Non-binding Decision-Making

Korthagen et al. introduce six quite successful digital tools used in decision-making processes. Three tools are used by political parties (Spanish Podemos, the Italian Five Star Movement and the German Pirate Party) aiming for direct democracy and more transparency. The other three tools are online participatory budgeting (PB) tools initiated by municipalities (the Brazilian city Belo Horizonte, the French capital Paris and the Icelandic capital Reykjavik). The authors place a strong focus on the participatory process and practical experiences. For a better understanding of these tools and how they are used in practice, interviews were conducted with administrators and researchers familiar with the respective tools. All tools except one, the Podemos tool, prove to have an impact on decision-making processes with the party or municipality. The authors explain this success rate by the fact that the tools are embedded in actual formal decision-making processes: the internal party decision-making process and the budgeting process of local democracies.


Introduction
The Pirate Party Germany was founded in 2006 by 53 inhabitants of Berlin.In 2009, the German Pirate Party got officially represented in politics, with one seat in the European Parliament and seats in four state parliaments.Because of scandals and internal disputes, the Pirate Party lost the trust of the voters and is no longer represented in any state parliament.Key party issues of the German Pirate Party are direct democracy, copyright, digital communication, privacy and transparency (www.piratenpartei.de).Their internal approach is characterized by many possibilities to participate (online) in the decision-making process.The Pirate Party makes extensive use of new technology to communicate and collaborate on policymaking.The technology facilitates the intension of the Pirate Party to-in terms of Bolleyer et al. (2015: 159)-maximize the democratic equality both between followers and members and between members and elites.
The use of technologies as Liquid Feedback was introduced to cater to the aims of this new party: democracy and transparency.Because of these values, anyone could come to the party conventions and have a say: "This rendered the convention to a point where you weren't able to make a decision, because people were always discussing formalities.It wasn't really effective.At that point liquid feedback came in.If we want to be different from the other parties and if we take seriously that people want to participate via the Internet, then they have to be able to do that and join the decision making of the party.If we want to propagate this as the big difference of the Pirate Party, we need usable software, which is able to do this.So, they started creating the software, because we exploded so fast in members" (interview with a politician).
The Liquid Feedback software has been the backbone for the inter-party decisionmaking process and incorporates four steps (Paulin 2014: 221-223, see also Klimowicz 2016).It all starts with an initiative that can be proposed by any registered member.This proposal must receive support of at least 10% of the registered users within a predefined period of time.If the proposal has received enough support, time for discussion and eventual modifications of the initiative is allocated, again within a defined time frame.Eventually, members can cast their vote on the final proposal directly or (temporarily) delegate their votes to a confidential whom they trust.Liquid Feedback has been used from May 2010 till May 2015, although use of the tool had been gradually declining since then (interview with a researcher).
German law prescribes that all important decisions have to be democratically decided on at party conventions-so not online.These regulations mean that input from participants through the Pirate Party's Liquid Feedback tool is not necessarily of influence on decisions of the Party's members of parliament."If you decide something in Liquid Feedback, it can never be binding" (interview with a researcher).This had its impact on the effectivity of the tool for party members.The real decision-making process was taking place outside Liquid Feedback (interview with a researcher).
Internal procedures are written down in the Statute of the Pirate Party, which can be found on their website.About internal party policies and candidates, the Statute postulates: "In the German Pirates, candidates are elected by members on the regional or local level (Section 1, §10.2).The national executive is elected individually on an annual basis during the national membership meeting (Section 1, §9a.3).Changes to party policy and statutes can be initiated by all members and require a two-thirds majority during the annual membership meeting (Section A, § §12.1 and 12.3).Further, all members, if they have sufficient support, can initiate a membership referendum, the outcome of which is binding and equal to a membership meeting vote (Section A, § §16.1-16.6)"(Bolleyer et al. 2015: 165).The finances of the party can be found on a website; the main financial sources are the contributions of the State and membership fees.

Participants
The target group of participants are people who share the main values of the German Pirate Party.The voter profile for the Pirate Party is young, male, Internet-savvy and well educated.The German pirates attracted many protest votes and only 14% of the voters for the Pirate Party made this choice because of concrete programmatic or content-based issues (Greenstein 2013).Membership of the German Pirate Party is inclusive; it is not a problem if you are already a member of another political party.Decisions about membership are made at the lowest organizational level.The regional branch is allowed to reject membership applications."The concept of 'membership' in their national constitution explicitly refers to the obligation to work for the party as well as to the rights linked to membership.[. ..] members must be Germans who live in Germany, and must be aged over 16 (Section A, §2.1)" (Bolleyer et al. 2015, 164-165).
At first, the Pirate Party in Germany mainly attracted a community of people interested in the Internet itself, in issues around the Internet (such as copyright), and in digital instruments to realize horizontal decision-making.In 2009, the party gained more attention and popular support due to a few events: an Internet censorship law proposed by minister Van der Leyen (CDU) that would ban German Internet users from accessing child pornography (and the fear that the government would ban other Internet sites in the future), the SPD parliamentarian Tauss who switched over to the Pirate Party and the success for the Pirate Party in Sweden.In addition, the Pirate Party principles of openness helped to gain public support as well as the wider attractiveness of the digital instruments (Hensel 2014: 248).The number of members increased more than tenfold in 2009.After the party won seats in Berlin's state parliament in 2011, they also won seats in Saarland, Schleswig-Holstein and North Rhine-Westphalia in 2012.
The online design of the tool has always been intended to open sharing of all information and towards horizontal decision-making procedures, striving to include participants beyond the participation elite.But when the party managed to attract a broader and more diverse member base, the result was not that inclusive after all.In practice, it appeared that making meaningful interventions within online conversations presumed some digital competences, knowledge on the subject and strong connections within the Pirate Party.Not only did the new group lack abilities or knowledge to contribute to the internal discussion, their contributions in the debates were also seen as light-spirited and unprofessional (Hensel 2014: 249).What actually happened was that informal structures between a small group of members overruled the digital horizontal structures in the party decision-making processes, although this practice is in contradiction with all principles of the Pirate Party.
Despite the aims of the party of participatory decision-making, the German Pirate Party [as well as the Swedish (Bolleyer et al. 2015) and the Dutch Pirate Party (interview with two members of the Dutch national Pirate Party)] shows relatively low member commitment in participation."In the German Pirate Party, recent figures indicate that only 28 percent of members on the national level pay their fees and have full voting rights" (Bolleyer et al. 2015: 170).Another problem is that the German Pirate Party is a very male-dominated party; only few women are represented (Kulick 2013).Also, questions are raised with regard to the representativeness of participants in relation to the "super voters": "In a discussion on the effect on super-voters-i.e.users with a large share of incoming delegations-the democratic nature of the system was questioned, and many users became inactive" (Kling et al. 2015: 3).

