Journal of Comparative Politics 44 DEMISE OF POLITICAL PARTIES: AN ANALYSIS OF COALITION-BUILDING ON THE LOCAL LEVEL IN SLOVENIA Miro HAČEK, Marjan BREZOVŠEK and Irena BAČLIJA 1 In parliamentarian democracies one can only rarely witness a single party winning a majority in a representative body; therefore, the processes of building a coalition and coalition government are especially important. We depart from the presupposition that, when formulating their priorities, political parties pursue three particular objectives: to gain power, to acquire a policy-making influence and to gain the greatest possible voter support at the next elections. We presume that political parties wish to have as few coalition partners as possible, they especially follow ideological proximity and the possibility of carrying out their political programme when choosing a coalition partner and they often choose a coalition partner in order to distribute the accountability for managing a municipality and to thus reduce the possibility of being punished by voters at the following elections due to mismanagement of the municipality. The article analyses the process of post-electoral coalition-building in Slovenian municipalities from the standpoint of the cohabitation of the executive and legislative branches of government within a municipality from 1998 until 2006. The main research question is whether various forms of ruling coalition can represent an obstacle to efficient government within a municipality. We ascertain that project-based co-operation is gradually becoming the prevailing form of co-operation within Slovenian municipalities, meaning that local councillors are being less and less led by party interests. 1 Miro Haček is an Associate Professor and Head of the Policy Analysis and Public Administration Department at the Faculty of Social Sciences, University of Ljubljana. His research interests are local government studies, civil servant systems, politico-administrative relations and comparative political systems. Contact: miro.hacek@fdv.uni-lj.si. Marjan Brezovšek is an Associate Professor at the Faculty of Social Sciences, University of Ljubljana. His research interests are comparative public administration and administrative culture studies. Contact: marjan.brezovsek@fdv.uni-lj.si. Irena Bačlija is a Teaching Assistant at the Faculty of Social Sciences, University of Ljubljana, a Ph.D. student and a young research fellow at the Centre for Political Science Research. Her research interests are urban management, local government studies and comparative political systems. Contact: irena.baclija@fdv.uni-lj.si. Journal of Comparative Politics 45 Theoretical approaches to the problématique of local- level coalition-building Coalitions involved in the formation of various forms of local authorities are – especially on the European continent – a relatively common and well-known phenomenon. It is therefore even more surprising 2 that the phenomenon of co- alition-building at the local government level has so far not been given much sig- nificant attention in the scientific-professional realm. The theoretical approach- es to the study of coalitions and coalition-building are of course immense and frequently appear in studies of the mentioned phenomena at the national level. Within the framework of our research and this article we seek to establish how to make use of the relatively extensive literature referring to coalition-building at the national level of government for coalition-building at the local level. By the term coalition we understand any combination of separate players (such as political parties) to win a voting game. The most common form of coalition arises where legislation requires a majority to pass but no single part controls at least half of the seats in the assembly. A coalition partner is any actor (political party, independent list, individual) that continuously supports the coalition with all the votes at its disposal in a representative body. Any actor within a repre- sentative body can be either a member of a coalition or opposition and there is no option (at least according to the evidence suggested by coalition theories) in between. 3 However, in the real world such “intermediary” actors do exist and are also well known in the Slovenian political arena. Of course, coalitions can form at any time but they generally nevertheless form prior to or after elections. During the last few decades, theoretical research on coalitions and coalition- building has developed within two relatively differing and mutually independent analytical traditions. The first is the tradition usually referred to by the majority of authors as the “European politics tradition” , which we follow in our observa- tion and explanation of coalition-building at the municipal level of Slovenian local self-government. This approach to the study of coalitions resorts to empirical data analysis when studying the great issues of the politics of (foremost) Eu- ropean states. The style of theoretical approach within this research tradition is mostly inductive in nature. Its purpose is to provide useful and interesting generalisations of coalition-binding and coalition-building on the basis of a sys- tematic analysis of data deriving from the real world or actual coalition-building in different countries. Examples of the application of this approach are found 2 See Michael Laver, “Theories of coalition formation and local government, ” in Political parties and coalitions in European local government, ed. Colin Mellors and Bert Pijnenburg (London and New York: Routledge, 1989), 15. 