Citizens inter-municipal political orientations: evidence from Sweden

Transcription

1 1 Citizens inter-municipal political orientations: evidence from Sweden Anders Lidström Umeå universitet Statsvetenskapliga institutionen Umeå Sverige Papper presenterat på XVI Nordiska kommunforskarkonferensen, november 2006, Göteborgs universitet

2 2 Citizens inter-municipal political orientations: evidence from Swedish city-regions. Abstract: Theories of public participation, such as the Civic Voluntarism model, typically assume that local participation takes place in the polity where one resides. However, as citizens have become increasingly mobile, this assumption should not be taken for granted. Particularly in the city-regions (cities with functionally interrelated hinterlands) many citizens travel to other municipalities for work, studies, shopping or leisure activities. Many of these journeys take place between the central city and its suburbs. To what extent is there a corresponding intermunicipal political orientation among citizens? Can this be explained by the Civic Voluntarism model or are other theories needed? These questions are investigated on the basis of survey data from 7200 citizens from seven Swedish city-regions, belonging to three different size classes. The study is unique as investigations of these matters have not been previously carried out. The analysis shows that the Civic Voluntarism model is only partly relevant. Citizens that have a strong political engagement and belong to recruiting networks are more likely to have a political orientation that cuts across the municipal borders between cities and hinterlands, but access to extensive resources will generally not make such an orientation more common. Apart from revising the Civic Voluntarism model along these lines, this model also needs to be complemented by other factors. To have links across the city-suburb borderline, and to live in a large city region enhances the inter-municipal political orientation. Suburban dwellers also turn out to be more interested in city politics than vice versa. The findings have both theoretical and practical implications. The Civic Voluntarism model should be revised and expanded so it can account for cross-municipal political orientations. The findings challenges traditional territorial boundaries and calls for reforms that open up for new means of inter-municipal democracy, particularly in the larger city-regions.

3 3 Introduction 1 In modern societies, fewer citizens spend the whole day in the municipality where they reside. As urban populations and urban areas grow, more people have to adjust to the typical specialization of functions between different parts of the city-region. Increasingly, places for living, work, consumption and leisure are becoming separated. As more people are commuting, shopping in out of town shopping centres and spending their leisure time away from home, this has profound effects on urban social life. Travelling becomes an integrated part of the urban life style. Commuting extends the working day. Attachments to the local place may become weaker which can undermine not only social capital (Putnam 2000) but also people s identity and responsibility for the local community (Wellman 1999). The city-regions, consisting of cities with functionally interrelated hinderlands (Parr 2005), are typically politically fragmented. The common pattern in most of the western world is that a city-region consists of several independent municipalities. Apart from the city itself (or sometimes two or more cities), there are a number of other municipalities in the suburbs or near surroundings in which significant parts of the population are commuting or have other economic ties with the city/cities. Although there may be a second tier of local government for the whole of the city-region, major decisions about local policies are being taken at municipal level (Herrschel & Newman 2002, Hendriks et al 2005, Hoffmann-Martinot & Sellers 2005). Greater territorial mobility has consequences for urban democracy. Traditional administrative borders and the role of the municipality as the local polity are being challenged. The municipal citizenship, which is fundamentally linked to the area where one resides, weakens when a growing number of citizens spend their day elsewhere (Miller, Dickson & Stoker 2000). The connection between paying for and using collective services becomes more elastic as citizens from one municipality may use facilities such as roads, cultural institutions and leisure centres, financed by other municipalities (Lowndes 1995, Westerlund & Wyzan 1995). Attempts to establish regional tiers of self-government, covering the whole city-region, and transfers of resources between municipalities in an urban area represent attempts to handle some of the financial consequences of these changes. In an increasingly mobile society, the question inevitably arises why should citizen rights, such as voting and political activism be limited to the municipality where one spends the night? In the long run, urban mobility challenges present territorial borders and may generate a need for some form of intermunicipal democracy. The tendency of increasing mobility in the city-regions characterizes most of the Western world (OECD 2005, compare also Urry 2000 and LeGalès 2002). Despite the importance of these changes, no empirical studies have been carried out of citizen participation across municipal borders. Theories of political participation, whether emphasizing that this can be explained by institutions, rational actors, social capital, socio-economic status or civil voluntarism (compare Verba & Nye 1972, Barnes & Kaase 1979, Brady et al 1995, Verba et al 1995, Norris 2002, Pattie et al 2003) tend to be based on the assumption that citizens are static. The question of the location of these activities is not problematized at all or is expected to take place in one s own polity (national or local). The Civic Voluntarism model has turned out to be particularly fruitful for explaining why people participate in their own locality, but the model may not be sufficient it we want to understand why some people do so elsewhere. This may require that traditional theories are either complemented or replaced by other