Participatory Process
Members of the German Pirate Party can give many different sorts of input online.They can discuss issues, formulate proposals, provide feedback on proposals and they can vote on a proposal or delegate their vote to another party member they trust.For a long time, Liquid Feedback was the most important tool for internal decisionmaking processes.Kling et al. (2015: 3) described the Liquid Feedback process clearly: "In Liquid Feedback as used in the German Pirate Party, members can create initiatives which are to be voted on to obtain the current opinion of the party members, e.g. for collaboratively developing the party program.Initiatives are grouped into issues which group competing initiatives for the same issue.[. ..]Furthermore, issues belong to areas which represent main topics such as environmental policies.Each user can create new initiatives, which need a minimum first quorum of supporters for being voted upon.In Liquid Feedback, votes can be delegated to other voters on three levels: On the global level, meaning that all initiatives can be voted for by the delegate on behalf of the delegating user; on the area level, so that delegations are restricted on an area; or on the issue level.The actions of every voter are recorded and public, allowing the control of delegates at the expense of non-secret votes."The threshold of a minimum support by 10% of party members also effectively obstructs counterproductive contributions, because this threshold makes sure that unsupported views can be ignored without further discussion."The only possibility to give feedback in Liquid Feedback is to do it in a positive and constructive way.[. ..]If you want to change something, then you have to suggest what has to be changed" (interview with a politician).Voting on proposals in Liquid Feedback works through a preference system, in which multiple, prioritized votes can be cast.Participants can thus choose and prioritize several policy options at the same time (interview with a politician).
The use of the Liquid Feedback tool had been contested within the party since the beginning (interview with a researcher; interview with a politician).Many Pirates did not perceive the security to be well organized.Also, practical failures and fundamental democratic issues within the systems were noticed (Hensel 2014: 248).On complex matters, the tool does not provide enough information to enable sufficient participation, according to the interviewed researcher.This has had the effect that too few people actually participated for decisions to be legitimate.The system itself was quite complicated as well, which made contributing difficult for some people: "I'm one of these supposed digital natives because [. ..]I grew up with computers and internet more or less.[But] I find it very hard to find my way around this Liquid Feedback and to find what's discussed at the moment and who is voting for what and why.And, for example, if you delegate your votes, you will never see where your votes land in the end.[. ..]It looked like it was programmed by nerds, for nerds, and not for the general public" (interview with a researcher).Eventually many of the supporters of the Liquid Feedback tool left the party and it fell out of use (interview with a researcher).Hensel (2014) argued that the most important tools for the communication within the Pirate Party are the Wiki, mailing lists, a party-related blogosphere, digital party magazines, video and streaming services and the video conferencing software Mumble.
• Mumble "is an open and free voice conferencing software.Mumble allows us to carry out decentralized meetings and short action planning, without having to travel across the country."• PiratenPad "is a widely used virtual notepad in which several people can leave messages synchronously.This makes it possible to construct complete texts, action plans or protocols for meetings.The Piratenpad server can be reached at www.piratenpad.deand is available to everyone.This way, this Pirate Party software has become a popular tool for companies, NGOs and even other political parties."• Piratenwiki functions as "the information and coordination platform of the Pirate Party Germany.Here, content that has been previously constructed through other means (for example, in Piratenpad) is collated and archived-it is, so to speak, a kind of Wikipedia for the Pirate Party.On the Wiki you can find our principle programs, minutes of meetings of all subdivisions, plans for events, application links to party conferences, lists of roundtables and working groups, and much more" (www.piratenpartei.de).
As noted earlier, the Liquid Feedback software did not leave much room for discussion, only for alternative proposals (Paulin 2014).Discussions were deliberately excluded from the system of Liquid Feedback and needed to find their way to the PiratenPad online platform, which provides chatrooms, wikis, mailing lists or other forums that are not related to the party.This software ensures constructive discussion and decision-making: "You have quite a dynamic discussion process.You have to propose something, you can't say 'Everything is bad' [. . .].You have to say: 'I want it this way'" (interview with politician).

Results
Liquid Feedback makes it possible for political parties to improve the accessibility and transparency of decision-making and the quality of members' participation (Edwards and de Kool 2016).But although direct democracy is an important aim of the German Pirate Party, the results of the Liquid Feedback system are not binding for the party leadership (Paulin 2014).Sometimes the party leaders do make a different consideration than the members.This is an unresolved normative conflict in the party."Should a parliamentarian act as a trustee, accountable only to his own conscience (a norm constitutionally enshrined in Germany)?Or should the parliamentarian act as a delegate of party members, who can indicate their policy preferences through Internet devices such as Liquid Feedback, thereby implementing the equality between office holder and 'ordinary' members?The representatives in Berlin were most open towards the 'delegate model' and committed themselves to considering the membership position via Liquid Feedback in parliamentary decisions.However, member feedback is considered to be non-binding advice even in this case" (Bolleyer et al. 2015: 173).
Because transparency within the decision-making processes is highly valued, in principle members discuss their own standpoints out in the open.That was also the idea of the developers of the software.Edwards and de Kool (2016) saw this, however, as an important dilemma, and a very sensitive issue within the Pirate Party.Being transparent about party members' opinions and decision-making behaviour may generate information that those concerned regard as a violation of privacy.The Pirate Party therefore offered an opportunity to participate anonymously.The developers of the software (Jan Behrens, Björn and Andreas Nitsche) subsequently distanced themselves from the application of the software by the German Pirate Party in an open letter.They argue that the federal board has not installed the system properly.Anonymous participation complicates the process overall.The developers do not think that anonymous votes (with a pseudonym) and an accountable, transparent decision-making process can go together.
Liquid democracy has inspired other actors to use parts or variation as a system of collective, online decision-making.But the interviewee related to the German Pirate Party warned: Do not use the Liquid Feedback software for show.Nothing is more frustrating than participating in a process in which ultimately nothing happens with the result-this makes it very hard to motivate people to participate again, he emphasizes.In reaction to the Pirate Party and its use of digital technologies, Chancellor Merkel set up the 'Kanzler Dialog'."They set up an Internet website which was thoroughly programmed and which had a voting system.It was just clicks.You opened a new window in your browser and you could vote again.It was with cookies.One topic on this website was legalizing marihuana.What happens?Nothing" (interview with politician).This can damage the trust people have in politics.

Introduction
Officially founded in 2009, the 5 Star Movement (Movimento Cinque Stelle, hereafter 5SM) is currently Italy's largest political party with 14 European Parliament representatives.The meteoric rise of the 5SM in Italian and European politics has been closely connected to the movement's ability to organize and mobilize using web-based communication platforms.In this case study, we attempt to identify the strategies and tools used by the movement to connect online community-building with real-life mobilization in a successful bid for political power.Our focus in this study is especially on the phase before the movement's electoral successes, in which an online "fad" around the movement's charismatic founder was harnessed and structured into an effective political organization.
Online communication was central to 5SM from the beginning.The movement was co-founded by Italian comedian Beppe Grillo and Gianroberto Casaleggio, an entrepreneur acquainted with web strategies.Grillo has been the face of the movement, leveraging his existing public profile in a turn to political critic.At the same time, Casaleggio's communications company (Casaleggio Associati Ltd) has been the practical initiator and coordinator of the movement's digital infrastructure.Today the company also maintains the web platform used by 5SM on the national level.The movement's online presence is essential to its structure.Saebø et al. (2014: 244) cited movement members who describe the movement as a "non-association" with a "non-statute" and with "its headquarters on the web".The 5SM is a selfdeclared 'populist' movement (IlFattoQuotidiano.ti2013) with a "post-ideological" approach (Kirchgaessner 2016).The aspirations of the movement to give voice and influence to "people" come together in the movement's championing of "direct democracy", which for the movement is embodied in its use of a number of different online tools for communication and mobilization.
The web platform of 5SM combines different tools, which are used for a variety of purposes such as debate, knowledge sharing, voting, e-learning and more.At the time of our study (2016)(2017), all online tools rested on a common platform named Rosseau.This platform provides the resources for different services.One such resource, LEX, provides a space to share documents.Over the years, practices have emerged for how to use this space.These practices are key to how the movement works.One example is that elected officials representing the movement will share all reports that they must respond to with movement's members well in advance.This enables the possibility of an open dialogue process about how the movement should react to that report (interview with platform developer).This is one way that the movement seeks to operationalize the central populist tenet of exerting a more direct form of control over representatives than is traditional in representative democracy.
Another, perhaps even more central service is the voting function.Here, all registered and eligible members of the movement can announce their candidacies for upcoming elections, and the group of voting members decides who becomes listed as official candidates (interview with platform developer).Other services include e-learning modules designed to prepare newly elected candidates for their official duties and to introduce the tools at their disposal (interview with platform developer), as well as facilities for online streaming of official meetings.New services are added on a regular basis (interview with platform developer).
The use of online tools has been central to the movement from the beginning, but this has developed gradually.Only in recent years have the different tools become integrated in a common platform.The movement emerged around Grillo's blog, which he started in 2005.Here, Grillo continued a history of publicly sharing antiestablishment views and criticisms as part of his comedy.Once followers of the blog began to identify themselves as a political movement, they adopted Meetup1 as a means of coordinating small meetings among movement supporters.From around 2007, Grillo's blog supplied space for contact and debate within and between these groups (Saebø et al. 2014: 245).
The centrality of the blog, which until 2015 was still formally Grillo's, as well as the allegedly opaque internal processes of the movement, led to criticism of centralism and even autocratic tendencies in the movement (De Rosa 2013a).Conceding the point, in 2015, the movement changed the URL of the blog (to www.movimento5stelle.it).While the movement may thus have its "headquarters" online, it has combined this virtual structure from early on with "real-world" mobilization of supporters.In this way, according to De Rosa (2013b: 20), "the web [is] the connective tissue, the megaphone and the organizing principle behind a campaign that offers seamless movement between different reality spaces (online/offline)".Meetings and rallies have been central to the movement since its inaugural "Vday" in 2007, where Grillo began the process of gathering signatures for a proposed change to the Italian electoral law (Beppegrillo.it: 2007).At around the same time, some of the Meetup groups began to organize to run for local elections on civic lists162.This turn to local politics culminated in 2016 with the election of 5SM mayors in Rome and Turin.This turn also started the process by which the movement began to turn into a party (Kirchgaessner 2016).The "non-association" was formally founded as an association in 2009 to provide an organizational platform for the coordination of civic lists and to make it legally possible for movement members to run for the Italian Parliament and, from 2014, for the European Parliamentary elections.
Regarding the political context of 5SM and its online participatory tools, the choice to seek representation in political institutions has been a crucial milestone for the movement."The 5SM has started to come to terms with representative democracy, choosing to act from within political institutions" (Bordignon and Ceccarini 2015: 456).As already noted, this choice has been highly successful with 5SM now being represented in the European and Italian parliaments as well as many Italian municipalities (ibid.).Saebø et al. (2014) noted that while the movement has thus had to make concessions to the "offline" world of institutions, the statute of 5SM reiterates that "the Internet plays a crucial role for 5SM", that "the headquarters of the 5SM is the blog itself", and that the movement uses the Internet to "let citizens enter into the movement for consulting, deliberating, decision-making and electing purposes" as well as "governance" of the association (246).Italian regulations for party associations give extensive leeway for the design of internal processes and, thus, provide a legal space that the movement has sought to utilize to give as much direct influence to movement members as is possible and feasible (interview with platform designer).
The online platform of 5SM claims to be funded largely by crowdfunding (IlFattoQuotidiano.ti2013).The funds received by parliamentarians for support staff are directed back to the organization to support these and other communication efforts (ibid.).

