Political Deadlock: A Network Analysis of Decision Processes in Urban Politics

Urban stakeholders face a diverse range of socio-economic, cultural, and technical challenges that each demand diverging and partially conflicting answers. In this context, local and regional public actors’ capacity to smoothly govern decision-making processes is structurally limited. Rather typically, conflicts may occur between public administration and the business community or the civil society. Moreover, actors within public administration notoriously clash over priorities. In this paper, we address the case of decision blockades and non-action in urban development politics. More specifically, we advance a network analytical perspective to explore the dynamics that lead to a political deadlock and persistent non-action in the case of Deutz Harbour in Cologne, Germany. Our findings indicate that a distribution of power and a non-prioritisation of frames effectively hinder action towards decision-making.


1
A network perspective on urban politics Social structure and the built environment of cities mutually influence each other: On the one hand, specific actor constellations pursue distinct goals in developing the city and pushing ahead with projects of eminent symbolic value (like culture centers or libraries, e.g.Balke et al., 2017); on the other hand, the perception of distinct urban places and spaces shape the behavior of actors and their options to act (Lelong et al., 2016: 6; see also Davis, 2006).Negotiation processes in the urban arena routinely encompass the allocation of public goods, like the water supply or public transport systems, for example.Due to the public good character of these infrastructures, planning processes are inherently political.And since a diversity of stakeholders pursue conflicting agendas, local and regional public actors' ability to smoothly govern decision-making processes is structurally limited.Conflicts typically occur between public administration on the one hand, and the business community and civil society on the other.Moreover, actors within public administration, like different departments or governmental levels, clash over priorities.Confrontational decision-making processes and partisan politics notoriously lead to delays and even political deadlocks and persistent non-action.In this paper we seek to analyze the dynamics that culminate in such deadlocks.
How can we analyze and explain instances of political deadlock?The paper aims to answer this question by advancing a network analytical approach.We chose this particular analytical strategy since the network perspective enables the investigation of a broad range of social phenomena through a differentiated set of concepts and methodological tools.Most generally, network science focuses on the relational properties of social entities and their embeddedness in social structures (Wasserman & Faust, 1994).By pushing beyond a pure structuralist perspective, more recent strands of network analysis seek to extend the analytical scope of relational analysis by attending to the importance of culture and language (e.g.Emirbayer & Goodwin, 1994;Mische & White, 1998;Mützel, 2009).This cultural constructivist advancement of network analysis provides conceptual tools to analyze discursive patterns and dynamics as well as the resulting constellations of growth coalitions (e.g.Logan & Molotch, 2007) and blockade alliances (Leifeld & Haunss, 2012;Lelong, 2015;Nagel, 2016;Ingold & Leifeld, 2016).This systematical investigation in our case study emphasizes that the unresolved conflicts over land-use may not be ascribed to a single (type of) actor or a direct confrontation but to a complex distribution of power and intra-actor ambiguities of framing.
Following this line of reasoning, we analytically distinguish between a social-structural and a cultural context of action that either facilitates or constrains actors' opportunities (cf.Emirbayer & Goodwin, 1996;Imbroscio, 1999).The social-structural context comprises enabling and constraining properties emerging from the social relations between collective and individual actors, and the specific position of actors in the network (e.g.Scott, 2000;Obstfeld, 2005;Vedres & Stark, 2010).The cultural context encompasses the actors' perceptions, "normative commitments and their understandings of the world and of their own possibilities within it" (Emirbayer & Goodwin, 1996: 365).The actors' perceptions and interpretations, then, guide action and, therefore, may cause actors to open or block certain courses of action (Kahneman & Tversky, 1984;Emirbayer & Goodwin, 1994;White, 2008;Fontdevila et al., 2011).
By framing our analysis in social-structural and cultural terms, we refer to network analysis as well as urban studies.As regards the network science, this paper employs an analytical perspective informed by social network analysis (SNA), which perceives networks as "a specific set of linkages among a defined set of actors, with the additional property that the characteristics of these linkages as a whole may be used to interpret the social behavior of the actors involved" (Mitchell, 1969: 2).Our network research neither takes a normative perspective that praises networks as a superior mode of governance (e.g.Davies, 2011) nor claims to describe historically novel forms of decision-making (e.g.Kenis & Schneider, 1991).As to urban studies, network analytical strategies up until now have only rarely been utilized to investigate the dynamics of urban political decision-making (Dempwolf & Lyles, 2012).Previous studies have mainly been restricted to the topics of social capital and community development (e.g.Laumann & Pappi, 1976;Wellman, 1979;Maloney et al., 2000;Middleton et al., 2005) or participation (e.g.Borich, 2010).In our opinion, network analysis is especially helpful for analyzing questions of urban studies because this relational perspective offers an integrated approach to analyze actors' constellations and power structures as well as argumentative patterns and frames of actors in a dynamic perspective (Lelong & Nagel, 2017: 16).
Our article is structured as follows.We first elucidate our network analytical framework (chapter 2).Subsequently, we outline the methodology employed and introduce the case study on the specific decision-making process in Cologne, Germany (chapter 3).By presenting key findings, we demonstrate how the actors' social-structural context and the framing of urban places and spaces influence their behavior (chapter 4).We finally draw broader conclusions on the political dynamics that culminate in non-action.

