Citizens preferences for Stealth democracy, responsiveness and direct democracy in Western Europe 1 :

Transcription

1 Citizens preferences for Stealth democracy, responsiveness and direct democracy in Western Europe 1 : Comparing evidence across four countries Jose Luis Fernandez 2 Spanish National Research Council Institute of Advanced Social Studies. jlfernandez@iesa.csic.es DRAFT VERSION. ENGLISH UNCORRECTED. DO NOT QUOTE. Abstract Since its publication in 2002, the arguments raised in Stealth Democracy by Hibbing and Theiss-Morse have been widely discussed from several angles and different countries. Its groundbreaking, and even provocative, nature meant having to review the bases on which the debate on citizens preferences for political processes of decisionmaking was firmly established. Stealth Democracy showed that at least in the United States- this demand or support for participatory practices were very far from universal. Where do citizen preferences for democratic processes come from? The blooming of this specific public opinion part of the debate incorporates the risk of becoming a new separate table (Almond, 1988), that creates its own isolate debates. That is why, one of the main goals of this paper is to enrich this literature about how public opinion wants political decisions to be taken (process preferences) with the incorporation of the analysis of an important characteristic of democracy, as is the case of responsiveness. Our analysis relies on the 6th wave of European Social Survey 2012 (module on Democracy) and on four national surveys (Finland, the Netherlands, Spain and the United Kingdom) where attempts to replicate the Stealth Democracy findings have already been carried out. The goal of this paper is to move forward on the understanding of the models of democracy making emphasis on those nuances which reflect a certain amount of internal diversity in both models, the participatory and elite-technocratic ones. The main findings presented in this paper show that when analyzing the support to responsiveness and to direct democracy, two different groups -in terms of size and qualities- of core supporters emerge. On the other hand, when focusing on the explanatory social and political bases for supporting to business people or for supporting to non-elected experts, results show that ideology plays an important role. 1 Paper presented at ECPR General Conference, August, Montreal. Section: Democratic Innovation Research: Bridging Approaches, Theories, and Methods for the Study of Participation and Deliberation; Panel: Stealth Democracy? Comparing Conceptions of Democracy across Countries and Social Groups. 2 I am greatful to all comments and contributions that Joan Font and Jose Luis Cañadas made for the preparation of this paper. Of course, mistakes fall under my own responsibility.

2 1. Introduction The literature about models of democracy is one of the central fields in political science. From normative theory discussing features that different democratic models have included (Budge, 1996; Held, 2006) to common packages of institutions that reflect different democratic logics (Lijphart, 1999), this discussion has covered a rich field of theoretical and empirical domains. Over the last decades, this normative and institutional centred debate has been enriched by the incorporation of a public opinion perspective: How do citizens want our democracies to work? Which democratic models are distinguished by citizens and why do they prefer one or another? The blooming of this specific public opinion part of the debate incorporates the risk of becoming a new separate table (Almond, 1988), that creates its own isolate debates. This could be the case even if many of its core discussions (e.g., the role of citizens in democratic systems) are quite similar to others developing in neighboring subjects, but grow without having much mutual influence with them. Thus, one of the main goals of this paper is to enrich this literature about how public opinion wants political decisions to be taken (process preferences) with the incorporation of the analysis of an important characteristic of democracy, as is the case of responsiveness. The goal of this paper is twofold. First, from a rather theoretical viewpoint, the aim is to move forward on the understanding of the models of democracy making emphasis on those nuances which reflect a certain amount of internal diversity in both models, the participatory and elite-technocratic ones. Previous researches (Font and Fernandez, 2014) have already hinted the need to make progress in this direction: We know a few things about citizens like and dislike about representation, less about their participatory preferences (beyond referendum) and really few about which actors, form, processes and values would like those citizens that choose to have more expert government (2014:19). In other words, when comparing (within the participatory model) support to direct democracy and support to responsiveness, may different patterns be observed? And when comparing (within the expert-elite model) support to business people and support to nonelected experts, are the same kind of patterns found? Second, from a rather empirical and methodological perspective, this paper attempts to connect and compare most of the empirical evidence accumulated until now at the European level around the academic field of citizens preferences for political decision making processes, and specifically, around the concept of Stealth Democracy. The analysis relies on the data provided by the 6th wave of the European Social Survey (ESS), which included an interesting module on understandings and evaluations of democracy. Since the research aims to touch on the new European data with previously existing national surveys on process preferences and more specifically in what has been termed as stealth democracy model, we mainly focus on four European countries (Finland, the Netherlands, Spain Stealth Democracy and the United Kingdom) where this kind of national surveys exist. In addition, these countries represent a unique opportunity because of their diversity in terms of political and organizational cultures. The paper proceeds as follows. First, we develop our theoretical arguments and present some hypothesis. Second, the research strategy, data and variables are discussed. We then present the results of our first research question: how support to responsiveness and to direct democracy relates to each other and what factors explain these different supports. Next, we present some preliminary findings regarding with our second research question: do successful business people and not elected experts have the same supporters 2

