An Empirical Typology of Local Government Systems in Eastern Europe

Abstract

The most widely used typologies of European local government systems are based on research conducted in the 1980s. The most popular are those of Page and Goldsmith (1987), distinguishing between Northern and Southern European systems, and Hesse and Sharpe (1991), distinguishing between Southern, Northern and Anglo-Saxon models. The rare attempts to include the Eastern part of the continent are far from comprehensive or satisfactory. They usually view the whole region as a distinct group, referring to its specific historical background and recent radical decentralisation (Bennett 1993, Heinelt and Hlepas 2006). Disappointingly, the same approach is presented in the most recent comprehensive analysis of European local government systems (Loughlin et al. 2010). This article tries to fill the gap produced by this simplification, by offering a comprehensive picture of the variation within the Eastern European region and suggesting a first attempt at a typology of around 20 countries of the region. The criteria for this typology refers to those used in earlier classifications of the Western European systems and include: (i) territorial organisation and tiers of elected local governments, (ii) scope of functions provided by local governments (functional decentralisation), (iii) financial autonomy, (iv) horizontal power relations within local government institutions (election systems and relationships between mayors and councils).

Key Words:

• Local government
• decentralisation
• central-local relations
• typology
• Eastern Europe

1. Introduction

There are several well-established typologies of European local government systems (for example, Page and Goldsmith 1987, Hesse and Sharpe 1991, Heinelt and Hlepas 2006, Loughlin 2003, Loughlin et al. 2010. Some are referred to later in this article). Most of the academic literature on local governments treats Eastern Europe ¹ either as terra incognita, requiring exploratory investigation in the future, or puts the whole region into one basket described as ‘new local democracies’, coupled with the accompanying stereotypes. Existing typologies are methodologically sophisticated, but they focus almost entirely on the western part of the continent, usually ignoring the eastern part of Europe. This article is an attempt to take at least the first few steps towards filling this gap in the knowledge of the variation of European local government systems. It starts with a brief review of what existing typologies have had to say about Eastern Europe. The following section provides important examples of the variation among the group of Eastern European countries. This leads us to the conclusion that what is needed is either incorporation of local governments from the ‘New Europe’ into existing typologies developed for the Western European countries, or preparation of a completely new typology. This article does not aspire to respond to this demand in full, but it offers at least the first steps, showing the variation of the central-local government architecture in Eastern Europe and describing the main dimensions of this variation. In the following section it suggests a method of typology which is based on empirical evidence that is already available. Finally, it presents the results of such a typology conducted on the basis of hierarchical cluster analysis.

2. Eastern Europe in existing typologies of local government systems

The typologies most often referred to in the academic literature are:

•	Page and Goldsmith's (1987) research on central-local government relations (updated and developed later by John 2001), which distinguishes between Southern and Northern European systems. It is based on analysis of three dimensions of central-local relations: allocation of functions, discretion (local autonomy) and access (of local governments to central level politics);
•	Hesse and Sharpe (1991), who distinguish between three models: the Napoleonic tradition of Southern Europe, the Anglo-Saxon model and the model of Middle and Northern Europe.

In spite of the many years of their development and significant changes in both legal systems and management models in many countries, ² both typologies are often recalled in the contemporary literature and provide the foundation for much empirical research. A more recent typology by Loughlin (2003) identifies four models:

1.	the Franco model of Southern Europe;
2.	the Anglo-Saxon model;
3.	the Germanic model of Central Europe (later called the Rhinelandic model);
4.	the Nordic (or Scandinavian) model.

This typology has been repeated and developed in the recently published Oxford Handbook of Local and Regional Democracy in Europe (Loughlin et al. 2010), and extended to cover some of the Eastern European countries. The set of criteria used by Loughlin includes the constitutional position of local governments, state–society relations, the basis of policy style and the form of decentralisation and the dominant approach to discipline of public administration (2003, p. 5).

Recently, a new typology (although based on earlier results) was presented by Heinelt and Hlepas (2006). In their research they focused on vertical (tiered) and horizontal power relations (with a special focus on the position of the mayor). The first dimension refers to the earlier typology by Hesse and Sharpe (1991), while the latter one refers to the classification of leadership styles by Mouritzen and Svara (2002). By adding these two dimensions they arrive at a matrix which represents several blends of horizontal and vertical power relations. What is important from our point of view is that Heinelt and Hlepas (2006) include three countries of Eastern Europe in their typology (Czech Republic, Hungary and Poland), which in their opinion should be classified as a separate type.

