there is no requirement that resource be allocated to the normatives in the formula (though there is some pressure from the different sectors to do so). In many countries, expenditure needs are incorporated into an equalization formula, which will be discussed below. 66. Conditional Capital Grants. Although the amounts are small, virtually every central government in the region provides some grants to lower level governments for investment. However, it is rare for a country to have a systematic method for doing so. The much more common situation is that line ministries: education, energy, transport, etc. are given a particular amount of resources, which they then allocate to lower levels of government. In the more advanced countries in the CEE countries, as in Hungary, rules are in place which provide for the submission of applications, presentation of feasibility studies, and then a comparative evaluation of projects within and across sectors. In these cases localities almost always need to provide a matching component. However, implementation has not fully caught up with the rules on the books. Often, there are many source of funds for investment to which localities can apply and this can lead to a type of "grantsmanship" in which localities spend a good deal of effort filling out applications for investment grants. With coordination among ministries lacking, those localities that are effective can sometimes acquire funding for the same project from different sources, whereas other localities may not be able to attract any resources at all. In the Caucasus, Central CIS, and Central Asian Economies there appear to be few specific methods for allocating investment 33 grants. In these cases, allocation for investment is ad hoc, carried out through lobbying and negotiation. 67. Block grants and equalization. Almost all countries for which data is available (except for the Czech Republic) use block grants to provide for equalization. In light of the fact that much of the recent reform effort in the transition countries is specifically concerned with horizontal equalization, it seems worthwhile to consider some of the guidelines specifically related to the design and operation of equalization grant systems. The design of an equalization grant involves two key aspects, determining the size of grant pool and establishing criteria for the distribution of grants to individual subnational jurisdictions. There are several options available for each of these aspects. 68. Assuming the grant pool is to be funded, at least in part, "vertically", i.e. with a contribution from the central government, the criteria of predictability and transparency argue for financing the pool through a fixed share of some national tax or taxes. If fiscal conditions make this too constraining for central authorities, then the preferred option is to determine the size of the pool through an open and transparent process of negotiation involving all parties. 69. With regard to the allocation of shares across jurisdictions, the preferred option is a relatively simple and transparent formula taking into account both expenditure needs and revenue capacity, whenever reliable data permit both to be incorporated. To prevent perverse strategic behavior by potential recipients, the formula should incorporate factors over which jurisdictions have relatively little or no control. There are many possible variations with regard to components of the equalization formula. To begin with, there is the question of choosing a standard against which equalization takes place and deciding on the degree of equalization to the chosen standard. This decision will necessarily reflect political and financial constraints. The choice of factors determining expenditure needs should be parsimonious. 70. On the revenue side, data limitations make it difficult to get good measures of fiscal capacity such as tax base information. Often, formulas have to try to use actual tax collection data. Depending on how these actual collections data are used, this gives rise to a higher risk perverse strategic behavior, including a disincentive for revenue mobilization. Recognizing this risk, some authors recommend incorporating a measure of tax effort. However, in the transition context, tax effort is often viewed as less relevant, since as seen above subnational authorities have little discretionary taxing authority. 71. Table 5 indicates that there are a wide variety of approaches taken in transition economies. In terms of determining the overall resource pool, this is most commonly determined in the annual budget law. In Hungary, the normative grant is tied to revenues from the Personal income tax. With respect to determining the allocation of funds among jurisdictions there are a variety of approaches have been pursued. Armenia employs a system based on estimated expenditure need per capita, adjusted for certain characteristics, such as border regions, earthquakes, etc. Hungary, Romania, Russia and Ukraine, all employ a formula that takes into account expenditure needs and fiscal capacity, and in which all funding comes from the center. In these systems there is typically a problem with having a substantial number o f indicators of expenditure need, often over 50, which makes 34 the formula, more complicated, less transparent and harder to calculate. Over time countries are making efforts to improve and rationalize, as in Russia, but typically the first steps in developing these formulas create systems that are too complex. Kazakhstan, Latvia and Lithuania also rely on systems that use a measure of expenditure needs and revenue capacity, but their systems are "fraternal" in which the wealthier regions or Republican cities transfer resources to the less well off in order to promote equalization. 72. Croatia and Poland use a slightly different approach, they calculate the average income across regions and/or counties and then transfer resources to those with lower income in order to reach a certain proportion of average income. In Croatia, counties are brought up to 75 percent of the national average, provided they are using all the revenue instruments at their disposal. In Poland, the equalization system brings per capita incomes of jurisdictions up to 85 percent of the national average. 73. Finally there remain countries, in which there is no specific approach to equalization and negotiated transfers remain the norm. These include Azerbaijan, Belarus, and some of the Central Asian economies including Tajikistan and Turkmenistan. Further strengthening is needed 74. While there is no doubt that a great number of actions and efforts have been taken to transform the system of transfers in most transition countries, clearly more needs to be done to make sure that these systems create the correct incentives and have the desired economic impact. First, what is fairly basic, but often disregarded, is the need to use the mechanism created as the basis for determining the resources transferred. Countries will go to great effort to create formula based systems and then carry out actions which completely undermine these systems. For example, Bulgaria has created a formula for allocating resources, but typically, towards the end of the year it provides and extra, unplanned transfer to localities to finance unpaid bills. This action i) indicates that the formula itself must be inappropriate and ii) creates an incentive for all jurisdictions to stop paying their bills and to anticipate that additional resources may become available towards the end of the year. In Kazakhstan an new system for formula-based equalization transfers was devised and introduced in 1999, but "negotiations around the edges" undermined the formula, and the approach and in 2000 it was not applied. The Government is now in the process of redesigning the system. The crucial point is that once an approach for allocating resources is agreed upon, it is critical that it be used and there not be other ad hoc elements for providing resources, that undermine the incentives of the system. 75. Most countries in the region could benefit from stepping back and thinking through the total effect of the transfer system that is evolving. Typically, grant system are developed in a rather ad hoc way over time. Conditional grants are created to serve one purpose, a range of capital grants mechanisms exist, equalization adds yet another component. As was discussed above typically one grant mechanism is used to address a particular purpose or objective, but it may also have indirect impacts on other systems. Recent proposals in Russia set out two separate mechanisms for conditional targeted grants (with some equalization elements), a separate equalization elements, an investment grant window and a window to reward regions pursuing fiscal reform. As systems with multiple 35 transfer mechanisms develop, it is quite important to assess their overall impact and whether those remain with the overall objectives of the system. 