The challenge of fiscal decentralisation in transition countries.

Dabla-Norris, Era

JEL Classifications: E62, H20, H30, H40

INTRODUCTION

A number of developed and developing countries have undertaken comprehensive reforms of their intergovernmental fiscal systems over the past two decades, and this trend has extended to the transition countries in Eastern Europe and the Former Soviet Union (FSU) as well. Several countries in the region are already in the process of considering or implementing far-reaching 'second-generation' reforms in this area. (2) However, persistent macroeconomic instability in some, and the entrenched legacy of socialism and central planning in others, have generated additional challenges which are complicating the design of an effective decentralised system.

The virtually global trend towards devolving service implementation to lower levels of government is based on the wide acceptance of the subsidiarity principle and on the view that it results in improved efficiency in the delivery of public services and, hence, a more efficient allocation of resources in the economy. (3) The current movement towards democratic forms of governance is also closely associated with the demand for decentralised government. (4) At the same time, increased fiscal decentralisation, in itself, is seen as an important means of increasing democratic participation in the decision-making process, thereby, enhancing accountability and transparency of government actions. In yet other countries, the trend towards greater autonomy for subnational levels of government is driven by the need for national coherence in the face of ethnic or regional centrifugal forces or conflicts. In many transition economies, the political and economic failure of autocratic, highly centralised socialist regimes may have provided an impetus to the subsequent decentralisation initiatives.

Whatever the precise background and the motivation in different countries, it is generally acknowledged that the manner in which decentralisation is carried out can have a significant impact on macroeconomic management and performance. A common view in the literature is that decentralisation may aggravate fiscal imbalances, thereby, endangering overall macroeconomic stability, unless sub-national governments are committed to fiscal discipline and the decentralisation package includes incentives for prudence in debt and expenditure management. (5) Empirical evidence on the relationship between decentralisation and macroeconomic instability is mixed. Shah (1998) and King and Ma (2001) find that a decentralised fiscal systems have a better record in controlling inflation and deficits. However, Fornasari et al. (2000) and DeMelo (2000) find that increases in subnational spending and deficits lead to an increase in spending and deficits at the central level.

Recent studies suggest that the design and implementation of a multi-tier system of government can significantly affect overall resource allocation in the economy and, hence, economic efficiency, growth, and welfare {Davoodi and Zou, 1998; Martinez-Vazquez and McNab, 2003; Akai and Sakata, 2002). A central argument for fiscal decentralisation leading to improved resource allocation rests on the assumption that fiscal decentralisation increases local influence over the public sector. However, in theory, there is an equal possibility that fiscal decentralisation simply transfers power from national to local elites and that improved access of local elites to public resources increases opportunities for corruption (Bardhan and Mookherjee, 2000). (6) In general, the impact of fiscal decentralisation on corruption depends to a large extent on the quality of the supporting institutional framework and in particular, the degree that subnational governments and/or officials can be held accountable.

In light of the possible effects--that depend on the institutional design --of fiscal decentralisation on economic growth, macroeconomic management and corruption, a key challenge for many transition economies has been to reap the economic benefits of decentralisation while maintaining control over public expenditures and borrowing, restoring growth and improving accountability of local governments and officials to limit corruption.

In retrospect, for many former socialist countries, the combination of efforts aimed at consolidating macroeconomic stabilisation during the early years of the transition, together with the fundamental structural changes in the economy, in some cases strong centrifugal forces, and political and ethnic conflicts, created an extremely complex setting for fiscal decentralisation. This goes a long way in explaining why the fiscal decentralisation process in many transition countries has been rapid, haphazard and largely nontransparent, with the emerging system of intergovernmental relations having important implications for budgetary developments.

In recent years, there has been progress in efforts to reform intergovernmental relations in many of the transition economies. However, this progress has been uneven across countries and across various components of intergovernmental fiscal relations. It is widely agreed that the Czech Republic, Hungary, Poland and the Baltics have been the more active reformers in many of the key aspects of intergovernmental fiscal relations. They have made considerable progress in carrying out fiscal decentralisation and have promoted institutional settings and processes that allow for the articulation of interests and policymaking based on consensus building. However even in these countries, there remain challenges and areas for needed progress. Many other countries, including Russia, Ukraine and the Central Asian States still need substantial reforms of the incentive structures that govern intergovernmental fiscal relations in order to obtain an efficient and well-functioning multi-tier system of government.

Bird et al. (1995) provided a comprehensive comparative assessment of the decentralisation in early years of transition. A recent study by Wetzel and Dunn (2001) examined the uneven progress on fiscal decentralisation and key challenges to effective decentralisation in the transition economies of Europe and Central Asia. They adopt a traditional approach that focuses on different structural elements that together make up a system of intergovernmental fiscal relations (the legal and institutional framework, expenditure and revenue assignments, transfers and subnational borrowing). Our paper differs from theirs in that we provide a more systematic framework to identify and assess key adverse incentive mechanisms inherent in the design of the system of intergovernmental fiscal relations in these countries. In this respect, our paper is similar to Dabla-Norris et al. (2000), which analysed key disincentives in intergovernmental fiscal systems in a smaller subset of transition countries, namely, Russia, Ukraine, and Kazakhstan. Our paper provides a simple analytical framework to examine areas where key principles of sound fiscal decentralisation have been breached with their potentially negative impact on the effectiveness of service delivery as well as macroeconomic performance. This includes an assessment of the strength of institutions pertinent to the intergovernmental financial relations, relevant laws, regulatory frameworks and other incentive structures that may facilitate or undermine effective fiscal decentralisation in transition economies. In the section following the next, we identify three critical principles of sound decentralisation which are then contrasted with actual policies and practices in select transition countries. Finally, the last section concludes.

OVERVIEW OF FISCAL DECENTRALISATION

Initial conditions

To set the stage for the discussion that follows, some of the essential characteristics of the fiscal adjustment that occurred during the early years of transition is presented. During the initial years of transition, fiscal imbalances quickly emerged as output collapsed due to structural dislocations with concomitant loss of revenues, particularly for the transfer-dependent states of the former Soviet Union. The negative impact on growth from structural dislocations was further aggravated by high inflation resulting from price liberalisation and the monetisation of large fiscal deficits to sustain output and employment (Tanzi and Tsibouris, 2000; Alam and Sundberg, 2001; Valdivieso, 1998).

Significant progress has been made towards achieving fiscal and macroeconomic stabilisation since the transition in most countries--progress which has importantly changed the environment in which the decentralisation process is taking place. There has, however, been substantial variation in the nature and pace of reforms across countries. In general, the Eastern European and Baltic countries, the most advanced reformers, have made rapid progress. The intermediate reformers--the South-Eastern European countries of Albania, Bulgaria, and Romania, and the BRO countries of Russia, Ukraine, Moldova, Kazakhstan, and the Kyrgyz Republic, and the slow reformers--Armenia, Azerbaijan, Belarus, Georgia, Tajikistan, Turkmenistan, Uzbekistan have been less successful in establishing fiscal institutions, controlling fiscal imbalances, and redefining the role of the state (see EBRD, 1998; Valdivieso, 1998).

The adjustments that took place in the context of economic stabilization represented only the first phase of a more substantial fiscal reform process. A key remaining challenge for these countries is to implement reforms aimed at improving the quality and efficiency of government, including improvements in institutional arrangements that underpin fiscal policy, and to enhance transparency and accountability at all levels of government. The reform of the structure of intergovernmental relations remains an important element of this reform agenda.

Decentralisation experiences

A common feature of almost all transition economies is that they began from a legacy of a highly centralised system of public finances with subnational governments acting mainly as administrative units with little independent fiscal responsibility. While originating from the similar economic structures and political systems, these countries have chosen very different routes and approaches to decentralisation, with some countries being considerably more centralised than others. In addition, progress with implementing intergovernmental fiscal reform has varied across the region.

One difficulty in comparing the degree of decentralisation to date across the various countries is that fiscal decentralisation is a multidimensional phenomenon, involving not only the assignment of expenditure and revenue responsibilities among different levels of government, but also the extent of subnational policymaking autonomy. For instance, the share of subnational spending in total government spending, which is a common measure of decentralisation, on average about 25 per cent among the transition countries, has varied from 15 per cent in Albania and Macedonia to over 50 per cent in Russia and Kazakhstan (Wetzel and Dunn, 2001). This standard, however, fails to take into account the effective decision-making authority of subnational governments. It also does not take into account whether subnational governments have the financial resources required to meet their assigned expenditures.

