Bias Crime Policing: ‘The Graveyard Shift’

Bias crime is crime that is motivated by prejudice or bias towards an attribute of the victim, such as race, religion or sexuality. Police have been criticised for failing to take bias crime seriously, and there is a pressing need to understand the reasons for this failure. This article aims to address this gap by presenting the results of the first empirical study of bias crime policing in the Australian state of New South Wales (NSW). Drawing on interviews with the NSW Police Force (NSWPF), the study found that sustainable reform in this domain has proven elusive. This can be attributed to a number of key challenges including reporting, recording, identification, framing, community engagement and leadership. The lessons that emerge from the findings have important ramifications for all police organisations. This study’s findings the research by showing that police tasked with the day-to-day delivery of bias crime reform place immense value on leadership endorsement. In the eyes of many respondents, the NSWPF executive sees bias crime as an ‘unwanted secret’, a ‘political inconvenience’ or ‘the graveyard shift’. These respondents suggest that the organisation’s commitment to bias crime is more about legitimising its public image with minority communities than setting a genuine agenda for change. These perspectives are corroborated by the lack of a structural framework to support the work of the BCU, such as a dedicated policy, sustained evidence-based training, a formal community-engagement process or a fiscal commitment to ongoing staffing. Frontline officers cannot make sense of new concepts and new procedures without leadership at the most senior level Gulledge and Stewart 2007). Not only do they need to be trained to identify, record and investigate bias crime, they also need to be motivated by their leaders to do so. At a minimum, this requires agency leaders to legitimise bias crime as ‘real’ police work for example, through public statements that are emphatic about its value to routine policing. The findings suggest that this dearth of symbolic and practical support trickles down to a lack of capacity, preparedness and confidence among frontline officers to identify bias crime and comply with recording rules. The findings also suggest that a leadership vacuum about the purpose of bias crime policing can be too readily filled by individual staff preferences rather than evidence-based research on appropriate priorities in the NSWPF vacuum been by an offender-centred approach that with organised groups, victim-oriented all


Introduction
Bias crime is crime that is motivated by prejudice, bias or hatred towards a presumed attribute of the victim, such as race, religion, sexuality, disability or transgender identity. 1 Police have been criticised since the 1980s for their failure to acknowledge and address bias crime (Grattet and Jenness 2005;Human Rights and Equal Opportunity Commission [HREOC] 1991;Macpherson 1999). Some police organisations have responded by introducing specialist bias crime policies and procedures, while others have done little to usher in fundamental change in how bias crime is policed.
Australia is one region where bias crime has gained little traction as a legitimate policing concern. In Australia, law enforcement agencies are largely autonomous organisations that interpret and enforce the criminal law of each state and territory. Although they are subject to political influence, these agencies have considerable discretion in policy formation and implementation (Sarre and Prenzler 2018). Recent research indicates that bias crime continues to be a problem in Australia (Benier, Wickes and Higginson 2015;Dwyer et al. 2017;Mason et al. 2017), and this creates a pressing need to understand the factors that contribute to the failure of police taking this crisis seriously.
Bias crime policing must be understood in the context of police relationships with minority communities. Research shows that these communities have comparatively high levels of alienation from and low levels of confidence in police (Cunneen 2001;Poynting 2002;Sivasubramaniam and Goodman-Delahunty 2008;Tomsen 2010). Community policing mechanisms designed to improve relationships with minority groups have had mixed success, especially when diluted by 'command and control' styles of policing (Bartkowiak-Théron and Asquith 2015;Chan 1997). There is a small body of research that examines the responses of Australian police services to bias crime (Mason et al. 2017;Wiedlitzka et al. 2018), but what is missing from the literature is an examination of the reasons why bias crime is yet to make a 'successful migration' (Perry 2016) to some policing organisations.
This article aims to help address this gap by presenting the results of a study into the challenges of implementing a bias crime initiative in one policing organisation. The New South Wales (NSW) Police Force (NSWPF), which serves the state of NSW in Australia, was chosen as the research site because it has the longest standing bias crime initiative nationwide. Using in-depth interviews, data were collected from a purposive sample of key informants employed by the NSWPF who had experience with the agency's approach to bias crime policing. The study found that the potential of the bias crime initiative to significantly change how bias crime is policed in NSW has never been fully realised. Drawing on these 'insider' perspectives, we argue that the reasons why sustainable reform in this domain has proven elusive can be attributed to a number of key challenges, which include reporting, recording, identification, framing, community engagement and leadership.
First is an overview of the literature that informs the study. Before presenting the results, background on the research site and the methods used to investigate bias crime policing are provided. This article concludes by considering the implications of the findings for bias crime policing more broadly. The key lessons that emerge from the study have important ramifications for all police organisations with nascent bias crime initiatives.

