Shaping Migrants as Threats: Multilayered Discretion, Criminalization, and Risk Assessment Tools

This article examines Operation Migrant, initiated by the Norwegian police following the so-called migration crises in Europe in 2015. One of its central aims was, by predicting challenges related to increased migration, to improve resource allocation and prevent crime. By drawing on research on risk and threat assessment as a form of power, this article aims to analyze how risk categories are distributed and translated into a multilayered institutional arrangement where migration is policed as a potential crime. The article examines the indicators that the risk assessments are based on, and the measures applied and investigates how discretionary practices make immigrants objects for law enforcement and policing. The article contributes to research on migration control in an ordinary police context, where immigration identity checks become part of the crime reduction strategy. Applying the concept of interpretive flexibility (Collins 1981), I will identify the steps in this chain of translation to explore the leap from targeting potentially criminal asylum seekers to targeting broader groups with temporary residency in Norway. The article analyzes the conditions determining how policing, technologies, and migrants are “co - constructed” in a chain of mediation and translation, which reinforces the view of migrants as risky and criminal. The final section discusses how risk and threat analysis is affected by the notion of the “crimmigrant other” (Franko 2020). In Norway, selectively targeting unwanted migrants as criminals has become dominant in police decision-making at a policy level and everyday practices affecting not only third country nationals but also unwanted eastern Europeans.


Introduction
This article examines Operation Migrant, initiated by the Norwegian police following the so-called migration crises in Europe in 2015. One of its central aims was, by predicting challenges related to increased migration, to improve resource allocation and prevent crime. Part of the strategy was to initiate a tactical law-enforcement operation in the police districts. By drawing on research on risk and threat assessments as forms of power, this article aims to analyze how risk categories are distributed and translated into a multilayered institutional arrangement where migration is policed as a potential crime. The article will examine the indicators that the risk assessments are based on, and the measures applied and investigate how discretionary practices make immigrants objects for law enforcement and policing. In Europe, the multilayered legal and governance structure allows legal, policy, and institutional discretion, making national views important for border practices (van der Woude 2020) and bordered penality (Aas 2013). Therefore, the article contributes to research on migration control in an ordinary police context, where immigration identity checks become part of the crime reduction strategy. The prevailing intelligence-led policing model, implemented by Frontex at the European level and by local police districts in Norway, together with high Norwegian performance indicators for deportation, demonstrates why risk profiling has net-widening effects. Applying the concept of interpretive flexibility (Collins 1981), I will identify the steps in this chain of translation to explore the leap from targeting potentially criminal asylum seekers to targeting broader groups with temporary residency in Norway.
The article has four parts. First, I describe the theoretical context of research on risk and discretion and then present the contextual aspects of policing migration in Europe and Norway, especially the social construction of risk. After the methodology section, the results section gives empirical examples of risk assessment tools and decision-making processes in Operation Migrant. The argument will focus on the leap from strategy to implementation into everyday policing practices in Norway. The article analyzes the conditions determining how policing, technologies, and migrants are "co-constructed" in a chain of mediation and translation, which reinforces the view of migrants as risky and criminal. The final section discusses how risk and threat analysis is affected by the notion of the "crimmigrant other" (Franko 2020). In Norway, selectively targeting unwanted migrants as criminals has become dominant in police decisionmaking at the policy level and everyday practices affecting not only third country nationals but also unwanted eastern Europeans.

Multilayered Discretion and Risk Assessment
In this article, the concept of "translation" from science and technology studies (STS) is applied to investigate the processes shaping risk-based assessment, from the European level to the streets of police districts. The concept is important for understanding the steps in decision-making as a "chain of translation" (Callon 1984;Latour 1999: 311). Egbert and Leese (2020: 45) argue that understanding translation in a policing context allows them to investigate the "hinges" of predictive policing processes, such as how data is turned into crime. I will use this approach to look at how risk estimates produced by intelligence become translated, criticized, confirmed, and communicated and how organizational logics bring such insights into policing on the street. The STS concept "interpretive flexibility" will be helpful here in pointing to the "capacity of a specific technology (or other knowledge system) to sustain the divergent interpretations of multiple groups" (Sahay and Robey 1996: 260). The role of user engagement in interpreting and appropriating the flexibility of knowledge systems has been important in this school of thought (i.e., see Doherty, Coombs, and Loan-Clarke 2006;Orlikowski 1992). Therefore, I will examine how different layers and risk assessment tools are brought together to manage migration.
