Output democracy in local government

Transcription

1 Output democracy in local government Jacob Aars and Dag Arne Christensen Affiliations: Jacob Aars Department of Administration and Organization theory University of Bergen Christiesgt. 17, N-5007 Bergen Dag Arne Christensen The Uni Research Rokkan Centre Nygårdsgt. 5, N-5015 Bergen Abstract Individual social resources are generally considered to be among the strongest individual-level predictors of political participation. When the way in which citizens relates to the political systems changes it is crucial to ask how this alters the sociodemographic make-up of the participants. Previous research has discussed this question related to different types of input related political mobilization. Far less had been said about the increased importance of participation directed towards government outputs. The aim of our paper is to examine the effects of individual social resources on both input- and out-related participation, and the relationship between these two forms of participation. We employ a comprehensive Norwegian survey comprising more than 10,000 respondents. We distinguish between three forms of input participation (party membership, voting and contacting) and two forms of output participation (contact with municipal officials and complaining) in the analysis. The analysis reveals that the more resource demanding forms of input participation is also most strongly related to individual resources. Contacting of councillors has, among other things, a distinct ethnic bias. We also find that the two forms of participation are partly overlapping. However, whereas Norwegian-born respondents are more likely to contact local politicians, there is not difference between Norwegian and foreign born respondents when it comes to contact with administrative personnel. Complaints do not seem to motivate or demotivate citizens to engage themselves in conventional politics. 1

2 Introduction In what ways do changes in patterns of political participation affect the social and demographical composition of the ones who participate? When certain forms of participation are on the rise and others decline, what consequences does this have for the socio-demographic make-up of those who participate? Do such changes contribute to even out or reinforce differences in participation rates between various socio-demographic groups? While these questions have been investigated with reference to the relationship between various forms of input participation, there has been little research on participatory acts directed at the output side of government. A number of authors have argued that an increasing share of public engagement is directed toward the output side of government. According to Peters (2010: 211), democracy have shifted significantly toward the output side of the public sector and The bureaucracy has become an important locus for public involvement and for democracy. At the same time, public sector offers opportunities for influencing one s everyday life in the capacity of user or consumer of public services. In addition, citizens seem to care more about the effectiveness of government than about representative democracy (Dahlberg and Holmberg 2013, p. 17). Hence, the legitimacy of government rests on output as well as on input factors (Rothstein 2009a). Moreover, it is argued that people are more often in touch with members of the public bureaucracy than with their elected officials. Our contribution is to raise the question of socio-demographic bias in participation with a particular focus on the alleged rise of output democracy, i.e. participatory acts that aim at influencing public administration rather than elected officials. We explore what characterizes the participants of this output democracy. First, to what degree do efforts at influencing public services constitute a distinct form of participation, different from traditional input participation? Second, since public sector is frequently inciting participation directed at government s output side, is the threshold lower for this kind of participation? Our empirical focus is participatory acts directed at influencing local government services in Norway. Local government is of particular interest since it increasingly is a provider of welfare services in western countries. Norway is relevant because Norwegian local government represents a system with comprehensive responsibilities within welfare service provision (Sellers & Lidström 2007: 617). In other words, Norway represents a case of 2

