Black and Minority Ethnic Boys and Custody in England and Wales: Understanding Subjective Experiences through an Analysis of Official Data

Recent years have seen a dramatic shift in youth justice outcomes and a fall in the number of children drawn into the youth justice system in England and Wales. However, it appears that children from some backgrounds have not benefited as much as others from this change. There is a wealth of academic literature on processes of criminalisation, policies, and practices of youth justice and the experiences of children, particularly boys, in custody. However, there is little detailed understanding of how these processes, policies, and practices affect children from different backgrounds. This paper examines the most intrusive aspect of youth justice, namely, custodial sentences. Through an examination of the Inspectorate of Prisons’ reports and associated surveys, this paper seeks to explore black and minority ethnic boys’ perceptions of their experiences of custody.


Background
For some decades now, the disproportionate rate at which black and minority ethnic (BAME) 1 boys are given a custodial sentence has been highlighted in government statistics in England and Wales. Surprisingly, however, there is a general paucity of scholarly work in the specific area of BAME boys and custody. Almost none explores the different experiences both prior to and in custody of the distinct ethnic, racial, and religious minorities that make up the broad categories the literature describes. This is an oddity given that the discipline of criminology is built upon the study of youth crime. In both the United States and the United Kingdom, there is a considerable amount of literature on youth crime, including studies that have explored risk factors such as socioeconomic background, poor educational attainment, family breakdown, and risk-taking behaviours (Farrington 1998;Loeber and Farrington 2012). Research related to black youth and racialisation provides us with some understanding of the processes at play in relation to vulnerability, marginality, and otherisation. 1 This paper explores the experience of children in custody that the Inspectorate of Prisons (hereinafter "Inspectorate") describes in its reports as being from a "black or minority ethnic background". In its surveys of children in custody, the Inspectorate asks children to define themselves as being from one of 17 categories of "ethnic origin" or to self-define. In its reports, it amalgamates all responses from children who do not describe themselves as "white British", "white Irish", or "white other" as being from a "black or minority ethnic group" (although the raw data is available for individual ethnic groups in surveys of individual establishments). We recognize the enormous variation that resides within this generic category of "black and minority ethnic" and indeed one of the conclusions of our paper is that this broad categorisation hides very varied experiences amongst the many different individuals and subgroups of whom it is comprised. Nevertheless, with this caveat, we believe examining the different experiences of even the broad categories used by the Inspectorate is of value.
The paper uses the definition of a child stipulated in the United Nations Convention on the Rights of a Child as "every human being under the age of 18" (United Nations 1989, Article 1), which recognises the obligation to ensure that "[e]very child deprived of liberty shall be treated with humanity and respect for the inherent dignity of the human person, and in a manner which takes into account the needs of persons of his or her age" (United Nations 1989, Article 37 (c)). It should be noted that much of the literature we examined describes the experience of young adults or older "youth" in custody and that should be borne in mind when considering its relevance to the age group with which we are concerned. It also serves to highlight the significant gap in the existing literature on the experiences of BAME children in custody.
Some of the earliest studies focused on the relationship between the police and black youth, and pointed to the inherent tensions in a context of discrimination, disadvantage, and institutional racism (John 1970;Lambert and Jenkinson 1970). Subsequent research identified the ways in which the media and the educational and welfare systems problematized black young people (Coard 1971;Hall et al. 1978; Barn 1990Barn , 1993Barn , 2001Solomos 1993;Sewell 1997). Here, key issues and concerns involved the processes by which black youth become labelled as "mad, bad, and dangerous" and in need of discipline and governmentality (Barn 2001). For example, whilst Coard (1971) documented how the black child is made educationally subnormal, Hall et al. (1978) identified the ways in which the crime of mugging became synonymous with black youth in the media and public discourse. Other scholars pointed to specific negative experiences of black young people at the hands of the police, including being categorised as a problem group by the police, and therefore more likely to be questioned or arrested, and allegations that the police used excessive physical force in their dealings with black suspects (Solomos 1993).
In relation to the overrepresentation of black youth in custody, Walker (1988) provided a first reference point. In her analysis of the outcomes of prosecutions of white, black, and Asian young men aged 17-25 in 198317-25 in , Walker (1988 showed that black youth were more likely to receive custodial sentences than their white counterparts. By citing the earlier 1986 Home Office (1986) study into black prisoners having fewer previous convictions than whites, Walker argued that higher rates of custodial outcomes combined with longer sentences may well account for the higher representation of black people in prison. This was further supported by Hood's (1992) seminal study on racial disparities in sentencing in the Crown Court, which confirmed that black males were more likely to be sentenced to custody and longer custodial sentences than white and Asian males.
In the first comprehensive analysis of youth justice statistics, Feilzer and Hood (2004) identified evidence of unjustifiable differential treatment of BAME (particularly those of African-Caribbean and mixed-parentage background) young people across various stages of the youth justice system, a finding which was confirmed by May et al. (2010). In 2017, David Lammy, Labour Member of Parliament for Tottenham (London), carried out an independent review, "commissioned by two Prime Ministers" (Lammy 2017, p. 3)-originally in January 2016 by David Cameron and subsequently supported by Theresa May. David Cameron, the then Prime Minister in 2016, stated: If you're black, you're more likely to be in a prison cell than studying at a top university. And if you're black, it seems you're more likely to be sentenced to custody for a crime than if you're white. We should investigate why this is and how we can end this possible discrimination.