76. Additional efforts are also needed to improve the quality of data underlying specific transfer mechanism, the ability to monitor and the transparency of the system as a whole. In countries that are less far along in the development of transfer mechanisms, it can be quite a challenge to collect the information needed in a consistent and coherent way. Agreement on a formula or capitation grants require adequate information to determine expenditure needs in a manner that focuses on resources required to provide the desired outcomes, as opposed to being based on input requirements which has traditionally been the case. Developing an ability to monitor the use of funds is an important element of conditional grant systems, otherwise the "conditions" become meaningless. Finally, a key element of developing formula based system is to provide local governments with the predictability and stability that using such a system can provide. A prerequisite for this, however, is that the system be transparent. It should be absolutely clear how and why resources re allocated to different jurisdictions. Although many countries have worked hard in developing systematic transfer systems, these are typically less transparent than desirable for having the desired impact on incentives. Often lack of transparency comes in the form of excessively complicated formulas, but it also may result from the difficulties localities may have in obtaining the formulas and information used in calculating them. CHALLENGE 6: MANAGING AND MONITORING LOCAL BORROWING 77. There are sound reasons for allowing subnational governments access to financial markets. First, subnational governments often have responsibility for financing capital investments that are large relative to the current revenues. In the absence of borrowing, financing these "lumpy" investments would require an inefficient increase in taxes. Also, as the benefits of these capital investments are typically spread over time, even generations, it is appropriate to accumulate some debt in their financing so that the costs of providing the services are also spread over time. A second reason to allow for subnational borrowing concerns the seasonality of cash flows. If the timing of outlays is poorly matched with the timing of tax revenues, then borrowing can help smooth spending patterns over the course of the year and prevent inefficient shifts in the level of provision for key public services. A third possible reason to encourage subnational reliance on borrowing relates to accountability. Certainly, in comparing the alternatives of grant financing versus borrowing for either capital investment or smoothing cash flows, borrowing is far more likely to encourage greater accountability on the part of the recipient with regard to the level of funding required and how it gets put to use. However, the appropriate conditions for managing and monitoring borrowing need to be in place. Methods of Local Borrowing 78. Accepting one or more of the arguments for borrowing by subnational authorities, the question then becomes how they obtain access to debt financing. There are a variety of options, including borrowing from a higher level of government, borrowing from 36 subnational (or national) public financial institutions, or borrowing from capital markets either in the form of loans or by issuing bonds. All of these options are, for better or for worse, employed in the transition countries. 79. Borrowing from the central government has traditionally been the most widely used option in the transition countries, and this remains the standard approach in many countries where the preconditions for utilizing other sources remain unmet. Even in countries where subnational governments do have access to other sources of finance, short- term loans from a higher level of government remain a common and acceptable source for financing cash flow imbalances over the course of a fiscal year as well as some capital investment projects. Unfortunately, it is too often the case that those loans never get repaid and essentially get converted into grants. 80. Borrowing from state-owned or controlled financial institutions also remains an option in some countries, much to the detriment of sound fiscal management. International experience demonstrates the dangers involved, when political influence overrules proper financial considerations in determining lending practices. This type of borrowing undermines fiscal accountability and can quickly become a threat to macroeconomic stability. 81. A third option is for subnational authorities to have access to capital markets for their borrowing needs. Indeed, the participation of subnational governments in capital markets, mostly domestic but also foreign, has increased rapidly over the past five years. In some cases, where efforts to establish a suitable legal, regulatory and supervisory framework has progressed further, this trend toward commercial borrowing by subnational governments is a very positive development. With it comes the opportunity to significantly increase both the supply of financial resources available to subnational governments and the efficiency and accountability with which those resources are used. More generally, access by subnational governments to well developed capital markets as a means to financing their investments contributes to the broader restructuring of intergovernmental fiscal relations in which the rights and responsibilities of different levels of government are more appropriately delegated and more clearly understood. 82. However, in the absence of both market discipline and the necessary regulatory framework, subnational borrowing can lead to fiscal and economic disaster. Where capital markets are not free and open, where disclosure and transparency are not required, where the perception is that higher levels of government implicitly guarantee subnational debts, and where a clear set of rules and standards pertaining to budget practices, financial reporting or the level of indebtedness is not adequately enforced, access of subnational governments to capital markets may be ill advised. The challenge of establishing a satisfactory set of rules to guide and regulate subnational borrowing is one faced by all transition countries. Even where progress in that direction has been greatest, in Hungary and the Czech Republic, additional measures will still be necessary as market conditions evolve. At the other end of the spectrum among countries where subnational borrowing is already significant and growing but where both market discipline and the regulatory framework have been weak, e.g. Ukraine and the Russian Federation, the rapid growth in subnational borrowing has the potential to destabilize the entire economy. 37 Regulatory frameworks for subnational borrowing 83. There is a critical need for central and subnational authorities to agree on a framework for subnational borrowing, one which facilitates access and participation and reduces the risk of systemic failure. Recent developments in a growing number of transition countries show that an appropriate balance of market discipline, rules-based controls and administrative oversight, complemented by an essential degree of transparency, disclosure and, as necessary, sanctions can create the foundations for prudent subnational borrowing in domestic or international capital markets. 84. A review of subnational borrowing in transition countries reveals a widely varying degree of development that closely reflects the progress of intergovernmental fiscal reformn in all the other areas discussed in previous sections. In many of the countries where reform efforts have been the least extensive, explicit subnational borrowing is quite limited. Financing of subnational governments labeled as borrowing comes from one source, the central government. And as noted above, it is difficult to distinguish these loans from grants, given that repayment is rarely enforced, whether the funds are directed for covering short-term imbalances between current revenues and current expenditures or if the funds are intended to finance capital investments. Given the infrequency with which these "loans" get repaid, it is more appropriate to say that subnational borrowing does not really exist in countries such as Belarus, Moldova, all the Central Asian countries and the Caucuses. Not surprisingly, with the exception of the Kazakhstan and Kyrgyz Republic, these are all countries where intergovernmental reforms have been relatively marginal and decentralization is not a policy priority. 85. There are also cases where intergovernmental fiscal reforms are evident but also noticeably unsuccessful in establishing the foundation for a stable and clearly defined distribution of rights and responsibilities across levels of government. In this context, the growth of subnational borrowing is especially threatening to economic and fiscal performance, as demonstrated by the cases of Russia and Ukraine. The rapid growth of subnational borrowing in these two countries taking place in the absence of a well developed regulatory framework, has exacerbated existing fiscal pressures on both subnational and national governments. 86. At the other end of the spectrum are the Czech Republic, Hungary, Poland, and, perhaps to slightly lesser degree, Estonia, where financial market development has complemented by the establishment of a relatively complete legal, regulatory and supervisory framework to support subnational government borrowing. In these countries, subnational governments have access to both foreign and domestic sources of finance. The extent to which subnational governments utilize their freedom to borrow varies, with Hungary having perhaps the most open access but exhibiting relatively less borrowing activity. 