Comparing Similar Local Political Systems Empirically
Department of Social Sciences, Carl von Ossietzky University Oldenburg, Oldenburg, Germany
Hiltrud Nassmacher. Comparing Similar Local Political Systems Empirically. Social Sciences. Vol. 4, No. 3, 2015, pp. 61-67. doi: 10.11648/j.ss.20150403.15
Abstract: Empirical comparisons of political systems can hardly avoid shortcomings. This is also true in comparing local political systems. The quantitative research on the basis of mass data nowadays is most fashionable. As the scientist is confronted with a multitude of municipal authorities a selection by chance seems to be appropriate. However, the local political systems vary very much. Nowadays designs for case studies take into consideration that polities and politics of global players among the municipal authorities are different from those in smaller towns. In 1973 Dahl already stated that there is a positive correlation between size and democracy. In Germany the new institutional arrangements at the local level were expected to bring about more participation by the citizens. This is more likely in smaller towns than it is in larger cities. However, the planning of empirical case studies on large, middle-sized and small municipalities should be aware that other variables have to be considered besides the number of inhabitants. If such variables are taken care of the dilemma of too many variables occurs. As empirical comparisons come to the fore, the question arises, why there is so little progress in strategies of comparative analysis by case studies. Scholars of political science have chosen different ways without reaching a consensus about the one suitable solution, which variables a researcher has to take into consideration when he or she is answering specific questions. This paper will discuss this problem against the background of scholarly knowledge about the local level of the political system in Germany.
Keywords: Local Government, Political Systems, Methodology, Case Studies, Comparisons
It is common practice in social sciences to use case studies of political systems only to generate new hypotheses. However, can such investigation strategy also be applied to produce comparative knowledge, which is appropriate for generalization of the results?
Political systems are very complex. Therefore in case studies researchers have to collect data, using different methods, which are compatible with their hypothesis. Because this is always very time-consuming, most research concepts will end up with just a small number of cases to be studied. However, "the problems of reliability and validity are smaller", if the researcher analyzes a "smaller number of cases more thoroughly, and he is less dependent on data that he cannot properly evaluate" (Lijphart, 1975: 170). If we follow this advice it is very important, to select the cases based on the current knowledge that is related to political systems.
Earlier considerations on comparing political systems (Nassmacher 2008) have pointed out, how important findings that result from different approaches in political science are for the creation of a comparative design for case studies, which is promising results that do not only apply to the analyzed cases alone but also to all other cases that are similar with regard to the same variables. In this paper this option shall be discussed for the political systems on the local level in Germany.
First of all, findings on local political systems in Germany are summarized in a short overview. Second, dominant strategies in empirical research are discussed with the intention to point out shortcomings. Next, a critical evaluation of two recent pieces of research that use case studies is given. Finally, a conclusion sums up the recommendations that result from this approach for the use of case studies in analyzing politics on the local level.
2. State of the Art: Analysis on the Local Level in Germany
Since two decades the basic institutional structures at the local level in Germany have changed fundamentally. This applies to Lower Saxony, North Rhine-Westphalia and the new states of the former GDR. Whereas the latter introduced a democratic structure, the former changed their division of power that was similar to a parliamentary system (with a temporal division of power between parliamentary majority and opposing minority) to a more presidential system (with a separation of power between the mayor and the council). This is the traditional set-up for local government in the southern states of Germany. There are two counteracting centers of power, both drawing their legitimacy from the vote of the people. The mayor who is the head of the local municipality, is boss of the administration as well as entitled to represent his city or town. As a politician the mayor has only power if he is able to interact with councilors to get support of the majority in the council. In theory the councilors are supposed to act as the principal and the administration as the agent. Therefore it is important that the mayor is supported by the discipline of a majority group or by a stable coalition in the council. To manage decision-making according to the policies that should be brought to the fore and implemented in time, controlling the majority is a stepping stone to survive politically.
Since decades differences of institutional arrangements on the local level in the various states of Germany have converged. The smaller the town the more important informal networks among the local political actors have become. Therefore the institutional arrangements (laid down in the law) are of lesser importance in small municipalities than they are in big cities. In small towns political parties play a minor role in the decision-making process.
