The Power of Racialized Discretion in Policing Migration

Discretionary practices have often been put forward to explain the racially disproportionate patterns we see in policing. The focus on discretion rather than racism neatly shifts attention away from race and instead towards discretionary practices, which are notoriously amorphous and inscrutable. The attention towards discretion (rather than race) further allows race to operate without being explicitly named and, therefore, to operate as an absent present. In this article, I discuss how race and discretion work together when ordinary police officers are tasked with migration control duties to identify foreign national offenders. Drawing on empirical research conducted in England, I propose the concept of racialised discretion and argue that it holds merit because it recognises that certain discretionary practices and decisions are animated because of race, through race and with the effect (intentional or not) of racially disproportionate outcomes. The article argues for the need for racialised discretion to be seen as distinct from other forms of discretion both in policing and the criminal justice process more widely.


Introduction: Bordering Race
The filtering function of the border has been discussed by scholars (Aas 2011, Aas andBosworth 2013;Mezzadra and Neilson 2013;Weber 2013) as a process by which those who belong or do not belong are identified, selected for further checks, subjected to delays and refused entry by border agents and immigration officers when crossing borders. Despite the ostensible rigidity of these procedures and legislation that governs bordering and police practices, there is nevertheless a significant degree of discretion on the individual officer or citizen's part (van der Woude and van der Leun 2017), and the nature of discretion is complex and at times paradoxical in that it can both extend and delimit officers' decision-making powers (Côté-Boucher 2016). Further, bordering practices are ubiquitous and implemented every day, and social sorting and criminalising procedures are increasingly performed at diffuse sites beyond the physical border and within borders (Yuval- Davis, Wemyss and Cassidy 2017). Practices that criminalise migration have uneven effects on racial minorities who are more likely to be stopped, suspected as illegal and subjected to further checks and apprehension, and these practices have been shown to be entrenched (Bosworth, Parmar and Vázquez 2018). This paper explores the relationship between discretion and race by discussing the findings from an empirical research project that examined day-to-day police involvement in migration control in Britain. Operation Nexus (hereinafter Nexus) is a policy aiming to swiftly remove foreign national offenders in England and Wales by identifying foreign national offenders who have an uncertain immigration status as soon as they are brought into police custody suspected of a crime. Nexus introduced the practice of placing immigration officers into police custody suites or ensuring that ordinary police officers had smoother access to immigration data to verify a suspect's immigration status. If a suspect is confirmed not to have leave to remain in the UK, they are diverted to the immigration system to see if a case can be made for their removal or deportation. In doing this, the suspect is diverted from the criminal justice process to the civil immigration process (Vine 2014). The police and immigration authorities gain significant powers to determine a suspect's future course in the criminal or civil legal systems and many of the decisions involved. For example, deciding whether to check someone on the immigration database, whether to pass a case on to the criminal cases team and whether to assess someone as 'high harm' or dangerous enough to warrant a case being made for deportation all, to some degree, involve a mix of discretionary and obligatory procedures. This example of 'crimmigration' (Stumpf 2006) in practice in Britain has found that discretionary practices are employed particularly towards visible racial minorities. Despite the policy's prescription that every suspect who is booked into police custody should have an immigration check performed, the reality was that the policy was selectively applied at the officer's (police or immigration) discretion and particularly towards visible racial minorities who, according to the officers' perceptions, did not sound or look British (Parmar 2018a).
This article focuses on unpacking the relationship between discretion and race to add conceptual texture to the wealth of research on the role of discretion in a socio-legal context (Bushway and Forst 2013;Galligan 1986;Schneider 1992). I propose that the coalescence between race and discretion is crucial for understanding how uneven practices in immigration enforcement towards racial minorities persist, despite official claims that laws and policies are colourblind and ostensibly neutral in their application (Bonilla-Silva 2018). The article also suggests that when policies and legislative approaches are combined, as in the case of criminal procedures and migration control, there is more scope for discretion and, in this context, for racialised discretion. The aim of this article is not to detract from the rich body of research on legal discretion and discretion in border policing but rather to augment current scholarship about discretion and add nuance by exploring how racism and discretion fuse and multiply through borderpolicing practices.

Socio-legal Discretion
However rigid or precise the law, theory or policy might be, in practice, there is always a certain amount of flexibility, ambiguity or discretion in how it is applied or not applied. The role of discretion has long been the subject of socio-legal analysis, and scholarship has exposed the co-dependency between the law and discretion and by going as far as to suggest that discretion is necessary for the law to function (Galligan 1986). Discretion has often been conceptualised as an umbrella term for a range of actions, rule-influenced actions or showing deviation from the rules. Discretion is sometimes the difference between the formal position and actual practice. It can also refer to 'the freedom, power, authority, decision or leeway of an official, organisation or individual to decide, discern or determine to make a judgement, choice or decision, about alternative courses of action or inaction' (Gelsthorpe and Padfield 2003: 3).
