Democratic Innovations in Deliberative Systems – The Case of the Estonian Citizens ’ Assembly Process

With the proliferation and application of democratic innovations around the world, the empirical study of deliberative and participatory processes has shifted from small-scale environments and experiments to real-life political processes on a large scale. With this shift, there is also a need to explore new theoretical approaches in order to understand current developments. Instead of analyzing democratic innovations in isolation, the recent ‘systemic turn’ in the field encourages us to broaden our perspective and evaluate democratic innovations as complementary parts of a political system. This paper will draw upon a qualitative case study, based on interview and supported by survey data, of the ‘Estonian Citizens’ Assembly Process’ (ECA), in order to operationalize the systemic approach to deliberative democracy and illustrate how this can be applied to an analysis of democratic innovations. The ECA spanned more than a year (November 2012 to April 2014) and covered three political arenas: the public sphere, democratic innovations and representative institutions. The systemic analysis highlights the deliberative strengths and weaknesses of arenas and institutions, and illuminates how various arenas and democratic innovations did and did not complement one another in the creation of a deliberative process. The systemic analysis offers two possible interpretations of the ECA. The more affirmative interpretation is it constituted a deliberative system, as it did perform the three main functions fulfilled by different arenas and institutions. The more critical interpretation is that the ECA partly failed to be a deliberative system, due to social domination and decoupling of institutions.


Introduction
The most appropriate venue for citizen participation and deliberation has been a matter of ongoing debate.While some scholars have argued that we need safe havens for deliberation among smaller groups of citizens who can meet and discuss political issues face-to-face (e.g.Fishkin, 2009), others contend that deliberative processes must scale up to involve more citizens, and connect deliberative participation to decision-and policy-making processes (Blaug, 2002;Pateman, 2012).For quite some time, the proposals for new forms of participation and deliberation have pulled in different directions.
With the proliferation of democratic innovations around the world, the analytical focus has gradually shifted from the study of artificially created environments for participation and deliberation, to real world events involving a mix of action, participation, interests and deliberation (Geissel & Newton, 2012;Parkinson, 2006;Smith, 2009).
By way of response, an opportunity to explore a third middle-ground position concerning deliberation and participation has presented itself.By applying insights from various democratic models in order to understand the democratic system as a whole (Parkinson & Mansbridge 2012;Warren 2012), we will assess the parts not in isolation, but in terms of how they interact and contribute different functions to the democratic system.Instead of focusing on single aspects of democratic life, politics is thought to take place across a comprehensive system: incorporating a range of differentiated, but interrelated, spaces and actors, including deliberation, protests, negotiation and voting.By adopting a systemic approach, the parts are differentiated in the sense that they serve different functions, promote different modes of communication, bring in different kinds of knowledge and are open to varying degrees of participation and different sorts of actors.
This paper draws upon a qualitative case study, based on interviews and supported by survey data, of the Estonian Citizens' Assembly Process (ECA).The ECA is a unique case, given its combination of online and offline democratic innovations, including both self-selected and randomly selected participants.It originated in the wake of a political scandal involving a scheme of illegal party financing exposed in May 2012 by the former MP, Silver Meiker.Latent distrust towards the political system in Estonia developed into a crisis, characterized by antipolitical sentiments, antagonism and protests.This culminated in a pamphlet, Harta 12 (Charter 12), and an online petition with more than 18,000 signatures.In the process that followed, two innovative solutions were introduced: an online crowdsourcing platform designed to collect policy proposals from citizens; and a modified version of a 'deliberation day' called 'Rahvakogu'. 1 This decided upon 15 proposals for new legislation, handed over to Parliament by the President.In April 2014, three of these proposals were implemented by Parliament.
The ECA thus provides an interesting case through which to examine how various arenas offer different strengths and weaknesses within the democratic system; and also how the complementary functions which the public sphere, democratic innovations and representative institutions provide helps create a deliberative system.
The aim of this article is to operationalize the systemic approach, formulated by Mansbridge et al. (2012), into an analytical scheme.The arenas and institutions applied in the ECA will be assessed in terms of how they perform the epistemic, ethical and democratic functions that make up a deliberative system.The main focus will be upon the democratic innovations applied in the process; while a secondary focus will look at the public sphere and representative arena.
The paper is structured as follows.First, a discussion on the arenas involved in the process will be presented.Second, the systemic approach to democratic analysis is discussed.Third, the data and methods applied are set out.Fourth, the case of the ECA will be presented, based on interview and survey data.The article will end in an analysis and discussion of the application of a systemic approach to the analysis of processes such as ECA; as well as what conclusions can be drawn.