Participants
As of January 2017, close to 1500 Meetup groups with more than 145,000 members in more than 20 countries had been established.In principle, anyone registering on the 5SM website can participate in the preparation and discussion of the movement's decisions, although they must stick to the debate guidelines.To be able to participate in the different voting activities of the movement (see below), voters must be 18 years of age and eligible to vote in Italy (interview with platform designer).To document their eligibility, subscribers must submit a scanned ID (Federici et al. 2015: 288).Despite this openness, however, Grillo has been criticized for eliminating online users unilaterally and for using his administrative position to "blackmail" movement voters into voting according to his own preferences (Saebø et al. 2014: 247).These criticisms aside, the participatory functions of the platform are extensive, and with the electoral results achieved by the movement in recent years to the European Parliament, the Italian Parliament, and local and regional Italian governments, the movement claims to have "one of the most effective platforms in the world" (interview with platform designer).

Engagement and Communication Strategies
5SM is in a sense synonymous with its engagement and communication strategy.From the perspective of the movement itself, members do not participate in activities of the movement; they are the movement, and the central administration and elected officials merely serve the movement.This high-minded philosophy of radical engagement is implemented in a number of ways, as outlined under 'Participatory process' below.
A central claim of 5SM is to be more open and therefore more representative than traditional parties (Federici et al. 2015: 288).There is no data gathered on the statistical representativeness of the movement's members in relation to the Italian and/or European elections (interview with platform designer).However, the movement website as of 2015 had 800,000 unregistered "followers" and around 100,000 certified "subscribers"-numbers which compare favourably with traditional party membership in Italy, where the most-voted-for party (PD) boasted a membership in 2015 of around 240,000 (down from 800,000 in 2009) (Federici et al. 2015: 288).Up to 30,000-40,000 subscribers to the website usually participate actively in important online discussions (ibid.: 294).
While statistical representativeness is neither a goal, nor a point of observation for the movement, the movement's claim to represent "people"-as long as this is understood in terms of representing a partisan, self-selected part of the population-could on the face of it be on par with traditional parties.De Rosa (2013a) documents that for the 2012 Italian Parliamentary election, the group of 5SM candidates-of which only 13% were female and only 10% were below 29-was in no way sociologically representative of the Italian population.Again, this is noted here with the caveat that the movement does not aim or claim to achieve statistical representativeness among its candidates.

Participatory Process
Participation and deliberation are central to the 5SM platform design.All content in the LEX and on the blog is thus open to debate among registered subscribers.Debate activity levels among subscribers are "good" in comparison with similar platforms (interview with platform designer).Movement members, however, have options for participation and co-creation that go beyond debate.These options are outlined here.
A first option is agenda setting.From the outset, the digital platform of 5SM focused on agenda setting.Participants can post about ideas and views and debate posts by other participants and party/movement officials.There is also an element of providing input for policy options in that posts can be in the form of policy proposals, and participants can co-create law proposals in wiki-style processes.Results are listed on the main web platforms of 5SM.Subscribers who have participated in debates and/or votes are notified by email about outputs and outcomes.A second option is voting.A central function of the blog has been as a platform for what the movement calls "digital primaries" (interview with platform designer).Saebø et al. (2014) noted that this use of the blog has taken various forms, flexibly adapted to each purpose.In one example, digital primaries conforming to direct democracy principles of "one user, one vote" were held to select candidates for 5SM national Parliamentary election lists, whereas, in another example, the digital primary took on a more deliberative democratic form when movement members had to decide on which presidential candidate to support (246).A third option, which as far as we have been able to determine is not yet implemented, is assuming direct control of elected officials (proxy voting).Registered blog users may propose legislation andon the condition of 20% of the online vote-compel the elected officials to put forward the proposal in Parliament (interview with platform designer).This has been a projected goal since the founding of the movement, but had by mid-2016 not been implemented, causing some internal criticism (interview with researcher).Such criticism would be natural, since the possibility to assume direct control over representatives is at the core of the republican populist tradition, which seems to be the movement's ideological ancestor (see above).A fourth option is the monitoring of elected officials.Adopting what De Rosa (2013b) calls a "no-confidence stance" on the relationship between the movement's members and its representatives, extensive direct monitoring of the actions of elected officials is possible.The blog thus acts as a hub for transparency with video streaming ensuring the transparency of meetings, also meetings with political allies and opponents (De Rosa 2013a; Saebø et al. 2014: 247).The funds received by parliamentarians for support staff are directed back to the organization to support these and other communication efforts (op.cit.), while simultaneously 5SM parliamentarians actively use social media to disseminate information about their activities.
The fact that subscribers have to submit ID to be confirmed creates a potential privacy threat to subscribers.However, in terms of privacy protection, the statute of the association contains provisions for privacy that prevent the association from sharing personal information with third parties (Federici et al. 2015: 292).Extensive security measures have been put in place to ensure the integrity of the movement's voting results (interview with platform designer).

Results
The impact of the largely online mobilization and organization efforts of the movement has been substantial in terms of providing voter support for elected officials.The electoral successes of the movement seem to prove the hypothesis that by providing an online infrastructure that gives voice to citizens, political initiators can in fact organize and mobilize voters.
In terms of policy impacts, we have no hard data on the relationship between, on the one hand, the online deliberations among citizens, the proposals produced by citizens and the proposal votes among registered subscribers, and, on the other hand, the voting and negotiation behaviour of the elected officials of the movement.We thus cannot conclude either for or against the assumption that the elected officials of 5SM represent more directly the political will of their constituents.This will likely be a theme for evaluation of the movement's government record.As described above, extensive information is made available to movement members to allow them to judge for themselves the correspondence between debate outcomes within the movement and positions taken by elected officials.

Introduction
Podemos was founded as a political party in 2014, following the anti-austerity Indignados protests.Podemos is represented in both chambers of the Spanish Parliament, Cortes Generales, and has five seats in the European Parliament.By all accounts, the influence of the party remains on the rise.In this case study, we look at the different digital participation tools the party has adopted as part of developing its organizational structure.Some of these tools provide a day-to-day online infrastructure for organization and participation in the Party, while others are used ad hoc to support specific political processes.
Podemos has made inclusive and transparent decision-making as the main asset and indicator of its politics.The basic claim is that direct participation of citizens is indispensable for a system to be democratic and legitimate.One of the main channels for public participation is the Podemos website,2 which enables online voting and decision-making.Since its foundation in 2014, Podemos has launched several online platforms based on open source software, such as Reddit (adapted by Podemos under the name Plaza Podemos3 ), Agora Voting (now nVotes),4 Loomio,5 Titanpad (now abandoned6 ) and, more recently, the UN-backed citizen participation platform Consul. 7The latter is the software behind the new platform to which Plaza Podemos has been moved.Generally, the use of different digital tools is initiated and managed by the political party Podemos.Plaza Podemos specifically was initially created as a "subreddit" (a dedicated forum on Reddit) by a supporter of the Party in 2014.The Party later contacted the creator of Plaza Podemos in its search for new digital tools to connect with people.As a result, the page became the official subreddit of the Party and the creator joined the managing team of the Party (Borge and Santamarina 2015).The aim of the digital tools is to organize direct democratic involvement, transparency and accountability through the application of direct democracy ideals in practice.This is done inside the institutional framework of the political party of Podemos.
Podemos seeks bottom-up support for Party activities through crowdfunding and fundraising.Podemos is mainly financed through donations and crowdfunding for specific projects by individual citizens.To achieve financial transparency and "corruption-free politics", it is possible to scrutinize the accounts and balances of the party online (Klimowicz 2016: 67).The main website of Podemos contains an opportunity to financially support the Party's activities directly by donating to the Party or its specific activities.
The combined toolbox used by Podemos lets participants engage in different phases of Party decision-making; from agenda-setting to policy deliberation and prioritization among policy proposals.In setting the agenda, Podemos uses the debate blog platform Osoigo, and Plaza Podemos, for open as well as focused debates, providing platforms to influence and deliberate on the political agenda.In providing input for policy options, proposals can be submitted on these same platforms or follow from debates held there.Whenever a proposal reaches a certain number of positive votes, the proposal is moved to the Podemos participation portal (participa.podemos.info),where it is processed among registered users as an introductory policy proposal.The user proposals that gain sufficient support are turned over in their final phases to a work group, which drafts a proposal document to be put to a vote.A work group consists of officials as well as the initiator(s) of the proposal, giving a very tangible element of co-deciding along with the binding vote that follows.