The typologies recalled above, especially those developed before 2000, usually ignore the existence of local governments in Eastern Europe. Occasional exceptions to this rule treat countries of the eastern part of the continent as a uniform group. In their summary of earlier typologies, Heinelt and Hlepas (2006, p. 27) justify such an approach in following way:

Although local-central relations have some features in common with the North and Middle European [the term refers to Hesse and Sharpe's terminology] group with respect to local competencies and fiscal or financial discretion … they are considered as a distinct group because their historical background and in particular recent radical decentralization in these countries needs to be taken into account in discussion of vertical power relations. These power relations are (as different schools of neo-institutionalism emphasize) not just characterized by certain formal (legal) rules for the distribution of competencies and resources but also by particular meaning of what is perceived as appropriate – or inappropriate.

Similarly, Loughlin et al. (2010, p. 724) justify their grouping of all Eastern European countries into one type by the common history of countries that

Share a common experience of communist dictatorship … they also share a common experience of the transition to democracy and preparation for, and accession into, the European Union [their analysis ignored countries of Eastern Europe other than those who became new member states of the EU in the 2004–2007 period] … The legacy of that period was political systems marked by high levels of centralization and uniformity.

Although they also notice the variation of historical traditions (German, Ottoman, Russian), especially in the nineteenth century when the modern administrative systems were emerging. Loughlin et al. (2010, p. 725), in contrast to Heinelt and Hlepas, see similarities to the French Napoleonic tradition, referring to the pre-communist period of long-standing affinities between France and countries like Poland and Romania, Marxist-Leninist admiration for the French Jacobin tradition, and nineteenth-century Balkan and Eastern European nationalists referring to France as the exemplar of the modern progressive state.

Bennett also (1993) treats all Eastern European countries as one group, but his approach may be explained by the time in which he published his research – 1993 was a very early stage of political transformation in the region, when all local government systems in this part of the continent were still far from stable.

These attempts are hardly sufficient to understand the variety of the local government systems across the continent, since they ignore the huge variation within the group of Eastern European countries. The differences are so great that their local government systems cannot be treated as a single, uniform group.

3. Local government systems in Eastern Europe – variation and common issues

There is no doubt that local government systems in the eastern part of the continent share some common features and that their roots can be identified in history. Such similarities include:

1.	Belief in ideas of decentralisation, which has almost been identified with democratisation. This belief was common at the beginning of the transformation, but later disappeared, at least in some of the countries analysed. Coulson and Campbell (2006, p. 556) rightly noticed that: ‘In most of the transitional countries a De Tocquevillian myth of localism had flourished in opposition circles in the years preceding the fall of the communism, in which local self-government was to be the incarnation of civil society and everything that the regime was not.’ This common belief has not led to the creation of the common, uniform institutional model of local government (as illustrated by Peteri and Zentai 2002).
2.	Weakness of the meso (above municipal) tier of elected local government. This feature is almost opposite to the dominant trend in several Western democracies, where regional tiers had been created or strengthened within the last 20–30 years (see discussion in Sharpe 1993). In Eastern Europe the last 20 years has seen a frequent weakening of the role of the second tier. The lack of the regional or county tier is easy to understand in relatively small countries, such as Latvia, Lithuania, Estonia or Slovenia. The same phenomenon may also be noticed in larger countries. In Bulgaria there is no upper tier of elected sub-national governments, in spite of constitutional provisions for regional government (Kandeva 2001). In Hungary, 19 counties (megye) plus 22 cities of county status survived, but their scope of functions and competences have been seriously reduced compared to the earlier period (Soós and Kakai 2010). In Croatia, the position of 21 counties (županje) was weakened in the 1990s, when they lost their right to supervise the activities of municipal governments (Alibegović and Slijepčević 2010). In Poland, the Czech Republic and Slovakia, the first wave of decentralisation reform (in 1990) concerned only the municipal tier. The county and regional tiers were not reformed in Poland until 1999. Regional governments (kraj) were established in the Czech Republic in 2000, and in Slovakia only in 2002. Moreover, the functions of counties and regions are usually relatively narrow compared to municipalities. The budgets of the counties and regions rely mostly on transfers from the central level and their influence on activities (for example, through coordination) of the lower tiers’ governments is usually minimal. Most of the important public services are provided by municipalities, which are also the only tier with (limited) powers of taxation. Poland is an exception to this rule, but only to a limited extent. Polish regions were created (in 1998) as financially and functionally weak entities. However, giving them the status of NUTS-2 European Union (EU) statistical units allowed for effective lobbying for their important role in the implementation of European cohesion policies at the regional level. The role of regions in preparation and implementation of regional operating programmes (financed from EU structural funds) is perhaps the most important. The alternative model (to weak elected regional self-government) of territorial organisation is represented by Ukraine, and to a lesser extent, Moldova. In this case the role of the regional tier is significant both in terms of the scope of delivered functions, and in terms of the role of local entities in coordination. The Ukrainian oblast is a more deconcentrated unit of state administration than an elected regional government (see Swianiewicz 2006). To illustrate this claim one should note that the head of the regional administration does not have local electoral legitimacy, but is nominated by the president of Ukraine. The roots of the ‘meso level’ weakness may be found in the history of public administration before 1990. In that period, counties or regions usually exerted direct control over municipalities and their strong position was seen as a barrier to local autonomy and democracy. As a result, after 1990 the municipal level politicians were reluctant to provide municipalities with a wide scope of competencies. The authors of the reforms feared that the strong upper tier would continue to limit the autonomy of the lowest levels. Consequently, in countries which opted for deep decentralisation, the importance of upper tiers of sub-national governments was limited by this fear. In countries which remained strongly centralised, the regional tier has stayed part of the national administration, through which relatively weak local governments could be controlled.
3.	Decentralisation reforms in Eastern Europe took place in a period of dynamic change in local governments in Western Europe. These changes were related to new trends in management, which have often bifurcated to countries of Eastern Europe (frequently through the experts working for individual aid and development programmes 3 ). Coulson and Campbell (2006, p. 544) wrote about this process in the following way: EE countries have had to introduce the Rule of Law (in the Weberian sense) and the New Public Management in parallel, whereas in the West the one preceded the other by many decades, so that in the East the tension between legalism and managerialism may threaten to fragment the local authority as an institution.