87. The minimal level of subnational borrowing in Hungary, well under 1% of GDP and under 3% of subnational expenditures, is said to reflect the reluctance of local official to incur debt. Only Budapest has issued a bond. Central government regulations establishing a limit of debt service being no more than 70% of yearly own-source revenue is not 38 binding. Municipalities are allowed to use future income streams as guarantee and use municipal "non-core" properties as collateral. Hungary also stands out as the first country in Central Europe to introduce a municipal bankruptcy law that defines debt workout procedures in case of municipal default. This helps provide an effective deterrent against implicit sovereign guarantee of municipal debt. 88. Subnational borrowing has developed quite rapidly in the Czech Republic since 1992, with strong growth in commercial loans and bond issues. Between 1993 and 1996, total municipal debt grew by a factor of ten. Most municipal debt is now long-term. This is an indication of a maturing and stable market, with reduced pressures for rolling over short-term loans helping to reduce interest rates. Over 1000 municipalities have utilized commercial loans, which account for a third of total municipal debt. Municipal bonds now account for the largest share of total municipal debt. Increased subnational borrowing has allowed direct state subsidies to decline substantially. As a share of financing for municipal infrastructure, capital subsidies from the central government declined from 50% in 1993 to 25% in 1996. Currently there are very few restrictions on municipal borrowing. Law allows borrowing in both domestic and foreign markets without any restrictions regarding the purpose or structure of the loan. There have been no defaults reported on any commercial municipal credit, despite the rapid growth in municipal lending. Aggregate debt service ratios remain relatively low in the area of 4%. 89. In Poland, subnational borrowing emerged in 1994 with privatization of regional commercial banks loaning to selected larger cities. Maturities were initially relatively short at 3 to 5 years, but more recently have become longer at up to 15 years for largest and most creditworthy cities. Krakow, Lodz and 25 large to medium-sized cities have issued bonds. Standard&Poor's has given Krakow and Lodz credit ratings of BBB. With regard to commercial banks loans, the most active suppliers of credit are regional commercial banks, lending principally to their main cities. More recently, international investors and banks are entering the market, with both foreign currency and domestic currency loans to larger cities with maturities up to 15 years. Overall, there are signs of matunty in the municipal debt market in Poland, though it remains fairly small in volume with total municipal credit at less than 0.5 % of GDP in 1996. Steps are being taken to improve the legal and regulatory framework, with relevant legislation currently under consideration. Issues to be addressed include a municipal bankruptcy law, imposition of limits on the outstanding stock of debt and debt servicing uncertainty about intergovernmental fiscal relations, budgeting skills and standards, and clarification of property rights for assets of municipal companies. Finally, there is a strong need for improved monitoring and reporting on municipal debt. 90. Estonia too has witnessed a sharp increase in subnational borrowing since the end of 1993,when total loans and bonds stood at EEK 1.5 million. A year later, it was up to EEK 108 million, and by the middle of 1996 had climbed to EEK 900 million, not including EEK 800 million more in on-lent funds from the central government, mostly for energy related expenditures. In sum, subnational borrowing reached 2% of GDP in 1996. The rapid growth in subnational borrowing pushed the debt-to-revenue ratio up to 40% in 1996. There is a growing concern that as debt service costs rise, other essential local public spending will get crowded out. The high level of local borrowing may reflect a perception 39 that central guarantees will protect them from financial burden of repayment, especially with regard to on-lent funds. So far, commercial banks have reported no repayment problems with either loans or bond issues. However, rescheduling was necessary for 90 percent of the amount on-lent from World Bank and EU loans. Repayment requirements for some of local governments were very high relative to local budgets, well above the official limit of 20%. 91. Between the two extremes of well-developed subnational debt markets and complete reliance on soft central government loans are countries such as Bulgaria, Croatia and Romania, where some commercial borrowing activity is evident, with both loans and bonds in use. However, borrowing activity is constrained by a less developed financial markets and overall institutional framework. 92. Although it is difficult to compare across systems, a wide range of countries have been pursuing development of the regulatory framework for subnational borrowing (Table 6). All countries for which the information is available allow local government borrowing domestically and most allow for external borrowing (Armenia, Kazakhstan, and more recently Russia -after significant defaults following on from the 1998 crisis--have disallowed foreign borrowing by subnational units.) In most other countries, lower levels may borrow with the authorization of the Ministry of Finance, but typically the procedures and processes for authorization are not well established. 93. Almost all countries have established limitations on borrowing. Typically these are on debt repayments as a share of local government own revenues and shared taxes and on total debt. In most countries the use of borrowed funds is restricted to capital investment purposes. Finally, as mentioned above, of the countries only Hungary has developed a Municipal Bankruptcy legislation, although both Estonia and Latvia have been developing such legislation as well. 94. Several important observations arise from this brief overview of subnational borrowing and regulatory frameworks.. First, it is clear that in quite a few of the countries, there have been some very dramatic developments in the area of subnational borrowing over the past five years. Over a very short time period, borrowing in the form of loans from commercial banks and bond issue has evolved from a negligible factor into an important source of autonomous financing for subnational government, substituting for the traditional dependence on financing provided by the central government. Even where the volume of funds is not large, borrowing on commercial terms represents a qualitatively different type of financing than the traditional reliance on highly subsidized loans from central authorities. As such, the development of commercial borrowing imposes new demands on both subnational and national governments. Subnational authorities must adapt to a more rigorous standard of accountability associated with financing their activities on commercial terms. National authorities must adapt to the role of an effective regulator, moving away from the more traditional role of directly controlling the finances of subnational governments. 95. In addition a move towards increasing subnational borrowing requires increasingly more effective information and monitoring systems, as well as reporting mechanisms. Even 40 some of the more advanced reformers in the region might be hard pressed to specify total local government borrowing. Old systems make strong Table 6 Regulatory Framework for Subnational Borrowing in Transition Countries. 2. Is there 3. Are 4. Are the 1. Are local governments a there any purpose for 5. Is there allowed to borrow? regulatory limitation which money Purpose regulation for Country framewor s on Limitations can be requirements municipal k for borrowing borrowed bankruptcy? At Home Abroad borrowing 9 limited? 9 Further implementing regulations have never Albania YES 9 YES been issued because local govts do not appear to have takcn on any loans (Banks and Pigey p. 32) The annual interest payment for earlier borrowing shouldn't excecd 20% level of Local budget's reservc fund The latter is reflected in the Armenia YES NO YES YES administrative part of the NO NO local budget. Rescrve fund itself should be limited by 30% of the noted administrative budget These limitations belongs to all types of expenditures Loans cannot be Loans can be raised to used for general Bulgaria YES YES YES YES cI,ovcer only YES purposcs NO "insufficiencies" in the (paying salaries, N municipal budget operating expenses, ctc) I) (Jurlina A p 6) for 1999-- Up to 20% of their own revenues minus transfers received in the previous year 2) Only for (Council p 27-8) In financing capital 1997, according to the projccts and for Croatia YES YES YES YES Law on Execution of YES the coverage of NO State Budget, local units its budget can raise loans on deficit (council condition that their p 27-8) annual loan liabilities do not excced 30% of their budget expenditure in 1996 Czech YES NO YES NO NO NO Republic Y 41 Table 6 Regulatory Framework for Subnational Borrowing in Transition Countries-Contd. 