The policing of cannabis possession in London - National Police Library

Foreword by Allan Gibson, Commander, Specialist Crime Directorate and Mark Simmons, Commander, Territorial Policing

The reclassification of cannabis and the harm it causes continues to be a topical subject of debate within Government, the media and local communities.

Public concern about the use of cannabis is evident in the current high profile of the issue. This report contributes to the important debate about cannabis use and the way it is policed.

Cannabis is a harmful drug creating widespread concern through its increased strength and the long-term harm it has on personal well-being and mental health.

The issue is inextricably linked with that of stop and search, and the long and difficult debate about its disproportionate use. In the last year, following the Metropolitan Police Authority (MPA) Scrutiny on stop and search, the MPS has capitalised on the investment in stop and search to make significant changes in our approach.

Metropolitan Police Service (MPS) policy on the enforcement of cannabis possession and cannabis warnings is led by and reflects current Association of Chief Police Officers (ACPO) guidance. It has significantly decreased the number of people arrested for cannabis possession, reduced friction between the police and young people and enabled the police service to provide greater focus and resources to organisational and community-policing priorities. The Monitoring Report, 2006 indicated a disproportionate impact in disposal outcomes between London’s black and white communities. The MPS is keen to understand why this is happening and to address any underlying issues. I therefore commissioned detailed research resulting in this comprehensive report. This was a collaborative piece of work with respected academics providing an independent overview and transparency to the process. The research has not answered all the questions to what is clearly a complex issue. It does, however, provide a clearer indication of the factors that influence disposal outcomes, and the disparity in charges and cautions between the communities identified. The MPS is committed to ensuring any policy it implements does not disproportionately disadvantage a particular community and is fair.

Allan Gibson Commander Specialist Crime Directorate

The result of these interventions, so far, has been to increase the proportion of stops where a search does not take place to nearly 50%; to improve the recording of self-defined ethnicity; to increase the arrest rate by ensuring that officers focus on those who intelligence states should be searched; and to reduce the disproportionality. Most importantly we have opened the debate to communities to look at what we are doing, how we are doing it and to influence the way officers carry out this sensitive activity. The debate about cannabis usage is being played out in many forums and the truth is complex. What is clear is that Londoners tell us that the misuse of drugs, especially by young people, is a serious issue for them. The open use of drugs, including cannabis, is seen as a signal crime which causes fear and as such is one the MPS needs to robustly deal with. It is an issue which is important to communities and for policing in London, and features as a priority that communities set for the majority of Safer Neighbourhood Teams. This report provides us with the opportunity to continue to develop our response to these issues through a discussion which fully engages local communities in London.

Mark Simmons Commander Territorial Policing

NOT PROTECTIVELY MARKED

THE POLICING OF CANNABIS POSSESSION IN LONDON Examining Ethnic Disproportionality

Gavin Hales Metropolitan Police Strategic Research Unit and Senior Research Fellow, University of Portsmouth Institute of Criminal Justice Studies. July 2007


Acknowledgements The author is grateful to a number of people, without whose assistance this research could not have been completed. Redacted - 1

Gavin Hales MPS Strategic Research Unit Seconded from the University of Portsmouth Institute of Criminal Justice Studies

i


Contents Page Acknowledgements

i

Contents

ii

Figures and tables

iii

Executive Summary

v

1. Introduction

1

1.1 Research questions

1

1.2 Existing research evidence

2

1.3 Equality and proportionality

6

1.4 The relationship between policing and the number of cannabis possession offences

6

2. Methodology

7

2.1 Obtaining data

7

2.2 Data preparation

8

2.3 Data analysis

9

3. Findings

10

3.1 Historical trends 2000 – 2006

10

3.2 Cannabis possession accused October 2005 – December 2006: overview

12

3.3 Offender demographics January – December 2006

19

3.4 In-depth examination of a sample of 144 CRIS records

29

4. Conclusions and discussion

37

5. Appendix A: Proceedings codes

40

6. Appendix B: Cannabis warnings as a percentage of all Sanctioned Detections – 2006/07 by borough

41

7. Appendix C: Ethnic Groups

42

8. Appendix D: Factors impacting on cannabis policing outcomes

43

9. Bibliography

44

ii


Figures and tables Figures Page Figure 1

Percentage of school pupils (11 to 15yrs) reporting cannabis use in past year

2

Figure 2

Cannabis use reported to the British Crime Survey (1996 to 2005/06)

3

Figure 3

Total recorded cannabis possession offences in the MPS, January 2000 – December 2006

10

Figure 4

Examining the contribution of cannabis warnings to OBTJ performance (April 2005 to April 2007)

11

Figure 5

Cannabis Possession Accused – By Owning Borough (October 2005 – December 2006)

12

Figure 6

Cannabis Possession Accused – Cumulative Percentage By Electoral Ward

14

Figure 7

Examining the relationship between personal robbery and cannabis possession offences at a ward level

14

Figure 8

Cannabis Possession Accused – Recorded Date (with trend line & moving average)

15

Figure 9

Cannabis Possession Accused – By Committed On/From Day of Week

16

Figure 10

Cannabis Possession Accused – By Committed On/From Time

16

Figure 11

Cannabis Possession Accused – By Committed On/From Day and Time

17

Figure 12

Cannabis Possession Accused – By Committed On/From Time and Adult/Youth

17

Figure 13

Examining the gender balance – white and black Caribbean populations, London 2006

20

Figure 14

Cannabis Possession Accused by Age and Gender

23

Figure 15

London Male Age Structure – White, black Caribbean, black African and ‘other’ black – GLA Population Estimate Data for 2006

23

Figure 16

Age specific offence rates by self-defined ethnicity (males, 2006)

24

Figure 17

Age specific offence rates as a multiple of the white offence rate, by selfdefined ethnicity (males, 2006)

24

Figure 18

Comparing numbers of offences (total numbers accused) and offence rates: black Caribbean males in 2006

26

Figure 19

Factors impacting on cannabis policing outcomes

38

Figure 20

Cannabis warnings as a percentage of Sanctioned Detections – by borough, 2006/07

41

Figure 21

Scatter plot of the relationship between the number of cannabis warnings issued and cannabis warnings as a percentage of sanctioned detections – by borough, 2006/07

41

iii


Tables Page Table 1

Prevalence of cannabis use (percentage) (16 to 59 year olds) – grouped ethnicity (BCS 2001/02)

4

Table 2

Prevalence of cannabis use (percentage) (16 to 59 year olds) (BCS 2001/02)

4

Table 3

Proportion of individuals found in possession of cannabis who were from the borough in question

13

Table 4

Repeat cannabis possession offenders, Oct05 – Dec06

18

Table 5

Cannabis Possession Accused By Gender, Age Group and Self-defined Ethnicity – Recorded date January to December 2006

21

Table 6

Offence and offender rates by Self-Defined Ethnicity (all males, 2006)

21

Table 7

Offence and offender rates by Self-Defined Ethnicity (adult males, 2006)

22

Table 8

Offence and offender rates by Self-Defined Ethnicity (male youths, 2006)

22

Table 9

Comparing proceedings rates for black Caribbean and white adult males accused of cannabis possession (2006)

26

Tables 10 & 11

Proceedings for Cannabis Possession, by Self-Defined Ethnicity: male adults (18-69yrs) and youths (10-17yrs) (2006)

28

Table 12

Number and rate charged with cannabis possession, by 4-year age group and Self-Defined Ethnicity (males aged 10-69 years, 2006)

28

Table 13

Ethnicity of accused by borough (5 borough sample)

30

Table 14

Whether arrested and how proceeded against by ethnicity (5 borough sample)

31

Table 15

Whether arrested and how proceeded against by ethnicity – confirmed arrests (5 borough sample)

31

Table 16

Reason for arrest by ethnicity: arrests for cannabis possession only (5 borough sample)

32

Table 17

Reason for arrest by ethnicity: arrests in context of/for other offences (5 borough sample)

32

Table 18

Cases where the initial police / offender contact was justified (5 borough sample)

33

Table 19

Cases where the arrest decision was justified (5 borough sample) – only where an arrest was confirmed

34

Table 20

Cases where the proceedings decision was justified (5 borough sample)

34

Table 21

Cases where the proceedings decision was justified (5 borough sample) – only where an arrest was confirmed

34

Table 22

Grouping proceedings codes

40

Table 23

Aggregating ethnic groups

42

iv


Executive Summary Introduction 1. The Metropolitan Police Service (MPS) Drugs Directorate maintains and reviews all MPS policies relating to drugs and recently launched the MPS Drugs Strategy for 2007 to 2010 (MPS 2007a). This includes commitments to tackle blatant drug use in public places, combat the open selling of drugs and disrupt cannabis cultivation. 2. The possession of cannabis, the subject of this research, is policed because: cannabis is illegal and can be damaging to health; drug dealing and open drug use are identified by the public as activities that impact on their, or their children’s, quality of life; drug dealing and open drug use are associated with a range of anti-social behaviours; and, vulnerable members of the public should be protected. 3. The Drugs Directorate commissioned the MPS Strategic Research Unit to undertake in-depth research into the impact of the policing of cannabis possession on London’s diverse ethnic communities, examining both the scale of ethnic disproportionality and the factors underlying it. This followed the 2006 cannabis possession policy Monitoring Report, which raised concerns around apparent ethnic disproportionality in relation to policing outcomes (MPS, 2006). 4. Forty-seven thousand MPS Crime Recording Information System (CRIS) records have been examined covering October 2005 to December 2006, with a particular focus on males accused of cannabis possession during 2006. Greater London Authority (GLA) population estimates for London in 2006 have been used to examine age- and ethnicity-specific offence rates and disproportionality. Finally, a sample of 144 CRIS records has been examined in detail to explore the reasons for disproportionality, with implications for police supervision. Key findings 5. National data from the British Crime Survey (BCS) indicate that cannabis use differs somewhat by ethnicity, usage rates being highest for mixed ethnicity respondents (25 per cent had used cannabis in the past year), especially mixed white and black Caribbean ethnicity respondents (33 per cent) (Aust and Smith, 2003: 3). Overall cannabis usage rates are equal for white and black respondents (both 11 per cent); however, variance exists within the black ethnic category: self-reported cannabis usage rates are much higher for black Caribbean (17 per cent) than black African respondents (3 per cent). Cannabis use by black Caribbean respondents is somewhat higher than their white British counterparts (11 per cent). 6. The recent Home Affairs Committee report on Young Black People and the Criminal Justice System reported that, ‘Young people and their youth workers told us that cannabis use was a particular problem among young black people’ (House of Commons, 2007: 63). 7. National survey data also indicate that cannabis use has been falling. In contrast, the number of cannabis possession offences recorded by the MPS has risen rapidly since the decision in April 2004 to count them as sanctioned detections. In 2006/07 cannabis warnings constituted 15.6 per cent of all sanctioned detections in the MPS (borough range 4.9 per cent to 25.7 per cent). 8. Cannabis possession offences are spatially concentrated. Between October 2005 and December 2006 17.7 per cent of cannabis possession offences recorded by the MPS were recorded in two boroughs (Westminster and Brent) and 10.1 per cent of offences were recorded in only 7 out of 624 electoral wards (1.1 per cent of wards). 9. At a ward level, cannabis possession offences are correlated with personal robbery rates (Rsq=0.44), which suggests that levels of cannabis offences may at least partially reflect levels of policing targeting other crime problems. Robbery has been and remains an organisational priority for the MPS, in relation to which stop and search is one of the mainstays of the policing response and also a significant source of cannabis possession offence detections. In relation to this, three factors appear to be critical: levels of front-line police resources, the quality of local analysis in producing ‘suspect profiles’ and the interpretation by front-line officers of ‘reasonable suspicion’. 10. In 2006, 94.5 per cent of cannabis possession offenders were male and 8.4 per cent were youths. The age distribution of cannabis possession offences is highly concentrated in the late teens and early-20s; the peak age for offenders was 18 years old.