Ambiguity as to whether discretion is desirable or undesirable adds to the complexity of trying to understand it, and the need for social science perspectives to augment jurisprudential analyses for a better understanding of discretion has been proposed (Lacey 1992). Further examination of discretion that allows direct observation and quantification is imperative to go beyond the current legally focussed and hazy place it claims in theoretical and ethical police literature (Nickels 2007). Indeed, an unduly legalistic view of the administrative and political world, some claimed, has narrowed the intellectual horizons that were originally anticipated for the analysis of discretion (Campbell 1999;Lacey 1992). Concomitantly, the understanding of how race and discretion work together in practice remains inscrutable. Instead, the assumption is often, quite simplistically, that discrimination is the outcome of discretion. Other perspectives seem resigned to discretion's inexorability, for example Bronitt and Stenning (2011), suggested that the exercise of discretion inevitably involves discrimination, which can be both positive and negative. The law, therefore, both constrains and enables discretion, and one can envisage that these dual aspects are often in tension.

Discretion, Policing and Border Policing
The growth of police studies in the 1950s and 1960s was given impetus by the understanding that police work was highly discretionary. Goldstein (1977), one of the first scholars to examine discretion in police work, highlighted that it applies in the choices about objectives, methods of intervention, investigative measures and issuing tickets and licences. Davis defined discretion as being when the effective legal limits on a public officer 'leave him free to make a choice among possible courses of action or inaction' (Davis 1969: 4). More recent research has suggested that placing limits on police discretion can interfere with officers' notions of professionalism (Rowe 2007).
The reality of policing in the street is one of discriminatory and selective enforcement (Tieger 1971). A street-level bureaucrat uses discretion when they act in an official capacity based on the worker's judgement about the best course of action (Lipsky 1980). Lipsky's definition identifies an important connection between the application of judgement and, by implication, prejudgment by the officer in his or her exercise of discretion. Prejudgement or prejudice is both conceptually and practically related to racism, literally shaping how we see, categorise and treat racial others (Eberhardt 2019).
Others have argued that in trying to locate where discretion occurs and where it matters, we must consider not only when discretionary decisions are made but also who exercises discretion. Motomura (2011), for example, argued that arrest is the discretionary stage that is particularly significant because this is where enforcement discretion is applied. Markus Dubber (2006) used the term 'front end' discretion to describe where most of the power was located and exercised in procedural criminal law. This description effectively captures the idea that discretion is concentrated in stop and search, arrest and 'no further action' decisions made by actors such as the police. This also shifted attention from the argument that discretionary practices were most concentrated at the trial and judicial areas of the criminal process, which tended to be the prevailing focus of much of the discussion on legal analyses of discretion (Hawkins 1992a;Hawkins 1992b). Research on policing has also suggested that boredom in policework can lead to discretion to disrupt the tedious status quo and implement some agentic control by officers (Holdaway 1983;Phillips 2016).
Against this context, an understanding of Nexus (the empirical focus of this article) was of particular interest given that the policy effectively advances the concentration of discretionary power for ordinary police officers prior to the arrest stage, alongside legitimating racial profiling by providing the sanction for police officers to perform immigration status checks on those they suspect to be foreign national offenders. Police custody is a space where discretionary practices are particularly prevalent and where crucial criminal-justice process decisions are made, bringing together a range of choices that will determine the suspect's future. It is also where the interaction between a suspect's rights and police actions comes headto-head and where the agency of individual officers may be performed through discretionary choices (Skinns 2011).
The expansion of research on borders and bordering over the last decade (Aas 2016) has spurred scholarship that has examined discretion within police practices as part of border control and immigration enforcement. The policing of internal borders has been found to be highly dependent on discretionary power, yet it is a factor that we do not know enough about (van der Woude and van der Leun 2017). An extraordinary level of discretion remains over the production of illegal mobile subjects, yet despite the widespread permeation of discretion, an analysis of discretionary power has not been a major concern of legal theory. Boon-Kuo (2019) suggests that discretion is central and organised in the legal authority of policing migration. Her argument scrambles the notion that a clear distinction can be identified between discretion and the law and instead sees discretionary practices as woven into all aspects of decisionmaking in legal and bureaucratic institutions. Given that discretion is present throughout all parts of legal bureaucracy, controls to limit discretion may not manage discretion, but can also lead to more discretion. Squeeze in one place, and like toothpaste, discretion will occur at another (Boon-Kuo 2019). The deflection or potential for discretionary practices to remain and reappear elsewhere or to go underground is of particular interest in understanding how racialised practices operate and morph and provide a context for the focus of this paper, as will be discussed later.