Three Arenas for Political Participation
As the citizenry of post-industrial western democracies has become more educated, less focused on materialistic desires, more distrustful of politicians and political institutions, and more self-expressive (e.g.Dalton, 2004;Ingelhart & Welzer, 2005;Norris, 1999), new forms of political participation have emerged in various arenas: the representative arena, public sphere and democratic innovations arena.
The representative arena consists of the political institutions holding the legislative and executive power.Although all 'modern democracies exhibit a variety of formal government institutions' (Lijphart, 1999, p. 1), it consists of institutions whose members are elected through direct or indirect free and fair elections (Dahl, 2000, p. 85).Laws, policies and decisions on public action are taken and implemented inside representative institutions.The links between citizens and their representatives are therefore of great importance for the function and the legitimacy of the system.Traditionally, the main channel for citizens' participation is that of voting for political parties or candidates, becoming a member of a political party or a candidate themselves.
The public sphere arena, on the other hand, enables citizens to participate in a variety of political activities: joining a demonstration; signing a petition; writing a political blog; joining a strike; boycotting certain goods; or occupying public space.These activities do not enjoy legislative or executive power, but are perceived as featuring indirect power (e.g. over agenda-setting or putting pressure on actors in the representative arena).As well as citizens, a range of actors such as think tanks, NGOs, social movements, neighborhood and leisure associations, companies, unions, experts, artists and various forms of media, all participate in the public sphere too.Debate over what it contributes to democratic polity often departs from Jürgen Habermas' conception (1989): discussed by, amongst others, Calhoun (1992), or Crossley and Roberts (2004).
The democratic innovation arena has sprung out of new demands for citizens to become involved in the policy-making process.It supposed to function as a link between the public sphere and representative arena (Smith, 2009).While some democratic innovations are open for all citizens to participate (self-selected), others are only open for those selected by the government or particular organizers (randomly selected).
'Democratic innovation' does not have a single definition, but this article concurs with Kenneth Newton (2012) that a democratic innovation may be defined 'as the successful implementation of a new idea that is intended to change the structures or processes of democratic government and politics by improving them' (p.4).The core idea here is to connect citizens to the decision-making and policymaking processes.In Graham Smith (2009)'s definition, two general aspects are critical: 1) that institutions directly engage citizens; 2) innovation concerns institutionalized forms of participation, in which citizens have a formal role in policy, legislative or constitutional decision-making (p.2).Democratic innovations also take on different forms in order to serve different purposes in different contexts (Åström et al, 2013b).

A Systemic Approach to Deliberative Democracy
In order to get to grips with the novelty and importance of the systemic approach, it is vital to understand from where the analysis of deliberative and participatory practices and processes departs.As the impact of deliberative democratic theory in the social sciences has been so vast over recent decades, it is not possible to cover the whole discussion here.Yet we can discern three general phases, or generations, in the development of deliberative and participatory democracy (Elstub & McLaverty, 2014;Jonsson & Åström, 2014), discussed below.
This was summarized by Dryzek (2000): namely, that democratic theory had taken a 'deliberative turn'.
The third phase of deliberative research is still in its infancy, but aims to broaden its application onto ongoing political processes, such as democratic innovations in action, and 'deliberative systems' as a whole.According to Mansbridge et al. (2012), a systemic approach to deliberative democracy 'encompasses a talk-based approach to political conflict and problem-solving -through arguing, demonstrating, expressing and persuading' (p.5).By addressing central discussions in the field, such as the role of multiple stakeholders and citizens (Kahane et al., 2013), that of information in deliberative processes (Gudowsky & Bechtold, 2013), and the problem of scaling up such processes (Friedman, 2006;Parkinson, 2003), a systemic analysis is intended to provide a deeper understanding of how deliberative and, presumably, non-deliberative institutions complement each other and help create a sustainable democratic system.Mansbridge et al. (2012) continue by arguing that '[t]o understand the larger goal of deliberation […] it is necessary to go beyond the study of individual institutions and processes to examine their interaction in the system as a whole.' (p.2).With a systemic approach, all aspects of a political system have different, potentially deliberative, functions that, in relation with one another, create a deliberative system.Even those actions previously regarded as non-deliberative (such as voting, bargaining and negotiation), should be analyzed (Mansbridge et al 2010, p. 64).Protest, for example, is said to 'facilitate and promote the circulation of useful information, [and] ethically respectful interactions among citizens [and] correct inequalities in access to influence by bringing more voice and interest into the decision-making process' (Mansbridge et al, 2012, p. 18).Yet protests also need to be analyzed case by case, and in their societal context.Mansbridge et al. (2012) take the example of the Tea Party and Radical Left protests: where the systemic benefits of new information and engagement in the public sphere were 'outweighed by the partisan and aggressive tenor [that created] a toxic atmosphere for deliberation and thus, is not system enhancing over time ' (p. 19).
To analyze the potentially deliberative parts of a system, Mansbridge et al. (2012) distinguish three main deliberative functions: epistemic, ethical and democratic.The epistemic function is to 'produce preferences, opinions, and decisions that are appropriately informed by facts and logic ' (p. 11).Thus, to be considered as having an epistemic function, the practice or institution must either add information or knowledge to the process/system, or provide or create venues for opinion and will-formation in a convincing way.The ethical function is to promote mutual respect among citizens (p.11).Since mutual respect is a moral issue and lacks universality, a problem of interpretation and analysis arises.There is no clear-cut answer in the literature, but some arguments involve nondomination, openness towards other opinions, and the attribution of moral status to other stakeholders in the process (Mansbridge et al., 2012, p. 11).
The democratic function promotes an equal, inclusive political process in which multiple and plural voices, interests, concerns and claims are acknowledged and taken seriously (p.12).This focuses on issues such as the selection of participants (open, self-selected or randomly selected) in deliberative processes, representation in citizens' juries and the inclusion of NGOs in expert committees.
To conduct a systemic analysis, Mansbridge et al. (2012) argue that each part in the system up for inquiry must be analyzed through these three functions.The sum is then weighted in the analysis, where an deliberative system could be seen as a 'loosely coupled group of institutions and practices that together perform the three functions […] seeking truth, establishing mutual respect, and generating inclusive, egalitarian decision-making' (Mansbridge et al. 2012, p. 22).
The analytical framework operationalizes the main ideas of Mansbridge et al. (2012).The framework focuses on the strengths and weaknesses of the epistemic, ethical and democratic functions in a democratic system.Following this framework, this article will provide answers to two main research questions: 1) What are the strengths and weaknesses of the functions that the different arenas in the ECA contribute to the system?

