Security Governance – An Empirical Analysis of the Norwegian Context

Introduction¹

Since the mid-twentieth century, the field of security governance has been characterized by changes and transformations. Once the monopoly of the state, the governance of security around the world is now characterized by plurality (Crawford, Lister, Blackburn, & Burnett, 2005; Johnston & Shearing, 2003). State and non-state, commercial and non-commercial actors operate alongside each other in the production and delivery of security. This has led policing scholars to speak about network and partnerships (Dupont, 2004; 2006a; Dupont & Wood, 2007; Fleming & Rhodes, 2005) between actors in order to enhance efficiency in the security governance. Internationally, the polycentric and networked security governance has in recent years received a substantial amount of scholarly attention. Thus, a range of national contexts have been explored and examined empirically (e.g. Australia, South Africa, the Netherlands, and Britain), which has resulted in important knowledge about different systems of security governance (see for instance Dupont, 2006b; Froestad, 2013; Marks & Goldsmith, 2006; van Steden, Wood, Shearing, & Boutellier, 2013; Wakefield, 2003). The Norwegian context of pluralized security governance, however, has not received as much empirical attention as the international sphere (see Larsson & Gundhus, 2007; Lomell, 2007; Myhrer, 2011 for some notable exceptions). But these studies do not examine polycentric and networked security governance in detail. Accordingly, it can be argued that there is a need for more empirical studies of the Norwegian context. One of the main purposes of this article is therefore to contribute with an empirical analysis which enhances the understanding of security governance within this context.

In this article, I examine the governance of security in Norway, and the question I address is what characterizes security governance in this Norwegian context? In order to highlight the question I have conducted a case study of local security governance in the city of Bergen. The first section is concerned with the contextual and theoretical foundation of the article. The focus is on a brief description of the development of the pluralized security governance, and the concepts of nodal governance and state anchored pluralism are elaborated. This is followed by methodological consideration in relation to the case study. In the main part of the article, the attention is turned towards the empirical investigation of Norwegian security governance. I have made a descriptive account in which my emphasis is at the business of different actors involved in the production and delivery of security in Bergen. One guiding question within this part has been; how do the public and the private actors work with security in relation to crime? This examination is then followed by an assessment of the collaboration and establishment of partnerships between state and non-state actors.

Changes in security governance – From a dream of state monopoly to pluralization

Security governance consists of two distinct concepts, security and governance. These two concepts will in the following be shortly elaborated. Within the field of criminology the distinction between objective and subjective security is highlighted as significant (Aas, Strype, Bjørgo, & Runhovde, 2010; Egge, Berg, & Johansen, 2010; Johnston & Shearing, 2003; Zedner, 2003, 2009). The distinction thus serves as the point of departure in this article. Therefore, when I speak about security, it refers to the objective state of being without, or being protected against threats, and it is used to describe the subjective state of freedom from anxiety derived from threats. Furthermore, it is crime that is considered to be the threat against security; that is what Baker (2010) consider to be «personal or citizen security». It is emphasized that the subjective condition of security is as important to most of us as any objective state of security (Johnston & Shearing, 2003). Thus, in order to effectively govern security, measures must both meet the subjective perception as well as the objective threats to security. The second concept, governance, is understood rather broadly as the «intentional activities designed to shape the flow of events» (Wood & Shearing, 2007, p. 6). Put together, security governance refers to «programmes for promoting peace in the face of threats (either realized or anticipated) that arise from collective life rather than from non-human sources such as the weather or threats from other species.» (Johnston & Shearing, 2003, p. 9).

During the last decades, one can identify significant changes and transformations concerning all spheres of governance (Burris, Kempa, & Shearing, 2008; Kersbergen & Waarden, 2004), as traditional mechanisms of governance have been challenged and new arrangements have emerged. In the context of security governance, which is the task of this article, profound changes have also occurred. Within traditionally conceptions of governance – e.g. the Hobbesian ideal – governance was perceived as the prerogative of the state (government) (Burris, 2004), and it was seen as a top-down affair. According to Wood and Shearing (2007, p. 8) this concept of governance can be characterized as state-centered involving the paradigm of «governance through force». Therefore, within this Hobbesian understanding the dream of the state was to effectively monopolize the field of security governance (Bayley & Shearing, 1996; Shearing, 2008; Shearing & Marks, 2011). Furthermore, it was the use of legitimate force to establish and maintain order that was sought monopolized. Weber (1946) followed up Hobbes’ line of thought when he defined states in terms of a legitimate monopoly over use of physical force to impose social order within its spatial boundaries. What is important to note here is that this monopoly has been vested – at least in terms of internal threats to security – within the police institution. The establishment of the London Metropolitan Police or the «new police» in 1829 by Sir Robert Peel is often viewed as a crucial step towards the realization of this state monopoly over security governance. Throughout the years since the inception of the new police it is argued that the police have done very well in realizing their dream. Indeed, so well that policing/security governance became synonymous with what police did. Shearing (2008), for instance, borrows Nils Christie’s phrase when he describes how well the police have done; the police now «owned» policing. Thus, security provision from the nineteenth and through the second quarter of the twentieth century was successfully monopolized by the police, inspired by a Hobbesian ideal of state sovereignty (Froestad, 2013; Wood & Shearing, 2007).