4 4 theories. Hence, there is both a lack of empirical studies and a need for better theories that can account for this new feature of urban political life. The present study addresses this gap of knowledge by investigating these matters in a set of Swedish city-regions. The purpose is to analyse citizens inter-municipal political orientation across the city-suburb borderline, with regard to both how common it is and how this orientation can be explained. Ultimately the aim is to test and develop theories that may enhance our understanding of political participation across municipal borders. The paper continues with a discussion about the meaning of the concept of inter-municipal political orientation and introduces theories that may help us understand why this varies among citizens. This is followed by an introduction to the method and material of the study. After that, the relevance of the theories is tested, using mainly logistic regression analysis. The article ends with a summary of the main findings and a discussion about what the implications are for urban democracy and for further research. Theories of inter-municipal political orientation In this study, inter-municipal political orientation refers to citizens concern with local politics and public policies outside one s own municipality. Hence, it is analogous with an international or cosmopolitan political orientation, where the focus is on matters outside one s own country. Such an orientation can be thought of as consisting of both a mental and an action component: Citizens may not only be attentive to political developments and policies in other municipalities, they may also have made attempts to influence them. These two features inter-municipal political interest and inter-municipal political activism are likely to be related, as interest may trigger activism and vice versa. The two aspects will be treated as separate indicators of the dependent variable. One question that this study addresses is whether the Civic Voluntarism model can explain variation in political orientation across the municipal border, or if the model is best confined to participation in one s own locality. The Civic Voluntarism model, developed by Verba et al (1995) has become the dominant model in studies of political participation. It was originally developed as a contrast to earlier theories which suggested that differences in participation could be explained by socio-economic conditions. Verba et al argued that such explanations were too limited and claimed that the reason for people to not be active in political life was that they were unable to, did not want to or had not been asked to get involved. Therefore, resources, engagement and recruitment were seen as necessary requirements for citizens to take part in political life. Resources are about having time, money and civic skills. Engagement concerns the psychological orientation of citizens, for example whether they regard it possible to influence political decision-making, is interested in politics and supports a particular political party. The third requirement, recruitment, is about whether citizens have been exposed to social networks and voluntary organizations that can facilitate for them to participate. The Civic Voluntarism model, which has become the most commonly used model of public participation (Seyd, Whiteley et al 2001) has been supported by a large number of studies (for example Oliver 2001, Rose 2002, Pattie et al 2003, Barkan 2004, Lidström 2006, Eriksson 2007), carried out in different countries. Although the focus of the Civic Voluntarism model is on participation we will also expect it to be relevant in explaining cross-municipal political orientation (that includes not only participation but also political interest).

5 5 Another theory would emphasize the extent to which citizens have inter-municipal connections. As inter-municipal political orientation is beyond the traditional scope of public participation, it could be expected that general theories of participation would be unable to capture these matters. Instead, it could be argued that such an orientation is a direct result of the specific links that citizens may have with matters in one or several of the adjoining municipalities. Two such types of links are suggested to be of particular importance: First, citizens may be more or less rooted in a neighbouring municipality, for example by having previously resided in it, being an owner of a piece of property in it or having relatives that live in the municipality. Second, citizens may be a regular visitor to a neighbouring municipality. This includes regular commuting but also frequent visits for shopping or leisure. Our second theory assumes that these two types of inter-municipal connections create an interest in politics and an incentive for actively trying to influence policies in the other municipality. Finally, whether citizens have an inter-municipal political orientation may primarily be connected to where his or her home municipality is located. As the city has a core position in the city-regions, it is more likely that those living in the neighbouring municipalities will be politically oriented towards the city, than vice versa. The city functions as a magnet in its region, with specialized economic, cultural and administrative functions. The sheer importance of city politics and how it is covered by media is likely to attract interest and activism among those living in the suburbs (compare Williamson 2002). For similar reasons of importance it can be expected that there is a difference between city-regions of various sizes. In the largest city-regions, more citizens are likely to have a focus on political matters in neighbouring municipalities. The three theories that have been discussed, the Civic Voluntarism Model, a theory of intermunicipal connections and a contextual theory may not be alternative, but can instead complement each other. At least on a theoretical level, nothing suggests that they would contradict each other. Method, material and design of the analysis The validity of the three theories is tested on the basis of an empirical material, collected within The City-Region Democracy Survey. This is a citizen survey, carried out in seven Swedish city-regions during 2002 and The city-regions were selected in order to get reasonable variation in terms of population size, social characteristics and geographical location, but are, in a strict sense, not representative of all Swedish city-regions. The regions were Göteborg (large), Norrköping and Umeå (medium sized) and Karlskoga, Oskarshamn, Avesta and Värnamo (small) 3. Each city-region included the core city and the municipalities in the commuting hinterland 4. In each city-region, a sample of citizens aged between 18 and 74 was selected randomly. Postal questionnaires with questions about democracy and participation were sent out. The response rate was 62 percent. The data set consists of information from respondents. In the survey, most questions were similar for all respondents, but a section on intermunicipal relations was adjusted to if the respondents were city dwellers or lived in any of the adjoining municipalities. Citizens living in the city municipality were asked about their political orientation towards the surrounding municipalities in the city-region and about their other connections with matters in these municipalities. Those living outside the city