2. Is there 3. Are 4. Are the a there any purpose for 5. Is there Cr1. Are local governments ramewor Limitations which money Purpose regulation for Country allowed to borrow? kf or son can be requirements municipal borrowing borrowing borrowed bankruptcy? borwn ? limited? Local authonties can finance current costs only with short- tcrm loans, which should be paid back by the cnd of the budget year Loans should be used to financc investments forecast in a municipality's development plan. Loans arc primanly meant to finance capital investments to renew the infrastructure (capital spcnding benefits people for ycars to The total amount of loans come and the and other factors with repayments and loan associated cost (e.g interest on the Estonia YES YES YES YES interest) do not excecd YES loans will be NO Estonia YES YES YES YES ~~~~~~~~~75% of last budget's YS paid from taxes N revenue (excluding loans raised in years to and block grafts from the come) Loans state budget.) are inappropnate for a local authority's current spending because using them for current spending would enable the present citizens to shift the cost of the current spending which benefits them, to future citizens who will receive no benefit However, there are cxceptions to the guidelincs E g Local need to undertake current cxpenditure that was unforcsecn, for instancc-- natural disaster needs 42 Table 6 Regulatory Framework for Subnational Borrowing in Transition Countries-Contd. 2. Is there 3. Are 4. Are the a there any purpose for 5. Is there Country frw1. Are local governments ra or Limitations which money Purpose regulation for allowed to borrow? k for s on can be requirements municipal b or borrowing borrowed bankruptcy? borrowing ? limited? Sums for paying back loans and loan interest in all financial years must Hungary YES YES YES YES not exceed 20% of the NO YES revenue of the last accepted budget, loans excluded I Borrowmng in any year cannot exceed 10% of local revenues in that year, 2 Total debt cannot exceed 25% of local revenues in any year, 3 Ls almited to Kazakhstan YES NO YES YES Expenditures on debt YES isimte to NO service must not exceed YE inancimngcptal N 10% of local revenues in any year, 4 Total local programs only borrowing is limited annually in the Republican Budget (Joao p 93) Kyrgyz YES NO ? NO Republic Debt limit (yearly cap) on borrowings and Capital Latvia YES YES YES YES guarantees to be YES expenditures NO detenmincd by Cabinet of only Ministers in process of Budget Negotiations 'according to the municipal budgets must proecdures and Lithuania YES ? YES YES be balanced YES conditions NO establishcd by the govt' A gmina's annual debt servicc could not cxcccd 15% of planned revenues, minus the debt service on loans secured by real property Act on local (showtng that reformers govemmmt gave had not fully appreciated gminas the right that in the municipal to borrow for sector, loans secured by both cash flow rcal property are and investment considered less secure purposes Poland YES 9 YES YES than those secured by the YES (important to NO full faith and credit of the both the issuer). A provision also economic allowed the central govt operations of to guarantee gmina loans, local although procedures for govemments issuing these guarantees and their and the method of their political repayment in cas of independence.) default was never specified. Only a handful of gmmnas has actually applied to the \MoF for guarantees 43 Table 6 Regulatory Framework for Subnational Borrowing in Transition Countries-Contd. 2. Is there 3. Are 4. Are the a there any purpose for 5. Is there Country rmeoLmiao1. Are local governments knf limitation which money Purpose regulation for Country allowed to borrow? frmwr son Lmtioscan be requirements municipal boror borrowing borrowed bankruptcy? ? ? limited? Borrowings by local councils, which, according to the law, may borrow up to a limit Romania YES YES YES YES of the annual debt service ? NO of 20% of their currcnt income (Laurentiu Tachichi's local govemment paper p. 17) Purpose of subnational borrowing Limits to the overall restnctcd to level of debt for regional capitan and local govts imposed gntmextemal by the Budget Code: 1) borrowngxea Budget deficits of local grang govt limited to 3% of privilcges only budget revenues (before Bonfdsemay bevt transfers and loans). 2) iosseds oly for Debt service Issued only for expenditures limited to investment no more than 15% of purposes (as Russia YES NO YES YES annual expenditures (in YES opposed to NO any year). Recently, Law funding for on the financial expend Foundations of Local expendirucs) Govt in the Russian Municdpallye Federation of Sept 1997 owned banks are restricted the amount that banned (in order local govemmcnts may mora hazard ofe borrow to 15% of budget moit baorrd ot expenditur-es [this applies aond lendoing onlyto loca, not regional and lendsng govts] local bonds are not guaranteed byt he federal govennment) 44 Table 6 Regulatory Framework for Subnational Borrowing in Transition Countries-Contd. 2. Is there 3. Are 4. Are the a there any purpose for 5. Is there regulatory liiato Country 1. Are local governments framewor an Limitations which money Purpose regulation for allowed to borrow? k for s on can be requirements municipal borrowing borrowing borrowed bankruptcy? borrowing ? limited? [Municipalities have the right to raise loans without special approval of the state administration body. The towns Municipalities and towns and are not obliged to submit muhecipapotics their budget to the state have supported Slovakia YES YES NO NO admin bodies [Only the NO e t NO use of state earmarked ostabhshemi nt grants can be controlled of their own by state admin bodies bank, the Frst communal Bank in Zilina Municipalities may request loans also from branches of foreign banks in the Slovak Republic] Local communities are allowed to Debt levels havc to kept borrow short- within limits stnctly term to cover Slovenia YES ? YES YES determined by the YES current NO Ccntral govt (exact limits expenditure, not given in my sources while long-term so far) debt can be raised only for infrastructure _financing 45 Table 6 Regulatory Framework for Subnational Borrowing in Transition Countries-Contd. 2.Isthere 3. Are 4. Are the a there any purpose for 5. Is there 1. Ae lcalgovrnmnts regmewory Limitations which money Purpose regulation for Country allowed to borrow? framewor s on Limitations can be requirements municipal borrowing borrowing borrowed bankruptcy? borrowin ? limited? Process of Borrowing from intergovernmental loans banks is used is non-transparent, no only excusely clear govemment or cqusvly guidelines for deciding for liquidity when or how to allocate purposcs, paying intergovt loans, loose wages, and accounting method for smootha g out reporting inter-govt local loans, and unpaid loans expenditures o are de facto subsidies the facc of Officially, there are low unstable transfer requirements (not emstate mentioned cxactly what the State Ukraine YES YES NO 7 they are), and national YES vanable and tied NO acccsibility for loans YS vral n id N from higher levels of to natUonal bank government (the most ofUkinrae's widespread form of dlscount lete borrowing utilized by Overall level of local govemments) MoF such borrowing has dircct relationship rematis with the oblast financial relatively low offices, thcrefore greatcr short term, control on their re time replaymen rates, and repa( d on tmme results in lower percent (commercnal of unpaid balances for banks do not oblasts As of 1998, this make long tel n systcm remains the same loans to any Ientaty) ) I distinctions between taking loans and issuing debt and, the monitoring mechanisms do not always capture different types of local borrowing. Strengthening information and monitoring systems for subnational debt is a critical element for improved debt management overall. 96. Generally, there is a close correspondence between those countries that have been more or less active and successful in supporting the development of subnational borrowing and the countries that have demonstrated greater commitment and success in reshaping other aspects of intergovernmental fiscal relations. Beginning with the significant share of transition countries in which subnational borrowing has not emerged as a significant institution (Belarus, Moldova, Central Asia and the Caucasus), along a continuum through to those countries that have been most active in establishing a stable and effective regulatory framework for subnational borrowing (Hungary, Czech Republic, Poland and Estonia), the survey helps to highlight which countries are most committed to supporting the development of autonomous subnational governments. 46 CHALLENGE 7: BUILDING CAPACITY 97. While one of the common rationales for decentralization proposes that local governments' proximity to their constituents will allow them to be better than central governments at managing resources and matching their constituents preferences, it is not at all clear that local governments have the capacity to translate this information advantage into an efficiency advantage. Inexperienced, small local governments may not have the technical capacity to implement and maintain projects and they may not have the training to effectively manage larger budgets. The legacy of central planning leaves many countries with a limited degree of capacity even at the central level, much less at the local level. Assessing, improving, and accommodating varying degrees of local capacity has become more and more important as countries seek to transfer greater responsibilities as well as budgets from national governments to local governments and communities. Building such capacity is a challenge for those countries who seek to further decentralize. Assessing Local Capacity 98. Measuring local capacity is difficult and there are no ready variables to apply across the transition countries. Some of the basic components of planning, implementing and sustaining basic services include: analyzing and solving local problems; determining community needs; organizing local and national political support for programs; mobilizing resources for programs; raising tax revenues or collecting user fees; writing specifications for the technical elements of programs; maintaining and sustaining the services; evaluating the impact of programs on the local environment; providing for those affected adversely by the program; contracting for services and buying equipment. Uphoff (1997) focuses on four fundamental functions that organizations (and systems of organizations) must be able to do in order to reach their objectives: * decision-making, which includes budgeting, planning and evaluation; * resource mobilization and management; * communication and coordination; and, * conflict resolution. 