v

NOT PROTECTIVELY MARKED 11. Overall disproportionality for male offenders in 2006 was as follows (multiple of the offences per 1,000 population rate for white offenders): Ethnicity Bangladeshi Black African Black Caribbean Black ‘other’ Chinese Indian ‘Other’ ‘Other’ Asian Pakistani

Adults (ages 18-69) 2.4 2.6 7.6 11.4 0.4 1.0 2.4 2.9 2.0

Youths (ages 10-17) 1.0 1.4 4.3 3.3 0.5 0.5 1.3 1.4 0.7

All (ages 10-69) 2.1 2.4 7.1 8.9 0.4 0.9 2.3 2.7 1.8

12. Black Caribbean and black ‘other’ males were found in possession of cannabis at exceptionally high rates: up to 300 offences and 280 offenders per 1,000 in a single year for 21 year olds, and up to 10 times the rate of their white counterparts. It is clear that cannabis possession enforcement does not fall equally on London’s different ethnic communities. 13. Disproportionality increases with level of sanction and is greater for adults than youths. Having been found in possession of cannabis, black Caribbean male adults were less likely to receive cannabis warnings, more likely to receive cautions and much more likely to be charged than their white counterparts. 14. The audit trail for key police officer decisions – establishing reasonable grounds for initial suspicion, arrest and proceedings – is fragmented across multiple records and systems, and even then appears often to be incomplete. Using electronic records from the CRIS, Stops and Custody databases, the Police National Computer and CrimInt (MPS Criminal Intelligence database) to examine a sample of 144 records from five boroughs, the initial police/offender contact appeared reasonable in 52 per cent of cases, the arrest decision in 94 per cent of cases (where an arrest took place), and the proceedings decision in 72 per cent of cases (54 per cent of cases where an arrest took place). 15. The analysis raises a number of questions. Not least, whether the independent supervision of individual sequential decisions and the fragmented audit trail constitute effective end-to-end supervision that can ensure the equal treatment of different individuals and groups when they come into contact with and are dealt with by the police. Data problems 16. The ‘how arrested’ variable recorded in CRIS should be considered unreliable, therefore it is not possible at present to routinely monitor the impact of MPS policies and practices around the arrest decision. This is due to a combination of two factors: (a) Arrests not being recorded when they had taken place (e.g. in relation to offenders who had been charged and must therefore have been arrested). (b) Arrests being recorded when they did not take place, notably in relation to the issuing of cannabis warnings (this appears to have occurred systematically in some boroughs). 17. Evidence that unique individuals have multiple self-defined ethnicities raises questions about the reliability and validity of the self-defined ethnicity variable. For example, to what extent is it selfdefined? Summary of recommendations on data quality and recording 18. Guidance should be issued at the earliest opportunity that arrests should only (and always) be recorded in CRIS when an arrest actually took place. Cannabis warnings issued in the absence of an arrest should not be recorded as arrests, either in CRIS or in the Stops database. Furthermore, only valid custody numbers should be recorded in CRIS. 19. The MPS should consider commissioning research to examine the reliability and validity of the selfdefined ethnicity variable in CRIS. This information is critical to race equality impact assessments, as well as being of broader value to strategic intelligence products and other analysis.

vi

NOT PROTECTIVELY MARKED 20. Consideration should be given to making it a requirement that the CRIS record ‘Details’ (Dets) screen includes a comprehensive, transparent and unambiguous account of the grounds (where relevant) for: (a) Initial police/offender contact – why reasonable grounds for suspicion existed; (b) Arrest – why necessary; and (c) Proceedings decision – why the chosen proceeding was appropriate and why the alternatives were inappropriate. Conclusions 21. Many ethnic minority groups, most notably black Caribbean and ‘other’ black males, are both heavily over-represented overall and more likely to receive more punitive sanctions – two factors that are likely to be very closely related. 22. Six ‘domains’ of factors impacting on cannabis policing outcomes are identified, placing the impact of policing in its wider context. Profile of cannabis possession accused Organisational culture Legislative domain

Performance domain

Police officer domain

Location/ place domain

Offender domain

Lifestyle domain

Cannabis reclassified

PSA Targets: reduce overall crime

Professional experience

Demographic profile

Behaviour: visibility, attitude

Socio-cultural factors

Personal prejudices / bias

Local (criminal) economy

Quality of supervision

Crime problems and policing priorities

Reasonable grounds for suspicion Legislation used for stop/search

OBTJs: maximise SD rate Organisational priorities (robbery) Scrutiny (e.g. Operation Pennant)

Age, gender, ethnicity Relationship to suspect profiles

Economic factors Peer group effects

Quality of local analysis Level and deployment of police resources

23. The policing of cannabis in the MPS appears to be positioned at the intersection of three key issues: (a) Stop and search, especially as one of the mainstays of the policing response to robbery, in relation to which levels of front-line police resources, the profile of robbery suspects and the interpretation of ‘reasonable suspicion’ appear to be of fundamental significance. (b) Other efforts to tackle local crime and disorder problems, including drug dealing and open drug use. (c) Local interpretation of performance culture, in relation to which cannabis warnings constitute a significant source of sanctioned detections. 24. This research has contributed to developing a better understanding of both the scale of disproportionality and the factors that underlie it. It is clear that the impact of policing cannabis does not fall equally on different ethnic groups. It is also clear that the reasons for that inequality of impact are highly complex, encompassing a range of intersecting factors that include, but are not limited to, policing. 25. Specific recommendations have not been made about responding to these findings. However, it seems likely that any response needs to involve local communities in a conversation about the appropriate solutions to local problems of crime and disorder.

vii


1. Introduction The Metropolitan Police Service (MPS) Drugs Directorate maintains and reviews all MPS policies relating to drugs and recently launched the MPS Drugs Strategy for 2007 to 2010 (MPS, 2007a) 1 . This includes commitments to tackle blatant drug use in public places, combat the open selling of drugs and disrupt cannabis cultivation. The possession of cannabis, the subject of this research, is policed because: •

Cannabis is illegal and can be damaging to health;

•

Drug dealing and open drug use are identified by the public as activities that impact on their, or their children’s, quality of life;

•

Drug dealing and open drug use are associated with a range of anti-social behaviours; and

•

Vulnerable members of the public should be protected.

Like all public bodies, the MPS is obliged to undertake Equality Impact Assessments (EIAs) of its procedures and practices to ensure that they impact fairly on different groups in society. This report presents the findings of a research project examining the race equality impact of the policing of cannabis possession in the Metropolitan Police District (Greater London excluding the City of London and British Transport Police). This follows the MPS cannabis possession policy Monitoring Report produced by the MPS Drugs Directorate in May 2006 that concluded: “…the MPS is aware and has concerns around the disproportionality indicated by the analysis [conducted to inform the 2006 report], which has identified what could be construed as a policy that delivers a disproportionate impact to the African/Caribbean community.” (MPS, 2006: 7) The 2006 Monitoring Report committed the MPS to conducting “further in-depth analysis” on these matters (ibid.: 6), which ultimately led to the commissioning of the research presented here. This is based on an examination of the Metropolitan Police Crime Recording Information System (CRIS) records of almost 47,000 individuals accused of possession of cannabis between October 2005 and December 2006, as well as an in-depth examination of a sample of 144 individual records. Significant legislative changes have impacted on the policing of cannabis in recent years. These have seen cannabis reclassified as a Class C controlled drug from January 2004 and the introduction of Cannabis Warnings in the same year as time-efficient (and therefore cheaper) bureaucratic transactions for dealing with the discovery by police of individuals in possession of small quantities of the drug for their personal use. 2 Now that they are well established, this is an appropriate time to examine the race equality impact of these procedures. This report is divided into four chapters. The rest of Chapter 1 will cover the research questions before reviewing existing research evidence on cannabis use and the policing of cannabis. Consideration will then be given to the concepts of equality and proportionality, before a brief examination of the relationship between policing and the number of cannabis offences recorded by the police. Chapter 2 will provide an overview of the research methodology, including an initial consideration of data quality. Chapter 3 presents the findings, including a consideration of the historical context, an overview of CRIS records for the period October 2005 to December 2006, a more focussed examination of the records for 2006 with reference to the impact on different ethnic groups, and finally an in-depth review of a sample of 144 CRIS records. Chapter 4 presents the conclusions and a discussion about the implications of the research.

1.1 Research questions The research questions being addressed by this research are as follows: “What are the key variables accounting for the disproportionality of outcomes in the exercise of the MPS Guidelines on arrests for cannabis possession, and how much of this variance is accounted for by police discretion?” 1

The MPS Drugs Strategy for 2007 to 2010 is available at http://www.met.police.uk/drugs/publications/drugsstrategy.pdf See http://www.met.police.uk/foi/pdfs/policies/cannabis_policy.pdf for the MPS policy on cannabis and http://intranet.aware.mps/Corporate/Policy/Drugs/Cannabis-Policy.htm (internal MPS only) for the Standard Operating Procedures (SOPs). 2

1


1.2 Existing research evidence Evidence about cannabis use and its relationship to ethnicity and policing is available from a number of national surveys and previous academic research. This section is intended to provide a brief overview of some key findings.

1.2.1

Cannabis use trends

Data on cannabis use trends over time are available from a couple of national surveys: firstly a national survey of school pupils ages 11 to 15 and, secondly and in more detail, the British Crime Survey. 3 Both indicate that at the national level cannabis use is falling. National pupils survey A 2006 national survey of secondary school pupils aged 11 to 15 shows falling cannabis use since 2001 (National Centre for Social Research and National Foundation for Educational Research, 2007: Table 8) (see Figure 1). Figure 1: Percentage of school pupils (11 to 15yrs) reporting cannabis use in past year Percentage of Pupils (11 to 15yrs) Reporting Cannabis Use in Past Year (Source: National Centre for Social Research and National Foundation for Educational Research, 2007: Table 8) 16

14

13.4

13.2

13.3

12

11.3

11.7 10.1

Percentage

10

8

6

4

2

0 2001

2002

2003

2004

2005

2006

Year

British Crime Survey (BCS) The British Crime Survey for 2005/06 indicates falling cannabis use in the last year and in the last month, for both 16 to 24 year olds and 16 to 59 year olds (Roe and Man, 2006: 45 and 49) (see Figure 2).

3

Being sample-based surveys of self-reported offending, both of these sources are subject to limitations to their reliability and validity. This includes the potential for respondents to exaggerate or conceal the extent of their offending behaviour and problems relating to sampling ‘hard to reach’ groups such as those in temporary accommodation (BCS) or excluded or absent from mainstream education (national school survey). More importantly, analysis of the 1994, 1996 and 2000 sweeps of the BCS indicates that non-white respondents are more likely than their white counterparts to refuse to participate in answering self-reported drug use questions, with important implications for potential ‘non-response bias’ – that is the possibility that those refusing to participate differ in important (and non-random) ways from those who agree to participate. In the 2000 BCS, for example, refusal rates relating to the ‘drugs component’ of the questionnaire were 1.5 per cent for white respondents, 4.1 per cent for ‘all black groups’, 7.8 per cent for Indian respondents and 19.3 per cent for Pakistani/Bangladeshi respondents (Ramsey et al., 2001:48, table 4.1).

2

NOT PROTECTIVELY MARKED Figure 2: Cannabis use reported to the British Crime Survey (1996 to 2005/06) Cannabis Use Reported to the British Crime Survey (England and Wales) (Source: Roe and Man, 2006) 30

Percentage who used cannabis

25 16 to 24 - last month 16 to 24 - last year 16 to 59 - last month 16 to 59 - last year

20

15

10

5

0 1996

1998

2000

2001/02

2002/03

2003/04

2004/05

2005/06

Year

Nationally (England and Wales), the BCS data for 2005/06 indicates that 8.7 per cent of 16 to 59 year olds and 21.4 per cent of 16 to 24 year olds used cannabis at least once ‘during the last year’; usage was highest for 16 to 19 year olds (21.8 per cent) (not shown in Figure 2: see Table A4.1 in Roe and Man, 2006: 59). The equivalent figures for use ‘in the last month’ are 5.2 per cent (16 to 59) and 13.0 per cent (16 to 24), with usage being highest for 20 to 24 year olds (13.1 per cent). The BCS data also show that cannabis use was higher for males than females. Amongst 16 to 24 year olds, for example, 17.1 per cent of males and 8.9 per cent of females reported having used cannabis in the last month; 26.3 per cent and 16.6 per cent respectively in the last year; and 44.6 per cent and 35.6 per cent respectively in their lifetime. Finally, the BCS data indicate that cannabis use in the last year by 16 to 59 year olds was slightly higher in London (8.8 per cent) than nationally (8.7 per cent), but that for 16 to 24 year olds (the peak ages for use) usage in the last year in London (16.7 per cent) was substantially lower than nationally (21.4 per cent).

