Decentralization of Conflict?

Transcription

1 DRAFT Decentralization of Conflict? How different degrees of decisionmaking decentralization can cause variation in level of non-institutionalized political participation Til gruppedeltakerne! Bakgrunnen for dette paperet er et forsøk på å forklare noe av variasjonen i omfanget av ikke-institusjonalisert politisk deltakelse mellom land. Det er fremdeles på et veldig tidlig stadium, og jeg er ute etter kommentarer på om det i det hele tatt er et fruktbart prosjekt å fortsette på, litteraturtips og forslag på relevante operasjonaliseringer av desentralisering. Slik paperet er nå, inneholder det ikke data på den uavhengige variabelen og derfor heller ikke analyser. Ser frem til å treffe dere i Gøteborg! Eva Kvelland Nordiske Kommuneforskerkonferansen Gøteborg, november 2011 Eva Kvelland Universitetet i Agder eva.kvelland@uia.no 1

2 INTRODUCTION In addition to the formal democratic institutions, such as voting, representation and elections, the independent civil society's ability to interact with the political institutions and political processes, an essential part of all democracies. This interaction is a part of the way society is organized, and therefore at significance for the way political systems work. There are different modes of political participation. Various forms of participation differ in their political implications, and in the factors that stimulate individuals to act (Verba, Schlozman, & Brady, 1995). The literature argues that citizens are still supportive of democracy, but have become more critical to how it is functioning. That can be interpreted at as one of the explanations for why the total level of political participation doesn t change much, but are being transformed from the institutionalized to more non-institutionalized forms of action that aims at affecting political processes and decisionmaking from outside the formal structures (Norris, 1999, 2002). Politics takes place within and between political systems that contains both a set of independent parts and boundaries towards the outside environment (Almond, Powell, Strøm, & Dalton, 2004). Both within the formal institutions and different constituent tiers of a political system, and between the institutions and the outside environment, the distribution of power is a vital element in the political decisionmaking processes. It is in that context that features of political systems in this article will posed to have an effect on the level of non-institutionalized political participation across countries. The main research question can be placed in two different debates: The power distribution between different constitutional tiers of a country, and the different forms and levels of political participation. Empirical documentation across nations represents a major advance in the latter field, and has since the 1960s been extended to include the growth of non-institutionalized political participation, and thereby forms an important theoretical foundation of participation research (Dalton & Klingemann, 2007; Norris, 2002). Even though there is a lack of longitudinal data on the aggregate of activities from outside the traditional channels, there are indications saying that the frequency is increasing rather than decreasing (Rucht, 2007, p. 713). The overall increasing aggregated level of participation through for instance demonstrations, petitions and boycotts varies within the European countries (ESS, 2008; WVS, 2005). Once there is significant variation in level of non-institutionalized political participation across countries, it is interesting to study why this is. While valuable and insightful on 2

3 individual level determinants, research has to a lesser extent provided contextual explanations for the significant variation in unconventional political participation across countries. Therefore, this article tries to shed light on the possibility for an institutional explanation for the variation in non-institutionalized political participation. When aggregated, this becomes a feature on state level and it is because of that relevant to examine whether the variation has a connection to the different political systems. Comparing different political systems can be done on many dimensions. Knowing that political participation aims at affecting politics and policy, we also know that their participation is directed at affecting power holders. Still, power holders vary in the sense of how much and what kind of power they hold. The overall aim of this article is not to explain differences between certain countries, but rather to study how the degree of decentralization of decisionmaking power affects the level of non-institutionalized political. The causal mechanism and hypothesis that will be presented in the next section indicates a positive relationship between the two variables in the following research question: RQ: Is the level of non-institutionalized political participation affected by the degree of decisionmaking decentralization? Causal mechanisms Different features of a political system have been used to explain the actions citizens take toward states. Beginning with the work of Tocqueville, scholars have used the relative strength and weakness of political systems as possible explanatory factors. Using political opportunities as an argument, Tocqueville claimed that the difference in violent and peaceful protests could be rooted in a strong or weak state and civil society, when comparing respectively France and the United states (see Della Porta & Diani, 2006). The concept of political opportunities have undergone numerous operationalizations within the social movement literature both on micro and macro level - and we should anticipate that political opportunities are an important factor in the cost function of using protest activities as a mean to affect political processes and decision making. One set of definition of political opportunities are macro-openness, and can for instance relate to whether societies are democratic and the degree of decentralization as an institutional measure. In the literature there is a discussion on 3

