FIRST DRAFT PLEASE DO NOT QUOTE

Marion Reiser (University of Halle-Wittenberg) The local party system in Germany FIRST DRAFT PLEASE DO NOT QUOTE Dr. Marion Reiser SFB 580, Project A6 Non-partisan voter associations Institute of Political Science University of Halle-Wittenberg Emil-Abderhalden-Str.7 06108 Halle (Saale) Germany Phone ++ 49 345 55-24217 Fax. ++ 49 345 55-27145 Email: marion.reiser@politik.uni-halle.de Paper prepared for ECPR Joint Sessions, Helsinki, 7 12 May 2007 Workshop No.6. Comparative Perspectives on Local Party Politics

1 Introduction During the last decades the ability of the political parties to fulfil their function has been questioned. All over Europe today, the crisis of party politics seems to be evident (Zehetmair 2004; Katz/Mair 1995; Poguntke 2002; von Alemann 2000:146). The signs for political parties becoming more and more unpopular are widespread. Some clear indicators for this development are first of all, a growing peevishness with partitocracy ; furthermore, party identification is continuously weakening and party membership is shrinking; we have to face lower turnouts at national and regional elections on the one side and a rise of floating vote on the other side; finally, trust in party leaders` political competence is steadily moving back. Party elites are not more regarded as honest agents of societal interests but as an oligarchially clustered part of the state, out of touch with the ordinary man s daily world. Corresponding to the damaged common confidence in established parties, we can observe a sharper fragmentation, and sometimes an entire dealignment in European party systems. Despite of the discussion of a crisis of party politics, the importance of political parties isn t questioned fundamentally on the national and state level in Germany. However, on the local level political parties do not play the same role in Germany as they do in state or federal politics. Whereas here their position as mediators between citizens and the political system is uncontested, they have to compete on the local level with the local voter associations and in some states with independent candidates as well. Furthermore, the relevance of political parties on the local level has been discussed controversially in the scientific community as well as in practice. Local self-administration in Germany has been considered as pure factual and harmonic and as an apolitical zone. Hence, political parties have been seen as interfering. Although, due to a process of politicisation in the 1970s, political parties became important local actors also in local politics, the notion of a pure factual local selfgovernment is still very popular and the voter associations perceive themselves as protectionists of this ideology. This ideology, that the main focus of the local level is to manage the urgent factual issues to the well-being of the citizens, revived after the Wende also in the new federal states. Corresponding to the developments on the higher levels of the political system, we witness a decline of party politics also on the local level. During the last two decades, the election results in local elections show decreasing results for political parties and at the same time there is to see a stable and even growing position of non-partisan voters associations on the local level. But despite of these general trends, there is little empirical evidence on the local political party system in Germany. The few empirical studies, which exist about political parties on the local level, indicate however functional deficits of the political parties (Naßmacher 2006: 59; Mayntz 1959; Gabriel 2000). Also despite the recent revival of non-partisan voter associations, research done in this field is also scarce. Next to a few monographs and articles 2

dating back to the 1970s and 1980s (Luckmann 1970, Haller 1979, Möller 1981), newer research done after 1990 and the German reunification exists mainly in form of case studies of single states, cities or groups (Naßmacher 1996, 1997, Holtmann 1998: 220 et sqq., Holtmann 2002, Holtkamp/Eimer 2005; Reiser 2006; Holtmann/Reiser 2006). Comparative studies with a more general approach on the local party system are almost non-existent. Therefore, this contribution wants to focus on a rather broad and comparative (and rather descriptive) analysis of the local party system in Germany: How does the political party system look like on the local level in Germany? In which respects does it differ from the national and/or state party system? Do the national organized parties fulfil their functions on the local level? In this paper, it will be focussed on one central function of political parties: their participation in local elections. Hence, the importance of a political organization can be assessed by looking at its participation and success or failure in elections. Hence, two different aspects will be analysed: The presence and support for different political actors on the local level: Which political actors are present on the local level? Where and under which conditions do they run for election? And where do the citizens elect them? Which are the influencing factors on the local party system and on the local actors? In which respects does the local party system in East Germany differ from the party system found in West Germany? Do we find specific regional parties? In order to answer these questions, the local party system will be analysed. Thereby, in a first step, it will be analysed whether or not the political parties are present on the local level in order to be able to fulfil their functions. In a second step it will be analysed how they do in local elections. Do we find similar to the developments on the national and state level disenchantment from political parties? Is this reflected in higher results of voter associations? And from which endogenous and exogenous factors does depend the support and success of the three different groups of political actors on the local level: the political parties, voter associations and independent candidates. Information about the candidacy of these groups for elections and their election results are the primary sources for analysis. The second aspect, which will be analysed, refers to the specific understanding of local politics in Germany. What kind of understanding of local self-government do the German councillors have? What is their attitude towards political parties on the local level? Are there differences between non-partisans and partisans? And are there possible explanations for specifics of the local party system on the local level? Generally, it is stated in research on local politics, that the importance of party politicisation is increasing with size of community, type of municipal law and regional political culture (see Holtmann 1998; Holtkamp 2003; Schneider 1999; Wehling 1991; Hermann/Werle 1983). Partly in line with that, the main assumption of this contribution is that there isn t a uniform local party system. I argue, that despite of regional differences two different political 3

