The Responsibility of an Audience: Assessing the Legitimacy of Non-elected Representatives in Governance Networks

ABSTRACT Non-elected actors in governance networks are legitimate representatives when the constituency accepts their claims of representation. However, not all constituents have the resources to approve or oppose this representation. Consequently, I argue that the audience, often the decision-making authority, which enables non-elected actors to act as representatives has a responsibility to consider their legitimacy. Drawing on seven business and urban development networks in Norway, this article explores how the decision-making authority considers credibility, qualifications and connectedness to legitimise non-elected representatives in governance networks. Through interviews with civil servants and politicians organising and participating in the network, relevant documents and observations, this article demonstrates that the decision-making authority legitimises non-elected representatives based on credibility and qualifications rather than connectedness with the constituency. The decision-making authority believes that claims grounded in specialist expertise, self-representation and shared experiences with the constituency legitimise non-elected representatives. Similarly, truthful representatives are considered legitimate. Finally, the decision-making authority is divided with regards to how the interactive process between the non-elected representative and the constituency legitimises the content of the representation.


Introduction
In representative democracies today, elected representatives are no longer the only actors who speak on behalf of others (Knappe, 2017). Non-elected representatives are becoming a supplement to the formally elected representatives. Several non-elected actors from employers' organisations, trade unions, NGOs, celebrities, businesses, activists, and volunteer associations are taking a central role politically as representatives of affected groups, interests, values or causes (Maia 2012;Montanaro, 2019). Representation is more than territorial representation, and a representative is more than someone with a mandate from voters (Montanaro, 2017).
Involving these non-elected representatives in politics is considered a way of strengthening the representative democracy in crisis, beyond elections, political parties, voter turnout, and governing institutions (de Wilde, 2019;van de Bovenkamp and Vollaard, 2018). The increasing use of participatory arrangements, direct participation, public debates, and social movements in policymaking has created public spheres for nonelected actors to behave as representatives although they are not democratically elected (de Wilde, 2019;Guasti & Geissel, 2019a;Saward, 2010;van de Bovenkamp & Vollaard, 2018).
Non-elected representatives, who are exempt from electoral procedures, raise concerns about democratic legitimacy (Maia, 2012). Saward's (2010, p. 145) theory of representative claims suggests that non-elected representatives are legitimate when the constituency accepts the content of the representation. However, disadvantaged subgroups within the constituency may not have the resources to approve or oppose representation. I therefore argue that the audience which claims of representation are directed at, i.e. the decision-making authority, has a responsibility to consider the legitimacy of non-elected representatives (Chapman and Lowndes, 2014;de Wilde, 2019;Guasti & Geissel, 2019a;Knappe, 2017). Building on Chapman and Lowndes (2014, p. 288), the decision-making authority may ensure that accurate representation occurs through paying attention to the non-elected representative's credibility, qualifications and connectedness. Without these considerations, some constituents' interests might go entirely unrepresented (Montanaro, 2017, Ch. 3).
In governance networks, which are described as 'self-regulating horizontal articulations of interdependent, but operationally autonomous, actors from the public and/or private sectors' (Sørensen & Torfing, 2018, p. 304), Chapman & Lowndes (2014) find that non-elected actors function as representatives. Given that governance networks struggle with inclusiveness, it becomes essential to understand the participating elite or sub-elite in discussing networks as a supplement to representative democracy (Hendriks, 2008;Klinke, 2017;Sørensen & Torfing, 2018;Torfing et al., 2012). If these networks are without actors who are legitimate representatives, they might produce undesirable ideas, solutions and even policies (Torfing et al., 2009). Having a decisionmaking authority that assesses legitimacy of participating non-elected representatives can ultimately reduce the democratic problem of governance networks. This paper therefore explores how the decision-making authority considers credibility, qualifications and connectedness to legitimise non-elected representatives in governance networks.