Policing Bias Crime
Social identity mediates the relationship that citizens have with police. Minority groups tend to have less confidence in police and lower levels of trust (Bradford et al. 2017;Murphy 2013). Mistrust in police and disillusionment with their capacity to treat victims with fairness and respect is pronounced among marginalised communities with histories of over-policing and IJCJ&SD 3 Online version via www.crimejusticejournal.com © 2019 8(2) recent experiences of abuse of authority. These include Indigenous people, ethnic and religious minorities, immigrants and LGBTI communities (Cunneen 2001;Miles-Johnson 2016a;Murphy and Cherney 2011;Poynting 2002;Sivasubramaniam and Goodman-Delahunty 2008;Tomsen 2010). However, cooperation between minorities and police requires reciprocal trust (Trinkner, Tyler and Goff 2016). Research shows that, for their part, police continue to hold negative perceptions, suspicions and stereotypes about some minorities (Miles-Johnson 2016b;. This mutual mistrust means that members of minority groups are less likely than non-minorities to report victimisation or contact police if they need assistance (Bradford et al. 2017;Dwyer et al. 2017;Murphy and Cherney 2011). More specifically, bias crime is less likely to be reported to police than parallel crime (Giannasi 2015;Wiedlitzka et al. 2018).
Open and timely collaboration can help overcome the trust deficit between police and minority groups by providing the latter with an 'instrumental voice' (Murphy and Cherney 2011: 251) that enables them to articulate localised concerns. Such collaboration is built-in to community policing models, which seek to recalibrate policing priorities and procedures through key partnerships with civil organisations (Sarre and Prenzler 2018). They enable police to develop tactics for responding, not just to specific incidents, but also to wider problems that engender community perceptions of vulnerability and disorder. This is important in the context of bias crime, in which research consistently demonstrates that only a small proportion of offences involve severe violence by individuals or groups motivated by deep-seated hatred (Tomsen 2010). While the impact of such violence can be profound, most bias crimes involve less severe forms of assault, property damage and harassment that are committed by 'ordinary' offenders when the opportunity arises (Benier, Wickes and Higginson 2016;HREOC 1991;Iganski 2008;Mason et al. 2017). Nevertheless, these 'everyday' incidents can cause inordinate harm because minority groups tend to experience them within a larger continuum of discrimination and hostility (Iganski 2008).
Principles of diversity policing are also applicable to bias crime. Diversity policing centres on the provision of specialised strategies to enhance service provision to marginalised and vulnerable communities (Bartkowiak-Théron and Asquith 2015). In Australia, diversity policing includes the introduction of multicultural policies, cultural sensitivity training and auxiliary liaison programs (Chan 1997;Cherney and Chiu 2010;Miles-Johnson 2016b). Such initiatives can make an important contribution to the capacity of police to respond to the needs of minority communities. However, research shows that progress is undermined by the failure to normalise cultural competency across the organisation or to integrate specific interventions into operational practice (Bartkowiak-Théron and Asquith 2017; Cherney and Chiu 2010;Miles-Johnson 2016b). This makes it difficult to break down the cynicism towards reform that continues to be associated with police culture (Loftus 2010). For example, Chan (1997) documents how both institutional and individual resistance within the NSWPF contributed to the failure of strategies to address a culture of racism in the 1980s. Like all forms of diversity policing, bias crime initiatives are met with resistance and cynicism from serving officers who see them as 'pink and fluffy', in contrast to the 'glamorous and sexy' (Foster, Newburn and Souhami 2005: 91) work in other departments.
There is no one model for policing bias crime. Typically, police have been compelled to introduce bias crime initiatives in response to public outcry about institutional prejudice or acts of extreme violence (Grattet and Jenness 2005;Macpherson 1999). Most bias crime policing initiatives translate bias crime statutes into policy and procedure, but how they do this is variable. In the United States (US), police tend to maintain authority over the way bias crime is defined and recorded, relying on objective checklists geared towards successful prosecution (Mason et al. 2017). Yet, the capacity and preparedness of frontline officers to identify and record bias crime remain core challenges. or any witness perceives it to be so (College of Policing 2014; Macpherson 1999). Third party reporting mechanisms have also been introduced in the United Kingdom (UK), allowing victims to report bias crime through an online portal hosted by a community organisation rather than reporting directly to the police (e.g., True Vision). Recent national guidelines argue that this democratic approach encourages reporting by putting the interests of victims at the heart of bias crime policing and circumventing the failure of police to record bias crime (College of Policing 2014). This is borne from data that show recorded bias crime is far higher in the UK than it is in other nations (Giannasi 2015).