3 service-intensive local government in the sense that a major share of public services are produced and delivered at municipal level. The paper commences with a discussion of the relationship between equality and different forms of participation. We then proceed with developing some assumptions about the sociodemographic effects of participation that is directly aimed at public services. Subsequently, we present our data and methods before entering into the empirical analysis. The paper ends with a discussion of the democratic implications of the rise of output democracy. Participation and socio-demographic equality In democratic systems there are several ways in which citizens may influence public policies. First of all, influence may be exerted through the conventional representative channels. These include voting and party based activities. Second, citizens may employ less conventional forms of participation in order to influence policies. These are less clearly linked to the institutions of representative democracy. Cases in point are demonstrations, petitions and other forms of protest as well as consumerist political behaviour. The former has been labelled institutional participation whereas the latter has been categorized as noninstitutional participation (Marien, Hooghe & Quintelier 2010). The distinction between institutional and non-institutional forms of participation is similar to Norris (2007: 639) distinction between citizen-oriented and cause-oriented activities. The former is of a more general kind and relates mainly to elections and parties whereas the latter is more restricted in scope and relates to specific issues. Recently, it has been acknowledged that efforts at influencing government output should be counted as participatory acts. People may influence their life conditions by interacting directly with bureaucracies, e.g. through contacting administrative personnel, filing complaints or even responding to a user survey. Government on its side has been compelled, through various forms of marketization reforms, to increase its emphasis on performance and results. Accordingly, complaints and appeals are, to a greater degree, considered to be valuable inputs to the organization. Procedures have been established for handling complaints, and users are even encouraged to complain (Gilliatt, Fenwick & Alford 2000: 334). In sum, public administration has come to be a more important target for acts of participation. 3

4 Acts of participation that are aimed at influencing government services - output democracy - ought to be scrutinized according to the same kinds of criteria as other forms of political involvement. According to Schattschneider (1975) all political action entails a mobilization of bias. Attention is a scarce resource (Cyert& March 1963) and all citizens cannot take part in all activities at all times. Because equality is an essential principle underlying democratic government (Beetham 1999, Dahl 1991) the distribution of participation between various socio-demographic groups is a prime concern for all students of political participation in democratic systems. Democratic equality is most distinctly expressed in the maxim of one person one vote, but this idea in turn rests on the notion that all citizens are entitled to have a say in the governing of a state or another political entity. And even if all citizens cannot take part directly in governing, the norms of social and substantive representativeness (Birch 1964, Phillips 1995, Sartori 1968) in elected assemblies are manifestations of the same basic idea; that all citizens are possessors of genuine experiences that constitute valuable contributions to a political community. But even if there is a strong norm of equality, the reality is that certain groups tend to participate more actively than others. The right to participate is equal but the exercise of this right is unequally distributed within the citizenry. One of the most solid findings in political studies is the tendency for political participation to be skewed in favour of persons with high social-status rankings (Almond & Verba 1963, Barnes & Kaase 1979, Parry, Moyser & Day 1992, Verba & Nie 1972, Verba, Nie, & Kim, 1978, Verba, Schlozman & Brady 1995). Social bias in participation is explained first and foremost by differences in cognitive resources that are presumably needed to comprehend the intricacies of political processes and to undertake effective political action. In addition, high-status citizens are more likely to be part of networks in which political involvement is more familiar. In this respect, education (as the predominant status indicator) provides a sorting mechanism to link high-status individuals to networks of political relevance (Campbell 2013, Nie, Junn & Stehlik-Barry 1996). Finally, sociodemographic elites are thought to have higher stakes in the outcomes of political decisions (Pettersen & Rose 1996: 56). In turn, the uneven utilization of political rights may create further inequalities in political outcomes. Those who are inactive and silent in the political process are less likely to have their interests taken into account when important policy decisions are being made. 4