Research has confirmed that when established criminal-justice practices intersect with migration control, discretionary behaviour capitalises on the blurriness that ensues and can also create moral ambivalence on the part of individual officers implementing the policies (Aliverti 2020). Canadian research, for example, has found that policing migration at land borders and in everyday street policing is inevitably based on race, because officers have less background information to go on in the way that they might, for example, at airports. The lack of agreed definition about racial profiling and ambiguity about what it means in practice has an 'enabling' outcome, whereby racialised practices at the border are protected from scrutiny and facilitate the use of race as a marker of risk at the border (Pratt and Thompson 2008).
The capacity for border policing to enact social-sorting practices informed by stereotypes has also signalled the extraordinary power of discretion, specifically in migration control (Lyon 2007;Weber 2011). In England and Wales, nationality acts as a proxy for race through programs such as Nexus, whereby mostly visible, racial minorities are asked to confirm their British nationality or right to remain in the UK as it is assumed that they do not legitimately belong to the country (Parmar 2018a). Recourse to immigration enforcement is often opportunistic and forms a web of invisible policing in Australia (Boon-Kuo 2019), and global surveillance tactics are multilayered, ensnaring racial minorities through visa checks and no-fly lists (Bowling and Westenra 2018; Nagra and Maurutto 2019). Police officers who apply discretion can feel powerful, 'good' or as though they are challenging a system that does not make sense, particularly when officers recognise that the suspect may have tried to comply with the law. The agentic dimensions of criminal justice agents are also illustrated by officers who have belatedly felt regretful for not having exercised discretion to the suspect's advantage in situations where they could have (Armenta 2017).

Discretion and Race
Exploration of the relationship between discretion and racial discrimination has drawn less scholarly attention. This may be because of the inherent difficulties in pinpointing exactly when discretion occurs alongside the ambiguity of when, where and how race operates (Hodes 2003). The multiplicative aspects of this ambiguity may have stifled research that tries to untangle the dynamic between race and discretion.
In the US, despite evidence about how discretionary practices are shaped by race and produce uneven case outcomes, such practices have been described as impervious to change (Baumgartner 1992). However, the modes through which discretionary practices operate have arguably evolved. For example, recent scholarship on the criminal justice system suggests that racial bias can be explained by processes of prediction and discretion (Baradaran 2013) and artificial intelligence in the form of policing risk algorithms, which often rely on discretionary and subjective decisions (Ferguson 2017). Research has also underlined the need to understand the complex contours through which racial power operates rather than seeing racism as episodic or incidental, as has previously been the case (Murukawa and Beckett 2010). Discretion, therefore, similarly requires a lens that acknowledges its systematic character rather than one that only captures and conceptualises discretionary decisions as exceptional or intermittent.
Analyses of race have suggested that race is slippery and oscillates between reality and non-reality, and shifts and changes and, therefore, has been described as holding an absent presence. Normatively, race is an absent presence because it is excluded from discourse and often operates without being explicitly named and is regarded as a relic of the past (Goldberg 2015;M'charek, Schramm and Skinner 2014a). Race is methodologically an absent presence because of the difficulty at times in grasping it and its ability to remain something that is difficult to put your finger on. Race can be both subtle and explicit in its operation and difficult to trace. These conceptualisations of race have applicability to the ways in which discretion operates and are, therefore, useful to consider in relation to the analysis in this paper. Studies on policing migration in Britain have thus far explained that the focus on nationality legitimises the targeting of visible racial minorities and allows for the denial of the racist premise and consequences of policies such as Nexus (Parmar 2019b). As criminal and immigration enforcement become increasingly co-dependent, race is selectively conflated with citizenship and national origin, highlighting the discretionary and contingent nature of race in the context of crimmigration (Kaufman 2018).
In the aim to understand the relationship between race and discretion in the policing of migration, this article draws on this disparate literature to analyse the findings from the research conducted. The article has considered above whether our current understanding of socio-legal discretion adequately accounts for uneven criminal-justice outcomes regarding race. It also considers the relationship between race and discretion and how this relationship might be conceptualised, illustrated and substantiated to accurately reflect my findings relating to migration policing. The remainder of the article unfolds as follows: first, I discuss the methodology adopted for my empirical research, followed by a thematic discussion of the research findings. The conclusion reflects on the potential analytical value of the concept of racialised discretion for advancing our understanding of how race and discretion collude and interact to produce racialised outcomes that are harder to trace and attribute.