Effecting change through the introduction of democratic styles of policing requires officers to internalise the values that underpin such reform (President's Task Force 2015; Trinkner, Tyler and Goff 2016). Frontline police are better equipped to make sense of, and comply with, new bias crime rules when institutional leaders foster and endorse values of fairness, respect, diversity and community participation as legitimate ideals of the organisation (Bradford and Quinton 2014;Chan 2007). Yet, police organisations have a history of maintaining internal and external control through rigid top-down styles of management (Sarre and Prenzler 2018). While most espouse community-policing values, few services have enacted the organisational change needed to work collaboratively with communities in practice (Sarre and Prenzler 2018). Bias crime initiatives must compete with traditional crime control functions, shrinking budgets and resultdriven measures of productivity (Bartkowiak-Théron and Asquith 2017; Bradford and Quinton 2014;Chan 2007). They must also compete with the recent drift towards more aggressive 'command and control' styles of policing that some have characterised as 'homeland security' policing (Oliver 2006;Sentas and Grewcock 2018). In this environment, policing innovation of any kind becomes difficult (Sarre and Prenzler 2018).

The research site
In 1991, a national inquiry into racist violence in Australia recommended that police develop systems for recording and monitoring bias crime (HREOC 1991). More than two decades later, only two Australian agencies have taken a proactive approach to bias crime: Victoria Police and the NSWPF. The NSWPF first began recording particular forms of bias crime (homophobic and anti-Semitic violence) in the early 1990s. This included the option of third party reporting through community organisations such as the Lesbian and Gay Anti-Violence Project (Asquith 2012). In 2007, following racially driven riots in the city of Sydney (Mason et al. 2017 Officers are advised to flag a 'bias motivation' associated factor when recording suspected bias crime events. The BCU then reviews each record, categorises the bias (e.g., race or religion), assesses the evidence and determines further action. Although bias motivation is not an element of any offence in NSW, it is an aggravating factor at sentencing (Crimes (Sentencing Procedure) Act 1999 s 21A(2)(h)). This places an obligation on police to collect evidence of bias crime.
The NSWPF has a history of poor relations with minority communities (Cunneen 2001;Poynting 2002;Tomsen 2010). Since the 1980s, they have introduced various community-oriented reform programs (e.g., Multicultural, Aboriginal, and Gay and Lesbian Community Liaison Officers). However, research conducted in the 1990s found that reforms to address persistent conflict with ethnic minorities had done little to ameliorate racism among NSWPF officers, largely because they targeted police culture in isolation from structural conditions (Chan 1997). While the organisation has enjoyed periods of enhanced confidence with the LGBTI community, this has been short-lived (Ellis 2018;Tomsen 2010). Current programs that aim to recalibrate service delivery to minority communities have been hampered by the organisation's return to a more authoritarian 'crime management' style of policing in the mid-2000s (Chan 2007: 329;Sentas and Grewcock 2018).

Research methods
The NSWPF was chosen as the site for study because it was the first police organisation in Australia to introduce a bias crime initiative and the only agency to have a dedicated bias crime unit (Mason et al. 2017). The aims of the study were, first, to identify the challenges involved in implementing a bias crime initiative in the NSWPF and, second, to advance our understanding of the impediments to sustained reform in bias crime policing.
In-depth, semi-structured interviews were conducted in late 2017 with personnel employed by the NSWPF. Interviews were the most appropriate method of data collection, as the researchers sought to understand the professional experiences and perspectives of people who had firsthand knowledge of how the NSWPF responds to bias crime (Ezzy 2002). The sample was purposive, and the criteria for selection concerned respondents who were currently or had been recently employed in positions with responsibility for bias crime or communities most vulnerable to bias crime. The pool of potential respondents was identified through collaboration with a designed contact person in the NSWPF. To avoid perceived coercion, potential respondents were invited to participate in an email sent by the researchers. This recruitment strategy elicited 10 respondents (N = 10), which was a high response rate, as only one person did not take up the invitation for interview. Respondents included a mix of personnel in operational, policy and management roles. Together, they represented a diversity of demographic characteristics, including age, gender, sexuality, ethnicity/race and employment experience with the NSWPF (some of which spanned over 10 years). To preserve the anonymity of respondents no further information is provided hereafter.
A semi-structured interview schedule was used, with questions focusing on major themes identified in the extent literature: reporting, recording, identification, framing, community engagement and leadership. Interviews lasted between one and two and a half hours, and were IJCJ&SD 6 Online version via www.crimejusticejournal.com © 2019 8(2) recorded and analysed using inductive thematic analysis to identify recurrent patterns and divergences among respondents (Ezzy 2002).