5 Potentially, differences in individual resources can be counterbalanced by collective resources. For instance if party organizations are membership based and mobilize citizens on a broad basis such organizations may compensate for lack of individual resources (Campbell & Rokkan 1960, Parry, Moyser & Day 1992). Moreover, collective organizations can offer alternative networks that may serve as leverage to political involvement. Collective organizations may lastly formulate interests of low-status groups in a manner that rises these groups claims in political outcomes. However, party membership is generally in decline (Biezen, Mair & Poguntke 2012) and the effect of individual-level resources seems to persist. The various social status arguments presented above however imply a differentiation with respect to forms of participation. In other words, the threshold for involving in political activities varies between different channels of participation. The cognitive resource argument suggests that participatory acts that are more demanding in terms of cognitive efforts and time consumption will tend to be more socially restricted. Hence, voting is expected to be less socially exclusive than for instance initiating a petition or contacting an elected official. Moreover, a consequence of the network argument is that participatory forms that require membership in a network is assumed to be more socially biased (NBNB SJEKK DENNE). Finally, with respect to political outcomes, participatory forms that are likely to have more consequential outcomes are also likely to be more skewed towards elite participation (NBNB DENNE OGSÅ). As a consequence, certain participatory channels are more liable to be dominated by social elites than others. This argument also pertains to fluctuations in participation between different channels over time. If the social profile of participants varies from one participatory form to another, one implication of the argument above is that the decline or rise of a particular form of participation will have potential repercussions for the overall sociodemographic composition of those who participate. On the one hand, the rise of new participatory forms may have the potential to reduce bias in participation but the opposite may be just as likely. A major part of the literature on political participation has explored the changing relationship between the older and the newer forms of political engagement. One 5

6 crucial question has been whether various forms of participation attract different participants or if new forms of participation just offer the established elites new ways to exert their influence. In other words, is participation differentiated or cumulative (Pettersen & Rose 1996, REF.)? From the perspective of equality, it is less harmful to democracy if the pattern is differentiated rather than cumulative. When different social groups dominate different participatory forms, lack of presence in one arena is counterbalanced by presence in another. The empirical literature on patterns of participation does not fully agree as to whether certain elite groups tend to dominate all participatory forms or whether the picture is more differentiated. Barnes & Kaase (1979) found that there were great overlaps between institutional and non-institutional forms of participation, i.e. more or less the same kind of people takes part in both arenas. In a recent comparative study including 25 countries Marien et al. (2010) also conclude that particular groups are more or less omnipresent in all political activities but still there are evidently some compensatory tendencies in the material: While certain groups are less active in some forms of participation, they tend to make up for their paucity by being more active through other channels. This is for instance the case with women, who tend to be underrepresented in the conventional, representative arenas but at the same time they are more active than their male compatriots when it comes to non-institutional forms of participation (Marien et al. 2010: 203). These conclusions are mainly supported by the findings of Pettersen & Rose (1996: 79). In an earlier Norwegian study, Olsen & Sætren (1980: 55) observe that non-institutional forms of participation are more dominated by education elites than is the case for institutional participation. Output democracy and equal participation What are reasonable expectations for the socio-demographic distribution of output-directed forms of participation? We relate the answer to this question mainly to the resource argument presented earlier in the paper. Concerning individual resources, there is surely a high threshold for participating among many residents. Bureaucratic structures are indeed intricate and may require a great deal of skills to figure out what is the appropriate address for one s concerns. What is more, the 6

7 concerns of many users are multiple and thus require the attention from various public offices. The task of linking together deliveries from various providers in order to obtain tailored services is potentially demanding. On the other hand, as a result of NPM-inspired ideas, public bureaucracies are subject to competition, either from private providers or from alternative public providers. Competitive tendering entails a greater sensitivity to users needs. Therefore, public agencies are expected to facilitate contacts and user influence. In short, public service providers are urged to become more responsive. User surveys have become standard instruments of measuring user satisfaction (Bouckaert & Van de Walle 2003). Post-NPM reforms, for their part, are aimed at achieving more holistic government structures. Recently, various public agencies have introduced schemes in order to customize services to the individual user s needs. In sum, public service providers are enhancing consultation with their users as well as assisting users in their contact with public offices. With regard to collective resources, we find few of the mass-based collective organizations that have characterized electoral politics. Political parties or mass-member interest organizations are rarely aimed at the output side of government. Thus, in output democracy, mass-based collective organizations hardly serve as compensatory resources for those who are short of individual resources. On the other hand, a number of more associations aimed at representing specific user groups interest are indeed present in the field of welfare policies. Many such organizations are deliberately established in order to support groups with little resources of their own. In other words, user associations may constitute extensions of individual resources. The question of motivation is also relevant in this context. Political acts are often the result of particular initiatives, prompted by a feeling of injustice. Engagement may be spurred by ideological convictions or by the fact that a person or group is affected by a specific issue. Output directed participation is rarely provoked by ideological considerations. On the contrary, such participatory acts are prompted by very specific and limited concerns. By this we do not imply that such concerns are limited in importance for individual users. But they are limited in scope, restricted to the particular issue in question. Moreover, contacts with public service providers are sometimes more or less inevitable, driven by the necessities of social hardship. Some such contacts are even automatic, at least if some initial qualification criteria have been fulfilled. Pension payments or child benefits are cases in point. In sum, 7