The majority of studies on non-electoral representation are theoretical. The few empirical studies focus on justifications for non-elected representatives from the perspective of certain actors such as civil society groups, social movements or faith representatives. Empirical studies also investigate representatives surrounding elections and political parties, or they focus on non-elected representatives within a given policy area (de Wilde, 2019;Guasti & Almeida, 2019;Heinisch & Werner, 2019;van de Bovenkamp & Vollaard, 2018). While the number of empirical studies is growing, the function of the audience in legitimising non-elected representatives, which I intend to investigate in this paper, is an area sparsely examined (Chapman & Lowndes, 2014;de Wilde, 2019;Guasti & Geissel, 2019b). In the context of governance networks, empirical studies have so far only considered non-electoral representation expressed by the network actors themselves (Chapman & Lowndes, 2014;Derkzen & Bock, 2009;Hendriks, 2009;Torfing et al., 2009) or the democratic legitimacy of the entire network in terms of input, throughput and output legitimacy (Ayres, 2020). In this paper, I add an audience's perspective on non-electoral representatives in governance networks.
To understand how the decision-making authority legitimises non-elected representatives in governance networks, I explore seven networks associated with business and urban development. This study focuses on the decision-making authority's acknowledgement of organised non-elected representatives, the reason for this being that they often have more resources than individuals to influence policymaking (Montanaro, 2017). This resource-rich economic sub-elite, highly present in business and urban development networks, may also be motivated to promote self-interests or the interests of advantaged sub-groups (Dür & Mateo, 2014;Hendriks, 2009;Montanaro, 2017). To cover the diversity of organised actors, the networks selected include different types of organisations, e.g. labour unions, private businesses, non-profit foundations and employers' organisations.
The paper proceeds as follows: in the next section, I introduce the theoretical framework of representative claims, non-electoral representation and governance networks before describing the methods used to answer my research question. After that, I present the results of how decision-making authorities consider the credibility, qualifications and connectedness of organised non-elected representatives. Before concluding the paper, I discuss how the decision-making authority considers the legitimacy of nonelected representatives.

The Legitimacy of Non-electoral Representation in Governance Networks
The literature on representation often uses Pitkin's (1972) concepts of formalistic, descriptive, symbolic, and substantive representation originating from electoral procedures. These concepts are also used to study representation in governance networks (Derkzen & Bock, 2009;Hendriks, 2009). Saward (2010, Ch. 2) in his recent theory defines representation as the activity of making claims to represent someone. The claims of representation are constructed in an interactive process between the representative and a constituency where organisations and individuals make, receive, accept, or reject representative claims. This study is among the first to apply this theory to organised interests in governance networks, describing representation in networks as the outcome of the dynamic relationship between a governance network actor and the constituents these actors portray.
Within this framework, individual or collective actors make claims that they themselves or their organisation (subject) represent affected groups, interests, or causes in public through an event or a series of events in more or less formal political processes (de Wilde, 2013;Saward, 2010;van de Bovenkamp & Vollaard, 2018, p.101). Elected officials are therefore no longer the only actors who represent others (Knappe, 2017). Non-elected representatives are taking this role, reducing unequal participation and strengthening the representative democracy (van de Bovenkamp & Vollaard, 2019). Building on Schlozman et al. (2015) this study focuses on organised non-elected representatives (p. 1018). In the investigated networks, I distinguish between interest organisations which represent their individual members and supporters, and politically active businesses, universities or other institutions without a membership basis which represent any group or individuals affected (Berkhout, 2013;Binderkrantz, 2009;Schlozman et al., 2015). Using the theory of representative claim broadens the concept of representation. The theory allows organisations without members to represent interests and for membership groups to represent more broadly than their members (Berkhout, 2013;Montanaro, 2017;Schlozman et al., 2015). A labour union participating in the network may also represent non-unionised workers or those with membership in another organisation. Thus, stakeholders or members of the network participants are internally affected, while all those influenced by the network's outputs are externally affected (Knappe, 2017).
To function as a representative and potentially influence policy, one must be recognised as such by the targets of representation, i.e. the audience. Thus, for a nonelected representative to exist, the representative claim needs to be acknowledged by the audience (de Wilde, 2013; 2019). To further specify, I will base my definition of the audience on Guasti and Geissel (2019b) as the decision-making authority which listens to, includes and responds to representatives (p. 104). In a governance network, the decision-making authority rests with the civil servants and politicians initiating and organising the network, selecting participants and participating in the network. The decision-making authority may recognise actors because of their ability to represent others, but also due to their status, fame or position (Montanaro, 2017). Selecting participants for the governance network can indicate recognition of actors as representatives (Berkhout, 2013;Knappe, 2017). Whether politicians and civil servants include actors because of their representativeness and whether they have a responsibility to consider the accuracy of representation is explored in this paper.