99. The discussion of revenues above indicated that in some countries local governments have begun to increase their ability to mobilize revenues. Unfortunately there is little data available in transition countnes to help in assessing these other characteristics, such as decision-making and communication and coordination More generally, however, given their history of central planning, most countries have begun with very limited local capacities. Efforts to improve management, build institutions and strengthen civil societies have therefore been under way in most countries in the region. Building capacity 100. The traditional approach to decentralization has been to build capacity through training and technical assistance before transferring responsibilities or revenues. This 47 method was based on concerns about irresponsible spending, local corruption, regional inequities, and service collapse as well as many central governments' reluctance to devolve authority and inability to monitor decentralization. However, it is difficult to expect local governments to change behaviors and respond differently based solely on training and technical assistance. Localities may understand better how to do things to suit their citizens' needs, but if the incentive to implement effective decision-making, resource mobilization, and coordination are not in place, there will be little reason for localities to pursue them. 101. Building capacity therefore requires two key elements: creating the incentives for localities to undertake the actions that support effective capacity, and giving support and guidance as they build their skills at decision-making, planning, resource mobilization, coordination and the like. An appropriate framework for decentralization combined with support to individual local governments can help give local governments the incentive to strengthen their abilities. Rather than plan and make large up-front investments in local capacity building as a prerequisite for devolution of responsibility, it is likely to be quicker and more cost-effective to begin the process of devolution, to permit learning by doing and to build up capacity through practice. The key is to make sure that capacities that are built from the bottom-up are consistent with the strategies of decentralization being proposed from the top down. 102. Before considering key areas in which capacity needs to be built, it is important to recognize that there are two very different types of localities in most transition countries: urban center or municipalities and rural communities. While serving the same ultimate functions, the two types of localities are quite different. 103. Urban centers and larger municipalities have inherited a substantial set of assets and a wide range of economic issues to address as part of the transition process27 In countries that are decentralizing that have also received a substantial portion of service responsibilities. These include sustaining the social safety net, renewing urban services such as the utilities and transport, managing the housing stock and often providing education and health services to the surrounding area. As discussed above, urban centers and larger municipalities typically have to take on functions in which the basic parameters for management are not within their control. For example, they may be responsible for managing basic utilities, with little say in the price that is paid for those tariffs. Or they typically have responsibilities, such as financing social assistance, for which comparative advantage is at the central level of government. Effectively, capacity at the urban level is often overwhelmed by a multiplicity of functions. This combined with limitations on actual decision-making can hinder the development of effective planning and decision-making. In terms of basic capacity, urban centers are often able to attract more educated and skilled individuals and in some countries will at least have the beginnings of strengthened civil society. Hence, the potential for strengthening capacity is quite good, as long as the overall incentives faced by urban areas support those capacities. 104. Rural communities create alternative issues in terms of capacity building. Not only are they likely to have less developed human capital and systems, but they face more 27 See World Bank (2000) From Commissars to Mayors: Cities in Transition Economies. 48 keenly issues related to economies of scale, networks and the ability to coordinate with other rural communities. Rural communities have often come less far in building civil society and strengthening economic and fiscal management. Building capacity in rural communities calls for special focus on issues related to economies of scale, as well as an emphasis on training and assistance to support better economic and fiscal management, decision-making, coordination, communication and conflict resolution. 105. In both urban and rural localities, three key areas in which setting the appropriate framework and supporting capacity building from the bottom up can have big payoffs are: * Basic budget management. This includes: basic elements of formulating fiscal objectives, evaluating tradeoff and creating a budget that will help meet those objectives; assuring a link between current and capital expenditures; developing open and participatory budget processes; strengthening budget execution and building revenue bases. * Contracting out, procurement, fiduciary control. Service management in many areas can be contracted out, although this requires developing the experience and knowledge of contracting and pricing methods. Improved and transparent processes for procurement can improve transparency and accountability. Fiduciary controls ensure that systems of checks and balances are in place with respect to the use of resources. * Monitoring & Information. Building capacity in monitoring and information is key for all levels of government. Otherwise, not only is accountability weakened, but it is impossible to determine whether overall objectives of the system are being met. Community Driven Development as a Tool for Building Capacity 106. Because if their history, most transition countries face a situation in which citizens show great reliance on government structures, but at the same time also show great mistrust of government structures - be they central, regional or local. This poses difficulties both for effective decentralization, which requires citizen oversight, and for capacity building which requires citizen engagement. In recent years, "community driven development" (CDD) initiatives have been proposed as an effective approach not only to achieving the objective of poverty reduction, but in helping countries to develop greater capacity, accountability and social capital at the lowest level of organization. In many ways, CDD efforts, which build skills and capacity "from the bottom up," are an important complement to decentralization policies "from the top down". However, there are some critical elements that are necessary for this type of capacity building to effectively support decentralization efforts. 107. CDD efforts are designed to focus on communities, which may be geographically based or organized around the provision of some function, such as water user associations, or condominium associations. The intent is to empower communities by channeling their demand, promoting participatory mechanisms, facilitating access to information and assuring social and gender inclusion. CDD projects try to promote the key elements of capacity building discussed above at the community level. 49 108. The benefits of CDD are that it can help to promote institutional change and support leaming by doing in many areas that are critical to effective decentralization (decision- making, voice, improved social capital). By strengthening community groups and their capacity to express demands and provide feedback to local governments, CDD efforts help to strengthen the accountability mechanisms that make decentralization meaningful. In so doing they help to support better governance overall. 109. However, there are also some important risks that need to be taken into account. CDD efforts need to be careful not to create parallel mechanisms of government. While it is certainly true that roles and responsibilities of the different levels of government need to be clarified in most transition economies, community efforts can lead to ways of circumventing local governments and may potentially cause even greater problems for the ability of local govemments to effectively manage their responsibilities. As is the case in many local governments the level of community involvement may hinge a great deal on the personality of a single leader and often times, elites may be in a position to exert particular influence. These latter are issue that effect decentralization more broadly and require that mechanisms for transparency and oversight be developed. An additional aspect to which CDD efforts need to pay attention is to make sure that efforts across communities are coordinated in areas where fragmentation can lead to inefficient outcomes. For those services which go beyond a particular community, it is important that coordination occurs, optimally this would be through a level of community that encompasses all the relevant communities. 