Furthermore for decades empirical studies have pointed out that the electoral systems are an important factor in shaping the political system. Everywhere in Germany the mayor is elected by popular vote, with the trend, that elected is only the person, getting more than fifty percent of the vote. Since parties and other groups must not pass a minimum threshold in order to get seats in the council, the number of political groups in local councils has increased over time. Clear-cut majorities are rare. Ever since such changes majority-building has become harder for local politicians.
In addition the number of votes to be cast by an individual citizen differ among German states. Therefore in Germany it seems appropriate to compare only municipalities, which belong to the same state. Another factor to counteract the common picture is regional strongholds of parties based on specific socioeconomic structure as well as historical entities within individual states that are still influential.
This is also true for regional differences within municipalities, which complicate decision-making especially in local perspective. Specific bodies on the local level, e.g. strong interests of former independent parts of the municipality or citizen groups with the opportunity to organize special interests, have to be taken into account. With regard to the major interest groups embedded in special bodies for decision-making is of importance. This may also include informal participation in decision-making, e.g. in decisions concerning sports. It may even end up in corporatist structures as in social policy. Informal decision-making procedures (vertically with the state administration and horizontally in cooperation with surrounding cities) intervene, too. All of them lead to different political cultures, which reflect the political beliefs and values of people in a political system. The political culture is responsible for trust in and satisfaction with the political system. It leads also to typical procedures of decision-making, a more consensual style or a more majority oriented procedure. Besides formal responsibilities the decision-making styles are of great importance. This is also true for the current social and economic structure of a region, which are determining the resources on the local level.
Usually for a comparison of municipalities cases are selected by size (mega-cities, big, middle-sized cities, small towns) as measured by the number of inhabitants, because frequently these types of municipalities have to face different problems. Very large cities are global players, which bring about a special infra-structure and specific economy (Sassen, 2000, 2001). Middle-sized cities are fighting to defend what is left of their central functions against towns in the neighborhood, which compete for administrative authorities and provide a lavish infra-structure to serve the people better than the towns nearby. A mono-structured economy (or even just one large employer) in a city brings about a special lobby structure. The struggle among middle-sized cities and small towns concerns competition for inhabitants by offering building-sites for houses. However, the size of cities, as measured by the number of inhabitants, is not a sufficient indicator for all issues at stake. The location of the community has to be taken into account as well. In a rural area a middle-sized city has more central functions than its counterpart in an urban region, where other middle-sized and large cities of the neighborhood are the competitors for such functions.
3. Dominant Strategies in Empirical Research
Empirical research is often done by asking a sample of people via questionnaire in order to assess wider public opinion. Frequently the sample is very small. Some data analysts have significantly more experience with large data sets but they tend to have little training in how to interpret causal effects in the face of substantial selection (Grimmer 2015). In methodical matters political scientists still make use of the common methods of social science. These are provided by sociologists and thus they are not adapted to specific research problems, which have to be mastered in political science only. Moreover political scientists have also developed a wide range of strategies to measure quantities of interest from data carefully (Grimmer, 2015: 80). However, a lot of research questions cannot be brought to the fore.
A large field of interest for political scientists is the analysis of political decision-making. Decisions on policies and power are initiated, prepared, carried through and implemented in a given polity by special politics of politicians in communication with the people, using various options and windows of opportunities. For the local level this means that different politicians or people of the administration are selected for interviews by positions held or reputation as well as a sample of common people. Especially the research on the citizens’ participation relies solely on questionnaires. However, public opinion may influence the answers given by the people. If, for example, voluntary work in public affairs is valued high as it is right now, the interviewee may stress that he or she is prepared to participate in discussions and in the implementation of solutions. Therefore it is not enough to apply just one method only. Researchers have to contact the organizations, too, that rely on voluntary work. Their activists may provide a better picture on shortcomings in voluntary help. By now data on the willingness to participate and the real action differ very much (Nassmacher, 2006: 35-59).