A non-elected representative is successful when the audience considers the act of representation positively, but they are only democratically legitimate when accepted by the constituency of affected or potentially affected interests (Disch, 2015;Montanaro, 2017;Saward, 2010). Legitimacy also says something about the quality of representative claims. A good claim enables the constituency to be in favour of or opposed to the non-elected representative (de Wilde, 2019). To determine whether one supports the representatives, claims of representation must be explicit and rich in information. The necessary information consists of clarifications on accountability, justifications as to why the claimmaker represents a particular group or interest, and statements which create a collective political identity (Arnesen & Peters, 2018;de Wilde, 2019). Studying faith representatives in urban governance partnerships, Chapman and Lowndes (2014) suggest four indicators to determine the quality of non-electoral claims: added contribution of claims, authenticity, horizontal connectedness, and attributes and skills of the representative. Added contribution of claims refers to considerations regarding whether claims provide specific specialist knowledge, are dynamic over time, fluid without spatial boundaries, explicit for new audiences, and share values with constituents (Chapman & Lowndes, 2014, p. 286). Claims with these characteristics contain information on justification (de Wilde, 2019). Authenticity means how genuine the claimants are, the credibility of what is claimed, and whether the claims are made independently of formal electoral processes. Horizontal connectedness concerns the representatives' visibility towards, and their dialogue with, the constituency of affected groups, interests, communities, and organisations (Chapman & Lowndes, 2014, p. 286). Representative claims may include information on accountability to ensure horizontal connectedness (de Wilde, 2019). In other words, the representatives' responsiveness to feedback when formulating, explaining, and justifying claims (Maia, 2012). Finally, the skills and attributes of the representatives mean their communication skills, knowledge of and matching values with the constituency, openness, and their ability to accommodate input (Chapman & Lowndes, 2014, p. 286). In governance networks, as investigated in this article, legitimacy therefore refers to whether the constituency accepts individual network participants making claims and the content of their spoken claims.
Although representative claims must have a certain quality for the constituency to determine the legitimacy of the representative, not all constituents have the opportunity to voice their opinion when they decide the representative's legitimacy (Montanaro, 2012;van de Bovenkamp & Vollaard, 2018). Disadvantaged sub-groups in the constituency may not have the resources to express agreement or disagreement with the nonelected representative (Montanaro, 2017). Less privileged groups are also more unlikely to use organisational mechanisms such as leadership elections, withdrawal of membership or followership, meeting attendance and donations to control organised nonelected representatives (Fraussen & Halpin, 2018;Montanaro, 2017;Schlozman et al., 2015). Consequently, the disparity in power may result in non-electoral representation which is biased towards the most powerful in the constituency (Chapman & Lowndes, 2014;Knappe, 2017). Following Chapman and Lowndes (2014, p. 287), I, therefore, argue that responsibility to interpret the legitimacy of non-electoral representatives is placed on the targets of representation, i.e. the decision-making authority. This decision-making authority can, when paying attention to what is represented, secure that representation is accurate (Fraussen & Halpin, 2018).