6 6 municipality were asked the same questions, but with regard to the city municipality. Hence, the study of inter-municipal relationships is confined to orientations across the city-suburb borderline within the city-regions. The design does not capture orientations or connections between different suburban municipalities or directed towards municipalities outside the cityregion. Although this study does not capture the whole picture of citizens inter-municipal relationships, it is concerned with the link that is likely to be the most common between city and suburbs. A study of all inter-municipal relationships would have been very complex. The relative importance of the three theories is tested mainly with logistic regression analyses. These are carried out on a data set consisting of all respondents, from all selected regions. In order to compensate for differences in sample sizes between the city-regions and to correct for known differences between the data set and true values for the whole population in each cityregion the data has been weighted with regard to municipal population size, sex, age and electoral turnout. Dependent and independent variables As already mentioned the dependent variable is seen as consisting of two components interest in municipal politics across the city-suburb borderline and attempts to influence decisions about such matters. Table 1 summarizes the responses to all questions in the survey on political interest and activism for each of the different city-region size classes. As could be expected, the forms of political orientation that has an inter-municipal focus are much less common than interest and activism that concerns one s own municipality. Hence, the connection to the home community remains strong. Contrary to citizens orientation towards political matters in the home municipality, inter-municipal political orientation varies with the size of the city-region. Interest in municipal politics that take place on the other side of the city-suburb divide is very extensive in Greater Göteborg, where almost one third of the citizens express such an interest. Inter-municipal political activism is a less common phenomenon, although it is more significant in the largest city-region. General interest in politics is also related to city region size. Table 1: Interest in politics and political activism in different city-region sizes Percentages Political orientation City-region size Small Medium size Large (Göteborg) Interest in politics In other municipality In own municipality Generally Political activism In other municipality In own municipality Note: Interest in politics concerns the share of the respondents that are very or fairly interested in politics (in a municipality on the other side of the city-suburb borderline than where they live, in their own municipality or generally). Political activism refers to the share reporting having tried to influence a municipal political decision (in a municipality on the other side of the city-suburb borderline than where they live or in their own municipality, during the last two years). For the wordings of the questions, see Appendix.

7 7 A separate test indicates that the two measures on inter-municipal political orientation are related to each other, with correlation coefficients (r xy ) varying between the size classes from.220 to.297. However, as they are not completely overlapping, the two measures will be used as separate indicators of the dependent variable, rather than being replaced by a joint measure. The other measures of political orientation in the table are included in the subsequent analyses as independent variables. The independent variables in the analysis have been generated from the three previously discussed theories. The Civic Voluntarism model is operationalized according to its suggestion that resources, engagement and recruitment are conducive for public participation. Men are expected to have greater access to resources than women. Also those who are older, more highly educated, belong to the middle class, are non-immigrant and who participate in the workforce or are students are expected to have greater amounts of resources. Engagement is captured by four variables: Trust in the local political system, party identification and interest in politics, both generally and in one s home municipality. Recruitment, finally, is represented by the number of voluntary organisations that the citizen belongs to. The assumption is that memberships in these make it more likely for a citizen to be asked to participate. In addition to investigating whether the factors suggested by the Civic Voluntarism model are relevant, attention will also be given to what this model is expected to predict, i.e. the level of political participation in the home municipality. In this way it can be analyzed whether it is the suggested explanatory factors, rather than the actual extent of participation in one s own municipality that matters if we want to understand variation in inter-municipal political orientation. The second theory, of inter-municipal connections, assumes that being rooted in an adjoining municipality or being a regular visitor to it creates an interest in political matters and a propensity to try to influence political matters in that municipality. Roots are represented by three variables; if the respondent has previously lived, owns property or has relatives in a different municipality. Regular visitors are those that commute to a neighbouring municipality, buy groceries or capital goods in it or go to that municipality for social, sports or cultural activities. All indicators concern connections across the city-suburb borderline. As Table 2 illustrates, roots and regular visits are common, but some types of connections are more common than other. Also, all forms are more common in the larger city-region sizes. Some differences between size classes are quite striking. For example commuting it is almost four times as common in the largest size class than in the smallest and it is twice more common to shop and carry out social activities. This suggests that citizens in city and suburbs are more interlinked in the large city region which perhaps also means that they are less attached to their particular home municipality. Despite the differences between the city classes, there are also forms of connections that are clearly significant in the two smaller classes.