In a cross-party alliance, Lammy had been commissioned to make recommendations for improvements in the disproportionate representation of BAME people in the criminal justice system (Lammy 2017). Lammy set out his grave concern about the overrepresentation of BAME young people in the youth justice system, and in particular, in youth custody (Lammy 2017). The sluggishness of the youth justice sector and lack of systematic good practice led Lammy to conclude that "unless something changes, this cohort (that is BAME young people) will become the next generation of adult offenders" (Lammy 2017, p. 5). In particular, he recommended the need for local and community involvement in a quest for accountability and transparency, with a strong emphasis on parenting. Most recently, evidence suggesting that almost 80% of those registered on the Metropolitan Police Service's Gangs Matrix are black reinforces the perception of "racial profiling that criminalises and demonises (primarily) black youths for being black and poor" (Scott 2018, p. 23). A key criticism of the use of a matrix is that is based on a loosely defined use of the concept of "gang", that many of the young people included on the matrix have not been involved in violence, and that the police are monitoring young people's behaviour online based on belonging to certain subcultures (Amnesty International 2018).
Our review of the literature indicates that whilst quantitative evidence into ethnicity, disproportionality, and incarceration has been documented through official statistics and independent reports such as the Lammy Review, there is a dearth of theoretical and qualitative analysis into the subjective and lived experiences of BAME young people-and BAME children in particular-in custody (Phillips and Bowling 2007). Although official and academic research has highlighted high rates of custodial sentences, BAME young people's and children's lived experiences of custody have been largely neglected. In a rare ethnographic study involving 15 black young men, Wilson (2003) provided some insight on how black social lives are constructed inside a young offender institution (YOI). In particular, he pointed to the salience of ethnicity and religion in the incarcerated experiences of black youth. Strategies such as "keeping quiet", "sticking together", or "going nuts" were framed as central in dealing with situations of crisis and conflict. The following quote captures the essence of the accounts narrated in the Wilson study about control, regulation, and otherisation in the context of racial/ethnic hierarchy. In this study, a narrative of ethnic solidarity is in evidence in the face of racism and social exclusion within a young offender institution: What if it was all black Govs, and all black lads on the servery? What if it was all black lads that worked outside as orderlies and it was all the white lads that were banged up? What if all the black lads were on the highest regimes and all the white lads on the lowest, and who kept getting stitched up and twisted up? Then they'd know how it felt, because that's how it feels to me. (Wilson 2003, p. 422) Some previous work into prison staff and prisoners' perspectives, from the 1980s in Britain, had also identified similar issues regarding racial/ethnic hierarchy and ethnic solidarity (Genders et al. 1989). In one of the most recent ethnographic accounts of prisoner perspectives, Phillips (2008) in her two-site study of an adult prison and a YOI, showed social class and neighbourhood identities to be more salient and argued that multiculturalism discourses were evidently and actively negotiated in daily interethnic encounters through a shifting process of otherisation in the context of ethnicity, faith, class, and nationality.
Similarly, Qasim (2018) recounted the experiences of young adult Pakistani men in the prison system and emphasized the importance of ethnic solidarity and religion "on the inside". Religious devotion increased whilst the young men were in prison and fulfilled a number of roles-it provided a purpose and comfort but also offered a routine to pass the time as well as a sense of dignity (Qasim 2018). It is often claimed that prisons, including YOIs, are a site of conversion to Islam but data on the scale of conversions is difficult to come by (Qasim 2018;Spalek and El-Hassan 2007; see Veldhuis 2016, 140 for a critical review of these claims).
Research evidence shows that in addition to the disproportionality of BAME young people in custody, there are important concerns about their custodial experiences. Although scholarly contribution into incarcerated experiences remains rather limited, it does indicate some important concerns around young people's need to develop coping mechanisms to deal with racism and social exclusion in these spaces (Wilson 2003;Phillips 2008).

Youth Justice Policy Context
Youth justice policy and practice has undergone a dramatic shift over the last decade, which is most obvious in the fall of children entering the youth justice system. The number of children cautioned or sentenced in 2016/17 represented an 81% reduction over the last 10 years (Youth Justice Board 2018). Explanations for this remarkable trend have been varied but identified as a mixture of changes in policing practices, increased prevention and diversion work of youth offending services, as well as a possible reduction in youth offending behaviour (Sutherland et al. 2017).
A fall in juvenile first offences has been mirrored by a fall in the number or children in custody.  Table 2.7). Notably, boys and girls may be held in secure children's homes (SCHs) or secure training centres (STCs), but only boys may be held in YOIs.
Whilst the substantial fall in the number of custodial sentences over the last decade is welcome, the custody rate has remained stable at 6% and the length of custodial sentences has increased by almost 40% for indictable offences, from 11.5 months in 2007 to 16 months in 2017 (Youth Justice Board 2018). Additionally, children from different minority ethnic groups have not benefited to the same extent from the reductions in custody. Figure 1 shows that the greatest reduction of 76% between April 2005/06 and April 2018/19 (from 1945 to 470) in custodial sentences is evident for white children, with a 31% reduction between April 2005/06 and April 2018/19 for BAME children (from 656 to 452). This has led to an increase in the proportion of BAME children in custody from 27% in 2007 to 45% in 2017 (Youth Justice Board 2018). These figures refer to BAME boys. At present, there is no information available on BAME girls and custody. And this remains an important area of concern for future research.