3 Abram De Swaan, Coalition theories and cabinet formations (Amsterdam, New York and London: Elsevier scientific publishing company, 1973), 143. Journal of Comparative Politics 46 in the works of Axelrod 4 , Dodd 5 , Budge and Herman 6 , Paranjoy 7 and many oth- ers. The works of all these scholars are essentially theoretical as they deal with general explications regarding the formation of specific coalitions and the dis- tribution of power among their constitutive parts. However, at the same time the works of these authors are also relatively empirically oriented in the sense that the preset theoretical premises are “tested” against empirical data from various (European) states. The second tradition in the study of coalitions and coalition-building is the so- called “game theory tradition” which sees the coalition-building process as a special kind of social interaction, forcing its actors (political parties or their rep- resentatives) to negotiate since mastering this process is the necessary condi- tion for “winning the game” . The game theory tradition is based on deduction as it tries to form models of coalition-building on the basis of a priori determined anticipations about the negotiating positions of individual actors. Examples of the use of this research tradition are found in the works of Riker 8 , Grofman 9 , Schofield and Laver 10 , and many others. These works are theoretically con- ceived and largely abstract in nature so they can be used in various situations. 11 However, the key dynamics of these theories do not derive from their testing based on actual empirical data from different European states, but from their own internal logic. Yet both research traditions encounter similar difficulties; one of the most com- mon is the lack of useful empirical data, especially concerning the very pro- cess of forming a certain type of coalition in a strictly specified and specific national environment. At the same time, the variations among different states are so immense that they often require every theory to be tested within a single specific national environment. Consequently, this entails that even very similar or the same kinds of coalitions cannot simply be compared with one another and without any convincing and methodologically based argumentation. Such limitations on the research of coalitions at the national level open some fairly 4 See Robert Axelrod, Conflict of interest; a theory of divergent goals with application to politics (Chicago: Markham, 1970). 5 See Lawrence D. Dodd, Coalitions in parliamentary government (Princeton: Princeton University Press, 1976). 6 See Ian Budge and Valentine Herman, “Coalitional government formation: an empirical relevant theory, ” British Journal of Political Science 8, 8 (1978), 454–477 . 7 Guha T . Paranjoy, A time of coalitions: divided we stand (New Delhi and London: Thousand Oaks, SAGE, 2004). 8 William H. Riker, The theory of political coalitions (New Haven: Yale University Press, 1962). 9 Bernard Grofman, “ A dynamic model of protocoalition formation in ideological N-space, ” Behavioural science 27 , 1 (1982), 77–90. 10 Norman Schofield and Michael Laver, Multiparty government: the politics of coalition in Europe (Oxford: Oxford University Press, 1990). 11 For instance, coalitions of different companies within cartels, coalitions of states within international organisations and, of course, coalitions of political parties in the formation and leadership of governments. Journal of Comparative Politics 47 intriguing possibilities for the study of coalitions and coalition-building at the level of local government. Namely, in the terms of the research coalitions at the local level of government are, compared to those at the national level, still a true terra incognita. The existence of a greater number of coalition communications in a bigger number of different local settings, within one state and among the same political forces, offers a de facto possibility for a researcher to elaborate a sophisticated empirical analysis since the key factors of observation within the system (legislative setting, actors) are constant. Accordingly, some key notions which will enable us to more accurately analyse coalitions and coalition-building at the local government level in Slovenia must be defined. While a definition of government at the national level can probably be perfectly clear, this is not the case of the definition of government at the local level. The local-level executive is often represented foremost by a career bureaucracy (local administration) which in some countries (including Slovenia) is indirectly headed by an either directly or indirectly elected executive body of local self-government (a mayor). The legislative body within local government communities is most often represented by elected local politicians who per- form their function non-professionally. In the majority of states no constitutional or functional equivalent to a government or ministers at the national level can be found, even though some working bodies and their presidents have at their disposal a relatively large influence on specific areas of local policies. Such a lack of a clear local equivalent to national government opens up interesting theoretical problems on at least two levels. At the most general level, we have to foresee as clearly as possible the motivational factors of the various actors during the process of coalition-building; then, we have to foresee in which po- litical space and with what aim or intention “local” coalitions are to be formed. In order to understand coalition behaviour and conduct at the local government level, we must first define the institutional and political rules 12 which determine