8 this means that many acts of output directed participation are not initiated by the individual user of services. In turn, this potentially lowers the cost of engaging. Data and method The data are collected from the Norwegian Citizen Survey The survey consists of answers from a sample of individuals (response rate 41 percent). The survey was conducted by the Agency for Public Management and egovernment (DIFI) (DIFI report 2015:5). The survey is mainly designed to measure citizens satisfaction with public services, but it includes questions about political behaviour as well. Before turning to the method used in the analysis, we look at the key indicators we use to measure input and outputrelated participation. Our primary objective is to study political equality across the two modes of participation. How equal is the participation patterns when it comes to social background? Do output related participation channels increase or decrease inequalities among citizens? We derive our measures of the two modes of participation from the Citizen survey which includes a limited number of questions about participation. Nevertheless, the questions included in the survey enable us to analyse both input and output based participation. We use active party membership, contact with (local) politicians and electoral turnout in the 2013 parliamentary election as indicators of input participation. As indicators of output participation we use contact with the local public administration and whether or not the respondent has complained about a public service. We do, however, lack measures of non-institutionalized participation like participation in manifestations, writing newspaper letters, and thus we are unable to compare with such forms of political activism (see Verba and Nie 1972; Petersson 1991; Parry et al. 1992; Verba et al. 1995). That said, contact with local politician on the input side could conceivably be defined as belonging to non-conventional forms of political participation. Two input indicators, active party membership and contacting, are based on simple yes/no questions in the survey. When it comes to party membership the respondents are asked if they are active in some of the following organizations, where political parties are one of the categories. They are also asked if they, in the last 12 months, have had contact with a politician in the municipality about issues that concern you. The third and final input 8

9 indicator, turnout, is measured using a question aimed to tap the respondents party preferences in the 2013 parliamentary elections. Respondents answering that they did not vote has been separated from the rest in the analysis. Turning to the output indicators, we rely on the above mentioned contact question which also asked respondents if they had been in contact with an employee in the municipality the last 12 months. Finally, complaints are used as our second output indicator. This is derived from a question in the survey were respondents are asked: Have you ever filed a complaint to the municipality, the county or a state agency on a decision? The item has four response categories: 1. Yes, 2. No, have not had any reason, 3. No, did not know how to complain, and 4. No, did not think it would help. In the analysis we have separated respondents answering yes from the rest. Table 1 shows the frequencies for both modes of participation. As can be seen from the table citizens seem to be more in contact with the public bureaucracy than with their elected officials. A minority of respondents (5 per cent) are active party members, and 16 per cent have been in contact with a local politician about issues of concern to them in the last 12 months. The latter can be compared to the 34 percent that say that they have been in contact with a municipal employee about issues that concern them. Around eleven percent of the respond have filed a complaint. Turnout is not only the least costly form of participation, but also the most common. 95 percent of the respondents say that they voted for one of the parties in the 2013 parliamentary election (the actual turnout rate was 78,2 percent). Even if it is not a simple yes/no question the percentage, as generally is the case, of people who say they voted by far exceeds the actual turnout rate (Granberg & Holmberg 1991). However, we are not interested in the level of participation as such, but how social background impacts different modes of participation. Compared to for instance contacting turnout demands little of citizens resources. This in turn makes it interesting to compare how the different modes of participation are linked to social background. 9