Participants
The target group of Plaza Podemos users is quite broadly defined.At 14 years of age, citizens can register in Podemos and gain access to the digital tools.However, some of the different platforms also contain opportunities to participate without registering, as well as a great deal of information on ongoing debates and topics of possible relevance to visiting non-members.Although certain citizen profiles may be more specifically targeted that others, Podemos and the digital participatory tools of the party address Spanish citizens on a broad scale: "That was the aim of this Party-you know-to reach to la gente, the people. . .So, in that sense, the target group was not, you know, a specific minority group or a specific section of the population, it was the whole society" (interview with a researcher).
Initially, this broad outreach strategy may have contributed to the popularity of and public interest in Plaza Podemos.Following the launch of the party and Plaza Podemos on Reddit, the novelty of a broadly inclusive and participatory forum for political debate gained media attention: "Plaza Podemos was very famous in the beginning.When they started in 2014, some traditional mass media (. ..) were following the debates in Plaza Podemos.It was very extraordinary in the sense that an online platform within a Party was given a lot of voice and a lot of echo in mass media" (interview with a researcher).However, as Podemos has become more established as a party, the Plaza Podemos platform seems to have become less influential.Although Podemos has gained great numbers of members and participants-estimated at around 400,000 users (interview with researcher)-media attention as well as participant activity on Plaza Podemos has decreased.It could be simply that the platform has lost its novelty.However, both our interviewees expressed considerations that something else might be at play: "It's not very lively now. . .I suppose because [there] was a lot of enthusiasm before, and now people are a little bit fed up because no proposal (. ..) has achieved to be voted on by the whole of the party. . .It was not accepted by the executives of the party.People, I think, are maybe disappointed in that sense" (interview with a researcher).
Podemos has also raised barriers to participation as an effect of increasing safety and Party control of the census of participants: "[. ..] they are becoming more restrictive, and so now this year you have to register as a member of Podemos and then you can participate in Plaza Podemos.But at the beginning, they were so open that it was possible to participate in Plaza Podemos and to have their say, even to vote. . .With only a nickname and without any restriction" (interview with a researcher).Until recently, Plaza Podemos users have all been handled as one census, including passive and active members alike.This has caused some internal discussion (Borge and Santamarina 2015), resulting in a distinction between the census in its entirety and a census of active participants.The point of this distinction was to qualify the group of people contributing to binding decisions inside the Party: "One of the main questions when you take into consideration decision-making mechanism is the census.If you don't have an accurate census, then it is difficult to give to this tool. . .binding decisions.And that was one of the main problems with Reddit: We didn't control the census" (interview with politician).This restriction, adopted in the period from 2014 to 2015, limits the right to participate in internal votes to only the active census.Problems of primaries being compromised by organized groups of right-wing voters also made higher-level security necessary (Borge and Santamarina 2015).In practice, the requirement for gaining voting privileges is connected to one's additional participation: "So, if in the last year you have done anything related to Party processes(. ..), you enter into the active census.(. ..)I think the requirement is to at least have participated electronically or physically once in the last year" (interview with a researcher).Being fully dependent on self-selection among participants, there is no mechanism of representative selection, and both our interviewees believe that it is skewed towards highly educated men: "In Plaza Podemos, the participants [. ..]I think they are not very representative of the whole Spanish population because I think that most of them are men for example, and I think that most of them are high-educated (. ..)But it is only an illustrative impression that I have" (interview with a researcher).
There is also a matter of digital and social divide: "I must say the majority is men and not all age groups are equally savvy with technology.There is also something about the economic status [. ..]The facility to have access to Internet and use it regularly, still nowadays-not only in Spain -depends on your cultural background and economic possibilities."(interview with a politician).In turn, the mechanism of self-selection and the generally flat structure of discussion can provide opportunities for minority groups to make themselves heard: "I think we are a in a way a party full of minority groups with a vocation of being a majority party.If you see data from polls and demographics and so on, we may have the most diverse base of supporters. . .Including maybe the strongest base of for instance disabled people" (interview with a politician).

Participatory Process
Participants provide input in several ways on the platform.First and foremost, participants can submit proposals to the deliberative platforms.Such proposals are subject to comments and being voted up or down by participants.In later phases of the decision-making process, participants can support a given proposal to get it accepted for proposal development by the organization and initiator.Anyone can post a proposal and the phases of making a successful proposal on Plaza Podemos are as follows: When a proposal has gained a certain number of positive votes (equivalent to 0.2% of registered users), it is moved to another section of the Participa platform.A proposal must then gain support from 2% of members before a notification is sent by email to all registered Podemos members.If the proposal then reaches 10% support within a set deadline (typically around 3 months), the proposal reaches the final draft phase.Here, the initiator(s) enter into a work group with party officials to co-draft a final proposal document, which is then voted upon by registered members.If agreement cannot be reached in this phase, two competing draft proposals are put to a binding vote on the Agora Voting platform.
"Agora is the software for online voting.(. ..)This is the tool that we use for primaries and this will be the tool that we use also for consultations.For instance, it is mandatory to ask our members about any pact or any agreement for instance to get or not into our government.Agora Voting is basically the interface that gathers the data from our census (. ..) to have online voting anonymously and all these guarantees" (interview with a researcher).Recently, the number of channels for participation has been reduced and simplified: "Apart from Plaza Podemos, they have Appgree [. ..].They have ListenToYou (Osoigo), which is a direct channel of citizens [to connect] with parliamentarians.They have these three channels now [. ..]One year ago, they had like, I don't know, ten different channels [. ..]Maybe it is smarter to simplify" (ibid).
In these ways, proposals are aggregated in several phases.First, proposals are voted up or down on the Agora Voting platform in order to gather initial support.Proposals that prove popular can then progress to a second platform, Participa, where only registered members can contribute.This accumulation of support continues as participants discuss, challenge, counter and qualify proposals.Successful proposals are ultimately aggregated-voted upon-in a binding vote.Between participating citizens, the "two most important proposals discussed [on Plaza Podemos] shows a high level of discourse equality, reciprocity, justification and civility" (Borge and Santamarina 2015: 33).Generally, proposals experience substantial discussion, suggesting an active group of participants.
When a proposal is sent for development in a work group, the author(s) of the proposal is included, and participants and Party decision-makers can exchange views in the process of developing the proposal.In this way, successful proposals from the online platform must see some interaction and deliberation between officials and initiators.The final vote is binding towards the political representatives, by which the aggregated view of participants is communicated to decision-makers.No initial information is given on the issues at stake, other than what the initiator submits to his/her proposal.However, information relevant to an issue can often be found because of the structure of the platform.Podemos itself merely provides information on the tools and processes, although our interviewee representing the Party expresses some concern about leaving documentation entirely to users: "I think we provide basic information about how to use the tool.I think when it comes [to] raising issues for discussion or for decision, maybe we should be more methodic. . .Explaining the context, the process, the pros, the cons and so on" (interview with politician).