1.2.2

Cannabis use by ethnicity

The recent Home Affairs Committee report on Young Black People and the Criminal Justice System reported that, ‘Young people and their youth workers told us that cannabis use was a particular problem among young black people’ (House of Commons, 2007: 63). The Committee went on to recommend that, ‘the Department of Health explore ways to determine effectively the extent of drug use among young people of different ethnicities’, further noting that, ‘We also believe there is a need for a more detailed study of cannabis use and its use by, and effects on, young people of different ethnicities’ (ibid.: 64). Evidence about the relationship between ethnicity and cannabis use is, however, already available from at least three sources: the British Crime Survey (BCS); the Offending, Crime and Justice Survey (OCJS); and the New English and Welsh Arrestee Drug Abuse Monitoring (NEW-ADAM) programme, looking for evidence of prior drug use in the urine of people arrested by the police. British Crime Survey Data on cannabis use by ethnicity indicate a complex picture at the national level. Analysis of the 2001/02 British Crime Survey (the most recent such analysis available, covering 16 to 59 year olds in England and Wales) showed overall levels of cannabis use were highest amongst mixed race individuals (25 per cent had used cannabis in the last year, 16 per cent in the last month), followed by white and black individuals (both 11 per cent and 7 per cent respectively) (Aust and Smith, 2003: 2) – see Table 1 below. 4 These findings broadly concur with previous Home Office analyses of national survey data (for example: Mott 4

This analysis has not yet been repeated for subsequent sweeps of the BCS.

3

NOT PROTECTIVELY MARKED and Mirrlees-Black, 1995; Graham and Bowling, 1995; Ramsay and Spiller, 1997; Ramsey, Baker, Goulden, Sharp and Sondhi, 2001; see also Phillips and Bowling, 2003: 30). Table 1: Prevalence of cannabis use (percentage) (16 to 59 year olds) – grouped ethnicity (BCS 2001/02)

Last year Last month

White

Black

Asian

Mixed

11 7

11 7

4 2

25 16

Chinese/ other 6 4

A more detailed breakdown of the ethnicity of respondents in 2001/02 shows, amongst other things, that cannabis use was higher among black Caribbean individuals than white British, but both were higher than black African respondents and much lower than mixed white and black Caribbean and mixed white and black African respondents (Table 2) (Aust and Smith, 2003: 3). The difference between black Caribbean and black African respondents is noteworthy and may highlight important cultural factors, although it is also notable that relatively little difference exists between mixed white and black Caribbean and mixed white and black African respondents. Table 2: Prevalence of cannabis use (percentage) (16 to 59 year olds) (BCS 2001/02) Last Year 33 29 17 15 17 3 13 11 9 8 4 5 7 4 3 7

Mixed – white and black Caribbean Mixed – white and black African Mixed – white and Asian Mixed – any other mixed background Black – Caribbean Black – African Black – other black background White – British White – Irish White – other white background Asian – Indian Asian – Pakistani Asian – Bangladeshi Asian – other Asian background Chinese Other ethnic group

Ever 54 44 36 27 33 8 42 30 20 28 9 10 13 12 10 17

Two issues merit further comment. Firstly, the detailed breakdown of ethnic groups comes at the cost of small sample sizes, which implies larger confidence intervals – that is, the prevalence use figures for smaller groups (especially the ‘mixed’ groups) should be treated with caution. Secondly, these data are not informative about the frequency of use. For example, two respondents could both have used cannabis in the last year but one of them could have done so several times per day and the other only once. Unfortunately, the data on use in the past month, which would provide some additional clues on frequency of use, could not be analysed in terms of a detailed breakdown of ethnicity due to the small numbers involved (Aust and Smith, 2003: 3). Offending, Crime and Justice Survey Elsewhere, the Offending, Crime and Justice Survey (2003), a national survey of self-reported offending by 10 to 65 year olds, found similar patterns, with cannabis use ‘in the last year’ highest for mixed race respondents (16 per cent), followed by white respondents (12 per cent), black or black British respondents (10 per cent), individuals from ‘other’ ethnic groups (6 per cent) and Asian or Asian British respondents (4 per cent) (Sharp and Budd, 2005: Table 4.1, p18). The ‘other’ and Asian or Asian British rates of use were statistically significantly lower than the white rate (at the 5% level).

4

NOT PROTECTIVELY MARKED NEW-ADAM: Urine testing of arrestees Finally, some evidence about cannabis use rates by ethnicity can be found in research analysing the urine of individuals in police custody for evidence of previous drug consumption (NEW-ADAM: the New English and Welsh Arrestee Drug Abuse Monitoring Programme). The published data indicates that there was no significant difference in the percentage of arrestees testing positive for cannabis between white arrestees (48 per cent) and their non-white counterparts (52 per cent) (Holloway and Bennett, 2004: 13, table 2.1b). Unpublished data breaks this data down into disaggregated ethnic groups and indicates that white and black arrestees tested positive for cannabis at roughly similar rates (Phillips and Bowling, 2003: 33), although the samples for the non-white ethnic groups were very small and therefore likely to be especially unreliable when viewed separately. Moreover, it should be said that the individuals tested for the NEW-ADAM programme were not randomly drawn from the wider population and therefore the sample is likely to have been influenced by non-random factors. Conclusions At the very least, it seems reasonable to conclude that at a national level cannabis use rates (in terms of having used the drug at least once in the last year) are broadly similar for white and black individuals, although both are lower than mixed ethnicity individuals. It is also clear, however, that variance exists within the broader ethnic categories, most notably in relation to the difference between black Caribbean and black African individuals.

1.2.3

The policing of cannabis

A wide range of existing research has a bearing on the analysis presented here, including studies of stop and search (e.g. FitzGerald, 1999; Quinton, Bland and Miller, 2000; MVA and Miller, 2000; Waddington et al., 2004) and the operation of the criminal justice system (e.g. Bowling and Phillips, 2002). These raise a number of issues, including the need to consider differences between the population ‘available’ to be stopped and the residential population of particular areas, the use of police discretion, the relationship between police activity and legislation, and the entrenchment of disadvantage through the successive stages of the criminal justice system. Two particular projects, conducted under the auspices of the Joseph Rowntree Foundation (JRF), have focused specifically on the policing of cannabis. The first, published in 2002 (May et al., 2002) when cannabis was a Class B controlled drug, highlighted important differences in the approaches taken by individual police officers. In particular, across the four research sites 11 per cent of officers who made any arrests for possession of cannabis accounted for 37 per cent of arrests, while 3 per cent of officers accounted for 20 per cent of arrests. To some extent these differences were explained by reference to the varying nature of the sites, however other factors identified included: using arrests for possession as a ‘door opener’ to other offences (in 8 per cent of cases a cannabis arrest led to the detection of another offence – including a serious offence in 1 per cent of cases); arresting persistent offenders to inconvenience them; and, new (probationer) officers being encouraged to make arrests to ‘learn the ropes’, in relation to which cannabis possession provides a ready supply of suspects. Importantly, the research also found that 69 per cent of officers had dealt informally with the discovery of cannabis at some point in their careers – that is, without giving a formal sanction. Members of the 2002 research team then revisited the policing of cannabis post-reclassification (May et al., 2007) when they found considerable inconsistency in the application of what were then termed ‘street warnings’ (now known as cannabis warnings to reflect the fact that they can be given anywhere), as well as variance in the circumstances of cannabis finds. In relation to the latter, one of the sites was characterised by the specific targeting of cannabis possession offences, while the other three were dominated by cannabis offences discovered in the context of arrests for other offences. The 2007 research found that the decision as to whether to arrest offenders was influenced by a range of factors, including: the official guidelines (the ‘rule book’); the amount of cannabis found; the location of the offence; and, the offender’s offending history, age and attitude. Importantly, the researchers found an over-representation of black and minority ethnic (BME) offenders that could not be explained, and highlighted the impact of government-set targets on the policing of cannabis, with officers ‘trawling for possession offences’ (May et al., 2007: x) in order to increase the number of sanctioned detections

5

NOT PROTECTIVELY MARKED recorded. They also found that, ‘the amount of money saved [by reclassification and the introduction of cannabis warnings] is smaller than originally expected, reflecting the fact that arrests for possession are still commonplace’, adding, ‘It is questionable whether savings… are being redirected to address class A drug problems’, and that, ‘In reality, the savings are likely to result in officers responding slightly quicker or slightly better to public demand’ (May et al., 2007: x).

1.3 Equality and proportionality It needs to be stated that equality and proportionality are not necessarily the same thing. If, for example, systematic differences exist between population groups – perhaps one group uses drugs at a higher rate or more visibly than their counterparts from another group – then they can be treated equally in respect of the application of laws, policies and procedures to the objective facts of each case, but with the result that disproportionate (that is, differential) outcomes are seen. It should also be said that these matters are not easily disentangled. It may be relatively straightforward to establish whether disproportionality exists – that is, outcomes appears to differ by ethnicity. However, it will almost always be much more difficult to understand whether that is as a result of systematic differences in the characteristics of different groups, or rather because they are treated differently on a systematic basis. A range of factors will have a bearing on such analysis, including the rate and visibility of offending, the intensity and style of policing, population structures (age profiles), the broader policing context (e.g. other crime problems), and the use of individual officer discretion. To some extent this complexity will be beyond the capacity of this research, which inevitably means that questions will remain. Finally, it is important to note a specific point that relates to the way ‘disproportionality’ is thought about. It will be seen that the possession of cannabis is highly concentrated within a fairly narrow age range and as such comparisons with entire populations may be especially inappropriate because of the way population structures differ by ethnicity. As a consequence, this research will seek to examine proportionality on an age- and ethnicity-specific basis.

1.4 The relationship between policing and the number of cannabis possession offences Following the brief discussion of the relationship between equality and disproportionality, it also needs to acknowledged from the outset that the number of cannabis possession offences recorded by the police will not be directly or solely related to the rate of cannabis possession in London. Indeed, it will be seen that the number of offences recorded by the police has risen rapidly while, as shown above, the British Crime Survey has recorded a fall in the percentage of respondents having used cannabis in the past year and the past month (in England and Wales). In general, the number and profile of individuals found in possession of cannabis by the police will be the consequence of the complex interaction of a number of important factors. These include: -

The number and profile of cannabis users;

-

The locations and methods by which cannabis users store, transport and consume their drugs (e.g. private vs. public locations, carried on the person vs. stored in a private or secure location);

-

The scale and distribution of front-line policing resources (spatially and temporally);

-

The specific policing tactics used in a given location (in particular the ratio between stop and search and other activities) – which will to some extent be a reflection of the local crime problems the police are seeking to tackle;

-

The use of officer discretion; and

-

The population structures of specific ethnic groups.

In addition to being explored in the quantitative analysis, these matters will ultimately be considered in more detail in the conclusions and discussion, when an effort will be made to tie the quantitative analysis to a more descriptive narrative situating the policing of cannabis in its broader context.