Methodology
Between 2013 and 2018, I conducted empirical research to understand how the police in England included migration duties as part of their everyday work because of the Nexus regime. I focussed on examining Nexus policy by exploring the inner workings of the formal collaboration between the police and immigration officers. In practical terms, Nexus involves immigration officers being stationed in police custody suites so that they identify foreign national offenders who are eligible for deportation more efficiently and swiftly than if they were processed through the criminal justice system. 1 My aim was to understand how policing was affected or indeed challenged by the requirement to verify a suspect's immigration status and if or how this affected ethnic groups in different ways. Overall, I was interested in documenting the consequences of the collaboration between policing and immigration through the lens of IJCJSD 10(3) 2021 race. As part of the research, I conducted 250 hours of observation and carried out over 80 interviews with police officers, custody personnel and immigration enforcement officers. The interviews focussed on asking officers about their views on migration duties being performed by the police, how they felt about the work they did, their relationships with the other agencies and actors and the procedural aspects of their everyday work. 2 I gained access to the police custody suites by contacting police forces and requesting meetings with key personnel and those who had responsibility for developing police-research relationships. I outlined the aims of the research and the significance of Nexus being better understood by all. Police ambivalence towards the policy assisted in their curiosity towards the research and the agreement to be given access to observe the custody suites. While conducting the research, I took fieldwork notes, which were digitised and coded in NVivo along with the interviews and document analysis. The police, immigration and custody detention officers were ethnically diverse, with many people from minority ethnic groups in immigration and detention roles. I have written in detail elsewhere about the complexities of this, given that many of those I observed and interviewed were themselves from ethnic minority backgrounds and had their own histories of migration, as well as my own positioning and experiences as a minority ethnic group woman researching race and policing (Parmar 2018b). All names and places that might identify where the research took place have been anonymised to protect the identity of the participants in the study. However, the pseudonyms reflect the ethnic identities of the real participants. Informed consent was obtained, and all participants were free to withdraw from the research at any time. I did not encounter major ethical challenges, although my position as a woman from a minority ethnic background meant that I encountered uncomfortable feelings in observing the treatment of suspects with whom I shared a migratory history and racial background. 3

The Shifting Presence of Race
As the research progressed, it became apparent to me that race was animated in certain situations and able to maintain its force and produce its effects through discretion in particular contexts. For example, a suspect's case at times was heavily leveraged to justify an extra layer of scrutiny to legitimise the act of further checking his right to remain and to perceive his verbal statement as disingenuous. In one instance, the suspect that the police officer was booking in was a black male who appeared in his mid-thirties, and the police officer, in his conversation with the immigration officer, remarked: Later, when talking with me, police officer Manny explained that police work under Nexus meant that sometimes it was necessary to go over or beyond prescribed rules to get results. 'We have a huge range of options at our hands-we can do nothing even if someone has insecure immigration status, we can pass them on or we can focus on building a case for prosecution and leave it to them to decide', Manny confidently stated. His comment confirmed the choices and room for manoeuvre that previous literature on discretion in police work has highlighted (Davis 1969;Nickels 2007). The reference made to how the suspect looked in the quote above and how somatic features related to the assessment of his probability of holding British citizenship is where the discretionary power was arguably racialised. And indeed, this is perhaps where racial power was enacted. In ordinary situations, a focus like this on somatic features would not be possible, but here, it was clear that Nexus policy and the combination between immigration and policing enabled race to be activated in this discretionary manner. Racialised discretion here was where actions were influenced by racial thinking, stereotypes and assumptions in one way or another and the prejudgement involved in the course of action taken by street-level bureaucrats (Lipsky 1980;Eberhardt 2019).