Concepts derived from risk assessment and risk governance theory can help to understand how the multilayered border and use of risk assessment tools are affecting discretionary practices (van Asselt and Renn 2011), here defined as complex decisions taken at different levels (Lempert 1992). The assumption that controlling discretion and decision-making are linked to control at the street level ignores how decision-making often involves many complex judgments (Lempert 1992). Such a top-down approach assumes that change means management strengthening its government control at the street level. A more holistic approach to discretion includes decisions made at the level of legislation and policy as well as of enforcement (Bushway and Forst 2013;van der Woude and van der Leun 2017), and here, digital data processing tools are important. Galligan (1986) argues that the most prominent factor in the "discretionary process lies within legislative and policy decisions," but most studies only address enforcement and, thus, ignore policy-making. In police studies, street-level discretionary powers are portrayed as obstacles to the implementation of new strategies (Gundhus 2017). Therefore, in this article, I will examine discretionary practices and interpretive flexibility at different levels.
Precautionary approaches to managing migration are guided by securitization logics, framing it as something to avoid, and are informed by models of security and the threat nexus (Aradau and van Munster 2007;Buzan, Waever and de Wilde 1998;Huysmans 2000;Ibrahim and Howarth 2017). Van Munster (2009) posits that, through improved information coordination and dissemination, Frontex analyzes migration and border control risks in a way that makes certain migrants objects of suspicion and targets of precautionary risk-management practices.
A strategic, operational, and tactical intelligence operation is deployed in the police districts, supported by a national criminal intelligence system. The National Criminal Investigation Service (NCIS) is responsible for producing intelligence and risk analyses, working in close collaboration with Frontex, the European Border and Coast Guard Agency, and Europol. Combating what Frontex (2019: 7) terms "illegal immigration" is an important area for cooperation locally, nationally, and internationally. The declared aim of Frontex is to prevent, detect, and suppress cross-border crime and irregular immigration, and it declares that intelligence and collaboration with EU states will support decisions on how to achieve this. At the same time, the fundamental rights of refugees are high on its agenda, as noted by several researchers (Aas and Gundhus 2014;Perkowski 2016). 1 However, in practice, Frontex applies a pragmatic management approach to risk analysis, emphasizing state security: A key aspect developed in the CIRAM [Common Integrated Risk Analysis Model] is the adoption of a management approach to risk analysis that defines risk as a function of threat, vulnerability and impact . . . According to the model, a "threat" is a force or pressure acting upon the external borders that is characterised by both its magnitude and likelihood; "vulnerability" is defined as the capacity of a system to mitigate the threat; and "impact" is determined as the potential consequences of the threat. (Frontex 2021) The risk or threat language that is used is instrumental, and, as the quote highlights, vulnerability means threats to borders, not migrants' vulnerability. The indicators generating the risk analysis are, therefore, criticized for protecting states, not migrants (Aas and Gundhus 2014;Gundhus 2018). Applying the precautionary risk perception to migration and border control could thus mean that risk avoidance causes state security to take precedence over human security and human rights obligations (Gundhus 2018;Gundhus and Jansen 2020). Gundhus (2018) points out, for instance, that in using CIRAM to analyze the risks related to migration and border control, the Frontex framework takes vulnerabilities and threats as indicators that lead to migration being seen as a threat. CIRAM-based Frontex operations thus lack a systematic evaluation of the consequences of practices and fail to take migrant security into account. The criminalization of migrants caused by closing borders is obscured. The language used by Frontex, like the formulation of their approach to risk analysis, is intended to influence national decision-makers when they set priorities, formulate countermeasures, and designate operational targets (Frontex 2021). Andersson (2016a) argues: Through its large Risk Analysis Unit and its Europe-wide network for collating data on border crossings (Frontex Risk Analysis Network), Frontex "exports" risk thinking to member state agencies, reinforcing the threat and security frames in operation on [the] national level, while contributing to the prioritisation of migration controls at the external borders above other tasks. (Andersson 2016a(Andersson : 1061 This exporting of risk-thinking is intended to affect discretionary practices in member states. The multilayered arrangements for Schengen members are designed to achieve common standards and share knowledge, information, and communication technologies. Eurosur aims to support the more effective use of national border control resources by creating European uniformity (Jeandesboz 2016). Risk analysis and Eurosur seek common understandings of risks and threats and impose a language of risk (Aas 2011;Bigo and Guild 2005;Ibrahim and Howarth 2017;Jeandesboz 2016). As this research shows, intelligence and multiagency arrangements support decisions to prevent, detect, and suppress crossborder crime and irregular migration. In this paper, we follow how risk assessments of migration are transferred from European to ordinary police districts. However, discretion in decision-making also opens the way for the politicization of police processes. Van der Woude (2020) argues: The very existence of discretion, on different levels, therefore allows states and state-actors to engage in the "scale jumping practices" and "jurisdictional games" to shape present national policies and practices in such a way that they are seen as most beneficial for the wellbeing and security of the country. (van der Woude 2020: 6) Risk analysis revealing border vulnerabilities and threats is, therefore, important for supporting decisions about resource allocation at both the EU and Norwegian level (Gundhus and Jansen 2020;Horii 2016). Given the nexus between migrants and threats, such topics as the deviant immigrant and migration and securitization have been exhaustively studied in recent years (for an overview, see Franko 2020; Huysmans 2006Huysmans , 2014. Frontex has securitized managerial and humanitarian rationalities (Perkowski 2016). It is also deeply involved in national security policies, and Horii's (2016) view of Frontex risk analysis as a "form of power" is significant. Risk analysis gives decision-makers arguments and reveals understandings of what security is, how it is threatened, and what solutions are necessary and appropriate (Horii 2016). As we will see, this problematization sometimes diverges due to interpretive flexibility and local understandings of risks and threats. Therefore, I will investigate what the multilayered structuring of risk intelligence in European networks and bureaucracies means for local understandings of "crimmigrant deviance" and "risky identity" and the resultant exercise of discretion.