8 8 Table 2: Connections across the city-suburb borderline in city-regions of various sizes Percentages Type of connection across the citysuburb City-region size borderline Small Medium size Large (Göteborg) Roots: Previously lived Owns property Having relatives Regular visitor: Commuting, at least weekly Shopping, at least monthly Leisure activities, at least monthly See Appendix for the wordings of the questions. One could have assumed that the six different types of connections would be closely interrelated, hence reflecting one underlying dimension. However, both factor analyses and an examination of correlation coefficients between these measures indicate that this is not the case. Among roots, the only significant correlation is between having previously lived in a different municipality and having relatives in it. Citizens that are regular visitors, on the other hand, are more likely to combine different types of visits. A particular common combination is to both shop and carry out leisure activities such at attending sports or cultural events in a municipality on the other side of the city-suburb borderline (r xy in the whole sample is.444). Finally, two aspects referring to the territorial context are also investigated. The first is whether citizens live in the city municipality or in any of the surrounding municipalities and the second concerns the size of the city-region. Analyzing inter-municipal political orientation The relevance of the different theories is investigated with the help of logistic regression analysis. This is suitable, as the two indicators of the dependent variable are dichotomies. Logistic regression analysis is also appropriate as it does not require that the dependent variable have a normal or equal distribution. As Table 1 has indicated, both measures of intermunicipal political orientation are skewed. Initially, separate regression analyses were carried out in the different size classes of cityregions. It turned out that the pattern of explanation that emerged was practically the same irrespective of size class and therefore, it was decided to conduct the analyses jointly, on a data set consisting of all respondents. Tables 3 and 4 summarize the results of the logistic regression analyses for each of the indicators of the dependent variable. First, the three theories are tested separately (model 1-3, 4 and 5). In model 6, all theories are examined jointly and model 7 represents the best fit with regard to the value of Cox & Snell R square.

9 9 Table 3: Explaining inter-municipal political interest across the city-suburb borderline Logistic regression analyses Inter-municipal political interest Independent variables The Civic Voluntarism model: Resources: Gender 0.153*** 0.148** Age Level of education 0.136*** 0.132*** 0.119** Social class 0.276*** 0.268*** 0.176** Immigrant Employment Engagement: Trust in local rep democracy 0.002* 0.002* 0.004*** 0.006*** Party identification 0.039*** 0.041*** 0.062*** 0.045*** Interest in politics generally 0.641*** 0.644*** 0.821*** 0.819*** Interest in politics in own municip 0.982*** 0.950*** 1.110*** 1.152*** Recruitment: Organizational memberships 0.141*** 0.133*** 0.100*** 0.146*** Political participation in own municipality 0.566*** 0.063* Connections across the citysuburb borderline: Roots: Previously lived 0.587*** 0.837*** 0.809*** Owns property 1.112*** 0.978*** 0.992*** Relatives 0.156*** 0.387*** 0.302*** Regular visitor: Commuting 0.418*** 0.561*** 0.602*** Shopping 0.241*** 0.479*** 0.439*** Leisure activities 0.556*** 0.554*** 0.561*** The territorial context: Suburb (dummy) *** 0.934*** City-region size 0.264*** 0.138*** 0.143*** Constant *** *** *** *** *** Cox & Snell R square Nagelkerke R square Number of cases Notes: The coefficients are B-values from logistic regressions. Levels of significance: * 0.05; ** 0.01; *** Data has been weighted for differences in population sizes between the city-regions and for differences between the data set and true values for the whole population in each city-region with regard to sex, age and voting turnout.