Although Chapman and Lowndes (2014, p. 286-287) emphasise that decision-makers have a responsibility to recognise legitimate non-elected representatives, they do not develop their indicators with the audience in mind. Neither does de Wilde (2019). Building on this work, but adjusting it to the audience's determining the legitimacy in the field of business and urban development, I suggest that non-elected representatives are legitimate if representative claims are credible and the representative is qualified and connected to a constituency. These three factors are arguments that the decision-making authority may use to legitimise network actors as representatives. Credibility is defined as justifications of representation referred to in representative claims. It is operationalised using non-elected representatives' explanations of why they represent what they do. This argument is similar to Chapman and Lowndes' (2014, p. 286) indicator of the added contribution of claims. Representative claims are justified on the grounds of i) position in the society; ii) tradition and standing on moral issues; iii) specialist expertise; iv) shared experiences, values and identity; v) self-representation or vi) popular support and unheard perspectives (Saward, 2010, Ch. 4;van de Bovenkamp & Vollaard, 2018, p. 101). Position, expertise, competence or status are found to be reasons why decision-makers include non-elected actors in networks (Hendriks, 2008;). Decision-makers ascribe legitimacy to interest organisations because of their expertise and information about public preferences (Flöthe, 2019). Altogether, information is a valued resource among all network actors (Torfing et al., 2012). Qualifications legitimise non-elected representatives based on their attributes and skills. A representative with a character and representative claims that 'ring true' can qualify representatives and make them legitimate (Saward, 2010, p. 103-104). Decision-makers' trust in network actors is crucial for a functioning governance network (Klijn & Koppenjan, 2012). Studies show that decision-makers also acknowledge non-elected representatives who are truthful, charismatic, well networked and with good communication skills (Chapman, 2012;Chapman & Lowndes, 2014). Thus, qualification is operationalised as the network actor's truthfulness and collaborative character. The argument combines Chapman and Lowndes' (2014, p. 286) indicators of authenticity with the skills and attributes of the representative. Finally, connectedness is the relationship between the representative and the constituency. An important part of this relationship is the representatives' visibility, responsiveness to feedback and that claims contain clarifications on accountability. Non-elected representatives are responsive when they explain their behaviour to the constituency and adjust it in line with the views of the constituency (de Wilde, 2019; van de Bovenkamp & Vollaard, 2018). Connectedness is operationalised as the network actors' perceived interaction with the constituency and how this interaction affects what is represented in the network. This argument shares characteristics with Chapman and Lowndes' (2014, p. 286) indicator of horizontal connectedness. Governance networks, interest organisations and private businesses are all found to struggle with accountability (Aarsaether et al., 2009;Hendriks, 2008;Schlozman et al., 2015). However, decision-makers use high membership, supporter or even subscriber density, which indicates a large constituency, to legitimise organised actors. Internal democratic processes in interest groups signal accurate representation which may also have a legitimising function (Binderkrantz, 2009;Fraussen et al., 2015;Fraussen & Halpin, 2018). In sum, I have applied general arguments on the legitimacy of non-elected representatives to explore how the decision-making authority might use credibility, qualifications, and connectedness to legitimise actors in governance networks as non-elected representatives.

Methods
This study investigates business and urban development networks in Norway organised around the functioning of commercial and social activities in city areas. The networks located in Oslo, Kristiansand and Tromsø touch upon issues such as climate and environment, transportation, retail, tourism and liveability. Thus, they are essential in dealing with the conditions for how industry, tourists and citizens use urban areas. These complex cross-sectional issues are dependent on the coordination of public and private resources (Pierre, 2005(Pierre, , 2016. In this context, the cities of Oslo, Kristiansand and Tromsø organise governance networks to facilitate collaboration between public and private actors. These cities are selected because of their richness in available information, although all are regional centres in the northern, southern and eastern part of Norway, respectively. In this study, I explore how the decision-making authority argues to legitimise organised non-elected representatives in governance networks. Thus, I aim to contribute to a theory of audience acknowledgement of non-elected representatives. To cover the diversity of organised non-elected representatives, I study three types of governance networks which include different organised actors. The organised actors can be split into two groups: interest organisations with a membership basis and politically active organisations without members. The latter are private businesses, public entities such as universities or hospitals, business clusters and non-profit foundations. Organisations with individual members are mainly economic interest organisations such as employers' organisations, trade unions, chambers of commerce, real estate associations, and citycentre, retail, business and residents' associations (Castiglione and Warren, 2019;Schlozman et al., 2015). One of the networks includes organisations without members, another includes primarily membership organisations, and the third consists of a combination of organisations with and without members. The first network, which I have labelled the inclusive informational network, is comprised of businesses, non-profit foundations, hospitals and universities, the exception being an employer organisation and city-centre association. The second network named the qualified consensus network restricts participation to selected interest organisations. Finally, the third network, which I have called the professional action network, includes a mix of interest organisations and businesses. All three networks are present in Oslo. The two networks in Kristiansand take the form of inclusive informational networks, while a qualified consensus network and a professional action network are present in Tromsø. In summary, the seven business and urban development networks selected operate within the same context, that is they are organised around issues which affect the same group of people and similar departments and municipal agencies govern them. They are different in terms of types of organised non-elected representatives included.