4 what counts as open and closed institution. Della Porta (2006) presents three categories that is being argued to fall into the openness category: territorial decentralization; the functional dispersal of power; and the extent of power in the hands of the state. The general argument will be that the more access points of power that is available, the more the number of attempts to influence will increase. While this argument is made in the social movement literature, it points at attempts to influence a political power and can therefore be suitable as a theoretical analog for the puzzle in this article illustrated in the model displayed below. Figure 1: Causal model Independent variable Political system Dependent variable Political participation Degree of decisionmaking decentralization Degree of noninstitutionalized political participation The model can be specified the following empirical hypothesis: H: The higher degree of vertical decentralization of decisionmaking power within a state, the higher occurrence of non-institutionalized political participation. The hypothesis argues that the more available access-points of decisionmaking power, the more attempts to affect decisionmaking. Theoretically it builds on the political opportunity argument from the social movement literature as presented above originally dating back to Tocqueville s assertion about the English decentralized colonialism that lead to more self-initiatives among citizens claiming that the relative strength and weakness of a political system can explain much of the action citizens take toward states. One of the operationalizations within the political opportunity approach is decentralization as a part of what can lead to macro openness, arguing that the more power passed from the national government and down to lower tiers, the more open the political system will be to pressure from below (Della Porta & Diani, 2006). A high degree of vertical decentralization of power means that lower tiers of government have high level of freedom from central government regulation. This can be 4

5 viewed as giving a high level of discretion and flexibility to a local level tier. Focusing on power and power distribution this argument can refer to a conflict of interest; in this case between power holders and those without formal power. This would also imply that high level of power centralization will give low non-institutionalized political participation since protest activities largely occurs on local and regional levels. Hence, removing or preservation of power far away from those affected by the political decisions, also removes the risk of conflict. Turning it the other way around as the hypothesis does implies that decentralization of power also creates conflict. Hence, when power is placed in a constituent tier close to us, it enables us to act politically also from outside the formal political system when there is a conflict of interest. When power is removed from this tier and leaves local authorities left with standard procedures for decision making, there is literally nothing left to protest. We could also argue that high level of decentralization of power could lead to high legitimacy of democratic decision making due to the principle of subsidiarity and by that a lower level of non-institutionalized political participation. The argument for this possible hypothesis is that the closer individuals are to the constituent tiers that makes decisions that affects their life, the more they trust in the political system. As will be showed later in this article is that there is a significant positive correlation between level of non-institutionalized political participation and variables measuring trust in politicians and satisfaction with the way democracy works. High degree of local power concentration can also be argued to enable the local political institutions to establish more or less institutionalized arenas and mechanisms that involve people in the decisionmaking processes. Hence, they participate mainly within the formal political system, not from the outside. This can thereby hinder conflict. What can exclude this from being an opposite hypothesis from the previously introduced is that the latter argument removes the conflict and disagreement aspect, which is an essential component for the existence of protest activities in the first place. It would in that regard be possible to suggest an argument stating that the more you integrate people in the political processes, the more knowledge and engagement you create and the possibility for them to act out their disagreement through protest activities increases. 5

6 CONCEPTS Non-institutionalized forms of political participation While the 19th and 20th century was characterized by strong popular movements in close proximity to and hierarchical marching orders from political parties, we can see in the 21st century, a transition from elite mobilization directed to a more elite challenging form of participation, where the strong movements has been weakened, and the respect for authority is changing (Bjørklund & Saglie, 2009, p. 190). The role of governments is transforming, and the challenges they face have for the recent years become more and more dependent on societal actors to achieve their goals (Klijn, 2008, p. 506). Policy making is taking place not just within the hierarchical structures and organizations, but also through subsystems operating in parallel fashion, but more or less independently of one another. The systems consist of a large number of actors working within specific policy issues (Adam & Kriesi, 2007, p. 129). The term of these networks refers to linkages between governmental and non-governmental actors structured around shared interests in policymaking where policy emerge from the bargaining between the actors in the network (Rhodes, 2007, p. 1244). The word protest implies an expression of dissent to something or someone. It occurs in many forms and places, but the term protest politics usually means the public use of protest to seek influence on a political decision or process. The use of protest as a mean to seek influence is used by a wide range of groups especially those standing outside the formal institutions and political processes (Rucht, 2007). Different theories and perspectives can be used to analyze protest politics, and the different approaches have changed over the last century and decades. In the late nineteenth and early twenties the mass psychology (Gustave Le Bon) and collective behavior was dominating the perspectives on protest politics. During the 1960s, the view of protest politics as irrational behavior was replaced by a more rational and instrumental look at protesting as a mean for reaching a goal. Protesting was now seen more as Michael Lipsky (1968) defined it; a political resource that under certain conditions could be an effective mean (Gamson, 1975), and also through the rational choice theories focusing on the individuals choice to take part in or abstain from protest activities. The distinction between institutionalized and non-institutionalized political participation is one of the most common classification of political participation. They both share the aim to influence the political processes and decision making, but can be 6