6


2. Methodology The research questions require a two-phase approach: firstly to understand the degree of disproportionality and where it arises, and secondly to examine in detail the factors contributing to this disproportionality, particularly around the arrest decision. The first research phase, therefore, is based on a strategic analysis of recorded crime data, which includes a range of variables relating to offences, offenders, and the criminal justice process. The present research is particularly concerned with the demographics and treatment of individuals found to have been in possession of cannabis and who have been accused as such. It will also, however, consider such questions as where the offences were detected. The second phase of the research has involved an in-depth analysis of a small sample of CRIS records, seeking a better understanding of the factors underlying the disproportionate outcomes observed in the overall data. The latter has also drawn on other documentary sources (e.g. MPS Stops Database, MPS Custody Database and Police National Computer) to explore the full circumstances of each case. In the early stages of this project, the MPS Custody Database was considered as a means of identifying individuals who had been arrested for possession of cannabis. This was rejected, however, when it became apparent that (a) it is not possible to identify custody records in relation to which cannabis possession was not the primary offence (e.g. cannabis found in the possession of someone arrested for theft), and (b) the code relating to cannabis possession is used very inconsistently – and not at all by many boroughs.

2.1 Obtaining data In general terms the records identified and data extracted from CRIS were selected on the following basis 5 : -

One or more individuals entered as Accused (that is entered on the CRIS Accused screen); with

-

An Original Classification 6 of 92/66 (possession of cannabis) 7,8 ; and with

-

A Recorded Date between 1st October 2005 and 31st December 2006 (inclusive).

The recorded date is the date on which the CRIS record was created. The start date of 1st October 2005 was selected as the date from which cannabis warnings were introduced as a separate ‘proceedings code’ in CRIS. The end date of 31st December 2006 was chosen partly to allow coverage of a full 15 months of records, and also to ensure that sufficient time had elapsed between the offences being recorded and the CRIS records being extracted (26th February 2007) that essentially all records had been fully classified and assigned a proceedings code (to indicate how the offender had been dealt with). Because personalised data (name and date of birth) were obtained, each row of data relates to one individual (more than one of whom can appear in an individual CRIS record), although individuals may appear more than once in the data if they had been found in possession of cannabis on more than one separate occasion. A total of 48,805 individual records (rows of data) were obtained from the live version of CRIS, current as at 26th February 2007 9 .

5

The author is most grateful for the assistance of Redacted – 2 in extracting the data from FBOC, the live version of CRIS. This is the classification of the offence used for official crime counting purposes. 1,470 records with an Original Classification of 92/86 (possession of cannabis with intent to supply) were also extracted, but are not analysed in this report. 8 Note that records were not extracted and analysed here where the possession of cannabis was recorded as a Subsidiary Code alongside a different principal offence – for example Theft & Handling or Violence Against the Person. A previous examination of CRIS data indicated that only around 0.5 per cent of cannabis possession offences were recorded in that way (an initial analysis of CRIS records for the period October 2005 to September 2006, Performance and Information Bureau request 2665). 9 Being a live system the data may be subject to minor changes over time. However, having been extracted eight weeks after the end of period under consideration, the vast majority of any required changes should have already taken place. 6 7

7


2.2 Data preparation A number of additional variables were added to the data to allow different analyses to be undertaken, for example combining ages and proceedings into groups (see Appendix A for detail concerning the latter). Whether arrested Most problematically, it was necessary to construct a ‘corrected’ variable to indicate whether individuals had been arrested or not. Although CRIS includes a variable recording how an individual was arrested (the ‘how arrested’ variable), this is not a mandatory field and consequently was not always completed when it should have been; conversely, it was also sometimes completed when it should not have been – as will be seen in section 3.4 below. For example, individual CRIS records may have a custody number (indicating that the accused individual was placed in police custody and must therefore have been arrested) or a proceedings code indicating that the individual was charged with an offence (and must therefore have been arrested), but not have a ‘how arrested code’. Initially, however, individuals were recorded as having been arrested in the following cases: (i)

Where a ‘how arrested’ code was entered;

(ii)

Where a valid custody number was entered 10 ;

(iii) Where one of the following proceedings codes requiring arrest was entered: -

AB – Summons AF –1st Time Charged AS – 2nd Time Charged AT – 3rd Time (or more) Charged BA – Youth Reprimand (1st Offence) BB – Youth Warning (1st Offence) BC – Youth Warning (2nd Offence) BD – Youth Warning (2 Years+) BJ – Adult Caution BK – Adult Conditional Caution CH – Taken Into Consideration D6 – CPS Decline Prosecution D7 – Police Decline Prosecution D8 – Time Expired

On the basis of this logic, 7.3 per cent of the 46,966 CRIS records analysed overall (see Data Analysis below) lacked a ‘how arrested’ code when one should have been completed (17.5 per cent of all arrests). 11 It should also be noted that subsequent in-depth analysis of a sample of CRIS records (see section 3.4 below) has highlighted a systematic problem with arrests being recorded when none took place. Overall, therefore, the ‘how arrested’ variable in CRIS should almost certainly be considered unreliable and not suitable on its own for analysis of ‘arrests’. It is recommended that this matter is given urgent attention because it significantly impairs the ability of the MPS to monitor the impact of its policies and procedures around the arrest decision.

10 A close inspection of the custody numbers entered in the CRIS records under consideration indicated that in some cases boroughs appear to have used a ‘holding number’ when the accused was not actually taken into custody. For example, out of a final total of 46,966 CRIS records examined, 1,121 had a custody number that did not appear to be valid (ending 00000 or 99999), including 976 entered by Wandsworth borough as “WW/00000”. This is believed to have resulted from a misunderstanding about correct procedures, most notably in respect of individuals who received a Cannabis Warning (97.5 per cent of cases). 11 This issue could be resolved in one of two ways: (i) CRIS could be modified to ensure that a ‘how arrested’ code is mandatory in cases where a custody number and/or a relevant proceedings code is entered; or (ii) CRIS enquiries could be developed to identify records requiring further attention with (a) a valid custody number, and/or a proceedings code requiring an arrest, but no “how arrested” code, or (b) a “how arrested” code but no custody number.

8

NOT PROTECTIVELY MARKED Identifying individuals Finally, the personalised data (names and dates of birth) were used to construct a variable intended to allow the identification of repeat offenders. This was achieved by merging the surname, first three letters of the forename and date of birth (as an integer) – as the following hypothetical example illustrates: Surname: Forename: Date of birth: Personal ID:

BLOGGS Joseph 1st January 1984 (as an integer in Microsoft Excel = 30682) BLOGGSJos30682

Although there will inevitably be limitations to this approach, particularly where typographical errors have been introduced in the recording of name information, the overall impression is that the personal data recorded in CRIS are substantially accurate. One notable individual appears at least 10 times in the 15 months of data – each time with the exact same surname, forename and date of birth.

2.3 Data Analysis The Pivot Table function in Microsoft Excel was used to tabulate the 48,805 rows of data. However, not all of the rows initially extracted were included for analysis, and the Pivot Table functionality was used to screen out certain rows of data. Specifically, the following were excluded from the analysis: -

1,713 rows of data with the following proceedings codes: o o o o

-

XE – Entry Cancelled (e.g. entered in error) XF – Accused Screen Created in Error XT – Crime Transferred to Another Force ‘Blank’ – no proceedings code entered.

A further 126 individuals with an age recorded as less than 10 years or greater than 69 years old.

The exclusion of certain ages was necessary because of a combination of two related factors: firstly, policies relating to the policing of cannabis differ depending on whether the individual found in possession of cannabis is an adult (aged 18 years plus) or a youth (aged 17 years or less); secondly, the extreme age groups (0 to 9 and 70 plus) were affected by a relatively high incidence of errors in the recorded ages, most notably 107 one year olds and seven individuals aged over 100. These exclusions collectively brought the total number of rows of data down to 46,966.

9


3. Findings This section will present the analysis in four parts: 3.1

An overview of historical trend data (2000–2006).

3.2

An overview of the 46,966 records analysed for the period October 2005 to December 2006, with reference to offence characteristics (e.g. location, time), offender demographics and policing.

3.3

A detailed consideration of the degree to which the recorded crime data indicates that the policing of cannabis possession impacts equally (or otherwise) on different ethnic groups. This will focus on males accused during 2006. Males comprised 95 per cent of all individuals accused of cannabis possession that year, and the single year focus allows the use of GLA demographic data for 2006 as the denominator in calculating offence rates.

3.4

An in-depth examination of a sample of 144 records from 2006, seeking to understand the factors underlying ethnic disproportionality.

3.1 Historical trends 2000 – 2006 The number of cannabis offences recorded by the MPS has risen steeply since May 2004 and the total for 2006 was 2.5 times higher than that for 2000 (see Figure 3). This is against a background of falling cannabis use in England and Wales (see section 1.2 above). Figure 3: Total recorded cannabis possession offences in the MPS, January 2000 – December 2006 (source: MPS Performance Information Bureau, 28th March 2007). Total Recorded Monthly Cannabis Possession Offences in the MPS January 2000 - December 2006 Total n=172,983

4500 Average in 2006: 3,448 offences per month

4000

Number of offences per month

3500 Seven year trend line 3000

2500

Average over full seven years: 2,059 offences per month

2000

1500

1000 Street crime initiative launched April 2002

0

Cannabis reclassified January 2004

Cannabis warnings counted as sanctioned detections from April 2004

Jan-00 Feb-00 Mar-00 Apr-00 May-00 Jun-00 Jul-00 Aug-00 Sep-00 Oct-00 Nov-00 Dec-00 Jan-01 Feb-01 Mar-01 Apr-01 May-01 Jun-01 Jul-01 Aug-01 Sep-01 Oct-01 Nov-01 Dec-01 Jan-02 Feb-02 Mar-02 Apr-02 May-02 Jun-02 Jul-02 Aug-02 Sep-02 Oct-02 Nov-02 Dec-02 Jan-03 Feb-03 Mar-03 Apr-03 May-03 Jun-03 Jul-03 Aug-03 Sep-03 Oct-03 Nov-03 Dec-03 Jan-04 Feb-04 Mar-04 Apr-04 May-04 Jun-04 Jul-04 Aug-04 Sep-04 Oct-04 Nov-04 Dec-04 Jan-05 Feb-05 Mar-05 Apr-05 May-05 Jun-05 Jul-05 Aug-05 Sep-05 Oct-05 Nov-05 Dec-05 Jan-06 Feb-06 Mar-06 Apr-06 May-06 Jun-06 Jul-06 Aug-06 Sep-06 Oct-06 Nov-06 Dec-06

500

2000 (n=16,323)

2001 (n=16,374)

2002 (n=23,389)

2003 (n=23,803)

2004 (n=21,609)

2005 (n=30,112)

2006 (n=41,373)

Three distinct periods can be seen in the historical trend data. (i)

For the first two years, the number of cannabis possession offences remained fairly constant at around 1,400 offences per month.

(ii)

A sharp rise is then observed in early 2002 to around 2,000 offences recorded per month, coinciding with street crime being identified – initially by the media, and then by the police and government – as a national priority, and the subsequent introduction of the Street Crime Initiative. This eventually tailed off in early 2004 following the reclassification of cannabis, with a fall back to 2000/01 levels. It is believed that this fall is related to the introduction of cannabis warnings; procedures took some time to become established and a period of uncertainty is thought (by the MPS Drugs Directorate) to have led to a fall in the number of offences recorded.