In another example, it was clear that the stage in the criminal justice process where Nexus procedures were or were not applied provided officers with significant discretion to determine one suspect's case over another and how this was especially dangerous because it could be based on initial assessments of a person such as their race. Stuart, a police officer whom I interviewed, told me: Before Nexus, we used to rarely ask about immigration and actually so early on in a case it wasn't like something that we really focussed on. We'd be interested in whether we could charge someone or not and whether a strong enough case could be made for the CPS [Crown Prosecution Service]. But now, with immigration here, we have another strand, if you like another way in which we can pursue a case and we can choose that on the basis of whether we suspect someone is a foreign national or not. (Stuart [white British male], police officer, interview #32) At other times, it was precisely through discretion that race was masked, tacit and displaced (Wade 2010), yet also able to intervene and influence situations, decisions and choices to act or not act. In this respect, the effect of race created a heavy presence that could be felt if not always seen. In one interaction that I witnessed, for example, four black men were arrested on suspicion of affray and robbery. While being booked in, the officer established that two of the men were from Somalia and the other two stated that they were black British Caribbean. The men were placed in the cells, and the investigation was initiated. The police custody sergeant in this instance talked to the criminal investigation team about the case and recommended that there may be enough evidence for common assault (a more serious charge than affray with a higher chance of a custodial sentence if convicted) and that gang membership ought to be a concern. All four men were also passed on to the Nexus hub team for immigration status checks and potential high harm designation. 4 Thus, race was never directly spoken about but surrounded the case, the atmosphere of the suite and the course of action for these four black men. In contrast, a few days later, I saw six white British men brought in under similar case circumstances; the charge of affray was not questioned, and they were released with no further action. One of this group of six was a black mixed-race man, according to the custody record, which recorded his ethnicity. This is an example of how the discretion to act and not to act was informed by race and how race was visible and present in the former case and yet was cast aside in the latter, perhaps because most of the suspects were white and also because the black mixed-race man somatically appeared white. Scholars have acknowledged the evolving force of race in different contexts and its 'absent present' quality, which captures how race oscillates between reality and non-reality, surfaces in one moment and then hides underground (M'charek, Schramm and Skinner 2014a), or persists in its effects without being explicitly named (Goldberg 2009). In reflecting on the situation with the two similar groups in police custody, I was also reminded of the obdurate presence of race despite and alongside the fact that earlier in the same week, I, along with other police officers from the station, had attended a race equality presentation that underscored the need for officers to ensure fair and equal treatment and to think twice about whether the case may have been influenced by bias at various stages. However, in practice, I saw that recognition of religious and cultural differences was often erroneously elided with anti-racism and by evidencing fair procedure as the following exchange between Nick (fieldnote 309: [white British male] police custody sergeant) and Faye ([black British female] external diversity training consultant) highlighted:

Faye: What anti-racist practices do you consider you currently have in place?
Nick: We adhere to the principle of fair treatment to everyone regardless of the colour of their skin. And we ensure that suspect's cultural and religious wishes are respected as far as is practicable. We provide interpreters, halal, kosher, vegetarian food, we are mindful of suspects who may want to pray and so on.
In my observations, the influence of race and how it was felt and applied in the police custody context varied greatly. Though perhaps an obvious point to underline, race was not constant but rather felt extremely heavy and relevant in some interactions throughout the entirety of some shifts, whereas in others, it had an invisibility or weightlessness felt through moments of convivial interaction. For example, one evening, as the shift turned over, a group of custody detention officers and the police custody sergeant talked about how they were going out for a curry together. 'I've been craving this all week', Nigel, the custody sergeant, said to Anish, a British Asian custody detention officer. Nigel (white British) went on to ask, 'I bet you are not as you're used to the best having home-cooked Indian food every night I guess?' Anish replied and talked about how his grandmother's cooking would put all the restaurant cooking to shame but that it would be so spicy that Nigel would be red as a beetroot if he tried it as his tongue was accustomed to the blandness of English food. They both laughed. This, to me, was an example of the lightness of race-open recognition of the racial and cultural differences between Nigel and Anish in a jovial encounter, yet an interaction that was also tinged with the subtext of cultural assumptions based on race.
At other times, the presence of race was hyper-visible, heavy, exacting and burdensome. A black mixedrace man (Fieldnote 23, mid-twenties) was brought into the station one evening and confronted the police officer who had arrested him as being racist. He shouted to him, 'you stopped me because I'm black, I'm standing here in custody because I am black. My whole life I've been chased by you lot'. These accusations resounded throughout the halls of the custody suite as the suspect continued while sat in his cell. The accusation of racism seemed to weigh heavily on the police officer (black British). He told me he was tired of being called names for just doing his job, it was wearing him down, and he considered leaving his job daily, which is a common sentiment among black and Asian police officers (Holdaway and Barron 1997).
As already discussed above, race had a lightness, an ethereality and something you could feel but not quite put your finger on, also resonant with the absent present metaphor (M'charek, Schramm and Skinner 2014a;Tate 2016). For example, one of the custody suites I visited had many staff from visible, minority ethnic groups, which was often the case in this sector. Custody detention officers were often men from black, Asian or Eastern European backgrounds. The work was often taken up by minority ethnic groups because of the locations of the custody suites in ethnically diverse cities and because the type of work, which was physically demanding, required shift work and was in the lower-paid sector, which was more often taken up by minority ethnic groups. On the face of it, therefore, the make-up of the suites were often some of the most ethnically diverse workplaces (and spaces) I had witnessed. The lack of gender diversity stood in stark contrast to this observation as it was an environment dominated by men. The high level of ethnic minority representation from my perspective was indicative of the lightness of race that I mentioned in that its presence was there, accepted, assumed, and taken for granted in many ways and had a normalised embodied presence, which was silent and hardly remarked upon. This also highlighted the diffuse yet important sources of racialised discretion, chiming with Motomora's (2011) focus on who exercises discretion and where. I saw that sometimes custody detention officers, immigration officers and at other times more senior police officers were able to apply discretion and how racialised discretionary power was further complicated when minority ethnic groups were able to exercise it towards racialised individuals.