Results
Results of the online participation of members contribute to the decision-making process in two main ways.First, the deliberation on proposals contributes to political agenda-setting in the Party.Second, the results of binding votes on proposals contribute to the policy output of decision-making processes.This output, however, is to be understood as policy aims of the Party.To be implemented as a policy, Podemos is still dependent on reaching majorities in relevant parliaments.Whenever a proposal from Plaza Podemos enters a representative democracy institution, obvious limits are posed to the political output efficiency of participation.Similar considerations are relevant when designing an online platform like that of Podemos.In the words of our interviewee: "If you don't give the possibility to have an actual impact on the outcome, then they are not interested in the mid-term. . .They lose interest, and we need to find ways to make it more binding and meaningful" (interview with a politician).
The interviewed researcher was more critical towards the political efficiency of participation on Plaza Podemos, and the feedback on the influence on policy: "In an indirect way, [Podemos] are paying attention to the discussion on Plaza Podemos, but it's in a very indirect way.[. ..]They take decisions without explaining themselves and without entering in Plaza Podemos and saying 'Well, you see, we are listening to you'.So, the listening is very indirect.It's not that those participatory tools have a direct impact.They have a kind of indirect impact that you can see between the lines in some decisions" (interview with a researcher).Regarding the different thresholds of support and the process of participant proposals, the interviewee continues: "The only way you can monitor [the process], is if the organizers tell you about it" (interview with a researcher).The researcher of Podemos finds that this seeming lack of political efficacy could be influencing the engagement of members: "I think that in the case of Podemos, people are very disappointed with these instruments.They were very fond of them at the beginning and now I think that the registered members are kind of. . .Well, they are abandoning.(. ..)If you enter in Plaza Podemos you will see that participation is very low.And if you compare this to Plaza Podemos in 2015[. ..]That was incredible" (interview with a researcher).

Introduction
In 2006, the Brazilian city of Belo Horizonte took a decisive step forward in moving the Porto Alegre model of participatory budgeting (PB) online.The Belo Horizonte digital participatory budget (DPB) is a process for discussing and voting on preselected projects proposed by City Hall.During the DPB period, participants can post comments in five discussion forums on the DPB website as well as vote on the proposed projects.Each citizen can vote for one project in each of the municipality's administrative regions (Coleman and Sampaio 2016).Through a platform accessible not only via the Internet, but also via voting kiosks and voting by phone, the city managed to engage 10% of the voting population in decisions on budget allocation.The participatory platform, however, was exclusively digital, meaning that voting would only take place online.Voting kiosks were set up to make ICT hardware available to all, as well as to provide guidance: "To minimize problems related to digital divide in the project, the City Hall established several voting kiosks throughout the city.Associations' headquarters, cooperatives and schools were also listed as official voting locations" (Sampaio et al. 2011).
The city has applied participatory budget processes since 1993, following the Porto Alegre PB model, which is based on physical deliberation meetings and neighbourhood representatives collaborating with city officials.In this model of annually recurring PB, online tools play a secondary, supplementary role.But for 2006, the Belo Horizonte municipality decided to add to the Porto Alegre model by having a parallel PB process that would run entirely online, that is, a Digital Participatory Budgeting (DPB) process.In this parallel process that ran on a dedicated website "any citizen with his or her voter's registration number from Belo Horizonte could choose 9 out of 36 projects (being one project per region), pre-selected by the City Hall, and by the associations linked to the PB" (Sampaio et al. 2011).The funds for the DPB were additional to that already dedicated to the ongoing Porto Alegre-style annual PB process.In 2006, "City Hall would invest US$ 11.25 million in its nine regions [while the] US$ 44.2 million budget for offline PB were maintained" (ibid.)The first application and creation of the tool was initiated by then Mayor Fernando Pimentel of the Workers' Party (Partido dos Trabalhadores) and the Belo Horizonte municipality were involved in all phases of the (DPB) process.City officials were not, however, involved in development or management of the website, which was performed by the communications agency Nitrato (ibid.).
After the initial DPB in 2006, further changes have distinguished it from the original Porto Alegre model.From 2008 onward, the process no longer focused on issues pertaining to specific neighbourhood but focused instead on a single issue on a city level.Here, "voters from Belo Horizonte should choose one among five preselected projects.All the options referred to road projects, with the goal to improve Belo Horizonte's traffic" (ibid.).Again, all project proposals were preselected without citizen involvement.Later applications of the DPB in Belo Horizonte were undertaken in 2011 and 2013 (Coleman and Sampaio 2016).The Municipality of Belo Horizonte finances both use of the DPB web tool and the implementation of policies following the process.The distribution of funds via the DPB is based on an index created by the municipality in partnership with a local university "high priority areas receive more funding, in order to stimulate economic growth and social prosperity" (Sampaio et al. 2011).
As a local government initiative, there are some legal constraints as to what sort of decisions are delegated to the citizens.However, this is no different from the legal limitations that the ordinary participatory budget of the city functions under: "the legal framework is kind of important in this sense that [. ..] there are some limits as to what City Hall can actually do.[. ..]So, people could not vote for, let's say, laws or bills or something like that.[. ..]As we had the face-to-face PB for several years, I guess it was not a big concern for anyone" (interview with a researcher).

Participants
In the design of the DPB process, openness to all citizens of the city was prioritized over sociological representativeness.Thus, in contrast to the Porto Alegre model, no selection mechanisms were applied to engage participants in a representative way.Nor were voting results weighted to reflect the demographics of the population.Instead, efforts were made to ensure that voters were actual citizens.
The general target group of the Belo Horizonte DPB includes all adult citizens with a voter registration ID.This automatically means that participants must be a minimum of 18 years of age.In comparison with the original Porto Alegre model for PB, in Belo Horizonte, it has been an aim to broaden participation beyond those already affiliated to the PB, and make participation more inviting, specifically for young people and the middle class-both groups that were underrepresented in the face-to-face PB processes (Coleman and Sampaio 2016): "We wished both to promote the expansion of popular participation and extend the participatory budget process to segments of the population that usually don't get involved, such as the middle class and youth" (Veronica Campos Sales, PB coordinator, personal interview in Coleman and Sampaio 2016: 5).This is supported by our interviewed city official who was affiliated with the 2006 and 2008 DPB's: "We were interested in increasing involvement by the population.We had already the presential participatory budget (. ..).The main target groups [of the DPB] were youth and middle class, who did not take part in the presential process." Thus, citizens of Belo Horizonte with a voter ID could already register to participate in 2006 (using their voter ID).In 2008, participants were also required to type in a random series of numbers (CAPTCHAs) as a security measure against computer bots.Although thorough registration processes were in place in 2006, security requirements were introduced by legal officers of the state in response to allegations of corrupt voting practices in the 2006 and 2008 DPBs (Coleman and Sampaio 2016).In 2011 and 2013, requirements for registration were supplemented by the download of a security app, dual voter ID (electoral and personal), email confirmation and questions on age and gender (Coleman and Sampaio 2016).
The number of participants has varied over the applications of the DPB tools between 2006 and 2013.The population of Belo Horizonte is just under 1.5 million.In the first DPB in 2006, well over 10% of the city's population, precisely 172,938 people, participated online.This resulted in 192,229 website visits and over 500,000 votes.In later applications, participation has fallen significantly from 124,320 citizens in 2008to 25,378 in 2011, and 8900 in 2013(Coleman and Sampaio 2016: 6)."It was a huge success in this first edition.I mean, the whole city was commenting about these topics and ideas.And you could see in the streets, people engaging and mobilizing for this case.(. ..)You have a tipping point after 2008.Because it was considered by everyone as a very good success and now today you see the same program with 100,000 less participants.There was a huge failure between 2008 and 2011" (interview a with researcher).
What happened between 2008 and 2011 to explain such a steep decline?The scientific researcher interviewed points to one particularly damaging error: the failure to implement the 2008 project chosen by the voters: "In 2008, the framework was just about choosing one single work in the whole city.And in the end, this work was not carried out" (interview with a researcher).In the researcher's view, this meant that citizens in general and participants in particular, lost trust in the process: "When the 2011 process started, people felt very frustrated about the DPB. . .Because they had mobilized, they had, you know, made everything to gain this work, to win this election.And then the work was not carried out.So, most people did not want to participate any more.Even though City Hall tried to explain that it was not its fault that the work was still to be done, but in the end, people did not trust the process anymore" (interview with a researcher).
The application of the online platform was first and foremost an attempt to involve younger citizens and the middle class, who were disaffiliated with PB.The main reason given for this was the assertion that these groups simply found participation in the face-to-face process too extensive and time-consuming, and that minimizing this would increase engagement: "The digital participation, we thought, was already the instrument to bring in the target group" (interview with a city official).On this basis, the process was simplified to allow quick and easy voting, although maintaining channels for more extensive participation as well: "If somebody wanted only to vote, he could vote in 30 seconds, but if he wanted also to participate, to discuss, to give his opinion, to lobby for what he thought was the best option, he had many instruments [online] to discuss with City Hall and also to promote discussion in-between citizens" (interview with a city official).However, no data was collected in these years to support conclusions that the DPB succeeded in engaging citizens in the target groups.

