Negotiating Municipal Merger

This article contributes to organizational research by describing the format and space of negotiation arenas – temporary arenas with diverse social order characteristics. The article also contributes to our understanding of the dynamics of different logics of negotiation. We compare two cases from the Norwegian Local Government Reform. The empirical phenomena of a negotiation arena, where municipalities negotiate potential mergers, provides relevant knowledge. Results from the study show that, despite similar conditional factors, we find significant differences in how the municipalities approach and organize the process, how it evolves and how it ends. We argue that the motivation and design of the negotiation arena not only affect the dominating logic of negotiation through the processes, but also the foundation for any potential mergers. Logics of negotiation are related to the initial characteristics of the negotiation arena, but also become an important variable in the development of this negotiation arena. In this way, the article illustrates a two-way interaction between social order and logics of negotiation, and shows why negotiation arenas differ and change over time.


Introduction
Municipal mergers 1 are a well-known phenomenon in most European countries where the number of municipalities changes over time (Baldersheim & Rose 2010). Such reform processes are to varying degrees, voluntary or enforced by national government (Loughlin et al. 2011). Norway faced a significant Local Government Reform (2014Reform ( -2017 as part of the coalition agreement made by the conservative government that gained power This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. after the 2013 parliamentary election. The aim of the reform was to reduce the number of municipalities through voluntary mergers, and to allocate broader responsibilities to municipalities (Ministry of local Government and Modernisation 2015). In the autumn of 2014, all municipalities received an invitation to start 'neighbour talks' . This was the minister's informal term for a negotiation between two or more municipalities considering merger (Klausen et al. 2016). In this article we refer to 'neighbour talk' as a negotiation arena where municipalities meet and negotiate preconditions for a possible future institution.
Understanding negotiation arenas between municipalities relates to the growing literature on spaces between organizations (Ahrne & Brunsson 2011;Haug 2013;Ahrne et al. 2016). Important contributions have led to progress in understanding this space through the lenses of three main social orders; networks, organizations and institutions. Network literature has contributed to an understanding of the interdependence between equal actors referring to collaborative governance/ network governance (Agranoff 2007;Sørensen & Torfing 2007;Ansell & Gash 2008;Klijn & Koppenjan 2014). Ahrne and Brunsson's article (2011) introducing the 'partial organization' was a significant beginning to several important contributions considering the links between an organization and its environment and the arenas that develop between that organization and others (Haug 2013;Ahrne et al. 2016). Institutional perspectives have also considered how collaborations and institutionalization of loose networks develop into more stable institutional settings (Holmen 2011). However, negotiation arenas between municipalities are temporary and unsettled arenas. Municipalities are free to participate in a number of different negotiation arenas at the same time. They can also leave or start new negotiations. Conditions in the negotiating arena can, therefore, change from day to day and lead to different outcomes. We know less about these temporary and unsettled arenas although they can be considered important for understanding future institutions. By asking how such negotiation arenas evolve and what causes the different outcomes, we can provide more insight into significant early phases of future institutions which are formed by the merger of several municipalities.
This article refers to a study exploring two negotiation arenas with similar conditions, but with different outcomes. In order to understand and explain why these different outcomes occurred, we look at the structural characteristics of the arenas as well how they evolve through processes of negotiation.
Through this study we address two main arguments and contributions. First, by exploring arenas for negotiation we contribute to a deeper understanding of their characteristics and development. We argue that such negotiating arenas can be characterized by features from all three social orders. However, they are dynamic and can evolve through a negotiating process.
Second, we argue that there is a two-way relation between social orders and logics of negotiation. By combining social orders through the framework of Haug (2013) and Ahrne and Brunsson (2011) with Rommetvedt's (2006) dynamics and variations of negotiation logics, we suggest a conceptual model to explain variation in outcomes and evolution of negotiation arenas.