2
Analysing asymmetries: Power and framing in urban politics In a recent overview, MacLeod (2011: 2630) reconsidered urban politics and governance and reflected several trends, one of them is "a 'renaissance' of redundant docklands and former industrial inner-city zones into mixed-use creative cultural quarters, buzzing economic districts, heritage and tourism villages and gentrified apartments (…)".This stands in the contrast to the deadlock situation in the case of Cologne Deutz Harbour.In the tradition of empirical work on urban studies, the investigation of power, actors, and interests in urban decision-making have been a central issue (McCann, 2017: 314).The investigation of the framing of the decision-making process related to strategic urban projects is indeed not unusual (cf.Satel, 2008) but the combination with network analysis is different.
In this sense, our network theoretical concept intends to provide a methodological frame that is able to capture the possible conditions of action or non-action in urban politics.This perspective can supplement (Blanco, 2013) the urban regime concept (Stone, 1989) and the growth machine approach (Molotch, 1976).Regime theory and urban growth machine offer valuable concepts and analytical categories for understanding the "collective action problems that have to be overcome for effective urban governance to emerge" (Stoker & Mossberger, 1994: 195).However, from our point of view, it is reasonable to move to a higher level of abstraction in theory as suggested by Mossberger and Stoker (2001), if a coalition of public and private business actors is not detectable, in cases of non-action or subsystems of urban politics.The abstraction of the network concept can include all types of actors.Thus, urban regimes may be understood as specific policy networks (Mossberger & Stoker, 2001).
Urban policy-making is characterized by high complexity and interdependency of decisions.In order to assess the social-structural context of political action (or non-action), we reconstruct the constellation of actors who exert significant power on the decision-making process by commanding key resources, i.e. the actor network (see chapter 4.1).Our understanding of power follows Max Weber's classical definition of power as "the probability that one actor within a social relationship will be in a position to carry out his own will despite resistance, regardless of the basis on which this probability rests" (Weber, 1947: 152 [trans. by Knoke & Kostiuchenko, 2016: 2]).The actual or potential exercise of power, then, comprises asymmetric relations between two or more actors or groups of actors (Knoke & Kostiuchenko, 2016).We differentiate institutional resources (related to official positions and functions), material resources (raw materials, machinery, infrastructure or financial resources), and knowledge resources (insider information, professional or procedural know-how) which can be leveraged to change the course of action.However, rather than determining action in an unequivocal fashion, ressources merely afford opportunities for action (Bathelt & Glückler, 2005).Moreover, the cultural context of action (or non-action) has to be taken into account because ideologies, norms, beliefs, and values play a central role in decision-making processes (e. g.Fischer & Forester, 1993;Williams, 1995;Mische, 2008).As Michel Foucault maintains, the production of discourse is controlled, selected, and channeled by society, and the discourse, in return, exerts control over society (Focault, 1991).Analyses of policy-decision processes show the complex interplay of actors' perception and interpretation of problems, and their problem-solving activities which continuously change during the discourse (Janning et al., 2009: 60).In this paper, we conceptualise the actors' political perceptions and interpretations by referring to the notion of frames (see chapter 4.2).Frames are 'schemata of interpretation' (Goffman, 1974), which construct meaning and guide action and, hence, may cause actors to open or block certain courses of action (Kahneman & Tversky, 1984;Emirbayer & Goodwin, 1994;Hays, 1994).By combining the analysis of the actor network with an analysis of the actors' frames, we aim to elucidate the dynamics of a specific decision-making process and to gain insights into the dynamics that lead up to non-action more generally.