10 10 Table 4: Explaining inter-municipal political activism across the city-suburb borderline Logistic regression analyses Inter-municipal political activism Independent variables The Civic Voluntarism model: Resources: Gender 0.340*** 0.391*** 0.267*** Age Level of education Social class Immigrant Employment 0.295** 0.201* Engagement: Trust in loc rep dem ** 0.007*** 0.008*** Party identification ** * * Interest pol generally 0.493*** 0.454*** 0.548*** 0.548*** Interest pol municip 0.458*** 0.221*** 0.208** Recruitment: Org memberships 0.218*** 0.158*** 0.120*** 0.110*** Political participation in own municipality 0.972*** 0.696*** 0.690*** 0.735*** Connections across the citysuburb borderline: Roots: Previously lived 0.338*** 0.356*** 0.327*** Owns property 1.019*** 1.034*** 0.951*** Relatives Regular visitor: Commuting 0.583*** 0.485*** 0.533*** Shopping 0.378*** 0.398*** 0.343*** Leisure activities 0.388*** 0.335*** 0.351*** The territorial context: Suburb (dummy) *** City region size 0.197*** 0.079*** 0.057** Constant *** *** *** *** *** *** *** Cox & Snell R square Nagelkerke R square Number of cases Notes: The coefficients are B-values from logistic regressions. Levels of significance: * 0.05; ** 0.01; *** Data has been weighted for differences in population sizes between the city-regions and for differences between the data set and true values for the whole population in each city-region with regard to sex, age and voting turnout.

11 11 Although it is more common for people to take an interest in political matters in neighbouring municipalities than to have attempted to influence these, the patterns of explanation are very similar for the two different forms of inter-municipal political orientation. On the whole, all the three theories that have been suggested as explanations turn out to be relevant. The Civic Voluntarism model, the theory of inter-municipal connections and a contextual theory contribute to explain variation in inter-municipal political orientation. The theories tend to be complementary, rather than conflicting although they are much more successful in explaining inter-municipal political interest than activism across the municipal border. Hence, the Civic Voluntarism model is not only relevant for explaining participation in one s home municipality, it also helps us to understand why citizens have an inter-municipal political orientation. In particular, being politically engaged and a member of recruiting networks enhance citizens focus on public policies and politics in other municipalities. Trusting the system of local democracy, being interested in politics generally and to be member of several organisations also contributes. However, the resources of citizens are generally less important for such an orientation. There are exceptions those that are more highly educated and belong to a more wealthy social class tend to be more interested in politics outside one s own municipality, and men more than women have attempted to influence decisions in neighbouring municipalities, but apart from that, extensive personal resources is no requirement for developing a focus on political conditions in other municipalities. Not only are some of the variables in the Civic Voluntarism Model important as explanations to inter-municipal political orientations, but also the actual level of participation in one s own municipality contributes to explaining variations in inter-municipal political orientations. This is illustrated in models 1-3 in the regression analyses. Hence, those that participate in political activities in one s own municipality are more likely to have an inter-municipal political orientation. This is particularly relevant for inter-municipal political activism, where local participation remains being highly significant when also the Civic Voluntarism variables are taken into account. Hence, there is no trade-off between participating in one s own municipality and to have an inter-municipal political orientation. The theory of inter-municipal connections is also highly relevant. If you are linked to another municipality, you will also be more interested in political matters and will attempt to influence decisions in that municipality. Generally, both roots and to be a regular visitor are important. Citizens are likely to have a more extensive inter-municipal orientation if they have several types of links, rather than just one or two. The only exception is that those that have relatives in a neighbouring municipality is not more likely to attempt to influence decisions in that municipality, which is a bit surprising as these relatives may be ageing parents in need of social care. Finally, where the municipality is located is also important. The conclusion from Table 1, that inter-municipal interest and activism increases with city-region size, remains significant in the multi-variate analysis. Also, suburbians tend to be more interested in city politics than vice versa. City dwellers are more likely to attempt to influence decisions in suburbian municipalities, but this relationship disappears when other factors are being taken into account. Hence, the combination of a limited version of the Civic Voluntarism model, extensive intermunicipal connections and contextual factors is best suited to explain variation in inter-