party systems can be observed on the local system in Germany, depending on size of community: In small communities, we find a low presence of political parties and a dominance of voter associations and independent candidates and councillors with an attitude towards a rather concordant and harmonic self-administration. In bigger cities we find and on the other side a politicised and established party system with competition. 2 Conceptual framework and methods 2.1 Local politics and party politics On the local level in Germany the relevance of political parties in the local decision-making process has always been a controversial question. The relationship between local politics and party politics is a topic of ongoing controversy in the scientific community, based on two explanatory models. Until the 1970s, local self-administration was considered to be outside of the political sphere and was seen as pure factual and harmonic. In this view, the political sphere existed only on the state and federal levels, mainly characterized by competition the between political parties. The local level however was characterized by concordance and harmony between the actors and their orientation towards the one best solution (Luckmann 1970; Ueltzhöffer 1975). Today, this notion of local politics as an apolitical zone, in which political parties are just foreign bodies, is still represented by the older, judicial influenced scientific community of local politics (vgl. Wollmann 1998:50ff; Holtmann 1998:208 ff; Knemeyer 1982:204; Gabriel/Ahlstich/Kunz: 1997:342f). Also, in practice, the idea that good local politics is not compatible with party politics is still remarkable popular. The non-partisan voter associations stand within this tradition. We define non-partisan voter associations 1 as those groups running for elections, which are no political parties in terms of the German party law. These groups see themselves as protectionists of a harmonist factual political style and understand themselves as non-parties or anti-parties. In the last few decades, this more administrative understanding of the local sphere, based on its legal regulations, has been widely challenged by a more political understanding. (see Knemeyer 1982:204; Gabriel/Ahlstich/Kunz: 1997:342f). Therefore, we find on the other side, a newer group in the discourse, which consists mainly of political scientists. They do not only refer to the status of the municipalities under constitutional law like the other group, but also argue rather along the local realities (Wollmann 1999; Schmidt-Eichstaedt 1989; Holtmann 1998; Gabriel/Holtmann 1996). It has been pointed out, that the realities of the local decision-making process show that this is a political process as well, dominated by different interests and factions. Due to a process of politicisation, noticeable at least since the 1970s, parties have become an important actor in local politics as well. Already in 1960, the Federal Constitutional Court (Bundesverfassungsgericht) assessed that political parties had 1 The term local lists is used as a synonym of voter associations within this contribution. 4

become an important subject of local politics. Therefore, today (cf. Holtkamp 2003), competitive attitudes seem to be characteristic for local political parties as well as parliamentary modes of interaction are dominating the procedures in local councils (Gabriel/ Holtmann 1996), especially in bigger cities. Generally, a mixture of a harmonic factual style (concordance) and a party-competitive style is seen as typical for German local self-government. This development is also evident in Eastern Germany (Pollach et al 2000). Despite of this overall development on the local level, according to the paradigm of Wehling (1991; see also Holtkamp 2003), party politicisation on the local level in Germany depends on three main factors: size of community according to Wehling, the party politicisations increases with the size of community (1), local political culture (2) and election system and municipal law (3). Therefore, it is assumed that competitive attitudes and parliamentary modes can be found in the first line in bigger cities and communities. In contrast to that, local self-administration in small communities seems to have a more harmonic factual style: Especially in these small communities the local branches of political parties seem to look comparing Janus, the twofaced god of doors and beginnings, at two opposite directions in the local context, they act factual and deny their partisanship, while they have to communicate politically as regards elections on the federal and state level (see Lehmbruch 1979: 329ff.; Schneider 1999:134). After this process of party politicisation in the 1970s and 1980s that led to a stronger involvement of established national parties on the local level, we now, as mentioned above, observe a contradictory process: the decline of party politics, combined with the increasing disenchantment and alienation with the established national parties, also on the local level. In contrast to the everywhere declining prestige and power of national and regional party politics, there is to see a stable and even growing position of non-partisan voters associations on the local level. It could be argued that these non-partisan voters associations make profit of a general trend of detachment of party politics, embedded in a similar mood of diffuse antiparty-sentiment. 2.2. Institutional framework on the local level The general framework of local politics in Germany (see for an overview Wollmann 1998; Naßmacher/Naßmacher 1999; Gabriel/Ahlstich/Kunz 1997) is determined by two principles: federalism and municipal self-administration. Municipalities do not constitute third level in the federal system next to the Länder and the federal government. Instead, their affairs are regulated and supervised by the Länder, each for its own territory. That includes such different matters like regional development and local structure, municipal duties or institutional arrangements and even the basic rules of local politics. However, the German constitution guarantees municipal self administration in article 28 II the right of municipalities to regulate and organize their own affairs including to a certain de- 5