This view, however, has been challenged. One can no longer conceive of the state and their police as the monopolist of security governance. Now, in most modern societies, a whole range of non-state actors has undertaken the tasks of governance. Thus, state and non-state, commercial and non-commercial actors operate alongside each other in the production and delivery of security. Bayley and Shearing (2001) argue that those who authorize policing (auspices) and those who are conducting policing (providers) are now being separated from each other. And more importantly, both of these functions have increasingly been taking place away from the governmental domain. This, however, is not to say that states no longer are important actors of governance, as Shearing and Wood (2003, p. 405) point out, «at certain times and places state governments are empirically significant and powerful». Rather, it is suggested that the Hobbesian idea of a state sovereignty can no longer be perceived as true (see Foucault, 1990; Rhodes, 1997; Wood & Shearing, 2007). Loader (2000) put forth a similar point when he notes that the changes can be characterized as a shift from the police to policing. As a result of the restructuring within the security field, one can argue that the world of security governance is now being viewed and characterized as polycentric and fluid (Bayley & Shearing, 1996, 2001; Crawford et al., 2005; Johnston & Shearing, 2003; McGinnis, 1999; Wood & Dupont, 2006b), consisting of multiple actors operating alongside each other under the operations of governance.

Regarding the pluralization one important question arises; what can explain the expansion of the policing sphere? Within the policing literature, a range of possible answers has been highlighted. Jones and Newburn (1998), for instance, emphasize two sets of explanations; fiscal constraint theory and structuralist theory. The first is related to the assumption that the state has limited resources at its disposal, thus the police are experiencing fiscal restrictions leading to an increase in other policing actors. A further refinement of this argument can be linked to neo-liberal thoughts, that is, a number of governmental functions are being outsourced to non-state actors (Johnston, 2006; Johnston & Shearing, 2003). In terms of policing, neo-liberal thoughts have led to an outsourcing of some peripheral security tasks to non-state actors, and at the same time core functions have remained with the public police. The structuralist explanation is fully elaborated by Shearing and Stenning (1981) and their argument is concerned with the establishment of «mass private property». They argue that when one see changes or shifts in property relationships towards the notion of mass private property, one also finds a shift in the policing arrangement – from public to private policing. Myhrer (2011) points to another cause; he argues that the need or demand for security provision (mainly private provision) has been greater than the modern police have been able and willing to offer. These explanations of the pluralization seem all to be plausible. However, I would argue that the pluralization cannot be explained by one factor alone; it is rather a combination of factors that have led to the expansion of security governance.

As a consequence of the transformations within the policing realm, it is argued that one can identify a decoupling of the police and the state, and more importantly, questions concerning the implications of such a progressive decoupling have been raised (Loader & Walker, 2001). During the nineteenth and part of the twentieth centuries, policing was pursued in the name of a collective or public interest. Accordingly, the development of the modern police was influenced by a principle of publicness. The expansion of the security provision, however, may have challenged this view of policing as a public good. Loader and Walker (2001, 2006, 2007), however, have been occupied with this question throughout a series of works, and they argue that a positive connection between policing and the state can be formulated under contemporary conditions, and their argument is related to the publicness of the good of security. Their argument will be further elaborated below. Another aspect of the decoupling is related to how one frames the relationship between the state and the police and between state and non-state actors. It is argued that the relationships should be viewed as an open empirical question, rather than one that is decided a priori. Thus, by framing this empirical approach, one is open to the idea that important policing actors can arise from all spheres. Furthermore, it becomes possible to conduct a comprehensive (empirical) mapping of actors involved in the security governance.

Nodal governance and state anchored pluralism

The changing patterns of security provision have led scholars to characterize this provision in many western societies as increasingly complex (see Johnston & Shearing, 2003). In addition, one can identify an ongoing theoretical and normative debate between policing scholars about how to govern security in a mixed market of public and private provision. It seems that most scholars agree that pluralism is a general trend (Wood & Dupont, 2006a). But what differs is the ways in which they explain and describe this plurality. Hence, different theoretical frameworks for capturing the conditions for effective and democratic security governance in pluralized and networked societies are presented.