10 Table 1. Input and output participation (percentages) Input-participation Percent N Party membership Contact local politician Turnout (2013 election) Output-participation Contact local public administration Complained (any public service) The statistical analysis builds on standard logistic regression. The dependent variables are coded straightforward, and based on the survey questions reported in table 1. Respondents who reported that they have participated in the different input and output related indicators are coded 1, all others 0. In order to obtain reliable estimates of the different individual social background variables the analysis reports probabilities with municipal clustered standard errors. We perform the analysis stepwise. First, we run separate models for the two modes of participation and next we study the overlap between them in the form of a regression analysis of the three input indicators as a function of the output indicators controlled for social background. Empirical Results Do the two modes of participation vary when it comes to the impact of social background? Is the variation greater within or between them? Tables 2 A and B depicts the results in form of logistic regression models for each of the two modes of participation. The first table (A) shows the results for our input related indicators, whilst the second (B) shows the corresponding results for the two output indicators. With respect to the input indicators we see that individual resources are more closely tied to contacting compared to party membership and turnout. Starting with party membership the analysis depicts that men and Norwegian-born respondents are more likely to be active party members. Since the entries in the table are probabilities, they should be easy to interpret. For instance, the probability is 2.4 per cent higher for men being active party 10

11 members compared to female respondents. Foreign born respondents, on the other hand, are 2 per cent less likely to be active party members compared to native Norwegians. The probabilities for the age variable should be compared to the reference category which is the youngest respondents (18-24 years of age). Being young (18-24) and old (over 67) seems to increase the probability of party membership. Turning to contacting the analysis reveals larger gender differences compared to party memberships. Male respondents have a 4.6 percent higher probability of having contacted a local politician compared to female respondents. Furthermore, older respondents as well as those with high income and education have an increased probability of having contacted a local politician. Note also that foreign born respondents are close to 8 percent less likely to have been in contact with a local politician the last 12 months compared to Norwegian born respondents. Finally, the gender differences are not significant for turnout, whilst the age effect seems to be more or less linear. The probability of voting increases as respondents grows older. The analysis includes a measure of contact with the Norwegian welfare administration (NAV). NAV is responsible for a host of welfare services, amongst them social security benefits and social insurance. Having personal experience with NAV seems to play a minor role when it comes to input related participation. Personal experience with NAV has no significant effect on party membership and voting, but such experience increases the probability of having been in contact with a local politician (3 percent higher compared to respondents without this experience). Thus, this first step of the analysis shows that among the three input indicators in the analysis individual resources are more closely tied to contacting than it is to active party membership and turnout. Therefore, participation that is more strongly linked to traditional representative institutions is less socially skewed than participation with a weaker link to such institutions. 11

12 Table 2. Logistic regressions: Input and output participation (predicted probabilities with municipal clustered standard errors). A) Input participation Party membership Contact with local politician Turnout 2013 Constant Male.024*.046* Age: ** *.080*.026** *.037* 67>.021*.076*.057* University education.019*.038*.028* Income *.009* Foreign born -.020* -.077* -.041* NAV User * Pseudo R N B) Output participation Contact: local public administration Complained Constant Male.044*.018* Age: * *.063* *.039* 67>.082*.021*** University education.081*.021** Income.010*.000 Foreign born NAV user.107*.067* Pseudo R N *p<.001,**p<.05, ***p<.10 12