3 within and across different countries? Lastly, a brief conclusion closes the paper summarizing the main ideas and the future avenues. 2. Adding nuances to participatory and expert models of democracy. One of the common conclusions of the process preference literature is the confirmation that the classic normative debate between representation and participation 3 is also perceived by citizens. Most citizens perceive these two components of democracy, as having a certain trade-off between them: most democracies will have some representation and some participation, but each of them adopts a specific set of institutions and practices that will give priority to one or the other (Bengtsson, 2012; Coffé and Michels, 2014; Font, Wojcieszak and Navarro, 2015; Webb, 2013). On the other hand, when representation and participation are not presented as alternative models in the same bipolar question, but as part of a larger set of alternative models of governance the public opinion picture gets often more blurred. In fact, the explanatory factors of support to both representative and participatory modes of governance appear as quite similar in some of the accounts (Ceka and Magalhães, forthcoming). This result is not completely contradictory with all the process preferences research, especially when a third expert-stealth-elitist option is also considered. For example, Hibbing and Theiss- Morse (2002) perceived their Stealth model as an alternative to both, participation and representation. In fact, the representation and participation models would share the idea of government based on values and conflict, whereas any expert based model of democracy would be guided, in contrast, by the idea of a common good that should be pursued through efficient and impartial (value-free) strategies 4. The dialogue of this process preference literature with neighboring democratic debates has been mostly limited to some discussion with the deliberative-participatory literature (e.g., Neblo et al, 2010; Jacobs et al, 2009) or with part of the political behaviour research (e.g., Bengtsson and Christensen, 2014). However, the dialogue has been quite less common with other close thematic areas, such as the rich literature about responsiveness (Manza et al, 2002; Morales et al, 2015; Yvette and Sander 2015). The focus of this separate responsiveness table is partially distinct (when and why are citizens heard in the policy-making process? Should they be heard? Through which mechanisms or for which policies?), and has not fully addressed the question of which are the citizens that have stronger preferences in favour of responsiveness or the explanatory factors of these preferences. In any case, both tables share a central common question: which is the role of citizen preferences in a democracy and how and when should they be incorporated 5? In order to bridge these different sets of literature, previous research (Font and Fernandez, 2015) has put into question an idea which is implicit in most of the process preferences literature: the existence of a unique path that goes from an elite based representative democracy to a fully-fledged participatory democracy. In fact, this is not a completely new idea: there are good reasons to think that a model based on assemblies, on mini-publics or on referenda could be very different from each other in practice (Gronlund, Bachtiger and Setala 2014; Smith, 2009), but empirical evidence is divided 3 Through the paper we will use participation, participatory model or direct democracy as representing the same basic democratic model, where an active role for citizens in policy-making is central. 4 Font, Wojcieszak and Navarro (2015) test this hypothesis for the Spanish case and do not find empirical support to the idea of a main opposition between an expert model versus a (partially) common representative-participatory valued led model. 5 See Denters et al (2015) for another attempt to incorporate responsiveness into the discussion of the process preferences of citizens. 3

4 on whether citizens distinguish these differences in their preferences (Webb, 2013; Font, Wojcieszak and Navarro, 2015). Something similar would happen if we think in a model based on the crucial role of successful business people, independent experts, civil servants or recognized technocrats. We can expect to find a large variety of sceneries when focusing on each of these different actors related to the elite-expertise-technocratic model. From a point of view more related to the deliberative approach (Majdik and Keith, 2011), expertise may be seen as argument, not just knowledge. Under this assumption, we could understand that some citizens that demand more opportunities to participate, agreed with the idea of giving a central role to experts as long as accountability is provided, which seems to be more likely to occur, for example, in the case of civil servants than in central bankers. In fact, the nuances would be even larger according to the levels of government that we refer to: local, national and supranational. For example, Lequesne and Rivaud (2003) focus their attention on the Committees of Independent Experts at European level. These authors suggest that these mechanisms are the results of an incomplete parliamentarian system because official control authorities lacked both expertise and independence (2003:701). Nevertheless, they state that the new model of post parliamentary democracy based on not elected wise men and women involves a serious problem of fragile legitimacy and accountability. Indeed, the picture becomes even more blurry when analyzing in depth the participatory mechanisms at local level. For example, the debate about the tensions between citizen participation, the role of experts and politicians, and accountability practices has been developed under the hypothesis of cherry-picking (Smith, 2009:93). Do politicians and experts cherry-pick among the participatory proposals? The central idea is that while some proposals emerged from participatory mechanisms become policies (participatory budgeting, citizen councils, strategic plans and surveys, among others), other participatory proposals are rejected, abandoned or substantially modified without providing any explanation by politicians or civil servants The origins of support to direct democracy and responsiveness. In contrast with these previous discussions that distinguished between different forms of participation and expertise, we move a step back. What if Hibbing and Thiess- Morse (2002) were right and some citizens want to be heard without necessarily being involved? In our understanding, this would be the main factor distinguishing preferences for direct democracy or for governments that are responsive to citizen preferences. While the majority of citizens may consider that both characteristics are strongly desirable in any democracy, others may choose only direct democracy or only responsive governments, because some of them want to be heard without necessarily being involved (or without their neighbors being involved), while others want to experience their direct input into the democratic process. Thus, for example, supporters of Stealth Democracy could not be interested at all in an option that requires their active involvement, but support government decisions in accordance with citizens preferences. According to this second perspective, one could favor only responsiveness or only direct democracy as adequate strategies to expand citizens role 6. Therefore, our first research question is how do the preferences for responsiveness and direct democracy relate to each other? 6 These two different relationships have certain familiarity with the typology suggested by Kriesi and Ferrin (2014). Their proposal incorporates other democratic goals that include not only processes (as we do), but also outcomes. According to 4

5 Hypothesis 1: There are groups that support exclusively one of the two options (direct democracy / responsiveness) to enlarge citizens voices. In any case, if some kind of differential preferences for political processes exist, where do they come from? The relationship between these preferences and variables such as education, ideology or political trust have been previously examined (Anderson and Goodyear, 2010; Bengtsson and Mattila, 2009; Bengtsson, 2012, Coffé and Michels, 2014). However, we still know very little about how citizens build these preferences. Are they based on their own experience and satisfaction with representative democracy and/or with their participatory experiences? Are they based on values such as responsiveness, efficiency or participation or based on the expected outcomes of each type of political processes? The starting point of our explanation is the difference that direct citizen involvement introduces between responsiveness and direct democratic arrangements. While both of them really rely on citizen preferences, we expect that direct democracy is really perceived as a system that gives more weight to citizens in policy making. First, citizens need to be more active and vote (and campaign) to make their voices heard. Second, the role of politicians is also different: they do not only ask questions (like in direct democracy), but a system somehow based on responsiveness also assumes that politicians need to read and interpret citizen preferences that have been gathered somehow (surveys, media, protest and civil society organizations 7 ). In sum, our second research question is how much these two types of variables (perception of how existing democratic institutions work, interest in politics and trust in citizens as decisionmakers) explain the preference for more or less responsiveness and direct democracy. Thus, we claim that two main types of explanations are important. First, the perception of the political actors and processes. Since a system based on direct democratic procedure would imply a larger break with the existing arrangements, a more positive perception of their performance should make citizens less inclined to support a more important break with current practices. In fact, this idea is quite close to what was been called the dissatisfaction hypothesis, claiming that larger support to more participatory practices should be concentrated on those sectors with have a more negative perception of how representative democracy works (Shuck and de Vreese, 2015; Webb, 2013). At the same time our idea goes beyond pure dissatisfaction: the central argument is that the main driver of these preferences is our assessment of how different democratic arrangements and procedures work. Hypothesis 1.1: We expect to find an explanatory role of our two main independent sets of variables. Hypothesis 1.2: Perception of how different democratic practices constitute the most important driving force explaining support to responsiveness and direct democracy. Hypothesis 1.3: The more negative perception of how democracy works the more support to direct democracy. them, many citizens could be divided in a less versus more demanding scale (from uncommitted to complete democrats ), while others would choose to incorporate only certain democratic characteristics as a priority. 7 The different risks and biases that responsiveness may introduce in the perception of citizens voices depending on how theses voices are captured is discussed in Grimes and Esaiasson (2014). 5