Shearing and colleagues present a «nodal governance» approach as a way of capturing the plurality (Burris, 2004; Burris, Drahos, & Shearing, 2005; Johnston & Shearing, 2003; Shearing, 2006; Shearing & Wood, 2003; Wood & Dupont, 2006b; Wood & Shearing, 2007). Nodal governance scholars argue that one must move away from a state-centered view in order to better understand the pluralized security governance. Thus, within a nodal perspective, no set of actors is given conceptual priority (Johnston & Shearing, 2003). This means that the exact composition and contribution of different nodes or actors should be an empirically open question and not decided a priori. The nodal approach places emphasis on the relationships and linkages between nodes. Networks are, as such, of central importance to the approach. Castells (1998, p. 332) argues that «A network, by definition, has nodes, not a centre». Scholars of nodal governance use this definition of network to focus on the nodes themselves in order to enrich the network theory; «nodal governance is intended to enrich network theory by focusing attention on and bringing more clarity to the internal characteristics of nodes (…).» (Burris, 2004, p. 341). Hence, the approach places its emphasis on the characteristics of nodes rather than on the networks themselves. In the nodal perspective, a node is conceived as «a site of governance exhibiting four essential characteristics» (Burris, 2004, p. 341). First, a mentality refers to a mental framework that shapes the ways we think about the world and the way we decide to act. In terms of the nodal approach a mentality, then, refers to the ways of thinking about the matters and issues that the node has emerged to govern (Burris et al., 2005). Within the security literature, two particular mentalities are highlighted as important – a punishment mentality and a risk mentality (Johnston and Shearing 2003). The first is concerned with reactive strategies based on punishment, reaction and retribution in relation to the wrongdoing (crime), and is connected to the criminal justice system. The latter one, Johnston and Shearing (2003, p. 16, see also Shearing, 2001; Shearing & Johnston, 2005) argue, is connected to pro-active strategies based on a chain of risk, anticipation, and prevention. This risk-based mentality is oriented towards the prevention of problems in the future. Central elements in this future-focused logic are instrumental calculations and techniques intended to reduce crime, danger and risk. In addition, emphasis is placed on the sources of opportunity for the crime to happen rather than on the wrongdoing itself. Second, technologies refer to the set of methods used for exerting influence over the course of events. Third, the implementation of technologies depends in large part upon nodes resources. Hence, resources will affect nodes capacity to exert influence over the course of events. And fourth, a node needs an institutional structure that enables the directed mobilization of resources, mentalities and technologies over time (Burris, 2004; Burris et al., 2005). Since not all nodes are created equal, nodes will therefore vary in their accessibility and their efficacy.

Another position on how to govern security in a world of public and private provision is introduced by Loader and Walker (2006, 2007). Loader and Walker’s position is based upon an idea of security as a «thick public good». Policing has long been characterized as a public good and the ideal has been that there should be universal equity in distribution and provision of this good, that is, security is equally available to all. As highlighted above, changes in the security governance may have challenge this notion of security as a public good. Still, Loader and Walker argue for an understanding of security as a public good. Their argument can be summed up in three dimensions. First, the instrumental dimension of security is concerned with «the sense in which security is seen as prerequisite to the effective liberty of individuals, which in turn is seen as prerequisite to the «good life», however conceived.» (Loader & Walker, 2006, p. 184). This means that security is perceived as a foundational element in the realization of freedom. However, the dimension is situated at the thinnest level. In order to perceive security as a «thick» public good, Loader and Walker argue that one has to move to a thicker level. The second dimension, called the social, is concerned with the notion that security of one individual depends in some distinct manner upon the security of others. Thus, Loader and Walker (2007, p. 161) suggest «That there is a tendency for the quality of security, (…), to be enhanced in the case of any particular individual when the security of those with whom that individual shares a social environment is also reasonably attended to». And more importantly, a system of public provision to guarantee this is considered to be one possible solution. At the thickest level one find the third, called the constitutive dimension of security. What Loader and Walker (2007, p. 162) are concerned with here is how security as a social good «is implicated in the very process of constituting the ‘social’ or the ‘public’». Thus, Loader and Walker (2007) argue that it is only when appreciating these dimensions of security one is capable of understanding the state’s implication in the production of security. Furthermore, they argue that security conceived as a «thick public good» is attainable only if one is able or willing to departure from an a priori state skepticism. In order for collective security to become possible and to be perceived as a public good, the state must be perceived as the main anchor in its provision. Hence, Loader and Walker introduce the concept of state-anchored pluralism, «The state (…) should remain the anchor of collective security provision, but there should be as much pluralism as possible both, internally, (…), and externally.» (Loader & Walker, 2007, p. 193). In terms of external pluralism, Loader and Walker argue, one has to recognize the appropriate place of other sites of cultural and regulatory production above, below and otherwise beyond the state. In this dimension the role of the state, they argue, should be as a meta-regulator. Marks and Goldsmith (2006) follow a similar line of thought as they maintain that the state must strengthen and reaffirm its primacy in the provision of security. Marks and Goldsmith’s argument is based upon the need to improve security for those citizens who are socio-economically disadvantaged and who reside in communities where the informal social control is weak. In their view the state is best placed in terms of capacity, legitimacy and effectiveness to provide equitable policing services for everyone. Thus, Marks and Goldsmith agree with Loader and Walker upon the idea that state should assert itself as the anchor of collective security provision, though at the same time be open to as much pluralism as possible.

There are clearly some differences between the «nodal approach» and the «state anchored pluralism». Nevertheless, both positions – though in different manners – accept the concept of pluralism in security governance. However, one can argue that some scholars are more concerned with the anchor than they are with pluralism and vice versa. According to Dupont and Wood (2006), anchoring comes in many forms. Their view is based on an idea that national context is a decisive factor for the anchoring project. In the context of Argentina, for instance, the anchoring project is a focus for human rights NGOs rather than state institutions (Wood & Font, 2004). But in contexts where states are viewed as «strong» – such as the Norwegian context this article interrogates – the anchoring project may differ, and more emphasis is often placed on the anchoring within the state.