10

NOT PROTECTIVELY MARKED (iii) Finally, a rapid and sustained rise in offences is observed from May 2004 onwards, peaking at almost 4,000 offences per month in November 2006. Of particular note, the decision was taken to count cannabis warnings as sanctioned detections from April 2004. The full reasons behind the rise since early 2004 are not known, but are likely to have arisen from a combination of factors, notably changes in police activity; it is also possible that other factors may have had an impact, for example reclassification having lead to a change in public behaviour, although it is worth repeating that cannabis use was falling at the time – and continues to fall – and that the number of offences has continued to rise since reclassification took place. In relation to policing, two closely related factors are likely to have been important: firstly, the introduction of cannabis warnings will have freed up police time to detect more cannabis offences (whether intentionally, or as a by-product of other policing activities); secondly, a combination of offences brought to justice (OBTJ) performance targets and the decision that cannabis warnings could be counted as sanctioned detections (SDs), which have seen cannabis warnings driving sanctioned detection performance improvements. Between 2005/06 and 2006/07, for example, the increasing number of cannabis warnings issued by the MPS accounted for 34 per cent of the increase in the number of offences brought to justice – see Figure 4. In 2006/07, the contribution of cannabis warnings to total sanctioned detections at a borough level ranged from 25.7 per cent in Wandsworth to 4.9 per cent in Sutton; the MPS average was 15.6 per cent (see Appendix B for fuller details). 12 Figure 4: Examining the contribution of cannabis warnings to OBTJ performance (April 2005 to April 2007) Cannabis Warnings and Total Offences Brought to Justice - Apr05 to Apr07 Source: PIB Report 17th May 2007 "Offences Brought to Justice (OBTJ) - April 2007 " (NB: Data for Mar07 and Apr07 are estimates ) 21.0

350.0 Total OBTJ indexed to Apr05 Cannabis warnings indexed to Apr05

18.0

OBTJs excluding cannabis warnings indexed to Apr 05 Cannabis warnings as a % of total OBTJs

15.0

250.0

April 2007 vs April 2005 Total OBTJs: Up 22% Cannabis warnings: Up 193% OBTJs minus cannabis warnings: Up 7%

200.0

12.0

9.0

150.0

6.0

100.0 2005/06 vs 2006/07 Total OBTJs: Up 12.8% Cannabis warnings: Up 44.4% OBTJs minus cannabis warnings: Up 8.4% Contribution of cannabis warnings to OBTJ improvement: 34.1%

50.0

Cannabis warnings as a % of total OBTJs

OBTJs and cannabis warnings indexed to Apr05

300.0

3.0

0.0 Apr-07

Mar-07

Jan-07

Feb-07

Dec-06

Oct-06

Nov-06

Sep-06

Jul-06

Aug-06

Jun-06

Apr-06

May-06

Mar-06

Jan-06

Feb-06

Dec-05

Oct-05

Nov-05

Sep-05

Jul-05

Aug-05

Jun-05

Apr-05

May-05

0.0

Month

12

Source: MPS Performance Information Bureau Redacted - 3 Report 20th April 2007 “Monthly Offences and Detections Ordered by % Detections”.

11


3.2 Cannabis possession accused October 2005 – December 2006: overview The 46,966 records are considered under the following categories:

•

Locations

•

Temporal patterns

•

Repeat offenders

It was originally hoped that it would also be possible to examine the concentration of offences at an individual officer level, however this proved impossible. Most notably, the use in some boroughs of Telephone Crime Recording Units (TCRUs) or Telephone Investigation Bureaus (TIBs) to create CRIS records on behalf of front-line officers has served to obscure (at a strategic level) the details of the officer(s) who originally discovered the cannabis possession offences. In such cases, although this information was recorded in the free-text Details (‘Dets’), the ‘Initial Investigating Officer’ was recorded as being the TCRU/TIB data in-putter rather than the front-line officer. In Brent, for example, 1,327 out of 3,627 (36.6 per cent) cannabis records created during the period October 2005 to December 2006 have the same ‘Initial Investigating Officer’.

3.2.1

Locations

Cannabis possession offences are not evenly distributed across London, but are heavily concentrated in certain locations, most notably Westminster (9.9 per cent) and Brent (7.7 per cent) – see Figure 5. This is important because: (i)

The boroughs recording the greatest number of offences will have a disproportionate impact on London as a whole; and

(ii)

Different boroughs have different demographic profiles (ethnicity, age profiles).

The prominence of Westminster is perhaps unsurprising given the levels of front-line police resources in that borough. On initial inspection, Brent would appear to be more anomalous, an observation that is reinforced by other evidence uncovered during this research (e.g. the contribution of cannabis warnings to overall sanctioned detections, individual offenders repeatedly warned). Figure 5: Cannabis Possession Accused – By Owning Borough (October 2005 – December 2006) Cannabis Possession Accused - By Owning Borough Recorded Date Oct05 - Dec06, Ages 10-69 Total n=46,966 5000 4500 4000

3000 2500 2000 1500 1000 500

Heathrow Airport

Merton

Richmond upon Thames

Kingston upon Thames

Sutton

Havering

Bexley

Harrow

Barnet

Barking and Dagenham

Bromley

Hounslow

Redbridge

Kensington and Chelsea

Greenwich

Hammersmith and Fulham

Islington

12

Croydon

Lewisham

Tower Hamlets

Camden

Hillingdon

Hackney

Newham

Waltham Forest

Ealing

Wandsworth

Enfield

Haringay

Southwark

Brent

Lambeth

0 Westminster

Number Accused

3500

NOT PROTECTIVELY MARKED At a borough level it is possible to examine the proportion of individuals found in possession of cannabis who were from the borough in question by using a combination of the ‘owning borough’ and ‘home intelligence unit’ data recorded in CRIS (see Table 3). This is likely to be informative, for example, in relation to the nature of the local cannabis ‘problem’ and the individuals being targeted by the police. In relation to the latter, it can be seen that Brent’s offenders are almost all from Brent (73.4 per cent), in contrast to boroughs such as Westminster (26.1 per cent) and Kensington and Chelsea (38.1 per cent). This data is also of potential value in terms of understanding the impact of the policing of cannabis on local communities, particularly where it can be triangulated with other sources. For example, community tensions have previously been identified in Brent around the use of Stop and Search, and this analysis would seem to support the assertion that the policing of cannabis is both disproportionate relative to most other boroughs (in absolute terms), and also heavily falls on local residents. This type of analysis could usefully inform local discussions about policing priorities and tactics – for example in the context of Independent Advisory Groups (IAGs) or Police Community Consultative Groups (PCCGs). Table 3: Proportion of individuals found in possession of cannabis who were from the borough in question (recorded date Oct05 – Dec06, ages 10 to 69 years old) Owning Borough Brent Havering Merton Hackney Enfield Harrow Hounslow Waltham Forest Hillingdon Greenwich Barnet Croydon Newham Barking and Dagenham Ealing Tower Hamlets Haringay Bromley Lewisham Redbridge Hammersmith and Fulham Southwark Lambeth Bexley Wandsworth Sutton Kingston upon Thames Richmond upon Thames Islington Camden Kensington and Chelsea Westminster Heathrow Airport Total

Total Offences 3627 462 403 1893 2104 672 968 1950 1616 1244 804 1357 1735 711 2097 1538 2288 813 1415 1012 1140 2146 2387 494 1971 478 423 332 1374 1684 1131 4669 28 46966

% offenders from borough 73.4 71.6 70.0 68.5 68.3 67.4 67.4 67.0 66.2 65.5 65.3 65.1 63.2 62.2 62.0 61.9 61.1 60.6 58.7 56.2 56.0 55.7 54.5 54.3 53.6 53.3 49.9 47.6 43.2 38.7 38.1 26.1 3.6 Average: 57.0

At the electoral ward level it can be seen that the locations where cannabis possession offences are recorded are even more spatially concentrated than at borough level. Only 7 out of the 624 wards in London (1.1 per cent) account for 10.1 per cent of all cannabis possession offences and 38 wards (6.1 per cent) account for 25.3 per cent of all offences (Figure 6). Again it is important to note that these high incidence locations will have a disproportionate impact on the London-wide picture and in all likelihood will have demographic characteristics that differ from London viewed as a whole.

13

NOT PROTECTIVELY MARKED Figure 6: Cannabis Possession Accused – Cumulative Percentage By Electoral Ward Cannabis Possession Accused - Cumulative Percentage By Electoral Ward Recorded Date Oct05 - Dec06, Ages 10-69 Total n=46,966 Geocoded Data: 3,224 (6.9%) accused records ward unknown 100 95

Wards: London has a total of 624 electoral wards. In the present data, cannabis accused were recorded in 611 of these.

90 85 80 75 Cumulative Percentage

70 Concentration of offences: - 10% of cannabis accused recorded in only 7 wards (1.1%) - 25% of cannabis accused recorded in 38 wards (6.1%) - 50% of cannabis accused recorded in 125 wards (20.0%) - 75% of cannabis accused recorded in 287 wards (46.0%) Top 7 wards: 1. West End (Westminster) - 1,216 accused 2. St. James's (Westminster) - 996 3. Camden Town with Primrose Hill (Camden) - 697 4. Harlesden (Brent) - 517 5. Tottenham Green (Haringey) - 484 6. Stonebridge (Brent) - 452 7. Coldharbour (Lambeth) - 385

65 60 55 50 45 40 35 30 25 20 15 10 5

609

593

577

561

545

529

513

497

481

465

449

433

417

401

385

369

353

337

321

305

289

273

257

241

225

209

193

177

161

145

129

97

113

81

65

49

33

1

17

0

Wards in rank order

It is notable that within the top 7 wards, the top three are in major visitor locations and the next four are in areas with large and longstanding black communities. Furthermore, at least two (Camden Town and Coldharbour) are well known for their ‘open’ cannabis markets. It is likely that this clustering of offences goes some way to explain the over-representation of black individuals that will be highlighted later in the analysis. Further qualitative work might be merited to examine in greater detail the characteristics of these locations and the policing processes that have resulted in this marked concentration of offences. Robbery hypothesis One hypothesis examined in the course of this research is that high levels of cannabis offences may simply reflect high levels of policing targeting other crime problems – most notably robbery of personal property, which has been an organisational priority for the MPS for a number of years. An initial investigation suggests that there is a correlation between cannabis possession and personal robbery rates at a ward level (Rsq=0.44) – see Figure 7. Figure 7: Examining the relationship between personal robbery and cannabis possession offences at a ward level (personal robbery data source: MetStats, data extracted 22 May 2007) Examining the relationship between personal robbery and cannabis possession offences Ward level - MPS Oct05 to Dec06 Personal robbery n=52,936; cannabis accused n=43,749 (nb: 3,217 cannabis offences - ward unknown) 1400

Number Cannabis Possession Accused

1200

1000

800

600 y = 0.9431x - 9.6908 2 R = 0.4355 400

200

0 0

50

100

150

200

250

Number Personal Robbery Offences

14

300

350

400

NOT PROTECTIVELY MARKED Whilst this correlation is not an indication of direct causation, it does at least suggest that areas with high robbery rates are typically also areas where cannabis possession is detected more often. It is likely that this is at least partly a consequence of the policing tactics used to tackle personal robbery – most notably stop and search/account. This point will be reflected upon in the conclusions and discussion. Interestingly, it can be seen that the seven wards that account for 10.1 per cent of all cannabis possession offences (Figure 6) appear in Figure 7 as outliers above the best-fit line. This might suggest that in those locations the styles and intensity of policing are not related to robbery, but to other characteristics of those wards. Alternatively, it is possible that the relationship between personal robbery offences and cannabis possession offences is not linear, but perhaps curvilinear or even exponential. This suggests a hypothesis about the relationship between robbery rates and policing intensity, for example that the policing resources directed to tackle robbery increase as a multiple of the robbery rate (high robbery locations have much higher levels of policing resources). Certainly this appears a possibility: conventional problem-solving approaches to policing, at least in London, do tend to focus particularly on ‘hot spot’ locations.