3
Examining deadlock: The case of Cologne Deutz Harbour In our case study, we elucidate a situation of a persistent political deadlock despite seemingly favorable conditions for action on the structural macro level.Rather than offering an empirically detailed historiographic reconstruction of the case of Cologne Deutz, we primarily seek to advance and gauge the usefulness of our conceptual approach to analyze the political deadlock (Grabher, 2004).

3.1
The contested site: Cologne Deutz Harbour The conversion or preservation of the harbour site in Cologne Deutz had been subject to political debates over more than two decades (Stadt Köln, 2009).The conflict echoes a global trend to leverage the value of centrally-located waterfront sites by transforming them from industrial locations into attractive new urban quarters (see for example, Hoyle & Pinder, 1992;Desfor et al., 2010).Derelict or under-used inner-city docklands have become key targets of urban redevelopment initiatives for prime residential, commercial, and touristic purposes at conspicuous sites (e.g.Evans, 2003;Smith & von Krogh Strand, 2011;Balke et al., 2017).In the case of Cologne Deutz Harbour, we focus on the persistent non-action from the 1990s until 2011.Only in 2015, the council of Cologne decided to convert the industrial harbour site into a mixed-used urban quarter (Stadt Köln, 2015) 1 .
The Deutz Harbour site, one of the six public ports of Cologne, is located at the river Rhine right opposite the highly symbolic landmark of the Cologne Cathedral.At the time of the investigation, land-use comprised port facilities for bulk goods and general cargo as well as manufacturing with no port-related activities.20 per cent of the area was derelict and only five per cent of the total turnover of the Port of Cologne was generated on this particular site (Stadt Köln, 2008).Due to the attractive central location and the decreasing level of economic activity, a controversial debate on the future development already went off in the 1990s (Stadt Köln, 2009) and revolved around the socio-economic potentials and ecological benefits of the site (see table 1).

3.2
The method: Data and analysis The case study design is based on a mixed-method approach (Hollstein, 2011) that combines semi-structured expert interviews, a quantitative survey, document analysis 2 , and participant observation.We interviewed 16 key actors in local politics, public administration, municipal companies, the private sector, and pressure groups 3 .The key issues in the interviews were an assessment of the power of the actors; their potential, and actual courses of action and strategies; general urban development discourses; alternative land-uses and political, legal and technical obstacles for such uses (such as flood protection, zoning law, planning regulations, logistics, or the need to relocate companies, for example).
We analyzed the interviews by applying a category system derived from network theory and frame analysis in order to identifiy the members of the network and their respective priorities for urban development.On the basis 1 This paper focuses on the political deadlock.It's dissolution may be subject for further research and may be adressed in a second paper.

2
We analysed various documents, e.g.city council minutes, newspaper articles, company press releases and publications of the City of Cologne.