The theoretical framework that has inspired this article – and thus serves as a point of departure – is the concept of state-anchored pluralism. Regarding this approach, it is important to highlight that there seems to be a lack of empirical analyses which examine the conditions for how such a system would function as a good output-producing system in different contexts. This lack of empirical examination gives rise to an important question; what are the conditions of state-anchored pluralism? As viewed above, the objective of Loader and Walker is primarily to defend the concept of state-anchored pluralism rather than to explore the conditions empirically. Hence, this article seeks to make an empirical contribution to the theoretical framework of state-anchored pluralism. However, elements from Shearing and colleagues’ framework are also of great relevance to the empirical analysis conducted here, in particular the notion of underlying mentalities found in the different actors.

Methodology

The empirical analysis is based on a case study of the local security governance in the city of Bergen, Norway. The data collection builds on a combination of qualitative and quantitative methods. Further, the study was designed to identify important agencies involved in the production of security in Bergen and to map the relationships between them. In this study, the main sources of data have been derived from qualitative interviews and a survey, but documental analysis has functioned as a complement. In the period from September 2013 to February 2014, 14 interviews were conducted with key respondents belonging to 6 different organizations involved in the governance of security in the city center of Bergen. Interviews were conducted with representatives of state and non-state organization, and commercial as well as non-profit organizations. Some of these organizations had security as their primary concern – such as the police and private security companies – whilst others saw security as one concern among many others. All of the organizations, however, had experiences and knowledge that informed the research question. Interviews were semi-structured following a general interview guide; however, questions were tailored to the specific interviewee. In addition, follow-up questions on themes that emerged in the course of the interviews were also present. Interviews lasted from 30 minutes to 1.5 hours. During the interviews, a range of themes were covered, such as the specific involvement in the production and delivery of security (e.g. what activities are you intended to maintain, how do you go about these activities); perception of security issues in Bergen and how to solve them; and cooperation and relationships with other actors. Thus, the interviews generated a great deal of data about how the different actors worked with and thought about crime and security. The interviews also revealed much about the collaboration these actors were involved in. One guiding principle for the analysis of interview data was derived from the nodal governance approach, that is, the four characteristics depicted above were used in order to structure the data.

During the same period as the interviews, a questionnaire (survey) was sent out to 32 organizations in order to capture ties or partnerships which these organizations maintained. The data on the security network in this article is derived from this survey. These 32 organizations had in common that their main task or part of their main task was the production and/or delivery of security for the benefits of individuals or other organizations in the municipality of Bergen. Relevant entities were identified through membership lists of professional security associations, the registration data of the police who regulate the security industry, internet searches, and snowball sampling during the qualitative interviews. Thus, an exhaustive list of actors involved in the governance of security in Bergen was produced. Representatives were asked to name organizations with which partnerships – both formal and informal – had been maintained over the past twelve months. The methodological framework for mapping the security network draws inspiration from Dupont’s (2006a) considerations on how to map and model complex organizational sets. These two methods of data collection – interviews and survey – were complemented with a documental analysis. During the data collection, documents from different sources were collected, such as governmental documents, internal documents from organizations, and newspapers. The various methods used in the study served different purposes, but altogether they enabled a comprehensive assessment of the security governance in Bergen and the actors involved.

The municipality and the public police – guarantors of public security

In the framework of state-anchored pluralism elaborated above, it is expected that the state undertake the role as a guarantor of public goods, and in particular, public security. It can be argued that this idea is based on values derived from the rule of law, where the police and criminal justice system are seen as important institutions for citizens’ overall safety. In many parts of the world, however, it is not necessarily clear that the state and the public police undertake such a role – as guarantors of public goods. The reason can be ascribed to the constant threat of corruption and exploitation of public office – which is a feature of governance in a range of countries across the globe. Corruption, it is argued, may undermine good governance and the rule of law, and thus poses threats to the principles of democracy and justice (Graycar & Sidebottom, 2012). In addition, it has been argued above that the relationship between the police and state should be framed as an empirical question. Thus, it is of importance to examine how the municipality of Bergen and the public police perceive their role in society and in particular their role concerning crime and security.