Data is collected from multiple sources: interviews with decision-making authorities, relevant documents from the networks, and observations from network meetings. Interview data were collected between December 2019 and November 2020 from 26 semistructured interviews with actors from the decision-making authorities in Oslo, Kristiansand and Tromsø. The respondents are the civil servants and politicians who have initiated, organised, and, on several occasions, participated in the business and urban development networks I have investigated. They have had a say in the final decisions on the content of the networks, who has been invited to participate, the function of the networks, and how network discussions have been used in local policymaking. The respondents are four actors from national government agencies (directorates), one politician and twenty-one civil servants. The interviewed respondents were identified through established professional contacts, minutes from network meetings, searches on the municipal website, and through snowballing. Using semi-structured interviews allows me to investigate individual attitudes towards the concept of legitimacy in detail and to explore the reasons behind the respondents' views. The flexibility to adjust and add questions as the interview progresses is valuable to capture these theoretical concepts (Mosley, 2013). The decision-making authority was asked for its response on network actors' behaviour such as their actions, spoken arguments, opinion and comments expressed in the network. The interviews lasted for approximately 45 min. To supplement what several of the respondents already had explained in the interviews, I analysed relevant publicly available documents from 2015 to 2020. These 69 documents include meeting minutes, agendas, presentations held at the meetings and external seminars, and evaluation reports. The documents are listed in Appendix A. To enrich the data further, I participated in and observed two network meetings in February and March 2020. The sessions lasted between two and three hours and were organised as a combination of presentations and group discussions. The number of network members present varied from around 30 participants in the first meeting to 70 participants in the second.
Using NVivo, all the data sources have been coded according to the three arguments of legitimacy, i.e. credibility, qualifications, and connectedness.

Findings
The primary purpose of this empirical analysis is to understand how credibility, qualifications, and connectedness legitimise governance network actors as non-elected representatives. The results revealed that the decision-making authority believes network actors make representative claims. Thus, the network actors are non-elected representatives. This recognition applies regardless of whether the network actor is an organisation with or without members. However, there is a tendency among the respondents that organisations which do not have members to a lesser degree are acknowledged as representatives compared to membership organisations.
Furthermore, the decision-making authority does not view individual network participants as the representative. Instead, it is the whole organisation which is acknowledged as a non-elected representative; thus, what the individual participant communicates is considered the content of what the organisation represents. Therefore, the findings present how the decision-making authority legitimises organisations as non-elected representatives, although the respondents were asked about individual network participants. All network minutes present the network or the private actors in the network as unitary in their view on particular issues. However, when references to individual statements are made, the minutes emphasise the organisation the individual is affiliated with, rather than the person. In general, I find that public documents concerning governance networks are objective. The content therefore rarely reveals the decision-making authority's response to non-elected representatives.

Credibility
Credibility is the decision-making authorities' understanding of how representation is explained in claims of representation. The representative claims may contain references to organisational position, tradition and standing on moral issues, specialist expertise, shared experiences, values and identity, self-representation, or popular support and unheard perspectives for the decision-making authority to legitimise non-elected representatives.
First, the decision-making authority legitimises non-elected representatives in governance networks because they have specialist expertise. The respondents explain that this means they believe the representative's opinions or arguments contain relevant knowledge they themselves do not hold. The interviews and documents show that specialist expertise is most often used to legitimise non-elected representatives. Organisations with no members offer their expertise by describing employees', students' or customers' typical workdays. Membership organisations have access to their expertise indirectly through their members. Interaction with employees, customers, students or members gives the non-elected representatives an understanding of the needs of those affected by network outcomes. As the senior advisor in a municipal agency described, the representatives deliver first and foremost local expertise, they represent others that report back to them when things don't work. They also have knowledge about the industry, what trends occur, changes that happenvaluable information for the decision-makers.