In the next section we lay out the proposed theoretical approach. This is followed by a section on methods. Here we also describe the background of the Norwegian Local Government Reform and address selected cases and data sources. The subsequent section attempts to analyse these detailed accounts, and to explain the observed differences through the prism of the proposed theoretical approach.

Theoretical Framework
In order to understand the concept and condition of a negotiation arena, we need conceptual clarification and a broad theoretical framework. This article combines insight from two theoretical approaches. First, social orders help us to explore the structure and conditions of the negotiation arena. Second, we use different logics of negotiation to explain the development of and the process within the negotiation arena. The theoretical framework is elaborated below.

Negotiation Arena as Social Order
Organizations, institutions and networks can be seen as different types of social order. In recent years, progress has been made in understanding the organized spaces between formal organizations through the lenses of social orders (Ahrne & Brunsson 2011;Haug 2013;Ahrne et al. 2016). Ahrne and Brunsson (2011, 83) claim that organizational research is largely directed towards formal organizations, while the outside world is referred to as network, market or institutions. To understand organizations at the intersection between formal organization and 'outside world', they present and describe the option of partial organizations. Formal organizations make decisions about membership, rules and sanctions. They also include a hierarchy and the right to monitor compliance with commands and rules (Ahrne & Brunsson 2011, 86). In the same way that formal organizations exist when all elements are present, partial organizations exist when one or more elements are present. Networks and institutions can also consist of organizational elements, but in an organized order the relevant elements have been decided (Ahrne & Brunsson 2011, 85). In the words of Ahrne and Brunsson (2011, 85) an organization is a '(…) decided order in which people use elements that are constitutive of formal organizations' . Outlined from this; Defining institutions as social orders is more complex as there is a wide range of institutional terms (Jessop 2001). Some political science literature uses 'institution' to mean what organization theorists call formal organizations, like schools and hospitals. Anthropologists link 'institution' to knowledge and culture. From a sociological point of view, an institution can even be a handshake (Ahrne & Brunsson 2011, 89). Jessop (2001Jessop ( , 1220 describes institutions as 'social practices that are regularly and continuously repeated, that are linked to defined roles and social relations, that are sanctioned and maintained by social norms, and have a major significance in the social structure' . Further, Jessop linked institution to the concept of 'institutional thickness' (Amin & Thrift 1995). From this concept, institutions become a social order where members 'interact intensively generating shared understandings, socializing costs, and developing mutual awareness of being involved in a common project to promote and sustain local or regional development' (Amin & Thrift, cited by Jessop 2001, 1221. Moving on to networks as an alternative social order, we find approaches built on several theoretical traditions, such as organizational theory and public administration (Klijn & Koppenjan 2000). The main characteristics of a network are interpersonal relations of trust, non-hierarchical coordination, non-boundaries, exchange of resources and a high degree of reciprocity (Börzel & Panke 2007;Ahrne & Brunsson 2011;Haug 2013, 712). As Ahrne and Brunsson (2011, 88) put it: In its genuine form, network is a form of interaction that is qualitatively different from organization and a network is often defined in terms of its informality, lack of boundaries and hierarchical relations, and is ascribed with qualities such as spontaneity and flexibility.
'A network consists of informal structures of relationship linking social actors; which may be persons, teams or organizations' (Ahrne & Brunsson 2011, 88). Another network concept is governance network. A governance network highlights how autonomous and mutually dependent actors (public and private) voluntarily collaborate in order to reach a desired outcome (Sørensen & Torfing 2007). Negotiation is the main tool for finding common and legitimate solutions and this is a delicate process to keep the members together (Sørensen & Torfing 2007;Røiseland & Vabo 2016). The key to success -realized policy -is co-operation (Klijn & Koppenjan 2000).
The three social orders; organizations, institutions and networks are different in nature but at the same time, overlapping, as all orders include interaction between people. However, whereas organizations are decided, networks and institutions emerge (Klijn & Koppenjan 2000; Haug (2013) attempts to rip down borders between classical social orders by trying to understand social movements as spaces where face-to-face meetings play a key role for the inter-organizational domain of meso-mobilization. In more detail, Haug (2013, 707) shows how meetings became the natural way to establish and stabilize social relations and to create a social order that facilitates collective action and fosters social change.