3
In 2010 and 2011, we interviewed the urban development policy spokespersons of four political parties; the mayor and his advisors; the head of the Urban Planning Department; an employee of the Economic Promotion Department; two employees of qualitative content analysis (Gläser & Laudel, 2009), we reconstructed the actor network 4 and the network of frames.We employed 'mutual relevance' as criteria to identify the possible network members5 , i.e. we regarded those as network members who attribute significant power on the decision-making process to one another.Our sample comprises all actors that "regard themselves as relevant and consider each other when calculating actions" (Janning et al., 2009: 6 [trans. by authors]).We further identified the frames by coding the text passages of the interview transcripts that contain statements about preferred land-use options and their relevance for the future development of Cologne (see table 1).We related our findings to data from an earlier exploratory case study in which we had the opportunity to participate in sessions of the Cologne public administration and the involved federal state ministries which increased our sensitivity towards the actors' diverging frames and the interpretability of seemingly objective facts.We further triangulated our qualitative findings with a quantitative questionnaire6 .

4
The anatomy of deadlock: Structural and cultural contexts of non-action Which constraining factors can explain the local stakeholders' deadlock in the Cologne Deutz Harbour case?We start by discussing the evidence of the actor network, subsequently focus on the frames, and finally provide a synthesis of both analytical perspectives.

4
Frequently, network analysis is associated with quantitative methods.However, researchers all along applied "less structured approaches to data collection, and interpretive methods in describing and analyzing social networks" (Hollstein, 2011: 404).

4.1
Social structure: The actor network The analysis of the actor network indicates social-structural constraints on both the network level and the level of individual actors.In order to understand the social-structural context of action, we reconstruct the constellation of actors (as mentioned by the interviewees) which employ the most influence on the decision-making process (i.e. the power network).Therefore, relations of power are the basis for the network reconstruction.We assessed each actor's strength of influence through a qualitative analysis which we then triangulated quantitatively (strength of influence, scale 1 to 5: unimportant to very important), resulting in minor adjustments.The actors' position on the concentric circles refers to their degree of influence on the process as a whole (center = high influence, periphery = little influence).By visualizing the most powerful actors in the decision-making process concerned with Cologne Deutz Harbour (see figure 1), a star-shaped network emerges.The city council is the center of the star due to its decision-making power in the eventual selection of urban development options 7 .Its power rests on the constitutional planning autonomy of the German municipalities.And yet, as a heterogeneous collective actor, the city council has maneuvered itself into a blockade since no majority for any of the available options could be mobilized.Several other actors prepare the decision-making of the council's members who, in our case, are almost equally powerful (see figure 1, marked in grey): • The political parties: At the time of the investigation, the SPD (Social Democratic Party, 25 seats) commands the largest number of seats in the council and displays a high degree centrality 8 (total number of relations) and, hence, can significantly shape decision processes.The Green Party (20 seats) forms a governmental coalition with the SPD.Both parties together could take the lead in the decision-making process, however, so far they could not agree on a joint position.The CDU's (Christian Democratic Union, 25 seats) role as political opposition naturally constrains its capability to implement its preferred option that is to convert the industrial site into a mixed-use urban quarter.Moreover, the head of the urban planning department is one of the CDU's influential members and rejects the party line of a conversion of the site.• The mayor: By definition, the mayor is connected to a multitude of actors.
The analysis of his asymmetrical relations, however, suggests that he hardly can leverage his central position to push through his preferences, but rather seems structurally constrained by his centrality.