These changes in the proportions of BAME children in youth custody have not been spread evenly through the children's custodial estate. The proportions of boys reporting they are from a BAME background in the Inspectorate of Prisons' reports of survey findings in YOIs vary from 14% to 60%. In three of the six YOIs surveyed, 50% or more of the boys report they are from a BAME background (YJB 2017, Appendix B1). As Lammy points out, the overall racial disparity is concerning in itself and we do not know how the shifting proportions of children from different backgrounds affect their perceptions and experiences of custody.  Table 2.7). Notably, boys and girls may be held in secure children's homes (SCHs) or secure training centres (STCs), but only boys may be held in YOIs.
Whilst the substantial fall in the number of custodial sentences over the last decade is welcome, the custody rate has remained stable at 6% and the length of custodial sentences has increased by almost 40% for indictable offences, from 11.5 months in 2007 to 16 months in 2017 (YJB 2018). Additionally, children from different minority ethnic groups have not benefited to the same extent from the reductions in custody. Figure 1 shows that the greatest reduction of 76% between April 2005/06 and April 2018/19 (from 1945 to 470) in custodial sentences is evident for white children, with a 31% reduction between April 2005/06 and April 2018/19 for BAME children (from 656 to 452). This has led to an increase in the proportion of BAME children in custody from 27% in 2007 to 45% in 2017 (YJB 2018). These figures refer to BAME boys. At present, there is no information available on BAME girls and custody. And this remains an important area of concern for future research.
These changes in the proportions of BAME children in youth custody have not been spread evenly through the children's custodial estate. The proportions of boys reporting they are from a BAME background in the Inspectorate of Prisons' reports of survey findings in YOIs varies from 14% to 60%. In three of the six YOIs surveyed, 50% or more of the boys report they are from a BAME background (YJB 2017, Appendix B1). As Lammy points out, the overall racial disparity is concerning in itself and we do not know how the shifting proportions of children from different backgrounds affect their perceptions and experiences of custody. This broad trend hides some other changes in the composition of the secure and YOI estate. We have made reference here to changes experienced in one particular YOI, HMYOI Feltham, to illustrate these demographic changes. Figure 2 below uses Inspectorate survey data to illustrate how the population of one young offender institution, HMYOI Feltham A (the part of Feltham that holds boys aged 15-18), has changed over the years. The raw data that is available for individual institutions enables, at least, some insight into how the different groups that make up the BAME category in this institution have changed. The proportion of boys describing themselves as white fell sharply between 2002 and 2009 from 45% to 24%. Boys who described themselves having Caribbean heritage increased from 19% to 33% between 2002 and 2009 before falling back to 25% in 2018. The proportion of boys who described themselves as "Black or Black British-African" rose steadily from 9% in 2002 to 13% This broad trend hides some other changes in the composition of the secure and YOI estate. We have made reference here to changes experienced in one particular YOI, HMYOI Feltham, to illustrate these demographic changes. Figure 2 below uses Inspectorate survey data to illustrate how the population of one young offender institution, HMYOI Feltham A (the part of Feltham that holds boys aged 15-18), has changed over the years. The raw data that is available for individual institutions enables, at least, some insight into how the different groups that make up the BAME category in this institution have changed. The proportion of boys describing themselves as white fell sharply between 2002 and 2009 from 45% to 24%. Boys who described themselves having Caribbean heritage increased from 19% to 33% between 2002 and 2009 before falling back to 25% in 2018. The proportion of boys who described themselves as "Black or Black British-African" rose steadily from 9% in 2002 to 13% in 2009 to 17% in 2017 and boys who were from other black, Asian, or mixed-race backgrounds also rose from 17% to 30% between 2002 and 2017. Turning to faith, the proportion of boys who described themselves as having no religion fell from 37% to 11% between 2002 and 2017. Until 2009, this was replaced by an increase in the proportion of boys who described their religion as Church of England (up from 19% to 31%) and Catholic (up from 18% to 27%), but in the period to 2017, the proportion of boys in both these categories fell to 25% and 19%, respectively. The proportion of boys who described themselves as Muslim remained reasonably static between 2002 and 2009 (18%-22%) before rising sharply to 37% in 2017. Soc. Sci. 2018, 7, x FOR PEER REVIEW 5 of 17 in 2009 to 17% in 2017 and boys who were from other black, Asian, or mixed-race backgrounds also rose from 17% to 30% between 2002 and 2017. Turning to faith, the proportion of boys who described themselves as having no religion fell from 37% to 11% between 2002 and 2017. Until 2009, this was replaced by an increase in the proportion of boys who described their religion as Church of England (up from 19% to 31%) and Catholic (up from 18% to 27%), but in the period to 2017, the proportion of boys in both these categories fell to 25% and 19%, respectively. The proportion of boys who described themselves as Muslim remained reasonably static between 2002 and 2009 (18%-22%) before rising sharply to 37% in 2017.