In 2005, through the publication of «police in the local community», the National Police Directorate (POD) promoted a division of work in which the public police would deal with the symptoms of crime, while the underlying causes would primarily be solved by other governmental agencies, particularly municipal agencies. This means that all municipalities in Norway are perceived to be important actors concerning the work with crime and security. As a consequence of this role, the municipality of Bergen has a stated objective that the city of Bergen should be a secure and safe city consisting of secure urban spaces for all its inhabitants (Bergen kommune, 2012). According to respondents affiliated to the municipality, this objective is sought to be achieved partly through a range of services that is directed towards children and adolescents with the aim of preventing crime within these groups (Respondent 12). One of the most important measures concerning the preventive actions directed to the youths is the program named Coordination of Local Drug and Crime Prevention (SLT). According to some of the respondents (1 and 12) the aim of the SLT is that when different actors meet on a regularly basis to exchange experiences and knowledge and to coordinate initiatives and responsibilities, it will result in a more effective crime prevention. However, SLT in Bergen is limited to public actors only. Another area where the local authority of Bergen has the opportunity to affect security is through its city planning and design of physical spaces. There is an assumption that thefts, violence and other vandalism occur where the physical spaces create blind spots in which there are few opportunities for insight and control (Aas et al., 2010). The consequences of this are seen through the Planning and Building Act of 2009 in which the municipality is required to take into account crime prevention when they handle planning and building applications. Thus, the municipality of Bergen may through thoughtful planning of the urban physical spaces affect how its citizens perceive security. The open drug scenes in Bergen are one particularly area where the municipality has started to work with the improvement of the physical space in order to reduce crime (mostly drug-related crime) and enhance security. Part of Nygårdsparken,² for instance, was in 2014 closed to be upgraded and a range of actions is now being implemented (e.g. new lighting, increased visibility) which will make the park a safer place. In both these cases – SLT and city planning/design – inter-agency cooperation is an important feature which will be explored further below. Furthermore, it can be argued that the preventive strategies of the municipality build on two different mentalities – SLT is based on person-oriented prevention, while the urban planning is oriented towards situational crime prevention (Clarke & Mayhew, 1980; Lie, 2011).

In Norway, police services and activities are regulated by the Police Act (Lov om politiet). The main goal and responsibility is derived from the Police Act 1; «The police shall through preventive, enforcing and helping activities contribute to society’s overall effort to promote and consolidate the citizens’ security under the law, safety and welfare in general». This implies that the Norwegian police, in general, play an important role in the society, and that the police have an overall responsibility for providing security to the citizens. The main objective of the Norwegian police is further specified in the White Paper no. 42 Police role and tasks (2004 – 2005). Here, it is stated that the police objective is to contribute to increased safety and security within the local community. According to Aas et al. (2010), the police is the one institution in Norway which is most explicitly held responsible for general security in the society. Accordingly, the Norwegian police are both by law and politics ascribed a role with a particular responsibility – a guarantor – for the maintenance of public security. This is a role that has important consequences for the practices of local police services in Bergen.

When examining the case of the local police in Bergen, one can see that the described objective reflects the core tasks which the police are expected to maintain. One of the respondents, for example, states that their main task is «to maintain law and order, to create a sense of security, to prosecute criminal offences, but also to prevent crime.» (Respondent 2). In the Police Act, it is further stressed that the guiding principle for police work should be crime prevention. Thus, it is thought that public security in Norway is best maintained through the prevention of crime. The leading strategies for preventive actions have in the last decade been «problem-oriented policing» and «knowledge-based policing» (Politidirektoratet, 2002, 2008), where the latter is now the most prominent. Knowledge-based policing within the Norwegian context can be described as the systematic and methodical gathering of relevant information and knowledge, which is then analyzed in order to make strategic decisions about preventive measures with the aim to increase security (Finstad, 2000). According to Gundhus (2009), the gathered information is also perceived to be of a more scientific character compared to the traditional experience-based knowledge. Furthermore, in this strategy, it is a clear expectation that the police to a greater extent should search for knowledge and information outside the internal organization of the police (Justis- og politidepartementet, 2009). For the local police in Bergen, knowledge-based policing has resulted in an increased use of different analytical and mapping tools in order to prevent crime and enhance security, as Respondent 1 articulates

Clearly, the use of analysis, for instance, to really analyze what happens, when it happens and the like, in order to identify crimes and to implement measures against it has now become prominent. And these analytical tools were not previously available.

Analysis of information and knowledge about what types of crime are occurring, where it happens, how it happens, and which actors are involved have become important for the local police in Bergen regarding crime prevention. This suggests that the local police in Bergen want to work proactively with crime, which is also expressed by several of the respondents. This is also in line with one of the main objectives stated by POD during the last decade, that is, to get the local police to work more proactively and less reactively in terms of crime and security. Despite such an ideal, several of the respondents from the police in Bergen report that they still work more reactively than proactively, «We have to acknowledge that we probably are more reactive than proactive in much of the work we do» (Respondent 2). And, «In general, my officers are more on the reactive side; something happens and we get called to the site, and gather the threads from there» (Respondent 3). Accordingly, the scientific-based knowledge derived from PODs ideal seems to be overshadowed by the experience knowledge:

I am the one who is out in the field and it is I who experience it. And should I then hear later on … or should the statistics tell me this and that, for example that it is not that bad. Offences of violence in Bergen has decreased for example, oh well, the statistics says that. Yet, I’m the one who is out there and fight every Saturday night and feel it on my body. Although it [violence] has decreased it is still there. (Respondent 3).

This is similar to what Gundhus (2006, 2009) found in her study: the police staff still ranked knowledge and information based on experience more highly than the scientific. And more importantly, Gundhus suggest that this can be explained by the police culture, that is, real police work is perceived to be bandit-catching and not analyzing information.