One can also find a few other (relatively minor) common features for local governments in the region. Although they are more visible in the first years of political transformation, what is much more striking is the number of very strong variations within the analysed group of countries. Some dimensions of this variation include:

Figure 1. Functional decentralisation in Central and Eastern Europe – local government spending as a percentage of GDP (2007)

Figure 2. Distribution of municipal governments in Eastern Europe according to their population size

Figure 3. Percentage of local politicians not affiliated with national political parties (2001–2002)

1.	Functional decentralisation. The scope of functions local governments are responsible for in individual countries of the region is very different. This is illustrated in Figure 1 by simplifying the index of elected local government spending expressed as a percentage of gross domestic product (GDP). In none of the countries is the value close to the value noted in Scandinavian countries. It is also lower than in the case of most federal Germanic countries. But in the Czech Republic, Hungary or Poland it is higher than average for the remaining types of Western European systems. At the same time, in some other countries of the region the value of the index is among the lowest in Europe. (For easier comparison, the averages for four types of Western European systems identified by Loughlin (2003) are presented on the right side of the figure.) The variation within the region is very significant.
2.	Territorial organisation, and – especially – the level of territorial fragmentation of the municipal tier. Territorial organisation is not necessarily a stable feature of the local government system. During the last two decades, we witnessed strong territorial fragmentation in several countries (for example, Czech Republic, Hungary, Slovakia, Croatia and Macedonia) as well as territorial consolidation reforms (for example, Lithuania, Georgia and Macedonia). 4 As a result of both processes, the current size of local governments is very much diversified among the countries of the region (Figure 2). Countries like Lithuania, Georgia, Serbia and Bulgaria are among the largest in Europe in terms of local government size, while at the same time countries like Czech Republic, Slovakia, Hungary, Armenia and Azerbaijan are among the most territorially fragmented.
3.	Local electoral systems and position of the mayor. Among the countries of Eastern Europe there are countries in which local councils are elected according to both proportional and majoritarian systems. There are also countries in which both systems are applied, depending on the population size of the municipality – majoritarian systems being preferred in smaller units, and proportional systems enjoying more popularity in the larger units (for example, Hungary, Poland, Slovenia). The eastern part of the continent has followed the ‘fashion’ of personalised, individual leadership by a directly elected mayor. 5 In some countries this model had already been applied in 1990 (for example, Slovakia and Bulgaria), while in others it was introduced in the following years (for example, Hungary, Poland and Croatia). However, there is still a group of countries which maintain another model based on a collective board appointed by the local council (Czech Republic, Lithuania, Latvia, Estonia, Serbia and, since 2006, Georgia).
4.	The role of national political parties in local governments. After the political turnover, the beginning of the 1990s was in a natural way a period of radical (but lengthy) reconstruction of party systems. The new national parties have infiltrated local governments to a different extent. Poland, and to a lesser extent also Hungary, belong to the group in which the role of national parties in local government is the lowest among all European countries. This is reflected in the low party membership among both local councilors and mayors. Independent local politicians may be found mostly in small municipalities, but mayors or significant groups of councillors who are not formally affiliated with any of major political parties can also be found in some of the largest cities (for example, Wrocław, Poland). At the same time, in the majority of countries in the region, local governments are strongly partisan (for example, Bulgaria, Albania, Georgia and many others). This variation is briefly illustrated in Figure 3 (which includes a group of seven countries for which comparable data can be found).