Along with Haug (2013), we argue that the negotiation arena is a dynamic hybrid of all three forms of social order that can change character over time. Previous studies have argued that different logics of negotiation are significant for interaction, decision making and the dynamics of change in negotiation arenas (Rommetvedt 2006;Hanssen et al. 2014). Following this argument, we present Rommetvedt's three logics of negotiation.
Three Logics of Negotiation Habermas' (1984) distinguishes between strategic and communicative actions. He argues that a broader spectrum of logics of action (decision-making) should be taken into consideration, and that the logics of decision-making should be related to the decision situation or constellations of benefits and preferences. He also argues that there are decision-making processes that lie in the boundary between strategy and deliberation. From this argument, Rommetvedt (2006, 206) offers a middle perspective by introducing the term 'deliberative negotiation' . As he puts it: '(…) negotiation can be both more than and different from pure strategic processes' (Rommetvedt 2006, 207). Central characteristics in strategic bargaining are the action's aim to succeed. Actors' interests are given beforehand and actors act strategically in order to satisfy given preferences. The primary characters of the negotiation are political strategies. 'Divisible goods or material values can become an object for strategic bargaining aimed at achieving a compromise -a compromise that can, for example, be based on the actors sharing the goods 50/50' (Rommetvedt 2006, 201). Crucial elements in strategic bargaining are political strategies and power (Rommetvedt 2006). While bargaining is a central characteristic of the strategic bargaining, arguing becomes central in deliberative negotiation. Arguing is based on arguments that the actors perceive to be valid -'honest arguing' (Rommetvedt 2006, 205). The aim is to establish a mutual understanding and goal. The deliberative negotiation process is open and sincere, as the actors believe in shaping and reshaping common preferences, meaning that preferences develop through communicative processes (Rommetvedt 2006, 205). Moving on to the category of deliberation, the negotiation argument is based on the public good, and politicians regard themselves as representatives of the people. Conditions for deliberation are stronger in situations characterized by institutional practice (Rommetvedt 2006).

Combining Social Orders and Logics of Negotiation
Inspired by Haug's model of social arena (2013, 713), Figure 1 illustrates the theoretical combination between social orders and logics of negotiation.
The triangle illustrates the negotiation arena characteristics from all three social orders (in each corner). The social orders are interlinked, meaning that they have potential for development over time. The dotted circle placing 'logics of negotiation' inside the negotiation arena, illustrate different negotiation logics in play. The arrows between the structure and the logic of negotiation illustrate a two-way relation between the two different theoretical approaches.
First, along with Haug (2013), we expect negotiation arenas to be characterized by all three social orders and that social orders can change over time. Second, based on Rommetvedt's typology for public decision-making (2006) we expect that different negotiation logics have different potential to shape and reshape institutional practice and, as a consequence, influence the characteristics of the negotiation arena towards a common platform of a new municipality. More specifically, we expect negotiation arenas dominated by 'arguing' and deliberative negotiations to promote a higher potential for evolving the social arena. An open, deliberative negotiation process is characterized by a mutual interest in shaping and reshaping common preferences. Under such circumstances a social order can be expected to develop  (Rommetvedt 2006;Holmen 2011). Based on the same argument, negotiation arenas dominated by strategic bargaining, are expected to be dominated by strategic actors defending their autonomy, making it less likely that the negotiation arena will evolve in terms of social order (Rommetvedt 2006). These expectations are followed up in the analysis.

Data and Methods
The article is based on a comparative case study. 2 The selection of cases was strategic: in order to explain variations in negotiation arenas regarding characteristics, evolution and outcome, similar cases were chosen. The two cases chosen were located in the same county and city region 3 and had, because of this, similar contextual factors affecting them. Both cases were a response to the same central government order and were organized by the same number of participants, the same negotiation agenda and the same timeline.