Consequently, there seems to be a discrepancy between the expressed ideal of POD and the practices of the local police in terms of crime prevention. Some of the respondents from the police express that a part of the explanation may be ascribed to a lack of time and resources (Respondents 2 and 3). But another, and probably a more important reason, is that the police in Norway are bound by performance measurements which are derived from POD (Respondents 3 and 4). These measurements are used to assess how effectively the police conduct their work, and it influences the allocation of resources to the local police. The police have to report on a variety of reactive measures, however, according to the respondents there are no measurements of the preventive work conducted by the police. As a consequence, the preventive actions are often given a lower priority compared to the reactive work of the police. This strong focus on the reactive actions may be perceived as a paradox: the stated objective of POD is that the police should emphasize preventive actions, yet it is the reactive work on which the police are evaluated, which is also derived from POD. This is in line with what Larsson (2010) observes, that is, the crime prevention perspective of the police has been more concerned with rhetoric and ideology at a management and political level than with the practices of the police.

The findings so far indicate that the municipality undertakes a preventive role concerning crime and security in Bergen. The local police in Bergen express that they want to work proactively with crime and security, however, they are constrained by a variety of factors and consequently the reactive work seems to be given a higher priority. Nevertheless, the municipality and the local police perceive themselves as actors with a particular responsibility for the maintenance of security and crime in Bergen. These actors are therefore perceived as guarantors of the public good of security.

Network and partnerships – the security network of Bergen

A security network is defined as a set of institutional or individual nodes (actors) that are directly or indirectly connected in order to authorize and/or provide security for internal or external actors (Dupont, 2004; Shearing & Wood, 2000). Dupont (2004) presents four ideal types of security network, and one can classify the identified network in Bergen as a «local security network». This local security network comprises 32 actors who work directly or indirectly with local security and criminal challenges in the municipality of Bergen. These 32 actors belong to seven different sectors: public police, public (other public activities), private security, in-house security, voluntary, hybrid, and professional associations. Most often, local networks function as information exchanges concerning local crime and security issues, and emphasis is placed on the resources that can be mobilized to solve them (Dupont, 2004). Hence, local knowledge is of great importance. One important strategy for increasing the amount and quality of available resources for nodes is to enter into partnerships. It is the development and maintenance of these partnerships that represent the «bones» of the network. Thus, partnerships are conceived as (1) physical interactions between actors on issues concerning security; (2) the transfer of material and non-material resources for security purposes, such as information exchange and the pooling of CCTV and communication equipment; and (3) formal security roles dictated through the power and authority.

Figure 1 represents the patterns of partnerships of the local security network in Bergen.³ The actors reported a total of 304 active partnerships, and each line in the figure represents a partnership that was active over the previous twelve months. When exploring the visualization of the network (Figure 1), one can see that it is a dense network and that three of the actors within the public police sector possess a central position in the network – these actors have many outgoing and incoming partnerships.

Figure 1: The local security network in Bergen (spring embedding).

One challenge in examining the visualization of a network is that the larger and more complex the network is, the harder it is to conduct a good analysis (Grønmo, 2004). Due to the large number of actors and partnerships the network in Bergen is considered to be complex. One strategy to overcome this challenge is to make use of statistical measures derived from social network analysis. These measures can be used regardless of the size and complexity of the network (Breiger, 2004; López & Scott, 2000; Wasserman & Faust, 1994). In network analysis one can place emphasis at different levels of analysis; the network as a whole (macro-level) or at the individual actors (micro-level). Measure of density, average geodesic distance and centralization belong to the macro level, while the concept of centrality is a property of a node’s position in a network.

The security network in Bergen has an average density of 30.6 %, which means that 30.6% of all possible partnerships between the actors are considered to be active. To determine what could be considered as high or low density one needs to look at what type of actors are included and the number of actors. The density of the security network in Bergen is considered to be quite high due to the number of actors (32) and their heterogeneity, both in terms of their tasks and size. The density described here is the average density of the network as a whole, however, it is also possible to explore how the density is distributed between and within the different sectors (Borgatti, Everett, & Johnson, 2013). Above, inter-agency partnership was highlighted as an important feature of the preventive work of the municipality of Bergen. Inter-agency partnerships are defined as collaborative ties between public actors (density between the public police and the public sector). The analysis shows that the density between these sectors is 83%, suggesting that 83% of all possible partnerships between the actors within these two sectors are active. This density is perceived to be very high, indicating that there exists close links between the public actors. When examining the density of the network it is also of interest to analyze the connectivity and distance between nodes, more precisely through the average geodesic distance.⁴ The analysis shows that the average geodesic distance in the network is 1.8, which is considered a low figure. This implies that nodes may come in contact with all the other nodes of the network through just under two intermediaries. In such a dense network, where actors have many alternatives to reach the same nodes, the potential for information and resource exchange is high, and the information travels fast through the boundaries of the network. This may help explaining why actors want to belong to a network.