The study advances the literature on bias crime policing by providing firsthand police perspectives on the barriers encountered in crafting bias crime into a legitimate and sustained policing concern. Yet, the study has limitations. Respondents were employed by the NSWPF and may have felt constrained in expressing their opinions (although, we note that there are some candid responses). The sample was necessarily small, as criteria for inclusion were experience and/or knowledge of the bias crime initiative. Caution must be exercised in generalising from the subjective experiences of this group of specialised respondents employed by one police organisation in Australia. Despite these limitations, the results provide rare insights into the reasons bias crime policing initiatives may flounder. This makes the results germane to all police forces, especially those with nascent bias crime programs.

Results
All respondents in the study spoke of the achievements that have been made since the NSWPF formalised its response to bias crime in 2007. These achievements include approval for the Standard Operating Procedures, establishment of the BCU, inclusion of the 'bias motivation' associated factor in the recording system, and delivery of specialised training packages. Respondents in management positions pointed to examples of quality assurance measures, such as real-time training exercises and 'mystery shopper' tests of customer service. These were said to demonstrate internal capacity for crisis response and day-to-day service delivery in relation to bias crime. Yet, without denying this progress, most respondents observed that the NSWPF has not achieved significant and sustained change in how bias crime is policed. The evidence and explanations for this lack of reform can be grouped into the core interview themes.

Reports of bias crime
Nearly all respondents identified under-reporting of bias crime as a significant problem. Data provided by the NSWPF show that between July 2007 and January 2017, 1,818 reported cases were categorised as bias crime or suspected bias crime. This is an average of 24.9 reports each month for the whole state. 2 It is likely that these reports are only the tip of the iceberg. Research that compares official reports of bias crime with victimisation studies in the UK and the US demonstrates that bias crime is 'massively under-reported and under-recorded' (Giannasi 2015: 332). Drawing on a national probability sample, recent Australian research shows that bias crime is less likely to be reported to police than non-bias crime (Wiedlitzka et al. 2018).
There are many reasons people do not report bias crime, including language barriers, feelings of shame, apathy or failure to recognise the incident as a crime at all (Asquith 2012). Research also points to a lack of trust and confidence in police as a major barrier to complaint-making among minority groups (European Union Agency for Fundamental Rights [FRA] 2016; Murphy and Cherney 2011;Wiedlitzka et al. 2018). The results support this research, in that respondents consistently referred to under-reporting as a sign of the lack of trust that minority communities place in the NSWPF. For example, one respondent stated, 'if the community doesn't trust you they don't talk to you … they cut you off' (R9).
The respondents were acutely aware of the effect of this lack of trust on reporting: We know that the Islamic community is not reporting crime because they're that disillusioned. Nearly half the respondents attributed this trust deficit to the effect of prejudice or 'unconscious' bias among police personnel at all levels of the organisation. This bias was said to be evident in the ways that police treat members of some minority communities. For example, one respondent observed that some racial minorities are treated as perpetual offenders rather than potential victims: 'Once police … stereotype you as a "regular client" it's really hard for them to see you as anything else' (R10). Respondents welcomed a future rise in reporting as a sign of enhanced confidence in the NSWPF.

Recording bias crime
Under-recording has been identified as 'one of the greatest operational challenges' (Giannasi 2015: 333) in policing bias crime. The NSWPF Standard Operating Procedures for bias crime provide guidance for officers in recording such offences. They state that the 'essence' of bias crime is 'the motivation or intent of the offender-not necessarily the perception or belief of the victim' (NSWPF n.d.: 4). Respondents were clear that this approach to recording is deliberately strict, requiring firm evidence of a biased motive on the part of the offender. Respondents drew a distinction between the US and the UK models of recording bias crime (described in the literature review above). The NSWPF follows the US model because it requires 'clear evidence' of the 'offender's intent' (R8). This 'investigative approach' was said to be 'more accurate' (R3) than the UK 'perception test', which gives authority to victims to determine whether an incident is recorded as a bias crime. Respondents described two problems with the UK model. First, they characterised it as overly sympathetic to 'angry' communities whose perceptions 'can be way off' (R3): Are we just keeping the community happy to keep them happy or are we keeping them happy because we're doing our job effectively? … I don't want to use the word 'kowtow' but we can be too sympathetic. (R8) Second, this overly sympathetic approach was said to lead to the 'over-representation' of recorded bias crime that 'blows' the UK figures 'through the roof' (R8, R1). In contrast, authority to name and record an incident as bias crime in NSW remains the sole preserve of police. This is so, even if there are conflicting community perceptions. Only one dissenting respondent (R4) noted that the UK model helps send the message that police take victim perspectives seriously and, thus, is better equipped to alleviate fear of crime-a service delivery priority in the NSWPF (2016) Corporate Plan 2016-2018.