As illustrated by the few issues discussed above, the variation among the local government systems in the region is very significant and must not be ignored. This provides a dramatic mismatch between the actual variation and the uniform treatment of the region which dominates the mainstream academic literature at the moment.

4. Construction of the typology

The critical comments formulated above do not mean that ‘old’ typologies are useless. They help us to define the dimensions we should take into consideration to build the typology of Eastern European local government systems. As Heinelt and Hlepas (2006) suggest, the ‘new typology’ should take into account both horizontal and vertical power relations. An important point of reference should be the three dimensions of central-local relations identified by Page and Goldsmith (1987): functions which are allocated to local governments, discretion in managing their functions and access to policies formulated at the national level. Because of limited access to comparable, qualitative characteristics of the system (collection of such information would require extensive studies in all of the involved countries), the typology presented in this article concentrates on measurable features of the institutional systems. The discussion below, to a small extent, refers to differences of political cultures or informal power relations that are more difficult to measure.

This typology has been conducted on the basis of measurable indicators, using the Ward method of cluster analysis. In this method one does not create ideal types which are compared to actual cases. Instead, one looks for clusters of cases which are relatively close to each other in terms of applied criteria. This means that the typology is inductive rather than deductive in nature (although it is based on theoretical assumptions concerning the set of criteria taken into consideration).

The typology covers 19 countries in what is broadly defined as Eastern Europe:

1.	Albania
2.	Armenia
3.	Azerbaijan
4.	Bulgaria
5.	Croatia
6.	Czech Republic
7.	Estonia
8.	Georgia
9.	Hungary
10.	Latvia
11.	Lithuania
12.	Macedonia
13.	Moldova
14.	Poland
15.	Romania
16.	Serbia
17.	Slovakia
18.	Slovenia
19.	Ukraine

As one can see, the definition of ‘Eastern Europe’ has also been extended to countries of South Caucasus (Armenia, Azerbaijan and Georgia), which according to most of the strict geographical definitions are not located in Europe. There are at least two important arguments for including them in our discussion. First, their joint heritage of transition from the Soviet system of public administration makes them comparable to other East European countries. Second, all of these countries often refer to their European cultural traditions, have ratified the European Charter of Local Governments and are members of the Council of Europe, which has in many cases played an important role in steering the process of decentralisation reforms. The list of countries is not complete. The countries still missing include Belarus, Bosnia and Herzegovina, Kosovo and Montenegro (due to the scarcity of comparable data). Because of its incomparable scale, Russia has also been excluded from the typology. Nonetheless, the set of countries which is considered in the typology is nearly complete.

This typology method refers to data on four dimensions of local government institutional systems, which are described below. Each of them is measured by one or more indicators. The list refers to two out of three dimensions discussed by Page and Goldsmith (1987). The ‘access’ dimension is missing, since it is the most difficult to measure without in-depth qualitative research.

The typology has been constructed on the basis of the following indicators:

1.	Territorial organisation, which is illustrated by two variables: • Number of tiers of elected local and regional governments. In this variable, only those tiers which have both elected representatives and locally appointed, locally accountable executives are taken into account. For example, the Ukrainian oblast and rayon level have not been considered. Although there is an elected council on these levels, the administration is vertically subordinated towards the centre, and local autonomy of personal nominations and decision-making is very limited. • Territorial fragmentation, measured by the average population size of municipal government unit. The level of territorial fragmentation proved to be an important explanatory factor for the horizontal power relations in several studies, including Page and Goldsmith (1987), discussed already in this article. 6
2.	Functional decentralisation – measured by the share of sub-national expenditures in GDP. The variable used to measure functional decentralisation provides a simplified picture. One could argue that the share of local spending in total public spending would be a more accurate alternative measure. However, difficulty in comparing public finance systems in individual countries (which include not only public budgets but also the variety of non-budgetary funds dedicated to health care, social insurance, and so on) makes this alternative index extremely difficult to compose. In such circumstances the decision was made to use the variable which is simpler but much easier to interpret and to avoid methodological traps.
3.	Financial discretion of local governments. This dimension is measured by three variables: • Financial decentralisation – measured by the share of locally controlled taxes in total revenues. Only revenues which may be directly influenced by local governments (through decisions on tax rates or tax exemptions) are considered. This approach is to some extent a simplification, since it does not take into account what the scope of discretion is. In particular, it does not distinguish local taxes, which can be freely shaped by local decisions, from those in which the local autonomy is seriously limited (for example, there are maximal rates which are set on a national level). Nevertheless, this measure seems to be precise enough for our purposes. • The shape of the grant allocation system. To what extent is it based on a stable and measurable formula (which is resistant to the political manipulations of both the central and local levels) and to what extent does the allocation depend on arbitrary decisions made by the grantor? This is the only variable which to some extent has a quantitative character. Individual countries receive scores between zero and four, and the score is based on the literature concerning individual countries and individual research experience of the author. The score ‘0’ means a dominant role of arbitrary decisions made by central level bureaucrats and frequent political manipulations concerning the size of the grant transfers. Score ‘4’ means a formalised formula, which is not vulnerable to political manipulations. Scores ‘1’ to ‘3’ mean situations in between the two extremes, where elements of the formula-based approach mix with some arbitrary decisions. 7 Traditionally, grant allocation was one of the most important means for exercising power and control by the higher tiers towards the lower tiers of government. Therefore, the extent to which the system has been transformed into a formula-based one is a very important criterion. • Local government debt as a percentage of GDP. This variable requires additional justification. Although access to local borrowing is guaranteed in the European Charter of Local Governments, in practice the regulations in this respect were adopted with considerable delay in several countries. The actual implementation of the charter has been even more problematic. In practice, access to borrowing had been strongly restricted or even impossible in several countries until recently. In some other countries, these restrictions are still in place. 8 At the same time, in none of the considered countries has the level of local government debt ever exceeded the level which could have negative macroeconomic consequences. The cases of excessive indebtedness of individual local governments are also very rare. Taking this into account, I have decided to consider the level of local government debt (expressed as a percentage of GDP) as one of the indicators of local fiscal autonomy, and as a measure of financial decentralisation – the higher the value, the more local financial discretion. 9
4.	Horizontal power relations. This dimension is measured by two variables: • The position of the mayor in the municipal government. The main difference is between a directly elected strong mayor, and the collective form of a board appointed by the local council. • The electoral system of the local council (majoritarian, proportional or mixed).

As is clear from the list of variables presented above, the special focus is on the municipal level of government. It should be recalled once again that in several countries there is only one sub-national tier, namely the municipal one, while in others (Croatia, Czech Republic, Hungary, Poland and Slovakia) the upper tiers are much less important for public service provision and have more restricted discretion than municipalities to determine their own policies.

The sources of information used to measure values of individual variables are very much dispersed in various publications. The most useful sources include the comprehensive set of publications prepared within the framework of the LGI programme (Local Government and Public Service Initiative), which has operated under the auspices of the Open Society Institute funded by George Soros in the 1995–2010 period (see especially: Hogye 2000, Horvath 2000, Kandeva 2001, Peteri and Zentai 2002), as well as the recently published book edited by Šević (2008) on local government finance in Eastern Europe.

5. Results of the typology

The typology conducted according to the criteria and procedures described above, allows identification of five clusters of countries.

5.1. Type I: Hungary, Poland and Slovakia

Taking into account the characteristic features of local government systems in these three countries (see also the summary in Table 1) Type I may be called champions of decentralisation. The wide scope of functions provided by local governments (especially in Poland and Hungary, and to a lesser extent, Slovakia) locates this group closer to Northern European countries, known from earlier academic typologies of Western European local government systems.