Data was collected through observation of meetings. 4 Arena A was established in September 2015. From this point we generated data based on 14 meetings. Arena B was established in November 2015. As we experienced Arena B becoming relevant for our research, we started to observe meetings at the beginning of 2016. As a result, we missed the first three meetings (8.25 hours). As described in Table 1, four of seven meetings were observed. This gap was compensated for by using meeting reports, newspaper articles and informal conversations with actors in the field. We were not informed of executive group meetings in Arena B.
The observation guide was divided in two columns. On the left, we wrote the agenda of the meeting. On the right, we described the process, actions and arguments. At the end of the observation guide we had a space where we noted social climate and comments about the process in general.
For analysis and the systematic search through our observation notes, we look to Spradley (1980, 78). The nine dimensions given are not equally important for every case, but they give a rich insight and overview of the observed situations. We selected eight relevant dimensions related to characteristics of social arenas and modes of negotiation. First, to be able to describe the structure of the negotiation arena we look at the physical place (space) and the people involved (actors). We noted the meeting place and members of the groups, and also the responsibilities distributed between the actors. Further, we tried to describe how the negotiation arenas arise and develop through the actions in the meetings (act and activity), the set of related activities that actors carry out (event) and how the events are scheduled over time (time) and, thereby, how the negotiation arenas become central. Through the observation notes we find expressed emotions, tensions (feelings) and different objectives the actors tried to accomplished (priorities and goals). In summary, we studied the characteristics and evolvement of the arena during the negotiation process in two similar cases in order to understand the outcomes and the progress towards forming new municipalities. National policy documents describing terms of the Norwegian Local Governance Reform were also important data sources, as were meeting reports. The documents helped strengthen our contextual knowledge and our understanding the goals, arguments, political procedures and timelines. In all, the documents included knowledge about elements that would be likely to affect the local negotiation arena. In the following section we present the reform and the two selected cases.

Case Presentation
The story of the two negotiation arenas is based in the context of the Norwegian Local Governance Reform. The reform was structured into two different timelines. First, municipalities which decided to merge in the autumn of 2015 had their merger formalized by royal decree in the spring of 2016. These new municipalities became operative on January 1, 2018. Second, municipalities which decided to merge by July 1, 2016 were considered in the new proposition to the Norwegian Parliament in the spring of 2017. The new municipalities then became operative, by January 1, 2020 at the latest (Ministry of local Government and Modernisation 2015). Our cases belong in the second timeline and both are located in the county of Rogaland, in the south-west of Norway. Each case, understood as a negotiating arena, consisted of three municipalities. The county administrator had the responsibility for guiding municipalities through the negotiations. Despite the established timeline and expectation for local negotiations, there were no instructions or principles from central government to follow. Municipalities had to design the structure and content of the negotiation arena on their own; the number of meetings, the meeting agenda, roles, rules, internal deadlines and end product.
Arena A consisted of three municipalities; Time, Klepp and Hå. Nine local politicians (mayors, deputy mayors and the leader of the largest opposition party) and three chief municipal executives formed the operation group. 5 They also hired one external project leader. His role was to support the chief municipal executives and structure the interaction within the negotiation arena. Results from a common resolution in all three municipalities showed a mutual desire to start negotiating for possible merger. Based on this, their common aim was to establish a Letter of Intent for a future municipality.
The municipalities had a lot in common and shared several features. First, they shared a common identity and culture. Second, they had equal areas, similar industry, geography and populations. They also had a shared history, with earlier attempts at merger and collaboration through several inter-municipal companies.
Arena B consisted of three municipalities: Stavanger, Sola and Sandnes. Nine local politicians (mayors, deputy mayors and leader of the largest opposition party) and three chief municipal executives representing the municipalities formed the operation group. 6 Based on a common decision, they hired an external facilitator from the 'KS consultancy' owned by the Norwegian Association of Local and Regional Authorities.