Policing Migration in Europe and Norway-A Multilayered Arrangement
The Norwegian police participate in Frontex at the European level. There is only one police service in Norway, and the agencies included in it are the National Police Directorate, the NCIS, the National Police Immigration Service, and the police districts; all were involved in Operation Migrant. In 2014, the Norwegian police adopted intelligence-led policing as a general practice, and Operation Migrant was implemented within that framework as the first national intelligence-led policing project. The Norwegian model of intelligence-led policing is constructed as a business model that defines intelligence as an organizational process ensuring that verified information guides managers' decision-making (Police Directorate 2014). A major objective is to match the use of force and resources to the actual threat, putting actionable knowledge to use.
Since 2012, in larger cities, the Norwegian police have made use of coercive and investigative powers granted by immigration law in criminal cases (Franko 2020). Studies of crimmigration processes in Norway find that officers speak of "changing track" and the advantages of being able to choose between criminal and administrative tracks when policing target groups, depending on their objectives, the groups involved, and the availability of resources (Aas 2014: 524;Gundhus and Franko 2016).
Although there is collaboration with travel agencies on passenger name record systems and customs (Nøkleberg 2019), immigration policing is mainly carried out by the police districts; they are responsible for law enforcement and crime prevention, along with territorial, border, and immigration control. The Norwegian Immigration Directorate is separate from the police and is the civil public administration authority responsible for handling immigration cases. It works closely with the Norwegian Police Immigration Service. The Norwegian Immigration Directorate relies on the police as its executive organ for control and implementing decisions. The Norwegian Police Immigration Service is responsible for registering asylum seekers and carrying out deportations, including forced returns, and runs the only closed detention center in Norway, Trandum (Ugelvik 2016). Almost all aspects of border control, including deportation orders, are carried out by local police units in police districts.
Through its participation in Schengen, Norway's approach to policing migration is closely affected by global and EU border policies and practices. Frontex 2 was established in 2004 but has gradually expanded its autonomy since then, and in 2016, it became the European Border and Coast Guard Agency. It works increasingly closely with Europol and promotes itself as a law enforcement agency combating human smuggling and irregular migration (Franko 2020). Sharing data through "smart solutions" is widespread and ever increasing in controlling mobility (Bigo 2014;Jeandesboz 2016). However, as Dekkers (2019) argues, it varied to what degree the information technology affects the decision-making by border officers since it is perceived too general. It shapes decision-making on a political and higher organizational level, but experience is more important for frontline decision-making.
Through translation and operationalization through the Frontex Risk Analysis Network, the discourse of risk is reinforced and negotiated in a highly political context (Andersson 2016a(Andersson , 2016b. The nature of contemporary forms of cross-border police cooperation demonstrates the expansion of its geographic scope, the intensification of information exchange, and the increased integration of previously relatively discrete systems (see inter alia Bigo 2008;Hufnagel 2013;Jeandesboz 2016).
As argued by van der Woude (2020), these multiscalar arrangements also allow flexibility and a measure of discretion. In Norway, this mainly manifests in how territorial control and national border control are carried out. At the same time, police discretion needs to be limited to achieve the target figures for deportation set by the politicians and the police. The decision-making process is not one-dimensional and static but dynamic and variable. Regarding risk assessment and management practices as a form of power moving across different actors and being translated, I will approach the multilayered system involved in the social production of the crimmigrant other by examining Operation Migrant. The operation was initiated by the police and encouraged by knowledge transfers from Frontex (Gundhus and Jansen 2020). Its aim was to produce risk and threat reports for police management to help them make better decisions on resource allocation, priorities, and tactics in their management of the influx of migrants. A central aim was to predict developments in crime related to increased migration so that resource allocation could be improved.