The uneven nature of the degree of decentralisation as well as the extent and scope of intergovernmental fiscal reform among the former socialist states largely reflects, among other things historical, political, ethnic, geographic, and demographic differences (see Table 1). For instance, countries with larger populations or geographic areas (such as Russia, Kazakhstan, and Poland) require a greater decentralisation of public service provision to subnational governments as compared to smaller countries like Moldova and the Kyrgyz Republic. Similarly, ethnically diverse countries such as Russia also would have fairly high need for fiscal decentralisation compared to other, smaller and ethnically more homogeneous transition economies (such as Slovenia). At the same time, political factors such as the accession to the EU can be seen as providing an impetus for reform in countries like the Baltics, the Czech Republic, Poland, and Hungary. In countries such as Croatia and Bosnia and Herzegovina, ethnic conflict may have played a significant role in shaping the nature of fiscal decentralisation that has evolved. Differences in institutional, economic and political development can also be expected to influence the extent of decentralisation across these countries. (7)

These factors not only have influenced the observed degree of decentralisation across the region, but also whether decentralisation policies are effective and have had their desired impact. At the same time, macroeconomic and fiscal policies adopted at the outset of transition have had an impact on the system of intergovernmental relations that has evolved in many countries. Many of the more advanced reformers, with more stable macroeconomic conditions, including low overall public sector deficits, have made considerable progress in carrying out fiscal decentralisation and, in particular, have promoted institutional settings that are supportive of effective decentralisation. However even in these countries, weaknesses in existing systems have prevented them from fully exploiting the potential public finance and service delivery benefits from decentralisation. In countries that are regarded as intermediate reformers in the broad fiscal arena, the current design and structure of the intergovernmental systems contains many inappropriate incentive structures that are adverse to the sound working of the fiscal system and, hence, to overall macroeconomic performance. Among the slow reformers, both in terms of the timing and actual implementation of fiscal reforms, intergovernmental fiscal relations preserves many features, patterns and structures of the Soviet system.

PRINCIPLES FOR SOUND AND EFFICIENT DECENTRALISATION

The importance of the relationships between decentralisation and macroeconomic performance underscore the need of identifying the factors that would ensure a sound and efficient decentralised fiscal system. We outline three basic principles that capture some key aspects of the incentive mechanisms needed for sound and effective decentralisation. Throughout the paper we use the term sound fiscal decentralisation with regard to the extent it is conducive to macroeconomic stability. The term efficient fiscal decentralisation refers to the extent it enhances microeconomic efficiency in the input and output mix of public service delivery.

Clarity, transparency, stability and well-defined rules of the game are paramount for achieving accountability that efficient and sound decentralisation requires. Given the interdependence among the various components of the system of intergovernmental relation, this requires a clear and effective delegation of functions by central government, with revenue assignments that are transparent, unambiguous, and commensurate with subnational governments' expenditure responsibilities. It also requires transfers that are based on stable principles and specified by legal formulas that support hard budget constraints.

A measure of autonomy for subnational governments on the expenditure and revenue side is crucial for realising the efficiency gains of decentralised government and supporting macroeconomic stability. On the expenditure side, this requires subnational budget flexibility to decide--within limits --expenditure priorities and the choice of both the output mix and techniques of production. On the revenue side, this requires that subnational governments have the authority to own-finance locally provided services at the margin. More complete revenue autonomy requires a minimum of authority to set tax rates and an assignment of at least one significant tax source.

Achieving meaningful and sustainable autonomy at the subnational level and avoiding an inequitable geographical reallocation of resources requires a 'level playing field' such that local and regional councils can provide relatively uniform service levels at reasonable levels of tax effort. Such equalisation of fiscal disparities is possible only when effective sizeable and well-designed equalisation transfers are put in place.

While autonomy should be explicit and well-defined, it must also be circumscribed with respect to the access to borrowing by subnational governments in order to support hard budget constraints and reduce moral hazard. In particular, subnational borrowing in the absence of both market discipline and a sound, effective and strictly enforced regulatory framework (as seen in some transition economies) can undermine achieving fiscal targets for the general government and hence pose a risk to macro-fiscal stability. (8) For instance, if central governments lack credibility with regard to not bailing out subnational governments with debt servicing problems, given the implicit moral hazard, it cannot solely rely on the market to enforce sufficient fiscal discipline on subnational governments (Ter-Minassian, 1997a; Rodden et al., 2003). As a result, the framework for subnational borrowing requires an appropriate balance of market discipline, rules-based controls, and administrative oversight and supervision.

Institution building is the last of the three pillars. A prerequisite for successful decentralisation is that subnational governments possess the administrative and technical capacity required to effectively carry out their assigned responsibilities. Supporting institutions, including democratic representation, sound budget processes, local government revenue collection capacity, and mechanisms to ensure coordination and cooperation between different levels of government--both at the political and the technical level --are crucial for the functioning of a multi-tier system of government. In the remainder of this section we provide an overview from a broad range of transition economies to illustrate instances where these principles are lacking, with potentially adverse impacts on economic efficiency, macroeconomic stabilisation, and growth.

Clarity of roles

Legal and institutional structure

In many transition countries, the evolution of the legal and institutional framework has been subject to a fairly continuous series of revisions, reversals, and shifts in focus and has reflected political compromises rather than consistently applied rules and principles. The degree to which the legal and institutional framework supports a well-defined system of intergovernmental fiscal relations, however, varies. Countries like Hungary, the Czech Republic, and Poland have pioneered reforms in the legal and institutional framework required for decentralisation, and have had considerable success in defining the role of intermediate levels of government. However, current legislation in these countries has set the stage for the fragmentation of municipalities into entities of inefficient sizes. (9) Some countries in Central Asia (eg, Kyrgyz Republic, and Tajikistan) still lack a well-specified legal and institutional basis for decentralisation.

Other countries (Russia, Ukraine) have made significant progress in recent years to clarify the legal and institutional framework for decentralisation between the centre and regional (oblast) governments as well as subprovincial governments. (10) In other countries (eg, Kazakhstan) the distribution of functions to local governments remains ambiguous, with important gaps in the legislation arising from the lack of well-defined criteria to determine the assignment of functions and powers across regional and local governments. In some countries (Belarus and Azerbaijan), the practice of resolving contradictions between the various legislation in an ad hoc manner, with crucial provisions often decided in the annual budget laws, has imparted a measure of unpredictability and instability to the system of intergovernmental relations. In addition, overlapping and poorly defined roles, and unclear divisions of power between different levels of government has created confusion about the functions and modes of interaction of different parts of government.

Expenditure assignment

In recent years, but to varying degrees, progress has been made in clarifying expenditure assignments. While actual assignments often broadly correspond to the principle of subsidiarity (Table 2), further progress is needed in a number of countries. The ambiguity stemming from shared responsibilities between the centre and subnational government needs to be reduced. There is also a wide variation across countries with respect to the distribution of specific expenditure responsibilities. In most countries, regional and local governments bear a substantial portion of expenditures on education and social insurance and health services. This figure is higher for the Baltics and the Central Asian countries, reflecting, in part, differences in physical and demographic characteristics, institutional capacities, political preferences and the degree of centralisation. In other countries, local governments are routinely mandated to fund and provide redistributive social services, funding for which could be more effectively provided by the central government.

The advanced reformers, (Czech Republic, Hungary, Poland, Estonia, and Lithuania), have developed clear assignments of expenditure responsibilities. However, the efficiency of service delivery is often compromised due to the excessive fragmentation of municipalities in countries such as Hungary and the Czech Republic, as many small local governments are required to provide a broad range of services. In addition, local governments are frequently given greater flexibility in providing certain voluntary services with the option of passing this responsibility to a higher level of government. (11)

In a number of countries, such as Albania, Moldova, Georgia, Romania, and Azerbaijan, the distribution of spending responsibilities remains ambiguous. In Russia, for instance, ambiguity in the assignment of the authority to regulate and issue declarative norms and the resultant proliferation of unfunded mandates, until recently, compromised subnational budgetary positions (Litwack et al., 2002). In general, the lack of clarity and stability in expenditure assignments have detracted from accountability at all levels of government and undermined the efficiency of public expenditures. Moreover, in many countries, it has reduced incentives to prioritise budgets, lower costs of service delivery, eliminate excess physical capacity, and properly maintain capital infrastructure at the subnational level. (12)

Subnational governments in Russia, Ukraine, and other countries have also been slow to rid their budgets of private market interventions, with subsidies for housing and communal services, including public utilities, accounting for a large proportion of their expenditures. For instance, subnational governments in Russia spend close to one-third of their total resources on subsidies to consumers and in Kazakhstan the relevant figure was 10 per cent (Dabla-Norris et al., 2004) Among the more advanced reformers, such as Hungary, Poland, and the Baltics, subnational governments have been more successful with privatisation and contracting out of service provision. For other countries with limited private sector capacity and a weak legal and institutional environment for private providers, the private sector response has been marginal.

In many cases, the problem of unclear expenditure assignments has been especially acute at the regional-local level, with regional governments enjoying a high degree of discretion, with the result that they exert a certain arbitrariness over expenditure assignments to their subordinate local governments. For instance, local governments in Georgia and Kazakhstan lack a formal assignment of expenditure responsibilities which may negatively affect the accountability of both regional and local governments to taxpayers and subject local governments to added budget uncertainty.

Tax assignment

While progress to varying degrees has been made in recent years in many countries in the region (eg, Hungary, Poland, the Czech Republic, the Baltic's, and Russia, and to a lesser extent in Kazakhstan, Kyrgyz Republic, and more recently Ukraine) to make revenue assignment more transparent and predictable by formalising revenue sharing arrangements and adopting stable and uniform sharing rates at the central-regional level, in other countries, revenue assignments continue to rely on the 'regulating approach.' Under this approach, the formal basis for setting tax sharing rates is the central government estimates of each region's 'minimum' expenditures needs. This practice has resulted in customised and yearly changing sharing rates and compensations through non-transparent transfers to fill the subnational budget gaps as well as assignments being ad hoc and negotiated with individual regions through bilateral bargaining.