The decision-making authority emphasises that it cannot handle the professional expertise required within business and urban development in terms of complex issues such as transport, logistics, retail, tourism and real-estate development by itself. The respondents agree that they need experienced private actors to provide them with insights into the challenges they might face as a result of existing policy. Network minutes and reports show that the non-elected representatives repeatedly are encouraged to present their experiences in the network as input to policy.
Secondly, the interviewees legitimised organisations without members as representatives based on self-representation. These non-membership organisations are seen as nonelected representatives of their own interests, which are not heard elsewhere. Using stories from employees', students' or customers' workdays to justify why they oppose or support an issue discussed in the network also contains an element of self-representation. In a policy area such as business and urban development, which is often driven by economic interests, it would be strange if one did not self-represent. Among several other respondents, a senior advisor in a municipal agency with several years of experience within business development opined, of course, the advantaged [private business] see opportunities for their own business, to position themselves in the market and represent their interests. Still, my experience is that many of the issues that emerge, also those promoted by a particular company, apply to several of the others. The respondents further mentioned that self-representation in itself is not sufficient to legitimise non-elected representatives. In combination with specialist expertise, self-representation becomes a reason to legitimise representatives.
Thirdly, the respondents use membership organisations' references to shared experiences, values, and identity to legitimise them as non-elected representatives. Interest organisations are legitimised as non-elected representatives who act for those with material stakes in network discussions, i.e. their members or at least a majority of their members. The decision-making authority believes membership organisations' speech acts are on behalf of those who work in transportation, retail, restaurants, hotels and tourism. All of these entities have a stake in business and urban development. City council minutes refer to organisations with individual members as representatives of 'the industry'. A senior advisor in a municipal agency explained that they, as the decision-making authority, challenge membership organisations to not only be available to their members. Thus, membership organisations may be expected to represent more broadly than their membership basis. Several respondents also acknowledge non-elected representatives in networks based on them sharing identity with the constituency. Previous work experience, within for example transport of goods and services, indicates a shared identity with these members. In combination, the interviews and documents show that shared experiences also justify claims of representation by organisations without members. The organisations without members represent business companions with coinciding interests as themselves. A director of a municipal agency emphasised that 'the organisations I talk about are competent, advantaged industry organisations, which are sort of trusted by the industry itself as their representatives'. Finally, respondents from the inclusive informational network legitimise non-elected representatives in networks because their representation contains recognition of popular support and unheard perspectives. In their view, the representatives represent important issues that have broad popular support, but are not voiced elsewhere. Climate, environment, the greening of industry, circular economy, liveable cities and end-user experiences, in other words sustainable development, are causes which according to the respondents require additional representation, thereby legitimising non-elected representatives.

Qualifications
Qualifications refer to the perceived attributes and skills of non-elected representatives. Hence, qualifications legitimise the makers of the representative claims. Since the decision-making authority considers the organisation the non-elected representative, charisma, personal communication and networking skills rarely legitimise non-elected representatives. Nonetheless, some respondents value confident and enthusiastic network participants, a clear organisational mandate, and the individual's position within the organisation. Having the head of an organisation present is perceived to increase the legitimacy of the representative because the head makes final decisions.
Findings reveal that truthful arguments, stories and assertions presented in the network are a qualification which legitimises organised actors as non-elected representatives. When asked, the respondents considered both organisation with and without members to be truthful non-elected representatives in issues related to the sector they operate in, their mandate and the expertise they hold, e.g. transport, real estate, or retail. A senior advisor in the municipality commented that 'they are truthful in the sense that they have expert knowledge'. Two senior advisors with several years of experience within urban development found the non-elected representatives in governance networks to be trustworthy when: i) they adjust their views as additional knowledge is produced, ii) it is known who they represent, and iii) they make it clear if they are expressing something mainly for their personal advantage.
The interviews uncover that trust is constructed through previous positive collaborative experiences. Thus, previous collaboration with public authorities is indirectly a qualification which may legitimise non-elected representatives. In presentations held in the network, participating organisations are encouraged to engage in other areas in local politics, e.g. in consultations and workshops. Previous collaborative experiences as a legitimising quality apply to both organisations with and without membership.