110. CDD efforts are likely to be a tool for supporting changes in institutions and accountability, and in strengthening social capital. By building capacity at the lowest levels, CDD can be an important element of effective decentralization. However, as with other types of capacity building, it is critical that CDD be developed in the context of a consistent and coherent strategy for decentralization, ideally as part of that overall strategy. This assures that the learning and capacity developed will contribute to effective local governments. 50 Box 3 Community Driven Development in Central Asia Countries in Central Asia highlight the potential benefits of community driven development as a means of building capacity, but also its inherent difficulties. In particular, three factors distinguish Central Asian communities: i) many communities, especially in rural areas, are artificial creations from the past, ii) new state institutions have only partially replaced collapsed soviet institutions and therefore there may be questions of legitimacy; and iii) migration and other demographic changes have disrupted social bonds and networks, complicated resource allocations and roles, creating difficulties for community participation. As a result is not always straightforward to identify what constitutes a community in these countnes. Nevertheless there are efforts in the Central Asian countries that have focused on strengthening communities and their capacities. In Kyrgyz Republic, several projects have community activities: The Sheep and Wool Improvement Project (FYI 996) works through sheep farmers associations; the Rural Finance Project I (FY1997) has a small-scale group credit facility, based on Agricultural Consumers Cooperatives; an On-Farm Irrgation Project (FY2000) water user associations are integral element of the project. All of these help in various ways to build the capacities of these groups. The Rural Water Supply Project (FY2001) will be demand driven, responding to community groups who request support to build or rehabilitate rural water systems that the groups will operate and maintain. In Kazakhstan, the Drainage and Irrigation Improvement Project (FY1996) works through water users associations. Similar projects exist or are being proposed in Uzbekistan and Tajikistan. While these efforts have had inimediate impact on the given area on which they focus, it is too early to assess the overall impact on building social capital and accountability. The Country unit has proposed a four prong strategy for the next three years: demonstrate value by assessing and improving existing operations; lay the groundwork for new approaches and operations; disseminate results and findings to build constituencies and advocate to create new opportunities and mobilize resources. Only in the last few years have some of the Central Asian countries begun to contemplate whether more decentralized approaches to governance and service delivery are appropriate and a major issue is the capacity of lower levels to manage greater responsibilities. Despite the many difficulties and issues to be addressed, these countries offer an opportunity to build coherent strategies that integrate decentralization and capacity building through community development, so that they each can reinforce the development of the other. Source: CDD Strategy for Central Asia 51 THE ROAD AHEAD 111. Transition countries are in the midst of monumental changes . For political, economic and fiscal reasons many have started down the road of decentralization to find that it is both fraught with challenges and does not always live up to expectations. While many countries are used to asking lower levels to admnister functions, it is much more difficult, and the conditions are much more stringent, to really devolve decision making in an effective way. All countries in the region continue to grapple with the numerous challenges that stand in the way of effective decentralization. 112. The rationale for pursuing such a path is not always straightforward, but there is reason to believe that decentralization in an accountable environment can help to improve the perfornance of the public sector, assist in combating corruption and help countries in their efforts to reduce poverty (see Box 4). In terms of public sector reform, decentralization is at the nexus of key forces that help to strengthen public sector institutions overall. It can be a vehicle for establishing the appropriate rules and restraints Box 4 Decentralization and Poverty Reduction The Bank's recent study, Making the Transition Workfor Everyone: Poverty and Inequalhty in Europe and Central Asia documents the rise in poverty and inequality since the start of the transition and suggests an agenda for poverty reduction. Much of this agenda suggests the important role of effective intergovernmental financial relations and the potential for fiscal decentralization to help address poverty when it occurs in an accountable environment. The study highlights the necessity of building public institutions to support poverty reduction strategies. Key aspects include increasing political accountability, strengthening institutional restraints, developmg the participation of civil society and media oversight, creating a competitive private sector and reforming public sector management. Decentralization can be an effective means of putting public officials in closer touch with the people and giving the poor greater opportunities for voice - provided that they operate in an accountable and transparent environment. Decentralization can also contribute to building civil society and providing new channels for community and social cohesion. Poverty reduction is also served by helping to create opportunities, strengthening security and reducing inequality.. If decentralization is done well, it may contribute to better resource use and therefore to growth. In most transition countries, the system of social assistance is administered and sometimes financed by local governments. The bulk of education and many health services are delivered through subnational governments, so the extent which the system of devolved or decentralized service delivery functions, can also influence prospects for the poor. Finally, systems of equalization can be utilized to help offset growing inequality and to assure that the poor receive the basic public services they need. for all levels of government. It is a mechanism for creating competitive pressures, sometimes among governments, and sometimes among private contractors who provide 52 services to the public sector. Finally, it is a mechanism for strengthening voice and partnership. 113. When done well, fiscal decentralization can support hard budget constraints and macroeconomic stability, as well as reducing moral hazard and generating responsive, effective and sustainable service delivery. This in turn can promote increased incomes and productivity, improved literacy, better health and strengthened civil society. The problem is when done badly, it can have the opposite results. The key question then becomes, how to "do decentralization" well. Given the diverse starting positions and idiosyncrasies that exist in the transition countries, the answer is likely to differ, but all can be considered in the context of what we have learned about the challenges that transition countries have faced so far. 114. Figure 7 sets out key elements of a fiscal decentralization strategy. The figure reflects three different layers or areas which play an important role in effective decentralization. The first layer reflects aspects of the contextual environment ( the "overpinnings", so to speak) that go beyond the system of intergovernmental finance, but that have an important influence on how effective decentralization may be. These include the structures of government and several dimensions of accountability, such as political accountability, freedom of information, rule of law, the extent of state capture and even the extent to which fiscal management, both on the expenditure and revenue side, have advanced . The second layer sets out the constituent parts of a system of intergovernmental finance: expenditure assignments, revenue assignments, the transfer system and the legal and regulatory framework for subnational borrowing. It is important that this system be coherent and create the right incentives for responsible behavior for each part of the system. Finally the third layer is composed of elements such as capacity building, strengthening of civil society, specific systems of information and monitoring-- the "underpinnings" that need to be developed from the bottom up. 115. In the best of all possible worlds, fiscal decentralization would occur when all of these elements are "in place". It would start with a clear administrative and territorial structure, explained in the legal and regulatory framework. Mechanisms supporting accountability would be in place. Then the system of intergovernmental finance would clearly set out which level is responsible for what, how it is to be financed and what transfers are needed to pursue specific objectives and how those transfers would be allocated. It would also set out a framework for subnational borrowing based on the strength of subnational finances and the depth of financial markets. Finally, such a system would be supported by the appropriate degree of capacity at each level and in each type of government. This would be accompanied by systems for information collection and monitoring. A strengthened civil society and enhanced social capital would effectively express local preferences and reinforce accountability. 