gree even financial matters thereby acknowledging the local community as the cornerstone of a successful society. To regulate their own affairs, as it is written in the constitution, points towards a much broader understanding of this right beyond mere administrative issues, because it requires a determination of what these affairs are at first. The right to local self-administration therefore covers the local political process as a process of competing interests, general agenda setting and the resulting course of self-administration as well. This constitutional right is of course not limitless but restricted by the constitution and federal and state law. In the end, we look at a framework of local politics in Germany that is determined by some commonalities and many differences among the states, including such defining factors like municipal law, administration structure, electoral system and local political culture. Municipal law and electoral system Traditionally, four types of municipal law have been identified, that laid out the ground for institutions, administration structure and electoral system. Within the last fifteen years reforms took place, which afflicted especially the electoral systems and thereby strengthened civic influence (see Kost/Wehling 2003). Mayoral elections for instance are today in almost every state direct elections. Regarding city council elections, the electoral law throughout the states differs in some aspects. In general, proportional elections are being held everywhere. Majority elections, however, take place when no or just one list of candidates exists. These lists are then regularly not composed by a political group but of independent candidates. This is the case in the first line in the smallest communities. Candidates are usually proposed on lists, assembled by political parties and voter associations. Independent candidates are only allowed to run in five states separately. Within the last two decades most states have adopted the practice of open lists. Hereby, the voter has the opportunity not only to vote for the list as a whole, but also to combine or distribute his votes on one or more candidates even among different lists, hence to cumulate and cross-vote. That requires of course for the voter to have more than one vote. In about half of the states with the open-list-rule voters have three votes. In the other states Baden-Wurttemberg, Bavaria, Hesse, Rhineland- Palatinate citizens have as much votes as there are seats in the city council this can be as high as 80 votes in a city like Munich. Only in North-Rhine-Westphalia, Saarland and Schleswig-Holstein still have closed lists, meaning that a list as a whole can only be voted. It is self-evident that the practice of open lists greatly enhances the options of the voter and favours candidates instead of organisations (see for an overview on electoral systems Kost/Wehling 2003). In some northern states voter associations were factual forbidden during the 1950s (Andersen/Bovermann 2002:70f.; Eith 1997:398f.), since the local electoral law limited the right to propose candidates to political parties. The federal constitutional court has outlawed such 6

practice in 1960. Today, non-party lists are allowed to participate in elections in every state as long as they can provide the required signature of supporters for fulfilling other requirements like a written statute. These provisions differ from state to state. In federal and state elections the restrictive clause of 5% is a well-established rule. Parties have to poll 5 percent of the votes in order to win seats. However, today, most states do not require such a restriction for their local elections (anymore), only Thuringia and the Saarland still have this threshold. In some cases, these restriction clauses have been outlawed by the state constitutional courts; others have dropped or changed these rules in order to raise participation. In the past, there have been more states with a restrictive clause, which impeded especially small parties and voter associations. In Hesse, there has even been a 15% hurdle until 1948. How these restrictions affect especially smaller parties or voter associations is hard to tell. It is reasonable to argue that such rules constrain such groups especially. However, among the few states, which still have a restrictive clause in their election laws, is Thuringia, which ranks among the top states regarding electoral performance of local voter associations. Structural characteristics The structural characteristics of the local level in Germany are not even less diverse. There are about 13.000 municipalities and cities in Germany, of which 114 are politically and administratively independent urban districts excluding the city states Berlin, Hamburg and Bremen. The majority of municipalities are organized in rural districts and municipal associations, staying politically independent but sharing administrative duties. Table 1: Size of communities in year of election (inhabitants) Size of community Number Percent < 500 2196 19,5 500 < 1.000 1980 17,6 1.000 < 2.000 2008 17,8 2.000 < 5.000 2359 21,0 5.000 < 10.000 1219 10,8 10.000 < 20.000 806 7,2 20.000 < 50.000 470 4,2 50.000 < 100.000 104,9 100.000 < 200.000 44,4 200.000 < 500.000 23,2 >= 500.000 9,1 Total 11218 99,7 Table 1 shows that the size of the communities, measured by the number of inhabitants is extremely right-skewed: More than half of all communities in Germany have less than 2,000 inhabitants, and almost three quarters have less than 5,000 inhabitants. On the other side, 7