Participatory Processes
In 2008, two online features for further participation in the DPB were implemented.One was a chat feature, where citizens could contact city officials responsible for the PB.The other was a comment section on each project, where anyone (registered user or not) could submit comments (ibid.).
The five discussion forums make up the foremost platform for exchange of views.In 2006, 1209 messages were posted.A study of two of these forums (including two issues from each forum) indicated that only around 30% of messages were dialogical, while around 70% were monological (Sampaio et al. 2011).This suggests a somewhat low degree of "deliberativeness" in the forums.City officials and incumbents did not participate in these forums to encourage deliberation nor to answer citizens' questions: "People would address the City Hall in the questions, 'Why don't you do this work instead of that?', 'Why don't you select works in my region?', and the City Hall would just leave these messages unanswered.This was very frustrating for people" (interview with a researcher).
The 2006 DPB website provided basic information about each of the projects available.This included information on location and costs, as well as pictures.In the 2008 process, which dealt exclusively with city infrastructure projects, background information included visualizations of the roads "before" and "after" reconstruction, as well as informational videos.Virtual maps were also employed to visualize project locations, as well as where to find voting spots around the city (Sampaio et al. 2011).

User-Friendliness
Generally, our interviewees expressed satisfaction with the user-friendliness and efficacy of this tool for online participatory budgeting, and dissatisfaction with the scope and ambition of its application.The technological simplicity of the tool is considered to motivate more participants to take part and make participation more efficient: "The main strength is that it was easy.Digital democracy platforms can be kind of hard to understand, or even demand too much, especially if it's based on comments-that you have to comment, to read, and to respond-it is not everyone that has the time or the energy to participate.So, it was kind of a platform for, let's say, ordinary people who wanted to say what they prefer" (interview with a researcher).
This consideration runs parallel to the initial idea that minimizing the inconvenience of participation would increase participation among the specific target groups.One could also view this in the context of citizen empowerment versus simplicity of use: "The technology was very, very simple (. ..), but the empowerment of people, on the other hand, was very high" (interview with a researcher).This seems to mostly be the case concerning the voting process and less so regarding the deliberative processes: "I think we had a weak participation of citizens in terms of discussing which public works should be voted.This is the weak part of the process" (interview with a city official).DPB voting results are intended to become policy.This occurs in most instances, but "The [only project] voted in 2008 has not yet been finished because there is a problem in terms of land use and land property -it's a legal problem that involves federal government and local government and this has not been [finalized]" (interview with a city official).Although the outcome of the process is in this way a disappointment, participants did influence the output of decision-making.The failure to deliver their choice lies in the configuration of the successful project: "It was not that [the DPB participation] did not influence the decision making. . .Yes, it did influence, but later on, City Hall had administrative and legal problems [in the implementation]" (interview with a city official).This may explain a measurable decline in feelings of political efficacy among participants.Expressions of external efficacy in posted messages fell from 40% in 2008 to 20% in 2011 (Coleman and Sampaio 2016).In total, 6.7% of messages in 2011 specifically mentioned the 2008 project: "I encouraged many people to vote for this Project of São Vicente Square and where is the work?It's only deception [...] Why should I vote again?" (Participant in a discussion forum on 28 November 2011, DPB 2011 as quoted in Coleman and Sampaio 2016: 11;Sampaio et al. 2011).
In evaluating the methodology of the tool, one interviewee emphasizes the limitations in political scope of participation in the DPB: "I think the participatory budget process in Belo Horizonte should change (. ..) into a process that discusses which city we want.(. ..)The methodology should discuss largely with citizens what are the public works because actually it is not 'what are the public works' but 'which is the city that we want?'" (interview with a city official).This in turn leads to the conclusion, that although the DPB in Belo Horizonte has had some positive effects, it has done little to establish new trust in political actors and institutions: "I think the tool has not contributed in terms of changing trust of the target group in local politics.I think what they thought before is the same they think after the process" (interview with a city official).Factors that are emphasized as inhibitory in this regard are the focus on city planning projects (rather than policy) and especially the breach of trust that occurred when not implementing the 2008 project.To our other interviewee, the latter indeed makes the Belo Horizonte DPB most illustrative as "an example of what not to do" (interview with a researcher).

Introduction
In 2014, Paris began a multi-annual process in which Parisians are invited to participate in the distribution of an annually increasing share of the city's budget.The general purpose of Budget Participatif (BP) is to promote democratic innovation in the city, giving citizens the opportunity to participate in decisions that affect their daily lives.BP should be "an inclusive device, (. ..) a tool against social inequalities" (translated from Mairie de Paris 2016).Citizens can be involved in the phases of agenda setting, giving input for (and co-selecting) spending options and in deciding by public vote which ideas are to be implemented.The participation process runs at four levels: the City, the 20 arrondissements (district level), some low-income neighbourhoods and all public schools at primary, college and lycées levels (Cabannes 2017).
In the first year, 17.7 million euros were spent on the BP.Parisians were able to vote on 15 projects designed by the City of Paris (Budget Participatif 2016).One year later, in 2015, Parisian residents were able to submit proposals themselves.They initially submitted more than 5000 proposals.A process of preselection reduced the number to approximately 600 projects, out of which 8 were selected at the level of the city for a total budget of about 35 million euros.In addition, 181 less expensive projects were selected at the level of Paris' 20 arrondissements for a total budget of about 59 million euros (Holston et al. 2016: 6).In 2016, the total budget reached a 100 million euros (compared to Paris' overall budget of close to 10 billion euros).An innovative aspect of the Paris participatory budgeting project is that an overall value of 500 million euros for the whole 2014-2020 period was announced; that way trying to help to build confidence between the city and the citizens of Paris.
Participants can propose their ideas on the BP website.When submitted, proposals are available for comment and support by participants.A permanent team of nine people conducts day-to-day activities of the participation process.This team also connects with many other employees of the city of Paris administration, which help out in different ways.There are different commissions at the district and city levels that are responsible for the preselection of the proposals; they evaluate all the proposals on cost, feasibility and eligibility.Both commissions are composed of an elected body and a citizen body.There are four criteria for selection (decided by the City, not by citizens): the project needs to (1) be proposed by a Parisian, meaning a resident, (2) satisfy general interest, (3) be part of the city's responsibility and (4) the running costs of the investments related to projects need to be limited and primarily should not imply generating a public job (Cabannes 2017).After the selected projects are published online, candidates can run promotional campaigns for their ideas.The city provides an online toolkit to make flyers, posters, social media posts and organizes in-person workshops to help the candidates.Each project is presented during city sponsored events.In the end, the proposals are put to a public vote on the BP website as well as on paper ballots.Participants have 10 votes, from which they can assign five to city-level proposals and five to district level (interview with an organizer).Following the final selection, citizens can monitor the implementation of winning projects on the website.
The participatory process is closely tied to the political and administrative institutions of Paris.First, online participation takes place on the city's web platform.The city is obliged to process citizen proposals and to follow the results of the voting processes.Following the first PB-presented by Mayor Anne Hidalgo in 2014-the application of participatory budgeting was approved by the Paris Assembly as a rule of law.This means that the next mayor of Paris cannot reverse the process of participatory budgeting unless the Paris Assembly does so (Napolitano 2015).In this way, Paris is legally committed to budget participative "every year until 2020.And if we get a new mayor, the new mayor [must] still ask the council to vote on the cancellation of the participatory budget" (interview with an organizer).This legal framework is especially relevant as it depoliticizes the use of Budget Participatif and commits the city to its outcome.As the interviewed journalist points out: "So now that it's a rule of law, the next mayor cannot just decide to scrap the participatory budgeting.It has to go through the assembly of the city again.(. ..)Now, it's a thing of the city and not just of the mayor.(. ..)I think it contributed very positively in this sense." Procedures and responsibilities are formalized and communicated by the City, often on the website.As Paris is divided in 20 local districts (arrondissements), a lot of the groundwork to engage the public is delegated to districts.The district mayors are not all a member of the Socialist Party of Paris Mayor Anne Hidalgo and could thus be thought to be less enthusiastic about the participatory budget.The Hidalgo administration solved this by financial motivation: "We actually made a very interesting deal with them, at the very beginning, which was that any district that would get into the game will have their local budget doubled. . .So they're all playing the game, and are actually very involved" (interview with an organizer).

Participants
Budget Participatif as a participatory tool targets all residents of Paris.Anyone can participate by way of self-selection-regardless of age and nationality.There seems to be an acceptance that not all citizens have access to the Internet or are comfortable using the platform for one reason or another.Some district city halls and public buildings have offered personal help and/or workshops (Napolitano 2015).
And Parisians do participate.Around 40,000 citizens participated by voting in 2014 and around 67,000 voted in 2015 (Joignot 2016) and 92,809 in 2016, which represent 5% of the total urban population of Paris (Cabannes 2017).This "is a lot compared to other cities but not that much compared to the entire Parisian population" (interview with an organizer).To put it in perspective, 1,018,280 Parisians registered to vote for the local elections in 2014, of which close to 600,000 voted.