13 What about our two output indicators, contact with the local public administration and service complaints? Starting with the gender differences, we see that the result more or less echoes the results for the three input measures. Men have a higher probability of both contacting local administrative officials and complaining but the gender differences are not great. Age seems to be more important for contacting than for complaining. The two age groups between 35 and 66 have around 17 percent higher probability of having contact with the local public administration compared to the youngest respondents (18-24). This should come as no surprise since the two age groups consume more local public services compared to the youngest respondents. The probability that respondents with a university degree have contacted the local public administration is 8 percent higher than the less educated, but the educational differences are less pronounced when it comes to complaining (2.1 percent). Income seems to have minor impact on output related participation. Two findings stand out as especially interesting compared to the results of the input indicators. First, ethnicity seems to play no role when it comes to output participation. Contact with the public sector as measured here is not related to country of origin. Second, personal experience with the welfare state is far more important for output participation than for input participation. Personal experience with NAV increases the probability of both having had contact with the local public administration (10 percent higher compared to non- NAV user) and having filed a complaint (7 percent higher compared non-nav user). Whilst it is natural that service use is important for being in contact with the public sector, it seems to play a minor role for political engagement. So far, the results suggest that output related participation can both increase and decrease patterns of inequality. In the analysis we have compared very different forms of participation, but in figure 1 we take a closer look at contacting. Contacting related to the input side (contact with local politicians) and the output side (contact with local administrative employees) are especially interesting to compare. The two questions are part of the same battery in the survey and the actual survey question is identical apart from the two words politician and employee. They also intend to tap more demanding forms of participation. Figure 1 compares them in detail, and as can be seen only foreign born respondents are less likely to have contacted a local politician whilst ethnicity play no role when it comes to contacting a local employee. Again, and less surprisingly, NAV users have a 13

14 higher probability of contacting an employee, but such personal experience also matters for having been in contact with a local politician. Figure 1. Contacting: Local politicians and local employees (N=9503, marginal probabilities with 95% confidence intervals) Politician NAV-user Employee NAV-user Foreign born Foreign born Income University education Income University education 67> Age: > Age: Male Male Do input and output participation overlap? In the final step of the analysis we relate the three input participation measures to the two output measures. Previous studies show that policies may play a role in shaping political participation (Shore 2014). Policies may motivate or demotivate citizens to become engaged in politics due to the importance policies can have for them. For instance, having filed a complaint on a public service can either motivate or demotivate citizens to engage in politics. Figure 2 shows how having had contact with the local public administration (contact employee) and having filed a complaint relate to the three input indicators. In the analysis we have controlled for all social background variables included in the previous analysis. The results reveal that contact with a local public servant overlap all the three input indicators. Respondents having contacted a local employee have a higher probability of being party member, having contacted a local politician and voting in elections. The overlap is especially strong between input and output contacting. The probability of respondents having contacted a local politician is 30 percent higher for respondents that have been in contact with a local public employee in the last 12 months. 14

15 Respondents having filed a complaint are, on the other hand, not more likely to be a party member or to vote in elections. Complaints do not seem to motivate or demotivate citizens to engage themselves in conventional politics. Complainers do, however, have a higher probability of being in contact with a local politician. Figur 2. Input participation as a function of output participation (Marginal probabilities with 95% confidence intervals controlled for social background) Active party membership Contact employee Contact employee Contact local politician Complained Complained Turnout (2013 election) Contact employee Complained Conclusion Our concern in this paper is with output democracy, i.e. direct attempts aimed at influencing public service delivery. More precisely, we have analysed the socio-demographic profile of output-directed participation. Whereas a number of studies have compared the sociodemographic makeup of different forms of input participation (institutional vs. noninstitutional) very few have given attention to participatory acts directed towards public bureaucracies. We have concentrated on two indicators of output participation: Contacting local administrative officials and filing a complaint. The results have been contrasted with three forms of input-directed participation: party membership, contacting locally elected officials, and turnout in elections. 15