116. Of course, as seen above, in the real world, countries find themselves at different starting points, with some elements of such a strategy in place and some not. Other countries may have few of these elements in place. While each country is different, the information set out in the preceding sections suggests that the transition economies can be grouped into four categories. Each requiring a slightly different approach to the elements of 53 a strategy. The four groups are as follows: "keen decentralizers", "uncertain decentralizers", "non-decentralizers" and "decentralizers by necessity". Figure 7 Elements of a Fiscal Decentralization Strategy (tacityccountability -administrative reform\ -legal & regulatory -electoral system framework -dcforation flow H on wha they are aiming for. Structures--althoughevolving-areinplace-rule of law t k \ ~~~~~~~~~~~~~~-state capture/ discussion about how and if they might be changed.Sy-fiscal management j / ~~~IGF System\ majo elements of the syst expenditure assignment f a focused ondeel w -revenue asslgnment s o and reuligifat e(op-transfers C i tp Capacity need Insormas on t sa s i l o -rural H--- & m i n Social c 117. The "keen decentralizers" compnse Czech Republic, Croatia, Estonia, Hungary, Latvia, Lithuania, Poland, Slovak Republic and Slovenia. While these countries have chosen different degrees of decentralization. Most have achieved some degree of consensus on what they are aiming for. Structures --although evolvag-are in place and there is open discussion about how and if they might be changed. Systems for accountability, such as elections, freedom of the press, etc, are largely in place. In most of these countTies the major elements of the system of intergovernmental finance are in place - efforts need to be focused on developing own revenues and on developing sources of financing for renewing and rebuilding infrastructure (to bring it up to EU standard). Capacity in these countries is typically an area that needs work because of the small size of many of the local governuments. However, information & monitoring systems and social capital are relatively well developed. While there is work to be done to improve elements of the system sin these countries, they are well on their way. 54 118. Therefore in the "keen decentralizers", a pnority will be providing support to countries in this area in the run-up to EU accession. Specific issues that will likely need to be addressed include: the role of regional governments and implications for expenditure assignments; developing "own revenues" through property taxes, personal income tax surcharges, motor vehicle taxes; transfer mechanisms for the allocation of funds, especially investment funds; assistance in helping to contract out infrastructure and support to help build capacity in the smallest communities. 119. The "uncertain decentralizers" include Albania, Bulgaria, Kazakhstan, Kyrgyz Republic, Romania, the Russian Federation and Ukraine. Many of these countries are working wholeheartedly to improve their systems of intergovernmental finance, but do not always have a consensus on where they are going, and have weaknesses in the "overpinnings" and the "underpinnings" for effective decentralization. Administrative reforms and the legal and regulatory frameworks are less far along and the elements necessary to support accountability are in need of further development. Typically in these countries the framework for intergovernmental finance requires more than just changes on the margin, but some major realignments to improve incentives. Significant efforts are needed to build capacity at lower levels of government, to strengthen information and monitoring, and to support the development of civil society. The specific entry point and sequencing of measures will vary in each country, depending on the progress made so far in each area and the degree of political will. An additional consideration is that many of these countnes have not yet fully come to grips with macro stabilization and therefore efforts at decentralization must be careful not to cause further deterioration on the macro front. 120. The "non-decentralizers" include: Armenia, Azerbaijan, Belarus, Georgia, Moldova, Tajikistan, Turkmenistan and Uzbekistan. While some of these countries such as Georgia and Moldova have begun to pursue decentralization recently, others remain highly centralized. In these countries, effective decentralization is unlikely until the "overpinning" and "underpinnings" can be significantly strengthened. Structures may be developed, but accountability is particularly weak among a number of its dimensions. System of Intergovernmental finance inmost of these countries continue to largely reflect old modes of doing business. Local capacity is minimal and civil society and strengthening of social capital is in its infancy. In these countries, priority should be placed on strengthening accountability and supporting the development of civil society. Gradually, if the situation permits, countries may then consider devolving more authorities. 121. The "decentralizers by necessity" include Bosnia & Herzegovina, the Former Republic of Yugoslavia and, potentially, Macedonia. These countries are not cohesive states that choose to decentralize, but rather countries in which decentralization pursues specifically political purposes. On this context, some of the "standard rules" of intergovernmental finance are over-ridden, so efforts at continued reform need to focus on strengthening accountability, building capacity and developing civil society. 122. In sum ,there are many different roads to decentralization. To the present, most countries have made great strides, particularly given the monumental changes underway in these societies. Some have come quite far, but only very few countries in the region can 55 truly be said to have decentralized. Real, effective fiscal decentralization will take more effort and more time on a variety of fronts. 56 REFERENCES Bahl, Roy. 1999. Fiscal Policy in China: Taxation and Intergovernmental Fiscal Relations. San Francisco, CA: The 1990 Institute. Bahl, Roy and Johannes Linn. 1992. Urban Public Finance in Developing Countries. New York: Oxford University Press. Banks, C. and J. M. Pigey. 1998. "Republic of Albania: Opportunities and Issues for Municipal Reforms." Urban Institute Project 06610-113, Prepared for USAID. Bird, Richard M., and Susan Banta. Fiscal Sustainability and Fiscal Indicators in Transitional Countries. Proceedings of the conference on Central and Eastern Europe and New Independent States. Instanbul. June 15-18, 1999. USAID. Bird, Richard M., Robert D. Ebel, and Christine I. Wallich, eds. 1995. Decentralization of the Socialist State. A Regional and Sectoral Study. Washington D.C.: The World Bank. Bird, Richard M., and Francois Vaillancourt. 1997. Fiscal Decentralization in Developing Countries. Cambridge, U.K. Cambridge University Press. Burki, Javed, and Guillermo Perry, ed.s. 2000. Decentralization and Accountability of the Public Sector. Annual World Bank Conference on Development in Latin America and the Caribbean 1999. Washington: D.C. : The World Bank. Burki, Javed, and Guillermo Perry and William Dillinger. ed.s 1999. Beyond the Center: Decentralizing the State. World Bank Latin American and Caribbean Studies. Washington, DC: World Bank. Fisman, Rayrnond and Roberta Gatti. 2000. Decentralization and Corruption: evidence Across Countries. World Bank Policy Research Working Paper 2290. Washington, DC. The World Bank. Fiszbein, Ariel. 2001. Decentralizing Education in Transition Societies: Case studies form Central and Eastern Europe. World Bank Institute Learning Resources Series. Washington, DC: The World Bank. Hellman, Joel; Geraint Jones; and Daniel Kaufman. 2000. Seize the State, Seize the Day: An Empirical Analysis of State Capture and Corruption in Transition. World Bank. Presented at the Annual Bank Conference in Development Economics. April 18-20, 2000. Horvath, Tamas. 2000. Decentralization: Experiments and Reforms. Local Governments in Central and Eastern Europe. Local Government and Public Service Reform Initiative, Budapest: Open Society Institute. 57 Ivanov, Stefan. 1999. Local Finance. Sofia. Club Economika 2000. Ma, Jun. 1997. Intergovernmental Fiscal Transfers in Nine Countries: Lessons for Developing Countries. Policy Research Working Paper 1822. Washington, D.C.: World Bank. Kurlyandskaya, Galina 2001. Budgetary Pluralism of Russian Authorities. Discussion Paper Number 17. Local Government and Public Service Reform Initiative. Budapest: Open Society Institute. OECD. 1999. Taxing Powers of State and Local Government. OECD Tax Policy Studies, No. 1. Paris: OECD. OECD. 2001. Fiscal Design Across Levels of Government: Year 2000 Surveys - Summary Note. Paris. OECD Shah, Anwar. 