the boundaries that exert a key influence on the processes of establishing, formation and operating of coalitions. Here the question surfaces of whether subtle and unstable “arrangements” , which seem to be even more frequent than formal, transparent and written coalition agreements, can be understood 12 The Slovenian local electoral system regulates the election of various local self-government bodies – the legislation regulates not only the elections of municipal councils but also the elections of mayors (throughout Slovenia mayors are elected by a two-round voting system with an absolute majority) and the elections of local, village and district community councils. The elections of municipality councils use two completely different voting systems. For municipalities with a lower number of members of municipal councils (twelve or less; there were 60 such municipalities at the last local elections) the majority voting system is used. It is a single-round voting system with relative majority (i.e. first-past-the-post) enabling better chances for individuals and reducing political room for parties, which also facilitates the election of non-partisan candidates. Municipalities with a larger number of municipal council members use the proportional voting system which is about voting for lists of candidates. Here, according to the principles of local self-government a certain level of voter influence on the choice of persons is also provided for and done by a vote of preference. Journal of Comparative Politics 48 as coalition arrangements at all and whether this is the case of coalition or some other form of co-operation. The differences between coalitions at the na- tional and local levels of government concern both the nature of actors and the nature of their aims. 13 At the national level, the actors involved in the coalition- building process are, in the vast majority of cases, political parties; this is less the case with the local level where, apart from political parties, a multitude of other actors is present, ranging from various civil society and pressure groups to more or less independent candidates and candidates nominated by groups of voters. At the local level, one can always find territorial areas where political parties play a less important role and where they might even be absent. Local politics thus often “happens” within a more intimate, less formal atmosphere where local political leaders and their personalities are often more important than internal party rules as far as the coalition-building process is concerned. Certainly, coalition-building at the local level is often heavily influenced by both the institutional framework determined by state institutions (electoral system, the degree of financial autonomy, legislative responsibilities, limitations of com- petencies etc.) as well as by political parties’ centrals (that prefer or exclude the possibility of co-operation with specific political parties and/or other actors at the national level and transfer these preferences to the local level). Analysis of coalition-building in Slovenian municipal councils – the 1998–2002 and 2002–2006 terms The results of local elections in Slovenia reveal the success of independent candidates which can, in the first place, be attributed to the use of a two-round, absolute majority voting system for the elections of mayors. At the same time, local elections are more suitable for the assertion of the passive suffrage on the part of independent, non-partisan candidates. Because of a fairly strong tradition of non-partisanship, voters, especially at the local level, would rather choose to support a candidate they personally know and who is not aligned with any party. 14 The research on post-electoral coalition-building at the local level in Slovenian municipalities during the 1998–2002 and 2002–2006 terms took place from November 2006 to February 2007 . The research included all of the then 193 municipalities established either prior to or in 2002. Sample realisation ranged from 79.3% to 81 .3% depending on the variables observed. The data analysis sought to look for the presence or absence of a causal relationship between the mayors’ party (non-)allegiance and the forms of coalition-building within the 13 See Colin Mellors and Bert Pijnenburg, Political parties and coalitions in European local government (London and New York: Routledge, 1989), 12. 14 See Miro Haček, “The phenomena of independent candidates on local elections, ” in Local democracy: the analysis of local elections 2002, ed. Marjan Brezovšek et al (Ljubljana: Faculty of Social Sciences, 2004), 71–72. Journal of Comparative Politics 49 municipal councils of Slovenian municipalities. A provable relationship between two variables can enable us to shape an applied model for predicting the most probable form of coalition-building depending on a mayor’s party allegiance. Further, we also observed the relationship of both variables 15 with regard to the size of a municipality (as a number of inhabitants). Namely, this last compari- son can demonstrate the main differences among Slovenian municipalities. A combination of various methods was used to collect data for all three variables: (1) the mayor’s party allegiance; (2) the form of coalition-building in a municipal council; and (3) the municipality’s size. The analysis of coalition-building in Slovenian municipalities involves four key research questions: (1) the percentage of a certain form of coalition-building in Slovenian munici- palities; (2) the form of coalition-building according to the mayor’s party allegiance (does the mayor’s party allegiance prejudice a form of coalition co-oper- ation); (3) the form of coalition-building in respect of the municipality’s size (does a certain size of a municipality prejudice the form of coalition-building); and (4) a mayor’s party allegiance regarding the size of the municipality (does a municipality’s size prejudice the party allegiance of the mayor). The data on the mayors’ party allegiance (1) in all Slovenian municipalities during both terms included in this part of analysis (1998–2002 and 2002–2006) are part- ly accessible at Državna volilna komisija (the National Electoral Commission). 