6 Second, since a larger reliance in direct democratic practices would imply to provide a more explicit role to citizens in the policy making process, we claim that both, trust in other fellow citizens as well as desire to be a participant would be the second large explanatory factors of these preferences. Perceiving other fellow citizens as appropriate centers of policy-making should be an important part of these perceptions, as it has been shown in the Spanish case for general desire for more participatory practices (Navarro, 2011; Del Rio, Navarro and Font, 2015). On the other hand, it is not only a general perception that citizens can be valuable decision-makers that matters, but also desire to be a participant in the process. This idea has also a clear connection to two previous hypotheses tested in the literature: one, the idea of citizens being reluctant to participatory practices due to their own aversion to deal with political conflict 8 and two, the cognitive mobilization hypothesis, relating more skills and abilities to participate with a stronger desire for a system that gives a central role to citizen intervention (Donovan and Karp, 2006; Shuck and de Vreese, 2015). Hypothesis 1.4: The more cognitive mobilization the more the more support to direct democracy The origins of support to business people and experts as policy makers. The majority of researches focus on trying to explain support Stealth Democracy preferences have addressed the following potential explanatory variables: ideology (Bengtsson and Matilla, 2009); education (Coffé and Michels, 2014); political interest and efficacy and partisan identification (Gangl, 2015; Webb, 2013) and political trust and types of personalities (Font and Alarcón, ). Nevertheless, in this paper, the attention is focused on the role of ideology to explore possible different supporters within Stealth Democracy model. The rest of variables are mainly considered as control variables, which exploratory findings will be useful to guide futures analysis. As mentioned above, the literature in this respect seems to conclude that ideology plays a clear role to understand citizens preferences for political decision-making processes. In the case of Spain, research based on two surveys 9 shows that having a left-wing ideology is an important factor to understand lack of support to Stealth Democracy model (Font and Fernandez, 2014). Indeed, Spanish and Finish right-wing voters show a significant positive effect in their support to Stealth Democracy. In sum, and contrary to the first idea of Hibbing and Theiss-Morse that support to stealth democracy should be completely independent of left and right, research developed in Europe suggests that ideology would be the main factor to support stealth or participatory options, as an alternatives to representation model. Hypothesis 2: A clear and consistent relation between left-right ideology and support to Stealth Democracy exists when comparing the core supporters and core opponents of this model across four Western European countries. However, if such were the case, we still know very little about whether citizens clearly distinguish between business people and independent experts. Do citizens draw a distinction between non-partisan institution such as the European Central Bank, a successful businessman like Ross Perot or a politician with remarkable technical competence and limited partisan attachment like Monti. To be able to answer all these questions it would be necessary to carry out a specific survey focuses on the different 8 Two very different versions of the idea in Hibbing and Theiss-Morse (2002) and in Eliasoph (1998) 9 One national survey before 15M movement and one local survey two months after 15M. 6

7 social and political actors related to this model as well as with the others models. In this line, recent research highlights (Del Río, 2014) the importance of analyzing citizens opinions about the technical and moral qualities of citizens, politicians and businessexperts-technocrats. Nevertheless, we first need to know whether ideology, and other features, varies significantly and constantly between supporters of business and expertise. Hypothesis 2.1: Core supporters to business people and core supporters to non-elected experts differ, across countries, in their ideological selfplacement. The next section explains the research strategy we follow to check how much support these hypotheses receive in the ESS. 3. Methods Research strategy Since the goal of this research is to analyze together the set of questions addressed in the 6th wave of the European Social Survey in dialogue with the available databases which have more in depth measurements of process preferences, I focus on four European countries (Finland, the Netherlands, Spain and the United Kingdom) where attempts to replicate parts of the Stealth Democracy thesis have already been carried out. This selection of countries also represents a good representation of political diversity from both an institutional and a cultural perspective. Thus, I have from the clearest presence of majority politics in Europe (UK) to one of the countries better representing consociational practices (the Netherlands) and from countries with very high levels of political trust (Finland) to some of the lowest (Spain). This diversity also appears in our two dependent variables. On the one hand, the Netherlands has the lowest level of support to direct democracy among the 29 countries included by the ESS, whereas Spain has one of the largest levels of support. On the other hand, the statements included in the Stealth Democracy Index receive some different support among the four countries analyzed. It is important to remember that the diffusion of democratic ideals and practices has followed different trajectories in each country. This is especially clear in Eastern Europe, where the incorporation of new democratic practices was the result of the break with their communist past or in Germany where this diffusion went mostly from East to West following the same events (Vetter, 2009). In Spain, for example, the use of referenda during the dictatorship resulted in a strong left reluctance to them during the early democratic period and the 2015 British perception of referenda is probably very strongly based on each citizens perception of the two referenda that have British politics in the recent period: the European and Scottish ones. Possibly, a similar story could be explained regarding the idea of responsiveness. For example, in a country like the Netherlands, with a sharp contrast between a long tradition of consociational politics and the more recent presence of a strong anti-immigration party, the public debate about following public preferences very closely may be sharper than in other countries Data. European Social Survey 7