3.2.2

Temporal Patterns

The first thing to note is that the number of offences recorded by the police has increased over the duration of the data under consideration, although this has been subject to considerable ‘noise’ (Figure 8). Figure 8: Cannabis Possession Accused – Recorded Date (with trend line & moving average) Cannabis Possession Accused - Recorded Date Recorded Date Oct05 - Dec06, Ages 10-69 n=46,966 200 180

Number Accused per Day

160 140 120 100 80 60 40

Total 7-day moving average Linear (Total)

20

23/12/2006

09/12/2006

25/11/2006

11/11/2006

28/10/2006

14/10/2006

30/09/2006

16/09/2006

02/09/2006

19/08/2006

05/08/2006

22/07/2006

08/07/2006

24/06/2006

10/06/2006

27/05/2006

13/05/2006

29/04/2006

15/04/2006

01/04/2006

18/03/2006

04/03/2006

18/02/2006

04/02/2006

21/01/2006

07/01/2006

24/12/2005

10/12/2005

26/11/2005

12/11/2005

29/10/2005

15/10/2005

01/10/2005

0

Recorded Date

Cannabis possession offences are typically detected in the context of police activity in public locations, most notably the use of stop and search powers. The profile of the public ‘available’ to policing will differ depending on the time or day and day of week, as will both police resources and the competing demands on those police resources. It is therefore of interest that the number of cannabis possession offences recorded by the police peaks in the middle of the week and at 4pm and 8pm (Figures 9 and 10). Analysis previously conducted within the MPS (September 2005) found that overall frontline policing was overresourced mid-week – that is, the percentage of weekly police resources exceeded the percentage of weekly crime – and under-resourced at the weekend (internal MPS analysis conducted by SM&PD, September 2005). The 4pm and 8pm peaks may well reflect peaks in the street population of individuals in their late-teens and early-20s (the peak ages for cannabis possession offences), reflecting both after school/college and after dinner social patterns. It is also possible that they may relate to police behaviour relative to shift patterns, although further work would be required to explore this in more detail.

15

NOT PROTECTIVELY MARKED Figure 9: Cannabis Possession Accused – By Committed On/From Day of Week Cannabis Possession Accused - Committed On/From Day of Wk Recorded Date Oct05 - Dec06, Ages 10-69 Total n=46,966

9000

8000

Number Accused

7000

6000

5000

4000

3000

2000

1000

0

MON

TUE

WED

THU

FRI

SAT

SUN

Figure 10: Cannabis Possession Accused – By Committed On/From Time Cannabis Possession Accused - By Committed On/From Time Recorded Date Oct05 - Dec06, Ages 10-69 n=46,966 (inc. 8 where time was blank)

4000 3500

Number Accused

3000 2500 2000 1500 1000 500 0 0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

Time (Hour)

At the weekend a much greater proportion of cannabis possession offences are discovered in the early hours of the morning than is the case on other days of the week (Figure 11), which is likely to reflect a larger night time population commensurate with recreational lifestyles and the night time economy. It is also interesting to note that the mid-afternoon peak commented on earlier is observed on both Saturdays (at 3pm) and Sundays (at 4pm) as well as during the week, suggesting that this may not simply be an ‘after school’ effect.

16

NOT PROTECTIVELY MARKED Figure 11: Cannabis Possession Accused – By Committed On/From Day and Time Cannabis Possession Accused - By Committed On/From Day and Time Recorded Date Oct05 - Dec06, Ages 10-69 n=46,966 (inc. 8 with no day or time recorded) 10.0 9.0

Percentage of Day of Week

8.0 7.0 6.0 5.0 MON TUE WED THU FRI SAT SUN Total

4.0 3.0 2.0 1.0 0.0 0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

Time (Hour)

Finally, it is interesting to note that the offence time does not differ substantially between youths and adults, with the exceptions that a greater percentage of adult offences are detected in the early hours of the morning and a greater percentage of youth offences are detected in the afternoon, notably between 3pm and 5pm (Figure 12). Taken together, this analysis suggests that the temporal distribution of offences may be significantly influenced by the rate and styles of policing. Figure 12: Cannabis Possession Accused – By Committed On/From Time and Adult/Youth Cannabis Possession Accused - Committed On/From Time by Adult/Youth Recorded Date Oct05 - Dec06, Ages 10-69 n=46,966 (inc. 8 with no time recorded) 10.0 9.0 8.0

% of Age Group

7.0 6.0 5.0 Youths 4.0 Adults 3.0 2.0 1.0 0.0 0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

Time (Hour)

3.2.3

Repeat offenders

Finally, in this overview, it is worth noting that offences are not simply concentrated in time and place, but also at an individual level. In the 15 months of data being analysed in this section, the 46,966 offences were committed by a maximum of 41,426 individuals – at an average of 1.13 offences per offender. Table 4 illustrates the distribution of repeat offenders, with 92 individuals accused of cannabis possession at least 5 times in 15 months.

17

NOT PROTECTIVELY MARKED Table 4: Repeat cannabis possession offenders, Oct05 – Dec06 13 Total number of offences 10 9 8 7 6 5 4 3 2 1 Total

Number of offenders 1 1 7 9 15 59 180 674 3221 37259 41426

Following this, on a more detailed examination (allowing for spelling mistakes and other data recording problems in CRIS), the most prolific offender appears in CRIS accused of cannabis possession no less than 13 times in 15 months – all in his home borough. It is notable that he was only charged on two occasions, having received 11 cannabis warnings (followed by a 12th in February 2007). 14 At the very least, this suggests that the checks required to establish whether a cannabis warning should not be issued either did not take place or were ineffectual.

13 Unfortunately, it proved impossible to effectively repeat this analysis broken down by self-defined ethnicity (SDE), due to the fact that some individual offenders have more than one SDE recorded. 14 Indeed, the records relating to this individual are informative in a range of ways, not least the unreliability of some fields in CRIS. An example would be the ‘PNC Result’ and ‘CRO Result’ fields, which the author has seen used elsewhere as a proxy for ‘existing criminal record’. These remain blank in all of the 13 records mentioned here – including the two occasions when the individual in question was charged. A separate charge for possession with intent to supply cannabis does, however, provide his full details and confirms that he has both CRO and PNC numbers.

18


3.3 Offender demographics January – December 2006

3.3.1

•

Introduction

•

Overview

•

Age Profile

•

Proceedings

•

Identifying where disproportionality arises

Introduction

The CRIS database records data on the age, gender and ethnicity of offenders, with the latter recorded using both police-defined Ethnic Appearance (EA) codes and the Census 16+1 Self-Defined Ethnicity (SDE) codes. The SDE codes will be used throughout this analysis in order to allow a direct and meaningful comparison with population demographic data for London. 15 In order to permit direct reference to Greater London Authority (GLA) population estimates for 2006 (based on projections from the 2001 Census), this section will focus on 38,442 cannabis possession accused records recorded between January and December 2006 (inclusive) where the accused was aged between 10 and 69 years old. Introduction to GLA data The Greater London Authority (GLA) Data Management and Analysis Group (DMAG) produces population projections for London. The data used in the present analysis were obtained directly from the GLA in March 2007 and are the 2006 figures for males in Greater London from the GLA 2006 Round Ethnic Group Population Projection (EGPP) figures by single years of age from 2001 to 2026 [RLP High] (Greater London Authority, 2006). As with the data produced in previous Rounds, these data are projections (and therefore estimates) that have been ‘demographically modelled using 2001 Census data’ (Bains and Klodawski, 2006: 1). That is to say that ‘demographic trends present as at 2001’ (ibid.) have been applied to the 2001 Census data to make allowances for migration, births and deaths (ibid.: Appendix 2, pp29-33). In addition, adjustments have been made to (i) incorporate the 2006 mid-year population estimate for London produced by the Office for National Statistics (ONS), and (ii) accommodate a known under-count of young males in the 2001 Census data (personal communication with Redacted – 4 ). It must be noted that these data are subject to a number of caveats and should not be treated uncritically. Arguably, however, they are the best available. One specific acknowledged limitation of the data worth mentioning is that ‘more recent migration flows such as those resulting from the expansion of the European Union since 2004 are not reflected in the modelling methodology’ (Bains and Klodawski, 2006: 3), which means in particular that the estimate of the white population in London in 2006 is likely to represent an undercount. A possible undercount of black males is discussed below. More generally, it should be said that as the data are examined at a finer level of granularity (that is, in more detail) the confidence intervals get wider; that is, users can be less confident in the accuracy of the data. Finally, it should be noted that the GLA data use ethnic categories that in some cases aggregate categories used in the 2001 Census (see Appendix C). Bearing in mind that the projections include data on a borough-by-borough basis, this aggregation is intended to avoid difficulties arising from very small populations (for more detail see Bains and Klodawski, 2006).

15

This research has treated the self-defined ethnicity data relatively uncritically and a separate piece of work would be required to establish the degree to which self-defined ethnicity data are actually self-defined and therefore reliable. It is nevertheless worth commenting that the data under consideration include individuals found in possession of cannabis on more than one occasion who have more than one ‘self-defined’ ethnicity. For example, one individual who appears 9 times is recorded as black Caribbean 5 times, ‘any other black background’ twice, and mixed white and black Caribbean and black African each once.

19

NOT PROTECTIVELY MARKED GLA data: possible undercount of black Caribbean males? It is worth commenting briefly on one feature of the demographic data for London that may have a bearing on the analysis presented above and later in this report, namely the apparent undercount of black males, most notably in relation to London’s black Caribbean population. Whereas the 2006 GLA estimates for Greater London, based on projections from the 2001 Census, indicate that the white population in 2006 had a male-to-female ratio of 49:51, for black Caribbeans the ratio is 44:56 (see Figure 13 below); for reference the black African ratio is 48:52. If this anomaly, reflected in the gap between the male and female lines in Figure 13 that opens in the early-20s, does represent a genuine undercount, it will impact on any calculations examining rates per 1,000 population and therefore disproportionality, with the effect that rates relating to black Caribbean males in particular would be overstated. It should be said, however, that it appears that taking this phenomenon into account would not make much of an impact on the peak offence age (21 years – see section 3.3.3 below), but rather would impact on ages from the mid-20s and upwards. Even then, however, it would be unlikely to do more than somewhat narrow the gap between black Caribbean males and their white counterparts evident in the analysis below. In any case, this apparent anomaly highlights the complexity of using demographic data and thinking about offending rates by ethnicity, and suggests that the demographic data should not be treated uncritically. Figure 13: Examining the gender balance – white and black Caribbean populations, London 2006 Examining the gender balance - white and black Caribbean populations, London 2006 Source: GLA 6000

70000

60000

Number white

50000

4000

40000 3000 30000 2000

Number black Caribbean

5000 White Females White Males Black Caribbean Females Black Caribbean Males

20000

1000

Gender balance: White: 50.9% female, 49.1% male Black Caribbean: 55.6% female, 44.4% male

10000

0

0 0

5

10

15

20

25

30

35

40

45

50

55

60

65

70

75

80

85

90+

Age

3.3.2

Overview

From Table 5 it can be seen that overall 94.5 per cent of the individuals accused of cannabis possession were male, and 8.4 per cent were youths (10.7 per cent of females and 8.3 per cent of males). Although this research is concerned specifically with ethnic disproportionality, the significant gender disproportionality evident in the data – a 19:1 male-to-female ratio – at the very least merits further consideration. For example, according the 2005/06 British Crime Survey data presented above in section 1.2.1 the male-to-female ratio for self-reported cannabis use in the 16-24 year old group was only 1.9:1 for usage ‘in the last month’, 1.6:1 for use ‘in the last year’ and 1.3:1 for use ‘ever’. Although these data do not provide insights into matters such as the location of drug consumption and possession behaviours, they do nevertheless suggest that the 19:1 ratio evident in the police data indicates an over-representation of males.