Methodology
The author was granted access to all the intelligence reports and the strategic and concluding documents describing Operation Migrant, which were used to analyze the production and negotiation of risks and threats. The reports on unaccompanied minors are of particular interest to this article. These reports are confidential, so they cannot be referred to; the empirical analysis is based on interviews. 3 The interviewees from the Operation Migrant project were mostly managers, coordinators, and operational analysts trained as police officers, but also included analysts with academic backgrounds and managers with legal training. They had various roles in the project; some were contracting entities, intelligence managers, information managers, analysts (operational and strategic), and data collectors. Because of confidentiality issues regarding sensitive empirical data, including the intelligence reports, the project was authorized by the National Police Directorate. The topics discussed with interviewees included their role in the operation, how they felt it developed, the deployment of the intelligence cycle, the relationship between intelligence products and the measures, collaboration with other police districts and agencies, and lessons learned from the operation. Interviews were designed to find out how the operation was carried out, different ways of understanding risks and threats, how the actors perceived their roles, how daily routines were developed, how well the overall process was coordinated, and how it supported decision-makers in planning measures. We also asked questions about the content of the thematic reports and the strengths and weaknesses of the intelligence doctrine and technological tools. The move from risks to action plans was discussed in detail. We also participated in a process evaluation of the operation, and this might have increased trust during the interviews, which were marked by a self-critical and constructive spirit. The interviews were examined using a combination of thematic and narrative analyses, which is a good way of looking at accounts of risks since thematic analysis considers "what" the interviewees are talking about, while narrative analysis considers "why" and "how"-justifications and reasoning (Sandberg 2010).
This study is part of research exploring shifts in the understanding of policing and risk or threat assessment, which is why the findings are analyzed in a broader context. This broader project includes focus group interviews conducted between 2017 and 2018 with 24 police officers (12 in frontline patrols and 12 in frontline investigation) and 16 key informants 4 following police reforms initiated after the terror attacks in Oslo and Utøya on 22 July 2011. Also included are fieldwork and interviews with ordinary patrols and special immigrations units in the police districts and a project using intelligence measures to prevent youth crime in an area of Oslo (n = 35). 5 The broader project provides data, through retrospective interviews, on how Operation Migrant was operationalized on the ground.

Policing Migrants-Policing Crimmigrant Others
Drawing on interviews with participants in Operation Migrant, I will first introduce the strategic part of the intelligence operation and then describe how it was carried out in the police districts. As mentioned, the operation was initiated by the National Police Directorate in response to the influx of migrants in the summer of 2015. This influx reached a peak of 31,145 asylum seekers, 70 percent arriving between September 2015 and January 2016. 6 All European states responded to the increased number of migrants by developing ways to combat irregular mobility, and territorial borders were closed. Talk about a "crisis" raised concerns about the security of nation-states amid increased anxiety about terrorist attacks and issues regarding crime and public safety that might result from the influx and lack of identity checks (Gundhus and Jansen 2020). Public opinion was marked by an overblown fear caused by the high number of migrants. Operation Migrant, according to the interviewees, was, therefore, initiated as a rational and "cool" response to a heated and emotional debate in a "hot" atmosphere.
To a certain extent, the National Police Directorate followed the initiatives taken by Frontex. Some of the information gathering was carried out by Frontex, and the Norwegian operation made use of Frontex risk assessment products, which calculated the likelihood of an increased number of migrants arriving in Norway. With the core aim of intelligence being to reduce decision-makers' uncertainty, Operation Migrant was conceived as a national intelligence project; it was, therefore, necessary to have a common knowledge base so that resource allocation could be better targeted in the way described by this interviewee: Ideally, we should help to anticipate threats. The Police Directorate wanted something they can use to base their decisions and priorities on. And they wanted information about events before things arise, which is not so easy. We also used intelligence reports from the other Nordic countries, and reports from Frontex and Europol. (NPIC 9) As we were told in the interviews, the intelligence hub at NCIS had first to provide figures and predictions about the number of asylum seekers or migrants arriving in Norway, their nationalities, and the percentage of single unaccompanied asylum seekers, and how this compared with the figures in other European countries. The second step was to provide reports on the number of asylum seekers or migrants expected to arrive in Norway and on which border crossings were being used and expected to be used in the future. The third question was about the crime and public insecurity that could be expected from migration: did crime result from increased inflows, and did they affect public safety? Could the police see any trends or instances of crime connected with the flow of migrants, such as disturbances or unrest in the rest of the population (Police Directorate 2015)? While considering these questions, the intelligence hub was required to report weekly on police resource utilization and resource needs, requirements for coordination, the use of police effort, and needs for additional grants to handle the migrant situation, for example, by closing borders. A police analyst describes the difficulty of identifying threats: There were several concerns about who the people coming were. Were they terrorists? The focus for a long time was on whether terrorists hide among immigrants. Many of them are traumatized, come from areas where certainly, or at least very probably, they have been involved in a war. There are questions about the person's identity and there is a lack of certainty about it, etc. These were issues that one was supposed to try to include properly in reports. (NPIC 9) The intelligence hub identified 132 information requirements to provide intelligence on these matters, which police districts had to report on. Some of these were perceived as rather vague: In the beginning