7 distinguished by working respectively inside and outside the electoral channels and formal political systems (Marien, Hooghe, & Quintelier, 2010). Institutionalized political participation is closely related to the electoral process and involves officials who have been appointed as a result of the electoral process, like voting, being an official elected and party membership. The different forms of non-institutionalized political participation also aim at affecting political processes and decision makers, but in a more indirect way, through in example petitions, demonstrations and boycotts that are time limited, targeted and aimed at one single case with a low degree of organizational structure (Olsen & Sætren, 1980). As for the institutionalized political participation forms, it has a greater possibility to influence on political decisions and the composition of elected bodies, than the non-institutionalized forms. Hence, it is not possible to equate political participation and political influence (Martinussen, 1973), and the aim of this article is just to look at explanations for various levels of participation and will not consider whether it has influence or not. Decentralization The principle of local autonomy has regularly being expressed and debated in almost every Western country. Also in the European context the Council of Europe members entered a treaty in 1985 to protect local autonomy from central governments 1, and the treaty that suggested establishing a constitution for the European Union also contained recognition of local and regional autonomy. Most Western countries have for the last decades had general discussions on decentralization. Many of them have one or more times ended in a shift of fiscal, political and personnel power from the central government to a lower constituent tie. Federalism is often interpreted as a system placing fiscal, decisionmaking and personnel power at local and regional tiers. Still, when attempting to measure the degree of local autonomy, different measures of decentralization are more accurate, since federalism has been defined in many ways by scholars and a scholarly consensus is difficult to find. Federalism as William H. Riker (1975, p. 101) defines it, reads: Federalism is a political organization in which the activities of government are divided between regional governments and a central government in such a way that each kind of government has 1 European Charter of Local Self-government 7

8 some activities on which it makes final decisions. This is similar to the definition given by Robert Dahl. Another definition is the one preferred by Daniel J. Elazar: the fundamental distribution of power among multiple centers ( ), not the devolution of powers from a single center or down a pyramid. Using either of these two definitions, that Lijphart (1999) uses in his work on division of power, federalism as a concept does not theoretically imply a certain power distribution just a division of power between governmental tiers. What Lijphart also points at is that a federal state can be both centralized and decentralized, and that the same counts for unitary states. One example is Norway being unitary and decentralized, while Switzerland, in Lijphart s definition, is federal and decentralized. One of Daniel Treisman s six conceptions of decentralization is what he name decisionmaking decentralization, focusing on which tier decides what, and relates the following definition to the concept: If authority to decide all questions belongs to the central government, the system is maximally centralized in this sense; if the lowest-tier governments have all decisionmaking rights, it is maximally decentralized. (Treisman, 2002, pp. 8-9) Different types of decentralization have interested political scholars in the past, but the concept has not been very accurately defined during the years of research on the division of power between levels of the state. However, some approaches have been suggested to present a more systematic way of thinking about the vertical power structure. Scholars have presented different approaches to the concepts of decentralization and local autonomy, and many of them have related the two concepts causally to each other, saying that the more power you decentralize, the more local autonomy you get (Fleurke & Willemse, 2004). This assumption has often been used in a strong proximity to a normative aspect stating that local autonomy is a value of absolute good. Hence, the difficult concept of decentralization are being perceived as a mean to achieve local autonomy and by that a value in its own right, as James W. Fesler noted in his review on the concept of decentralization (1965, p. 538). Decentralization occurs in different political settings and therefore for various reasons. It can be a result of a political imperative; a need to improve and promote local and regional development; or as an incentive to preserve and strengthen local decisionmaking power. It is in the latter 8