20

NOT PROTECTIVELY MARKED Further examination of gender disproportionality would be a matter for further research. The analysis below will focus on male offenders found in possession of cannabis during 2006 and use age- and ethnicspecific population estimates relating to the number of males to calculate offence and offender rates. In terms of the ethnic profile of male offenders in 2006, 29.8 per cent were white British, 17.3 per cent black Caribbean and 10.1 per cent ‘any other’ white background. It is worth noting that the SDE of the offender was not recorded in 6.5 per cent of cases (n=2,483), most commonly where the offender refused to define their ethnicity (5.4 per cent). The profile of individuals whose SDE was not recorded appears to differ somewhat from the overall profile of individuals whose SDE was recorded, which suggests that non-random effects may be involved: 34.7 per cent were recorded with the Ethnic Appearance code ‘EA3: Black’, 20.2 per cent ‘EA1: White, North European’, 15.4 per cent ‘EA2: White, South European’, and 15.3 per cent ‘EA4: Asian’. Table 5: Cannabis Possession Accused By Gender, Age Group and Self-defined Ethnicity – Recorded date January to December 2006

Self Defined Ethnicity Any other ethnic group Asian - Any other background Asian - Bangladeshi Asian - Indian Asian - Pakistani Black - African Black - Any other black background Black - Caribbean Chinese or other - Chinese Mixed - Any other mixed background Mixed - White & Asian Mixed - White and black African Mixed - White and black Caribbean Officer Urgently required elsewhere Person declines to define ethnicity Person does not understand Situation involving public disorder White - Any other white background White - British White - Irish Grand Total % of Gender % of All Accused

Female Adult (18 Youth (10 to 69) to 17) 33 4 41 2 10 2 28 20 2 70 5 149 16 274 27 2 29 4 8 1 1 6 35 20 4 87 5 18 2 1 245 9 762 116 27 1 1844 222 89.3 10.7 4.8 0.6

Total 37 43 12 28 22 75 165 301 2 33 9 7 55 4 92 20 1 254 878 28 2066 100.0 5.4

Male Adult (18 Youth (10 to 69) to 17) 911 68 1551 98 837 84 1311 76 832 50 2514 285 2738 217 5819 534 115 12 407 43 120 28 109 24 464 147 28 3 1904 85 307 9 14 1 3492 115 9452 1106 388 34 33313 3019 91.7 8.3 86.7 7.9

Total 979 1649 921 1387 882 2799 2955 6353 127 450 148 133 611 31 1989 316 15 3607 10558 422 36332 100.0 94.5

Unknown Adult (18 Youth (10 to 69) to 17) 1 2 3 1 1 2 8

1 13

3 1 1 2 8

1

3 8 42 95.5 0.1

Grand Total 1017 1694 933 1418 905 2875 3122 6662 129 483 157 140 666 36 2095 336 16 3864 11444 450 38442

Total 1 2

1 14

3 8 2 4.5 0.0

44 100.0 0.1

% of All Accused 2.6 4.4 2.4 3.7 2.4 7.5 8.1 17.3 0.3 1.3 0.4 0.4 1.7 0.1 5.4 0.9 0.0 10.1 29.8 1.2 100.0

100.0

A closer examination of males aged 10 to 69 accused of cannabis possession in 2006 (Tables 6 to 8) highlights the over-representation of all of the black and minority ethnic groups, with the exception of Chinese and Indian individuals when all ages are examined, Chinese individuals in relation to the adult offenders, and Chinese, Indian and Pakistani individuals in relation to youths. It should be noted that the rates per 1,000 population are specific to the age groups being considered and use GLA population data for males in 2006 as the denominator. 16 Table 6: Offence and offender rates by Self-Defined Ethnicity (all males, 2006) Offence rates

All (10-69) 2006 Self-defined ethnicity Bangladeshi Black African Black Caribbean Black other Chinese Indian Other Other Asian Pakistani Unknown White Total

Total number Rate per 1,000 accused Population 921 2799 6353 3699 127 1387 1429 1797 882 2351 14587 36332

Multiple of white rate

14.7 16.9 50.7 63.2 3.1 6.7 16.1 19.3 13.0 n/a 7.1 12.3

2.1 2.4 7.1 8.9 0.4 0.9 2.3 2.7 1.8 n/a 1.0 1.7

16

Offender rates Total number of unique Rate per 1,000 individuals Population accused 779 12.5 2427 14.7 5440 43.4 3212 54.9 121 3.0 1178 5.7 1258 14.2 1597 17.1 745 11.0 2103 n/a 13363 6.5 32223 10.9


Average offences per offender

1.9 2.2 6.6 8.4 0.5 0.9 2.2 2.6 1.7 n/a 1.0 1.7

1.18 1.15 1.17 1.15 1.05 1.18 1.14 1.13 1.18 1.12 1.09 1.13

A methodological point is that ordinarily calculations of offence/offender rates and disproportionality do not distinguish between males and females, resulting in population averages. As has been shown, cannabis possession offences are overwhelmingly (95 per cent) male. Narrowing the focus to males only is likely to result in offence/offender rates almost twice as high as for total populations, although it should not impact significantly on disproportionality ratios. This has important implications in terms of understanding, for example, rates of police contact with male members of the public (e.g. as suspects).

21

NOT PROTECTIVELY MARKED Table 7: Offence and offender rates by Self-Defined Ethnicity (adult males, 2006) Offence rates

Male adults (18-69) 2006 Self-defined ethnicity Bangladeshi Black African Black Caribbean Black other Chinese Indian Other Other Asian Pakistani Unknown White Total


17.1 18.8 55.2 82.6 3.1 7.2 17.4 21.0 14.8 n/a 7.2 12.8

Multiple of white rate 2.4 2.6 7.6 11.4 0.4 1.0 2.4 2.9 2.0 n/a 1.0 1.8




2.2 2.5 7.1 10.8 0.5 0.9 2.3 2.8 1.9 n/a 1.0 1.7

1.19 1.15 1.17 1.16 1.03 1.18 1.14 1.13 1.18 1.12 1.09 1.13



1.0 1.3 4.1 3.3 0.4 0.5 1.3 1.4 0.6 n/a 1.0 1.3

1.12 1.18 1.13 1.10 1.33 1.15 1.08 1.10 1.32 1.14 1.08 1.11

Table 8: Offence and offender rates by Self-Defined Ethnicity (male youths, 2006) Offence rates

Male youths (10-17) 2006 Self-defined ethnicity Bangladeshi Black African Black Caribbean Black other Chinese Indian Other Other Asian Pakistani Unknown White Total



6.3 9.1 27.1 21.1 3.2 3.1 8.5 9.1 4.3 n/a 6.3 8.7

1.0 1.4 4.3 3.3 0.5 0.5 1.3 1.4 0.7 n/a 1.0 1.4


Overall, it can be seen that black Caribbean and black ‘other’ ethnic groups are most significantly overrepresented relative to their white counterparts. The offence rate for all black groups (ages 10-69) combined is 5.2 times that for their white counterparts. It is also worth commenting briefly on the final column, the average number of offences per offender. This is important because, where individuals are found in possession of cannabis on multiple occasions the level of sanction should be increased accordingly, and this may therefore provide some clues about differential sanctioning. When looking at all males accused of cannabis aged 10-69, it can be seen that only Chinese individuals have a lower average number of offences per offender than white individuals, while the highest rates are seen in relation to Bangladeshi, Indian and Pakistani individuals, followed by black Caribbeans. The pictures for adults and youths differ subtly, notably with adults committing a slightly higher average number of offences in 2006, but overall it does not appear that there are any particularly marked differences between ethnic groups.

3.3.3

Age profile

It can be seen from Figure 14 below that cannabis possession offences are heavily skewed towards the late-teens and early-20s, with a marked peak at 18 years old. Interestingly, this peak is more marked for female offenders than their male counterparts.

22

NOT PROTECTIVELY MARKED Figure 14: Cannabis Possession Accused by Age and Gender Cannabis Possession Accused - Age Profile By Gender Recorded Date Jan - Dec06, Ages 10-69 Total n=38,442 (Male=36,332; Female=2,066; U/K=44) 14.0

12.0

10.0

% of Gender

Female 8.0 Male

6.0

4.0

2.0

0.0 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69

Age

One very important implication of this age profile is that when considering offence rates by ethnicity it is imperative to understand the fact that different ethnic groups have very different age structures, and that as a consequence using total population figures to calculate offence and other rates may be particularly inappropriate. Figure 15 below illustrates this by demonstrating that there were important differences between the white and black male population age structures (ages 0 to 90+) in London in 2006. 17 These included the different ‘peaks’ (modes), the marked ‘dip’ in the black groups in the mid-20s and the fact that the black African and black other groups made up an increasing proportion of the London population for ages less than about 22 years old. To accommodate these differences, age specific offence rates will be stated where possible. 18 Figure 15: London Male Age Structure – White, black Caribbean, black African and ‘other’ black – GLA Population Estimate Data for 2006 (n.b: white scale on left-hand axis, black scale on right-hand axis) GLA Population Estimates - Greater London Males by Ethnicity 2006 6000

50000

5000 White Black Caribbean Black African

Number white males

40000

4000

Black Other

30000

3000

20000

2000

10000

1000

0

Number black Caribbean / African / 'other' males

60000

0 0

5

10

15

20

25

30

35

40

45

50

55

60

65

70

75

80

85

90+

Age

17

Black other includes the following ethnicities: ‘mixed: white and black Caribbean’, ‘mixed: white and black African’ and ‘black or black British: other black’. 18 Previous research on stop and search has discussed the importance of taking into account the demographic profile of the population available to be stopped and searched – i.e. the street population (e.g. see MVA and Miller, 2000). Such analysis presents significant methodological challenges and is beyond the scope of this pan-London study. However, the examination of age-specific offence rates should control for the street population to some extent, as one of the key influences on the street population is likely to be age.

23


Figure 16 below illustrates overall offence rates by age and self-defined ethnicity and highlights at least two significant areas for concern. Firstly, it can be seen that black Caribbean and black ‘other’ individuals are accused of cannabis possession at exceptionally high rates in their late-teens and early-20s. Secondly, the accused rate ‘differential’ (that is, the rate multiple) between black Caribbean and black ‘other’ individuals and their white counterparts rises to around 10-fold in the early-20s and more-or-less stays at that level thereafter, as is more clearly illustrated in Figure 17. Figure 16: Age specific offence rates by self-defined ethnicity (males, 2006) Male Cannabis Possession Accused Per 1,000 Male Population - by Self-Defined Ethnicity Recorded Date January - December 2006, Ages 10 to 69 Using GLA Aggregate Ethnic Group Data n=36,332

350

Bangladeshi Black African Black Caribbean Black Other Chinese Indian Other Other Asian Pakistani White Grand Total

Accused Per 1,000 Population

300

250

200

150

100

50

0 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69

Age

Figure 17: Age specific offence rates as a multiple of the white offence rate, by self-defined ethnicity (males, 2006) 19 Male Cannabis Possession Accused Per 1,000 Male Population - by Self-Defined Ethnicity Multiple of the Accused Rate for White Males Recorded Date January - December 2006, Ages 12 to 45 Using GLA Aggregate Ethnic Group Data n=35,419 (97.5% of total)

14 13

Multiple of White Accused Rate

12

Bangladeshi Black African Black Caribbean

11

Black Other

10

Chinese

9

Indian Other

8

Other Asian

7

Pakistani

6

White Grand Total

5 4 3 2 1 0 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 Age

Of particular note, the data indicate that 21-year old black Caribbean males were accused of cannabis possession in 2006 at a rate of 300 per 1,000; that is, for every 1,000 black Caribbean males aged 21 in 19 Note that only ages 12 to 45 are illustrated in Figure 17. This is for two reasons: (i) there were no white offenders aged 10 or 11; and (ii) above 45 years old the number of offenders is very small and the ratio data very volatile, which if included in the graph would have the effect of concealing the detail for the key age groups – the teens and twenties.