12 12 municipal political orientation. Although each of these theories is significant, it should be emphasized even when they are combined, they only capture a limited share of the total variation. Hence, other factors, not included in this analysis must also be important. This is particularly relevant for explanations to cross-municipal political activism. The theoretical implications are obvious. Understanding inter-municipal political orientation requires more than the existing theories of public participation. The factors that generate local activism are only partly relevant as explanations to inter-municipal political orientation. In addition, it is crucial to be linked to neighbouring municipalities and to live in the right kind of environment, such as a suburb in a large city-region. Hence, a version of the Civic Voluntarism model that is revised along these lines would be able to capture orientations towards politics and policies outside one s own local polity. Conclusions and implications As citizens have become increasingly mobile, theories of public participation needs to be adjusted to take these new preconditions into account. This study has demonstrated that the Civic Voluntarism model, that has proven to be successful for explaining public participation generally, is only partly relevant as a basis for understanding citizens inter-municipal political orientations in Swedish city-regions. As no previous studies of such orientations have been carried out in other countries, it is difficult to say whether these conclusions are confined to Sweden, or if they would be relevant also elsewhere. Hence, there is a challenge for urban analysts generally to include the inter-municipal dimension into their studies. Empirical investigations in other countries could shed additional light on the political consequences of this increasingly important feature of urban life. Eventually, this may provide better and more generally applicable theories. Comparative studies would need to take into account that there are country differences. Cityregions in Sweden are different from those in many other Western countries, and in a way that may affect patterns of inter-municipal political orientation. First, inter-municipal political orientation has turned out to vary between city-regions of different sizes and to be most common in the larger city-regions. As Sweden is a sparsely populated country, and most cityregions are comparatively small, it will be important to investigate whether this orientation is even more pronounced in very large metropolitan areas. Second, despite the fragmented structure of Swedish city-regions, with numerous independent municipalities, they remain more coherent than in many other Western countries. Smaller municipal sizes in countries like the US, France, Germany and Switzerland make urban areas even more fragmented (Hoffmann-Martinot & Sellers 2005). This could make patterns of inter-municipal orientation more complex in these countries. On the other hand, if there are significant second-tier local authorities, with coordinating functions for whole urban areas, citizens political orientations may be multi-layered, rather than cross-municipal. Third, Swedish city-regions are mainly mono-centric. In more urbanized countries, metropolitan areas may be more poly-centric, consisting of several cities, overlapping commuting areas and extensive commuting patterns across the whole area (Kloosterman & Musterd 2001, Hall & Pain 2006). This may also contribute to more complex patterns of inter-municipal orientation. Tendencies towards inter-municipal political orientations have consequences for our understanding of local democracy. As the local citizenship and associated voting rights are normally connected to the polity where one resides, increasing mobility may challenge traditional forms of representative democracy. It has been suggested that citizens should be

13 13 given multiple voting rights in local elections and that these votes could be spread between municipalities of one s choice (compare Bauböck 2003). Referenda that cover several municipalities may also be carried out. Less formalized arrangements may include hearings and consultations that are not only open for the inhabitants of a particular municipality, but also for citizens who take an interest in that area for other reasons. The alternative to developing new means of inter-municipal democracy would be to change the territorial borders. Amalgamations of municipalities into large units that cover the whole urban region would efficiently eliminate problems associated with how borders are drawn. This would extend the local citizenship to the whole city-region and make it possible for citizens to exert influence over territories that were previously outside their own municipalities. Indeed, the last comprehensive amalgamation reform in Sweden, completed in 1974, aimed precisely at creating units that encompassed both a central city or town and its commuting hinterland. As commuting has now extended beyond these borders, larger municipalities would be a way of resolving this mismatch. However, municipal mergers may also have drawbacks. Larger municipalities may result in weaker connections between citizens and decision-makers, they may not correspond with citizens territorial identities and they are not necessarily more efficient than smaller units as service providers (compare Dahl & Tufte 1973, Rose 2002, Dollery & Crase 2004). There is also the public choice argument that amalgamations will reduce the scope for citizens to choose between municipalities with different combinations of services and tax levels, thereby limiting inter-municipal competition (compare Tiebout 1956, Ostrom 1972). In addition, if commuting areas continue to expand, many newly amalgamated municipalities would soon be obsolete. Hence, citizens political orientations across municipal borders pose challenges for both how local politics can be understood and how democracy can be organized in urban conglomerations. This makes it even more important for scholars to take this challenge into account in future studies of urban democracy and urban politics.