Another important dimension of the network analysis is centrality. The centrality measure is tightly coupled with the concept of power and how power is distributed within the boundaries of the network (Borgatti, 2005; Borgatti et al., 2013; Freeman, 1979). According to Dupont (2006a, p. 175) the more central a node’s position in the network is, the more opportunities it will have, the fewer constraints it will experience, and the more influence it will derive from its position in the network. Nodes with a central position will therefore have a greater potential of influencing the course of events. A range of centrality concepts have been put forth in the literature of network analysis. However, in this study two different centrality concepts is used; degree centrality and betweenness centrality.

The network data used in this article is directed, thus an outdegree and indegree of degree centrality is defined.⁵ The analysis of degree centrality in the network shows that the two actors with the highest outdegree come from the public police sector; one of them has 26 out of 31 possible active outgoing partnerships. The same actor has an indegree centrality of 29. Out of the top five actors with the highest outdegree, three of them are found in the public police sector. Examining the same figures of indegree one find that two of the top five are from the public police sector. Thus, the most central authorities within the network are the public police sector. To further investigate this assumption one can look at the percentage of ties maintained to the public police sector. In the network, 52% of all the actors state that they maintain a partnership with actors belonging to the public police sector. By way of comparison, the average percentage for the non-police actors in the network was 27%.

The analysis of degree centrality is valuable in that it pinpoint the most connected actors, but it does not capture the fact that some actors may be key intermediaries linking other nodes (Brewer, 2014, p. see also; Granovetter, 1973). In order to measure strategically placed actors within the network, betweenness centrality is more useful. The analysis of betweenness centrality shows that the most strategically placed actor is one from the public police sector, with a score of 30.39. The score is remarkably higher than the next one, which is a private security company with a score of 9.34. This statistic indicates that the public police actor possesses an important brokerage function within the boundaries of the network, and it is critical to the overall effectiveness of the network.

Centralization is a measurement at the macro-level, but it is tightly coupled with the centrality concept. A network with a high centralization index implies that the network is dominated by one or a few nodes (Wasserman & Faust, 1994). Hence, these nodes have a greater potential of controlling the flow of information or resources compared to others. Centralization is of interest because the analysis of centrality above indicated that the public police sector possesses a central position in the network. Since the data is directed, the centralization index is also divided into an out- and an in-centralization index. The analysis shows that the security network has an out-index of 56.8% and an in-index of 67%. These figures are perceived to be high, suggesting that the network in Bergen is dominated by one or a few nodes. Due to the fact that centralization and centrality measures are related – and that the public police have the highest centrality scores – it is interesting to explore the centralization index when the public police sector is removed from the analysis. Now, the analysis shows an out-centralization index in the network of 26% and an in-centralization of 29.8%. This is a substantial reduction, confirming that the network in Bergen is dominated by one particular group of actors – the public police sector.

The analysis of the local security network – both at macro- and micro-level – indicates that the public police possess a central position in the network in Bergen. The police maintain a substantial amount of outgoing partnerships as well as having many incoming partnerships. This suggests, to borrow Dupont’s (2006a) words, that the public police are both «in search» and «in demand» of partnerships. A central question, then, is what can explain the central position of the public police in the network? In terms of outgoing partnerships, one can identify two possible explanations. First, in the questionnaire, respondents had the opportunity to come with their own reflections on three questions. It is relevant to highlight the views of the respondents in terms of what role the police should have in collaborative networks. The general perception by the respondents was that the police should have an initiating role in terms of partnerships. Second, POD stresses that the local police in Norway should seek to form partnerships with other actors regarding crime prevention. In sum, these aspects indicate that the public police should have many outgoing partnerships – they are in search. With respect of the amount of incoming partnerships the public police are controlling two essential resources; first, the police possess a monopoly on the legitimate use of force; this is expressed in the Police Act 6. And second, the police control the access to crime-related information through a variety of records, such as the criminal-intelligence register, electronic-criminal records, and the personal-identity register. If other actors want/need to use force (excluding the principles of necessity and self-defense) or need crime-related information they must come into contact with the public police, and thus create a partnership with the police – hence the police is in demand.

Furthermore, the network analysis shows that partnerships seem to be an integral part of the preventive work in Bergen. This resembles what Gundhus, Egge, Strype, and Myrher (2008) found in their evaluation of SLT, that is, collaboration via SLT was important for the overall work with crime prevention in Norway. The evaluation showed that there exist close collaborative ties between the police, municipality and other public actors – suggesting that public participation is a key aspect in preventive actions. This is also in line with the findings presented above. However, the analysis of Gundhus et al. (2008) highlights another element regarding the cooperation, that is, private participation within SLT seems to be lacking. Aspects of public-private cooperation in Bergen will be further explored below.

Through the exploration and discussion of the security network it is illustrated that the public police sector possesses a central position and function in the network, and thus has a greater potential to influence security governance than other actors. While the statistical network analysis is valuable in identifying the potential of influence, it does not necessarily capture whether the potential is successfully exploited by the actor(s). Consequently, one needs to explore whether the potential that the public police possess is fully utilized in Bergen.