only 2.7 percent of all communities have 50,000 and more inhabitants. According to that, about two thirds of the German population live in such rural districts and one third in those 114 urban districts. This shows the heterogeneous size structure of the local level, but furthermore it differs distinctly between the different states (see table 2). During the 1970s local administration reforms took place in most of the West German Länder. As a result there has been a more or less radical amalgamation of small communities in the different states. The most radical ones took place in Hesse and North Rhine-Westphalia, leading to the fact that in North Rhine-Westphalia today live about 18 million people in not even 400 cities. On the other hand we have a state like Rhineland-Palatine, where about 4 million people live in roughly 2.300 municipalities and where 85% of the communities have less than 500 inhabitants. In East Germany, on the other side, in most states there hasn t been an amalgamation process so far. There, most states are characterized by peripheral and fragmented structures. According to the general assumption of this contribution that the local party system is influenced by the size of community ( size matters ), these different structural characteristics of the different states in Germany have to be kept in mind. Table 2: Size of communities differentiated by state Region West Germany East Germany State < 2,000 Number of inhabitants 2,000-9,999 10,000-49,999 >50.000 Number of Communities Year of last elections Lower Saxony 48.5% 31.6% 18.1% 1.9% 1,023 2001 North-Rhine-Westphalia.0% 13.4% 67.2% 19.4% 396 2004 Hesse 2.6% 57.0% 37.6% 2.8% 426 2006 Rhineland-Palatinate 85.1% 12.9% 1.6%.4% 2,306 2004 Baden-Wurttemberg 17.1% 60.3% 20.5% 2.1% 1,110 2004 Bavaria 35.5% 53.6% 10.0%.9% 2,058 2002 Brandenburg 54.5% 31.2% 13.3% 1.0% 421 2003 Mecklenburg-Vorpommern 84.5% 12.6% 2.3%.6% 873 2004 Saxony 22.4% 64.3% 12.2% 1.2% 518 2004 Saxony-Anhalt 85.3% 11.5% 2.9%.3% 1,121 2004 Thuringia 77.3% 19.6% 2.7%.4% 998 2004 Total 55,1% 31.9% 11.4% 1.6% 11,218 2.3 Methods Empirically, this analysis is based on the one side on a data analysis of the results of the last local elections in all communities in 11 states 2. The data is from the Statistical offices of the several states. The statistics of the different states have totally different standards and a different empirical basis: Sometimes the results are only relating to the seats, sometimes to all votes. In some Länder non-partisan voter associations are not even expelled separately, but 2 The three city states Hamburg, Bremen and Berlin have been excluded since they have a different status. The two states Saarland and Schleswig-Holstein had to be excluded because the data is not available in the needed form. 8

are included into miscellaneous together with minor parties and independent cities. Therefore, so far, it was not possible to compare the results over states. Due to these restrictions for a comparative research, the results of local elections are so far only analysed on an aggregated level for the different states (see Kost/Wehling 2003; Möller 1981; Holtkamp/Eimer 2005). In order to be able to compare these data, we collected, edited and analysed the data for each community separately within the project Local voter associations in East and West Germany at the University of Halle-Wittenberg. On the other side, the analysis is based on individual data, gathered in two surveys. The interviews were conducted by telephone (CATI computer assisted telephone interviews) in 548 city councils, representing a layered random sample of all German municipalities. The three layers were based on population, each covering a specific size range: less than 5,000 inhabitants, 5,000 to 20,000 inhabitants and bigger ones. Therefore, all 4,085 councillors of local lists in these 548 councils have been interviewed (return rate: 69%). On basis of this survey, the motives and reasons of non-partisans will be analysed. In the second survey, all leaders of the factions of all political parties and local lists in the same selected councils have been interviewed (N = 1.955; return rate: 80%). This survey focused on a comparative analysis of the specific attitudes of local party politicians and representatives of local lists towards local politics. 3. Local party system in Germany What is the nature and structure of the local party system in Germany look like? Which political parties and local lists (and independent candidates) are organized on the local level and run for election? Are the national parties a represented on the local level and do parties fulfil their functions? In which respects does it differ from the national and state party systems? On the basis of the aggregated data, it is generally assumed that political parties do play an important role regarding the recruiting of political personnel, despite of the existence of local lists. From the empirical perspective, one of the most interesting questions regarding political parties and local lists is their performance in elections. In order to analyse the local political system, two phases have to be distinguished analytically: (1) The presence of the (collective) actors on the local level: Where are political parties, local lists and independent candidates present and run for elections? (2) Support/Success of local lists: How successful are the political parties and the local lists in local elections? 3.1 Presence of political parties and local lists on the local level On the local level in Germany, three types of local actors can be distinguished: non-partisan local lists, political parties and independent candidates. The decision of these (collective) 9

actors to run for election produces in each community an own field of candidates with a different degree of fragmentation. This field of candidates will be analysed regarding the following questions: How can this local party system be described? Who runs for elections? What is the proportion between the types of local actors, which run on the local level, and how is the situation of competition (see figure 1 for indicators)? What is the share of local lists in the field of candidates and what is the share of political parties? Where are political parties dominant and where local lists in the field of candidates and how can their presence be explained? Figure 1: Indicators for fragmentation and proportion n _ va Share of voter associations q _ va = n _ cand n_va Number of voter associations n_cand Number of candidate lists n _ part Share of political parties; q _ part = n _ cand n_part Number of political parties n_cand Number of candidate lists q _ kwg n _ kwg q_comp ratio of competition q _ comp = = q _ part n _ part Source: Reiser/Rademacher 2007 Non-partisan voter associations are defined as those groups running for elections, which are no political parties in terms of the German party law. Therefore, an analytical differentiation can be made between political parties and non-parties. As explained above (see 2.1), according to the paradigm of Wehling (1991), party politicisation and the presence of political parties on the local level in Germany depends on three main factors: size of community, local political culture and the election system and municipal law. Referring to the antagonism of political parties and local lists, it could be assumed that local lists are more present in those communities in which political parties are not or less present. This means that those factors, which are favourable for the presence of political parties, could be negative for the presence of non-partisan lists and vice versa. Averagely, 3.2 groups run for local elections in each German community, which means that the degree of fragmentation is quite low on the local level. However, there are differences between the different states: Most groups run for elections in Brandenburg (6.5), North-Rhine Westphalia (5.1) and Hesse (4.4), while in Rhineland-Palatinate only 1.9 lists run averagely for elections. All other states group around an average of 3 lists running for elections. Differentiated by type of those groups, averagely, 1.7 lists of political parties, 1.1 local lists and 0.33 independent candidates run for elections in each community. Political Parties: The average number of political parties running for election differs to a high extent between the different states from averagely 0.76 per community in Rhineland- Palatinate, 1.22 in Thuringia and 1.40 in Bavaria to 3.13 in Hesse and 3.93 in North-Rhine 10