In line with findings in previous literature about discretionary acts being displaced (Boon-Kuo 2019), it was clear that the Nexus policy provided discretion where it had been curtailed in other areas of policing. For example, Nigel, an older police officer who had worked in police custody prior to the computerisation of records and booking-in, told me that computer procedures meant that they had to follow all the questions on the screen and that the assessment of risk determined most of the decisions that were made about particular suspects. Although Nexus procedures were also focussed on whether the suspect was deemed to be 'high harm', that is, whether they represented a threat to the public good, the term 'high harm' was largely open to interpretation on the officers' part. Only belatedly in 2017 was guidance given on what 'high harm' meant, and even then, it remained vague and referred to cases where the conduct of foreign national offenders had resulted in significant adverse impacts, whether physical, emotional or financial, on other individuals or the wider community (Home Office 2017).

Layered and Blurred Discretion
One of the aims of Nexus was to foster collaborative work between the police and immigration authorities by placing immigration officers physically within police custody suites in England and Wales (Vine 2014). The intention was to improve the process that police officers had of checking a person's immigration status when they were brought into police custody if they were suspected to be a foreign national. By identifying foreign national criminals more swiftly, they could be diverted to the criminal cases directorate arm of the Home Office, which could remove and deport offenders without it being necessary for them to proceed through the criminal justice process. The imperative to deal with foreign national offenders increased in scale both procedurally and symbolically. Practically, collaborative working between the police and immigration was fostered by Nexus, and, symbolically, the police were vested with a direct mandate to manage migration with having immigration officers for consultation at hand and later the Immigration Act 2016, 5 which granted the police immigration powers that widened their existing immigration enforcement remit. The effect was to blur the criminal and immigration enforcement powers, and indeed, the layers of discretion at the officers' disposal multiplied as a result, which chimed with the findings discussed by Pratt and Thompson (2008).
As established in the review of literature above, discretion in policing is not unusual and, rather, is expected. My research observations, interviews and conversations indicated that collaborative working arrangements between the police and immigration, made possible by Nexus, enabled or augmented the discretionary powers that were already held by the police and was a means through which race was animated but without being detected. For example, when a suspect was brought into the police station, he or she would be 'booked in', and it was during this process that the police custody sergeant had the discretion to engage Nexus procedures or overlook them. On one occasion, a black man was brought into custody and arrested on suspicion of possession with intent to supply drugs. During the booking-in process, the custody sergeant suggested that the immigration officer who was sitting in the station might 'check him'. When I talked with the custody sergeant afterwards, he mentioned that he saw that the immigration officer was not particularly busy and so thought he would ask him to perform an immigration status check immediately. This showed an opportunistic recourse to immigration enforcement, which worked effectively to ensnare racial minorities (Boon-Kuo 2019). A similar situation arose with a South Asian Bangladeshi man who was brought in on suspicion of burglary. When I asked the custody sergeant whether he would have carried out the check were the immigration officer not there, he replied: Probably not, I don't think I would bother with the rigmarole of calling. But having immig there in front of you, means you think, well go on, why not, particularly when someone seems suspicious. Also I feel I have a back-up, a second pair of eyes to look over things. (Corlan [white British male], police officer, interview #28) Being in police custody confirmed well-established findings on racial disproportionality that showed black and Asian men were regarded as inherently suspicious by the police (Bowling, Parmar and Phillips 2008;Parmar 2011;Quinton 2011). They were less likely to be given the benefit of the doubt-or, in other words, less likely to benefit from discretionary moments that were to their advantage or that would lessen the likelihood of the case being put forward for prosecution. For example, the suspicion directed towards young black and Asian men as being caught up in gangs translated to them being arrested for harsher criminal charges, such as assault rather than disorder, or robbery rather than theft. As the examples suggest, this too was discretionary (to frame a crime and its context as more serious) and something that I saw often when the suspects were visible racial minorities. Discretionary power mapped onto racialised assumptions about who may not have British citizenship and a reminder of Sivanandan's familiar trope: 'All of us non-whites, at first sight, are terrorists or illegals. We wear our passports on our faces' (Sivanandan 2007: 48). So often were black and Asian men British citizens but assumed not to be and, therefore, subject to status checks, which at that time were framed as mandatory or a case of simply having to follow procedure. In this respect, one could say that with legal and policy tools such as Nexus, discretionary power multiplied when racialised groups came before officers. The regular forms of discretionary power that they might have had at their disposal would then be augmented by stereotypes founded upon racial differences and assumptions about a person's nationality based on their presumed race.