16 However, the data were further checked at the municipalities’ headquarters (in some places mayors had resigned, died or ceased to perform their function in some other way; in such cases, we included the party allegiance of the mayor with the longest tenure). Candidates who did not run in elections as candidates of any political party were deemed independent candidates; mayors supported by more than one political party at elections were designated mayors of coali- tions (a C-R coalition – a coalition of centre-right parties; and a C-L coalition – a coalition of centre-left parties; and a mixed coalition, whereby a candidate received support from at least one party from each political pole at elections), 15 In this case, the variable of the “mayor’s party allegiance” was a dependent one as well, even though we treated it as an independent one when we compared it with the variable of “form of coalition-building” . 16 Državna volilna komisija (National Electoral Commission), www.rvk.si (accessed on 10 September 2008). Journal of Comparative Politics 50 while the remaining mayors were included according to the list of the political party they belonged to during elections. 17 To better present the gathered data, we categorised them as follows: centre-left (mayors of the LDS, ZLSD/SD and of centre-left coalitions), centre-right (mayors of the SDS, SKD/NSi, SLS and of centre-right coalitions), independent (mayors, elected from non-partisan or independent lists) and other (SNS, DeSUS, DS). The data on the forms of coalition-building in municipalities during the two studied terms (2) were gathered using the interview method in the municipali- ties involved, whereby we specifically stressed the descriptive classification of the form of coalition-building; the survey respondents were not burdened by predefined types of answers but were left to freely describe the coalition co- operation. Here, we must emphasise the initial finding that in the vast majority of cases of municipal coalition co-operation no coalition agreements or similar written documents were entered into. On the basis of the gathered informa- tion, we could then form a categorised list of different forms of coalition co- operation: - a majority coalition (controlling at least one vote more than half of all possible votes in a representative body) of centre-right parties; - a majority coalition of centre-left parties; - a majority coalition of maximum distance (a coalition of political parties from the left and the right poles); - a majority coalition of the independent (a coalition only comprising councillors from independent lists); - an “unknown” majority coalition (we could not gather data on who constituted the majority coalition); - a minority (it controls less than half of all the seats in a representative body, yet controls the power within a local community) centre-right coalition; - a minority centre-left coalition; - a minority coalition of maximum distance (a coalition of political parties from the right and the left poles); - a minority coalition of the independent (a coalition only composed of councillors from independent lists); - a “unknown” minority coalition (we failed to gather data about the members of a minority coalition); and - no coalition – project-based co-operation. 17 LDS – Liberalna demokracija Slovenije (Liberal Democracy of Slovenia), SDS – Socialno demokratska stranka/Slovenska demokratska stranka (Social Democratic Party/Slovenian Democratic Party), ZLSD/SD – Združena lista socialnih demokratov/Socialni demokrati (United List of Social Democrats/Social Democrats), SLS – Slovenska ljudska stranka (Slovenian People’s Party), SKD – Slovenski krščanski demokrati (Slovenian Christian Democrats), NSi – Nova Slovenija (New Slovenia) and SNS – Slovenska nacionalna stranka (Slovenian National Party). Journal of Comparative Politics 51 For the further analysis we grouped the abovementioned forms into simpler and analytically more suitable forms of coalition co-operation. All the forms of majority coalitions were grouped in a new category of majority coalition and all the forms of minority coalitions were grouped in a new category of minority coalition; and those cases where no coalition was built were grouped in a new category labelled no coalition – project-based co-operation. The size of a municipality was an independent (control) variable (3), expressed as the number of inhabitants. On the basis of these data, we classified the mu- nicipalities in the following groups: (1) up to 2,000 inhabitants; (2) from 2,001 to 5,000 inhabitants; (3) from 5,001 to 10,000 inhabitants; (4) from 10,001 to 20,000 inhabitants; and (5) above 20,000 inhabitants. Forms of Coalition-building in Slovenian Municipalities Let us first look at the data on the form of coalition-building that is most fre- quently present in the municipal councils of Slovenian municipalities. The most common form of co-operation in the municipal councils was project-based co- operation, meaning that councillors voted on each individual proposal (project) separately; they had hence not entered into any (written) agreements to sup- port decisions according to their party allegiance and (as many survey respon- dents said) they were first and foremost led by local and not by party interests. During the 1998–2002 term, arranged co-operation on individual projects (proj- ect-based co-operation) was present in 58 percent of the municipalities and, in the following 2002–2006 term, these figures underwent a decrease in half of all municipalities. This is followed by the formation of a majority coalition; in the 1998–2002 term, majority coalitions were built in 34 percent of municipalities while in the 2002– 2006 term this occurred in 38 percent of cases. If we further analyse the forms of majority coalitions prevalent within this category, we may conclude that in 1998–2002 term, centre-right coalitions were the most common (in 20 munici- palities), whereas the remaining forms of majority coalitions were approximate- ly evenly represented (about 12 municipalities each 18 ). The 2002–2006 term witnessed a significant rise in the number of “coalitions of maximum distance” (from 11 to 18 cases), the number of centre-right majority coalitions fell from 20 to 18, while the number of centre-left majority coalitions was halved (from 12 during the 1998–2002 term to 6 during the 2002–2006 term). The remaining forms are underrepresented. Just as we expected, the minority coalitions were the least represented form of coalition-building. In the 1998–2002 term, these were found in 8 percent of all the municipalities observed 19 and, in the following term, they were found in 18 12 centre-left majority, 12 unknown majority and 11 maximum distance majority coalitions were built. 19 In 12 municipalities (N=146). Journal of Comparative Politics 52 12 percent of all municipalities. 20 In the first of the observed terms, centre-right, centre-left and unknown minority coalitions were equally represented (in three municipalities each) and a majority coalition of maximum distance and a major- ity coalition of the independent were present in a single municipality each. In the 2002–2006 term we found the relatively significant growth of centre-right coalitions (they appeared in nine municipalities), while the remaining categories were at approximately the same level as in the preceding term. It is interesting to see an increase in both minority and majority coalitions as re- gards project-based co-operation in the second researched term (2002–2006). Possible reasons for this change are: (a) a growth in the number of elected inde- pendent mayors who, as a countermeasure, “demand” a more organised mu- nicipal council; (b) the standard deviation of the gathered data that enables the oscillation of answers; (c) the growth of local political elites’ awareness of the importance of local political decisions, consequently entailing a more organised and professional political arena; and (d) the maturity of the local political elite. Forms of Coalition-building in Municipal Councils According to Mayors’ Party Allegiance If the primary data on the forms of coalition-building are upgraded with an analy- sis of the influences on the formation of connections among councillors, then the mayor’s party allegiance figures as one of the most important independent variables. A mayor is closely tied to a municipal council. He has the right of initiative on whose basis they propose that a municipal council adopt a budget, close accounts, make decrees and all other acts within a municipal council’s jurisdiction. He summons sessions of a municipal council and presides over them and, in the office of an executive body, provides for the public implemen- tation of decisions made by the municipal council. A mayor is also a guardian of legality and constitutionality and can, if the act of a municipal council is against the law and/or Constitution, withhold its publication. The role of mayors and municipal councils in decision-making processes is closely interwoven and in- terdependent. For the purpose of allowing a better overview we categorised both variables – the mayor’s party allegiance and the form of coalition co-operation. For the term 1998–2002 we concluded that centre-left mayors were more inclined to the formation of majority coalitions and even more so to the project-based govern- ing of municipalities. Obviously, they either looked for a strong foothold in mu- nicipal councils or, in the case of more salient projects, decided for the widest co-operation possible. There were some exceptions; for example, one of the municipalities with a mayor with allegiance to the Liberal Democrats witnessed the formation of a centre-right coalition and in two other such municipalities ma- jority coalitions of maximum distance were built. Therefore, it seems that the 20 In 19 municipalities (N=157). Journal of Comparative Politics 53 mayor’s party allegiance is certainly not the only factor that (can) affect(s) the form of co-operation within a municipality, yet the prevailing numbers of coali- tions encompassing politically-related parties indicate that there is some causal relation. On the other hand, centre-right mayors more frequently tended (albeit still relatively rarely) to decide on the formation of minority coalitions as far as the share of the latter is concerned. Of all the centre-right parties, the Slovenian People’s Party (SLS) stood out because maximum distance majority coalitions tended to form more frequently in those municipalities where its mayors had won the elections. In 69.5 percent (32 cases) of municipalities where independent mayors won the elections, the project-based governing of municipal councils was opted for. Independent candidates