Case Description and Methods
In the empirical literature on democratic innovations, case studies predominate.One of the main reasons is that cases are not strictly comparable: either the political context or combination and application of innovations differ from other cases.ECA must be considered as a unique case, and thus has an intrinsic value for research (Bryman, 2008).
The data come from five sources: interviews with key stakeholders in the process (N=9); a survey conducted among citizens participating in the crowdsourcing part (N=848); two workshops with stakeholders in the process during 2013 and 2014; observations at a workshop in Tallinn with civil society organizations; and a survey conducted with participants on site at the Rahvakogu (N=298).The article is mostly based on the interviews, while the surveys provide complementary data.
The interviews took place in three different environments.Six were conducted with key stakeholders in the ECA in Tallinn in September 2013.The interviewees were selected by a snowball method; all hold central positions within their organizations.One interview was conducted online (through Skype) in December 2013.The final two interviews were held during workshops in 2013.
The survey of citizens participating in the crowdsourcing part of the process was sent online (using Survey Monkey) in May 2014.It was sent out to every participant registered on the site (N=2042) and received (N=848) answers.The response rate (41.5%) was good.The survey included questions about the ECA process, experiences of political participation, trust in political institutions and satisfaction with democracy; as well as a number of questions relating to sociodemographic background.The data from the Rahvakogu was collected on site in April 2013 by Praxis Center for Policy Studies.The total number of participants was N=314, with a response rate of 95% (N=298).
The interviews were analyzed through qualitative text analysis.As some interviewees wished to be anonymous, they are always referred to according to their institutional or organizational affiliations instead.The interviews were recorded and transcribed into clean text (rather than verbatim CA-style).

The Estonian Citizens' Assembly Process (ECA)
The ECA consists of numerous actors and institutions spanning three political arenas: the public sphere arena, the democratic innovations arena and the representative arena.In this section I will describe and discuss the process by presenting the arenas, institutions and actors active in the ECA, based on the interviews and supported by survey data.

Public Sphere Arena
The first phase of the ECA was initiated in the public sphere and consisted of various actors: starting with the former MP, Silver Meiker, who became a whistleblower.To put the scandal he exposed into context, two general observations on political circumstances in Estonia and the Baltic region are necessary here.First, there is a tendency toward democratic backlash, manifested as low trust in political institutions and low levels of political participation.In The Economist Democracy Index 2012, Estonia is categorized as a 'flawed democracy' 2 , particularly due to low scores in political participation (Economist Intelligence Unit, 2013).Even though Estonians demonstrate 'decent levels of trust' in public institutions such as the legal system, the political parties are 'bitterly distrusted' (Lagerspetz & Vogt, 2013, p. 62).The president's office does, however, garner fairly high levels of trust (Duvold, 2006).Second, there has been a rash of political scandals in Estonia in the past decade, with the 2011 scandal, in which two party leaders were charged with corruption (Lagerspetz och Vogt, 2013, p. 57), freshest in the public memory.It is reasonable to assume that these recent scandals, in tandem with latent mistrust, helped leverage the legitimacy crisis.
In such a political climate, a spark can easily catch fire.In May 2012, Meiker published an inside account about how money for political campaigns was filtered through private accounts (for example his own).This is prohibited by law in Estonia.In his interview, Meiker alluded to his internal motivations for going public: he didn't 'think that one article can change society' [Interviewee 1], but 'believe [d] that there are small things and small changes that makes a bigger change.The reason why this article went so big, the scandal went so big, it was not because of what I wrote, it were known that these kind of schemes exist, No.99 theatre project 3 showed it' [Interviewee 1].
Yet it was not until autumn of the same year that the public reacted to the scandal by protesting in the streets.As suggested in an interview with an NGO activist, there seems to have been a common feeling that the lid was off when the protest began: 'I was very glad when there was this demonstration against the lying politicians.Really coming to the streets against corruption because it reached already the critical points where people could not stand all this fussing around' The combination of Charter 12's circulation by the media, the protests and the successful online petition certainly put pressure on the political establishment to act.The political parties had, thus far, done nothing concrete other than letting the Minister of Justice step down.Now, however, the President stepped in: after which the protesters left the streets and negotiations began.