Understanding BAME Experiences of Youth Custody
The focus in this paper is on exploring how the experiences of custody differ for BAME boys in YOIs using data from the regular surveys 2 of children in YOIs and STCs undertaken by the Inspectorate, either as part of an inspection or as a stand-alone exercise. The Inspectorate currently undertakes surveys of both STCs and YOIs. YOIs were chosen as the basis for this paper because they hold the greatest number of children and survey data has been collected from them for longer than STCs. Survey results may be published alongside individual inspection reports and in an annual report of Children in Custody, which includes all the surveys undertaken that year.
In his 2016/17 Annual Report of the Inspectorate's work and findings as a whole, the Chief Inspector of Prisons reported that: 2 For an explanation of the Inspectorate's sampling and analysis methodology, see, for example, "Report of an Unannounced Inspection of HMYOI Feltham (children and young people)" by HM Chief Inspector of Prisons 20-21 December 2017, 8-12 January 2018, pages 7 and 78. Almost all children in custody are boys, and we distinguish between when we are speaking of children of all sexes and boys only. We have used the term "significant" or "significantly" to describe differences in results that the Inspectorate determined, taking into account the sample size and response rate, were unlikely to have been caused by chance. We

Understanding BAME Experiences of Youth Custody
The focus in this paper is on exploring how the experiences of custody differ for BAME boys in YOIs using data from the regular surveys 2 of children in YOIs and STCs undertaken by the Inspectorate, either as part of an inspection or as a stand-alone exercise. The Inspectorate currently undertakes surveys of both STCs and YOIs. YOIs were chosen as the basis for this paper because they hold 2 For an explanation of the Inspectorate's sampling and analysis methodology, see, for example, "Report of an Unannounced Inspection of HMYOI Feltham (children and young people)" by HM Chief Inspector of Prisons 20-21 December 2017, 8-12 January 2018, pages 7 and 78. Almost all children in custody are boys, and we distinguish between when we are speaking of children of all sexes and boys only. We have used the term "significant" or "significantly" to describe differences in results that the Inspectorate determined, taking into account the sample size and response rate, were unlikely to have been caused by chance. We have pointed to other differences, but where we have not had access to the data, we have not been able to test for significance in the same way. the greatest number of children and survey data has been collected from them for longer than STCs. Survey results may be published alongside individual inspection reports and in an annual report of Children in Custody, which includes all the surveys undertaken that year.
In his 2016/17 Annual Report of the Inspectorate's work and findings as a whole, the Chief Inspector of Prisons reported that: "By February this year we had reached the conclusion that there was not a single establishment that we inspected in England and Wales in which it was safe to hold children and young people . . . we see establishments in which there seems to be something of a vicious circle. Violence leads to a restrictive regime and security measures which in turn frustrate those being held there. We have seen regimes where boys take every meal alone in their cell, where they are locked up for excessive amounts of time, where they do not get enough exercise, education or training, and where there do not appear to be any credible plans to break the cycle of violence." (HM Chief Inspector of Prisons 2017, p. 9) We know too little about how different groups of children in custody have reacted to this deterioration in safety and other standards or indeed whether small shifts in the characteristics of children in these groups-of age or offence, for instance-are reflected in how they cope with the custodial experience.
In England and Wales, the annual reports into Children in Custody help shed light on young people's perceptions and experiences of youth custody. The most recently published main report (2016/17) sets out the perception of boys and girls aged 12-18 in STCs and boys in YOIs. The Appendices to the main report set out in detail the survey responses which are split into a number of different groups. To date, the data presented in the Appendices have not been subjected to an external systematic analysis to understand the experiences of BAME boys in custody. This is an oversight given that BAME boys make up an increasing proportion of the youth custodial population whose needs may not be met sufficiently in the current system.
We scrutinised the Appendices and we highlight below some important aspects of how BAME boys experience youth custody. The Inspectorate's published reports make it possible to compare results between BAME and white boys, Muslim and non-Muslim boys, and examine the experience of other relevant categories, such as care leavers, boys with disabilities, Gypsy, Romany or Traveller boys, those with emotional and mental health problems, and carers. Published survey data do not make it possible to examine the important overlap between some of these categories and we suggest this is an area that would be worthy of further study if greater access to the data were possible. In addition, to help understand the heterogeneity of experience among diverse BAME groups, it would be useful if the Inspectorate reports disaggregated the recording and presentation of this data, if and where possible, without compromising confidentiality and anonymity.
We explored data from the Children in Custody Annual Report for 2016/17 and focused on boys aged 15-18 held in five YOIs and one specialist unit. The data presented in the report's appendices are derived from surveys carried out at all YOIs and may be used to inform YOI inspections. All children in youth custody were invited to contribute and the completion rate across STCs and YOIs was 85% (YJB 2017, p. 10).
The following account relates to survey responses from 600 boys covering all aspects of their experience, including receiving their sentence, their transfer to custody, their first days and daily life in the YOI, relationship with staff, applications and complaints, rewards and disciplines, safety, health services and activities, family and friends, and preparation for release (YJB 2017). Surveys, of course, only reveal what respondents are willing to report about their experiences and feelings. It is likely that there are differences in the willingness of different demographic groups to disclose what they might feel would be regarded as "weakness" or perhaps withhold information for fear of perceived possible negative repercussions (Jolliffe and Haque 2017).