Immigration officers too engaged their discretionary powers more often in relation to racial minorities, according to my observations. In one instance, I saw an immigration officer perusing a suspect's file following a request from a police custody sergeant. The immigration officer was able to see the suspect's name and details of the case, including the potential charges that the suspect was facing as well as their criminal history (including spent convictions, arrests with no further action and stop and search outcomes). At this point, immigration officers could decide whether to perform an immigration status check or to deem it unnecessary. The discretion afforded within policing regarding proceeding with cases in instances such as these mapped onto immigration powers, meaning that an immigration officer then had double the tools to apply (or not apply) in a given situation.
At other times, crime type was used as a guiding factor at the discretion of the police officer in collaboration with immigration. For example, if a spate of criminal credit card activity was reported, the criminal investigation team would be directed to trawl intelligence from Nigeria, Romania and Lithuania, and, in turn, any suspects would be checked against the immigration database. Crime typologies have been shown to be racialised, seamlessly linking categories of crime with nationality (and, by proxy, race) (Parmar 2019a). In another situation, an Italian national male suspect's case was passed to immigration, and an immigration status check was not deemed necessary. It was hard to locate any procedure or system that was followed with everyone or that identified who was subject to further checks and who was not. The mix of a relaxed approach by immigration alongside a procedurally guided approach at times by the police meant that there were ample spaces for discretion to creep in and compensate where police adherence to rules may have precluded certain types of discretion being applied. When pressed on what indicated whether a suspect should be subject to an immigration check or not, one immigration officer replied: Well, it's not a scientific assessment really, but you have a look at relevant factors-for example, offending history in another country, name, place of birth and then the offences can be a clue-drug trafficking, possession of drugs, firearms can all signal that [immigration] status may well be an issue. (Yacoub [British Pakistani male], immigration officer, interview #18) The reliance on hunches, gut feelings and instinct in police work has been well documented in police research (Dresser 2019;Reiner 2010) and specifically in studies of border policing (Pratt 2010). Public campaigns about reporting suspected terrorist activity have also tried to prompt the same from members of the public in encouraging reporting to the police if something does not feel right (British Transport Police 2016). Racialised hunches, as was the case in the Nexus practices I observed, have been widely cited in critiques of policing that examine racial disparities (Carbado and Song Richardson 2018). I saw that hunches were often used as a justification to legitimise the application of racialised discretion, as one officer I interviewed asserted: I go on my hunch about who to suggest for a check, and yes this might sometimes be based on stereotypes of where someone seems to be from, but 9 times out of 10 I am right, and the system is there to tell me if my instinct is wrong. Like a failsafe, you know. I do have a sort of checklist and the more flags that are raised in my mind, the more things are going to go one way than another. But I can decide if I wanna act it is my choice at the end of the day. I suppose that's the power … you know … to make a choice (or not) that you have when you are a police officer. (Sarnez [Eastern European Romanian], police officer, interview #37).
Underlining the human agency inherent to discretionary behaviour, the quote above also shows how flexibly the officer felt able to respond to his instincts or not. I thought to myself, was it this easy to calibrate who was deserving of more intense checks and who was not? Further, the absent presence of race was detectable in allusions to nationality or where someone seems to be from, as Sarnez's quote underlines. In making this comment, in my view, he was referring to visual, racial characteristics but presenting them in an anodyne way by referring to 'where someone was from'. The point he made about having the power to make a choice or not also underscored the sense of discretionary power he had become accustomed to in his occupational persona.
As mentioned at the beginning of the article, the official procedure promoted by Nexus was that every suspect booked into police custody should be checked for their immigration status. But this was regarded as unrealistic by one custody sergeant who claimed it would be a waste of time checking everyone. Indeed, my observations confirmed that the policy was selectively applied at the officer's discretion (police or immigration) and particularly towards visible racial minorities. It was through collaborative work that the contours of whiteness were also made explicit. For example, some white Europeans were accorded positional superiority-a privileged location that arises from having structural advantage within systems of white supremacy (Said 1979). White German and white Italian foreign national suspects were treated more favourably to white Romanians or white Polish suspects, for example, in that the latter were mostly checked on the immigration systems as part of Nexus policy, and the discretion not to perform further checks was more often disregarded. In this respect, the racialised discretionary practices upheld hierarchies of whiteness and discriminated between different white groups and presumed criminal whiteness. Further, an example of when national and ethnic boundaries crisscrossed involved a black Somali Dutch man, who was arrested on suspicion of a drugs offence and checked as part of Nexus to establish whether he was a high harm offender and indeed whether he was a Dutch national as he stated.