7
Options such as port modernisation, partial conversion or total conversion.

8
In social network analysis, centrality measures are understood as sources of power as they not only provide access to a larger choice of possible partners and sources of information but also allow to cut off paths of communication (Hanneman & Riddle, 2005).• Public and private sector companies: The public port authority's scope for autonomous action is limited despite its significant economic resources as a proprietor of the majority of the land (76 %) (Stadt Köln, 2008).Local politicians hold multiple positions on the supervisory boards of the port authority and the public utility company.Consequently, the port authority is constrained by policy goals which may overrule the commercial goals.
Although two influential private business companies can lobby for their preferred option, they have been unable or unwilling to speed up the process so far.One company, a private mill company, owns a large property on the site (10.1 %, Stadt Köln, 2008).It adopts a waiting attitude towards conversion or non-conversion as long as the future plans do not jeopardise its manufacturing activities.Due to the mill's noise emissions, advocates of harbour conversion propose a relocation of the mill to another public port.The proposal, however, has not materialized since relocation would incur significant costs.stakeholders but also involves superordinate government levels.On the federal level, a department of the ministry of construction and transport opposes a conversion of the site.The power of this department, however, is limited since it can only sanction the City of Cologne indirectly (e.g. by a non-approval of federal subsidies), since the harbour function has not been specified in any legally binding federal planning document.The power of the ministry is further weakened by internal conflicts regarding harbour conversion in general.On the district level, the Cologne Government Regional Office oversees the lawfulness of the planning proposals since the harbour is an integral part of a legally stipulated floodplain.However, since the German Water Resources Act is open to interpretation, the law does not determine the course of action 9 in an unequivocal fashion, and strategic knowledge resources may influence the decision of the Cologne Government Regional Office.
Although this analysis of the social-structural context provides key insights, it only provides a partial understanding of the tenacious deadlock in Cologne.To fully grasp the dynamics that lead to political stalemate, the social-structural analysis has to be complemented with an exploration of the cultural context.

4.2
Culture: The network of frames In our analysis of the actors' cultural context, we differentiate six key frames which implicate specific urban development goals (see table 1).Rather than prioritising a single frame, the stakeholders non-hierarchically employ several, partially conflicting frames which leads to a rather decentralized network of frames in the cultural context (see figure 2).The nodes represent frames that are connected if they are concurrently rated important by at least two actors (the more nominations, the thicker the relation).In the questionnaire, interviewees rated the frames on a scale of 1 to 5 (unimportant to very important).The matrix was transferred to a network depiction using the software visone.
As to the political parties, the SPD, despite its dominant position in the local polity, has not been able to enforce a particular course of action.The only partial realization of its potential power results from a deep-rooted ambivalence towards the various frames.The propositions and arguments of the SPD simultaneously revolve around the frames 'global inter-urban competition' and 'preservation of a variety of urban milieus' .However, whereas the former frame privileges, amongst others, measures to advance the urban attractivity for highly-skilled knowledge workers, the latter frame rather implies to secure traditional (manufacturing) activities.This ambivalence reflects unresolved conflicts between different factions within the party,

9
This was illustrated at a symposium when lawyers and other experts discussed the legality of a new urban quarter in the floodplain without arriving at a consensus.
i.e. intra-actor ambiguities of framing the second powerful party, the Green Party, displays a similarly ambivalent attitude by simultaneously championing the frames 'global inter-urban competition' and 'climate change mitigation and adaptation' .
As figure 2 reveals, there is no single frame equally prioritised by all actors.On the contrary, we identify a central triad consisting of the frames 'global inter-urban competition', 'adaptation to the service and knowledge society', and 'climate change mitigation and adaptation' .This decentralized network of frames adds to the blockades imposed by the actor network's properties which shape the social-structural context.While, for example, urban development measures to strengthen 'global inter-urban competition' and to promote the 'service and knowledge society' could easily be combined in a conversion and redevelopment plan for the site, the realization of 'climate change mitigation and adaptation' would require a rather different course of action.Firstly, flood protection subjects construction to severe restrictions, and, secondly, the aim to transfer freight traffic from roads to waterways would suggest preserving or even expanding the port function.Obviously, measures in line with the latter frame 'climate change mitigation and adaptation' would impose severe restrictions on a redevelopment scheme focused on premium real estate as the former two frames would suggest.