16 As for the input indicators, the more resource demanding form of participation contacting local councillors - is also most strongly related to individual resources. Persons with high income and/or education have significantly higher probabilities of contacting local politicians. Even more importantly, contacting of councillors has a highly distinct ethnic bias: Foreign-born citizens have an 8 per cent lower probability of contacting a local elected official than have Norwegian-born citizens. If we compare this pattern with the one that emerges when we analyse contacting of administrative officials, we can conclude that output-directed participation does not favour Norwegian-born residents in the same manner. On the contrary, contacting local bureaucrats is unrelated to country of origin. In aggregate, the results reveal patterns of participation that are partly overlapping, though still rather differentiated in some respects. The two forms of contacting (elected vs. administrative officials) are clearly related. Hence, contacting a local councillor increases the likelihood of contacting a local administrative officer. Both forms of contacting are skewed in the sense that education as an individual resource incites contacting. On the other hand, whereas Norwegian-born respondents are more likely to contact local politicians, there is no difference between Norwegian and foreign-born respondents when we observe contacting of administrative personnel. Output democracy thus is characterized by some of the same social biases as input participation but judging from our evidence it appears to be more open towards certain groups that are normally marginalized, namely foreign-born citizens. Complaints do not seem to motivate or demotivate citizens to engage themselves in conventional politics. References Barnes, S. H. & Kaase, M. (eds.) (1979). Political Action: Mass Participation in Five Western Democracies London: Sage. Beetham, D. (1999). Democracy and Human Rights. Cambridge, Oxford, Malden: Polity Press. Biezen, I.V, Mair, P., & Poguntke, T. (2012). Going, going,.gone? The decline of party membership in contemporary Europe. European Journal of Political Research, 51(1), Birch, A. H. (1964). Representative and responsible government: An essay on the British constitution. Toronto: University of Toronto Press. 16

17 Bouckaert, G., & Van de Walle, S. (2003). Comparing measures of citizen trust and user satisfaction as indicators of good governance : difficulties in linking trust and satisfaction indicators. International Review of Administrative Sciences, 69(3), Campbell, Angus, and Rokkan, Stein (1960). Citizen participation in political life: A comparison of data for Norway and the United States. International Social Science Journal 12: Campbell, D. E. (2013). Social networks and political participation. Annual Review of Political Science, 16, Cyert, R. M., & March, J. G. (1963). A behavioral theory of the firm. Englewood Cliffs, NJ, Prentice-Hall. Dahl, R. A. (1991). Democracy and its Critics. Yale University Press. Gilliatt, S., J. Fenwick and D. Alford. (2000). Public Services and the Consumer: Empowerment or Control? Social Policy & Administration, 34, 3, Granberg, D., & Holmberg, S. (1991). Self-reported turnout and voter validation. American Journal of Political Science, Marien, S., Hooghe, M., & Quintelier, E. (2010). Inequalities in Non-institutionalised Forms of Political Participation: A Multi-level Analysis of 25 countries. Political Studies, 58(1), Nie, N. H., Junn, J., & Stehlik-Barry, K. (1996). Education and democratic citizenship in America. Chicago: University of Chicago Press. Parry, G., Moyser, G., & Day, N. (1992). Political participation and democracy in Britain. Cambridge, UK: Cambridge University Press. Peters, B. G. (2010). Bureaucracy and democracy. Public Organization Review, 10(3), Pettersen, P. A., & Rose, L. E. (1996). Participation in local politics in Norway: Some do, some don't; some will, some won't. Political Behavior, 18(1), Phillips, A. (1995). The politics of presence. Oxford: Oxford University Press (SJEKK) Shore, J. (2014) How Welfare States Shape Participatory Patterns. In Kumlin, S., & Stadelmann-Steffen (Eds.), How Welfare States Shape the Democratic Public: Policy Feedback, Participation, Voting, and Attitudes. Cheltenham: Edward Elgar Publishing, Sartori, G. (1968). Representational systems. International Encyclopedia of the social sciences, 13, Schattschneider, E. E. (1960). The Semisovereign People: A Realist s View of Democracy in America. New York: Holt, Reinhart, and Winston. 17

18 Verba, S., & Nie, N. H. (1972). Participation in America. Political democracy and social equality. New York: Harper & Row. Verba, S., Schlozman, K. L., & Brady, H. E. (1995). Voice and equality: Civic voluntarism in American politics. Cambridge, MA: Harvard University Press. 18