Democratic Innovations Arena
'The Ice-Cellar Meeting'.President Toomas Hendrik Ilves began by inviting representatives from civil society organisations, political parties, social scientists, lawyers, and the signatories of Charter 12 to a meeting in the Jääkelder ('ice cellar': hereafter referred to as 'the Ice-Cellar Meeting').As a frequently proclaimed champion of discussion in the public sphere [Interview 7] and someone regarded as a mediator between the people and the politicians, the President chose to act as 'the so-called atmosphere or 'language' in the newspapers articles got really nasty and then the President finally thought that something need to be done' [Interviewee 4].
His move was, however, not uncontroversial.A civil society group highlighted '… how it all got so structured and institutionalised in a way that the President took it over and they created this whole concept and timeline and everything' [Interviewee 2], and how this cooled down the momentum created in the streets.
At the meeting, broadcast online [Interviewee 3], the President and an umbrella organization for civil society organizations explicitly referred to the ideas of James Fishkin; as well as to the Icelandic experience of crowdsourcing the constitution (Åström et al, 2013a;Landemore, 2014) as possible ways forward [Interviewee 7].One of the participants, representing the NGOs, says in retrospect that the process 'came out actually pretty much the way that we planned it.
[…] You could say that we designed this process and so we could put in everything that we wanted to try and that we were interested in [e.g.] [t]he idea to go for crowdsourcing Thus, from the beginning, there was a common understanding that deliberative features and ICT solutions should somehow be implemented in the process.Yet it was not decided how they should be implemented.During the negotiations, the political parties seemed uninterested, especially about crowdsourcing.They merely argued for 'politics as usual'.'It was clear that political parties will never take active part in solving this issue and if we wanted to get something done, then it's clear that we [the NGOs] have to be the initiators' [Interviewee 5].
Despite the approach taken by the parties, the Ice-Cellar Meeting had a huge impact on ECA proceedings.Decisions to both initiate a crowdsourcing process and establish a deliberation day were both taken.It was also agreed that five specific topics connected to party financing should be defined, in order to render discussions over crowdsourcing and the deliberation day more fruitful. 5 Crowdsourcing Proposals.The Estonian Cooperation Assembly 6 , along with civil society actors, set up and managed the crowdsourcing process.To connect this to the deliberation day, a web page was formulated around the five chosen topics.
Every contribution had, in some way, to deal with one of these topics.Among the requirements for citizens to be able to post material on the website was for them to log in with an electronic ID 7 , thus making the identity of contributors known, and their suggestions publicly accessible.An initial aim was to reduce public animosity; interestingly, the design choice of electronic IDs seems to have had this effect.'Really, 99% of the proposals and comments were, well in a neutral tone.Of course some used more colorful language, but it was seldom hostile.' [Interviewee 7] As the process ended, a total of 2,000 original proposals and 4,000 comments on those had been posted on the website [Interviewee 7].At this stage, an expert group consisting of representatives from civil society organizations started to categorize and organize the proposals by hand [Interviewee 4].
When looking in more detail at who each participant was, we find that those involved in crowdsourcing were already politically active citizens.As we see in Table 1 below (based on the survey data collected, N=848, from the total sample of 2042), these participants were over-represented in every form of political participation: from formal participation (such as contacting politicians and working in political organizations), to informal participation (such as signing petitions and boycotting certain goods).Expert Meetings.After categorization, experts in various fields such as economics, political science and law were invited to make an impact assessment, and contribute their own opinions on the collected proposals.This work was all conducted pro bono.In addition, a series of five deliberative seminars was held during one week in March 2013: in which political representatives, experts, and citizens who had contributed original proposals in the crowdsourcing process deliberated upon the five issues.One of the participants explained that the proposals was drafted and then 'discussed in the smaller meetings.
[…] The bundled ideas together with expertise were then rated whether they were relevant' [Interviewee 4].The crowdsourcing suggestions boiled down to 18 proposals in total, which were handed over for further deliberation on the Rahvakogu ('deliberation day').
Rahvakogu.The last democratic innovation, the Rahvakogu, was held on 6 April 2013.Randomly selected from all Estonian citizens, 550 individuals were called to participate.Of these, 314 chose to do so.
As the Rahvakogu was planned, attention turned to the expectations which stakeholders could have of the policymaking process.As the Rahvakogu was not formally sanctioned by the parties, but instead by the President, there were no clear-cut answers to this.One of the organizers recalls that: 'I tried to explain from the first moment of this process to all the participants that we can't expect the parliament to act like a rubber stamp to all the proposals coming from People's Assembly' [Interviewee 4] The day itself became a major media event, attracting a great deal of public attention.The 18 amendments were discussed and deliberated upon by the 314 randomly selected participants.At the end, a formal vote took place, and 15 of the amendments were passed.Following deliberation day, a survey was circulated, completed by a large majority (N=298 out of 314) (Appendix 1).The results suggest a consensus-oriented atmosphere, in which almost 90% 'agreed' or 'strongly agreed' that 'the voting results were similar to their own viewpoints'; and that they 'perceived an increased knowledge in the topics discussed'.Additionally, the vast majority of participants 'became more interested in politics in general'; while all those who completed the survey agreed that 'I am happy that I decided to participate', of whom 88% 'strongly agreed'.