Westphalia. Thereby, the number of political parties running per community ranges from 0 to a maximum of 11. In 30.7% and hence in almost one third of the German communities not even one single political party and in further 17.2% only one political party runs for local elections. In one quarter of the communities (24.6%), two political parties compete against each other. In less than 3% of the communities, 5 and more parties run for elections. But which political parties run for election on the local level? Are there different political parties running at the last local elections than on the national and state level or are the national political parties present area-wide and the dominant ones? Which differences do exist between East and West Germany? Table 3: Participation of political parties in last local elections by state (% of communities) Christian Democrats Social Democrats PDS Greens Liberals (CDU/CSU) Bavaria 63.1% 53.8% 0.2% 13.0% 5.4% Lower Saxony 84.6% 88.4% 2.8% 41.4% 38.7% North-Rhine- 100.0% 100.0% 6.1% 82.3% 88.1% Westphalia Hesse 96.9% 98.6% 4.7% 47.7% 54.9% Rhineland- 31,6% 32.2% 0.0% 4.5% 7.2% Palatinate Baden- 70.8% 58.5% 0.0% 17.2% 14.9% Wurttemberg Brandenburg 66.3% 60.6% 55.3% 16.2% 36.6% Mecklenburg- 68.0% 35.2% 39.5% 3.2% 13.7% Vorpommern Saxony 91.1% 53.9% 66.2% 4.8% 31.3% Saxony-Anhalt 60.3% 33.3% 28.9% 5.4% 20.6% Thuringia 54.6% 26.8% 24.3% 2.1% 13.2% Total 62.7% 50.7% 14.0 15.3% 19.7% As the figures in Table 3 show, the Christian Democrats are organized best: They ran in 63% of the communities in the last elections. But this means at the same time, that they didn t run in more than a third (37%) of all communities in Germany. The Social Democrats ran in half of the German communities. The other three national parties Liberals, Greens and PDS ran in less than 20% of the communities. Thereby, huge differences exist between the states: There are states such as North-Rhine-Westphalia and Hesse, where at the least the two biggest national political parties Social Democrats and Christian Democrats run nearly in every single community, and also the presence in local elections of Greens and Liberals is rather high. In contrast to that, the organizational degree of political parties in some states is very low and this is reflected in a low participation of political parties in the local elections: In Rhineland-Palatinate, the two biggest parties run in less than one third of the communities, but the degree of participation is also rather low in Thuringia, Bavaria and Saxony-Anhalt. Regarding the East-West comparison, two main differences regarding the political parties are obvious: On the one side, the PDS is not only on the local level still a regional party, 11

which is present in East German communities to a rather high extent, while the party actually doesn t even run for elections in the West German communities. On the other side, the Greens are hardly existent in East Germany with the exception of Brandenburg they ran in less than 5% of the communities for elections, while the participation is considerably higher in West German communities. Overall and interestingly, in the first line those political parties run for elections that are also present in the national parliament. Other, smaller parties hardly play a role. Even the ödp, which is a Bavarian regional party, only runs it 80 communities in Bavaria (out of 2,058) and hence in less than 4% of Bavarian communities. The right-wing parties are not present on a significant level either: The Republicans ran in 111 communities, which is equivalent to 1% of German communities, the NPD in 48 communities (0.4%). All other political parties run in less than 25 communities and are therefore irrelevant for the local party system, even within single states. But how can the different and low presence of political parties in the German communities be explained? Do the criteria of Wehling (1991) to explain the presence of political parties still hold true: size of community and political culture? Table 4: Political parties running for elections differentiated by size of community (in percent) N < 500 500-999 1000-1,999 2,000-4,999 5,000-9,999 10,000 20,000 > 20,000 Total 0 75.2 42.9 29.4 14.2 2.0 0.0 0.0 30.7 1 18.3 27.5 24.1 17.5 5.4 1.5 0.0 17.1 2 5.6 22.6 32.1 40.0 36.8 17.5 2.2 24.6 3 0.8 6.2 12.3 19.8 32.2 33.0 12.6 14.2 4 0.0 0.8 2.0 8.0 20.3 40.0 51.2 10.2 >= 5 0.0 0.0 0.1 0.5 3.3 8.1 34.0 3.0 The univariate distribution shows are clear relationship between the presence of political parties and the size of community: The bigger the community, the more political parties ran for elections in the last local elections: In 75.2% of all communities with less than 500 inhabitants, no political party ran for elections, in 43% of the communities with 500-1,000 inhabitants and in 29% of those communities with 1,000-2,000 inhabitants there haven t been political parties in the field of candidates. On the other side: There is no community with more than 10,000 inhabitants in which not at least one political party ran with a list at election. Communities in which more than 4 political parties participated in local elections can be found in the first line in cities with more than 20,000 inhabitants. This result is backed by the bivariate analysis, which shows a high correlation between the size of community and the share of political parties in the field of candidates (r =.574). The second explanation factor is according to Wehling (1991) the local political culture. Although Max Kaase (1983) penned the saying that measuring political culture is like "trying to nail jelly to the wall" and also Wehling (1985/1987) stated that local political culture cannot be 12