a lot was unclear. There were a lot of information needs that the Police Directorate required intelligence about, which did not necessarily prove to be anything real you could collect. There was no information on it, or one could not be as detailed as they required. (NPIC 9) This category included, for example, information about unstable people who had the capacity to cause great harm and about named terrorist groups. The main sources reported were internal intelligence items from police databases in the police districts. Other sources were social media, news, and reports from police associates in Europe, particularly Frontex and Europol, Interpol, embassies, and the Ministry of Foreign Affairs. The information gathered was coordinated by the intelligence project's information manager, broken down, and distributed to strategic analysts. Weekly reports were requested on various types of crime to identify the potential crimes that could be related to migration, such as human smuggling, false ID, drugs, and crime among unaccompanied minors, together with information on the prevalence of vigilante groups. Weekly reports were made on the likelihood of rises in the numbers of asylum seekers and on the increased crime related to this. Irregular migrants, including those with legitimate protection needs, were defined primarily in terms of their risk qualities-as threats. "Risk" was rarely mentioned by the Norwegian police. However, threat is a commonly used concept, as this quote from one of the project's information managers put it: But what we are good at is threats, but even that . . . I just think we do not theorize this very much. It is really about the social mission of the police, who may be doing something wrong? (NPIC 4) The intelligence hub collected data to predict whether future crimes or threats to public safety were likely caused by individuals or groups. The obtained data were generally collected manually without much support from software systems. To a certain degree, the idea of the border as being threatened has leached into perceptions of migrants as possible criminals due to the police logic referred to above, which reinforces perceptions of migrants as "crimmigrant others." However, participants from the intelligence hub felt strongly that their main task was to "negate concerns" coming from within police institutions and elsewhere, such as politicians, newspapers, and other media. Thus, an important finding was that, since public opinion was marked by exaggerated fears arising from the increased number of migrants, the intelligence hub saw their role as being to dispel myths and provide reassurance: One can see, at least the alertness was taken down. This kind of report where you can say, "Okay, that's what we actually know, we know about it," it creates a sense of security. (NPIC 8) They saw their mission as providing a calmer, more balanced, nuanced, and correct picture of the situation (Gundhus and Jansen 2020). They also felt that their reports helped calm fears about the big issues of crime carried out by youngsters (both violence and drug-related offenses), sexual harassment by asylum seekers, and crime connected with unaccompanied minors escaping from reception centers. This sensemaking of the situation also informed how the situation was interpreted. To a certain extent, they felt that, by being involved in the operation, they functioned as brakes on populist politics and the culture of control resulting from fear.

The Making of the Crimmigrant Youngster
Operation Migrant issued five thematic reports on sexual offenses; human trafficking; human smuggling; international war crimes; and the potential for serious violence, including that carried out by mentally unstable people and drug addicts and for serious damage. A sixth thematic report commissioned by the Ministry of Justice dealt with the situation of unaccompanied minors.
The interpretation of the potential criminality of unaccompanied minors makes clear how decisionmaking is seen differently by the intelligence hub and by managers. "Why are unaccompanied asylum seekers escaping from reception centres?" was a question asked in the sixth report, which illuminates dilemmas concerning not only such central concepts as vulnerability and threats but also potential criminality. On one hand, young boys are seen as threats to public safety, because they may commit sexual harassment, burglary and other types of crime. On the other, those identified as children are vulnerable and have special needs because of their violent experiences and tremendous losses. This report was one of the few making extensive use of empirical data and other research. As we heard from interviewees, its recommendations changed how the problem was defined: it is not the youngsters who are the problem, but the way they are treated. They concluded that current immigration policy is part of the problem. Previously, minors received permanent residence permits. It was the introduction of temporary residence permits, respondents argued, that made them escape from reception centers.

Translation from Strategic to Operational
According to the intelligence doctrine, managers were free to make decisions on risk analysis or anything else, and how far this changed the deployment of resources is discussed in interviews (Gundhus and Jansen 2020). One interviewee responsible for the operation said: No matter how good the product you get describing threats . . . So, then it's not that easy for a decision-maker to take decisions. Even if the products are great at describing the threats and great in terms of assessments linked to the threats, it is not certain that the decision-makers will use them. (NPD 1) Interviewees also explained that it was the strategic reports that revealed an internal need for the NPIC to provide operational and tactical support for decision-making in police districts to ensure that knowledge from the intelligence reports had been acted upon. Seeing risks in reports also brings responsibility for acting and taking precautions: The reason for this was that they saw that they had knowledge about issues that they clearly believed that the districts should take measures against. NPIC didn't want to sit on this knowledge without making sure that this was shared. (NPD 1) The operational part of the project was led by a manager supporting police districts with the knowledge necessary for coordinated tactical efforts and operations, and they said this about the reasons for launching this part: What I think was in the background was that, when they read the strategic products, the decision-makers started to think on top of that, "Okay, what if we are sitting on some potential ticking time bombs and we know about this now." (NPIS 8) As emerged from the interviews, risk and probabilities were not concepts in common use, which led to dilemmas, particularly in the operational project. Those put in the position of developing measures and acting on the recommendations were mainly familiar with such law enforcement measures as investigation and combating organized crime.