This is also true if the researcher tries to find out, how much political parties are rooted in their society. A lot of people will not tell you that they have quit the membership of a political party in order to save money. Others are unwilling to point out their disappointment with political procedures or some people in their local party, who are blocking elective positions over years. Some may never have thought about their frustration more closely. A lot of people regard democratic decision-making as too time consuming. They therefore prefer other ways to influence the political process, e.g. in a citizens’ group advocating specific issues. In any case the quantitative information on the decrease of party members is often seen as a major hint that democracies loose linkage to their grass-roots. However it is only a small item of practical information about this problem. Such impressions have to be confirmed by results of local elections and by visits to an appropriate number of different party meetings, not only at election time.
Another problem in local politics is to get hold of actual incidents of corruption. Anonymously distributed questionnaires may be useful for a start, as politicians and people of the administration will not be ready to talk about this subject. Investigative journalists using informal contacts may collect information by pure chance and social scientists must start at this end to dig up more precise material by interviews with persons in different positions.
For research on city development, a set of relevant newspapers has to be the starting point. Collecting newspaper articles can prepare for the study of other documents (e.g. minutes of the council) and observations of the location under study. A lot of interviews with people of the administration, politicians of several parties and groups as well as with citizens who are living nearby will be necessary to get a full picture (see Nassmacher 1987).
In most analyses a complex research design is needed. Individual case studies, which answer specific questions for smaller and larger municipalities have been done frequently in the last decades. Some of them have already become classic works, as Lehmbruch on local parties in a party democracy (e.g. Lehmbruch 1975). They provide in-depth views. Meanwhile results from one single case are considered to be of limited value. Nonetheless research in only one municipality is mostly done because of limited financial resources for the research. Comparisons are still rare, but have come to the fore recently (Holtkamp 2010).
A lot of research on local political systems is undertaken in an environment, which is easily accessible to the researcher. He or she may bring in personal knowledge and strong feelings concerning the municipality. Longstanding observation of political procedures and developments is of great value. Frequently case studies are done in the neighborhood of the university (Gabriel, Brettschneider, Vetter 1997) or in the local political party, to which the researcher is affiliated. However, very often an explanation, why specific cases were selected, is not given.
Many researchers are familiar over years with the development of a municipality as far as different policies are concerned. In order to gather valid information it seems uncontroversial that case studies in a lot of municipalities will be more promising. Therefore such case studies have to be well-founded in the knowledge of political science. Furthermore each method causes specific deficits. A complex design, which meets high standards, will combine different methods to get the full picture and to compensate for the weaknesses of a single method.
4. Designs and Deficits of Case Studies
At present a well composed design is missing in comparisons based on case studies with a small amount of municipalities. This might be the major reason, why empirical research based on case studies is held in low esteem among social scientists. Such studies are not considered to allow for generalization of the results, but are only esteemed as explorative approaches, which prepare the ground for quantitative analyses (Lijphart, 1975:160). However, as the examples above have already shown, for a lot of topics quantitative methods cannot be the only opportunity to follow. Therefore considerations have to be set in train, how to create a variable-oriented design for comparisons with a small N, which provides stepping stones for a generalization of the research results. When they adopt the concept of comparison, researchers in political science hope to achieve a better understanding of the functioning of political systems, to get information about inherent problems and of those issues, which may possibly arise in the future, e.g. as a consequence of new circumstances after the implementation of specific solutions.
In Germany researchers of the local level may have in mind that despite reforms, which have led to a convergence of institutional arrangements, there is a lot of difference among the local political systems under federalism in Germany. The formation of local government seems to be equal as a consequence of popularly elected mayors. However, the electoral systems for the council, at least as far as the number of votes each citizen may cast and the manner of participation in selecting personnel for the councils are concerned. In the south of Germany the opportunities of participation are common, in other states they are new or not yet implemented long enough. In North Rhine-Westphalia the dominance of the local party bureaucracy has survived and the party bosses still prepare the lists of candidates, which the voters cannot influence. In the south of Germany there is a long tradition for the popular election of the mayor, whereas elsewhere the new opportunity to elect the mayor by direct vote of the citizens has to be learned as a major step to influence local politics and policies. There is also a tradition of using instruments of direct democracy to influence decision-making at the local level in the south of Germany. Such opportunities are new to all other states since German reunification. Therefore it seems to be necessary to create a set of similar case studies, which are all located in the same state.
In the studies on local government, which have been published recently, some researchers had in mind most of these considerations, others had not.