Identifying bias crime
Most respondents cited the inability of frontline officers to identify bias crime as a major barrier to both reporting and recording. For example, one respondent commented that unless there is blatant evidence of a biased motive on the part of a suspect, the 'average cop' simply does not identify bias crime: The biggest challenge is police actually identifying that it could be … and sow that seed of thought in their head as to why [the offence was committed]. And it's a hard sell … from a general duties point of view. (R8) These respondents attributed the difficulty of 'selling' bias crime to the fact that it is not perceived as core business by frontline officers, who do not perceive the need to ask why an offence was committed when such evidence is not essential for any criminal charge. Some respondents stated that although the accuracy of recording among frontline officers has improved in recent years, there are still 'many, many wrong flags' (R9) entered into the system (e.g., an event may be recorded as a bias crime solely because the victim is from a racial minority). Research suggests that this confusion stems from a limited understanding of the concept of bias crime itself (Dalton and de Lint 2018; ). This was apparent in the responses of some respondents who struggled to identify incidents involving mixed motives or opportunism as 'real' bias crime. For example, one respondent discussed the case of an offender who spray-painted racist abuse on his neighbour's car following a dispute. In the respondent's view this was not bias crime because it was 'just the language' (R1) that was racially motivated, not the crime itself. Yet, the Bias Crime Standard Operating Procedures make it clear that bias need not be the only motive. Many respondents attributed this lack of internal understanding to the absence of systematic training (a point further explored in 'Framing Bias Crime').

Framing bias crime
Crime motivated by bias can have an aggravated effect on vulnerable communities (Iganski 2008). Accordingly, police are encouraged to take a 'big picture' approach to such offences by recording minor incidents, prioritising victim liaison and generating local intelligence about its nature and effect (Macpherson 1999).
The BCU has adopted a 'greater picture' (R9) approach to bias crime. However, this approach has two distinctive features that set it apart from best practice. The first is that the BCU frames the significance of bias crime in terms of the perceived risk it poses to public order: NSW is the only Australian state to have had three major public order incidents based on race and religion. (R7) Bias incidents precede bias crimes and the incident is the canary … it does end up in the big rallies in the main thoroughfare of Sydney … Many of the things that sit within the bias crime portfolio are drivers of that unrest … [T]hose ripples will emerge in your community and the worst-case scenario is you end up with significant public order issues. (R1) The second dimension of this 'greater picture' is the characterisation of organised hate groups, particularly right-wing extremists, as a key driving force behind bias crime: 'They may not necessarily be the ones that commit them but it's the frenzy they whip in people's minds' (R8).
Respondents explained that much of the work of the BCU has been to build trust with right-wing groups: They trusted us enough that we would allow them to do what their democratic rights allowed them to do, so long as they didn't overstep the line. They trusted that and they were thankful. (R8) Here, extremist groups function as a symbol of bias that has a 'ripple' effect on the minds and emotions of other people. The rationale for building trust with these groups is to gather intelligence and pre-empt public disorder. In effect, the value attached to bias crime is its potential to predict mass violence and disorder.
It is possible that framing bias crime as a problem of public order and extremism has the potential to dislodge its institutional status as a 'pink and fluffy' (Foster, Newburn and Souhami 2005: 91) aspect of police work. Yet, this search for internal legitimacy comes at a price. A high proportion of respondents questioned the prioritisation of organised extremist groups over everyday incidents of bias crime. More specifically, they questioned the relevance of organised hate groups to bias crime against LGBTI, disabled and homeless communities. As one respondent pointedly commented, 'at times it was like we were watching American History X' (R10). 3 IJCJ&SD 9 Online version via www.crimejusticejournal.com © 2019 8(2) In this context, the recent relocation of the BCU to the Fixated Persons Investigation Unit is significant. Some respondents suggested that the move to an operational command that manages individuals at risk of 'high level' extremism provides enhanced capacity for threat assessment as well as an opportunity for bias crime to achieve greater internal 'street cred' (R8). However, most expressed considerable reservations about this move. These concerns coalesced around the danger that the unit's narrow focus on individual pathology and counterterrorism will funnel the work of the BCU towards a handful of potential offenders with extremist motives at the expense of the everyday, opportunistic forms of intimidation experienced by individual bias crime victims. Respondents commented that effective bias crime policing requires both approaches. For example, one respondent stated: it's part of a police officer's job to deal with the verbal harassment, the bullying … I can't see how the Fixated Persons Unit is going to deal with any of that … Victim support is a really important part of police work and we need a bias crime initiative to do that. (R4)

Community engagement
The Bias Crime Standard Operating Procedures acknowledge that 'sustainable community engagement is the key to addressing, responding to and preventing' (NSWPF n.d.: 4) bias crime. Yet, community engagement has not been core business for the BCU. In the 10 years since the NSWPF formalised its bias crime initiative, it has not established a community outreach program dedicated to bias crime. Instead, senior managers explained that the role of the BCU has been to 'create the framework' within which Local Area Commands liaise with minority communities; for example, the BCU provides monthly advice to Commands on local patterns and trends (R2). This approach is consistent with the NSWPF's decentralised service delivery model, but it does little to build a conduit for sustainable community collaboration.