Conclusion – The unfulfilled potential of security governance in the city of Bergen

The question which has been addressed in this article and has guided the empirical analysis is; what characterizes security governance in a Norwegian context? In order to explore this I have undertaken a descriptive account and I have analyzed the local security governance in the city of Bergen. To summarize the findings from the empirical analysis; first, the security governance in Bergen consists of several actors, which is consistent with the pluralization. Each of the actors analyzed does possess important resources and expertise which can be mobilized in order to work effectively with crime and security. The two actors examined first, the municipality of Bergen and the local police, perceive themselves as guarantors of the public security, and both have a stated objective to work proactively with security and crime. The analysis, however, indicates that the public police are still situated in a reactive track concerning their work with crime. This seems in large part to be connected with the performance measurements that the police have to report on, which do not focus on preventive measures. Therefore, the crime prevention perspective is more concerned with rhetoric than with the practices of everyday police work. The private security companies are also viewed as central actors in the security governance in Bergen. The analysis showed that these actors have specialized themselves within the preventive mentality and it is suggested that they may be as good or even better that the public police at preventive and future focused measures regarding crime and security. In terms of prevention, however, an important difference has been identified in how these actors (public and private) perceive the role of prevention. The public actors are more concerned with a public interest, while the private actors have located themselves in the prevention of a private interest. One important question in this regard has been concerned with what sort of knowledge and expertise the private actors possess, which could be important for other actors. Here, the analysis suggests that the private actors could increase the «eyes» on the street, but also the use of statistical measures may contribute with important knowledge. Second, the security governance in Bergen is characterized by partnerships and collaboration. This has been explored through the network analysis, and it is suggested that the public police sector possess a very central position within the boundaries of the network. Such a position gives the public police a greater potential than others to influence security governance in Bergen. This potential has also been explored, and the analysis indicates that there is an unfulfilled potential in terms of collaboration in the security network, especially between the public police and the private actors. One important question that arises from this finding is how can one explain the unfulfilled potential in the security governance in Bergen? I will end this article by exploring this question further, and in order to seek an explanation, I will return to the theoretical framework of nodal governance and state anchored pluralism introduced earlier. When the unfulfilled potential is not fully utilized by the actors, is it caused by a lack of conditions for the nodal governance approach or the approach of state-anchored pluralism, or both?

Two of the most important conditions derived from state-anchored pluralism are, first, that the public police must possess a central position within security governance, and second, the public police must be oriented towards the maintenance of a public interest or value. The empirical findings presented in this article have shown that the public police (as well as the municipality) in Bergen view themselves as the main guarantor of public security and that they have oriented their work towards the maintenance of the public interests. Through the empirical analysis, it is also shown that the police possess a central position in the production and delivery of security. Consequently, it can be argued that the security governance is anchored within the public police in Bergen. Thus, the two conditions derived from the framework of state-anchored pluralism are fulfilled, and these conditions are, as such, perceived to be beneficial for the maintenance of security governance in Bergen.

Now we turn to the conditions of nodal governance. Within this approach emphasis is placed at the nodal character of the security governance, that is, the most important condition derived from this theoretical position is the acceptance of pluralism. Dupont, Grabosky and Shearing (2003) seek to identify means of strengthening social control and conflict resolution in weak and failing states. There are clearly differences between the context Dupont et al. examine and the Norwegian situation this article address. However, the idea presented by Dupont et al. (2003) is of importance for the assessment of this condition. One of the main arguments put forth is that the legitimacy and efficiency of a security system can be enhanced if one is capable of combining the qualities and capacities of different actors. This line of argument can also be found in the writing of Ayling et al. (2009). However, in order to be able to mobilize resources of other actors, the actors must first recognize and accept the pluralism in security governance.

In Norway, it is taken for granted that the police are and should be the central institution which maintains security (Larsson & Gundhus, 2007), this is also in line with the framework of state-anchored pluralism elaborated above. However, one important consequence of this close connection between the police and the governance of security may be that it can be difficult to imagine that other actors than the police can be involved in the work of security and crime. As this article illustrates, however, both state and non-state actors are in fact involved in the security governance in Bergen – the governance is pluralized. Through the empirical assessment above it is illustrated that the public police in Bergen do not view the private security companies as equal partners, and the police do not seek to establish collaborations where equality is deemed as a crucial value. Furthermore, it seems that the public police have problems with acknowledging the nodal or plural character of the security governance. This, in turn, indicates that the police in Bergen still possess a strong state-centered mentality regarding the security governance. As a result, the police view themselves as the only possible actor which can and should maintain the security and safety in Bergen. Such a view is thus related to the monopolization of the field of policing by the police. The explanation of this perception is partly related to formalities – such as laws and regulations – but the most important part of the explanation seems to be linked to the above mentioned police mentality. To create equal collaborations with private security companies in terms of security seems to be difficult for the public police to accept or admit. The nodal or plural character of the security system in Bergen is not acknowledged by the public police. Thus, the state-centric mentality of the police seems to hinder a more efficient and legitimate organization of the security governance. In sum, the challenges with security governance in Bergen – and this serves as the reason for why there exists an unfulfilled potential – are in large part related to the lack of conditions of the nodal approach.