Representative Arena
As the entire ECA was an ad-hoc solution, there were no clear-cut institutional regulations for how these proposals would reach Parliament.President Ilves therefore utilized his privilege: the right to propose Bills in the legislature, and handed over the 15 agreed proposals to the Parliamentary Constitutional Committee.
Yet despite this, Parliamentary procedures remained unclear in terms of both how to weigh up the proposals; and indeed, the entire purpose of the process from a Parliamentary perspective.
The ECA was conducted in order to restore order and engage citizens in the policy-making process.On this, the MP interviewed elaborated somewhat, arguing that citizens do not feel listened to, but that crowdsourcing could function as a channel for this.'It's about bringing the Parliament closer to the people.So that's important tactic.I mean, it's not like I think that the people are dumb, but I'm really sceptical about extremely smart things coming out of this.But it's more about helping you to draft [a law], which is already more or less there' [Interviewee 6].
However, Parliament kept its doors for further cooperation closed.One of the organizers of the process states that there was enthusiasm among citizens about Rahvakogu, but 'unfortunately, the Riigikogu turned away the offered support and maintained their preference to closed working processes and a rather dismissive attitude' (NENO 2014).By April 2014, three proposals had been implemented, and an additional four had been partly committed to by the Riigikogu (PRAXIS & NENO 2014).
The three proposals turned into law were as follows: (1) 'Legalisation of popular initiatives', meaning that proposals for the amendment of legal provisions and a better organisation of civil society may be submitted to the Riigikogu through a memorandum containing at least 1,000 signatures (PRAXIS & NENO 2014, p.2); (2) 'Allow for a political party to be founded with 200 members' (an amendment to the Political Parties Act, which was not fully accepted; instead, the limit was reduced to 500) (PRAXIS & NENO 2014, p.3); (3) 'Possibility to replace the election deposit with supporters' signatures', which reduced the previous deposit amount by half, making participation in the Riigikogu elections more accessible (PRAXIS & NENO 2014, p.4).The ECA cuts across several spheres in Estonian society.Through the systemic approach, it is possible to analyze the deliberative strengths and weaknesses of each part, but also to assess the process as a whole and identify how the various components interplay with each other as part of a democratic system.The following analysis will thus focus upon the strengths and weaknesses of each arena, how the various parts are complementary to one another (or otherwise), and how we might go about interpreting the ECA as a case of a deliberative system.

Contributions of Knowledge and Information to the System
It can be argued that the various activities in the public sphere brought knowledge and information into the system.Information from the media kick-started the protests: amplifying the message by bringing it into public awareness.This is in line with Mansbridge et al. (2012, p.18)'s argument that protests can play a constructive role.The main contribution from Charter 12, and the online petition which resulted, was to bring the issues into public awareness and establish an initial point of departure by compiling a list with suggestions for policy changes.
The applied democratic innovations also contributed knowledge and information.It is, however, important to value what form of knowledge was added, and from whom it originated.The 'Ice-Cellar Meeting' brought the perspectives of competing elites to the table: providing knowledge and information from a plural set of actors.The explicit references in the interviews to the academic work and experiences of crowdsourcing are illustrative examples of how the participants provided a base founded on research and experiences from other political systems and contexts.
Crowdsourcing was supposed to add to the multitude and plurality of perspectives and ideas of lay citizens, and would thus permeate the process.However, the data shows that those ideas which were tapped into did not emanate from Estonian people as a whole 8 ; but rather from an already politically active clique of society, with a bias towards professional Estonian men.This validates Brabham (2012, p.407)'s conclusion that crowds are not amateurs, but consist to a large degree of 'self-selected experts and what we might otherwise call professionals'.By refining the raw proposals through expert meetings, another dimension of knowledge was brought into the process: that of political craftsmanship.
The last instance, the Rahvakogu, also contributed knowledge, giving participants an opportunity to weigh and discuss the proposals in a non-coercive environment.Meanwhile, representative institutions contributed knowledge inasmuch of their adaptation of the Rahvakogu propositions to the existing political system and its mechanisms.