The efforts initiated, therefore, resulted in several investigations of criminal cases, also leading to the deportation of noncitizens. Another type of action taken was order maintenance by increased patrolling at "hot spots." 7 According to the manager of the operational part of Operation Migrant, not enough crime prevention measures were deployed. Actions were mainly reactive, leading to several investigations and criminal cases and the policing of asylum centers at risk.
The steps taken were at odds with advice in the report about unaccompanied minors. For example, contrary to the intelligence hub's suggestions that there should be more crime prevention efforts helping the minors, the chosen measures were police controls of open drug scenes, where young offenders with no formal connection to the city were part of the target group. The police followed a harsh strategy focused on deterrence to control what was seen as a disorderly and violent drug market. Young people escaping from asylum centers were indirectly part of the target group (Gundhus 2020). In addition to feeling the need to police the open drug scene in Oslo, this approach was influenced by the desire to achieve deportation performance targets (Franko 2020; Gundhus 2017). One way to obtain good statistics was to increase the number of ID checks. Police officers on patrol were required to report all minor offenses and public nuisances if the suspect might be a foreigner, which made it possible to increase the use of immigration law in tandem with criminal law (Gundhus 2020;Lundgaard 2019). The lack of connections, a high degree of mobility, unknown identity, and uncertainty about age and residence status all triggered from Operation Migrant control efforts differed from the soft policing normally found in youth crime prevention (Gundhus 2020) and went beyond more inclusive multiagency collaboration. They were aimed at milieus frequented by both adults and minors, and minors were not recognized as such (Lidén and Salvesen 2016;Tyldum et al. 2015). Such approaches are important in the social production of the "crimmigrant other." This situation goes against the rights of children and young people in Norway, which are supported by the UN Convention on the Rights of the Child, whereby children under the age of 18, particularly unaccompanied refugee children, have protection and rights. There are, thus, gaps in the legally required provision of protection to young people (Gundhus 2020). Moreover, the social production of the crimmigrant other justifies penal sanctions (Franko 2020). Indeed, collecting data for intelligence is framed by the risk and threat aspects of the police mission. By exclusively using law enforcement logic and relating it to combating organized crime, this strategy can further obscure migrants' needs and vulnerability. It also makes it more difficult to identify risks to them as victims of trafficking, for instance, or people with health problems or unaccompanied minors (Gundhus 2020). The discursive leap from vulnerability to organized crime resembles the one often observed in the discourse on trafficking (Aas and Gundhus 2014;Pickering 2011). As Pickering (2011) points out, combating crime does not meet the needs of victims, and may ultimately reduce attention to their vulnerability.

Dataveillance and Net-Widening
The intelligence-led approach at the national level inevitably affected the scope of policing in the police districts. Following the identification of threats, local police operations were initiated. The national operation developed information needs that had to be met by the police districts, and this meant their performance indicators were combined with more general deportation targets. A major concern, as the number of migrants grew during 2015-2016, was the inability to roll out a comprehensive registration procedure, which affected the identification and protection of vulnerable asylum seekers (Boysen and Viblemo 2018). The introduction of simplified routines, together with the general sense of crisis, led to tighter territorial control (Immigration Law §21) conducted by the ordinary police. This affected those classed as "foreigners," and control extended to satisfy the "information needs" of Operation Migrant. IJCJSD 10(3) 2021 Control efforts were entrusted to traditional policing and regulatory actors, leading to various combinations of border control and traditional law-enforcement rationalities, methods, and objectives (Gundhus 2020). As already mentioned, combining immigration law and criminal law determined which people were put under scrutiny. Minor offenses and fines were used as starting points for deportation processes. At the same time, Operation Migrant required extensive information gathering to satisfy the needs of the intelligence hub. Police patrols were ordered to prioritize monitoring and ensuring migrants' compliance with the conditions of their residence permits, and greater efforts were made to check their identity and legal situation by concentrating on passports and false documents and making arrests for petty crimes such as shoplifting (Franko 2020; Gundhus 2017).