or candidates of independent lists were obviously less prone to the (political) influences of municipal councils where political deals made in advance are possible. However, if municipal councils built coalitions in the presence of independent mayors there was a greater possibility of a centre- left coalition (be it a majority or minority). If we compare the situation during the following term (2002–2006) we can conclude that municipalities with centre-left mayors retained almost the same number of majority coalitions built (12 in the 1998–2002 term and 11 in the 2002–2006 term), while the number of municipalities with project-based co- operation rose somewhat (from 14 municipalities during the first term to 18 during the second). In those municipalities where centre-right mayors were elected, an elevated percentage of majority coalitions can be observed dur- ing the second term while the share of project-based co-operation decreased within these municipalities. If we look more closely at the prevalent forms of majority coalitions in municipalities with centre-right mayors, we can argue that these mayors usually built coalitions among centre-right parties. The probability of a centre-right coalition being formed in a municipality having a centre-right mayor was 82 percent, which is significant enough for us to corroborate a caus- al relationship between the mayors’ party allegiance and the party composition of coalitions within municipal councils. A possible explanation of this high figure is the development of the local organisation of a certain party which, in most cases, had composed both a list of candidates for members of the municipal council and nominated a candidate for the mayor. The presence of an individual party thus to some extent assumes (especially in smaller municipalities with fewer candidate lists) a certain share of elected municipal councillors who then, under the auspices of a mayor from an ideologically related political option, at- tempted to form a majority coalition in a municipal council to ensure municipal operations were as uninterrupted as possible. The second analysed term (2002–2006) witnessed a change in the proportion of the mentioned forms of coalitions with independent mayors elected to the function. The share of majority coalitions rose by almost one-tenth, the share Journal of Comparative Politics 54 of minority coalitions doubled, while the share of project-based co-operation dropped by more than 15 percent. In those municipalities with independent mayors, maximum distance majority coalitions were the most frequent. The marked increase in the share of coalition-building in almost all categories (the slight decrease among centre-left mayors being an exception) indicates that lo- cal political elites were undergoing a politicisation and organisation, that there were fewer and fewer projects left for coincidental co-operation and momen- tary support for the governing option, or even that the mayors’ experience told us that, to ensure uninterrupted work, it was better to form a consensual major- ity already at the beginning. However, the share of project-based co-operation, i.e. the non-formation of a coalition, still remained the biggest so this was the most common form of local-level decision-making, although the first signs of an opposite trend have been spotted. We can conclude that during the 2002–2006 term the connection between the centre-left and independent groupings was no longer so close. Those munici- palities where independent mayors were elected showed a three times higher number of maximum distance majority coalitions (for the 1998–2002 term the respective number of such coalitions in the presence of independent mayors was two and for the 2002–2006 term it was six). The trend of coalition-based co-operation was obviously slightly tilting towards the politicisation of local po- litical elites and towards a more general awareness that local policy-making can be undertaken provided that a consensual majority is built in advance. Forms of Coalition-building According to the Size of Municipalities The second independent variable used to analyse the specificities of coalition- building in Slovenia is the size of a municipality (in terms of the number of inhab- itants within a municipality). Namely, a municipality’s size directly influences the number of members of a municipal council which (could) consequently mean(s) that the ways of co-operation among councillors depend on their numbers. As regards the smallest municipalities (up to 2,000 inhabitants), the 1998–2002 term showed that project-based co-operation used to prevail. It seems that the small number of councillors somehow reduced the possibility of whatever form of coalition being built. The share of municipalities with somewhat more inhabitants (from 2,001 to 5,000 and from 5,001 to 10,000) whose councils did not form any coalitions (project-based co-operation) was the same as for their smaller counterparts. This supports the supposition that, in small municipalities with up to 10,000 inhabitants, there was not much space for political alliances or, to put it differently, that local interests prevailed over political ones. However, the same cannot be said of the two largest categories of municipali- ties (above 10,001 inhabitants) as project-based co-operation was revealed in “just” 46 percent of the municipalities with 10,001 to 20,000 inhabitants and in the largest municipalities (over 20,000 inhabitants), this share was further re- duced to 23 percent. At the expense of this form of municipal co-operation, the Journal of Comparative Politics 55 percentage of majority coalitions increased