9 setting that an increasing level of power concentration is being seen as an increasing level of democracy. The question that maintains is whether the different decentralization reforms have lived up to its expectations, and whether the commonly perception of a positive causal relationship between decentralization and local autonomy tests to be true. Although we can construct a theoretical distinction between local autonomy and local democracy, it is empirically a close proximity between the two concepts. It is in this intersection that decentralization can be posed to have an effect on the level of non-institutionalized forms of political participation. Degree of decentralization is difficult to both measure and compare across countries, especially when defined as decisionmaking power. Many have attempted to give a more precise definition of the concept, but are ending up with a rather unclear and uncertain concept, with lots of gradations and nuances. Therefore, when selecting decentralization as the cause variable in this article, the operationalization will be an important part. DATA AND MEASUREMENT Contextual data describes the environment in which individuals reside and behave (Johnson, Phillips Shively, & Stein, 2002, p. 220), and can be macro data on such as demography, economy and political systems. The individual behavior and attitudes measured through value surveys, as the European Social Survey (ESS) 2 are often contingent upon the social and political environment, and contextual data can therefore be at great use to try explain what external factors affects different forms of political behavior also on an aggregate level. The availability of contextual data have improved over the last couple of decades (Rydland, Arnesen, & Østensen, 2008, p. 2), and even though we now have massive access to data, there will always be questions and concerns about the quality and comparability 3, and there are still inaccuracies or errors in widely used datasets (Herrera & Kapur 2007). The selection of data has to be done out of some basic criteria but also with the possibility of missing important data that are available within the total universe of databases. While governments collect a great amount of data, they generally do not collect much on aspects of political systems, and 2 The European Social Survey (the ESS) is a multi-country survey covering over 30 nations. Its first round was conducted in 2002/2003, while the fourth was fielded in 2008/2009. Source: 3 See Rydland, et al., (2008) for a discussion on quality and comparability of contextual statistics. 9

10 they are therefore usually compiled by non-governmental organizations and individual research projects (Rydland, et al., 2008). When choosing data on both the independent and dependent variable it is important to be mindful to questions of data reliability in regard to the possible problems of cross country comparability. Since I lack longitudinal comparative panel data, the hypotheses needs to be tested by a combination of cross-sectional data and theoretical reasoning. Quantitative data on political institutions are often portrayed as formed by objective information. Still, it s important when using this kind of data to examine how the institutions are defined, operationalized and measured, in order to reveal a possible degree of subjective judgment and to find out what they really measure. Dependent variable: Non-institutionalized political participation The dependent variable is the level of non-institutionalized political participation. This variable will be operationalized by using data from the European Social Survey (ESS). ESS presents different measurements of political participation, and a selection of variables that fit the definition of being non-institutionalized political participation must be selected. There are at first glance seven variables needed to be considered in that context (all from the fourth ESS round (2008)) 4. For conceptual clarity, some forms of non-institutionalized political participation are left out. This is mainly to distinguish political interest from political action. Discussing politics with friends and family is an example of what can both be a measure of interest and a measure of action. Data reduction is needed also within the seven variables in order to find those that best represent the same underlying pattern of non-institutionalized participation. By conducting a factor analysis among the variables we can see a pattern that divides them in one component that describes non-institutionalized activities totally from the outside of the political system, and one component that are more leaning towards working inside or directly at the traditional structures of formal political processes. 4 ESS-data fra 2010-undersøkelsen er nylig frigitt. Den er foreløpig ikke benyttet her, siden denne delen av paperet ble skrevet før den nye dataen kom. 10

11 Table 1: Patterns of political participation Component 1 2 Signed petition last 12 months,751,150 Boycotted certain products last 12 months,742,002 Taken part in lawful public demonstration last 12 months,558,215 Worked in political party or action group last 12 months,011,787 Contacted politician or government official last 12 months,094,697 Worked in another organization or association last 12 months,335,542 Worn or displayed campaign badge/sticker last 12 months,420,467 Principal Component Analysis. Rotated Varimax in 3 iterations. Total variance explained: 48,7 % A widely used cutoff criteria in factor analysis is to exclude variables that scores higher than 0,30 on more than one component (Eikemo & Clausen, 2007, p. 231). If we were to extract the two variables fitting this criteria, the results would still show the same two components, with three variables on the first and two left on the other one. To underline the pattern of the two components the variable voted last national election was put into the matrix. It showed not surprisingly a high score on component two, and strengthens the understanding of the two components as indicators of political action on each side of a continuum measuring how institutionalized they are. The variable worn or displayed campaign badge/sticker last 12 months scores equally high on both components. In addition to that it needs to be left out by two other reasons. First of all it is hard to single out as a distinct form of participation since it is an often used action within many of the other forms described above. If you for instance attend a public demonstration, wearing a sticker of some kind is not unusual, and the same can be said when taking part in more conventional and institutionalized forms of participation. Second, it s not clear what wearing a badge or sticker means in this context. Wearing a badge or a sticker showing a protest of some kind could fall under the definition of being a non-institutionalized form of political participation, but also other kinds of badge display can have fallen under that category in the dataset. Wearing a badge showing where you work or displaying the symbol of your political party are two examples that shows that displaying a symbol doesn t need to imply activity, rather being a passive action. 11