8 The sixth round of the ESS represents an interesting material to explore these questions. In this occasion, the ESS develops a module on issues related to Europeans understandings and evaluations of democracy across twenty-nine countries 10. The complete set of variables of the democracy module covers a large number of concepts about democratic processes and practices such as rule of law, competition, participation, accountability and equality, among others. The dependent variable consists of two categories: exclusive core supporters of referendum 11 (value 1) and exclusive core supporters of responsiveness 12 (value 0). An exclusive core supporter of referendum has been empirically defined as (see table 1 and table 2 below) a person who considers very important for democracy the practice of referendum and only little or medium importance having a responsive government. To define an exclusive core supporter of responsiveness we use the same logic. For explanatory analysis, binary logistic regression, several independent variables often used in previous research in this field are incorporated. Following a forthcoming paper that uses the same data, we select some variables and codings used in Ceka and Magalhães. Socioeconomic information is captured through the following three dummies variables: income, (value 1 includes those individuals not included 13 in the highest quintile), relative education level, (value 1 means having an education level below the mean of their cohort in the respective countries) and gender (values 1 means male). In this socioeconomic set are also included age and ideology (self-placement leftright) in a 0-10 scale 14, both measured as continuous variables. All these variables work in the analysis as control variables (see table 1 in appendix). The evaluation of democratic performance includes four variables, two capturing support for the performance of representative institutions and two capturing the perception about the complementary processes addressed in this paper. First, political trust 15, measured as the mean score 16 of the trust in national parliament, legal system, politicians, political parties and police. Second, satisfaction with the way democracy works in their countries, in a 0-10 scale. Evaluation of the frequent use of referenda is measured in a 0 10 scale, where value 0 means that citizens do not have the final say 10 N= The fieldwork period was from 8/2012 to 8/2013. More information: 11 Support to referendum was originally measured in a 0-10 scale. And still thinking generally rather than about [country], how important do you think it is for democracy in general that citizens have the final say on the most important political issues by voting on them directly in referendums? 12 Responsiveness was originally measured as a categorical variable and responses were coded as follows:(1) the government should change its planned policies in response to what most people think, (2) the government should sticks to its planned policies regardless of what most people think and (5) it depends on the circumstances. Respondents who chose 1 or 2 were asked in the following items about the level of importance of their respective responses. In order to have the same scale for all the variables, we created a new variable in a 0-10 scale in which the value 0 means that maintaining the planned policies regardless of public opinion is extremely important for democracy; the value 5 includes those who consider that sticking or changing planned policies is not important at all (voting 0 in question one or two) as well as, the option it depends on the circumstances in the original qualitative variable, as the authors of the module have considered that these three groups of respondents are giving a similar answer (opinion). The value 10 means that changing planned policies in response to what most people is extremely important for democracy. We are extremely thankful to Hans Peter Kriesi and Mónica Ferrín for sharing their coding. 13 Value 1 also includes those who responded refusal or don t know. 14 Don t answer, Don t know and Refusal were recoded as value 5 to avoid a large number of missing cases. We tried the alternative format suggested by Ceka and Magalhaes dividing between moderates and radicals but results did no change, not reaching statistical significance in any country. 15 Cronbach s α > 0.83 in the four countries. 16 Among those who answered at least three of the five items. Those answering two or less are considered missing cases. 8

9 by voting on referendums. Lastly, evaluation of responsiveness is a scale 0 10, which was created from a combination of two variables 17. The last set of variables is related to attitudes towards citizen participation and political behavior. Social trust 18, is included as a mean score of the main three items commonly used in the literature. Interest in politics, is a scale in which value 4 means not at all interested. Mass Media captures watching news about politics and it is measured in a 0-7 scale, in which value 7 means more than three hours per day. Lastly, we include a deliberation variable, a scale 0-10, which focuses on the idea of the importance of discussing politics among citizens before voting. National surveys on Stealth Democracy Since its publication in 2002, the arguments raised in Stealth Democracy by Hibbing and Theiss-Morse have been widely discussed from several angles and different countries. In this section we jointly analyze the available databases (see table 1) of four countries in order to combine the evidence and observe the similarities and differences in the patterns of how European citizens want that the political decisions to be taken. Table 1.National Stealth Democracy Surveys Description Spain Finland 19 United Kingdom The Netherlands Title of the study Preferences for political decisionmaking processes. Finnish National Election Study. British Participation Survey, 2011 Public Attitudes Towards Representative, Direct and Stealth Democracy Universe Aged 18 and older. Finnish people entitled to vote in the 2007 parliamentary elections, excluding the Åland Islands Electors living in the UK during Probability sample of households drawn from the population register by Statistics Netherlands. Aged 16 and older Principal investigator Joan Font Åsa Bengtsson et al. Paul Webb H. Coffé and A. Michels Collection date January-February 2011 March May July July 2011 December 2011 Mode of data collection: Face-to-face interviews Face-to-face interviews and a supplementary, selfadministered questionnaire Self-completion; Web-based survey Online panel. LISS Panel (online) Number of cases The variable aims to capture how close is the perceived status of responsive practices to the desired arrangement of each respondent. Thus, as explained in footnote 9, we start from their preference for governments being or not responsive to public opinion. The follow-up question (e.g. How often you think the government in [country] today sticks to its planned policies regardless of what most people think? ) captures its degree of satisfaction, so that a person who responded 0 in the stick evaluation option and another person who responded 0 in the change evaluation option, both are considered 0 in the new variable of evaluation of responsiveness. The same logic was used for the rest of the values. 18 Cronbach s α > 0.70 in the four countries. 19 Paloheimo, Heikki: Finnish National Election Study 2007 [computer file]. FSD2269, version 1.1 ( ). Helsinki: Taloustutkimus [data collection], The Political Participation and Modes of Democracy: Finland in a Comparative Perspective research group [producer]. Tampere: Finnish Social Science Data Archive [distributor],