As seen in Table 3, subnational governments in most transition economies are assigned income taxes as the most important type of shared revenue with consumption taxes trailing far behind and generally being more important at the central government level. Property taxes, which are widely recognised as an important source of finance for local governments, are still a small and underutilised revenue source locally, often reflecting lags in regularising the property market as a basis for taxation.

The lack of clearly defined, stable and uniform revenue assignments between the centre and subnational governments inherent in this approach has weakened budgetary management at the subnational level and created perverse incentives for subnational governments to either hide locally mobilised revenue sources in extrabudgetary funds (eg at the municipal levels in Ukraine, and at the regional level in many Central Asian countries), or to simply reduce their efforts to mobilise revenues locally. Evidence from Bulgaria, Georgia and other countries suggests that central governments routinely revise tax sharing rates on an annual basis in face of political and economic pressures. (13) Punitive 'extractions' by higher level governments in the form of clawing back any additional revenues raised by lower level governments through reduced sharing rates have also created perverse incentives for revenue mobilisation, at the local levels in many countries. (14) In Albania, while there has been improvement in recent years, local governments' efforts to improve revenue collection in the past were discouraged since they were not free to determine the spending allocation of funds collected under the local (independent) budget. This led to surplus funds being trapped in the treasury system and captured by the central government by the end of the year. The resultant non-uniformity in revenue sharing and the absence of stability undermined sound fiscal management at the local level (Banks and Pigey, 1998).

Transfers

In most transition economies--the Slovak Republic, Poland, Hungary, and the Baltics are among the exceptions--transfers among the various levels of government remain discretionary, and negotiated, with transfers largely unconditional and determined ad hoc by the central government, often changing with each annual budget. Other countries, including Bulgaria, Croatia, Romania, Georgia, Russia, Kazakhstan, and more recently, Ukraine have moved towards more transparent formula-based systems, although weaknesses remain. While most formulas include indicators of expenditure needs and tax capacity, in Bulgaria and Estonia for instance, formulas continue to incorporate a fiscal gap component, creating negative incentives for revenue mobilisation (Table 4). In many countries, transfer formulas are undermined through end-year negotiated transfers that serve to soften budget constraints.

The gap-filling nature of the transfers between central and regional governments in many transition countries and between regional and local governments in most countries, however, provides negative incentives for revenue mobilisation by subnational authorities and the efficient provision of public services as any increase in regional own revenues or budgetary savings in the provision of public services triggers, often commensurate reductions in the level of transfers. This practice is seen in Albania, Bulgaria, Belarus, Moldova, Ukraine, and all Central Asian countries.

Apart from equalisation transfers, other types of grants and transfers are used across the region. Matching grants for funding centrally mandated services in the areas of education, health or social spending are used widely in Croatia, the Kyrgyz Republic, and Poland, and for investment purposes, in the Czech Republic and Hungary. In Russia and Ukraine, however, the use of other ad hoc and non-transparent transfers, such as mutual settlements, which accounted for over 75 per cent of all non-equalisation transfers in Russia in 1998, provided a soft budget constraint environment at the subnational level. In recent years, however, have witnessed significant improvement in the design and implementation of intergovernmental transfers (Martinez-Vazquez et al., 2004).

The use of specific or conditional grants (ie, earmarked for specific purposes) is also common. For example, in the Czech Republic all grants are specific, covering a number of delegated responsibilities (state administration and environmental protection) and own responsibilities (education, social care, and cultural activities). In Poland, the widespread use of specific grants has been criticised on the grounds that it provides negative incentives for local governments to exert tax-effort (OECD, 2001f). Other countries (Russia and Kyrgyz Republic) provide conditional grants on a per capita basis, while Armenia, Hungary, Kazakhstan, and the Baltics do not provide conditional grants for current expenditures.

A measure of autonomy

Expenditure autonomy

Sound and efficient decentralisation requires a close correspondence between responsibility and decision-making authority. However, effective expenditure autonomy at the subnational level has been limited in most transition economies. Norms and regulations (for instance with regard to quality and scope of service provision) emanating from central government agencies have interacted with explicit spending priorities set by the central governments, including spending mandates to severely constrain the authority of subnational governments to adjust current expenditures. For instance, in Bulgaria about 90 per cent of actual local expenditure in 1999 were not under the control of local authorities (McCullough et al., 2000) and in Albania this figure was 90-95 per cent (Banks and Pigey, 1998). Hence, the outcome in most transition economies has been considerable burdens imposed on subnational budgets.

This is in contrast to the situation in Hungary, Poland, Estonia, Latvia, and the Czech Republic, where subnational governments are granted greater flexibility in service delivery, backed by law. In Hungary, local governments are entitled to determine the ways and modes for service provision, depending on the requirements of the local population and their financial resources. Mandatory duties of service provision can only be imposed by law approved in Parliament and with the required funding provided by Parliament (OECD, 2001c). In Latvia, local governments face no output controls on their service provision. They are free to provide services according to law and norms in certain sectors (education, welfare, and environment), to structure the provision of services and, in general, and to decide the service level (OECD, 2001d). In the Czech Republic, local governments decide on the structure of local expenditures and the quality and quantity of the services to be provided (with exception of some earmarked activities financed by specific grants). However, in some areas such as education a considerable share of the local costs is determined by the central government while in other areas the local governments have considerable discretion (Oliveira and Martinez-Vazquez, 2001). While many of the expenditure norms developed by central government agencies in transition countries are merely indicative in nature, other regulations directly affect the expenditure positions of subnational governments. For instance, local governments in Armenia and Kazakhstan have little flexibility in setting wages, or the wage fund of public employees. The lack of effective autonomy over expenditures combined with weak capacity for budgetary management has encouraged the accumulation of payments arrears by subnational governments as a means of deficit financing, and eroded fiscal discipline and accountability. Even among the more advanced reformers, central government mandates on local governments with respect to employment and salaries often constrain budgetary autonomy of local governments (Ebel and Yilmaz, 2002). However, unlike in the intermediate and slow reformers, local government associations in these countries typically play an important role in negotiating mandates with the central government.

In a number of transition economies, local autonomy has also been constrained by the shifting down of subsidies and social services to local governments since the beginning of the transition. Minimum expenditure requirements for social services imposed by the central governments impinge upon the budgetary autonomy of local governments. In Bulgaria, municipalities have to fund 50 per cent of social welfare payments from their own revenue which results in significant disparities among municipalities in residual spending on other services (McCullough et al., 2000). In Russia, until recently, this problem was compounded by the existence of regional norms and regulations in conjunction with federal norms and mandates, and the failure of local governments to distinguish between funding for each (Litwack et al., 2002).

Overall, there is considerable spending autonomy in the Central and East European countries such as the Czech Republic, Hungary, Poland, and the Baltics whereas in Ukraine, Albania, Bulgaria, the Kyrgyz Republic, and Moldova the local spending level and priorities are mainly determined by the central government.

Revenue autonomy

Granting effective subnational revenue autonomy--the authority for subnational governments to determine tax rates and/or bases-remains a critical challenge for most transition countries. As seen in Table 5, only the advanced reformers have devolved limited revenue autonomy to subnational governments, although they still rely on the central government for the bulk of their revenues. (15) In the Czech Republic, the Slovak Republic, Hungary, and Poland the share of 'own' revenue (over which they have policy control and collect themselves) ranges from 33 to 40 per cent. In the Czech and Slovak Republics most of this revenue is non-tax revenue; in the former local governments have very limited tax authority. In Lithuania, legislation giving municipalities greater discretion in setting tax rates and fees accruing to local budgets--including a real estate tax of up to 1.5 per cent of assessed property values--is under consideration for 2002.

A few countries among the intermediate reformers also appear to have fairly high shares of 'own' revenue (Romania (33 per cent), Ukraine (31 per cent), and Russia (43 per cent)). In the remaining transition countries, subnational revenue autonomy remains virtually nonexistent. Even among the Baltics, that have made significant progress in other areas of intergovernmental finance, only a very small share of subnational revenue is controlled by subnational governments, which depend almost entirely on transfers from the central government.

The generally low level of revenue autonomy, particularly among the intermediate and slow reformers, reflects, in part, weak subnational administrative capacity, political constraints, pre-emption of the local tax base by central governments, and central limits on subnational tax rates. Taxes typically assigned to subnational governments include property taxes, user charges and a number of 'nuisance' taxes with little revenue potential. These raise only a small proportion of the total revenue of local governments even among the more advanced reformers--around 9 per cent in Estonia, and 6 per cent in Hungary, for instance (Ebrill and Havrylyshyn, 1999). In Poland and Romania, however, local taxes account for 26 and 17 per cent of local tax revenues, respectively.

Closely related to the issue of subnational tax autonomy is the widespread use of tax sharing arrangements, with revenues from taxes shared on a derivation basis, whose structures can only be changed at the central level, accounting for the largest share of regional revenue receipts. As seen in Table 5, in all transition countries tax sharing constitutes more than 50 per cent of total subnational tax revenue. Some of the advanced reformers in Central and Eastern Europe (Poland and Hungary) have the lowest tax sharing rates, whereas in Russia, Ukraine, and most Central Asian countries, revenues from taxes shared on a derivation basis continue to account for well over 60 per cent of regional revenue receipts. The minimal subnational autonomy to raise revenues and decide tax policies at the margin and the resultant mismatch between expenditure responsibilities and the real tax base, has important implications for accountability and responsibility at the subnational level. In many countries, limited formal revenue autonomy has encouraged the widespread use of informal revenue generating mechanisms, such as tax offsets and extrabudgetary funds.