Although there is a general perception of the organisations being truthful non-elected representatives, the respondents are unsure of the degree to which they can be trusted. The respondents have experienced that especially organisations without members sometimes portray the situation worse than necessary. In general, private businesses are less trusted when discussing society as a whole. A director of a municipal agency said that 'they are somewhat blind to general societal perspectives, more focused on their own interests'. The respondents also say that they have to pay attention to whether the organised non-elected representatives express self-interests. As one director of a municipal agency described, 'truthfulness is a tough question. They [the non-elected representatives] use arguments which they cannot document to influence the outcome for what it is worth'. Decision-making authorities find promotions of self-interest relatively easy to detect. Undocumented statements, statistics or documented arguments which benefit individual needs are often provided by the most active or economically well-off actors. To avoid such biased information, the decision-making authority orders research and consultancy reports and has the findings presented and discussed at meetings. The decision-making authority therefore emphasises the importance of interpreting what is being said and by whom to reduce the effect of self-interest. A few respondents emphasised that the willingness to practice this behaviour depends on whether the individual participant is perceived as trustworthy, rather than on the organisation.

Connectedness
Connectedness refers to the decision-making authority legitimising non-elected representatives because they interact with the constituency to negotiate what is represented. What became evident when interviewing respondents about the relationship between non-elected representatives and their constituency was that the respondents primarily considered membership organisations to have a constituency which needed to be engaged in constructing representation. The constituency is the membership basis of the organisation. The decision-making authority does not reflect upon if and how organisations without members interact with their constituency.
The findings, which apply to membership organisations, indicate that the decisionmaking authority responds to connectedness in two ways. One group of respondents gives this issue little thought, while the other group is more conscious. Thus, decisionmaking authorities do not necessarily legitimise non-elected representatives because they negotiate the content of representation with the constituency.
The first group of respondents take it for granted that membership organisations explain their actions to their members. The respondents assume that network actors actively inform and get feedback from their members about the discussions, conclusions and work conducted in the network. A senior advisor in a government agency with several years of experience working with private organisations said that I take it for granted that they have meetings with their membership basis or boards. Or some kind of internal newsletterwhat do I know.
[…] I assume that they do not participate without telling someone. Some of these respondents consider maintaining the relationship between the nonelected representative and a constituency to be the responsibility of the representatives themselves. The non-elected representatives' stories from members' workdays are sufficient evidence that a dynamic relationship between the representative and the constituency exists.
The second group of respondents, although significantly fewer, are conscious of the content of representation emerging through interaction between membership organisations and the constituency. The interviews reveal that it is the membership organisations' internal elections that make the decision-making authority rely on them being responsive to a constituency, thus legitimising them as representatives. Also, having membership organisations which include members in network meetings reinforces the perception of responsive representatives. Furthermore, these respondents also express the dangers of relying on membership organisations. As outsiders, the interviewed decision-making authorities acknowledge that they do not know whether the organisation equally presents the opinions of all their members. Hence, the process of negotiating the content of representation is unknown to the decision-making authority. At the same time, the respondents comment that some members may not find it necessary to interact with the representative on what needs representation. With experience from the private sector, a senior municipal advisor expressed that it may be difficult for the membership organisations to engage their members to continually get their input because the members might In addition, the interviewees identify membership meetings, individual contact, social media channels, newsletters, board meetings, websites, breakfast meetings and membership magazines as alternative arenas. The respondents found it difficult to provide an answer to whether these arrangements result in the representatives adjusting their perspectives, opinions and arguments. Some respondents had not participated in the network long enough to see adjustments happening, and those who did recognise changes in what was represented could not identify whether this was a result of the communication with the constituency, changing situations, societal trends or new knowledge.

Discussion and Conclusion
The findings from this study first of all support the theoretical assumption that actors in governance network function as non-elected representatives and that the decisionmaking authority, i.e. the audience, must recognise representatives and their claimmaking for representatives to have an impact on politics. What further emerged was that the respondents in all seven networks acknowledge the organisation and not the individual network participant as the representative of a constituency. From the perspective of the audience, the individual participant making claims and the organisation they present as the representative is the same. This echoes studies which argue that a distinction between the individual claim-maker and the subject, i.e. movement or organisation does not make sense (de Wilde, 2013).