This negotiation become possible because of a strategic invitation from Sola. Results from a resolution in Sandnes showed that they didn't want a future municipality formed by two big cities (Sandnes and Stavanger). To prevent this resolution from stopping the negotiation process, the municipality of Sola invited the other two to the negotiation arena.
All three municipalities have important areas for industry and they already collaborated through several inter-municipal companies. Between them, they have a major proportion of the total population of the county of Rogaland. Stavanger -known as the oil capital of Norway -is the biggest of these three municipalities. Sandnes is traditionally the 'little brother' of Stavanger but had experienced a growing number of inhabitants and an increase in industry since the last round of municipal amalgamations in the 1960s. Sola has a huge impact on policy and industry in the area by controlling both the airport and a large harbour area.

Findings and Discussion
Based on Rommetvedt's (2006) typology for public decision-making and Haug's argument of social orders as dynamic (2013), we expected that different logics of negotiation would dominate the negotiation arena. Following this we also expected different logics to have the potential to shape and reshape institutional practice and, as a consequence, evolve negotiation arenas into new social orders. The next section attempts to analyze these detailed accounts of two negotiation arenas and to explain the observed differences through the prism of the proposed theoretical framework.

Arena A: Building Bridges (from a Network to a Common Institution)
On 29th September 2015, three municipalities decided on the number of meetings, meeting agendas, who would be the leader of the operation group, the role of the chief executives and project leader. They also agreed that a letter of intent was the desired outcome and that this letter of intent should serve as the foundation for an upcoming referendum. In this way, the eight meetings that followed had a clear structure, division of responsibilities and goal. The arena consisted of municipalities, all of which wanted to negotiate and saw the reform as an opportunity. In more detail, the data shows three mutually dependent municipalities with a declared common goal to become Norway's biggest agricultural municipality. They were autonomous actors representing their municipality, but at the same time one unit representing the citizens of the geographical area 'Jaeren' . Despite this, some members in the arena were not convinced, in the early phase of the negotiation process, that a merger was the solution.
Initially, the chief municipal executive asked the question; 'How would you like it to be when we talk with each other? How deep and how broad will we go?' By this, the chief municipal executive established 'rules' for the negotiation between the participants, but also clarified the common product, i.e., how much information they wanted to present to the inhabitants in the letter of intent. Every meeting started with the operation group leader welcoming everyone and asking for comments on the report from the previous meeting. They then continued with the meeting agenda. Representatives from all municipalities had their turn to speak and all contributions were appreciated. If the topic was of a character that required more information or investigation, they wrote in the letter of intent that it was up to the new municipal council to decide. In this way they closed arguments and continually moved on to the next topic. Through conversations and arguments -stage by stage processing -they agreed. In more detail, the data indicate negotiation as a process -progressing through suggestions for the future structure of a new municipality -rather than political standpoints. Arguments sounded like, 'In a new municipality we will have (…)', and 'in a new municipality we get (…)' . In summary, arguments reflected a common desire. Also, the members identified themselves as 'we' indicating a strong feeling of fellowship that would be expected to have an effect on the perceived relevance of moral and legal norms inside the arena (Rommetvedt 2006, 198). Along with serious arguments about a common future, we also observed a climate characterized by informal conversations and jokes.
In our observations of the climate inside this arena, the close relations between the administration and the politicians were prominent. There was a clear administrative desire to contribute in the negotiation process. As a result, the chief municipal executives gained an important role. While the politicians argued, suggested and agreed on decisions related to the future municipality, the chief municipal executives took notes. Between each operation group meeting, chief municipal executives and project leaders discussed the arguments and input from the previous operation group meeting. On the basis of this, and with help from one administrative section, they wrote and rewrote the letter of intent each time and forwarded it to the operation group. In this way the letter of intent describing the new municipality grew and developed throughout the process. Chief municipal executives and the project leader also designed the structure and subjects for upcoming operation group meetings.