Review of the data presented in Children in Custody 2016/17 revealed both positive and negative differences between the experiences of BAME and white, and Muslim and non-Muslim boys. Some key demographics highlighting the complexity of the population of young people held in YOIs included: It should be noted that the survey results differed significantly on some measures-such as perceptions of gang problems, feelings of safety, and relationship with staff-between the different YOIs, which may have affected boys from diverse ethnic groups differently. For example, as Table 1 shows, those YOIs with the highest proportion of boys from BAME backgrounds (Feltham 63%, Cookham Wood 60%, and Werrington 50%) also had the lowest proportion reporting they felt unsafe at the time of the survey: Cookham Wood (10%), Feltham (11%), and Werrington (13%). On the other hand, Parc (32%), the Keppel Unit (29%), and Wetherby (18%) all reported higher levels of current concerns. Feltham (21%) also held the highest proportion of boys reporting problems with gangs when they first arrived and Parc and Wetherby (11%) the lowest. A less clear pattern was evident in relation to the question, "Do most staff treat with you respect?" Responses from the Kepple Unit (83%) and Wetherby (74%) were most positive followed by Feltham (66%) and Cookham Wood (62%). Less positive were Parc (41%) and Werrington (57%) (YJB 2017, Appendix B1). We can only speculate about the reasons for these differences. Perhaps there was something about the dynamic between different groups, the consequences of gang culture, and their relationships with staff that explained these variations. It might be because of differences in the YOIs themselves rather than in the boys they held. Some YOIs had different functions: the Keppel Unit was a national specialist unit attached to Wetherby YOI, for example (YJB 2017, p. 27). Or, as we note above, perhaps some groups were simply more willing than others to admit their fears in a survey. The arrival in custody and the early days are crucially important for children, in particular those entering custody for the first time. Significant differences in the perception of treatment and access to support are apparent at this early stage. BAME boys were significantly less likely than white boys to report being treated with respect when it comes to being searched on entering a YOI, or being offered help and support with smoking, loss of property, feeling scared, or feeling worried and upset/needing someone to talk to. They were also less likely to have access to a chaplain or Childline/Samaritans.
Additionally, day-to-day life in YOIs appears to differ for BAME and white boys. Significantly fewer BAME boys reported being able to access showers every day if they wanted to, found the food good, or found a sufficient variety of goods in the shop. More worryingly, BAME boys' access to a member of the Independent Monitoring Board appeared more restricted, fewer BAME than white boys felt that they were treated with respect by staff (57% BAME vs. 73% white) and felt their personal officer was available to help them. There was general evidence of a perception of being treated unfairly through the complaints, rewards, and adjudication system. Of considerable concern is the use of physical restraint, with just over half of BAME boys reporting experiences of physical restraint (53% BAME vs. 36% white) and relatively high levels of victimization by other young people because of ethnicity (7% BAME vs. 2% white), religion (5% BAME vs. 1% white), or nationality (4% BAME vs. 0% white), both at the hands of young people but also by staff (victimization because of ethnicity-10% BAME vs. 2% white; victimization because of religion 7% BAME vs. 0% white). A lower proportion of BAME than white boys would report victimisation to staff (20% BAME vs. 34% white) and be confident in being taken seriously (20% BAME vs. 35% white). Muslim boys reported similar concerns, although their views on whether they could report their concerns to staff and whether those concerns would be taken seriously are not significantly different from non-Muslim boys. BAME boys are also significantly less likely than white boys to report they had association every day (37% BAME vs. 62% white), which might indicate they were subject to more formal or informal disciplinary sanctions, and fewer said their caseworker had helped prepare them for release (43% BAME vs. 55% white).
Overall, boys from a minority ethnic group tended to report more negatively than others (YJB 2017) in terms of their experiences of being held in a YOI. However, on other areas of concern, the survey showed a rather different picture. Fewer BAME than white boys reported having any problems on arrival, although proportions of reported problems were very high for all boys (71% BAME vs. 83% white reported having any problems). They reported they were significantly less likely to have skipped school before entering custody. Fewer BAME boys reported feeling worried or upset and needing someone to talk to, reported problems with drugs and alcohol, or emotional or mental health problems (20% BAME vs. 34% white). Fewer BAME boys reported that they considered they had a disability (12% BAME vs. 25% white). Significantly fewer BAME boys reported a problem with shouting out of windows (34% BAME vs. 49% white)-a key indicator of bullying. These possible contradictions suggest that we need a more in-depth understanding of the cohort of BAME boys in custody and how they differ from white boys in a number of factors. This is particularly important as some of the factors highlighted above have an impact on how custody is experienced by boys in YOIs. We have highlighted some of the most important aspects here but recognise the potential of differential reporting for some of the indicators set out above.
As the population in custody has changed, so have the conditions in custody themselves. For example, as the number of boys in custody has fallen, so have the number of institutions holding them, and so those held will be likely to be further away from their homes. We can find no evidence of the specific impact of distance from home on BAME and Muslim boys, such as whether they are more or less likely to experience vulnerability and isolation than white boys if they are further away from a familiar community environment and community or whether greater distance provides an opportunity to break free from harmful associates.