14 14 References Barkan, S. E., Explaining public support for the environmental movement: A civic voluntarism model. Social Science Quarterly, 85 (4), Barnes, S.H. & Kaase, M., Political Action: Mass participation in Five Western Democracies. Beverly Hills: Sage. Bauböck, Rainer, Reinventing Urban Citizenship. Citizenship Studies, vol 7 (2), Brady, H.E., Verba, S. & Schlozman, K.L., Beyond SES: A Resource Model of Political Participation. The American Political Science Review, vol 89 (2), Dahl, R.A. & Tufte, E.R., Size and Democracy. Stanford: Stanford University Press. Dollery, Brian & Crase, Linn, Is Bigger Local Government Better? An Evaluation of the Case for Australian Municipal Amalgamation Programes. Urban Policy and Research, vol 22 (3), Eriksson, K Spelar adressen någon roll? En studie av områdeseffekter på medborgares politiska deltagande. Academic dissertation, Department of Political Science, Umeå University. Hall, Peter & Pain, Kathy (eds), The Polycentric Metropolis. Learning from Mega-City Regions in Europe. London: euroscan. Hendriks, F., van Stipdonk, V. & Tops, P. (eds), Urban-Regional governance in the European Union. Practices and prospects, Elsevier. Herrschel, T. & Newman, P., Governance of Europe s City-Regions. London and New York: Routledge. Hoffmann-Martinot & Sellers, Metropolitanization and Political Change. Wiesbaden: VS verlag für Sozialwissenschaften. Kloosterman, Robert C. & Musterd, Sako, The Polycentric Urban Region: Towards a Research Agenda. Urban Studies, vol 38 (4), Le Galès, Patrick, European Cities. Social Conflicts and Governance. Oxford: Oxford University Press. Lidström, Anders, Commuting and Citizen Participation in Swedish City-Regions. Political Studies, vol 54(4), Lowndes, Vivien, Citizenship and Urban Politics. In Judge, David; Stoker, Gerry & Wolman, Harold (eds) Theories of Urban Politics. London: SAGE Publications. Miller, William, L., Dickson, Malcolm & Stoker, Gerry, Models of Local Governance. Public Opinion and Political Theory in Britain. Basingstoke: Palgrave.

15 15 Norris, Pippa, Democratic Phoenix. Reinventing Political Activism. Cambridge: Cambridge University Press. OECD, Employment Outlook. Paris: OECD. Oliver, E. J. (2001) Democracy in Suburbia. Princeton: Princeton University Press. Ostrom, E., Metropolitan Reform. Propositions Derived from Two Traditions. Social Science Quarterly, vol 53, pp Parr, John, Perspectives on the city-region. Regional Studies, vol 39 (5), Pattie, C., Seyd, P. & Whiteley, P., Citizenship and civic engagement: Attitudes and behaviour in Britain, Political Studies, 51 (3), Putnam, Robert D., Bowling Alone. The Collapse and Revival of American Community. New York, Simon & Schuster. Rose, L. E., Municipal size and local nonelectoral participation: findings from Denmark, the Netherlands, and Norway. Environment and Planning C: Government and Policy, 20(6), Seyd, P., Whiteley, P. & Pattie, C., Citizenship in Britain: Attitudes and Behaviour. The Political Quarterly, vol 72 (s1), Tiebout, C., A Pure theory of Local Expenditures. Journal of Political Economy, vol 64 (5), Urry, John, Sociology Beyond Socities. Mobilities for the Twenty-first Century. London and New York: Routledge. Verba, S. & Nie, N.H., Participation in America. New York: Harper & Row. Verba, S. Schlozman, K. & Brady, H. (1995) Voice and Equality. Cambridge: Harvard University Press. Wellman, Barry, The Network Community. An Introduction. In Wellman, B (eds.) Networks in the Global Village. Life in Contemporary Communities. Boulder, Col: Westview Press. Westerlund, O, & Wyzan, M.L., Household Migration and the Local Public Sector: Evidence from Sweden Regional Studies, vol 29, Williamson, Thad, Sprawl, Politics, and Participation: A Preliminary Analysis. National Civic Review, vol 91 (3),

16 16 Appendix Variables in the logistic regressions Variable Question/variable construction Variable values Indicators of the dependent variable Inter-municipal political interest For respondents in the city municipalities: How interested are you in political issues in any of the following municipalities: [a list of the surrounding municipalities in the city-region is presented]. 1 Not at all interested or not particularly interested 2 Very or fairly interested Inter-municipal political activism For respondents in the adjoining municipalities in the cityregions: How interested are you in political issues in [the name of the city municipality is presented]. For respondents in the city municipalities: Have you, during the last two years, tried to influence an issue on which politicians in any of the following municipalities [a list of the surrounding municipalities in the city-region is presented] make decisions? A list of 11 different policy areas was presented. Those indicating that they have attempted to influence at least one decision were coded as yes, the others as no. 1 No 2 Yes For respondents in the adjoining municipalities in the cityregions: Have you, during the last two years, tried to influence an issue on which politicians in [the name of the city municipality is presented] make decisions? A list of 11 different policy areas was presented. Those indicating that they have attempted to influence at least one decision were coded as yes, the others as no. Indicators of the independent variables Gender Are you a man or a woman? 1 Women 2 Man Age What year were you born? Varies between 18 and 74. Year of birth recalculated to age. Level of education What is your highest level of completed education? Seven response alternatives, in the analysis reduced to three. 1 Compulsory school, etc 2 Upper secondary school, etc Social class How would you describe your own household? Five response alternatives referring to social class, in the analysis these were reduced to two. 3 University, college 1 All other replies 2 White collar or higher white collar Immigrant Have you or any of your parents immigrated to Sweden? 1 No 2 Yes Employment What is your main occupation? 1 All other replies 2 Own business, full- or part-time salaried labour, or student Trust in local system of rep. democracy Additive index based on responses to three statements about the functioning of democracy in one s own municipality: - Municipal politicians do their best to make decisions that The index can vary between 0 and 9.