especially in the largest municipali- ties where 70 percent of all forms of party co-operation were due to this type. During the 2002–2006 term, the ratio among the forms of councillors’ co-op- eration remained almost the same in the smallest municipalities (up to 2,000 inhabitants). Growth in the formation of minority coalitions during this term can be detected in all categories of municipalities under scrutiny, possibly once again indicating the politicisation of the local decision-making arena. For mid- sized municipalities (from 2,001 to 5,000 inhabitants) the percentage of no co- alition within municipal councils, i.e. with co-operation on individual projects, was the same as for the small ones. Once again, the mid-sized municipalities (from 5,001 to 10,000 inhabitants) showed no change in trend (except for an elevated share of minority coalitions) and that the share of project-based co-op- eration was the same as in the previous two categories of municipalities. Yet if a comparison with the preceding term is made one can see that the percentage of majority coalitions was larger than the so-called project-based co-operation, indicating that the growing size of a municipality increased the chances of ma- jority coalitions being built. This comparative analysis of the two studied terms allows us to conclude that: (1) the proportion of minority coalitions was slowly rising regardless of the municipalities’ size; and (2) that the share of project-based co-operation was decreasing and the probability of majority coalitions was increasing with the growing size of municipalities. In the 2002–2006 term this characteristic was already expressed in large municipalities (10,001 to 20,000 inhabitants) and not, as in the preceding term, exclusively in the largest ones (more than 20,000 inhabitants). Mayors’ Party Allegiance According to the Size of Municipalities Apart from other correlations, we also observed the influence of the indepen- dent variable of the size of a municipality on the party allegiance of an elected mayor; one of the most often exposed correlations in public, as it has often been argued that a smaller municipality almost automatically entails a centre- right mayor and vice-versa. In the 1998–2002 term the prevailing mayor allegiance in the smallest munici- palities (up to 2,000 inhabitants) was to the centre-right (7), followed by the independent (6); the percentage of centre-left ones was significantly lower. A similar share was reflected in the moderately small and moderately large cat- egories (2,001 to 5,000 and 5,001 to 10,000). The centre-right mayors unques- tionably held the largest portion. The trend turned with the large (from 10,001 to 20,000 inhabitants) and the largest (over 20,000 inhabitants) municipalities. The former witnessed a complete equalisation of the percentage of centre-left and centre-right mayors and the latter had almost twice as many centre-left mayors Journal of Comparative Politics 56 than centre-right ones. We can conclude that, during this term, centre-right mayors were prevalent in smaller municipalities (these are also mostly rural). Compared to the 1998–2002 term, the 2002–2006 term had a significant in- crease in the percentage of independent mayors; as a rule, this was usually at the centre-right mayors’ expense. Especially successful were the independent mayors in the smallest municipalities (up to 2,000 inhabitants) where they ac- counted for 65% of all mayors. The share of centre-left mayors from the small- est municipalities remained the same for both terms. A similar percentage can be found with the categories of moderately small and moderately large munici- palities. Centre-right mayors overwhelmingly achieved the lion’s share and the difference between them on one side and the independent candidates on the other increased at the latter’s expense. In moderately small municipalities (from 2,001 to 5,000 inhabitants), the percentage of independent mayors rose as well (from 37% to 43%) with the proportion represented by other mayors remaining more or less intact. If the abovementioned data are taken into consideration, one can establish that, despite the growth in the trend of coalition-building within municipal councils indicating the politicisation and higher level of organisation of the local decision- making arena, the depoliticisation of the polity was taking place at the same time because, during the 1998-2006 period, voters increasingly opted for in- dependent candidates. If both trends are merged, the mayors’ de facto inde- pendence comes into question since as soon as they took up office they were eager to make some form of political connections. Conclusions As we predicted, the concluding findings may somewhat deviate from the statis- tically processed data especially because of individual municipal councils where the balance of power could change even during the current term for objective or subjective reasons, making the definition of the form of co-operation even more difficult. Since the analysis of coalition-building in Slovenian municipalities refers to the four main points: (1) to establish the percentage of individual forms of coalition-building in Slovenia; (2) to find out the form of coalition-building we can expect according to a mayor’s party allegiance in a municipality or whether a mayor’s party allegiance can prejudice a form of coalition co-operation; (3) to establish which form of coalition-building corresponds to the size of a munici- pality or whether municipality’s size somehow prejudices the form of coalition- building; and (4) what a mayor’s party allegiance is according to a municipality’s size or whether a