On 21st of January 2016, all members signed the Letter of Intent. The event was celebrated with great applause, cake, pictures and people full of expectations. The signing of this document illustrated the foundation and potential start of a new municipality.

Arena B: Three Organizations -Three Strategies
The first of seven meetings was held in November 2015. In this first meeting, politicians designed the structure and responsibilities for the negotiation arena. They discussed different topics for upcoming meetings, how to carry out a citizen survey and the final product -should it be a letter of intent or a principle document? While some representatives wanted a signed letter of intent, others wanted a more objective document free from obligations. They did not share the same understanding of the importance of the negotiations and the end product. Nor did they manage to create a common goal. In other words, in the initial meeting there was an imbalance of motivation, as a mayor summarized: We must clarify what this process is. We have said from the first day that we are not sitting here to make a letter of intent. We are making a document of how it can look if we agree.
Here are mayors representing those who are cheering -and those who are against a merger. We have different roles. We are here to do what the politicians have asked us to do. This statement (…) show a clear instrumental approach to the process where some municipalities entered the negotiation more strategically to answer expectations from central authorities. The debate dealing with defining the aim of the process also came up in later meetings.
The negotiation arena was controlled by formal rules and sanctions on how to act. There was a formal tone between the participants and they followed strict rules when they signed and made contributions or comments. The chief municipal executives had a neutral role during the operation group meetings and the administrative work. Tasks like writing the principle document were, to a greater degree, left to administrative employees in representative municipalities.
The collaborative climate and social norms in Arena B were based on tension and disagreement. Observations showed politicians bargaining for results which favoured their own municipality. Protection and lack of trust affected opportunities to find common solutions. For instance, a representative from Sandnes demanded that a potential new municipality build 'Gandsfjord bridge' and develop the area east of Sandnes, all located inside the representative's own municipality. There was also disagreement about © 2020 The Authors. Scandinavian Political Studies published by John Wiley & Sons Ltd on behalf of Nordic Political Science Association which questions should be included in the upcoming referendum. Should inhabitants be given options between different possible mergers or not? The tension in the operation group also reflected the fact that some participants were involved in a local resistance group named 'Yes to Sandnes as a municipality' (Olsen 2016). A reporter in a local newspaper described the disagreement and tension in the operation group like this: 'Stavanger, Sandnes and Sola apparently agree -at least for the first ten minutes (…)' (Ueland 2016, 8). A participant summarized the characteristics of the negotiation arena by saying: 'This is not group work, it's bargaining!' Under these circumstances, the hired facilitator had a clear role as a mediator trying to lower the level of conflict. Discussions related to naming a possible future municipality serve as an example. In short, Stavanger had decided in advance that they wanted the new municipality to be named 'Stavanger' . This suggestion met resistance from the two other municipalities. After it was argued back and forth for more than an hour the facilitator asked the politicians representing Stavanger to consider another name. The whole negotiation was at stake and some politicians indicated an intention to leave the negotiation table. In other words, we observed imminent danger of deadlock or breakdown. The success of strategic bargaining is when participants compromise '(…) where each of them is given a share of the "cake"' (Rommetvedt 2006, 202). On the facilitator's recommendation a compromise became the solution when they all agreed on 'North-Jaeren' as the future municipality name. 7 The hired facilitator also had a motivator role. He always started the meetings by welcoming all the participants and inviting them to contribute. When he experienced some 'delicacy' in the group he reminded them of the reason they were together and their goal for the meetings. By doing this he was praising them for continuing with the negotiation despite disagreement and different political standpoints.
Observations showed politicians' uncertainty about the process: they were afraid of being accused of statements that were not agreed by their political party. As one politician put it: 'I hope I don't read in the newspaper tomorrow that I agreed.' This reflected the ongoing debate about how detailed the principles in the principle document should be. One politician said: 'The question is, how far we should go? We have to use the word "should" instead of "shall" . It sounded like it already had been decided' . Some called for more details and decisions to be included. Others wanted to present short facts and leave the more demanding questions to any potential new municipality council. Another politician said: 'We must dare to talk about name and location. There are many ideas of names in the delivery room, but you don't need to decide before the christening' .