Such state-centered responses reduce discretion and tighten the social control network around crime and public nuisances. The interagency collaborations of Operation Migrant were mainly concerned with law enforcement, the maintenance of order, and collaboration with private security guards. Standard methods were deployed to police areas of risk. The places and persons to be monitored were identified by local intelligence reports, but the risk indicators guiding data collection were decided by the intelligence hub. The interagency collaboration took a pragmatic approach to crime and disorder; getting the job done was more important than legal and philosophical questions about whether issues were crimes or administrative violations (cf. Sklansky 2012;Weber 2013).
This led to a lopsided selection of those to be controlled, determined by nationality, class, gender, cultural membership, and race. It resulted in the detection of minor offenses and fines for disorderly conduct. In Norway, as in France, Sweden, and the UK (Franko 2020), it is striking how shifting discretion from the strategic to the operational or tactical level means that eastern Europeans (usually following their imprisonment) are prime candidates for expulsion, whether they are from EU countries or not (see more in Franko 2020: 118-162). The Operation Migrant focus on asylum seekers, therefore, supported existing orders that patrols should combine immigration law with criminal law to achieve high deportation targets. Strangely, efforts to control those coming from third countries outside Schengen, in practice, turn into intra-Schengen control efforts targeting eastern Europeans. This can be explained by the general political situation, which features strict immigration policies and strenuous efforts to deport foreign nationals and criminal-sanctioned Europe Economic Area (EEA) citizens.
Pre-arrival policing went hand in hand with checking people already in Norway. This particularly affected people from EEA countries who were guilty of minor offenses and antisocial behavior such as begging (see also Franko 2020). The police rely on "creative thinking" as well as the coercive measures available to deal with these people (Aas 2014;Gundhus and Franko 2016). These include arrest and remand in custody, searching of personal belongings or dwellings, together with surveillance and tracking people's networks.
This interpretive flexibility leading to discretionary practice at the policy level has, therefore, made immigration control ever more important in shaping the domestic law-enforcement regimes and ordermaintenance practices of the last two decades (see inter alia Aas 2014;Aliverti 2013;Franko 2020;Gundhus 2017;Leun 2003;Weber 2013;Weber and Bowling 2008). The use of penal power leading to the social production of the crimmigrant other justifies these practices in a welfare state (Franko 2020). Applying immigration law to target potential criminals goes beyond the intention of the law. Diverse processes both inside and outside the country make migration into a penal subject. The penal welfarism described by Barker (2018) is also affected by dynamic interactions with external processes that are not based on the internal logic of the welfare state alone. This move toward criminalizing migrants chimes with the upgrading of the European border guard service to make Frontex even more of a law enforcement tool for detecting and fighting cross-border crime (Franko 2020).

Discussion and Concluding Remarks
As we have seen, police immigration control, where identity checks are part of a crime reduction strategy, is shaping the risk analysis and measures deployed. The adoption of intelligence-led policing-by Frontex on the European level and by local police districts in Norway-together with high deportation targets, helps explain why risk profiling might have net-widening effects.
These findings support Franko's (2020: 4) argument that positioning migrants as penal subjects "is a multi-layered phenomenon, deeply rooted in the intricate connections between law, scientific knowledge, bureaucratic practices, politics, media and popular discourse." Operation Migrant combines policing organized crime and the objective to deport foreigners. Through the association of crime and migration, law enforcement becomes a natural response to the migrants and serves to justify exclusionary and preemptive measures. These findings point at the importance to research the translation involved in applying discretionary processes of risk analysis from European to national and local levels in police districts. Despite variations in the effects of risk analyses, due to interpretive flexibility, there seems to be what Andersson (2016b) terms a vicious cycle of risk, reinforcing the view that migrants are threats and "risky." Refugees are subject to digitalized border regimes and are highly monitored (Kasapoglu and Masso 2021;Metcalfe and Dencik 2019). Inside the country, risk and threat assessment tools target increasing numbers of suspects. The targets are no longer asylum seekers but potentially any foreigner: the net has widened (Cohen 1985).
Analysis reveals further how police law enforcement and order-maintenance efforts are reinforced by the dominant institutional logic in the police. Interpretive flexibility allows for adaptation to policy and political pressures. The message from Frontex that migrants are risky places this group on the spectrum of risk and threats. Constructions of threats are deeply embedded within power relations, as is shown by the fear of invasion by violent young men with a different cultural background. Discourses about the "dangerous immigrant" also carry with them racial ideas about the other, providing more general support for using penal power and helping to change the traditional criminal justice system within the nation-state (Bosworth, Franko, and Pickering 2018: 45-46).
Political directives designed to strengthen the return policy make it harder to argue for inclusive approaches to "young strangers," in line with the Convention on the Rights of the Child. As the welfare services withdraw, the police have more room for maneuver. As in other Western countries, the police are taking on a wider range of tasks, including those traditionally carried out by welfare agencies (McCarthy 2014: 165). The Ministry of Justice and the police have made deportation one of the central performance indicators in the new public management system (Gundhus 2017), setting similar deportation targets to those in countries such as France, Spain, and the UK.