Turning to the question of faith, a significantly higher proportion of BAME than white boys self-reported as adhering to the Muslim faith (44% vs. 4%). Muslim boys specifically reported more negatively on their relationship with YOI staff than non-Muslim boys. Their experiences of applications and complaints were more negative, they were more likely to have been restrained, more likely to have had adjudications, and more likely to have been victimized.
Evidence from the 2016/17 Children in Custody Report suggests that compared with non-Muslim boys, Muslim boys report that staff are less likely to check with them how they are getting on (38%/19%), they are less likely to feel applications are dealt with fairly (57%/40%) or complaints dealt with quickly (25%/9%), and they are more likely to have received an adjudication (61%/80%) or to have been physically restrained (41%/56%) (YJB 2017).The more negative perceptions of being treated fairly for BAME and Muslim boys highlight the importance of the intersectionality of certain demographic variables as mentioned above. We do not know whether religion, ethnicity, or other factors drive these experiences and perceptions. Crucially, if such differences are to be addressed, understanding the drivers is of key importance.
This section has highlighted the complexity of discussing BAME boys' experiences in custody. Experiences and perceptions clearly differ from those of white boys, but they are not exclusively more negative. There are other factors that influence differential experiences such as the YOI, disability, mental health problems, and substance misuse problems. And an in-depth analysis accounting for the different variables is key in developing a better understanding of how young people from different ethnic backgrounds experience custody.

A Detailed Case Study: HMYOI Feltham
In order to illustrate how boys experience custody differently over time, we explored changes in how boys from one institution, Feltham YOI, who were from BAME backgrounds or white, and who were Muslim or non-Muslim, experienced some key aspects of their imprisonment. We examined responses to questions in 10 surveys of HMYOI Feltham 3 through the data available in the Inspectorate reports conducted between 2005 and 2018, which we believe might indicate whether boys perceived they were treated fairly, gave insight into boys' relationships with and treatment by staff, and whether they felt safe.
HMYOI Feltham is a large young offender institution in West London. It is divided into two parts. The report of the most recent inspection in December 2017 described some improvements from the previous very critical report and a reduction in violence and the time boys spent locked in their cells from the previously very high levels (HM Chief Inspector of Prisons 2018). We were interested to see whether boys from BAME backgrounds experienced the variations in Feltham A's performance differently to white boys.
The questions we examined in detail were: • Is your cell bell normally answered within five minutes?  Cell bells should be answered within five minutes both as an important safety measure and also to ensure any urgent needs are met. Muslim children were only identified as a separate category in surveys from 2009 onwards, but we found that BAME boys responded significantly more negatively in 2009 and 2010 than white boys and Muslim boys did so in 2009, 2011, and 2013. Thereafter, their responses are more similar to those of white and non-Muslim boys, respectively, and in December 2013, BAME boys reported significantly more positively than white boys.
We used ease of making complaints as an indicator for institutional fairness. In most inspections, survey responses regarding ease of making a complaint are similar between BAME and white boys and between Muslim and non-Muslim boys. Fewer boys in all categories tend to say it is easy to make a complaint as time passes. We found a significant adverse difference in the response of Muslim boys in January 2013 and BAME boys in December 2013.
Responses to the question about whether staff treat them with respect differ from year to year and is a critical indicator of relationships between staff and children. In every survey, boys from BAME backgrounds were less likely to say that they had been treated with respect than white boys, and in four years-2005, 2007, 2013, and January 2017-the difference was statistically significant. In January 2013, for example, 53% of boys from BAME backgrounds said staff treated them with respect compared with 93% of white boys. Differences between Muslim and non-Muslim boys were not significant other than in January 2013, when Muslim boys reported more negatively than non-Muslim boys.
Physical restraint is the most extreme use of power by staff and that most open to abuse. Staff are more likely to use physical restraint if they are fearful and or do not believe other forms of de-escalation are likely to be effective (Gooch 2015). The proportion of boys from a BAME background who had been restrained increased in most surveys up to July 2015 and then declined. Boys from BAME backgrounds report being more likely to have been restrained than white boys in every survey up to January 2017, and in most of these, the difference was significant. This was not true for the two most recent years, when the proportion who said they had been restrained was falling. Indeed, in December 2017, more boys from a white than BAME background reported they had been restrained. That result is not statistically significant, but it does suggest a substantial change from the previous surveys. The inspection report that accompanied the December 2017 survey reported a marked improvement in safety in the establishment. Muslim boys also generally report they have been restrained more frequently than non-Muslim boys, but these differences are not statistically significant.
Perceptions of safety are both a response to physical threats and emotional insecurity. Boys from BAME backgrounds were significantly more likely than white boys to have reported ever feeling unsafe in 2005, 2011, and 2013, although in the last two years, responses were similar. Muslim boys also report feeling less safe than non-Muslim boys, although these results are not statistically significant, and as with BAME and white boys, the differences in the last two years have been very small.