4.3
Discussion and conclusion: The interplay of the social-structural and the cultural context of action In empirical terms, our case study reveals that the political deadlock can be traced back to both the constraining effect of the actor's relations (social-structural context) and their frames (cultural context).
First, the characteristics of the overall social network constrain actors who would prefer to speed up the decision-making process (city council, port authority).Second, some actors exploit their advantageous structural position to effectively prevent action and, thus, to preserve the status quo (the SPD, the Green Party).To keep all available options, both actors adopt an ambivalent attitude towards the future development of the port area (Padgett & Ansell, 1993).Other actors who advocate an accelerated change are too peripheral to significantly influence the course of action (the local administration, the liberal party FDP and the Christian democratic party CDU) or are too structurally 10 In an abstract and simplified version, figure 1 was published in an earlier paper in order to generally illustrate a network visualisation of power in urban planning (Levin-Keitel et al., 2016).constrained by other actors (the mayor).Third, the actors who prefer a conversion of the site are unable to mobilize a 'discourse coalition' (Hajer, 1995) which would strengthen their overall position with a compelling narrative (Lelong, 2015).Forth, private business actors adopt a neutral stance or exert only insignificant power; civil society actors are not at all involved in the process.The case study, then, reveals that the unresolved conflicts over land-use neither can be ascribed to a single (type of) actor nor a direct confrontation, but to a complex distribution of power and intra-actor ambiguities of framing.
In conceptual terms, our chief aim was to advance a social-structural and cultural network theoretical framework and to gauge the usefulness of such a framework for the analysis of deadlock in urban development politics.Examining the cultural context of action reveals the constructionist character of urban development politics.Urban politics in this perspective is not only about the power of resources but also about the power of stories: getting action, then, critically is "a matter of persuasive storytelling" (Throgmorton, 2003: 126;Lelong, 2015).In the case of competing stories, actors strive to advance their selected narrative into the hegemonial argument (Laclau & Mouffe, 1991;Throgmorton, 2003).Our framework foregrounds the power of controlling the discourse through persuasive narratives and, at the same time, seeks to fully appreciate the influence of material resources and the resulting power relations.
Despite its explanatory potential, the present network theoretical framework has shortcomings which might be addressed in future research.First, our investigation is static and limited to a specific point in time.The analysis without a doubt would benefit from a dynamic framework which accentuates significant reconfigurations of relations and transformations of the framing of the discourse over time (e.g.Lelong, 2015;Nagel, 2016).Second, in order to reconstruct the dyadic power relations and to visualise the structure of the actor network, we categorized relations in a binary fashion (e.g.Actor A has power over Actor B).The current framework thus is based on coarse-grained categorizations and, obviously, would benefit from more granular data on tie strength.Third, the explanatory scope and contribution to the literature of urban development politics and political deadlock is limited due to the single case study design.
Despite these limitations, we are confident that the analytical framework offers a fruitful approach for the systematic investigation of urban politics since it integrates the socio-structural and the cultural perspective on (non-) action.The research on political deadlocks also extends network research which so far has put cohesion, trust and cooperation center stage (Grabher, 2006), while cases of collective blockades and non-action have only rarely been explored from a network analytical perspective.

Figure 1 :
Figure 1: The actors' positions in the network

Figure 2 :
Figure 2: Network of frames 10 /low amount of nomination through actors Links = frames important for min.two actors concurrently Width of links = amount of concurrent nomination through actors Bettina Lelong, Melanie Nagel, Gernot Grabher: Political Deadlock: A Network Analysis of Decision Processes in Urban Politics.Yearbook of Swiss Administrative Sciences 2017: Pages 133-149

Table 1 :
Economic, social and ecological frames and goals

•
Multilevel governmental actors: The debate on the development of the Cologne Deutz Harbour is not limited to the confrontations between local