analysed in a cross-state analysis, two indicators could be applicable in order to get some information about the influence of local political culture: (1) If one assumes that the political culture in rural areas differs from the one in urban areas, the degree of urbanisation could be an indicator to analyse local political culture: According to the community-power-research (see for Germany Laumann/Pappi 1976) one could expect a higher presence in urban areas than in rural areas. The bivariate analysis regarding the degree of urbanisation shows a correlation, which, however, only can explain the share of political parties to a lower extent than size of community (r =.298). According to Holtkamp (2003: 21) political culture can also be defined as political behaviour of actors influenced by tradition. Hence, certain lines of tradition should be identified in the behaviour of the local actors. In other word, one could expect that political parties (and also local lists) run for election in those communities in which they participated already in the elections before. The bivariate analysis shows a high correlation between the presence of political parties and their presence in the previous elections (r=.473). Hence, the presence of political parties in the field of candidates can be explained by the structure and the political cultures and the paradigm of Wehling seems to be approved regarding these two indicators. Non-partisan voter associations: Non-partisan voter associations run in 73% of all German communities for election. In some states they run for election in almost every community for example in Bavaria (94%), Brandenburg (92%), Saxony and Baden-Wurttemberg (each 90%). There are only two states in which they participated significantly under-average: In Rhineland-Palatinate (38%) and in Lower Saxony (61%). But in general, one can state that non-partisan local lists are present almost area-wide and even have a higher presence than political parties on the local level. Table 5: Participation in last local elections by state (% of communities) Local lists Bavaria 94.2% Lower Saxony 60.8% North-Rhine-Westphalia 72.5% Hesse 82.4% Rhineland-Palatinate 37.7% Baden-Wurttemberg 89.5% Brandenburg 91.4% Mecklenburg-Vorpommern 72.3% Saxony 90.1% Saxony-Anhalt 73.8% Thuringia 84.4% Total 73.0% 13

The number of local lists running per community ranges from 0 to a maximum of 12. The figures in Table 5 show, that in 27% of the communities no local lists do exist. In half of the communities (49.6%), one local list and in 14.6% of the communities, two local lists ran for elections. In 8.8% of the German communities 3 and more local lists were in the field of candidates. But in contrast to the assumption that local lists are predominantly in those communities present where political parties are not organized, the analysis shows that the number of local lists, which ran for elections, increases also parallel to the increasing community size. In communities with less than 500 inhabitants, averagely only 0.55 lists of non-partisans ran for elections, in the size between 1,000-2,000 it is 1.25, and there are averagely 2.4 lists of non-partisan voter associations running in cities with more than 200,000 inhabitants. And according to that, the bivariate analysis shows hardly any correlation (-.061) between size of community and presence of local lists. The same is true for the degree of urbanisation, which also hardly can explain the presence of these non-partisans (-.113). The only factors, which can explain the presence, are the tradition: There is a correlation (.311) between the presence in the previous elections and their participation in the last elections. But despite this tradition, the presence of local lists in the field of candidacy cannot be explained by the same factors than the presence of political parties. Their presence seems to be dependent on heterogeneous reasons, which lead to the formation and the candidacy in local elections. However, this can t be explained by the structural data and will therefore be analysed on basis of the survey (see 4.5). The analysis of the presence has shown that political parties are not present in 30.7% and local lists are not present in 27% of the German communities. This lack of presence of those two political groups leads to the fact that in 14.6% of the German communities neither a political party nor a local list was in the field of candidates at the last local elections. On the one side independent candidates run for election in these communities, on the other side, in many small communities, it happens that no person or list was willing to candidate for elections. In that case, there is a blank list at the polling booth at the day of election and the voters then elect ordinary citizens of their community by majority. This can be found in the first line in Rhineland-Palatinate where in 58% (N = 1,375) of the communities only independent candidates and citizens can be elected by the citizens while in all other states this constellation is rare. In further 27.9% of the communities either only local lists or political parties run for elections partly with independent candidates. An actual situation of competition in the field of candidates in German communities can be found in 57.5% of the German communities: Here, at least one political party competes against at least one local list in order to get elected. In these communities, the political parties are dominant in the field of candidates: In 61.9% of these communities (N= 4,004 com- 14