To summarize: first, the performance management system is important for understanding these discretionary processes and how risk is governed. Targets for forced returns of nonnationals with criminal convictions and police districts' work on initiating expulsion cases have played a crucial role in increasing deportations (National Police Directorate 2014a). Second, the figure of the crimmigrant other is also important in triggering law enforcement and making it a natural response to migration (Franko 2020). It legitimizes the suspicious gaze that identifies potentially criminal migrants (Gundhus and Jansen 2020). Discursive and coercive operational practices are, therefore, co-constitutive in conceptualizing and pursuing security in the policing of migration in Norway. Policing migrants as signals of pre-crime not only serves to define the nature of a polity but also, as Brown (2010: 71) points out, offers a promise of protection and a way of distinguishing between friends and enemies by applying a language of criminalization. The risk assessment tools that distinguish members and nonmembers may be essential for the successful ordering of society but necessarily involve a suspicious gaze that produces suspect identities (Cole 2001). As mentioned in the background section, one objective of intelligence-led policing is to match the use of force and resources to the actual threat; it should generate actionable knowledge. This preemptive approach has unintended consequences for migrants' situations; they become disproportionately exposed to police control.
Third, Frontex's risk discourse has provided tools to reinforce the national security response through its language and practice of risk analysis. In the agency's definition of risk, the border is seen as "vulnerable," while the people crossing it are constructed as threats. Through its sizeable Risk Analysis Unit and its Europe-wide network for collating data on border crossings (the Frontex Risk Analysis Network), Frontex "exports" risk thinking to member state agencies, reinforcing the threat and security frames in operation on national levels (Andersson 2016b). However, the definition of risk and threats are also translated, and the way the different meanings of (in)security and vulnerability are related to police security practices at European, national, and local levels provides a framework of multiscalar bordering. The analyses show how the discretionary decisions related to risk assessments at various levels-Frontex, national, and local-are made more or less important, depending on the context in which they were co-constructed.
The article, therefore, speaks to the complexity of multiscalar bordering and how "things get lost/enhanced" in multiscalar translation. Risk assessments are drawn in different directions by different actors, but they lead to hot spots and "hot" people, law enforcement, and order-maintenance efforts. Intelligence-driven agencies produce criminalization not only through institutional arrangements, procedures, and law enforcement measures but also by how notions of security, risk, threats, and solutions are conceptualized. These institutional arrangements are multilayered and translated into national contexts, ideally dynamically, though practically through top-down processes (Paul 2017). As Andersson (2014) comments on Frontex's thought work, it influences EU policy-makers' perceptions (and policies), resource allocation, and member states' access to funding and defines the rationale of its own operations (see also Horii 2016). Risk assessment requires data to be collected, not just aggregated data, but more and more personal data, which then feeds back into risk assessment because of how it is subsequently evaluated (Gundhus 2018). The way risk and crime predictions are co-constructed and context-dependent also directs our attention to the understandings of crime that different actors and software work with and reproduce in policing (Kaufman 2018).
It will be of interest to future researchers to examine to what degree the ambiguities of who the targets are, and the possibility of being flexible about this, place a broader group somewhere on the membernonmember spectrum of risk and threat, with the understanding that everyone within this spectrum might become a suspect in the face of police requests for threat assessments. As Andrejevic (2018: 102) says, risk will score at some point, and being part of the group will pose a potential risk. This pushes the categorization beyond being a disciplinary system that separates deviants from nondeviants and introduces a more actuarial system, placing everyone along a broader spectrum. Potential dangerousness exists everywhere and is conditioned by anything connected to individuals' identity attributes, as it is an inherent part of the actuarial surveillance system (Andrejevic 2018).
Following a preemptive logic, migration control has become more interested in the criminal potential of groups and networks rather than in the individual's past and actual crime that has been carried out; a forward-looking approach has been adopted based on historical data (Gundhus and Jansen 2020). This change leads to a shift from discourses of punishment and moral considerations of legal offenses toward asking who a person is (i.e., toward attributes such as their citizenship, nationality, race, identity issues, and moral conduct). Seeing an uncertain identity as dangerous leads to the use of more coercive methods to secure it; this has been described as the criminalization of identity (Aas 2013;Aradau and Van Munster 2007). Controlling migration as a future risk connects it to crime and terrorism in ways that may make migrants and asylum seekers into potential criminals or terrorists (Gundhus and Jansen 2020). When the police take on new tasks, including ones traditionally carried out by welfare agencies for vulnerable groups, one may ask whether a control device is being constructed that is completely different from that used for the rest of the population, without any discussion of where this leads, what legal implications it has, and whether it is something we can recognize as part of the welfare state.