It is perhaps noteworthy that in its report of its January 2017 inspection, just as the Inspectorate was castigating Feltham YOI for its lack of safety (HM Chief Inspector of Prisons 2017), its survey of the boys held there showed a decline in the number who had felt unsafe and, for the first time, there was almost no difference in the perceptions of BAME and white boys. Results for Muslim and non-Muslim boys were also very similar. As Table 2 illustrates, the concerns of boys from BAME backgrounds about their safety had reduced from a high in 2015 to relatively typical levels for the period in which surveys are available, whereas the concerns of white boys about safety had increased in 2015 following a period at relatively low levels. Once again, we cannot provide any certain explanations, but this appears to be a very necessary area for further study. It is difficult to identify clear causes for the changes in the relationships between the experiences of these different groups from inspection reports. In 2013, when responses from BAME children were worse than those of white children in questions related to respect and safety, the Inspectorate report : Not asked; It is difficult to identify clear causes for the changes in the relationships between the experiences of these different groups from inspection reports. In 2013, when responses from BAME children were worse than those of white children in questions related to respect and safety, the Inspectorate report It is difficult to identify clear causes for the changes in the relationships between the experiences of these different groups from inspection reports. In 2013, when responses from BAME children were worse than those of white children in questions related to respect and safety, the Inspectorate report : Significantly better.
It is difficult to identify clear causes for the changes in the relationships between the experiences of these different groups from inspection reports. In 2013, when responses from BAME children were worse than those of white children in questions related to respect and safety, the Inspectorate report described concerns about serious gang-related violence and found that safety was not sufficiently good.
Inspection reports were most critical in 2015 and January 2017. In July 2015, over 50% of BAME boys and over 60% of Muslim boys reported feeling unsafe. Although not significantly different from white and non-Muslim boys, it is the highest result for this question from the categories. Over 60% of BAME boys report being physically restrained in this survey, which is significantly higher than the response by white boys. It appears that boys from all groups and staff did not feel safe and boys from BAME backgrounds were more likely to be seen as a threat than white boys.
What is clear from the survey results is that in Feltham, as elsewhere across the juvenile estate, BAME children, and to a lesser extent Muslim boys, were reporting different experiences from their white and non-Muslim counterparts. What is less clear is why. The survey results and inspection reports indicate that poor overall levels of safety in Feltham may have had a disproportionate impact on boys from BAME backgrounds, in particular among those who reported high levels of fears for their safety and experienced disproportionate use of restraint.

Discussion and Conclusions
Our analysis of data, from the children in custody annual reports and the report on HMYOI Feltham, confirms that whilst there has been a substantial fall in custody rates, children from different ethnic groups have not benefited to the same extent from this phenomenon in our juvenile justice sector (Lammy 2017). For example, for the 14-year period between 2005 and 2018/19, data show a 76% reduction for white children compared with a 31% reduction for BAME children in custodial sentences. The outcome of these trends is an actual increase in the proportion of BAME children in custody, from 27% in 2007to 45% in 2017(Youth Justice Board 2018. Through a focus on one YOI (namely, Feltham) Inspectorate's reports, since 2002, we have illustrated the dramatic increase in the proportion of BAME boys over the years and how this changes the composition of YOIs.
It is also clear looking at the survey data from across the children's custodial estate and from one institution, HMYOI Feltham in particular, that children from black and minority groups and Muslim and non-Muslim children do not experience custody in the same way.
In some concerning ways, BAME boys and Muslim boys reported their experiences more negatively than their counterparts. For example, BAME boys reported less support when they first arrived in the institution, had poorer access to the mechanisms of procedural justice such as the complaints system, were not as likely to feel respected by staff, and were more likely to have been restrained. On the other hand, BAME boys were less likely to report vulnerabilities or having problems than white boys. We argue that these responses are important in their own right, but further work needs to be done to understand what underlies them.
Muslim boys specifically reported more negatively on their relationship with YOI staff than non-Muslim boys. Their experiences of applications and complaints were more negative, they were more likely to have been restrained, more likely to have had adjudications, and more likely to have been victimized. This tallies with accounts that prisons were perceived as hostile and racist places by Muslim prisoners (Qasim 2018;Bhui 2010;Marranci 2009).
The experiences of young people in YOIs are shaped by some demographic background variables as well as conditions in the YOI. Some of the background variables clearly intersect, for example, being from a BAME background and Muslim or having local authority care experience and being white and disabled. On the basis of the analyses provided in the Children in Custody report, it is difficult to draw firm conclusions beyond the obvious differences between ethnic groups. It would be important to run multivariate analyses on the data to account for the intersection of various factors that impact on the experiences of young people in custody and to understand better how the custodial population differs between different ethnic groups but also other indicators of structural disadvantage in youth justice, such as experiences of care and disability. The lack of understanding of the profile of BAME young people in the youth justice system was highlighted in minutes from the IICSA Inquiry Children in Custodial Institutions Investigation in July 2018 and is particularly urgent given the significant increase in the proportion of BAME young people in custody.