Transfers

Countries in Eastern Europe and Baltics generally have relatively sound equalisation transfer systems (see Table 4). While the Czech Republic does not use equalisation transfers, it does however have formula-based features in its tax sharing system. However, the existing system has been criticised for being unwieldy, unstable, and nontransparent (Wetzel and Dunn, 2001). In a number of countries the equalisation transfer system, however, suffers from weaknesses that prevent if from achieving its goal of reducing the gap in fiscal revenue per capita across municipalities and intermediate level governments. For instance, in Romania the equalisation transfer system suffers from relying on an unpredictable pool of transfer funds, a formula that does not target funds to local councils with inadequate revenues to provide basic services; and lack of accountability at the county level for distributing funds according to published criteria. In Armenia, for example, equalisation transfers are supplemented by gap-filling transfers with associated drawbacks.

While a significant number of countries in the region have moved towards the increasing use of equalisation transfers, the actual volume of funds involved is either small or equalisation transfers only account for a small share of total transfers. For instance, given the relatively large fiscal disparities existent in Russia, the on-going level of funding for equalisation transfers in Russia (1.1 per cent of GDP in 1998) appears to be insufficient to bring about a significant level of equalisation. Although the overall funding for equalisation transfers is also quite limited in other countries such as Ukraine, Estonia, and Croatia, they have less pronounced fiscal disparities at the subnational level. In Hungary and Poland, where transfers account for half of local revenues, the overwhelming reliance on transfers is viewed as circumscribing local autonomy.

Gap-filling transfers, however, remain the norm in most FSU countries. This provides disincentives to local revenue mobilisation and cost savings through increased efficiency in delivery of services. In a number of countries (Bulgaria and most Central Asian countries), the allocation of transfers, in practice, has remained uneven and subjective, with a tendency for transfers to not get implemented as budgeted or, if they do, the actual flows have been unpredictable and subject to long delays.

Subnational borrowing

In recent years, subnational governments in a number of transition countries have been borrowing from a variety of sources, including, (i) the central government, (ii) subnational or national financial institutions (often regional institutions), and (iii) tapping domestic or international financial markets, by issuing domestic bonds or Eurobonds. However, there appears to be a wide variation in the subnational borrowing practices of countries and in the strictness of administrative controls over borrowing, largely reflecting the extent of development of financial markets and progress in other areas of intergovernmental fiscal relations (see Table 6). In many countries, while the overall level of subnational borrowing remains low, there has been an increasing trend towards greater subnational deficits, accumulation of debt, and loan guarantees.

In countries with relatively undeveloped financial markets (Belarus, Bulgaria, Moldova, and all Central Asian countries), subnational borrowing generally takes the form of central government loans. In many of these countries (such as Bulgaria and Albania), the market for municipal debt is underdeveloped because of the failure of municipalities to establish creditworthiness, which is a precondition for access to private credit. The limited revenue and spending autonomy in these countries and associated problems is reflected in the overwhelming reliance on the central government for financing subnational deficits. As a result, on the margin there is little difference between central government loans and transfers. One important threat to subnational budgetary discipline in all these countries is the moral hazard or impression of a soft budget constraint created by the practice of the central government granting loans that are eventually forgiven (Rodden et al., 2003).

Subnational borrowing from subnational or national state-controlled financial institutions, on the other hand, also poses significant macroeconomic risks. For instance, in Russia, commercial bank debt has become an important source of deficit finance, particularly since promissory notes (veksels) were disallowed since 1997. Subnational governments in Ukraine have also used veksels or bills of exchange as important means of financing. In 1998, over 16 per cent of all subnational tax collections in Ukraine were in the form of veksels and 17 per cent were collected in the form of tax offsets. However, these transactions are often non-transparent, subject to abuses and in many cases the loans are procured from commercial banks owned by regional governments.

A number of countries have gained access to private domestic and international sources of finance (such as the Czech Republic, Hungary, Poland, Russia, Ukraine, and Estonia). While liberalisation of subnational borrowing in Hungary, the Czech Republic, and Poland has been accompanied by relatively effective institutional and regulatory frameworks and increasing reliance on market-based discipline, in many countries such frameworks are virtually nonexistent. Other countries (such as Russia, Kazakhstan, and Estonia) have enacted legislation limits on overall debt as well as limits of the budget deficits of regions and require control and supervision of all subnational bond issues. However, the lack of effective monitoring and enforcement and the general absence of adequate municipal bankruptcy procedures (with the sole exception of Hungary) and financial emergency controls for defaulting governments pose important risks. For instance, while the regulatory framework in Estonia for municipal borrowings is sound, borrowing regulations are often ignored, there are no effective sanctions of violators and lenders may still perceive that there is an implicit sovereign guarantee. This has led to rapidly growing indebtedness of many municipalities with the potential need for future bailouts. In the Czech Republic there is no ex-ante control on local borrowing in place but the existing legislation states that the central government is not responsible for local debt. At the same time, determining the actual level of local indebtedness is often difficult due to the existence of contingent liabilities and other forms of off-budget operations (Oliveira and Martinez-Vazquez, 2001).

Building institutions for fiscal decentralisation

Effective implementation of fiscal decentralisation requires the presence of a comprehensive institutional framework. This holds in a number of important respects.

While local governments in a number of transition economies (eg, Hungary, Poland, the Czech Republic, the Baltics, and Russia) are led by democratically elected councils and governors/mayors, the heads of regional governments continue to be appointed by the central governments (eg, in Ukraine, Georgia, and all Central Asian states) (Table 7). Belarus, Azerbaijan, and Turkmenistan have no elections for subnational governments. The lack of democratic representation at lower levels of government may importantly affect the responsiveness of these governments, since conflicts between policies implemented and the preferences of local taxpayers are politically inconsequential. In cases of weak local influence on local officials decentralisation could simply lead to a transfer of power from national to local elites. With improved access of local elites to public resources this could increase opportunities for corruption. Furthermore, a system of appointed officials may imply that central political or other interests dominate local policy making, again adversely affecting the responsiveness of subnational governments to the interests of local citizens. Ultimately, the objective of enhanced accountability of regional and local governments is negatively affected.

International experiences have shown that a basic requirement for efficient multi-tier governments is the presence of intensive cooperation between the main stakeholders--the different levels of government. Countries have chosen very different ways of securing the required cooperation, but common experiences seem to indicate that an efficient system is characterised by transparent, regular, and comprehensive exchanges of information and discussions, and that cooperation must take place at the political as well as the technical level. Very few BRO countries have established such consultation mechanisms or cooperative bodies, with Estonia and Latvia being important exceptions. Other Eastern European economies, such as Hungary, Poland, and the Czech Republic also have such coordinating institutions in place.

At present, there is a clear lack of interaction and coordination between central government agencies (ministries of finance and line ministries) and regional government agencies (finance departments and sectoral departments) in many BRO countries. This lack of dialogue has occasionally led to unrealistic regulations, the proliferation of unfunded mandates, ineffective supervision and weak support and absence of performance evaluation of subnational programmes. More generally, it has encouraged conflicts and frictions in intergovernmental relations.

The lack of a modern tax administration has hampered both the day-to-day implementation of revenue assignments, and adversely affected general government revenue collections in many transition countries (eg, in Russia). In most BRO countries, the tax administration is a central government agency exclusively responsible for collecting taxes at all levels of government. Regional and local governments do not have their own tax administrations. However, the lack of effective control over the regional and local offices of the central tax administration and the de facto dual subordination of tax administrators to the central tax authorities and to subnational government officials has had an important impact on tax collections at all levels of government. Regional and local officials may be more interested in preserving the economic viability of local enterprises which provide employment and a tax base for subnational taxes than ensuring that federal taxes get paid. They may, therefore, pressure tax officials to be selective in their collection efforts. Furthermore, subnational officials may press tax administrators to employ more resources to the collection of subnational taxes than the low yield of these taxes may warrant. Even among the more advanced reformers, local capacity in tax collection is generally low, which reduces the effectiveness of tax collections (Ebel and Yilmaz, 2002).

With the exception of the more advanced reformers, budget process at the subnational level remain deficient in most transition economies and strengthening institutions for fiscal management remains a key challenge. Formulating budget objectives in a clear, transparent and realistic manner, strengthening budget execution, monitoring, and cash management at all levels of government, to varying degrees, are on the policy agenda of many of the intermediate reformers. However, progress in many of these areas has been slow.

CONCLUSION

Progress in reforming intergovernmental relations in the last decade has been uneven across the group of transition economies and across various components of intergovernmental fiscal relations. The nature and extent of decentralisation to date has been shaped in large measure by political, historical, and ethnic realities, and its effectiveness influenced by the institutional design and capacities of the various levels of government. In many countries, much has been accomplished in putting sound foundations in place, while, in others decentralisation reforms have been carried out without institutional and legal support mechanisms and appropriate intergovernmental fiscal arrangements to support a decentralised system.