10 The dependent variable has been created from the following two of the four items included in the Stealth Democracy Index because they have shown greater capacity to discriminate: leave political decisions to successful business people and leave political decisions to not elected experts. In the same way that we did for the creation of variable dependent utilized to analyze the European Social Survey, in this case we have identified groups internally homogenous and mutually exclusive, according to the answers given to the two items mentioned above. Following this reasoning, the dependent variable has four categories in each of the four countries: I) Core supporters of Stealth Democracy (those who have answered strongly agree (SA) or agree (A) to both questions (business people and not elected experts); II) Core supporters of successful business people (those who have answered SA/A to business item and strongly disagree (SD) or disagree (D) to experts item; III) Core supporters of not elected experts (those who have answered SA/A to experts item and SD/D to business item and, IV) Core opponents of Stealth Democracy (those who have answered SD/D ) to both questions items (see table 2). Next, we present the set of independent variables used for explanatory analysis 20, multinomial logistic regression. It is important to note that the analysis presented in this paper supposed a first exploratory approach. That is why I have included a large number of independent variables which appear in the most of the national surveys analyzed. Nevertheless, a similar set of independent variables were included by Amy Gangl (2007) in her research on citizens beliefs that government should run like business. Our independent variables address the following dimensions (see table 2 in appendix): Socioeconomic information is captured through the following three dummies variables: income, (value 1 includes those individuals not included 21 in the highest decile), education level, (value 1 includes those individuals that have attained at least vocational studies to highest) and gender (values 1 means male). In the same way that we did for the ESS we include in this set of variables age and ideology 22. In addition, we include an evaluation of the personal financial situation (1, better; 2; no change and 3, worse). A second set of variables seeks to capture the evaluation of democratic performance through the following two groups of items: the first one focusing on general aspects such as satisfaction with the way that democracy works (from 1, very satisfied to 4, not at all); political trust (from 1, strongly distrust to 5, strongly trust) and, country financial situation (1, better; 2 no change and 3, worse). The second group focuses on three dummy variables based on attitudes towards political processes and actors: support for referendum, partisan identification and support of citizens as policymaker. A third set of variables addresses the concept of political efficacy and influence through four dummies variables. First, internal political efficacy (1 means citizens think that politics are so complicated). Second, external political efficacy (1 includes those who think that politicians do care about the opinions of ordinary people. Third, internal political influence (1 means that citizens think they not have influence/say in politics) and interest in politics (1 means no interest). 20 Because of the different national surveys on stealth democracy (processes preferences) were not designed aiming to analyze them together, we preferred to recode most of the variables and transform them into dummies variables. 21 Value 1 also includes those who responded refusal or don t know. 22 Because ideology in the case of the United Kingdom was measured in a 1-7 scale we recoded the other three countries (measured originally in a scale 0-10 scale) as follows: 0-1, very left-wing; 2-3,fairly left-wing; 4,slightly left-centre; 5, centre; 6, slightly right-centre; 7-8, fairly right-wing; 9-10, very right-wing. 10

11 4. Results 4.1. How do the preferences for responsiveness and direct democracy relate to each other and where they are coming from? With this question in mind, our first step has been to create (in the four countries where we want to focus from now on) a similar three category variable for each of them: little or no importance for democracy includes values from 0 to 5; medium importance category, from 6 to 8 and, 23 very important, values 9 and 10 (see table 2). Table 2. High, medium and low support to direct democracy and responsiveness, by country Direct democracy Responsiveness Little or no importance for democracy 6,6 11,8 Spain Medium importance 24,2 25,1 Very important 69,2 63,1 Little or no importance for democracy 10,7 29,1 Finland Medium importance 46,5 30,6 Very important 42,9 40,3 The United Kingdom Little or no importance for democracy 13,1 19 Medium importance 36,4 32,3 Very important 50,5 48,8 The Netherlands Source: own elaboration Little or no importance for democracy 19,4 28,2 Medium importance 53,2 31,1 Very important 27,4 40,7 The next step was to compare the core supporters of each variable between them to see whether accumulative structure exists. That is, whether for example all citizens who consider direct democracy very important also consider that responsiveness is a central democratic trait. Results show that this accumulative structure does not exist. When we look at the relationship between direct democracy and responsiveness, two different groups of core supporters emerge, with some citizens considering that both are crucial characteristics, but many others choosing only one of them as more central. Table 3 shows the distribution of these groups, dividing the population from this point of view. Finland, Spain and the United Kingdom, in line with most of the ESS countries, show a similar pattern: there are two different groups of core supporters, two groups relatively similar in their size that choose that only one of these characteristics is a high priority for 23 Value 5 in responsiveness was considered as medium importance due to it is the option it depends on the circumstances in the filter variable. 11

12 them. On the contrary, the Netherlands is the only country in which the exclusive core supporters of responsiveness are clearly more numerous. Thus, the hypothesis 1 can be confirmed since direct democracy and responsiveness have some similar (in terms of quantity) but also some different core supporters. As a result, it makes sense to approach their explanatory analysis recognizing the distinctiveness of these two groups. Which are the characteristics that make some citizens exclusive supporters of one of these characteristics of democracy? This is the question we try to answer from now. Table 3. Exclusive core supporters of referendum and responsiveness Spain Finland United Kingdom The Netherlands ESS N % N % N % N % N % Core supporters responsiveness , , , , ,0 Core supporters referendum No core supporters , , , , , , , , , ,9 Total Source: own elaboration We use binary logistic regression to explore some of the potential explanatory variables of the pertaining to the exclusive club of supporters of only one of these two categories. Thus, we exclude the rest of the population (those who do not consider any of these two characteristics as centrally important for democracy, as well as those that consider that both of them are very important) to focus only on the differences between these two groups. In order to test our hypothesis 1.1, I include several models to see how the incorporation of new variables adds explanatory power: a basic model incorporating only the control variables (model I), a second model that adds the variables corresponding to the evaluations of the performance of different democratic arrangements (model II), a model that only includes control variables plus those measuring attitudes towards citizens participation (model III), as well as a final model that includes all of them (model IV) (see table 3 in appendix). Hypothesis 1.1 is confirmed in almost all cases. Only in Finland, citizen centered variables do not appear to add much explanatory power to the model, even if in the final model political interest has a significant effect in the expected direction. For all the remaining countries (and for Finland in the case of political variables), adding one of this sets of variables helps us to better understand the choice between direct democracy and responsiveness. As we expected in our hypothesis 1.2 political variables tend to add more to the explanation than those centered on citizens. Only in the UK, where the role of citizen variables is especially powerful and where the role of political variables is more limited different pattern is found. In general, the evaluation of the existence of referenda is a significant variable everywhere (except in Spain). The perception of responsiveness has 12