Van Deth and Tausendpfund chose the state of Hesse for their research on the importance of local politics. The size of municipalities under study was a criterion for their selection of cases. They selected case studies from 261 municipalities with 5,409 to 32,502 inhabitants and, in addition included the large cities of Kassel and Wiesbaden. The smaller cities and towns were selected at random (van Deth and Tausendpfund, 2013: 18) from the large number of local political systems, which are available for research in the state of Hesse. However, does this selection of cases represent an adequate way to respond to the problem of political orientation of the people on local politics, which is closely connected to the perception of the political performance by parties and politicians in the municipalities under study?
Time and again a lot of formerly independent smaller municipalities had to merge with neighboring ones. In Hesse this happened in the 1970s. The process of creating a new identity in the newly formed city or town often took several decades. Thus competition among the formerly independent districts of middle-sized cities as well as in larger towns is still strong, just as it was especially in the first decades after the merger. By selecting case studies for the analysis of attitudes towards the local level or of the decision-making process, you have to take this problem into account. To put middle-sized cities of some 25,000 inhabitants and towns starting with 5,000 inhabitants into just one sample may cause a problem and seems not appropriate.
A further item is to look at the location of the municipalities to one another. Politics of a metropolis or a city in the central position of a region and those of smaller towns surrounding such center in a conurbation always includes the options of conflict or cooperation. On the one hand the center offers a lot of services, e.g. in the fields of culture and higher education. On the other hand the center has to manage specific conflicts, which derive from social problems, those of transportation and the provision of workplaces. As the middle-class people as well as economic enterprises leave the central city either to live in their own house or to expand on cheaper grounds in suburbs or smaller towns nearby, the communication of these people and enterprises with the central city remain strong, for shopping, events, education and transportation opportunities alike. The central city has to provide the very costly infrastructure for the surrounding area, maybe without getting enough compensation in financial resources.
However, there are cities and towns in rural areas far away from conurbations. Sometimes these are middle-sized municipalities (e.g. of 50,000 inhabitants) and they have to perform as the center in the surrounding area with smaller towns and villages. If the research design puts together larger towns and those with some 5,000 inhabitants, this will cause a problem, especially as the financial situation is concerned. A lot of data is available for quantitative analysis. These are published in official statistics by the government of individual states as well as by organizations of the municipalities, e.g. the Association of German Cities or the Association von Cities and Towns which can be used by researchers. These publications provide data on the social structure as well as on income and expenditure of each municipality. Often larger municipalities provide such information for individual boroughs.
However, statistical methods reach the limits of usefulness rapidly, not only as the number of cases becomes too small. Nevertheless this material is helpful as a hint. For the test of hypotheses such data may not illustrate the problem in a proper way because they are not collected within the framework of the particular research project. Last but not least political research using mass data has to keep in mind the restriction that data can only reproduce facts, which are easy to quantify, e. g. expenditure figures. However, data collections never provide the information, how the figures are recorded. This can be demonstrated for figures on capital investment. There are poor municipalities with heavy debts and without any opportunity to invest for future generations or the years to come. Such debts may be caused by expenditure for prestige buildings, which were expensive to construct. Moreover in the following years will require ongoing subsidies from the public purse. Ordinary citizens are rather sensible and often judge this sort of expenses as a waste of money, being aware of potential financial problems and shortcomings in other fields of interest. Frustration about local politics is the consequence. Debt can also indicate that the relation between revenue (from municipal taxes and contributions by the state) and ongoing expenses (e.g. to fund measures on social problems and to provide for a suitable infrastructure) is not appropriate. The latter cause and effect is usually pointed out by top politicians of German towns and cities all around. If surveys aim at measuring political and social attitudes, values and behavior, the researcher has to be aware that public opinion will follow the line of interpretation for current problems, which dominates citizens’ observations and information. Considering this, statistical data may have "lured some researchers into a false sense of security" (Dogan, 1994:36) and may have prevented them from recognizing the real problems. Dogan (1994:37) stresses the problem of an "imbalance of quantitative data and the technological capacity of statistical treatment".