Contributions to Mutual Respect among Participants
In terms of the public sphere, it can, on the one hand, be argued that a 'we-thecitizens'-moment created an atmosphere of mutual respect.On the other, the political representatives were not included in this atmosphere.On the contrary, they perceived it as hostile and the public tone as harsh.Heated arguments from the protestors, and the high-pitched critique of Charter 12 were perceived as counter-productive to a constructive dialogue.This resembles the argument made by Mansbridge et al (2012), that some protest movements create a hostile, polarized environment instead of facilitating public dialogue.
The 'Ice-Cellar Meeting' exhibited the opposite problem.Given the invitation of indirect representatives of all stakeholders involved in the conflict, the meeting obviously served to create mutual respect among the latter.In turn, this provided a platform for creating mutual respect among citizens by designing democratic innovations.
The crowdsourcing platform seems to have contributed to that sense of mutual respect, through its application of electronic IDs and the reportedly positive or at least neutral tone employed in the arguments.It also filtered the protests in the streets into constructive ideas which could be used when formulating policies.The Rahvakogu also fostered mutual respect through its strong emphasis that participating citizens should be heard during the deliberations.
It is less easy, however, to identify any active strategies for the generation of mutual respect among participants within the representative institutions.Other than their attendance at the 'Ice-Cellar Meeting', there was no genuine effort on their part to participate in the rest of the process.The President, though, played an important role: such is the power of the office and his popularity among Estonians.Thus it appears that the application of democratic innovations was key in fostering mutual respect: something which neither the protests, discussions in the public sphere, nor the representative institutions could provide.

Contribution to Equal Democratic Participation
During the initial phase of the process in the public sphere, all political activities can be argued to have been inclusive.The loosely connected protest movement did not have any formal requirements on citizen participation, while the online petition was open for all to sign.
The raison d'être of democratic innovations is the linking of citizens to the policymaking process.Crowdsourcing is, theoretically, an inclusive innovation; yet the data shows that the crowd was not representative of the entire population.Instead, a majority were already active, professional Estonian men.
Crowdsourcing thus offered a platform for equal, inclusive participation, but failed to attract large groups of citizenry.
The main event, the Rahvakogu, did not face the same problem, as it was based on a random sample of citizens.Yet another problem becomes critical when applying mini-publics such as Rahvakogu: can such a small number of citizens be seen as legitimate representatives for the whole population, despite the fact that they are randomly selected (Parkinson 2006)?On the one hand, it certainly spared the process from the common critique of the dominance of special interests.On the other, it left the most vital aspect of the whole thing, the decision-making moment, or at least the to-be-handed-over-for-decision-making moment, to a small number of randomly chosen citizens.
The last stage of the process was that of the representative institutions; but the actors within these (politicians and civil servants) did not engage, other than at the Ice-Cellar Meeting.The main problem was the inability to involve and engage citizens in the final policy-making stage, due to the lack of formal institutional arrangements.
In Table 3 below, the ECA is summarized and analyzed through the analytical framework presented above.To summarize the results from the three research questions: the process as a whole did, in many respects, fill a democratic function, an ethical function and an epistemic function.While some arenas and institutions contributed more strongly by adding knowledge and information, others contributed mutual respect among participants; and others, democratic inclusion.Did not create respect between citizens and politicians/institutions.
Possibly high costs of participation in protests (location and resources).

Strengths
Plural forms of knowledge and information added.
Contributed to mutual respect through crowdsourcing and Rahvakogu.
Involved a mix of self-selection and random samples of the citizenry.

Weakness
No direct interaction between citizens, experts and representatives.
Only indirectly and vaguely contributed to bridge the citizen/ representative gap.
Better connected to the public sphere than representative arena.
Social and political bias in the crowdsourcing part.

Strength
Institutional expertise and the 'bigger picture' knowledge added.
Met the citizens by accepting 3 of the 15 proposals.
Involved various political parties in committee negotiations.

Weakness
No direct interaction with the stakeholders.
No deeper understanding of the proposals.
Did not actively participate in the process.
Did not engage in dialogue with citizens.