Notably, this is apparent in the top-down manner that the NSWPF has approached the process of defining bias crime. In many jurisdictions, the police definition of bias crime is constrained by an existing statutory definition. The NSWPF are under no such constraint, as there is no bias crime offence in NSW. Nonetheless, respondents explained that the organisation has deliberately maintained full internal control over the definition of bias crime. This was explained by one respondent, who said that if the police had to consider community views, 'you would get so many special interest groups wanting to have their own input into the definition it probably wouldn't be workable' (R2).
This paternalistic approach is at odds with recent research, which suggests that the process police use to define bias crime has an effect on the likelihood that members of minority communities will understand bias crime and report victimisation (Wickes et al. 2015). Engaging stakeholder groups at the developmental stage gives them confidence that bias crime programs are legitimate and not just cosmetic.
Strikingly, the community with whom the BCU has developed the greatest rapport, and arguably the most trust, is comprised of organised hate groups. Some respondents commented that this orientation has isolated the work of the BCU from diversity portfolios that focus on vulnerable and marginalised communities. These same respondents saw proactive outreach as the 'number one' priority if under-reporting is to be ameliorated in victim communities that have histories of poor relations with police, such as the Aboriginal, Muslim and LGBTI communities. For example, one respondent stated: we've got to see how they see the world because if we don't know how they see the world or how they see the experiences … then how can we manage the response to that? … [A]nd nowhere is that more important than in the bias space. (R6)

Leadership
Individual champions are essential to drive reform through organisational structure and police culture (Beech, Gulledge and Stewart 2015). Overwhelmingly, respondents attributed progress to date in bias crime policing to a handful of individuals at different levels of the organisation. As one respondent put it, 'it's the only way we've survived' (R7). Respondents in non-managerial positions contrasted the tenacity of this individual commitment to the paucity of institutional commitment. The following two comments demonstrate this perspective: The major challenge has been to get people to take it seriously … and getting endorsement from senior management. (R10) You need … multiple people throughout the hierarchy to realise that this is something important and to champion it. And not because they're told to or just because it's their role … Everyone has to take ownership … Realistically this has to come from the top … If the Commissioner views this as important, it'll get pushed up. If he doesn't … then it gets pushed down. I feel like we just haven't had the right push for this unit to go in the right direction. (R9) Three key examples illustrate these claims. First, in June 2009 the NSWPF moved the officer who held the Bias Crimes Coordinator role to other duties. The position was left vacant for three years until September 2012, when this same officer was reappointed. This striking neglect means that no bias crime data were collated and no bias crime reports were reviewed during this period. Several respondents observed that the position was only filled again when one high-ranking 'champion' returned to the force after a period of secondment. Second, the NSWPF (2016) has never produced a distinct bias crime policy, much less included bias crime in its service delivery priorities. 4 Research highlights the importance of bias crime policy for formalising purpose, standardising the exercise of discretion and tempering resistant policing culture (Grattett and Jenness 2005; Perry 2016). To some extent, the Corporate Sponsorship program within the NSWPF has provided a degree of support by ensuring that bias crime has had a dedicated sponsor at the Superintendent level. However, most respondents pointed out that this mid-level support is necessary but not sufficient, as sustainable reform requires 'buy in' (R3) at the most senior level. Third, there is no systematic or integrated training on bias crime for recruits or serving officers. With limited resources, the BCU has been able to offer only sporadic training at regional and command levels (e.g., through station training days). 5 While specialised training may not immediately improve recruit understanding of bias crime, research suggests that the integration of ongoing, evidence-based training into routine practice is essential to enhance cultural competency across the organisation (Miles Johnson et al. 2018).
Eventually, there is a need to move from individual to corporate drivers, and this requires senior managers to set a lasting change program (Beech, Gulledge and Stewart 2015). Respondents frequently remarked that this institutional shift has not been realised in the NSWPF. This was vividly captured by one respondent who stated that the NSWPF views bias crime as 'the graveyard shift: what you do when you've done everything else' (R6).
This disinterest was partly attributed to generic variables, such as competition for resources in a large institution. Yet, a lack of political will among senior executives was identified as a major barrier to the full migration of bias crime to operational policing. The depth of negative feeling among some respondents on this point is highlighted in the following two examples: Bias crime is the unwanted secret. We don't want to admit publicly here in NSW that we have tensions between various racial groups. (R8) The NSW police doesn't care. They've got no interest in this. It's a political inconvenience they have to worry about now and then … It's standard NSW Police.