In this paper we have provided an overview of key aspects of the ongoing decentralisation process in transition economies to illustrate instances where certain broad principles for sound and effective decentralisation are breached. To this end, it is important to note that there is no unique design or optimal degree of decentralisation that is 'appropriate' for all countries. The institutional context for decentralisation, including the overall level of economic development, ongoing economic and political reforms, existing technical and administrative capacity of subnational governments, geographic, demographic and other factors determines the design of intergovernmental fiscal system and ultimately affects the outcome of the fiscal decentralisation reform process. In general, institutional reforms that minimise adverse incentives and promote transparency, predictability, and accountability are key to an effective decentralised system. However, in the absence of strong institutional capacity and firm and transparent rules that regulate intergovernmental relations, forcing subnational governments to provide an adequate level of services and maintaining a sustainable decentralised system can pose a formidable challenge.

REFERENCES

Akai, N and Sakata, M. 2002: Fiscal decentralization contributes to economic growth: evidence from state-level cross-section data for the united states. Journal of Urban Economics 52: 93-108.

Alam, A and Sundberg, M. 2001: A Decade of Fiscal Transition. World Bank: Washington.

Banks, C and Pigey, J. 1998: Republic of Albania: opportunities and issues for municipal reform, Urban Institute Project 06610-113. Agency for International Development: Washington.

Bardhan, P and Mookherjee, D. 2000: Capture and governance at local and national levels. American Economic Review 90: 135-139.

Bird, R, Ebel, R and Wallich, C. 1995: Decentralization of the Socialist State. World Bank: Washington.

Dabla-Norris, E, Martinez-Vazquez, J and Norregaard, J. 2004: Fiscal Decentralization and Economic Performance: A Comparative Study of Russia, Ukraine, and Kazakhstan. International Monetary Fund: Washington, forthcoming.

Dabla-Norris, E, Martinez-Vazquez, J and Norregaard, J. 2000: Making decentralization work. Andrew Young School of Policy Studies Working Paper No. 9, Georgia State University.

Davoodi, H and Zou, H. 1998: Fiscal decentralization and economic growth: a cross-country study. Journal of Urban Economics 43: 244-257.

DeMelo, L. 2000: Fiscal decentralization and intergovernmental fiscal relations: a cross-country analysis. World Development 28: 365-380.

European Bank for Reconstruction and Development. 1998: Transition Report. EBRD: London.

Ebel, R and Yilmaz, S. 2002: On the measurement and impact of fiscal decentralization, Policy Research Working Paper No. 2809, World Bank: Washington DC.

Ebrill, L and Havrylyshyn, O. 1999: Tax Reforms in the Baltics, Russia, and Other Countries of the Former Soviet Union. IMF Occasional Paper No. 182, International Monetary Fund: Washington.

Fisman, R and Gatti, R. 2002: Decentralization and corruption: evidence across countries? Journal of Public Economics 83: 325-345.

Fornasari, F, Webb, S and Zou, H. 2000: The macroeconomic impact of decentralized spending and deficits: international evidence. Annals of Economics and Science 2: 403-433.

Gurgur, T and Shah, A. (2002): Localization and corruption: panacea or Pandora's box? In: Ahmad, E. and Tanzi, V. (eds). Managing Fiscal Decentralization. Routledge Press: London.

King, D and Ma, Y. 2001: Fiscal decentralization, Central Bank independence, and inflation. Economics Letters 72(1): 95-98.

Litwack, J, Sutherland, D and Lavrov, A. 2002: Fiscal federalist relations in Russia: a case for subnational autonomy. In: Ahmad, E. and Tanzi, V. (eds). Managing Fiscal Decentralization. Routledge Press: London.

Martinez-Vazquez, J, Timofeev, A and Boex, J. 2004: Russia's continuing transition toward a new federalism: subnational finance in the Russian Federation. World Bank Institute Learning Resources Series, Washington, DC.

Martinez-Vazquez, J and McNab, R. 2003: Fiscal decentralization and economic growth. World Development 31: 1597-1616.

Martinez-Vazquez, J and Boex, J. 2000: Russia's Transition to a New Federalism. World Bank: Washington.

McCullough, J, Spofford, T, Savoy, E and Ivanov, S. 2000: Bulgaria: comprehensive municipal finance and fiscal reform proposal. The Urban Institute (September).

McLure, C. 1995: Comment on 'The Dangers of Decentralization' by R. Prud'homme. World Bank Research Observer 10: 221-226.

OECD. 2001a: Fiscal Design Across Levels of Government. Country Report: Hungary.

OECD. 2001b: Fiscal Design Across Levels of Government. Country Report: Latvia.

OECD. 2001c: Fiscal Design Across Levels of Government. Country Report: Poland.

Oliveira, J and Martinez-Vazquez, J. 2001: Czech Republic: Intergovernmental fiscal relations in the transition. World Bank: Washington.

Prud'homme, R. 1995: The dangers of decentralization. World Bank Research Observer 10: 210-226.

Qian, Yand Weingast, B. 1997: Federalism as a commitment to preserving market incentives. Journal of Economic Perspectives 11: 83-92.

Rodden, J, Eskeland, G and Litwack, J. 2003: Fiscal Decentralization and the Challenge of Hard Budget Constraints. MIT Press: Cambridge.

Sewell, D. 1996: The dangers of decentralization according to Prud'homme: some further aspects. World Bank Research Observer 11: 143-150.

Shah, A. 1998: Fiscal federalism and macroeconomic governance--for better or for worse? Policy Research Working paper no. 2005, World Bank, Washington, DC.

Tanzi, V. 1995: Fiscal federalism and decentralization: a review of some efficiency and macroeconomic aspects. Proceedings of the World Bank Annual Conference on Development Economics.

Tanzi, Vand Tsibouris, G. 2000: Fiscal reform over ten years of transition. IMF Working paper 00/113, International Monetary Fund: Washington.

Ter-Minassian, T. 1997a: Decentralization and macroeconomic management. IMF Working Paper 97/115, International Monetary Fund: Washington.

Ter-Minassian, T. fed). 1997b: Fiscal Federalism in Theory and Practice. International Monetary Fund: Washington.

Treisman, D. 2000: The causes of corruption: A cross-national study. The Journal of Public Economics 76: 399-457.

Valdivieso, L. 1998: Macroeconomic developments in the Baltics, Russia, and other countries of the Former Soviet Union, 1992-97. IMF Occasional paper no. 175 International Monetary Fund: Washington.

Wallich, C. (ed). 1994: Russia and the Challenge of Fiscal Federalism. World Bank: Washington.

Wetzel, D and Papp, A. 2003: Strengthening hard budget constraints in hungary. In: Rodden, J., Eskeland, G. and Litwack, J. (eds). Fiscal Decentralization and the Challenge of Hard Budget Constraints. MIT Press.

Wetzel, D and Dunn, J. 2001: Decentralization in the Transition Economies: Challenges and the Road Ahead. PREM Unit Europe and Central Asia. World Bank: Washington, DC.

World Bank. 2000: World Development Report: Entering the 21st Century. Oxford University Press: Oxford.

Zhuravskaya, E. 2000: Incentives to provide local public goods: fiscal federalism, Russian style. Journal of Public Economics 76: 337-368.

(1) The author wishes to thank John Norregaard and Valerie Mercer-Blackman for sharing an early version of their dataset and participants at the OECD conference on Fiscal Decentralization in Russia for useful comments.

(2) See Bird et al. (1995), Wallich (1994), and Ter-Minassian (1997b) for a discussion of fiscal decentralisation experiences during the early years of transition.

(3) This principle suggests that provision of any given public service should be assigned to the lowest level of government that allows for the full internalisation of the costs and benefits associated with that service.

(4) The 1999 World Development Report notes that the proportion of countries with some form of democratic government rose from 28 per cent in 1974 to 61 per cent in 1988 (World Bank, 2000) partly as a response to the ongoing globalisation and integration of national economies with increasing circumscription of the powers of nation-states that it entails.

(5) See Prud'homme (1995), Tanzi (1995), and Ter-Minassian (1997a). McLure (1995) and Sewell (1996), among other, however, have questioned the validity of the adverse link between decentralisation and macroeconomic stability. Qian and Weingast (1997) stress the need for market-preserving federalism--a system of intergovernmental relations that is conducive to private sector development--as key to the relationship between decentralisation and sound macroeconomic performance.

(6) Fisman and Gatti (2002) report a strong negative relationship between expenditure decentralisation and corruption, Gurgur and Shah (2002) find that decentralisation results in greater public sector accountability and lower corruption in unitary rather than federal countries. However, Treisman (2000) does not find any significant relationship between fiscal decentralisation and corruption.

(7) Ebel and Yilmaz (2002) note that high-income OECD countries tend to be more decentralised than others, as measured by indicators of subnational share of revenues and expenditures.

(8) The experiences of Argentina, Brazil, and India with extensive borrowing by subnational governments illustrate clearly this problem. In these cases, the absence of effective limits on borrowing by subnational governments greatly complicated achieving overall fiscal tightening.

(9) The average size of Hungary's municipalities is 3,200 people and over half of the municipalities have a population below 1,000 (Wetzel and Papp, 2003). In the Czech Republic, 86 per cent of the municipalities have fewer than 1,500 inhabitants, and 42 per cent have fewer than 300 inhabitants (Oliveira and Martinez-Vazquez, 2001).