After seven meetings with varying amounts of debate, the outcome was an unsigned 'in principle' document with objective information that served as the foundation for the upcoming referendum. This document illustrates closure on discussions about a possible new municipality. It was a relieved group of politicians and administrative employees that left the negotiation table at Sola cultural centre on 30th of March 2016.

Negotiation Municipality Merger
Our presentation of the two negotiation arenas show that, despite similarities, we find significant differences in how they approached the process and organized it, how it evolved and the outcomes. The theoretical framework presented aims to help explain why. On the basis of the framework, we had expectations of deliberative negotiation as a logic with better potential for evolving the social arena. On the other hand, we expected processes dominated by strategic bargaining to have less effect on the evolvement of the social arena. Through the model we also argued that the initial characteristics of the social arena would have an effect on the climate and the negotiation logic which dominated. Figure 2 illustrates characteristics of the two negotiation arenas according to social orders and how they developed throughout the process. It also shows the logic of negotiation dominating the processes.
The characteristic of Arena A illustrates a network of autonomous municipalities with the goal of finding a common solution for future merger. Both administrative and political members of the network saw the reform as an opportunity to develop a new sustainable municipality. They were also mutually depending on each other to create this new municipality (Sørensen & Torfing 2007). The members had a strong interest in strengthening the collective interests between the municipalities through the negotiation process. Network can, because of this, be considered to be the dominating social order based on the informality, the lack of boundaries and the lack of hierarchical relationships as typical features (Ahrne & Brunsson 2011). Trust, based on previous collaboration and a common identity, maintains networks as social order (Røiseland & Vabo 2016). These characteristics are closely related to the type of negotiation logic dominating the process. Some traces of strategic bargaining were found in the early phases where the individual political strategies were presented but a more argumentative approach linked to deliberative negotiation dominated early. Members had different interests, but were able to take part in serious discussions to find '(…) how preferences can be united to the benefit of both parties' (Rommetvedt 2006, 201). Through an open and sincere process different topics were discussed, text was developed and brought into discussion again.
This deliberative approach evolved and provided the space for interpretation of new ideas, allowing individuals to change position during the process so that common preferences were reshaped (Rommetvedt 2006, 205). Those members who were unsure or against merger in the early phase changed their position during emerged in the arena. As one member said: 'I'm on my way, as I see that there are some interesting arguments here' . The combination of a network evolving common preferences (Røiseland & Vabo 2016), and an institutional strength growing, where members develop a mutual awareness of being a common unit (Amin & Thrift 1995;Jessop 2001), provides the argument for arena A's evolving from a network into an arena with institutional characteristics. Our findings also show deliberative negotiation as a significant explanation for this evolvement.

The Characteristics of Arena B
These illustrate a more stable situation where the three member municipalities had already positioned themselves with definite standpoints -considered as a lock-in with clear power interests. Few common preferences for the merger were present in the early phase, where each municipality strategically bargained for individual preferences. Still, preconditions like this bring political and personal conditions to the negotiation table where terms for a common unit are impaired and actors follow incompatible interests (Mouritzen 2006, 20). This also affected the organization of the arena.
Arena B as a negotiation arena can be considered as a network if we define it as an arena for seeking common and legitimate solutions. However, as we expand the definition of network by including voluntary collaboration and interdependence (Røiseland & Vabo 2016), our data exclude this negotiation arena as a reinforced network. In contrast to a negotiation arena of common goals, openness and mutually dependent members, we observed members participating as if in answer to a formal invitation, arguing as three independent and autonomous organizations. There were, however, some institutional characteristics that brought them to the negotiation table and made them interact because they felt an institutional commitment to completing the formal mandate. Because of the strategic and individual preferences, it is more likely that we can understand this negotiation arena as a network of three separate organizations. This is the basis on which we argue for Arena B's position in Figure 2.