munities) the share of political parties in the field of candidates is higher than those of the local lists. In only 776 communities (12%) local lists are dominant in the field of candidates and in 26.1%; there is an equal share of political parties and local lists in the field of candidates. Hence, in communities where both political actors run for elections, political parties generally have a higher share of the candidates then local lists. Table 6: Competition of political parties and local lists Ratio of competition Percent Valid percent Valid Dominance of political parties 35,6 61,9 Competition 15,0 26,1 Dominance of local lists 6,9 12,0 Total 57,5 100,0 In order to get some information on the development of the local party system and on the questions whether or not there is a decline of political parties visible, the shares of political parties and voter associations resp. in the field of candidacy have been compared: This comparison (see table 7) shows first, that the local party system seems to be rather stable: In more than 60% of the communities both types of political actors had the same share in the field of candidates in the last elections as well as in the previous elections. This univariate distribution seems to approve the hypothesis of tradition. However, the comparison also shows that voter associations gained ground in the field of candidates: They had a higher increase of their share in comparison to the political parties: While voter associations increased their share in 24% of the communities in comparison to the prior elections, the political parties increased their share only in 13.9% of the communities and lost at the same time in about one quarter of the communities their important position in the field of candidates. Table 7: Share of political groups in comparison to last election Voter associations Percent Political Parties Percent Increase 24,4 Increase 13,9 Stagnation 61,2 Stagnation 61,4 Decrease 14,3 Decrease 24,7 Total 100,0 Total 100,0 3.2 Support at election At local elections, the voters can decide between those political actors, which are present in their local community, those that run for elections. How is the support for the political parties and local lists on the local level in Germany? In the following, the election results will be presented in two steps: First, the overall election results on the local level will be presented and 15

in a second step, the results in those communities where the specific local actors run for elections will be analysed. Table 8: Average share of votes of main political actors by state State Political parties Local lists Independent Municipalities candidates Baden-Wurttemberg 78.7% 48.0% 7.0% 1,110 Bavaria 39.9% 45.5% 0.9% 2,058 Brandenburg 44.1% 45.0% 10.2% 421 Hesse 81.8% 15.9% 2.3% 426 Lower Saxony 78.7% 20.8% 4.7% 1,023 Mecklenburg-Vorpommern 43.3% 43.6% 13.0% 873 North Rhine-Westphalia 90.2% 9.6% 0.8% 396 Rhineland-Palatine 24.6% 17.9% 57.6% 2,306 Saxony 60.5% 39.4% 0.3% 518 Saxony-Anhalt 37.9% 41.0% 21.1% 1,121 Thuringia 39.1% 56.3% 1.5% 998 Total East 43.0% 45.7% 10.4% 3,931 Total West 46.4% 30.2% 19.3% 7,319 Total Germany 45.3% 35.4% 16.2% 11,302 The political parties win averagely the biggest share of votes in German communities: In the last local elections they won averagely 45.3% of the votes in each community with similar results in East and West Germany. Local lists have won 35% of the votes and independent candidates 16% of the votes. The differences between the states are, however, huge: Political parties are the clear winners in local elections in North-Rhine-Westphalia (90%), Hesse (82%) and Lower Saxony (79%). Here, voter associations as well as independent candidates have received averagely very low results per community. In Mecklenburg-Vorpommern and Saxony-Anhalt, local lists and political parties have received similar shares of votes in each community. Rhineland-Palatinate is a special case since here neither political parties nor voter associations are the dominant political actors, because they have received together averagely not even half of the votes per community. Here, independent candidates have averagely 58% of the votes. In all other states, local lists have won averagely more votes per community than political parties and are therefore the most important local actors in these communities. In Thuringia, voter associations were in comparison to all other states the most successful ones and won averagely 56% of the votes. But also in Baden-Wurttemberg (48%), Bavaria (46%) and Brandenburg (45%) local lists won averagely almost have of the votes in each community. Political parties won in these states averagely 40% of the share of the votes per community. 16