(10) In Russia, recommendations of the Kozak Commission have brought significant improvements in regional-local relations by mandating a uniform two-tier structure of local government and further clarifying assignments (Martinez-Vazquez et al., 2004). In Ukraine, much needed clarity was introduced in regional-local relations in the 2000-2001 Budget Code.

(11) For instance, the Law on Local Government in Hungary specifies certain potential responsibilities for local governments (urban development, housing, and waste management) that leave a degree of ambiguity in the system as local governments can 'pass up' a voluntary responsibility to a higher level if unable to meet the costs of providing that service (Wetzel and Papp, 2003).

(12) A closely related issue for transition countries has been budgetary management and control at different levels of government, including the way the budget is prepared and presented, its coverage, and the way the budget is executed. In many countries, weaknesses in budgetary management have interacted with opaque expenditure assignments to further constrain the effectiveness of expenditures.

(13) Dabla-Norris et el. (2004) and Wetzel and Dunn (2001).

(14) See Martinez-Vazquez and Boex (2000) and Zhuravskaya (2000) for the case of Russia.

(15) It is important to note that cross-country comparisons of subnational revenue autonomy as seen in a number of reports can in general be misleading. This is because such comparisons are based on GFS data for the share of subnational revenue in total general government revenue across countries. GFS data only measures the quantity of revenue that eventually ends up as being used by local governments and does not capture the extent of discretion or control by local governments associated with each type of revenue collected (Ebel and Yilmaz, 2002).

ERA DABLA-NORRIS (1)

IS 3-1600, International Monetary Fund, Washington, DC 20431, USA. E-mail: edablanorris@imf.org

Table 1: Structural aspects of fiscal decentralisation, 2001

Country Number of Number of Average
 subnational top tier population
 tiers (regions/
 province/
 oblast)

Armenia 2 11 336,000
Azerbaijan 2 71 107,000
Belarus 3 7 1,454,000
Estonia 2 15 96,000
Georgia 2 12 450,000
Kazakhstan 3 14 978,000
Kyrgyz Republic 3 7 (b) 645,000
Latvia 3 33 (d) 71,527
Lithuania 2 10 371,000
Moldova 2 11 390,000
Russia 3 89 1,652,000
Tajikistan 2 3 1,967,000
Ukraine 3 27 2,058,000
Uzbekistan 3 14 1,721,000
Albania 3 12 275,000
Bulgaria 2 9 921,000
Croatia 3 20 230,000
Czech Republic 39 14 740,000
Former Yugoslav 2
Republic of
Macedonia
Hungary 1 3,177 3,200
Poland 3 16 2,419,000
Romania 2 41 548,780
Yugoslavia 2

Country Number of Average
 lowest tier population
 (town/
 municipality)

Armenia 930 3,970
Azerbaijan -- --
Belarus 133 58,000
Estonia 247 6,000
Georgia ~1000 5,400
Kazakhstan 258 (a) 60,000
Kyrgyz Republic 549 (c) --
Latvia 541 2,219
Lithuania 56 66,000
Moldova 911 4,300
Russia 2337 63,000
Tajikistan 70 84,000
Ukraine 937 (e) 59,000
Uzbekistan 1749 14,000
Albania 374 (f) 9,000
Bulgaria 255 33,000
Croatia 423 10,900
Czech Republic 6292 1,700
Former Yugoslav 123 (h) 16,500
Republic of
Macedonia
Hungary 3177 3,200
Poland 2483 16,000
Romania 2948 (i) 7,632
Yugoslavia 187 55,500

Source: IMF country economists

(a) 173 rural rayons and 85 towns of rayon status.

(b) One independent city (Bishkek) functions as an oblast.

(c) 72 rayons, 19 towns and 458 village centres.

(d) 26 districts and 7 republican cities.

(e) 490 rayons and 447 municipalities.

(f) 65 communes and 309 municipalities.

(g) District offices have been phased out from January 2001.
This phase-out will be completed by the end of 2002.

(h) The number of municipalities may be reduced during 2002
to about 85.

(i) 2948 municipal councils (2686 communes, 182 towns and 80
municipalities).

Table 2: Structure of expenditures for each level of government,
most recent year (percent of total within each level of government)

 Level General Defense Social
 public and insurance
 services public and health
 order services (a)

Azerbaijan (1998) C 6.3 21.1 31.9
 R&L 4.3 0.0 22.7
Belarus (2000) C 4.6 8.5 41.7
 R&L 2.8 1.3 20.8
Estonia (2000) C 7.2 12.2 47.9
 R&L 11.2 0.3 12.1
Georgia (2000) C 7.4 12.6 35.4
 R&L 9.8 3.7 11.2
Kazakhstan (2000) C 6.6 14.2 42.6
 R&L 3.5 5.1 22.2
Kyrgyz Republic (2000) C 14.9 16.0 20.5
 R&L 9.4 1.5 28.4
Latvia (2000) C 5.4 9.5 52.8
 R&L 10.5 1.3 9.2
Lithuania (2000) C 4.2 11.4 52.5
 R&L 4.8 0.8 15.1
Moldova (2000) C 3.7 5.7 43.6
 R&L 9.4 3.5 26.0
Russia (2000) (c) C 7.6 17.8 32.5
 R&L 25.6 2.8 19.2
Tajikistan (2000) C 20.5 20.0 20.9
 R&L 10.2 3.0 18.2
Ukraine (2000) C 11.1 14.0 22.8
 R&L 4.0 1.4 34.7
Albania (1998) C 5.8 9.4 23.9
 R&L 10.4 0.0 33.5
Bulgaria (2000) (d,e) C 7.5 14.0 40.6
 R&L 7.4 0.7 36.0
Croatia (2000) C 3.7 12.9 55.0
 R&L 22.4 0.8 3.9
Czech Republic (2000) C 2.9 9.8 54.3
 R&L 15.4 3.9 16.6
Hungary (1999) C 4.9 5.5 32.3
 R&L 12.6 0.9 27.1
Poland (2000) C 4.0 7.5 51.9
 R&L 7.2 4.1 32.0
Romania (1999) C 4.1 9.1 43.4
 R&L 13.2 1.8 8.1

 Level Education Culture
 and
 recreation

Azerbaijan (1998) C 3.5 1.5
 R&L 60.2 4.4
Belarus (2000) C 3.9 1.6
 R&L 26.5 3.1
Estonia (2000) C 10.2 4.0
 R&L 38.9 11.4
Georgia (2000) C 3.6 2.9
 R&L 31.8 7.7
Kazakhstan (2000) C 3.6 1.2
 R&L 23.6 4.3
Kyrgyz Republic (2000) C 19.5 2.9
 R&L 45.0 3.4
Latvia (2000) C 5.6 2.1
 R&L 45.5 6.4
Lithuania (2000) C 5.8 1.9
 R&L 58.9 4.6
Moldova (2000) C 4.0 1.0
 R&L 38.2 2.7
Russia (2000) (c) C 2.3 0.7
 R&L 17.8 3.5
Tajikistan (2000) C 3.2 3.3
 R&L 39.8 2.5
Ukraine (2000) C 7.7 0.7
 R&L 23.3 3.5
Albania (1998) C 1.9 0.9
 R&L 41.6 2.6
Bulgaria (2000) (d,e) C 4.3 1.6
 R&L 31.0 2.8
Croatia (2000) C 7.9 1.2
 R&L 11.9 14.3
Czech Republic (2000) C 9.4 1.1
 R&L 7.5 7.0
Hungary (1999) C 7.7 1.6
 R&L 23.1 4.3
Poland (2000) C 4.7 0.5
 R&L 27.5 3.7
Romania (1999) C 9.8 1.0
 R&L 8.4 4.3

 Level Other (b) Total

Azerbaijan (1998) C 35.7 100.0
 R&L 8.4 100.0
Belarus (2000) C 39.8 100.0
 R&L 45.5 100.0
Estonia (2000) C 18.4 100.0
 R&L 26.0 100.0
Georgia (2000) C 38.1 100.0
 R&L 35.7 100.0
Kazakhstan (2000) C 31.7 100.0
 R&L 41.3 100.0
Kyrgyz Republic (2000) C 26.2 100.0
 R&L 12.3 100.0
Latvia (2000) C 24.5 100.0
 R&L 27.1 100.0
Lithuania (2000) C 24.3 100.0
 R&L 15.8 100.0
Moldova (2000) C 42.0 100.0
 R&L 20.2 100.0
Russia (2000) (c) C 39.1 100.0
 R&L 31.1 100.0
Tajikistan (2000) C 32.2 100.0
 R&L 26.1 100.0
Ukraine (2000) C 43.6 100.0
 R&L 33.1 100.0
Albania (1998) C 58.1 100.0
 R&L 11.9 100.0
Bulgaria (2000) (d,e) C 32.0 100.0
 R&L 22.2 100.0
Croatia (2000) C 19.3 100.0
 R&L 46.7 100.0
Czech Republic (2000) C 22.5 100.0
 R&L 49.5 100.0
Hungary (1999) C 48.0 100.0
 R&L 32.0 100.0
Poland (2000) C 31.2 100.0
 R&L 25.5 100.0
Romania (1999) C 32.6 100.0
 R&L 64.3 100.0

Sources: IMF country economists, Government Finance Statistics (IMF)

(a) GFS categories 'Health' and 'Social Security and Welfare.