High rates of physical restraint and lower rates of help-seeking do indicate low levels of trust and poor relationships with prison staff. However, the survey results from Feltham suggest the experience of boys from BAME and Muslim backgrounds varies over time as does the population and performance of the institution. The results and the comparison between results from YOIs with differing proportions of their populations from black and minority ethnic backgrounds suggest that, as those proportions shift, perhaps not surprisingly, boys as a whole and boys from BAME groups and Muslim boys begin to experience custody differently. For example, surveys in institutions with the highest proportions of BAME boys reported least concerns about safety; as the proportion of Muslim boys grew in HMYOI Feltham, fewer reported concerns about safety. This may be linked to changes in experiences of victimisation because of ethnicity or religion and is something that would need further research or analysis.
The survey results and inspection reports indicate that poor overall levels of safety may have a disproportionate impact on boys from BAME background, in particular among those who reported disproportionate use of restraint. Previous research studies have suggested differential strategies employed by BAME young people to deal with a hostile prison environment (Wilson 2003;Phillips 2008;Marranci 2009;Bhui 2010;Qasim 2018). It is possible that some prison officers seek refuge in negative and racist stereotypes when dealing with some BAME and Muslim boys (Graham and Lowery 2004).
It may be argued that unconscious bias training could help improve communication in interethnic encounters in the helping professions (Sue et al. 2007). A recent report into Black and Muslim prisoners by the Runnymede Trust suggests, however, that such unconscious bias training is an enormous challenge in the face of indifference and work pressures in prisons (Jolliffe and Haque 2017). Crucially, it is important to recognise that whilst in this so called postracial reality, a focus on unconscious bias training is on the rise; critics have argued that such a framework serves to conceal the harsh reality of structural racism, and that "unconscious bias is an alibi to diminish the recognition, analysis and salience of white supremacy in order to maintain it" (Tate and Page 2018, p. 143).
The contemporary attention to unconscious bias is reminiscent of race awareness training (RAT) programmes of 1980s Britain, which led the late Sivanandan to argue that the RAT industry and its focus on ethnic and cultural sensitivity is "able to mis-appropriate black politics and black history-and degrade black struggle" (Sivanandan 1985, p. 30). We would agree that in the absence of a review of systemic structures and practices in incarcerated spaces, individualised unconscious bias training may well be rather inadequate. In the words of Sivanandan 1985, p. 28), "My business is not to train the police officer out of his 'racism'-but to have him punished for it-if, that is, he is meant to be accountable to the community he serves".
Crucially, serious attention needs to be paid to the structural aspects of incarcerated spaces. Given that only 6% of prison officers in England and Wales are from a BAME background, Jolliffe and Haque (2017) argue that whilst BAME prisoners in their study did not regard a rise in this number as a panacea, they did believe that it could help in terms of "cultural understanding" between staff and prisoners. It would seem that there would need to be a sevenfold increase to address the disproportionality ratio of prison staff and BAME young people in our youth justice sector (Slawson 2017).
As mentioned above, HM Chief Inspector of Prisons (2017) identified the racial disparities in youth justice as an important concern. In doing so, he highlighted the need for local and community accountability and transparency. We have not addressed the factors that lead to the over-representation of BAME and Muslim boys in youth custody in this paper but agree that their over-representation remains an urgent issue for the youth justice sector to address with BAME communities and the component parts of the youth justice system.
The experience that BAME and Muslim boys have once in custody that we have described in this paper has received even less attention than the processes that lead to their incarceration. We suggest there is an urgent need for further research to understand the extent to which these differences are a cause rather than a symptom of the current problems in youth custody and to ensure boys from these backgrounds have the same opportunities as others to access the support and services they need to reduce the risk of recidivism.
We would suggest a number of factors that may together or singularly explain the experience of BAME and Muslim boys and why their experience might differ from that of white and non-Muslim boys: It is feasible that conscious or unconscious bias plays a part, and that systemic structures and processes reproduce structural racism that has a detrimental impact on the experiences of BAME and Muslim boys; It may also be that boys from different groups are more or less likely to frankly report their perceptions in the surveys or that their expectations of how they should be treated are different; Changes in the population of individual institutions and the proportions of boys from different background may influence its dynamics and hierarchies of and between different groups; Improvement or declines in the performance of YOIs as a whole and individual institutions may impact in different ways on different groups.
In line with previous qualitative research and our varied findings from HMYOI Feltham, whether BAME children adopt strategies of "sticking together", "keeping quiet", or "going nuts" in their daily interethnic encounters in a culture of fear, otherisation, and intimidation may well be contextual and/or temporal. Future research is vital in not only addressing the academic research gap but also the practice gap that is of utmost importance if we are to best serve the needs of BAME children in custody. Future research must focus on the mechanics of involving BAME communities and examine structures and processes within incarcerated spaces to understand the impact of institutional racism, through the lens of intersectionality as experienced by BAME children, in our youth justice system. A nuanced methodological focus and analysis that seeks to capture children's subjective and lived reality in these institutions is essential to help address key issues and concerns highlighted in this paper. Future research could employ a creative qualitative approach that goes beyond "inspection-level reports" to help give voice to BAME young people's lived experiences of incarceration. For example, a focus on BAME young people formerly in custody may help create a research environment that is less threatening and more conducive to eliciting subjective narratives without fear of perceived possible negative repercussions. Also, given the survey data available on a range of categories including ethnicity, religion, disability, and care background, we suggest that these data are made available for a more sophisticated intersectional analysis. Crucially, the heterogeneity of the BAME category must be studied to target specific policy and practice interventions.