17 17 Party identification Interest in politics generally Interest in politics own municipality Organizational memberships Political activism in own municipality Previously lived Owns property Relatives Commuting correspond to citizens wishes; - If one really wants to, it is possible to influence the politicians decisions in the municipality; - The politicians in the municipality listen to the views put forward by ordinary people. For all three statements: 1 Agree fully; 2 Agree, but not fully; 3 Disagree, but not fully; 4 Disagree fully; 5 No opinion/don t know. The last alternative has been coded as missing value. The order of the others have been turned around (1=3, 2=2, 3=1, 4=0) and then added to an index. Do you consider yourself to be a convinced supporter of any political party? How interested are you in political issues generally? How interested are you in political issues in your municipality? A list with 13 different types of voluntary organizations was presented. For each of them, the respondent was asked to tick one of the following alternatives: Not member, passive member, active member or board member. The number of types of organizations that the respondent is a passive or active member in, or is a board member of was added up. If more than 6 were indicated, this has been coded as 6 (in order to exclude the undue impact extreme values) Have you, during the last two years, tried to influence an issue on which politicians in your municipality make decisions? A list of 11 different policy areas was presented. For respondents in the city municipalities: Have you previously lived in any of the following adjoining municipalities [names]? For respondents in the adjoining municipalities in the cityregions: Have you previously lived in [the name of the city municipality]? For respondents in the city municipalities: Do you own a property/leisure cottage in any of the following adjoining municipalities [names]? For respondents in the adjoining municipalities in the cityregions: Do you own a property/leisure cottage in [the name of the city municipality]? For respondents in the city municipalities: Have you got family/relatives in any of the following adjoining municipalities [names]? For respondents in the adjoining municipalities in the cityregions: Have you got family/relatives in [the name of the city municipality]? For respondents in the city municipalities: How often do you commute to any of the following adjoining municipalities [names]? For respondents in the adjoining municipalities in the city- Seven-grade scale, from 1 not convinced at all to 7 - yes very convinced. 1 Not at all interested 2 Not particularly interested 3 Fairly interested 4 Very interested The same as above Varies from 0 to 6. 0 Not attempted 1 Attempted in one area 2 Attempted in several areas 1 No 2 Yes 1 No 2 Yes 1 No 2 Yes 0 Never 1 Less frequent than once per week 2 At least once per week

18 18 Shopping Leisure activities regions: How often do you commute to [name of the city municipality]? For respondents in the city municipalities: How often do you buy groceries or capital goods in any of the following adjoining municipalities [names]? For respondents in the adjoining municipalities in the cityregions: How often do you buy groceries or capital goods in [name of the city municipality]? For respondents in the city municipalities: How often do you do at least one of the following in any of these adjoining municipalities [names]: visit sports or cultural events, spend the leisure time or visit friends/relatives? For respondents in the adjoining municipalities in the cityregions: How often do you do at least one of the following in [name of the city municipality]: visit sports or cultural events, spend the leisure time or visit friends/relatives? 0 Never 1 Less frequent than once per month 2 At least once per month 0 Never 1 Less frequent than once per month 2 At least once per month Periphery (dummy) Respondents home municipality 1 City municipality 2 Any adjoining municipality in the cityregion City region size City region population size 1 Small 2 Medium size 3 Large

19 19 1 This research has been carried out within the Urban Design Research Project at Umeå University, Sweden. This has been jointly financed by the County Council and the County Administrative Board of Västerbotten, eight municipalities in Västerbotten and the EU Objective 1 structural fund. An early previous version of this paper was presented at the ECPR conference, Marburg September I am grateful for valuable comments from the conference participants. 2 For further information about The City-Region Democracy Survey, see Lidström 2006, particularly Appendix I. 3 The city region of Göteborg has approximately 900,000 inhabitants, the medium size city-regions has around 150,000 inhabitants and the smaller around 50,000 inhabitants. 4 With two exceptions, the city-regions correspond to the local labour market areas, as defined by the Swedish national statistics agency Statistics Sweden. Each local labour market area consists of a central municipality and the adjoining municipalities that have their main commuting directed towards this centre. In two of the smaller city-regions the local labour market area covers a smaller territory than the municipalities selected as city-region. This was mainly done in order to investigate the validity of the local labour market area as a proxy for the city-region.