(b) GFS categories 'Housing and Community Amenities,'
'Agriculture, Forestry, Fishing and Hunting; 'Mining,
Manufacturing and Construction; 'Transportation and
Communication; 'Other Economic Affairs and Services'
and 'Other Expenditures:

(c) Central government transfers to extrabudgetary funds
are included. General government includes wages and salaries.

(d) Some small contributions to Defense and Education by
both governments are included in the 'other' category.

(e) General government services comprises 'Wages and
Salaries' and 'Maintenance and Operations.'

Table 3: Structure of tax revenues at
each level of government, most recent year

Country Central/ Total subnational Income
 regional tax revenue taxes
 & local
 Percent Percent
 of GDP of total
 SNG
 revenues

Azerbaijan (1998) (a) C 22 25 40
 R&L 4 95 43
Belarus (2000) (b) C 12 41 42
 R&L 15 91 31
Estonia (2000) C 16 38 46
 R&L 5 83 89
Georgia (2000) C 9 22 62
 R&L 4 93 52
Kazakhstan (2000) C 31 8 53
 R&L 10 96 51
Kyrgyz Republic (2000) C 18 0 78
 R&L 2 75 38
Latvia (2000) C 13 39 46
 R&L 5 72 80
Lithuania (2000) C 13 35 51
 R&L 6 95 91
Moldova (2000) C 4 30 59
 R&L 4 77 55
Russia (2000) (c) C 14 34 37
 R&L 11 80 49
Tajikistan (2000) C 3 22 59
 R&L 4 94 40
Ukraine (2000) (d) C 24 9 60
 R&L 7 86 71
Albania (1998) C 9 18 52
 R&L 0 40 0
Bulgaria (2000) C 16 33 48
 R&L 3 77 90
Croatia (2000) C 10 34 49
 R&L 3 59 85
Czech Republic (2000) C 14 46 36
 R&L 5 71 91
Hungary (1999) C 22 35 38
 R&L 4 57 45
Poland (2000) C 21 35 42
 R&L 6 64 24
Romania (1999) (e) C 18 39 37
 R&L 3 84 77

Country Central/ Payroll taxes Consumption
 regional and social taxes
 & local security
 contributions

Azerbaijan (1998) (a) C 3 10
 R&L 0 40
Belarus (2000) (b) C 0.0 6
 R&L 0 57
Estonia (2000) C 0.0 0
 R&L 0 1
Georgia (2000) C 0.0 7
 R&L 0 11
Kazakhstan (2000) C 0.3 8
 R&L 31 8
Kyrgyz Republic (2000) C 0.0 4
 R&L 0 62
Latvia (2000) C 0.0 1
 R&L 0 2
Lithuania (2000) C 0.0 1
 R&L 0 0
Moldova (2000) C 0.0 7
 R&L 0 19
Russia (2000) (c) C 0.3 15
 R&L 0 30
Tajikistan (2000) C 1 15
 R&L 0 40
Ukraine (2000) (d) C 0.0 7
 R&L 0 13
Albania (1998) C 0.4 20
 R&L 0 92
Bulgaria (2000) C 0.0 3
 R&L 0 0
Croatia (2000) C 0.4 7
 R&L 0 4
Czech Republic (2000) C 1 2
 R&L 0 5
Hungary (1999) C 1 5
 R&L 0 44
Poland (2000) C 0.0 3
 R&L 55 2
Romania (1999) (e) C 0.0 6
 R&L 0 2

Country Central/ Property Other tax Total
 regional taxes revenue
 & local

Azerbaijan (1998) (a) C 100 11 100
 R&L 6
Belarus (2000) (b) C 100 5 100
 R&L 7
Estonia (2000) C 100 0 100
 R&L 10
Georgia (2000) C 100 11 100
 R&L 26
Kazakhstan (2000) C 100 0 100
 R&L 10
Kyrgyz Republic (2000) C 100 0 100
 R&L 0.0
Latvia (2000) C 100 0 100
 R&L 19
Lithuania (2000) C 100 0 100
 R&L 9
Moldova (2000) C 100 0 100
 R&L 26
Russia (2000) (c) C 100 11 100
 R&L 10
Tajikistan (2000) C 100 6 100
 R&L 14
Ukraine (2000) (d) C 100 5 100
 R&L 11
Albania (1998) C 100 8 100
 R&L 0.2
Bulgaria (2000) C 100 0 100
 R&L 10
Croatia (2000) C 100 0 100
 R&L 11
Czech Republic (2000) C 100 0 100
 R&L 5
Hungary (1999) C 100 0 100
 R&L 11
Poland (2000) C 100 0 100
 R&L 18
Romania (1999) (e) C 100 2 100
 R&L 18

Sources: IMF country economists, Government Finance Statistics

(a) Assumes that unaccounted-for local revenue is split
up evenly between 'other tax revenue' and transfers.

(b) Other revenue of central government includes property
taxes and other tax revenue.

(c) 'Other tax revenue' includes fees and other revenues.

(d) VAT and excises are no longer allocated to local budgets.

(e) Property taxes of subnational government includes land tax.

Table 4: Transfers (grants) from central to
subnational governments, most recent year

 Size Type of grant

 Percent Percent of General Specific
 of GDP total revenue

Azerbaijan (1998) 2.8 42 100 0

Belarus (2000) 3.0 15 100 0

Estonia (2000) 2.0 26 Yes Yes

Georgia (2000) 0.5 10 0 0

Kazakhstan (2000) 1.7 14 100 0

Kyrgyz Republic (2000) 2.6 49 100 0

Latvia (2000) 2.5 26 100 0

Lithuania (2000) 0.7 10 100 0

Moldova (2000) 2.7 34 100 0

Russia (2000) 1.4 9 Most Some

Tajikistan (2000) 1.2 24 90 10

Ukraine (2000) 2.5 23 60 40

Albania (1998) 5.4 96 100 0

Bulgaria (2000) 3.3 43 80 20

Croatia (2000) 0.2 5 Small Most

Czech Republic (2000) 2.3 25 0.0 100.0

Hungary (1999) 6.3 49 69 31

Poland (2000) 6.0 39 60 40

Romania (1999) 0.7 17 Yes Yes

 Properties of general grants

 Determinants of total amount

Azerbaijan (1998) Gap-filling.

Belarus (2000) Gap-filling.

Estonia (2000) Revenue equalisation.

Georgia (2000) 'Rich' regions make transfers to poor
 regions. Not transparent transfer system;
 based on negotiations between central and
 local government.

Kazakhstan (2000) Gap-filling after revenue sharing.
 'Fraternal' system introduced in 1999
 budget formulations which determines
 subventions to poorest regions and
 withdrawals from richest regions.

Kyrgyz Republic (2000) Gap-filling. Since 1998, categorical grants
 which fund wage spending for health and
 education at subnational level. Poor regions
 without sufficient tax revenues receive
 equalisation grant. Amount negotiated with
 Ministry of Finance.

Latvia (2000) Revenue equalisation; process initiated
 by the local governments.

Lithuania (2000) Expenditure needs based on inflation-adjusted
 previous year expenditures: 1. Allocated
 based on claims at the end of the year;
 2. Allocated in order to bring forecast
 tax revenues below national average and
 close to national average; and 3. Allocated
 on 'fraternal system' based relative measure
 of need adjusted with demographic coefficient.

Moldova (2000) Gap-filling, based on expenditure norms
 and revenue mobilisation capacity.

Russia (2000) Equalisation and gap-filling.

Tajikistan (2000) Gap-filling. Not transparent transfer system;
 based on negotiations between central and
 local government.

Ukraine (2000) Expenditure needs and population; Conditional
 grant.

Albania (1998) Gap filling.

Bulgaria (2000) Specific transfer for capital purposes and
 block grants. Transparent transfer system.

Croatia (2000) Equalization grant for specific purposes.
 Not transparent transfer system.

Czech Republic (2000) Grants are calculated each year when
 preparing the central government budget
 or distributed according to a specific
 legal frame work. The precise use of grants
 is defined by the central government. Typical
 criterion is per pupil, per bed etc.

Hungary (1999) Certain degree of equalisation. Dependency
 of municipal sector from grants is high.

Poland (2000) Equalisation block grant, capital
 expenditures. Equalisation grants bring
 localities up to 85% of national average
 per capita revenues. Transparent formula
 using objective criteria.

Romania (1999) Equalisation grant. Transparent transfer
 system.