12 All the three variables that scores high on component 1, scores low on component 2, and can be seen as a part of an expression of the total level of non-institutionalized forms of political participation. They can all be questioned in relates to what they really measure, in the same way as wearing a badge or sticker was in the previous section. When respondents answer yes on whether they have signed a petition, we don t know if that equals just paper signing or also signing an online petition. For the last years it would also be possible to ask whether liking a page supporting a political case on such as Facebook can be defined as signing a petition. The same can be said about the variable measuring boycott. What is the respondent s definition of boycotting? One might answer that it is to not buy a product for two months, while others adds a more active and longer committed political engagement to a boycott. We also know that there are possible high overlaps between the three selected variables. An example can be that you boycott a product, sign an online petition supporting your boycott and then attend a public demonstration following up your boycott. The strongest measure of noninstitutionalized political participation in this selection would obviously be the variable measuring the attendance in public demonstrations. It can clearly be interpreted as a physical appearance and participation in a public visible demonstration supporting a case. Knowing this, the three variables will in the following be used as operationalization of the dependent variable as indicators of level of non-institutionalized political participation. Table 2: Participation within all countries Activity % Signed petition last 12 months 18,1 Taken part in lawful public demonstration last 12 months 5,8 Boycotted certain products last 12 months 12,8 Source: ESS 2008, all 29 countries. Entries are percentages of respondents indicating that they have participated in this act during the past 12 months. 12

13 These three variables are all single indicators for non-institutionalized political participation, but it is also very likely that they correlate with each other. By running a bivariate correlation test, we find that all the three variables in table two correlate positive, significant on a 0.01 level. It is reasonable to assume that protesting a case (in example the adoption of a certain law) can contain two or more of these three actions at the same time, as means to reach a specific goal. Hence, in addition to be embedded in a common theoretical ground, being examples of non-institutionalized forms of political action, they also empirically correlate. The total within all countries in the ESS survey shows that the variable on signing a petition scores the highest with 18,1 % in the dataset from 2008, while the variables on taken part in public demonstration and boycotted certain products scores respectively 5,8 % and 12,8 %. Looking at each country separate we see that some of the western countries scores significant higher on one or more of the mentioned variables than the total within all countries. The Scandinavian countries scores overall high on all the mentioned variables. As do countries like France and Switzerland. Finland scores significantly lower on the variable measuring participation in public demonstration, then on the other two variables. While Sweden by large are the country with the highest numbers on both boycotting and signing petitions, they score just above average on participating in demonstrations. Most of the eastern European countries come out with numbers lower than the total within all countries on all the three protest variables. Croatia are maybe the country differing the most from the overall trend in this survey data, and scores significantly above average on all three variables. These numbers shows that there is a large variation within European countries, but also that participating in these kind of protest activities are significant in many countries. Looking at data from 2002 compared to 2008, there are small changes both in the total within all countries and in single countries isolated. The period of time is short, and it doesn t show a clear and significant direction. The figure displayed below shows an index of the three variables for all the 29 countries in the survey, ranging from high to low score on percentages participating in one or more protest activities for the last 12 months. 13