1 Results of typology – summary

	Type I	Type II	Type III	Type IV	Type V
Number of tiers of elected local governments	2–3	1–2	1–2	1–2	1
Territorial fragmentation of the municipal tier	Diversified (consolidation in Poland, small municipalities in Hungary and Slovakia)	Small municipalities	Moderate	Very big local government units	Small municipalities
Functional decentralisation (mean share of local spending in GDP – 2007*)	Wide scope of functions (10.5% of GDP)	Wide scope of functions (10.5% of GDP)	Moderate (6.6% of GDP)	Moderate (6.3% of GDP)	Narrow scope of functions (0.9% of GDP)
Financial decentralisation – mean share of locally controlled fees and taxes in GDP (2007*)	Considerable (3.1%)	Small (0.7%)	Small (0.8%)	Considerable (3.1%)	Negligible (0.1%)
Allocation of grant transfers	Based on the formula	Based on the formula	Partially based on the formula	Diversified	Major role of arbitrary decisions of central level politicians and bureaucrats
Local debt as % of GDP (2007*)	2.3%	2.8%	0.4%	0.6%	0%
Position of the mayor/type of leadership	Strong mayor, direct elections	Collective leadership	Strong mayor, direct elections	Collective leadership	Diversified
Council – electoral system	Majoritarian in small, proportional (or mixed) in the biggest municipalities	Proportional	Proportional	Proportional	Majoritarian
Countries	Hungary, Poland, Slovakia	Czech Republic, Estonia, Latvia	Albania, Bulgaria, Croatia, Macedonia, Moldova, Romania, Slovenia, Ukraine	Georgia, Lithuania, Serbia	Armenia, Azerbaijan
Notes: *Except for: 2006 – for Albania, Moldova, Serbia, 2005 for Armenia, Azerbaijan, Croatia, Georgia, Macedonia, Ukraine.
Source: Own calculations based on: Alibegović and Slijepčević (2010), Bajo and Jakir-Bajo (2008), Bercu and Onofrei 2006, Dąbrowska et al. (2009), Levitas and Peteri (2006), Local Government Finance (2010), Meekel (2008), Mshvidobadze (2006), Nikolov (2006), Pigey et al. (2008), Sedmihradská (2010), Slukhai (2008), Swianiewicz (2006), Swianiewicz and Mielczarek (2010), Taxation Trends (2010), Tumanyan et al. (2007).

The common feature is also the direct election of mayors – introduced in Slovakia in 1990, in Hungary in 1994 and in Poland in 2002. In most municipalities there is a majoritarian electoral system for the local councils. This includes all of Slovakia, local governments in Hungary with a population below 10,000, and local governments in Poland with a population below 20,000 ¹⁰ ).

This group is also characterised by the high level of financial autonomy, at least compared to other countries of the region. Local governments have a right to decide on local tax rates, grant transfers are usually allocated according to transparent (although often criticised) criteria and, for more than a decade, local governments have also enjoyed considerable autonomy in implementing their own borrowing policies in order to finance development projects.

In each of these countries, there is at least one tier of elected sub-national government above the municipal level (megye in Hungary, kraj in Slovakia, powiat and województwo in Poland). Only in Poland does the regional level play an important role in implementing European cohesion policy (which is possible due to the territorial overlap between the elected regions and NUTS-2 units). One of the main differences among the countries of this group is related to municipal territorial organisation. In Slovakia and Hungary, almost every town or village forms its own local government. Consequently, the average size of local government is several times lower than in Poland, where a single municipal government usually covers more than a dozen small towns and villages.

5.2. Type II: Czech Republic, Estonia and Latvia

This type may be called ‘relatively decentralised’. The scope of functions local governments are responsible for is similar to that identified in Type I countries above. The main difference, however, is the negligible role of locally controlled taxes in Type II countries. Tax revenues are most often shares in central government taxes, while discretion to decide upon local tax policies concerns a very tiny proportion of revenues.

The other common features of this group include: a high level of territorial fragmentation ¹¹ (most clearly visible in the Czech Republic, where over 6000 municipalities retain the French model of territorial organisation) and collective leadership, based on a board appointed by the local council.

5.3. Type III: Albania, Bulgaria, Croatia, Macedonia, Moldova, Romania, Slovenia and Ukraine

With some simplification, this cluster may be called Balkan – since most countries classified in this type are located in the Balkans (although the group also includes Ukraine). While in Types I and II we find only the new EU member states which joined the EU in 2004, Type III has only one country belonging to this group: Slovenia. The remaining countries either joined the EU a few years later (Bulgaria and Romania) or still remain outside the EU structure (Albania, Croatia, Macedonia, Moldova and Ukraine).

In each of these countries the attempts at decentralisation reforms have been undertaken in different periods and with varied levels of success. However, the scope of local government functions is usually much narrower than in Types I and II. Due to this criterion and also to geographical location, this type is much closer to the Southern European model identified in earlier academic classifications.

The level of financial autonomy in Type III is also lower than in Types I and II. Examples of late reforms in this respect are provided by Albania and Macedonia, where local government borrowing had only been allowed in 2008. Financing local investments through debt instruments is also still very rare in Ukraine and Moldova.

The common feature of the Type III countries group is direct election of mayors (in some countries – like Croatia – introduced very recently) and the model of strong personal leadership.