Therefore researchers who follow a different path tried to dig into the problem of local debts by combining quantitative and qualitative methods. Does exogenous or endogenous influence determine local budget deficits? (Bogumil, Holtkamp, Junkernheinrich, Wagschal 2014). In this study factors, which derive from the political procedures in municipalities, are taken into account: the number of groups in the council, some of them splinter groups, and the mayor’s party affiliation. The problem to compare quantitative data on debts arises from different organizations of the local administrations, which make the budgets highly fragmented. However, since 2011 the official statistics for Germany provides for common categories in budgeting (Bogumil, Holtkamp, Junkernheinrich, Wagschal, 2014: 617). These categories had to be implemented by all municipalities.
On this background the quantitative analysis of all 10,621 cities and towns in Germany took place. The result was that on average debts of municipalities in various states are very different. The highest figures were identified in the eastern part of Germany, whereas three states of the western part are listed as well with debts over average. This does not tell much about the reasons (Bogumil, Holtkamp, Junkernheinrich, Wagschal, 2014: 618-9). However, looking at the debts of individual states as compared to those of the municipalities located in these same states suggests that municipalities with high debts have to perform more tasks, which were transferred to them from the level above without financial compensation. Another hypothesis about high debts is, that problems of the socio-economic situation are important.
Moreover a strong left wing among the councilors or a mayor affiliated to the political left (Bogumil, Holtkamp, Junkernheinrich, Wagschal, 2014: 623-4) may cause high debts. The latter items highlight the hypothesis of party difference. However, the researchers cite other results, which point out that party difference on the local level does not matter much (p. 625) as far as debts are concerned. But this does not say anything about the size of municipalities, which might be important – as already mentioned above.
The study then turns to the decision-making process. The focus is on the mayor, whether he or she is in a strong position to seek solutions for specific policies, and the council, whether its leaders tend to act consensually or usually seek for majorities in decision-making (p. 625-6). There are quantitative results, that originate from the power, which is given to the mayor by the written constitution in individual states (p. 626). It does not seem adequate, to rely on this information to grasp the power of the mayor. (This will be discussed below in this paper). The answers to a questionnaire sent to all mayors in German cities and towns with more than 10,000 inhabitants were that the mayors see debts caused by exogenous reasons, such as the financial support of the state level, which does not meet the demands, caused by the problems in municipalities. However, looking more closely, endogenous socio-economic factors are also of great importance, e.g. the costs caused by unemployed people or by a large administration (p. 628-9). The number of members in the council, too, is seen as an important endogenous factor for debts. The researchers expect that the perception of success by councilors often is combined with expenditure. The problems of this quantitative analysis are that the results have not been controlled for different size of municipalities and for the coalitions, which govern them.
For the qualitative analysis only 16 case studies located in four states of Germany (two in the western and two in the eastern part) were selected: North Rhine-Westphalia and Baden-Wuerttemberg as well as Brandenburg and Saxony. The towns under study had between 20,000 and 100,000 inhabitants, however, they were similar in their socio-economic structure (p. 635). Statistical data pointed out that the cases are different concerning the debts: those with consensual decision-making have lower debts, those that prefer majority decisions have higher debts. Furthermore in consensual decision-making municipalities the mayor and the treasurer are in a strong position, which is confirmed in the institutional set-up as well as in interviews. Mayor and treasurer have the opportunity to bring together a majority that will save money, because usually there are free votes in the council, as happens in Baden-Wuerttemberg, Saxony and Brandenburg (pp. 637-8). Unfortunately information on the composition of the councils, on the number of inhabitants of the towns under study and on the persons who were interviewed is missing or only given in rudimentary form in the research report.
Despite the shortcomings, which were pointed out, the research by Bogumil, Holtkamp, Junkernheinrich, Wagschal is an important step towards a fruitful application of case studies in the comparison of local systems on a major problem, the financial deficits on the local level. Its major result is that against the mainstream hypothesis, for the financial problems of municipalities exogenous as well as endogenous factors may be important.