14 Figure 2: Index of non-institutionalized political participation (all countries). Index of non-institutionalized political participation Ukraine Romania Portugal Bulgaria Cyprus Turkey Poland Russian Federation Hungary Estonia Slovenia Latvia Israel Greece Czech Republic Slovakia Spain Croatia Netherlands Ireland Belgium United Kingdom Germany Denmark France Norway Finland Switzerland Sweden 7,4 7,5 8,2 9,1 10,5 10,6 10,9 11,2 11,8 12,3 12,6 13,7 14, ,2 22, ,2 28,8 31,6 33,3 46,1 46,8 47,4 47,5 47,6 47,7 48, Percent Source: ESS 2008 (all countries in the sample); percentages answering yes on participating in one or more of the three actions signed petition, taken part in lawful public demonstration and boycotted certain products last 12 months. Richard Rose argues that political behavior occurs within a three-dimensional space, as a function of when, where and who (Rose, 2007). Countries differ in their political institutions and in the timing of modernization, and it s therefor significantly important to take time and space into account when studying political behavior across countries. The study of political behavior often relates empirically to modern democratic societies (Rose, 2007, p. 283). The time factor when studying political participation and political 14

15 institutions are closely related to the democratization of the countries. Looking at the ESS survey data the countries that scores high on the index of the three variables, are by large the countries with the oldest democracies. The newer democracies i.e. many of the eastern European countries that have developed their democracies during the last 20 years, shows overall a score on the lower part of the scale. If removing all countries that have less than 70 years of democracy (between 1930 and 2000), it leaves 10 countries in the original group of 29. The countries are: Ireland, United Kingdom, Belgium, Netherlands, France, Sweden, Finland, Norway, Switzerland and Denmark. The aim of this article is to reveal possible explanations for the variation in level of noninstitutional political participation between countries in Europe. If the 10 countries were equally high on all the mentioned variables measuring this kind of participation, the selection would not be suitable to the aim of this article. However, these 10 countries varies quite a lot on all the three variables for itself, with a range of respectively 12,1; 24,2; and 27,7 percentage points. Extracting these 10 countries from the index in figure 2 gives overall the same picture. Figure 3: Index of non-institutionalized political participation (10 countries). Index of non-institutionalized political participation Netherlands Ireland Belgium United Kingdom Denmark France Norway Finland Switzerland Sweden 28,8 31,6 33,3 46,1 47,4 47,5 47,6 47,7 48, Percent Source: ESS 2008 (the 10 oldest democracies in the sample); percentages answering yes on participating in one or more of the three actions signed petition, taken part in lawful public demonstration and boycotted certain products last 12 months. 15

16 Individual aggregated data from the ESS survey also shows that these ten countries all can be placed on the upper part of the lists ranking the 28 original countries by means on the two variables trust in politicians and how satisfied with the way democracy works in country. This goes especially for the trust variable where 9 out of the 10 countries are to be found among the 11 top countries. The four Nordic countries Norway, Denmark, Sweden and Finland - are all among the top 6 on the two variables, with Denmark first on both. The variation in mean among the ten countries on these two variables exists, but most of them are gathered within a short range. Looking at each variable, the countries scores high on the satisfied with democracy variable, but centers just around the middle on the scale measuring trust. By selecting these 10 countries, we can then say that in addition to being long lasting democracies, the politicians are among the most trusted in Europe - but average on the low-high scale - and the democratic system is largely accepted by the people. Independent variable: Decentralization The use of indicators of decentralization as features of a political system can result in rather accidental conclusions, where you can completely change the result with a simple change of indicator (Fesler, 1965). Still, if you need indicators that can be compared across countries and in different contexts to measure decentralization as a feature of a political system, it is to some extent inevitable to do so (Fleurke & Willemse, 2004, p. 529). When doing that, it is crucial not to interpret the results normatively or automatically as an indicator that alone can measure local autonomy. To measure local autonomy as a feature of political systems you must choose a specific conceptualization of it, and then operationalize them by the use of specific indicators that are measurable. As presented earlier in this article, it is the concept of decisionmaking decentralization that will be used as the main independent variable. I interpret decentralization as something static not dynamic. It is also considered to be continuous not dichotomous. Hence, we talk about the degree of decisionmaking decentralization in a given time in history as one characteristic of a complex and compound in this case governmental system. To be continued 16

17 Control variables As already shown, variables like years of democracy; trust in politicians and satisfaction with the way democracy works, has already functioned as empirical cut offs when selecting units of analysis. But the universe of factors that scholars have claimed to affect political participation is enormous. Therefore, I will control for some of the most commonly discussed causal variables, as for example level of urbanization as a basic resource mobilization argument. To be continued ANALYSIS To be continued CONCLUSION To be continued 17