Further, Arena B is illustrated as a rectangle referring to a stable situation. Observations show some traces of deliberative negotiation, for instance, suggestions of package deals such as Stavanger becoming the new municipality name but Sandnes getting the city hall. Still, strategic bargaining dominated the negotiation arena with individual political strategies and previously decided arguments. The language during the process, 'they' and 'we', also indicated a strategic line with dominating individual preferences (Rommetvedt 2006, 198).
Our findings show that preconditions and structure of the negotiation arena gave limited opportunity for deliberative negotiation logics. The participants failed to gain a common preference for a future new municipality. Nor were they able to construct an open and sincere process for shaping and reshaping their arguments. Because of the structure, together with strategic bargaining as the dominating logic of negotiation, the prospect of finding common ground for a future merger was limited.

Concluding Remarks and Policy Implications
The objective of this article has been to describe and explain variations in spaces referred to as 'negotiation arenas' . A negotiation arena is not a defined and stable space, but is a temporary arena with diverse social order characteristics. This study on the negotiation arena between municipalities during a negotiation process, addresses the growing literature on spaces between organizations (Ahrne & Brunsson 2011;Haug 2013;Ahrne et al. 2016). This article has provided insight in premises for developing such negotiation arenas with characteristics that facilitates a possible future new institution.
Our study reports from two negotiation arenas with similar conditions, but quite different negotiation processes, developments and outcomes. Our initial expectations of deliberative negotiation as a logic with better potential for evolving the social arena, can be confirmed through this material. The cases show that deliberative negotiation allows ideas and arguments which lead to a reshaping of common preferences as argued by Rommetvedt (2006). Deliberative negotiations also contribute to strengthening institutionalized patterns between the actors, through a gradual change from independent members to developing institutional characteristics during the process.
On the other hand, our study shows limitations in negotiation arenas dominated by strategic bargaining, especially in situations where members do not have mutual interests and motivation. Strong individual preferences, combined with limited trust in relationships, turn the negotiation arenas into a formal forum consisting of three autonomous organizations with their own few promises to become a future partial organization (Ahrne & Brunsson 2011). In this way, domination either by strategic bargaining or deliberative negotiation will affect the social order during a process.
The two cases show how initial participant motivation has significant impact on the structure and design of the arena as social order. Arena A had an initial common interest and designed a space based on predictable collaborative processes. Such preconditions pave a path for a deliberative dialogue. Arena B initially had conflicts of interest and designed a space based on formalities and voting, which can reinforce a logic of negotiation based on strategic bargaining. Based on this, we confirm the expectation, raised in the theoretical figure, that the motivation and design of the negotiation arena affected the dominating logic of negotiation. In other words, the study illustrates an expected two-way relationship between social orders and logics of negotiation.
Our findings have clear implications for policy. First, insight into this early phase of processes between autonomous actors inside temporary arenas provides important understanding of the development of new institutions. Characteristics and conditions of a negotiation affect the possibility of autonomous actors, like municipalities, merging.
Second, our study shows how the government order of 'neighbour talk', does not necessarily lead to the desired outcome. Other than the expectation of 'talking' to neighbouring municipalities and considering merger, there were few clear directions from central authorities. This study also gives an insight into how actors can use 'neighbour talk' strategically, as an instrument to reinforce earlier decisions, rather than a tool for new insights and development. This tells us that the way in which national reforms guide local actions of negotiation, can have an impact on the effects of the reform and the achievement of the goals. This is useful information when designing future merging processes of autonomous public organizations. NOTES 1.
We use the concept merger to describe the possible goal of negotiation between municipalities. In other contexts the concept amalgamation is used.

2.
Data from the two cases were generated through a bigger piece of fieldwork following eight municipalities from March 2015 to June 2016 (in total 117 hours). The fieldwork can be divided in two periods. The first period (March 2015 to June 2015) consisted of observing municipalities writing a common report for potential merger in their county.