13 also a significant effect in Finland and Spain. For some of these variables the results are clearly in line with expectations (e.g., the evaluation of referendum, where those who think that they are already a powerful instrument are those most supportive of their democratic role). But the traditional variables that show the degree of political disaffection with representative democracy (political trust and satisfaction with democracy) have a less powerful effect than when they are simply used to explain support to referenda (hypothesis 1.3). Interest in politics is the variable from the citizens set that shows a more consistent effect: those more interested are also more clearly in favour of direct democracy (hypothesis 1.4), except in the country where supporters of direct democracy are a minority (the Netherlands), where the relationship does not reach statistical significance. Regarding control variables, education has the most surprising result, being the only variable that is significant in two countries in different directions. Whereas in Finland we find the pattern suggested by Anderson and Goodyear (2010), so that the most unconditional supporters of referendum are those citizens with lower educational degrees, the opposite is true in Spain Do successful business people and not elected experts have the same supporters within and across different countries? In the same way as performed in the former section, the four core supporters (and opponents) groups which compose the second dependent variable are compared between them (see table 4). Except in the United Kingdom, the core supporters of business option is the minority group. In fact, when focusing on Spain, we observe that 32 percent support the expertise option. This result is even more remarkable when comparing it to the support received by the business option. In the four countries citizens seem to differentiate between not elected experts and successful business people. Table 4. Distribution of core supporters and core opponents to Stealth Democracy. Spain Finland United Kingdom The Netherlands N % 24 N % N % N % Core supporters of Stealth Democracy Core supporters of successful business people Core supporters of not elected experts Core opponents of Stealth Democracy , , , , ,4 46 3, , , , , , , , ,1 Source: own elaboration 24 Percentages do not sum 100. The rest of respondents are those who answered to both questions don t know, refusal, etc. 13

14 Next, a multinomial logistic regression was performed in order to find out whether a difference pattern may be observed between those who support only one of the two options (business vs expertise). As mentioned above, the attention will be put on the effects of ideology and the rest of findings are considered as an exploratory approach (see table 4 in appendix). In line with previous research, it may be observed how ideology is clearly and consistently even when comparing between countries with a different political and organizational cultures. Thus, hypothesis 2 can be confirmed; in fact, ideology is the most powerful explanatory variable according to the model presented in table 5. Table 5. Ajustment of multinomial logistic regression models Spain Finland United Kingdom The Netherlands Sig. Sig. Sig. Sig. Age,627,012*,093~,000*** Sex,517,330,931,013* Incomes,037*,540,506,030* Education,000***,925,111,062~ Ideology,000***,000***,000***,000*** Personal Financial situation,928,012*,057~ Satisfaction democracy,302,650,000*** Political Trust,271,210,019*,293 Country Financial situation,002**,148,011* Political Party Identification,184,147,009** Referendum,701,211,231,619 Citizens should influence policy making,096~,797,577,000*** Interest in politics,648,187,473,460 Internal political efficacy,001**,752,097~ Political efficacy,202,001**,000*** Internal political influence,084~,709,000*** Nagelkerke,104,216,248,247 <,1 = ~; <,05 = *; <,01 = **;<,001 = *** The literature on process preferences seems to conclude that having a right-wing ideology is related to greater support to Stealth Democracy assumptions. Nevertheless, as it was developed in the former sections, it is still not clear to what extent this statement is totally consistent when focusing on the internal nuances of Steal Democracy model across four Western European countries. Results in Figure 1 follow the expectations suggested by hypothesis 2.1: core supporters to business people and core supporters to not elected experts show differences, across countries, in their ideological self-placement. In more detail, in both Finland and the Netherlands of expressing a left-wing ideology and to be a core supporter of Stealth Democracy at the same time is unlikely (P,05). However, in the cases of Spain and United Kingdom, although this tendency is maintained (specially in Spain), the probability is slightly higher. On the other hand, expressing a right-wing ideology implies a lower probability (P <,2) of belonging to the Stealth Democracy core opponents group with the exception of Finland, where expressing a very right-wing ideology still maintains high probability of being a core opponent. Although the largest differences are observed between core supporters of SD and core opponents of SD, other specific differences may be found, for example, an association between the expression of right-wing ideology and the support to leave political decisions to successful business people. This tendency is strongly observed in United Kingdom and the Netherlands. These findings, although moderated in the levels of 14

15 probability reached, gain more relevance when looking to the effects of ideology in the expertise option. In this case, the tendency is exactly the opposite. That is, the more leftwing ideology is expressed, the more support to leave political decisions to not elected experts. This tendency is strongly observed in Spain and the Netherlands. Figure 1. The effects of ideology on Stealth Democracy preferences. The shaded area in the graph illustrates standar desviation Source: own elaboration 15