18 REFERENCES Adam, S., & Kriesi, H. (2007). The network approach. In P. A. Sabatier (Ed.), Theories of the policy process (pp ). Colorado: Westview Press. Almond, G. A., Powell, J., G. Bingham, Strøm, K., & Dalton, R. J. (2004). Comparative politics: a theoretical framework. New York: Pearson/Longman. Bjørklund, T., & Saglie, J. (2009). Ensaksdeltakelse: om underskriftskampanjer og politikerkontakt. In J. Saglie (Ed.), Det nære demokratiet -lokalvalg og lokal deltakelse. Oslo: Abstrakt forlag AS. Dalton, R. J., & Klingemann, H.-D. (2007). Citizens and political behavior. In R. J. Dalton & H.-D. Klingemann (Eds.), The Oxford Handbook of Political Behavior. Oxford: Oxford University Press. Della Porta, D., & Diani, M. (2006). Social movements: an introduction. Malden, Mass.: Blackwell. Eikemo, T. A., & Clausen, T. H. (2007). Kvantitativ analyse med SPSS: en praktisk innføring i kvantitative analyseteknikker. Trondheim: Tapir akademisk forl. ESS. (2008). European Social Survey Round 4 Data (Data file edition 3.0. ed.): Norwegian Social Science Data Services, Norway - Data Archive and distributor of ESS data. Fesler, J. W. (1965). Approaches to the Understanding of Decentralization. The Journal of Politics, 27(03), Fleurke, F., & Willemse, R. (2004). Approaches to Decentralization and Local Autonomy: A Critical Appraisal. [Article]. Administrative Theory & Praxis (Administrative Theory & Praxis), 26(4), Gamson, W. A. (1975). The strategy of social protest. London. Herrera, Y. M., & Kapur, D. (2007). Improving data quality : actors, incentives, and capabilities. Political Analysis, 15(4), Johnson, M., Phillips Shively, W., & Stein, R. M. (2002). Contextual data and the study of elections and voting behavior: connecting individuals to environments. [Article]. Electoral Studies, 21(2), 219. Klijn, E.-H. (2008). Governance and Governance Networks in Europe. Public Management Review, 10(4), Lijphart, A. (1999). Patterns of democracy: government forms and performance in thirtysix countries. New Haven, Conn.: Yale University Press. 18

19 Lipsky, M. (1968). Protest as a Political Resource. The American Political Science Review, 62(4), Marien, S., Hooghe, M., & Quintelier, E. (2010). Inequalities in Non-institutionalised Forms of Political Participation: A Multi-level Analysis of 25 countries. [Article]. Political Studies, 58(1), Martinussen, W. (1973). Fjerndemokratiet: sosial ulikhet, politiske ressurser og politisk medvirkning i Norge. Oslo: Gyldendal. Norris, P. (1999). Critical citizens: global support for democratic government. New York: Oxford University Press. Norris, P. (2002). Democratic Phoenix: reinventing political activism. Cambridge: Cambridge University Press. Olsen, J. P., & Sætren, H. (1980). Aksjoner og demokrati. Bergen: Universitetsforlaget. Rhodes, R. A. W. (2007). Understanding Governance: Ten Years On. [Article]. Organization Studies ( ), 28(8), Riker, W. H. (1975). Federalism. In F. I. Greenstein & N. W. Polsby (Eds.), Handbook of Political Scence, 5: Governmental Institutions and Processes. Addison-Wesley: Reading, Mass. Rose, R. (2007). Perspectives on Political Behavior in Time and Space. In R. J. Dalton & H.-D. Klingemann (Eds.), The Oxford Handbook of Political Behavior. Oxford: Oxford University Press. Rucht, D. (2007). The spread of protest politics. In R. J. Dalton & H.-D. Klingemann (Eds.), The Oxford Hanbook of political behavior. Oxford: Oxford University Press. Rydland, L. T., Arnesen, S., & Østensen, Å. G. (2008). Contextual data for the European Social Survey. An overview and assessment of extant sources. Bergen: Norsk samfunnsvitenskapelig datatjeneste AS (NSD). Treisman, D. (2002). Defining and Measuring Decentralisation: A Global Perspective. UCLA/unpublished manuscript. Verba, S., Schlozman, K. L., & Brady, H. E. (1995). Voice and equality: civic voluntarism in American politics. Cambridge, Mass.: Harvard University Press. WVS. (2005). World Values Survey. Stockholm: The World Values Survey Association. 19