The level of territorial fragmentation at the municipal level is usually moderate. The two extreme cases in the Type III group are Bulgaria (with very big local government units) and Ukraine (which has a huge number of very small local governments). Macedonia has undergone territorial reforms twice – first, big municipalities inherited after Yugoslavia were divided into more than 120 smaller units, and later, in 2004, the number of municipalities was reduced to 84 (although they are still smaller than those that existed before 1990).

5.4. Type IV: Georgia, Lithuania and Serbia

The main feature these countries share is a specific territorial organisation, namely, a very high level of territorial consolidation. In Serbia, this model has remained unchanged for well over 20 years, in Lithuania it was introduced by 1998 reforms, and in Georgia in 2006. ¹² One of the official aims of territorial reforms is usually decentralisation and extension of local capacity to deliver public functions. But as is visible in Figure 1, the large size of local government units does not have this effect in the Type IV countries (the share of local spending in GDP varies from 8% in Lithuania to less than 7% in Serbia, and only 3.7% in Georgia).

The other common feature of these countries is the collective type of leadership ¹³ (with the board appointed by the council) and a proportional local electoral system. In the case of other variables considered in this typology, the variation among the three countries in this group is significant.

5.5. Type V: Armenia and Azerbaijan

In spite of the political and cultural differences between these two countries, they share several characteristics of their local government systems. These are, in particular: strong functional and financial centralisation as well as territorial fragmentation of local government units. Both the scope of functions and the level of local discretion are clearly smaller than in most countries classified in other types of this typology. ¹⁴

6. Summary and conclusions

The typology presented here reflects very significant variations of local government systems within the group of eastern European countries. It confirms the claim that treating the whole region as one homogenous group, which would constitute a single, separate type of European local government, is a mistake.

However, the results presented in this article provide only the first step towards the integration of our understanding of the variation of local government systems across Europe. There are at least three more steps to be taken in the future. The first step would be to carry out more in-depth research, based on a common methodological framework, which would allow inclusion of qualitative criteria related to the nature of central-local relations into our typology.

Second, one should ask the question, to what extent might results of this typology be linked with results of the typologies conducted earlier in Western European countries. The answer to this question is not an easy one. It should only be of a preliminary, tentative nature, especially keeping in mind the lack of sufficient, in-depth qualitative studies. The evolution of Western European systems after the publication of the most classic typologies provides additional complications. For example, in several northern European countries (for example, Germany and England) there are direct elections of strong mayors, which were almost absent in this part of the continent when Page and Goldsmith (1987) and Hesse and Sharpe (1991) were formulating their conclusions. The clear distinction between various models of territorial systems (within territorially fragmented Southern Europe) has been questioned by the territorial consolidation reforms in Greece (Hlepas 2010). One can easily multiply the number of complicated examples of such organisational evolution (see also discussion in Goldsmith and Page 2010). Although there are also some newer studies (some of them quoted above), the major framework of almost all of typologies still relies on the results of empirical research conducted at the end of the 1980s.

Nevertheless, one could pose a set of very tentative comparative observations.

i.	None of the Eastern European countries have local government systems which are even similar to that identified by Loughlin's Nordic (or Scandinavian) model. The very wide scope of functional decentralisation, the crucial role of local governments in maintaining the welfare state system, and the specific system of local financial autonomy (connected with strong horizontal equalisation) are very different from the reality seen in the eastern part of the continent.
ii.	There are no countries in Western Europe that are as centralised as the Type V countries, including Armenia and Azerbaijan. In a slightly less categorical tone, the same observation may be related to other strongly centralised countries classified within Type III.
iii.	Countries of Types I and II demonstrate several similarities to the Northern European model identified by Page and Goldsmith (1987) and Hesse and Sharpe (1991). This similarity has already been noted by Heinelt and Hlepas (2006) in their earlier study. Poland, with its territorial consolidation of the municipal tier, and having the highest share of local spending in GDP in the region, is probably the closest to the typical characteristics of this model. However, in Eastern Europe there are no examples of the federal systems, 15 which according to Loughlin is a characteristic element of the north-western section of the continent. At the same time there are features of Types I and II countries which retain some elements of the Southern European systems. These features can be found in some elements of the political culture, but also in some of the institutions (for example, Polish RIO – Regional Chambers of Accountancy, which have been constructed to a large extent on the model of the French Chambre Regionale des Comptes).
iv.	The Type III (covering mostly the Balkan countries) category shows several similarities to the Southern European or Franco model.