1994. "The Reform of Intergovernmental Fiscal Relations in Developing and Emerging Market Economies." Policy and Research Series 23. Washington, D.C.: World Bank. Spahn, Paul Bernd. 1995. "Local taxation: Principles and Scope." in Jayanta Roy, ed. Macroeconomic Management and Fiscal Decentralization. EDI Seminar Series. Washington, DC: World Bank. Stein, Ernesto. 1997. "Fiscal Decisionmaking in Decentralized Democracies." IADB. Can Decentralized Democracy deliver Fiscal Stability?: The Institutional Dimension. Washington D.C. Ter-Minassian, Teresa. Ed. 1997. Fiscal Federalism in Theory and Practice. Washington, DC: International Monetary Fund. Treisman, Daniel (2000). Decentralization and the Quality of Government. Paper presented at IMF conference on Fiscal Decentralization. November 20-21, 2000. Uphoff, Norman. 1997. Institutional capacity and decentralization for rural development. Technical Consultation on Decentralization, FAO, Rome, 16-18 December, 1997 World Bank. 1996. World Bank Development Report 1996: From Plan to Market. New York: Oxford University. --. 2000. Anticorruption in Transition: A Contribution to the Policy Debate. Washington, DC: The World Bank. --. (2000). Balancing Protection and Opportunity: A Strategy for Social Protection in Transition Economies. Washington, DC: World Bank. 58 ---. (2000) From Commissars to Mayors: Cities in Transition Economies. --. 2000. Entering the 215' Century. The World Development Report 1999/2000. New York: Oxford University Press for the World Bank. --. 2000. Hidden Challenges to Education Systems in Transition Economies. Washington, DC: The World Bank. 59 Annex Table 1: Data Year and Sources Country Year Source Albania 99 Banks and Pigey, (1998 Hegedus, BTO (2000) Armenia 99 Country office ( Vigen Sargsyan's EMs) Azerbaijan Data on elections from Freedom House, no local fiscal data available Bulgaria 98 Stefan Ivanov's book update Stefan Ivanov "Local Finance " Club 'Economika 2000' Sofia, march 1999 Bosnia & 99/00 1) Spahn, Bemd 2001 Intergovernmental Fiscal and financial relations and Institutional Aspects Herzegovina Affecting Municipalities in Bosnia and Herzegovina 2) FDI-CEE 2001 Bosnia & Herzegovina Fiscal Decentralization Forum World Bank Institute Fiscal Decentralization Initiative for Central and Eastern Europe. Croatia 99 1) Dubravka Jurlina Alibegovic "Croatia-Fiscal Decentralization and Local Government " First Draft 2) William F Fox and Dubravka Jurina-Alibegovic " Local Public Finance in Croatia Overview and Issues "June 1998 3) Council of Europe. "Croatia- Structure and operation of local and regional democracy Situation in 1998." CDLR, Council of Europe. Czech Republic 00 1) World Bank. 2001. Czech Republic-Intergovernmental Fiscal Relations During Transition 2)Lacina and Vajdova in Horvath (2000) 3) WB Country Report "Czech Republic Toward EU Accession Main Report, 1999 4) Vera Kamenickova "Fiscal Decentralization in the Czech Republic." The Urban Institute June 1999 Ul Project 06610-532 Estonia 97 1) Maeltsemes in Horvath (2000) 2) Tarvo Kungla. "Fiscal Decentralization in Estonia " Open Society Institute, Discussion Papers, No 13 Local Government and Public Service Reform Institute 3) IMF document (1997) Republic of Estonia Selected Issues and Statistical Appendix 1997 SM/97/277 4) World Bank (1995) Estonia Financing Local Governments. Report No 14925-EE Hungary 98 1) World Bank "Hungary-Subnational Modernization." Hungary Policy Note WB HU-P066937 2) Mihaly Kopanyi, Samir El Daher, Deborah Wetzel, Michael Noel, Anita Papp "Hungary. Modernizing the Subnational Government System " World Bank discussion paper No 41 May 2000 3) SNDP paper "Intergovernmental Finance in Hungary Continuous Progress, Continuous Change and Options for Reform nWB, EDI, BKHF, USAID, Ul etc 4) William F Fox. "Intergovernmental Finance in Hungary Summary and Evaluation June 12, 1998 5) Christine Allison et al "Hungary Poverty and Social Transfers " WB Kazakhstan 99 1) WB "Kazakhstan Public Expenditure Review." Vol. Il.: Main Report. ECA-PREM, June 27,2000 Report No 20489-KZ, Vol. III. Technical Annexes Kyrgyz Republic 99 Natalia Pisareva (World Bank consultant) via EM Latvia 99 1) World Bank. 2000 "Latvia Strengthening Fiscal Decentralization and Balanced Development Current situation and policy options for reform." Infrastructure Sector, ECA Apnl 10, 2000 2) Vanags and Vilka in Horvath (2000) Lithuania 97 1) Beksta and Petkevicius in Horvath (2000) 2) World Bank Country Report "Lithuania An Opportunity for Economic Success " Vol 2 Analytical Background Macedonia 97 Francis Conway et Al "Draft Report Macedonia: Local Government in Transition " January 1999 Urban Institute project 06610-903 Moldova 97 1) Luiz de Melo "Fiscal Federalism and Govemment Size in Transibon Economies The Case of Moldova December 1999 (IMF Fiscal Fed. Working Paper) 2) IMF "Republic of Moldova Recent Economic Development" September 1999 IMF Staff Country Report No 99/110 Poland 97 1) Kowalczyk in Horvath (2000) 2) Tony Levitas. "The Political Economy of Fiscal Decentralization and Local Government Finance in Poland 1989-1999" Research Tnangle Institute June 1999 Draft Romania 99 1) Data form Dena Ringold. 2) Romania Study of Local Social Service Delivery 2001 Russia 00 1) Jorge Martinez-Vazquez and Jameson Boex. 2001. Russia's Transition to a New Federalism World Bank Institute. 2) Jorge Martinez-Vazquez et al. "Fiscal Decentralization in the Russian Federation: Main Trends and Issues " December 1998 2) OECD "Economic Surveys Russian Federation." March 2000 Slovak Republic 98 1) Nemec, Bercik and Kuklis in Horvath (2000) 2) Thurzo, A. and Congress, N, eds (2000). "National Forum on Fiscal Decentralization Possibilities and Preconditions of Fiscal Decentralization in the Slovak Republic" Washington, D.C FDI-CEE 3) Council of Europe "Slovak Republic- Structure and operation of local and regional democracy Situation in 1998 n CDLR, Council of Europe Slovenia 98 1) Setnikar-Canka, VlaJ and Klun in HOrvath (2000). World Bank Country Study "Slovenia. Economic Transformation and EU Accession " Vol II Main Report Tajikistan Data on elections from Freedom house No local fiscal data available Turkmenistan Data on elections from Freedom house No local fiscal data available Ukraine 98 World Bank "Intergovernmental Finance in Ukraine An Agenda for Reform " April 15, 1999 World Bank FYR, (Serbia Mont) No fiscal data available 60 ANNEX TABLE 2: BASIC CHARACTERISTICS Population Urbanization Density Population Area (% of Pop. Living (People per GDP per Capita (millions) (I000s of sq km)' in Urban Areas) sq. km) (PPP, cur int'l $) Albania 3.3 27.4 404 12 9 2797 3 Armenia 3.8 28.2 69 4 134 6 2067.2 Azerbaijan 7.9 86 6 56.7 91.3 2169.3 Belarus 10.2 207.5 70.6 49 3 6088 2 Bosnia & Herz 3.8 51.0 42.2 73.9 Bulgaria 8.3 110 6 69.0 74 7 4795 5 Croatia 4.5 55 9 56 9 80.5 6790.0 Czech Rep. 10.3 77.3 74.6 1332 12331 5 Estonia 1.4 42 3 69.1 34 3 7663.4 Georgia 5.4 69.7 59 7 78 1 5000.5 Hungary 10.1 92.3 63.6 109.5 10349 8 Kazakhstan 15 6 2670 7 56.4 5.8 4339 8 Kyrgyz Rep. 4.7 191.8 33.9 24.5 2251 6 Latvia 2.4 62 1 69.1 39.5 6043 0 Lithuania 3.7 64.8 68.3 57 1 6413 1 Macedonia, FYR 2.0 25.4 61.2 79.0 4293 8 Moldova 4.3 33 0 46 1 130 4 1941.1 Poland 38.7 304.4 64.8 127 0 7601.0 Romania 22.5 230.3 55 7 97.7 5331.2 Russian Fed. 146.9 16888.5 77 0 8.7 6406 5 Slovak Republic 5.4 48.1 57 2 112 1 9674.6 Slovenia 2.0 20.1 50.3 98.5 14257.8 Tajikistan 6 1 140 6 27 5 43 5 949 2 Turkmenistan 4.7 469 9 44 7 10.0 2664 2 Ukraine 50.3 579.4 67.8 86.8 3178.2 Uzbekistan 24.1 414.2 37.5 58.1 1964 9 Yugoslavia, FR 10 6 102.0 51 9 104 1 (Serb./Mont.) Maximum 146.9 16888.5 77.0 134.6 14257.8 Minimum 1.4 20.1 27.5 5 8 949.2 Average 15 3 855.3 57.1 76.5 5494 5 Data Sources World Development Indicators (WDI) 2000, data for 1998, except-- (1) Area (in lOOOs of sq. km), 1997 61 I X_Z. I Z- Je J- I I I I Id I a _ z _ -- 21l11IIIII I Z ZZ ;=) 2 _ _ _ _ _ X z _ _ > __ Z' a J _ _ _ _I z 7' _7 1 _ L {t zIt l l - M Jt 1 14 S zz z z~ *V z '__z ,___**dd_z I .z d~ |d|z '| '||||d| d||| ii} x z z _ .t _I . ,: z L z z L * z | Z Z I Iz - | | | z z z z z I I L I I I I I~~~~~~~~; le z zz , 1)e z zt LZl .Z z l0 l - 8~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~zz- z z -Z~~~~ z z ' z Cd o a, 'o tn v cr, W) 0 00 t-: tri 6 0 0 tn C4 m 2 00 'n 2 C:, a, ;7) Q "I CN 91 en 0 0 Cd co 0 w -4 'O v 'o 07 .4 "I"'ti PTIMF 1":I* II kn w 0 ci 0 (I C q C 0 - %R - (R -) q o tn X C, -1) In 10 00 en C, C- N CD C,4 C-4 en0 C) M E 1.0 e4 tn f4 -Ir m C-4 I= o - kf) E 0 U Cd' Cd rA co 0 en r, 00 0, 00 00 P4 0 d 4.) 0 a, u . . . . . . . . . 10 tn "6 C; c. o, 06 tn 11 CD o 00 o IT m tn o6 06 r- m ci en r- 'IT C4 ol 'D 00 w c I 0 C 01 Cl . . . . . . . 0 o C Go cq \0 -'e 0, oo 'T oo 'Q "r 'ov t- en 00 02 In t- V W'V W'A C ON tn 10 01 N M a, M M C-4 0 C4 .6 C ON tf) en "o o en IT 'Ir Mtn m o c 10 W) a, w) oo o r- w oo r- (7, FI) 'o m 7-4 :3 4) rA r-.,o r- Ten a\ %O r-: 06 6 (Z i r, 0, WD 6 r 4 ci en 0, a, \0 en r t- Ft--m m w w) -o m r-w 3, oo -e c-4 xc' U" r- t- eq 'a 0\ WI 00 "r 'D cq c; 4i 0; o Ln (10 cn oc r- r- 0 -IT , C4 p, tn 00 r.. 'D ; -4 (4 11v C4 ION rn en 14 I- r- tn 00 17, 'IT r- r-: C' 14 IR Oq 11 mNr- v- m en - -f)0eq C\ t- ":r r, 00 t, om mm o mr- oo *0 10 tn r- %o tn 00 tn w nmw mo r- %C t c' 0, CD r- i e4 'o 00 r4 In r- m w a M IR a\ C) m 00 "R C> r en C r- 'n 00 'A 'D "6 wm 06 fn Mn fn M rn M 'n IN q, In 'Ir Vr4 0. CA EN w Ca 0, En r MC -z 2M 0 0 0\ < m UU w=s ':4 04 H 'U' 0 Z' Z' En- U)D