Furthermore, the analysis shows that voter associations have higher shares of votes in East Germany (46%) than in West German communities (30%). Within the political parties in all states the Christian Democrats are the clear winners since they have the highest share: Overall, they won 26% of the votes in each community with only minor differences between East and West Germany. They range between 47% in North- Rhine-Westphalia and 13% in Rhineland-Palatinate. In most states, however, they are grouped around 27% of the votes. The Social Democrats received averagely 14% of the votes. Their clear focal points are Hesse (38%) and Lower Saxony (33%). In the East German states, they have only very low results, ranging from 6% in Thuringia to 13% in Brandenburg. All other political parties are in this kind of analysis almost irrelevant, having 1.2% (Greens) to 2.3% (PDS) of the share per community. Table 9: Average share of votes by size of community Size ranges (inhabitants) Political Parties Local lists Independent N < 500 12.1% 34.0% 53.3% 2,196 500 1000 30.5% 42.4% 25.5% 1,980 1,000 2,000 42.6% 45.8% 5.0% 2,008 2,000 5,000 54.6% 38.5% 1.5% 2,359 5,000 10,000 70.5% 27.5% 0.3% 1,219 10,000 20,000 79.7% 19.1% 0.2% 806 20,000 50,000 85.2% 14.1% 0.2% 470 > 50,000 89.5% 9.9% 0.2% 180 But what are the reasons for these major differences? One explanation might lie in the size of community. As explained above, it is assumed that voter associations have their focal points in the small communities while political parties are the dominant actors in bigger cities. By looking at the different sizes of community (see Table 9), this assumption is approved: In all states, the relative share of political parties increases parallel to the size of community while the share of voter associations and independent candidates decreases. The independent candidates are in the first line dominant in communities with less than 500 inhabitants. Here, political parties receive averagely only 12% of the votes. In communities between 500 and 1,000 inhabitants, there is a different picture: Independent citizens receive only 25% of the votes. Here, local lists are the dominant actors receiving a relative share of 42% of the votes. The relative share of political parties is in comparison to the smallest communities clearly higher with 31% of the votes. In the next size range (1,000 2,000 inhabitants), the relative shares of local lists and political parties are equal high with 46% and 43% respectively. Independent candidates are in this size range only a residual category and are getting less and less important as bigger as the communities are. In communities with more than 17

2,000 inhabitants, but especially with more than 5,000 inhabitants the political parties have the clear majority of the votes and are the dominant actors in the local political system. Hence, one can observe that in communities with less than 2,000 inhabitants, voter associations and independent candidates are the central actors, while political parties are the central actors in the bigger communities and cities. Since this observation is true for all states, the differences between the states can at least partly be explained by the different community structures of the states. As demonstrated in Table 2, the community structure differs enormously between the different states with the extreme cases North-Rhine-Westphalia on the one side and Rhineland-Palatinate and Thuringia on the other side (see Kost 2003:198; Peter 2003:295). But why are political parties have such low election results in small communities and at the same time, why are independent candidates and voter associations successful in these small communities? There are two different assumptions possible: First, the reason could lie in a low presence of political parties in small communities. Hence, independent candidates and voter associations would serve as substitute for political parties. The second assumption would be that political parties are seen as interfering and not necessary for local politics in small communities, according to the notion of an apolitical self-government and hence, they are not elected by citizens. In bigger communities and cities, on the other side, political parties are assumed as important local actors and therefore receive better election results. The analysis of the supply showed, that political parties are hardly present in small communities. In order to get more information about the preferences of the voters, in the last step of the analysis of the election results, it will be analysed which relative share of votes the political actors receive in those communities they run for election. Table 10: Presence and success of political parties All communities Communities where at least 1 political party ran for election Size ranges (inhabitants) Share of votes Political Parties N Share of votes Political Parties N < 500 12.1% 2,196 48.9% 544 500 1000 30.5% 1,980 53.5% 1,131 1,000 2,000 42.6% 2,008 60.4% 1,417 2,000 5,000 54.6% 2,359 63.7% 2,023 5,000 10,000 70.5% 1,219 71.9% 1,195 10,000 20,000 79.7% 806 79.7% 806 20,000 50,000 85.2% 470 85.2% 470 > 50,000 89.5% 180 89.5% 180 18

The analysis of the election results based only on those communities where at least one political party ran for election shows that the low election results can be explained first and foremost by the presence of these local parties. The data presented in Table 10 shows clearly that political parties gain high share of votes in all size ranges if they run for elections. Hence, if they run, the citizens elect them. For instance, the Christian Democrats receive in those communities where they run for elections averagely a relative share of 41%. This share is rather constant ranging from 45% in the smallest size range to 40% in cities with more than 20,000 inhabitants; similar is true for the Social Democrats, the PDS and the Greens. The Liberals even have a very high election results in those small communities they run for elections (e.g. 34% in communities with less than 500 inhabitants, while they have similar to the results on the national and state level between 6 and 8% in cities with more than 5,000 inhabitants. Overall, it can be stated that they success of the political party is rather independent from the size of community. Their low shares in the small size ranges are in fact a result of the low presence in these small size ranges. Therefore, it doesn t seem to be a matter of disenchantment of political parties and neither a matter that citizens don t see a need for political parties in small communities. Therefore the success of the voter associations and independent citizens in small communities can be explained in the first line by the absence of political parties in these small communities. Therefore, voter associations are quite often the only groups running for election. It can be assumed that voter associations in small communities are not elected in the first line because they are perceived by the citizens as non-parties and anti-parties and are therefore. They seem to be rather elected by citizens because they are the only alternatives citizens have to vote for and are therefore in these small communities rather substitutes for political parties. Table 11: Existence of political parties and success of local lists Number of political Average share N (Number of communities) parties run for election of votes of NPVA 0 47.6% 3454 1 49.8% 1929 2 30.7% 2781 3 20.9% 1616 4 13.7% 1172 5 13.1% 255 6 and more 6.9% 95 19