Communication and Mobilization Strategies
The tool is marketed broadly and has seen a lot of media attention, especially in its first year."We do have a lot of press, but it's a concept that seems to be hard to explain.So, it's not that easy" (interview with an organizer).The advertisement from the city's hand has focused on visibility on city streets and online, as the interviewed organizer points out: "[We use] big advertisements and posters in the streets, because the city controls that.So, we use it a lot, and it's a big, big budget.[. . .]We started to buy some online advertisement which is quite new for the city" (interview with an organizer).
In 2014 and 2015, around 60% of the voters choose to use the online tool (Napolitano 2015 and interview with a journalist)."For me, that's still a lot of people voting offline" (interview with an organizer).The backup of an offline voting method, our expert interviewee believes, prevented critique on the web solution: "Probably, if it had been only an online vote, there could have been some kind of [public] discussion, but the fact there was also a paper ballot meant also that the people that didn't feel safe, instead of complaining, they said 'okay I'm just going to vote on paper'" (interview with a journalist).
It is an explicit goal of the tool to involve participants beyond the political or participatory elite.It seems to succeed in doing so.Whether this is caused by advertising the process and tool, by design characteristics or other factors is arguable.Some design characteristics that lead to more inclusion can be pointed out.First of all, the parallel process of physical voting and project submission is implemented to prevent a digital divide.The initiative to offer advice and support could qualify layman participation in comparison to that of elite actors.The same applies to the organization of project workshops.Finally, registration is kept very easy: "We don't want any obstacle on the participation and we know that the more detailed the form is. . .Like to register to vote. . .The less votes we're getting" (interview with an organizer).In fact, identification is so limited that "you just have to declare on your personal honour, when you say that you live in Paris.And you give an address, but it's fully possible to fake it" (interview with an organizer).Combined, these and potentially other characteristics seem to lower the barrier of participation, promoting broader involvement-as well as empowerment-of ordinary citizens.
No specific mechanisms have been implemented to ensure representativeness, and it does not seem a primary focus."In my perspective, the participants are representative in the view that, like, all those people were allowed to vote and all those people were involved.Of course, I don't know the people who voted. . .Nobody knows" (interview with a journalist).Little data is collected to pursue or assess the representativeness of participants, due to legal constraints and considerations towards participation barriers: "Because the French law is very strict, we cannot ask questions about for instance race.(. ..) it would be interesting to target minorities, but we cannot do that" (interview with an organizer).Officials do have some general ideas of socio-economic and demographic characteristics, although the basis of this is limited: "You can tell from the part of Paris where they live, what kind of social [level] it is.(. ..)Where there is more poverty or young people, [people] participate a lot.But surprisingly, the conservative areas and very family working areas are also very involved. . .Like in the 15th district" (interview with an organizer).

Participatory Process
On budgetparticipatif.paris.fr,participants can submit proposals on city level as well as on arrondissement level.These proposals are featured on the website, where participants can then offer their comment and/or support.At a fixed deadline, the City collects the proposals and assesses them with regard to technical and legal feasibility.Participants can vote on all the accepted projects either online or on paper ballots.Ten votes can be cast, divided in city level and arrondissement level: "You can choose five projects in the whole Paris area and five projects in one particular district that you get to choose, like wherever, where you live or you can choose the district where you work. .." (interview with an organizer).
The participatory process has changed already in its first years.In 2014, all proposed projects were designed by the City.A year later, in 2015, the proposals came from the citizens and 10 million euros was earmarked for projects concerning schools.This, however, did not stop participants from focusing on another topic: "one thing we were not happy about last year was that a lot of projects were addressing the same topic.(. ..)Anything green, which is good, but we have a lot of money already dedicated on green in the regular budget, so it was actually too much" (interview with an organizer).
In 2015, when citizens were encouraged to submit their own ideas, 5.115 proposals came in.Around half of these were rejected by the administration due to technical, legal and social criteria.A total of 71% of these proposals were sent in by individuals and 29% by associations, companies, etc. (Budget Participatif 2016).A lot of effort has been put into teaching participants the formalities of proposing projects, and the 2016 process has seen a greater success rate in meeting the criteria (interview with an organizer).Lessons from the 2015 process may have been illustrative in this regard."One of the main difficulties in (. ..) involving citizens in the process is that people usually-because it's not their job-don't know about feasibility of projects, they don't know what the city can do, what the local council can do, and also what the city cannot do, because sometimes it's a competence of the county or the region" (interview with a journalist).

Aggregation
The input of participants is aggregated in at least two ways.First and foremost, the process of voting is one of aggregating support for citizen projects to decide which to implement.In the phase of project assessment, there is also room for the City to combine, pair and interpret proposals, being another form of weighing and aggregating citizen input.In 2016, this phase was opened up for citizen involvement."We made a big effort to involve citizens into the merging phase and really encourage them to come and defend the project together.First of all, that way we have less projects to deal with, but also, we will have more comprehensive projects.And finally, because we need people to get more involved in the campaigning phase" (interview with an organizer).

Deliberation
Interaction and exchange of views between participants is encouraged by the design (e.g.comment and support options).Between participants and political decisionmakers, the exchange of views is arranged indirectly.The participants' proposals (including qualifying comments, etc.) are not given to the decision-makers, but to the broader population of Parisians to vote upon (albeit after being processed by the City Administration).The results (representing the aggregated views of the citizens) are given to the political decision-makers after the vote, who are obliged to adopt the results as policy measures under the constraints of the budget (Plesse 2014).

Information on the Process
Information is provided by the City of Paris in several phases of the BP process.First, the website provides a fair share of infographics, FAQs and information, which explain the process of BP and how to participate.In the proposal submission phase, information regarding the legal framework and support for financial approximations is provided by the administration to the submitting participant(s).After the submission of proposals, the reasons of rejection are explicitly explained to the owner of the project by the administration.
Furthermore, distribution of decision-making responsibilities is clarified in the thorough information material on the website and the different phases of the decision process are documented on the website.Participants and other interested readers can also monitor winning projects on a devoted section of the BP website, where progress is reported.Throughout the process, participants can give feedback on the result and impact of their contributions.Participants who have submitted proposals are informed by email about acceptance, changes, regrouping, rejections and the reasoning or regulation behind it that may be relevant to the proposal (Budget Participatif 2016)."We've done a lot of data visualization, videos, there is this chart where the rules are written. . .There is a big part of the website doing pédagogie.(. ..)We do a lot of meetings, explanations, drawings and everything" (interview with an organizer).

Results
In the first phase of public participation, citizens contribute to the proposals, from which the acceptable ones (with regards to technical and legal feasibilities) are selected and put to a public vote.In the second phase of participation, the participating citizens fully decide by way of voting what is to be the political output of the process.Voting results on city level and for the arrondissements are communicated publicly shortly after the vote.
Most of the proposals concern quality of life (like renovating micro public spaces like a street corner, a garden or a wall), transportation and mobility (like safety of pedestrians and alternative forms of transportation to cars) and the environment (like green energy, educational programmes on recycling, etc.).In 2017, the city started to encourage development of projects around less popular topics like sport, cleanliness, security, housing, solidarity, cleanliness, new technology, economy and citizen participation.
Citizens are not generally involved in the implementation and thus can merely monitor outcomes: "It's the city that's going to do the projects" (interview with a journalist).This is also reflected in the distribution of funds to winner projects.The city has pre-set a budget for each project and pre-distributed funds to the different arrondissements dependent on the number of inhabitants (interview with organizer) and on social development: "They didn't give the same amount of money to every local council, to every arrondissement.They gave more money to the underdeveloped areas, so the richest part, the central part, they get very little money comparatively" (interview with journalist).This priority is surely an expression of a social political position but can also be seen as a means to gain higher effect on development-or incremental value-of public investments.
The BP is not only thought of as a tool of ideation and voting-another level of abstraction is apparent: "For us, participatory budgeting is just a tool of trust, and a symbol and a demonstration of a new way to do politics.So that's something we don't measure.I mean, we measure the trust people have in [Paris Mayor] Anne Hidalgo and it's working well, but we don't know if it's related to the participatory budget" (interview with an organizer).Thus, institutional and incumbent trust is a key consideration.But the attraction of the process as well as the (digital) tool towards younger citizens is highlighted: "The main strength for me is transparency and also reaching a new audience for democracy and renew the population, that is a success for sure.If you go to a traditional local meeting with only retired old people, and you compare it now to the number of children who participated at public schools, it's striking, so that's an amazing success" (interview with an organizer).
city didn't have to change any policies.So, they wanted to do this project, but they would have had to change a lot of policies to run it themselves.[. ..]They just let this non-profit, the developers, run it and manage it.And they sent someone from the city to gather the top ideas that came out of that every month.And then that person physically put them on the agenda of the individual city boards that would address those ideas.So it was kind of a run-around of the regulatory environment of the city to take advantage of this new type of engagement."