16 5.Discussion From this set of preliminary analyses we can draw a few conclusions. The first goal of this paper has been to find out whether distinctive supporters groups exist within each of these modes of democracy: participatory and expertise. On the one hand, regarding the relationship between responsiveness and direct democracy: those supporting them as essential democratic characteristics are only partially the same people, but there are also particular clubs of supporters of these two different characteristics, with the Netherlands having a particularly large group of exclusive supporters of responsiveness, who are not so centrally interested in direct democratic practices. The different format of the responsiveness question in the questionnaire introduces some difficulties into this comparison of support, but the existence of these two different groups of supporters is quite clear and has important implications since it points in the direction suggested among others by Webb (2013) that there may be more than one way of moving towards more citizen centred politics and that not all of them would gather the same social supports. On the other hand, regarding the relationship between business and expertise options: the groups supporting exclusively one of them are asymmetric in their sizes. In fact, business supporters groups are especially small in Finland and in Spain. Nevertheless, the other side of the story is United Kingdom where the option represented by successful business people receives higher support. These different national stories are not totally contradictory with previous research, for example, Evans (2013) found in Australia an equivalent support to both options. The fact that this pattern is repeating in United Kingdom, Australia and United States, opens the door for further inquiries from a cross disciplinary and transcultural point of view. The second main goal of this paper has been precisely to add some nuances about the social and political bases of each different clubs. First, in order to explore the two distinctive supporters groups within the participatory model responsiveness vs direct democracy- we have explored the role played by two groups of variables, one dealing with the perceptions of political processes and institutions and one with the role of citizens in policy-making.the results have mostly confirmed our hypotheses, showing the relevance of both sets of variables, the dominance in most cases of the political variables and the existence and some similarities as well as some differences between countries. The existence of these relatively similar patterns (and their differences) is one of the issues that asks for further analysis incorporating more in-depth country knowledge to interpret the origins of these differences, which may be related to either the different sets of preferences (from a slightly larger support to direct democracy in Spain to a large preference for responsive governments in the Netherlands) as well as to the different stories and political meanings that these types of political processes have acquired in each of these countries. Second, we have tried to add knowledge about the social and political bases for supporting the least known model of democracy: stealth democracy. The results have confirmed that there are differences, at least ideological, between those who support the business or the expertise options. On the one hand, for some could seem obvious to find a positive association between right-wing ideology and support to business option. Nevertheless, as Gangl demonstrated (2007), the belief that there is a broad support to market mechanisms to political processes (populism market) is still a far too common feature among scholars and cultural critics (Sunstein 1999; Sandel 1996). On the contrary, she shows that conservatives and those who are identified as Republicans are expected to be more likely to evaluate the fairness and timeliness of the political process negatively than those who are identified as Liberals or Democrats (2007:666). Thus, the support to market mechanisms populism market- is very far to be 16

17 universal. On the other hand, we knew very little about where do preferences for nonelected -independent- experts came from? For some would be reasonable to expect that those opting for a central ideological position were the ones who more supported this option. In line with this view, while right-wing citizens would support the business option; central-wing citizens would support to independent experts and, left-wing citizens would feel closer to the citizen participatory model. In fact, it would not be surprising a close link between preferences for responsiveness and expertise. For example, last local elections in Spain (2015) have shown how many of those political parties (new left) that included citizen participation as their main sign of identity, have chosen non partisan and highly qualified citizens as candidates. One example of this is the Madrid City Lady Mayor, a former judge with a Twitter account where one can read: Govern is to listen. What do our comparative results add to previous knowledge? At first sight, they show some important differences with previous research about preferences for direct democracy and other forms of participatory arrangements. However, these differences are completely normal given the approach adopted here, which is based in trying to understand the different options between these four sectors of the public. Whereas most previous research tried to differentiate between two very opposite sectors of the population (e.g., supporters of representative versus participatory forms of democracy), the approach adopted here (including responsiveness into the picture of preferences for more citizen voice in democratic policy-making) has been to compare two groups which are probably much closer to each other in a global democratic mapping of preferences. This helps to explain, for example, that political trust or satisfaction with democracy are less important in these results than in previous research (Dalton, Bürklin and Drummond, 2001): Disaffected citizens are probably looking more intensely to improve democratic procedures, but their discontentment is not a central explanation of their preference for responsive governments vs a direct role of citizens. Even with these differences in the approach, the results confirm ideas that had already emerged in previous research: our preferences are based on what we know and on the more or less solid perceptions and judgments built based on these experiences (our set of political variables). Also, any move in the direction of more citizen role in democratic politics has to rely on how much we do like the idea of actively participating (Bowler, Donovan and Karp, 2007; Bengtsson and Christensen, 2014) and how much do we trust our fellow citizens as active policy-makers (Del Rio, Navarro and Font, 2015). References Almond, G (1988): Separate Tables: Schools and Sects in Political Science, PS: Political Science and Politics, Vol. 21, No. 4, pp Anderson, C. and Goodyear-Grant. E. (2010). Why are highly informed citizens sceptical of referenda? Electoral Studies. 29(2): Bengtsson, A. (2012). Citizens perceptions of political processes. A critical evaluation of preference consistency and survey items, Revista Internacional de Sociología, 70 (5). Bengtsson, A. and Mattila, M. (2009). Direct democracy and its critics: Support for direct democracy and stealth democracy in Finland. West European Politics, 32 (5), Bengtsson, Å- and Christensen, H. (2014) Ideals and Action: Do Citizens' Patterns of Political Participation Correspond to their Conceptions of Democracy?, Government and Opposition Bowler, S.; Donovan, T. and Karp, J. (2007). Enraged or Engaged? Preferences for Direct Citizen Participation in Affluent Democracies, Political Research Quarterly 60;