In exploring how the audience acknowledges non-elected representatives as legitimate representatives in the field of business and urban development policy, the empirical analysis revealed that credibility in connection with qualifications, rather than connectedness, legitimises non-elected representatives. Among credibility and qualification, the first is considered to be the most influential. I find that organisations with individual members are considered credible representatives when they explain how they represent using specialist expertise and shared experiences of their members. Organisations without members are justified as credible representatives when they claim to represent on the basis of self-representing their specialist expertise. Thus, the decision-makers do not only ascribe legitimacy to interest organisations because of their expertise (Flöthe, 2019), but also to private businesses, universities and non-profit foundations. What also became evident was that considerations regarding the non-elected representatives' credibility often are the same as the reasons why some non-elected actors are included in the network at the expense of others (Hendriks, 2008;. Contrary to what others have found, individual participants' qualifications such as charisma, communication skills or their professional network rarely legitimise organised non-elected representatives in business and urban development networks (Chapman & Lowndes, 2014). However, similar to other studies (Chapman & Lowndes, 2014;Saward, 2010) truthfulness qualifies non-elected representatives. There are no differences between organisations with and without members in the degree to which they are perceived as truthful non-elected representatives. Thus, one may argue that the audience legitimises organised non-elected representatives who they experience as truthful in claiming to represent others. Indeed, the decision-makers' trust in participating nonelected representatives is important for a functioning governance network (Klijn & Koppenjan, 2012).
Qualifications and credibility refer, albeit in different ways, to the information contained in claims the representatives present in the network and whether or not this information can be trusted. Findings show that through previous experience in collaborating with the representatives, the decision-making authority considers the representatives' information on their own or shared experiences and specialist expertise trustworthy. Information which the decision-making authority itself does not hold is a valued resource among non-elected representatives in governance networks. The relevant documents support this and show that several non-elected representatives in meetings present their experiences, current work or ongoing struggles related to issues discussed in the network. Hence, they provide the decision-making authority with information about, for instance, implemented policies, regulations or conflicts. Therefore, the information gained through representation legitimises non-elected representatives in governance networks. This finding supports de Wilde's (2019, p. 6) argument that legitimate non-elected representatives provide high-quality claims which are rich and detailed in information.
What further emerged is that the decision-making authority is divided when it comes to considering the relationship between the representative and the constituency when legitimising non-elected representatives in governance networks. Connectedness has the most variation in its use to legitimise non-elected representatives. The decisionmaking authority does not recognise organisations without members to interact with a constituency, which may be due to the fact that these organisations are found to self-represent. The decision-making authority does not automatically think of employees, customers or students as a constituency. Without questioning current practices, one group of respondents takes for granted that membership organisations explain their opinions, arguments and comments to a constituency, i.e. members. Interacting with the constituency to construct representative claims is deemed the responsibility of the representative. The other group of respondents expresses that internal democratic processes and inclusion of members in network meetings have a legitimising function of membership organisations as non-elected representatives. Altogether, this finding aligns with previous studies which suggest that governance networks and network actors struggle with their accountability to those outside the networks (Aars & Fimreite, 2005;Aarsaether et al., 2009). Signals from the constituency to the audience that non-elected representation is based upon constant consent are theoretically crucial for the legitimacy of a nonelected representative (Saward, 2010). This study indicates that this is not necessarily the case empirically. Accountability may not provide the audience with an argument to legitimise the non-elected representatives in governance networks. The information non-elected representatives bring to the network is the primary reason for an audience to legitimise non-elected representatives. Again, this raises questions about the transparency of governance networks, the opportunity of constituents to express judgement, and even networks' contribution to representative democracy.
Given these findings, future research should focus on whether the arguments that legitimise organised non-electoral representatives also apply to other types of nonelected, and even elected, representatives in governance networks. Activists, celebrities and experts as non-elected representatives might require different audience acknowledgement than organised non-elected representatives, and thus the results might differ elsewhere. Valuable insights can also be expected from a comparative study, on whether the findings apply to the audience of organised non-elected representatives in connection with, for example, elections, social movements, and political parties. These representatives have a different audience which might, in turn, legitimise non-elected representatives differently.