It is in these situations that the collaboration between ordinary policing and immigration enabled and advanced racialisation through discretionary practices and also animated racial power in the sense that it caused movement procedurally in a case. Officers often said they 'just knew' whom to pass on for further checks, and Nexus made it much easier to fulfil this gut feeling. Although some degree of discretion is expected and very much consciously practiced in policing, the concern was when the discretionary behaviour was enacted because of race. Collaborative working, then, allowed discretion to operate and to maintain the absent presence of race, as the examples above neatly illustrate. When I asked one police custody sergeant followed by an immigration officer as to who ultimately made the decision about whether someone was designated as 'high harm', there was a lack of clarity, and the immigration officer said that the police would need to make a case for high harm, and then this would be investigated further by the criminal cases team at the Home Office. What was clear from what I was told was that collaborative working served to make accountability for decisions less visible and ultimately served to diffuse responsibility. Some would be subject to the dual powers of police and immigration, while others simply would not be. Ultimately, this meant that collaborative police and immigration work expanded the possibilities for racialised discretion and reduced individual officer accountability.
As discussed at the beginning of the paper, research has tried to calibrate discretion according to whether it is 'weak' or 'strong' or to gauge where its application is most significant (Dworkin 1977;Galligan 1986;Motomura 2011). My observations of discretion found that when discretion melded with race, it became even more difficult to elucidate and grasp. Ostensibly, it was here and in discretion's messy ethereality that its force resided and where its collusion with racist practices and profiling became most dangerous. Racialised practices such as deciding from someone's non-British sounding name to put them forward for an immigration check meant that racism could hide more easily, as it could be justified as being part of Nexus procedure rather than a racial stereotype premised on the idea that a name sounding south Indian could not be British.

Conclusion: Racialised Discretion Through the Enactment of Borders
Analyses about the role and nature of discretion in criminal justice and migration control have tended to either neglect race or to perceive uneven racial patterns as evidence of discretion. In this article, I have tried to shed light on the need for conceptual specificity that enables the identification of discretionary practices that are inherently racialised. While scholars have previously identified discretion that leads to (or can lead to) racially uneven outcomes, my suggestion is that this is distinct from the type of racialised discretion I have discussed in this article, where racism is used to unlock discretionary processes or where discretion is used to animate or reignite race's presence at a time when explicit discussions about race have become less commonplace. As I have discussed and illustrated in this paper, discretion was used systematically as a compensatory tool to racialise certain individuals and not others when procedural safeguards in police practice delimited autonomy on the police officer's part. The fact that Nexus provided the police with an additional legal and procedural toolkit to access meant that the options for discretionary decisions based on race were wider. This also meant that, in contrast to the idea that discretion was momentary, incidental or episodic, racialised discretion was revealed as much more considered, systematic and embedded.
While this paper has focussed on one context of policing migration policy in practice, and the generalisability is, therefore, limited, the paper, nevertheless, highlights that our conceptual tools need to be capacious enough to capture the serpentine ways in which race operates in the criminal justice process (Murukawa and Beckett 2010) and signals that our current concepts may require critical attention to evaluate whether they are flexible enough to explain race and account for racialised practices. The absent present nature of race described in the paper also underlines the difficulties of pinpointing racialised discretion because of its ability to oscillate between being visible and invisible. As Bruno Latour tells us, to study race as an absent presence, we need to 'hover around it' (Latour 1987;M'Charek, Schramm and Skinner 2014b). Arguably, there is a need for discretion to be 'revisited and retold' through the lens of race, with acknowledgement about how discretion concurrently operates to normalise whiteness and uphold its structuring capacity. Future inquiries into racialised discretion could include investigating discretionary decisions throughout the criminal justice process and measuring their cumulative effects rather than envisaging them as discrete moments. Examination of the negative and racialising effects of removing or curtailing discretion should also be better considered in the criminal justice process. As formal guidelines and algorithmic assessments in the criminal justice system continue to expand their reach, reduction of the capacity for discretionary human agency and the effect this may have on racialising practices needs to be better understood.
The killing of George Floyd in the summer of 2020 at the hands of a Minneapolis police officer and the momentum gathered by the Black Lives Matter movement demonstrate the brutal, explicit and insidious realities of racism alongside the more subtle discretionary racialised practices discussed in this paper and indeed the need to understand how subtle and explicit forms of racism intersect and inform each other. We should also be attentive to the ambient way in which racism exists and is sustained, and schemes in the vein of Nexus are one example of how race is enacted in everyday police practices under the premise of bordering. Policies such as Nexus effectively deny the racial power that they wield by being framed as focussed on nationality and serious criminality, which in turn legitimises their basis and their enforcement. The close examination of Nexus policy, as shown in this paper, highlights the fact that bordering practices across different scales of analysis-from the street-level to middle-level decision-makers and from the state-level to cross-national policies and legislation-can sanction and animate racialised discretionary practices. Forms of discretion that effect racial power in this way need to be better understood through research that looks across the criminal justice system and that captures, unpacks and challenges the dynamic between race and discretion.