18 Budge, I. (1996): The new challenge of direct democracy. London: Polity Press. Ceka, B and Magalhaes, P (forthcoming): The meaning of democracy and its determinants, in Kriesi (ed): How Europeans view and evaluate democracy. Unpublished manuscript. Coffé, H. and Michels, A. (2014): Education and support for representative, direct and stealth democracy, Electoral Studies, doi:1o.1o16/electstud Dalton, R, Bürklin, W. and Drummond A. (2001). Public opinion and direct democracy, Journal of Democracy, 12, 4: Del Rio, A; Navarro, C and Font, J (2015): Ciudadanía, políticos y expertos en la toma de decisiones políticas: la percepción de las cualidades de los actores políticos importan, Revista Española de Investigaciones Sociológicas. Denters, B et al (2015): Procedure or performance: what do citizens expect from local government and how does this affect their satisfaction?., Paper presented at the ECPR Joint Sessions, Warsaw, March. Donovan, T. and. Karp, J. A. (2006). Popular Support for Direct Democracy, Party Politics,12(5): Eliasoph, N. (1998). Avoiding politics: How Americans produce apathy in everyday life. Cambridge University Press. Kriesi, H.P. and Ferrin, M (2014): How Europeans view and evaluate democracy, EUI, Unpublished presentation. Font, J. Wojcieszak, M ; Navarro, C. (2015): Participation, Representation and Expertise: Citizen Preferences for Political Decision-Making Processes, Political Studies, DOI: / Font, J. and Fernández, J.L. (2014) Who Supports Expert Governance?. Paper presented at the Jornades de Comportament Polític, ICPS, Barcelona. Font, J. and Fernández, J.L. (2015) Direct democracy and responsiveness: two paths towards enlarging citizens' voices in democratic politics?, Paper presented at the Conference Democracy: a citizen perspective, Abo (Finland), May Session: Citizens and representative democracy. Evans, M. et al. (2013) How do australians imagine their democracy, Anzsog, Institute for governance at the university of Canberra. Grimes, M and Esaiasson, P (2014): Government responsiveness: a democratic value with negative externalities?, Political Research Quarterly, 67 (4): Grönlund, K; Bächtiger, A and Setälä, M (2014): Deliberative mini-publics, Colchester: ECPR Press. Held, D (2006): Models of democracy, London: Polity Press. Hibbing, J. and Theiss-Morse, E. (2002). Stealth Democracy. Cambridge: Cambridge University Press. Jacobs, L. R; Lomax Cook, F. and Delli Carpini, M. (2009): Talking together. Public deliberation and political participation in America, Chicago: University of Chicago Press. Lijphart, A (1999): Patterns of democracy, New Haven: Yale Univeristy Press. Madjick, Z.P. and Keith, W.M. (2011) Expertise as Argument: Authority, Democracy,and Problem-Solving, Argumentationm DOI /s z Manza, J; Cook, F. L. and Page, B. (eds.) (2002). Navigating Public Opinion: Polls, Policy, and the Future of American Democracy. New York: Oxford University Press. Morales, L; Pilet, J.B. and Ruedin, D (2015): The Gap between Public Preferences and Policies on Immigration: A Comparative Examination of the Effect of Politicization on Policy Congruence, Journal of Ethnic and Migration Studies. Navarro, J. Clemente (2012). Procesos y confianza política: quiénes deben ser virtuosos? in J. Font; J. C. Navarro; M. Wojcieszak and P. Alarcón Democracia sigilosa en España? preferencias de la ciudadanía española sobre las formas de decisión política y sus factores explicativos, Madrid: Centro de Investigaciones Sociológicas. Neblo, M. et al (2010). Who Wants To Deliberate And Why?, American Political Science Review, 104 (3), Schuck, A and de Vreese, C (2015): Public support for referendums in Europe: a Cross-national comparison in 21 countries, Electoral Studies, 38:

19 Sandel, M. (1996) Democracy's Discontent: America in Search of Public Policy. Cambridge, Massachusetts: Belknap Press of Harvard University Press. Smith, G (2009): Democratic Innovations: Designing Institutions for Citizen Participation, Cambridge: Cambridge University Press. Sunstein, C. R. (1999) Free Markets and Social Justice. Princeton, NJ: Princeton University Press.Tyler, Tom. Vetter, A. (2009). Citizens versus parties: Explaining institutional change in German local government, Local Government Studies, 35 (1): Webb, P. (2013): Who is willing to participate? Dissatisfied democrats, stealth democrats and populists in the United Kingdom, European Journal of Political Research, 52: Yvette, P. and Sander, J.E. (2015) Differential Responsiveness in Europe: The Effects of Preference Difference and Electoral Participation, West European Politics, 38:3, , DOI: /

20 Appendix: Table 1 in appendix. Independent variables in the four countries analyzed. ESS data SPAIN FINLAND UNITED KINGDOM THE NETHERLANDS N Mean Stdev Min Max N Mean Stdev Min Max N Mean Stdev Min Max N Mean Stdev Min Max Age ,6 18, ,8 18, ,8 19, ,2 17, Sex (Male) Percent , , , ,4 Incomes (no fifth quintile) Percent , , , ,8 Relative education (below) Percent , , , ,2 Ideology ,6 2, ,6 2, ,1 1, ,3 2, Political Trust ,4 1, ,2 1, ,7 1, ,5 1, Satisfaction democracy ,0 2, ,8 1, ,5 2, ,2 1, Evaluation Referendum ,0 3, ,3 2, ,5 2, ,8 2, Evaluation responsiveness ,5 2, ,6 2, ,9 2, ,3 1, Social Trust ,1 1, ,4 1, ,7 1, ,0 1, Interest in politics ,8 0, ,4 0, ,6 0, ,3 0, Mass Media ,0 1, ,0 1, ,9 1, ,1 1, Deliberation ,3 2, ,6 2, ,9 2, ,3 2,

21 Table 2 in appendix : Independent variables implemented. National Surveys on Stealth Democracy. SPAIN FINLAND UNITED KINGDOM THE NETHERLANS N %/Mean N %/Mean N %/Mean N %/Mean Age Sex (dummy) 1=Male , , , ,3 Incomes (dummy) 1=No greater decile (includes dk ) , , , ,0 Education (dummy) 1=Vocational to highest , , , Ideology1=Very left; 7=very right (4 includes dk ) , , , ,23 Financial situation personal1=better;2=same; 3=worse , , ,44 No available Financial situation country 1=Better;2=same; 3=worse , , ,59 No available Satisfaction democracy1=very satisfied; 4=not at all , , , ,14 Political Trust1=Strongly Distrust; 5=Strongly Trust (3 includes dk) ,08 No available , ,94 Political Party Identification (dummy)1=no (includes dk ) , , , ,1 Referendum (dummy) 1=Support , , , ,9 Interest in politics (dummy) 1=No , , , ,0 Citizens should influence policy making (dummy) 1=Support , , , ,3 Internal political efficacy (dummy) 1= Politics are so complicated , ,3 No available ,2 Political efficacy (dummy) 1=Politicians do care about the opinions of people , ,6 No available ,3 Internal political influence (dummy) 1=I have no influence No available , , ,9

22 Table 3 in appendix. Logistic regressions. Dependent variable (1 core supporters of referedum; 0 core supporters of responsiveness Stealth democracy, responsiveness and direct democracy in Western Europe