Interaction among the Political Arenas: A Deliberative System?
Departing from Mansbridge et al (2012, p. 22)'s depiction of a deliberative system, it is possible to interpret the ECA as an example of such, judging from the overall process which did indeed perform the three main functions.Yet as Mansbridge et al. (2012) discuss, some pathologies could make the system fail: tight-coupling, decoupling, institutional domination, social domination and entrenched partnership (p.22).Two of these pathologies (decoupling and social domination) could be interpreted to have occurred in the ECA, and will therefore be discussed here.
Due to the current abstractness of the systemic approach and that this represents the first attempt to operationalize it, this analysis will provide two possible interpretations of the ECA: one that is more affirmative of ECA as an deliberative process, and one more critical of the ECA as an deliberative process.
The more affirmative interpretation acknowledges that even though not all arenas and institutions individually met all deliberative criteria set up by Manbridge et al.
(2012), the system as whole did so.The interplay between self-selection institutions (protests, online petition and crowdsourcing), randomly selected institutions (Rahvakogu) and representative/elite institutions (ice-cellar meeting, expert meetings and Parliament) does, as shown in the table above, contribute different functions to the system and makes the process as a whole deliberative.
The process can be seen as 'a loosely coupled group of institutions and practices that together perform the three functions […] seeking truth, establishing mutual respect, and generating inclusive, egalitarian decision-making' (Mansbridge et al., 2012, p. 22), thus a system with deliberative qualities.
What should be highlighted is that the large numbers in the public sphere and crowdsourcing fueled the process with information and knowledge and provided space for democratic inclusion.The expert meetings contributed their knowledge when drafting the proposals; Rahvakogu democratic inclusion and legitimacy through the random sample of citizens.Parliament then weighted the proposals and implemented those deemed suitable in the current political context.From this perspective, we can interpret the ECA as a process in which the demands of the people in the streets became refined and ended up as new legislation through the application of democratic innovations.
The more critical interpretation is based on the findings that the process was, in practice, framed, shaped, and executed by already active members of society/and or elites.We might view the initial protests as, in some sense, a 'citizen uprising'; yet without data and control for who participated, it is not possible to state that this was definitively the case.The data does reveal, though, that the frame for the whole process was set up by the President and a few key representatives from civil society, i.e. elite actors in the Estonian political system.This frame set the stakes for participating in crowdsourcing: which again according to the data, could be labeled as suffering from social domination, due to the strong majority of highly educated, professional men.
In the next stage, a group of experts transformed the ideas and opinions into policy proposals, adding to the professionalization of original demands from the citizens.According to such an interpretation, those proposals deliberated upon by the random sample of citizens at the Rahvakogu were already biased and unrepresentative.
The other major problem is that of the decoupling of the institutions involved in the process.Despite three of the proposals actually being implemented, the connection between the initial protests, democratic innovations and Parliament was rather weak.I think, however, that in the ECA's case, 'decoupling' is too strong a term to employ.If an deliberative system consists of 'loosely coupled' institutions, and systemic failure can be found in 'decoupled' institutions, the ECA could provide an example of 'too loosely coupled' institutions, i.e. not ideal but not a total failure either.If the President had not taken the Rahvakogu proposals into Parliament, there would not have been any connection to the decision-making institutions.The whole process could thus be captured in Dalton et al. (2004)'s concept of advocacy democracy: it is inclusive and deliberative, but representative institutions have the final, exclusive power to make the ultimate decisions.
The ECA provides us with an illuminating example of the complexity of deliberative processes, and how the systemic approach can be applied to better understand this.It also shows that such an approach can provide different interpretations, depending on how the process is viewed and what values are highlighted.
It is also interesting that the ECA is not a case of formal institutional change.
Although the ECA ended up in Parliament and did indeed have a policy impact, it did not penetrate or change the institutional structure or character.It provided platforms which channeled engagement from the public sphere to the representative institutions, but did not do likewise in terms of interaction and cocreation during the final stages of the decision-making process.The ECA can, however, to borrow Dryzek's (2009) criteria for how to judge a deliberative system, still be seen as consequential due to the fact that the process both had policy impact, and that it have had to some degree changed the relationship between representative institutions and citizens.One example on this changed relationship between representative institutions and citizens is the new tax legislation passed in February 2015 (an amendment to the Tax Administration Act), that was based on a petition created and signed in the e-petition system created by the ECA.The Parliament now, after the ECA, has a citizen-controlled channel to take into consideration when creating legislation, and while there are no formal requirements to do so, the Parliament obviously does so.
Thus, in its current state, the systemic approach offers at least two sorts of interpretations.To achieve more definitive findings, further empirical studies are required, while the theoretical discussion needs to deepen.Mansbridge et al. (2012) are aware of many of the problems, and do for example argue that the 'virtue of a deliberative system is that failures in one institution can be compensated for in another part' 22).Yet how should we understand deviations such as social domination within a single institution, if that in turn affects the deliberative process as a whole by framing the issue for deliberation?Should that be interpreted as a failure of a system or, as in this analysis, just a 'minus' in the judgment of the system as a whole?An additional question is about the potentially transitory nature of these kinds of processes; must a systemic analysis perhaps span over a longer time period to capture the consequences of the process on the representative institutions and political culture?To answer these questions, further research needs to address these issues empirically and contribute to the theoretical development of the systemic approach.