Nothing changes until somebody dies … Until the day a white supremacist walks into a mosque … The biggest issue with NSW police is we're run by old white men … I think the direction in this organisation has always been to shut the minorities up. (R7) Most respondents believed that until this leadership vacuum is addressed, bias crime would continue to stay on the 'backburner' (R9) where it is accorded little institutional worth.

Discussion
The results of the study show that there have been many challenges in 'migrating' (Perry 2016) bias crime to policing in NSW. This does not mean there has been no progress. Achievements have been made in bias crime recording, training, intelligence and investigation. Yet, respondents expressed divergent views about the extent of these achievements. This study focused on the major consistencies in their responses as a means of bringing validity to the findings. It concludes that these achievements have not brought about sustained and significant reform to the way that bias crime is policed in NSW. Although the study sample is small and confined to one research site, the findings build on existing research by presenting rare 'insider' perspectives on obstacles to bias crime policing. The findings have implications for policing services in other jurisdictions.
The study lends support to existing research, which shows that bias crime is under-reported (especially when compared with parallel crime), and that this under-reporting can be partly ascribed to a lack of trust that minorities place in police (Murphy and Cherney 2011;Wiedlitzka et al. 2018). Some respondents attributed this mistrust to minority experiences of prejudice at the hands of the NSWPF. This observation is consistent with policing research that reveals a persistent culture of negative attitudes and stereotypes towards marginalised groups, despite efforts to improve cultural competence through diversity programs (Loftus 2010;Miles Johnson 2016b). Among the respondents, there was an acute awareness of the effect that this trust deficit has on bias crime reporting in NSW.
The UK is widely regarded as a global leader in recording bias crime against minority communities with a history of mistrust in police (FRA 2016;Giannasi 2015). The UK adopts a 'perception' test that promotes openness and accountability by enabling victims to have a voice in the way their complaints are categorised and recorded (Macpherson 1999). The NSWPF has chosen to reject this community-oriented approach on the grounds that it is overly sympathetic to victim interests and inflates official reports. This finds some support in recent research that calls for the over-classification of bias crime to be taken as seriously as under-classification (Dalton and de Lint 2018). However, what is missing from the NSWPF approach is an appreciation of the reasons for the UK model. Naming an incident as a bias crime sets up a requirement to investigate it as such. This circumvents possible police failure to proceed on that basis and encourages reporting by building confidence that victim perspectives will be taken seriously (Macpherson 1999). This, in turn, helps strengthen community knowledge of bias crime, among both marginalised and wider communities. The study suggests that a police force that maintains sole control over how bias crime is recorded is likely to experience low rates of reporting.
The NSWPF has not equipped officers with the capacity and motivation to identify bias crime and do so correctly. A recent critique of the NSWPF's investigation of anti-gay homicides exposes the need for greater precision in identification through the adoption of bias crime indicators based on evidence and evaluation (Dalton and de Lint 2018). While this is crucial, the findings suggest that identification skills also need to be tackled long before an individual complaint is received.
To identify bias crime, police need to understand the meaning of the concept and appreciate its value to policing . In NSW, relatively few officers have been exposed to an educational package that fosters this understanding. At a minimum, bias crime training should be both integrated into the wider curricula and delivered through standalone training modules (Asquith 2012). Although training alone cannot guarantee improved identification , this analysis suggests that fledgling bias crime initiatives will make little progress in the absence of systematic and sustained education.
Little has been written about how police frame bias crime (Grattet and Jenness 2005). Framing bias crime primarily as a 'hard' policing problem of civil unrest or violent extremism helps coordinate counter-terrorism and bias crime intelligence to enhance the capacity for early intervention in cases where there is a high risk of significant injury to the public. It may also help overcome institutional indifference and individual cynicism towards bias crime policing. This may appeal to police services, such as the NSWPF, that prioritise public order and state security over community policing values (Chan 2007;Sentas and Grewcock 2018). However, the findings suggest that this approach to building internal legitimacy is misguided if it overlooks forms of bias crime that have little connection to public disorder or organised hate groups, such as the victimisation of sexual minorities, people with disabilities and homeless people (Asquith 2012). It isolates bias crime from diversity policing, which sees it as an everyday problem of social inequality, discrimination and a 'violation of the fundamental rights of individual citizens' (FRA 2016: 28), and sends the message that bias crime is a natural fit with policing tactics geared towards authoritarian styles of 'homeland security' (Oliver 2006). In the context of NSW, these concerns have come to fruition with the relocation of the BCU to the Fixated Persons Investigation Unit. This re-engineering signals a repudiation of international best practice that orients bias crime policing not only around the risk posed by potential extremist offenders but equally around the everyday safety needs of targeted communities.