municipality’s size somehow prejudices the mayor’s party adherence; the following can be stated: Journal of Comparative Politics 57 (5) During the two observed terms, project-based co-operation prevailed within Slovenian municipalities which could mean that councillors were led by local and not party interests. If we look at this phenomenon in time, we can detect growth in both the presence of minority and major- ity project-based coalitions at the expense of the absence of formal party coalition co-operation. The possible reasons behind this change were: (a) an increase in the number of elected independent mayors, hence a more organised municipal council was “needed” to act as a counterbalance; (b) the standard deviation of the collected data which enabled fluctuations of answers; (c) growth in local political elites’ awareness of local political de- cisions which led to a more organised and more professional polity; and (d) the maturity of the local political elite. (6) A mayor’s party allegiance and the form of coalition co-operation within a municipal council were not entirely causally dependent; however, some rules are indicated. We can say that during both terms centre-left mayors were more inclined to form majority coalitions and, to an even greater extent, to the project-based governing of municipalities. It is evident that they either looked to the municipal councils for a strong background or, on the other hand, decided for unified co-operation on bigger projects. Of all the centre-right parties, the Slovenian People’s Party was the most prominent one as the municipalities in which its mayors were elected more often witnessed the formation of maximum distance majority coali- tions. Municipalities with an independent mayor elected witnessed their municipal councils adopt project-based governing in 69.5% of cases. The marked increase in the share of coalition co-operation in almost all cat- egories for the 2002–2006 term, indicates that local political elites were undergoing politicisation and organisation and that an ever decreasing number of projects was left to coincidental co-operation and ad hoc sup- port. Nevertheless, the percentage of project-based co-operation, e.g. of the formation of no coalition, still reveals that this was the most frequent form of local-level decision-making. However, the first signs of a differing trend have been detected. (7) If we look at the influence of a municipality’s size on coalition-building during the past two terms we may conclude that in the smallest munici- palities (up to 2,000 inhabitants) project-based co-operation prevailed in municipal councils. Yet the larger the size of a municipality, the greater the probability there was of a majority coalition being formed. Since we are comparing the two terms, we can say that: (a) the share of minority coalitions was on the increase regardless of the size of municipalities; and (b) the proportion of project-based co-operation decreased as the size of municipalities increased and the probability of majority coalitions being built rose. Journal of Comparative Politics 58 (8) Similarly, a municipality’s size was related to a mayor’s party allegiance during both of the terms as we found that small or medium-sized munici- palities had a greater probability of having a centre-right mayor, while the large or the largest ones were more likely to have centre-left mayors. At the same time, the number of independent candidates was on the rise, which is inconsistent with the growth in the share represented by major- ity or minority coalition-building within Slovenian municipalities. References (2003). Research project “Standpoints on local democracy”. Ljubljana: Centre for Political Science Research, Faculty of Social Sciences. Axelrod, Robert. Conflict of interest; a theory of divergent goals with application to politics. Chicago: Markham, 1970. Brezovšek, Marjan. “Local democracy research and Slovenian’s perceptions. ” In Local democracy: the analysis of local elections 2002, ed. Marjan Brezovšek, Miro Haček and Alenka Krašovec. Ljubljana: Faculty of Social Sciences, 2004. Budge, Ian and Valentine Herman. “Coalitional government formation: an empirical relevant theory. ” British Journal of Political Science, 8, 8 (1978): 454–477 . De Swaan, Abram. Coalition theories and cabinet formations. Amsterdam, New York and London: Elsevier scientific publishing company, 1973. Dodd, D. Lawrence. Coalitions in parliamentary government. Princeton: Princeton University Press, 1976. Grofman, Bernard. “ A dynamic model of protocoalition formation in ideological N-space. ” Behavioural science, 27 , 1 (1982): 77–90. Haček, Miro. “The phenomena of independent candidates on local elections. ” In Local democracy: the analysis of local elections 2002, ed. Marjan Brezovšek, Miro Haček and Alenka Krašovec. Ljubljana: Faculty of Social Sciences, 2004. Laver, Michael. “Theories of coalition formation and local government. ” In Political parties and coalitions in European local government, ed. Colin Mellors and Bert Pijnenburg. London and New York: Routledge, 1989. Mellors, Colin and Bert Pijnenburg. Political parties and coalitions in European local government. London and New York: Routledge, 1989. Paranjoy, T . Guha. A time of coalitions: divided we stand. New Delhi and London: Thousand Oaks, SAGE, 2004. Riker, H. William. The theory of political coalitions. New Haven: Yale University Press, 1962. Schofield, Norman and Michael Laver. Multiparty government: the politics of coalition in Europe. Oxford: Oxford University Press, 1990. Von Neumann, John and Oskar Morgenstern. Theory of games and economic behaviour. New York: Wiley, 1967 .