Participants
Since its establishment, Betri Reykjavik has had over 22,800 unique visitors and now-October 2018-has over 23,000 active users.The tool is designed to be broadly applicable and relevant to all citizens of the municipality-be it stakeholders, experts, officials or laymen."One of the interesting things about the Better Reykjavik project, especially in its early days, is that they were targeting anyone who was literate and had access to a computer" (interview with a researcher).Participants register using their electronic citizen's ID.They do not go through any sort of selection.Apart from a minimum voting age of 16 years, no characteristic (dis) qualifies ideas or participants.In other words, the selection of participants is strictly self-selective.

Communication and Mobilization Strategies
The website was initially advertised broadly around the city and online.There was also a specific appeal to political parties to use the tool."This was a fairly highprofile project in the city.There was a lot of discussion in the mass media, so it was in the newspapers, on the radio. . .Of course, there was a social media campaign.They also did outdoor advertising, I think, so there were posters up.(. ..)It was a legitimate, serious effort to pull people in, I think" (interview with a researcher).

Representativeness
The tool does not weigh votes for representativeness.In fact, there has been little consideration regarding representativeness-the focus in this regard has been on mobilizing participants with applicable interest, knowledge and opinions (interview with an organizer).Some thought has gone into assessing representativeness in retrospect: "In our participatory budgeting system, where people have to vote with an electoral ID-card, we get anonymized, demographic-style data.Gender, age groups and things like that" (interview with an organizer).The expectation from the developers that an online tool would automatically attract young citizens was shot down: "Most problems are with young people.And we thought that after the first year, when we did the participatory budgeting, being an online tool and everything, that we would see a lot of action of 20-30 year olds" (interview with an organizer).The interviewed organizer finds this difficulty of engaging the youngest age groups almost a law of nature, as they often are not very interested in their neighbourhood: "When people start to have children, then it seems to be that the near neighborhood (. ..) becomes certainly a lot more important in people's lives. . .It's just one of those things, right?(. ..)You're never going to be able to have someone who's 20 years old, single (. ..) as interested as a 35-year old who has two children, you know?" (interview with an organizer).
There has been some attention paid to minority protection and participation on the Betri Reykjavik platform.This is done through the sorting of arguments-thus paying attention to the minority view-rather than the efforts of selection or mobilizing of participants from minority groups (see next paragraph).This is contrasted by the fact that the vast majority of the website is only available in the Icelandic language, making it difficult for some minorities to participate: "In Icelandic context, I think the biggest issue, the biggest barrier in a lot of cases, is the Icelandic language.So, the immigrant community uses mostly English but not everything was necessarily available in a language they could understand" (interview with a researcher).
It is an explicit focus of Betri Reykjavik to make participation in decision-making more egalitarian and inclusive to layman citizens.Design-wise, this is evident in "a direct connection with social media networks like Facebook and Twitter [which] reduces barriers to participation while situating policy discussions within the users' real social networks" (Lackaff et al. 2014).Participation is designed to be simple, quick and easy, adopting "like"-style mechanisms and limiting text contributions to 500 characters (Bjarnason and Grimsson 2016).

Participatory Process
Betri Reykjavik combines two main tools or platforms, tied to the two forms of processes.Your Voice at the City Council starts with a brainstorm phase, in which participants deliberate on the political agenda by proposing and debating policy ideas.The technology behind this is Your Priorities, an open-source idea generation platform.Here, participants can submit ideas in text and graphics, informally vote ideas up and down and debate ideas by placing arguments for and against.The arguments are sorted into columns to keep the debate on track and relevant to the proposal.Finally, after the most popular projects are determined (10-15 every month), citizens can have a final vote on the projects on the website using Open Active Voting software, which is also open source (interview with an organizer).New versions of both platforms are released on a somewhat regular basis in order to enhance user-friendliness and lower participation barriers: "We are always working on simplifying the process, in terms of how to participate.And that, I think, is in general a weakness of participatory processes, that they can be too complicated" (interview with an organizer).
During the online citizen deliberation on project proposals, the diversity of views is managed by allowing equal visual impact to both sides of the debate.The most popular arguments against any proposal are presented at the same level as the most popular arguments in favour of the proposal, regardless the distribution between the two sides (Lackaff et al. 2014): "What we tried to do was to split the screen in two so people who support the idea can write points for it on the left side of the screen (. ..), and on the right side of the screen, people who are against the idea can put their points. . .And almost overnight (. ..) the quality level of the debate increased a lot" (interview with an organizer).Thus, views are exchanged between participants strictly by arguing for or against proposals.This minimizes the extent to which a comment can refer to another comment rather than the proposal itself: "If you see a point, you don't agree with, there's no way to comment on it.You have to write a counterpoint" (interview with an organizer).Both our interviewees emphasize several advantages to this design: • Firstly, securing that minority views can gain attention, and "make sure that the minority view is heard.When you go to the website you can see the idea, then the best points for the idea are at the same level as the best points against it.Even if there's a thousand people who support the idea and only 10 that don't support it" (interview with an organizer).• Secondly, user-friendliness in the sense of debate overview for newcomers: "The thing that I like about it is that after a while, when a new participant comes to this particular issue, they can very quickly get a sense of the thinking on it, that's gone on previously" (interview with a researcher).• Thirdly, user-friendliness in the sense of minimizing the time necessary to participate: "Participants can participate very little or they can participate very much, depending on how interested they are, which I see as a strength" (interview with a researcher).• Finally, this way of designing the debate module works to improve ideas."This has been very helpful to help the city make the ideas better" (interview with an organizer).
It is clear from the dual process participation model which decisions are up to the City Council and which processes are subject to binding online votes.Reports of City Council meetings are publicly available.The public participation processes are public to everyone while ongoing and remain available in the website's archive until they are finished.
Participation results contribute (in some cases directly) to policy output.Citizens are informed about this output via the website platform of Betri Reykjavik and on the municipality's main website.The participants of a specific process are notified by email when there are developments in the decision-making process, as well as on implementation and later developments (Bjarnason and Grimsson 2016)."If there's an idea that is going into processing, people can track it on the website (. ..) and each time there's a status update, you know, when it goes into a committee and is discussed and there are meeting notes, they are sent to all the participants" (interview with an organizer).The interviewed organizer is quite insistent on the importance of proper feedback: "(. ..)And obviously at the end, when the idea is agreed on or rejected, then everybody gets an email as well.It's super important (. ..).Otherwise, you're really not respecting people's time."

Results
In the first 3 years of Better Reykjavik, over 150 citizen ideas were approved by City Council.Better Neighbourhoods has seen over 200 different project ideas being implemented, following an investment totalling 5.7 million euros from the city (Participedia.net2018).With Betri Reykjavik, results of the participatory process contribute directly to the formal decision-making process.This happens either by presenting and discussing the top proposals directly in the City Council's meetings or by providing construction projects, which are more or less co-created by citizens and civil servants, which are then voted on by the public and accepted by the municipality as binding.Of course, the results of binding votes are more directly influential on policy and/or planning output, but the participatory process in all cases is nevertheless directly tied to the decision-making process in City Hall or the municipal administration.Some reservations can be made regarding the genuine political influence of Betri Reykjavik: "The Reykjavik PB project is expertly focused on kind of small things.So, it's focused on, you know, park benches, parks, and repairs and things. . .So yes, within this kind of small domain, participants have a very direct impact.But it could also be argued that we're not talking about zoning or infrastructure development or what kind of. . .Bigger issues, that citizens might also be interested in" (interview with a researcher).In other words, the possibility of affecting city policy on a higher level-for example, an overall strategy of "what kind of city do we want"-is rather limited.
It is also important to note that even binding votes do leave some room for interpretation of decisions in the implementation phase: "The city does have some leeway to interpret the results" (interview with a researcher).Surely, this is even more so the case when it comes to the proposals and ideas from the ideation process Your Voice at the City Council being presented to City Hall.
Whatever political, strategic or planning level is in question, this tool does-as agreed by both our interviewees-seem to improve City Hall and administration decisions.Being the co-founder of the Citizen's Foundation, the organizer puts it like this: "That's really the biggest strength of using this tool: better decisions" (interview with an organizer).