The major problem linked to the qualitative results in the study presented above is that the collection of data is – as usually - limited to a small number of municipalities. Steps to get information about the power of the leading persons, especially that of the mayor, remain an open question. From electoral statistics it is well known, how many votes led to the victory or the failure of a candidate who is running for the position of the mayor. But this does neither tell anything about the person vying for votes nor about what went well or wrong in his or her campaign. How significant has been the support of a political party or of specific groups among civil society, just to name some potential influences. Data for the social milieus may be available for parts of the municipality. Such data supply information on the strongholds of local parties and potential citizens’ support for a candidate. Local newspapers may be helpful for candidates or criticize them by the wording in articles and thus influence public opinion. Case studies for a more in-depth analysis are of great value, as those that were done by Nassmacher (2006: 22-34; 2013: 49-51).
The intention of empirical comparative research must be to identify rules that apply to all similar cases or those of a particular type. Thus for decades a tendency has taken hold in the discipline to analyze and compare more most similar cases in a research project. This is often seen as a stepping stone for generalizations derived from a larger set of units. The effort of Gerring to reach this goal on the basis of a single unit, defining case studies as "an intensive study of a single unit for the purpose of understanding a larger class of (similar) units" seems to be a setback (Gerring, 2004: 342).
A widespread consensus on what is already known about the major features of given units, which are typical and essential for local political systems provides the background for the selection of cases. This helps to lay the emphasis on specific variables, which are classified as dependent or independent variables according to the research question. The assumption that the case study method is an approach, which is not variable-oriented (Ragin 2000: 22), has to be negated, because hypotheses for the interaction of variables are provided by the results of different approaches in political science. Scholars have studied the complexity of political systems from different points of view, formulating and testing similar hypotheses. Using different approaches the knowledge about the complexity of the municipalities became much more elaborated.
The entities to compare have to be similar to each other concerning most variables (the context variables). This opens up the opportunity to look at specific variables, which are selected according to the hypotheses that have to be verified or falsified. Because of the complex structure of local political systems the inclusion of all variables would be impossible. The smaller the sample, the more variables can be taken into account. The larger the sample the more likely it is, that all characteristics of the entirety will be in it and the selected entity will be a complete copy of it. Following this path of choosing a random sample, however on the basis of a limited number of cases only, the scholarly results of the analysis would be just a sort of piece meal work.
The discussion about the major variables, which are important for the structuring and the functioning of local political systems, is not finished yet. You never know, if new variables are examined, which are relevant. Therefore the variables have to be named by the researcher, when he or she is constructing the design for a comparison. Thus critical views on the selection and emphasis are possible. If the selection of cases is restricted to just one state in Germany the institutional set-up is largely equal. Researchers should have in mind that observations might detect differences in political culture, which shape and modify individual choices of politicians and members of the administration. For example in policy studies it is advisable to know whether politicians prefer stable coalitions as the normal way of decision-making, tend to form coalitions if possible from time to time issue by issue or usually prefer a consensual style of decision-making. As direct decision-making is an option in municipal codes there are towns where people often make use of it and thus the risks caused by this instrument influence the decision-making, e.g. to avoid a popular vote in order to save money. Decisions in many fields are the results of corporatist structures. New institutionalism combined more or less different approaches in political science. These have pointed out, that the impact of institutions is more important for political scientists than the rules that are laid down in the statute book.
All political decision-making in local political systems follow historical developments, the social and economic basis, the values of actors and the people as well as the dynamics of institutional arrangements. People acting within such arrangements as individual or collective actors are guided by the arrangements and in turn they can shape them. In large municipalities formal structures are more important than in small towns. In spite of a widespread consensus on the relevance of the variables just mentioned, researchers have to disclose the basis of their classification.
Usually researchers prefer most similar cases or a most similar system design. This means, that many variables, which are similar, can be ignored, while the number of variables, that are different, can be controlled more easily. Quite in this vein Sartori recommends to "choose entities that are similar, if possible, in all variables, with the exception of the phenomenon to be investigated" (Sartori, 1994:22). After the selection of cases on the basis of this theoretical background researchers have to adopt appropriate methods from the arsenal of quantitative and qualitative methods. Researchers on the local level may be sure, that there are enough entities to apply quantitative methods only. As discussed above this assumption has to be negated. Comparisons are far from being perfect from the point of view by researchers that only tend to use mass data for quantitative analyses. However, a consensus on the impacts of an increasing number of variables is on the rise. As the creation of more complex designs advances systematically selected case studies and the analysis of mass data will supplement each other, depending on the research questions or the hypotheses, which are to be verified or falsified.