Conclusion
This article analyzes the strengths and weaknesses of the institutions and arenas involved in the Estonian Citizens' Assembly Process (ECA), with regard to whether it performed the epistemic, ethical, and democratic functions prescribed by the systemic approach to deliberative democracy.As an illustrative case, the ECA reveals the analytical benefits and weaknesses of applying the systemic approach, by highlighting how the public sphere, democratic innovations and representative institutions can perform different functions and interplay in the creation and shaping of policy.
The systemic analysis offers two possible interpretations of the ECA.The more affirmative interpretation is it constituted an deliberative process, as it did perform the three main functions fulfilled by different arenas and institutions.The more critical interpretation is that the ECA partly failed to be an 'ideal' deliberative process, due to social domination and decoupling of institutions.
Overall, we might conclude that the strength of this approach is that, by including various arenas, institutions and actors, it provides a framework for understanding deliberation within a wider context; and that different arenas, institutions and actors provide different functions.Its weakness, though, is the unclearness in how we should weight deviations such as social domination within a single institution and institutional decoupling.The survey was conducted on the 6 th of April 2012, on site after the main Rahvakogu event.N=298 (out of 314 participants).Source: PRAXIS Policy Center, Tallinn, Estonia.
civil society must convene an alternative institution, in which the representatives of civil society might dominate.
First and foremost, we desire support to be found in Estonia's political system for the following principles, central to which is opening of the system of political parties and giving civil society actual influence in politics: • the public must have a clear overview of political organizations' sources of financing -both of the usage of public money and the origins of their other revenues; • the establishment, financing and operation of political parties must be transparent, and political parties must represent the public-, not group interest; • the system of governance must clearly reflect the will of the voter -e.g. a representative must be obliged to report constantly to the voters of his or her constituency; • political parties in parliament may not monopolize the path to powerthe establishment of political parties and access of non-party forces to Riigikogu must be simplified; • citizens must have broader opportunities for expressing their political will than regular elections -an instrument for public initiative must be created.
These are principles, to the defense of which many different people with very different worldviews have spoken up over the last few months.However, this is not enough, and therefore, we find that efforts by the wider public are necessary for healing Estonia's democracy.We call on all Estonians that agree with our positions to join this charter.
weaknesses are found in the arena(s) concerning the contribution of knowledge and information to the system?b) What strengths and weaknesses are found in the arena(s) concerning the contribution of mutual respect among participants?c) What strengths and weaknesses are found in the arena(s) concerning the contribution of equal democratic participation? 2) In what way do the various political arenas in the ECA complement each other in a systemic manner?

Figure 1 :
Figure 1: An Overview of the Estonian Citizens' Assembly Process (ECA) and 'those holding power in Estonia no longer feel the need to take heed of the public' [Appendix 2].Here, Charter 12 appeared to set the tone for the public debate: a finding confirmed in the interview data.The interviewee from the President's Office states that: 'The whole sentiment was very heated already, the discussions were very heated, academia took very fierce positions in some of the cases [..] all ended up in a piece called 'Charter 12' [..] and it was signed by several persons who are opinion leaders in Estonia' [Interviewee 3].With citizens protesting in the streets and opinion leaders in favor of political change, the pamphlet metamorphosed into a petition at the website petitsioon.ee 4 , and collected over 18,000 signatures.
[Interviewee 2].At this point, Harta 12 (Charter 12) was published in the newspaper, Postimees.The entire pamphlet constituted a full frontal attack on the political establishment in Estonia, adopting a harsh but straightforward tone.It included statements such as 'Estonia's democracy is crumbling before our eyes', 'democratic legitimation has ceased'

Table 1 : Prior Political Participation among Crowdsourcing Participants versus the General Public
Moreover, a majority of participants shared certain characteristics.As we can see in Table2below, a majority of higher educated, professional, Estonian men leaned towards right-wing views.In comparison with the general Estonian population, the sample of crowdsourcing participants was, thus, strongly biased towards already politically active citizens: who to a large degree, actually shared the characteristics of what we might term the 'usual suspects' in political participation.
Note: The figures displayed above represent the proportion of citizens answering 'yes' to the question: 'Have you, during the last 12 months…'.* Statistics taken from the European Social Survey 2012, random sample (N=2377).** Statistics taken from our own survey among crowdsourcing participants in May 2014 (N=848).Statistical significances are displayed as follows * : p<.05, ** .<.01.Significance measures were produced through bootstrapping (with 1000 random samples) for 95% and 99% confidence intervals.

Table 3 :
Summary and Analysis of the Citizens' Assembly Process in Estonia