Sources: Government Finance Statistics (IMF),
IMF country economists; Wetzel and Dunn (2001)

Table 5: Degree of tax sharing vs own financing
of subnational governments, most recent year

 Own taxes Own nontax Total own
 as % of revenue revenue
 total SNG as % of as % of
 revenue total SNG total SNG
 revenue revenue
 (ex grants)

Armenia 0.0 0.0

Azerbaijan (a) -- 8.0 --

Belarus (b) 6.0 4.1 10.1

Estonia 6.3 9.1 15.4

Georgia -- 6.2 --

Kazakhstan 0.0 3.2 3.2

Kyrgyz Rep. -- 12.7 --

Latvia 0.0 16.1 16.1

Lithuania 0.0 4.8 4.8

Moldova 15.4 12.4 27.8

Russia 34.4 9.1 43.5

Tajikistan -- --

Ukraine 30.9 30.9

Uzbekistan (d) -- --

Albania 0.0 1.6 1.6

Bulgaria -- 11.9 --

Croatia

Czech Republic 3.9 36.3 40.2

Macedonia, FYRM

Hungary 16.3 17.0 33.3

Poland (f) 10.6 24.6 35.2

Romania -- 12.6 --

Yugoslavia

Average

Standard deviation

 Total Distribution of total
 subnational subnational tax revenues
 tax revenue
 in percent Tax Own-
 of total sharing financing
 subnational
 revenues

Armenia 52.0 100.0 0.0

Azerbaijan (a) -- 55.0 45.0

Belarus (b) 96.0 93.8 6.2

Estonia 62.0 89.2 10.8

Georgia n.a. 95.0 5.0

Kazakhstan 82.0 100.0 0.0

Kyrgyz Rep. n.a. 64.6 35.4

Latvia 55.2 100.0 0.0

Lithuania -- --

Moldova 80.4 80.9 19.1

Russia 86.0 60.0 40.0

Tajikistan 39.4 95.0 5.0

Ukraine 87.0 64.4 35.6

Uzbekistan (d) 88.4 n.a. n.a.

Albania 35.7 100.0 0.0

Bulgaria 72.9 90.0 10.0

Croatia 55.8 85.0 15.0

Czech Republic 47.7 91.7 8.3

Macedonia, FYRM 86.4 -- --

Hungary 18.0 67.4 32.6

Poland (f) 40.0 57.7 42.3

Romania 73.7 75.0 25.0

Yugoslavia 71.5 8.0 92.0

Average 82.4

Standard deviation 16.1

 Degree of Autonomy

 To set To set
 tax base tax rate

Armenia None None

Azerbaijan (a) None None

Belarus (b) None Some discretion
 over local taxes

Estonia None Some discretion
 over local taxes

Georgia None (c) None (c)

Kazakhstan None None

Kyrgyz Rep. None None

Latvia None None

Lithuania None On land tax, up to
 4% ceiling.

Moldova None None

Russia None None

Tajikistan None Can set local tax
 rates

Ukraine None Very limited

Uzbekistan (d) Very limited Very limited

Albania None None

Bulgaria None None

Croatia Very limited Very limited

Czech Republic Limited Limited

Macedonia, FYRM None Local govt financing
 is currently under
 discussion

Hungary Some (e) Rate autonomy on
 substantial share
 of revenue

Poland (f) None Rate autonomy on
 substantial share
 of revenue

Romania Some (d) Some (g)

Yugoslavia None None

Average

Standard deviation

Source: IMF country economists

(a) Percent of local tax revenue from tax-sharing includes personal
income tax in which local governments' share is 100 percent.

(b) Revenue as a percent of GDP excludes Extrabudgetary Funds.

(c) Over shared taxes. Local governments impose a few minor taxes
and fees, and there is evidence of local governments creating tax
bases outside of the official mandate.

(d) Local authorities have discretion only with regard to granting
tax exemptions, delays and spacing out of tax payments.

(e) Building tax, land tax, communal tax on private individuals
and businesses, tourism tax, and local business tax.

(f) No autonomy with respect to share of income tax, and little
or none on real estate tax.

(g) Local councils can increase or decrease proposed tax and fee
rates within a range of 50 per cent.

Table 6: Borrowing by subnational governments

 Borrowing Actual Is there
 Allowed? amount regulation
 % of GDP for municipal
 At home Abroad bankruptcy?

Armenia Yes -- 0 No

Belarus Yes -- No

Estonia Yes Yes 2.6 (a) No

Georgia Yes

Kazakhstan Yes No n.a. No

Kyrgyz Republic Yes No

Latvia Yes Yes 0.40 No

Lithuania Yes No

Moldova Yes n.a.

Russia Yes No No

Tajikistan No Yes

Ukraine Yes Yes 0.9 No

Uzbekistan No

Albania Yes

Bulgaria Yes Yes (b) No

Croatia Yes

Czech Republic Yes Yes No

FYRM Yes No Very
Macedonia limited

Hungary (c) Yes Yes 0.1% Yes

Poland Yes No

Romania Yes Yes No

Slovakia Yes Yes No

Slovenia Yes No

Yugoslavia Yes No

Bosnia No

 Sources of financing

Armenia Only from home sources, not
 from abroad

Belarus In practice, build up of
 arrears.

Estonia Financial markets.

Georgia Can borrow from the central
 budget to meet temporary
 deficits.

Kazakhstan SNGs can borrow (1) from the
 Central Government to repay
 local debts; (2) from securities
 market for financing local
 investment programmes

Kyrgyz Republic

Latvia From central government;
 Minister of Finance is
 authorised to approve of
 another party to the loan
 agreement.

 Credits guaranteed by the
 government; other creditors.

Lithuania

Moldova Arrears to Social Funds and
 wage arrears.

Russia Bonds, bank loans and vexels.

Tajikistan Arrears

Ukraine In practice, build up of
 arrears. Most widespread:
 1. loans from central
 government. 2. Short-term
 loans from commercial banks.

Uzbekistan

Albania

Bulgaria

Croatia

Czech Republic Local government may borrow
 from any institution (even
 abroad) at any terms of
 interest rate, maturity etc.

FYRM
Macedonia

Hungary (c) Bonds and bank loans

Poland They can borrow from local
 market. Not clear if they can
 borrow from abroad.

Romania Bonds and bank loans. External
 credits guaranteed by the
 central government.

Slovakia Bonds and bank loans--both
 from local banks and domestic
 branches of foreign banks.

Slovenia

Yugoslavia Commercial banks

Bosnia

Source: IMF country economists

(a) Estimate is for 1998. This equivalent to
about 16 percent of local government revenue.

(b) Localities have not borrowed in the past,
except Sofia municipality.

(c) Refers to 2000.

Table 7: Political accountability in selected transition economies

Country Elections at Elections at
 regional level? Local level?

Armenia

 Executive No Yes

 Council Yes Yes

Azerbaijan No (district) Yes

Belarus No No

Estonia

 Executive No (district) Yes

 Council Yes Yes

Georgia

 Executive No Yes

 Council No Yes

Kazakhstan

 Executive No No (district)

 Council Yes Yes

Kyrgyz Republic

 Executive No No

 Council Yes Yes

Latvia

 Executive No (district) Yes

 Council Yes Yes

Lithuania No (district) Yes

Moldova No Yes

Russia Yes Yes

Tajikistan

 Executive No Yes

 Council Yes Yes

Ukraine No Yes

Uzbekistan

 Executive No Yes

 Council Yes Yes

Albania

 Executive No Yes

 Council Yes

Bulgaria No Yes

Czech Republic Yes Yes

FYRM Macedonia Local elections
 at the municipal
 level for mayors
 and local council
 members.

Hungary Yes (district) Yes

Poland

 Executive No Yes

 Council Yes Yes

Romania Yes (district) Yes

Country Comments on political
 accountability and elections:

Armenia

 Executive Central government appoints and dismisses
 heads of regional governments.

 Council

Azerbaijan Representatives of central government in
 lower levels.

Belarus President appoints regional governors, who
 then appoint subordinates.

Estonia

 Executive County governor appointed by central
 government on proposal of Prime minister.

 Council Local councils elected, local councils appoint
 mayor.

Georgia Top tier in general governed by central
 government appointed commissioners.

 Executive Two lowest tiers (municipal and district)
 have elected councils and executives.

 Council

Kazakhstan Oblast akims appointed by president. Local
 level akims appointed and removed by
 oblast-level. Regional and local councils
 elected.

 Executive

 Council

Kyrgyz Republic President appoints oblast and rayon
 governors. Heads of village councils
 are elected locally.

 Executive

 Council

Latvia Local governments are elected and form the
 district councils. District councils consist of
 lower levels chairs of municipal councils.
 Local councils are elected, local councils
 appoint mayor. No representatives of central
 government in local councils.

 Executive

 Council

Lithuania Central government appoints and dismisses
 governors of regional administrations. Local
 councils elected, local councils appoint
 mayor.

Moldova Local councils are elected. Executive selected
 locally but approved by the President.

Russia Representatives of central government in
 lower levels (super-districts).

Tajikistan Executive at regional and local level
 appointed by the president. President
 can remove local officials, including
 locally elected mayors.
 Executive

 Council

Ukraine Regional governors are appointed. Rayon and
 local level elected (but have in past been
 removed by the President).

Uzbekistan Regional executives are appointed. Local
 leaders selected from 'village elders:
 Regional and local councils are elected
 but elections influenced by higher level.

 Executive

 Council

Albania

 Executive Representatives of central government in 12
 regions replacing districts as administrative
 units by law in 2000.

 Council

Bulgaria Central government appoints governors of
 regional administrations. Local councils and
 local leaders (mayors) are directly elected for
 4-year terms.

Czech Republic

FYRM Macedonia

Hungary Representatives of central government in
 lower levels.

Poland

 Executive Vovoidship (regional council) elected, but
 head appointed by prime minister.

 Council

Romania Representatives of central government in
 lower levels.

Source: IMF