The third research challenge ahead is to develop a model which will explain the pattern of variation among Eastern European countries. One possible path may be to examine the relationship between EU accession and the shape of institutional frameworks for local governments. However, this approach does not provide any immediate answers. Both new EU member states and other countries of the region may be found in different clusters of our typology, and most decisions regarding institutional structures were made either before, or with no relation to, the requirements of EU accession. As Hughes et al. (2003, p. 82) note in relation to Poland (their observation might also be extended to other countries) ‘the final shape of the reform should be seen as inherently endogenous development’. An alternative explanation might refer to the variation in individual countries of the meaning of local governments. This approach would also require additional qualitative research. So far, more questions than answers have been raised, but by exposing the variation of local government systems, which has been thus far underestimated in the academic literature, one hopefully finds an inspiring starting point for critical discussion and deeper empirical analysis.

Notes

1. The term ‘Eastern Europe’ used in this article is often contested in the region. Using geographical arguments as well as referring to common historical traditions, at least part of the region (including Czech Republic, Hungary and Poland) prefers the term ‘Central’ rather than ‘Eastern Europe’. Not disregarding these arguments, I use the term ‘Eastern Europe’ to cover the whole spectrum of countries that was under communist rule before 1990.

2. Noted also by the authors of the classic studies themselves – see for example, Goldsmith and Page (2010).

3. The role of foreign experts in local government reforms in Eastern Europe could be a topic for separate analysis. They were usually financed by multilateral and bilateral donor programmes and they came into the ‘vacuum’ in the East European experience with the management of local issues within a market economy, so ‘internal demand’ for their services coincided with the willingness of external institutions (such as the World Bank, European Union) to influence the process of transformation. However, their actual role in shaping institutional structures and management practices varied among individual countries.

4. For more details on the territorial reforms see Swianiewicz (2010).

5. In some theoretical interpretations, this trend is identified with ‘New Political Culture’ – see Clark and Hoffmann-Martinot (1998).

6. Page and Goldsmith (1987) indicate more formalised central-local relations and more influential institutional lobbying of local governments in Northern Europe. They note that before consolidation reforms the nature of these relations was more similar to those observed in Southern Europe and included more direct and informal contacts. Central control was based more on direct influence of individual cases than on regulation through formal regulations on the standard of locally delivered services.

7. Armenia provides an anecdotal example of the arbitrary system. During research conducted in 2009, I interviewed the mayor of a city who was elected as an independent, non-partisan candidate (which is a rare phenomenon in this country). After two years, during which the city had not received any grants from the central budget, the mayor decided to join the ruling party (see Dąbrowska et al. 2009). Ukraine is an example of a situation ‘in between’ – 10 years ago the New Budget Code introduced the formula of fund allocation, but frequent changes of the formula parameters (and changes often introduced in alignment with the government's current political preferences) as well as arbitrary methods of ‘reserve fund’ (which is of a very significant size) allocation do not allow for a score of ‘4’.

8. Until recently, in some countries (for example, Ukraine, Moldova and Armenia) the term ‘local debt’ was understood by local politicians not as the level of borrowing incurred for development projects in banks (this sort of debt was hard to imagine), but rather it was identified by unpaid invoices for gas and electricity and/or unpaid salaries of local government employees. The issue of development of local government borrowing in Central and Eastern Europe is extensively discussed in Swianiewicz (2004).

9. Obviously this measure does not have a universal character. For a different group of countries or for a different moment in the history of Eastern European local governments it would be quite useless, or at least its significance would be very different.

10. Most probably, the majoritarian system will also be extended to larger cities in Poland from 2014.

11. It should be noted that Latvia has recently started territorial consolidation reform (see Vilka 2010). The data used in this article concern the situation before the reform. The reform, however, has not led to a radical change of the level of fragmentation (yet?).

12. For more on territorial reforms in Eastern European countries see Swianiewicz (2010).

13. In Georgia, the situation has been changing over the last 20 years. In some periods the mayor (gamgebele) was directly elected by the local population. This typology refers to the legal system introduced by the 2006 territorial reforms (see Swianiewicz and Mielczarek 2010). It should also be noted that the recent 2010 local elections introduced the direct election of the mayor in the Georgian capital city of Tbilisi.

14. In 2009, during field research on local democracy in South Caucasus countries, I discovered that employees of the local administration in Azerbaijan could not meet researchers for interviews without a lengthy and troublesome approval procedure from one of the central government ministries. This fact illustrates the lack of local government autonomy. Additionally, the level of centralisation does not seem to be as extreme in Armenia, but it is still higher than in most other countries in the region.

15. Bosnia and Herzegovina are exceptions in this category, and have not been included in the classification. Another formal case of the federal system is Russia, which has not been addressed here.