Democratic processes in police decision-making help build public trust and confidence (President's Task Force 2015; Trinkner, Tyler and Goff 2016). Accordingly, police corporate plans typically prioritise the need for 'sustained community engagement' (NSWPF 2016). At the same time, police work in an increasingly politicised and fiscally restrained environment (Sarre and Prenzler 2018). It has been suggested that this encourages agencies to turn inwards to concentrate on 'core business' rather than outwards to collaborate with community stakeholders (Bartkowiak-Théron and Asquith 2015). This is apparent in the NSWPF, which in the last 10 years has not built a sustained outreach program of consultation with targeted communities over how bias crime is defined, categorised, recorded or addressed. For example, third party reporting mechanisms initiated in NSW in the 1990s have been left to languish, no longer offering an accessible pathway to reporting for minority communities. This more authoritarian approach may reflect the lack of trust that police themselves place in minority communities (Miles-Johnson 2016b; Miles-Johnson and Pickering 2018). Murphy and Cherney (2011: 251) underscore the need to provide minority communities with an 'instrumental voice' that enables them to articulate local concerns, and shape policies and procedures that affect their members. The findings build on this research by providing empirical evidence to suggest that minority communities remain reluctant to report bias crime when police are the primary or, indeed, sole definers of the problem. A police-initiated system of ongoing partnership with well-respected stakeholder organisations is an essential foundation for building legitimacy and mediating between estranged communities and police. It enables vulnerable and diverse communities to have genuine input into how police define the problem, set priorities and develop tactics for enhancing safety. As one respondent put it, 'you can't turn to a white supremacist to provide this kind of leverage' (R10).
At the heart of the failure of the NSWPF to bring about significant change to bias crime policing is leadership. Research demonstrates that tailored change programs, including bias crime initiatives, must be set by senior executives who can inspire, motivate and communicate with rank and file officers (Beech, Gulledge and Stewart 2015;Grattet and Jenness 2005;Trinkner, Tyler and Goff 2016). This study's findings advance the research by showing that police tasked with the day-to-day delivery of bias crime reform place immense value on leadership endorsement. In the eyes of many respondents, the NSWPF executive sees bias crime as an 'unwanted secret', a 'political inconvenience' or 'the graveyard shift'. These respondents suggest that the organisation's commitment to bias crime is more about legitimising its public image with minority communities than setting a genuine agenda for change. These perspectives are corroborated by the lack of a structural framework to support the work of the BCU, such as a dedicated policy, sustained evidence-based training, a formal community-engagement process or a fiscal commitment to ongoing staffing. Frontline officers cannot make sense of new concepts and new procedures without leadership at the most senior level (Beech, Gulledge and Stewart 2015;Chan 2007). Not only do they need to be trained to identify, record and investigate bias crime, they also need to be motivated by their leaders to do so. At a minimum, this requires agency leaders to legitimise bias crime as 'real' police work , for example, through public statements that are emphatic about its value to routine policing. The findings suggest that this dearth of symbolic and practical support trickles down to a lack of capacity, preparedness and confidence among frontline officers (Bradford and Quinton 2014) to identify bias crime and comply with recording rules. The findings also suggest that a leadership vacuum about the purpose of bias crime policing can be too readily filled by individual staff preferences rather than evidence-based research on appropriate priorities (e.g., in the NSWPF this vacuum has been filled by an offender-centred approach that builds trust with organised hate groups, rather than a victim-oriented approach that builds trust with targeted communities). Like all forms of democratic policing, long-lasting change in bias crime policing depends on shifting internal norms (Trinkner, Tyler and Goff 2016) from a 'warrior' to a 'guardian' mindset (President's Taskforce 2015: 1). This must come from the top.

Conclusion
There is a lack of empirical research on bias crime policing from police perspectives. The perspectives presented in this study suggest that well-intentioned bias crime initiatives will flounder unless backed by an institutional framework that builds both internal and external confidence in the value of such policing. This is best achieved through democratic styles of policing that help repair relations with minority communities. Bias crime policing is but one innovation that will continue to struggle in institutions that remain committed to authoritarian approaches to law enforcement.
IJCJ&SD 14 Online version via www.crimejusticejournal.com © 2019 8(2) 1 Bias crime is also referred to as hate crime. 2 We have excluded the period from June 2009 to August 2012 when the Bias Crimes Coordinator role was vacant and no data were collated. 3 American History X is a 1998 film about the Neo-Nazi movement in the US. 4 Although bias crime is given appropriate attention in the NSWPF Strategy on Sexuality, Gender Diversity and Intersex 2016-2020, it is subsumed, briefly, into 'violent extremism' in the Multicultural Policies and Services Plan 2017-2020. 5 Bias crime has been well integrated into some other training packages, such as training for Gay and Lesbian Liaison Officers.