24

NOT PROTECTIVELY MARKED 2006, 300 offences were recorded as having been committed by a 21-year old black Caribbean male in that year. The raw data (which, it is worth stating, have been double-checked with the respective data owners and are subject to the caveats detailed above) indicate that in 2006 there were approximately 1,670 21 year old black Caribbean males in Greater London (GLA data); in the same year 503 offences of cannabis possession were committed by approximately 463 such individuals: an offender rate of 277 offenders per 1,000, and an offence rate of 301 offences per 1,000. 20 It should be noted that these are the rates for only one year, which suggests exceptionally high lifetime contact with policing – both in absolute and relative terms. This may represent a significant threat to police/community relations within these ethnic groups, and has the potential to significantly undermine police legitimacy within certain communities – and especially within particular cohorts within those communities (i.e. young black males). It should also be stated that these figures do only relate to incidents when cannabis was found, at least one implication of which is that overall levels of police contact, most notably in terms of stop and search/account, will in all likelihood be very much higher. 21 One potential caveat is that it is possible that there are systematic problems relating to the recording of the self-defined ethnicity data – especially if it is not in fact self-defined (note the comments in Footnote 15 above). This might, for example, result in individuals who are black African being recorded as black Caribbean, which would result in an over-count of the latter and an under-count of the former. Even if this were the case, however, the gap between black Caribbean and black ‘other’ males and their white counterparts would probably only be narrowed slightly by more accurate data recording. In any case, it may be appropriate for the MPS to consider commissioning a separate piece of work looking at the recording of self-defined ethnicity data to understand how reliable it is and therefore how robust are any conclusions based on analysing SDE data. This could consider a number of issues, such as the degree to which such data is actually self-defined and the challenges to policing presented by the self-definition process. Anecdotally, it has been suggested to the author that the process of officers asking suspects to define their ethnicity may at times aggravate already difficult situations and is therefore sometimes (perhaps often) overlooked. The implications for monitoring ethnic disproportionality should be obvious. Finally, using the data for black Caribbean males accused of cannabis possession in 2006 to illustrate, Figure 18 highlights the importance of considering age-specific offence rates and the impact of population age profiles. In this case, although the greatest number of cannabis possession offences (blue line) related to 18 year olds when viewed as a crude total (number accused), the peak age for offences stated as a rate per 1,000 population (red line) occurred at 21 years old. The difference is accounted for by the population age structure (grey line), in which 21 year olds make up a much smaller percentage of all black Caribbean males than their 18 year old counterparts.

20 In relation to the issue of a potential undercount of black Caribbean males discussed in section 3.3.1 above, it is worth noting that even if the undercount was as high as 50 per cent – that is there were actually 2,505 21 year old black Caribbean males in London in 2006 rather than 1,670 – the offence rate would still be 201 per 1,000 and the offender rate 185 per 1,000. In other words, almost one fifth of 21 year old black Caribbean males found in possession of cannabis in a single year. 21 For example, in the financial year April 2006 to March 2007, only around 11 per cent of stops and searches conducted by the MPS resulted in an arrest (37,359 out of 328,469), although the number of cannabis warnings issued where no arrest took place is unknown. During this period, 51 per cent of all stops and searches were for drugs. Note: these data exclude section 44 and section 60 stops and vehicle only searches. (MPS, 2007c)

25

NOT PROTECTIVELY MARKED Figure 18: Comparing numbers of offences (total numbers accused) and offence rates: black Caribbean males in 2006 Cannabis Possession Accused - Black Caribbean Males aged 10 to 69 in 2006 Number of Offences (n=6,353), Offence Rates (per 1,000 population) and Overall Age Profile (n=125,204) 800

4000

Peak number of offences at 18-years-old

3500

Number Offences / Offences Per 1,000 Population

Number of Offences 600

Offence Rate Per 1,000 Population

Peak offence rate at 21-years-old

3000

Number in London (GLA 2006)

500

2500

400

2000

Overall age profile of black Caribbean males in 2006 shows 3 distinct cohorts: peaks at 16, 41 and 68 yrs

300

1500

200

1000

100

500

0

Number Black Caribbean Males in London in 2006 (GLA data)

700

0 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 Age

3.3.4

Proceedings

The police have a range of options in terms of how to deal with (proceed against) individuals found illegally in possession of cannabis, although this differs between youths (aged 10-17) and adults (18+). There is a presumption against arrests in the case of adults, while the policy in relation to youths has recently changed: during the period covered by this analysis, there was a presumption that youths found in possession of cannabis would be arrested; recent changes to ACPO guidelines (January 2007) mean that this is no longer the case, and arrests must now be ‘necessary and justified based on individual circumstances’ (MPS, 2007b: para.8.1). Cannabis warnings can only be given to adults and should not normally be given to the same individual more than twice. Youths are dealt with under the provisions of s65 of the Crime and Disorder Act 1998 and can only be given a reprimand, a final warning or be charged. In the present analysis, the proceedings types have been summarised into four categories, namely caution, charge, cannabis warning and ‘other’ (see Appendix A for details). One of the key questions for this project concerns differential outcomes for different ethnic groups. Tables 9 (below), and 10, 11 and 12 (on page 28) illustrate a number of key areas for consideration: a. Disproportionality increases with the level of sanction. Comparing, for example, black Caribbean and white adult males accused of cannabis possession in 2006, it can be seen that the former are much less likely than the latter to be given a cannabis warning and much more likely to be charged (Table 9). In other words, disproportionality becomes more entrenched as the sanction increases. Table 9: Comparing proceedings rates for black Caribbean and white adult males accused of cannabis possession (2006) Proceeding Cannabis warning Caution Charge

Black Caribbean % of Rate per 1,000 accused population 73.9 40.8 10.5 5.8 14.0 7.7

% of accused 81.5 9.7 7.4

White Rate per 1,000 population 5.9 0.7 0.5

Ratio black Caribbean: white 6.9:1 8.3:1 14.4:1

b. Disproportionality is greater for adults than youths, and in relation to charging increases up until the late-20s. For example, the number of black Caribbean individuals charged was (proportionately) 11.4 times greater than whites at 18 to 21 years, 17.5 times greater at 22 to 25 years and 21.5 times greater at 26 to 29 years (Table 12).

26

NOT PROTECTIVELY MARKED It is worth stating that these issues are inherently related to each other due to the fact that the sanctions given to offenders are intended to increase with successive offences, and that these have a cumulative lifetime effect. In other words, according to the data, a young black man is more likely to be found in possession of cannabis in the first place, and also more likely to be found in possession of cannabis on subsequent occasions, which all else being equal means he is more likely than his white counterparts to receive a more punitive sanction. Other factors may also be important, for example the context of the discovery of cannabis, the offender’s willingness to admit responsibility, any previous criminal record (especially previous drugs offences), the quantity of drugs found (for example, where there is a suspicion of possession with intent to supply) and the nature of local crime problems (for example, local policies intended to tackle open drug dealing or use). These matters will be examined in section 3.4 below, and discussed in more detail in the conclusions. As a final couple of points, the data is informative about the degree to which standard procedures are followed. For example, it is interesting to note that 13.2 per cent of youths found in possession of cannabis in 2006 were given a cannabis warning, despite the fact that this is a proceeding only applicable to adult offenders. Elsewhere, analysis of the treatment of individual offenders has highlighted some notable departures from the standard operating procedures. For example, the individual who appears most often in the data under consideration has been accused of cannabis possession 13 times in fifteen months (October 2005 – December 2006) and received no less than 11 cannabis warnings – all in the same borough. 22

3.3.5

Identifying where disproportionality arises

It was originally envisaged that this research would pay particular attention to the decision as to whether an individual found in possession of cannabis was arrested, and subsequent proceedings decisions postarrest. This focus was intended to consider the factors underlying the apparent differential use of charges, cautions and cannabis warnings in the (relatively highly controlled) custody suite environment. To this end, the data were subdivided into mutually exclusive categories depending on whether the individuals accused of cannabis possession had been arrested or not, and then on how they were proceeded against. Further sub-divisions were incorporated breaking the data down further into age groups. As will become apparent from section 3.4, however, a subsequent in-depth examination of a sample of 144 CRIS records has highlighted systematic problems with the data on whether (how) individuals were arrested – most importantly in relation to the issuing of cannabis warnings. As a consequence, it is not presently possible to make any reliable statements about post-arrest proceedings decisions at a strategic or corporate level and therefore to know where in the case handling process disproportionality is arising. Consequently, the strategic analysis subdividing the data into those arrested and those not has been excluded from this report.

[Tables 10, 11 and 12 over page]

22 It may be of further interest to note that this same individual was stopped by the police at least 21 times between December 2005 and December 2006, including on four occasions (by four different officers) on the same day in July 2006.

27

NOT PROTECTIVELY MARKED Tables 10 and 11: Proceedings for Cannabis Possession, by Self-Defined Ethnicity: male adults (18-69yrs) and youths (10-17yrs) (2006) NB: RPKP = Rate Per 1,000 Population, using age-specific GLA population estimate data for males in Greater London in 2006. Male adults (18-69) 2006 Self-defined ethnicity Bangladeshi Black African Black Caribbean Black Other Chinese Indian Other Other Asian Pakistani Unknown White Total

Number 105 313 610 269 15 102 132 115 69 120 1291 3141

Cautioned % RPKP 12.5 2.1 12.5 2.3 10.5 5.8 8.1 6.7 13.0 0.4 7.8 0.6 10.0 1.7 6.9 1.4 8.3 1.2 5.3 n/a 9.7 0.7 9.4 1.2

Ratio:white 3.1 3.3 8.3 9.6 0.6 0.8 2.5 2.1 1.8 n/a 1.0 1.7

Number 129 373 814 379 13 80 121 87 80 124 991 3191

Charged % RPKP 15.4 2.6 14.8 2.8 14.0 7.7 11.4 9.5 11.3 0.3 6.1 0.4 9.2 1.6 5.2 1.1 9.6 1.4 5.5 n/a 7.4 0.5 9.6 1.2

Ratio:white 4.9 5.2 14.4 17.6 0.7 0.8 3.0 2.0 2.6 n/a 1.0 2.3

Number 591 1783 4298 2620 85 1118 1048 1450 676 1991 10861 26521

Cannabis Warning % RPKP 70.6 12.0 70.9 13.3 73.9 40.8 79.1 65.4 73.9 2.3 85.3 6.1 79.5 13.8 86.8 18.3 81.3 12.0 88.4 n/a 81.5 5.9 79.6 10.2

Ratio:white 2.0 2.3 6.9 11.1 0.4 1.0 2.3 3.1 2.0 n/a 1.0 1.7

Number 12 45 97 43 2 11 17 19 7 18 189 460

Other proceedings % RPKP 1.4 0.2 1.8 0.3 1.7 0.9 1.3 1.1 1.7 0.1 0.8 0.1 1.3 0.2 1.1 0.2 0.8 0.1 0.8 n/a 1.4 0.1 1.4 0.2

Ratio:white 2.4 3.3 9.0 10.5 0.5 0.6 2.2 2.3 1.2 n/a 1.0 1.7

Number 837 2514 5819 3311 115 1311 1318 1671 832 2253 13332 33313

Total RPKP 17.1 18.8 55.2 82.6 3.1 7.2 17.4 21.0 14.8 n/a 7.2 12.8

Ratio:white 2.4 2.6 7.6 11.4 0.4 1.0 2.4 2.9 2.0 n/a 1.0 1.8

Male youths (10-17) 2006 Self-defined ethnicity Bangladeshi Black African Black Caribbean Black Other Chinese Indian Other Other Asian Pakistani Unknown White Total

Number 38 86 188 121 5 32 45 50 22 36 554 1177

Cautioned % RPKP 45.2 2.8 30.2 2.8 35.2 9.5 31.2 6.6 41.7 1.3 42.1 1.3 40.5 3.4 39.7 3.6 44.0 1.9 36.7 n/a 44.1 2.8 39.0 3.4

Ratio:white 1.0 1.0 3.4 2.3 0.5 0.5 1.2 1.3 0.7 n/a 1.0 1.2

Number 31 149 217 161 7 20 35 44 12 19 400 1095

Charged % RPKP 36.9 2.3 52.3 4.8 40.6 11.0 41.5 8.7 58.3 1.9 26.3 0.8 31.5 2.7 34.9 3.2 24.0 1.0 19.4 n/a 31.9 2.0 36.3 3.1

Ratio:white 1.1 2.4 5.4 4.3 0.9 0.4 1.3 1.6 0.5 n/a 1.0 1.6

Number 7 25 52 62 0 12 14 19 11 25 170 397

Cannabis Warning % RPKP 8.3 0.5 8.8 0.8 9.7 2.6 16.0 3.4 0.0 0.0 15.8 0.5 12.6 1.1 15.1 1.4 22.0 1.0 25.5 n/a 13.5 0.9 13.2 1.1

Ratio:white 0.6 0.9 3.1 3.9 0.0 0.6 1.2 1.6 1.1 n/a 1.0 1.3

Number 8 25 77 44 0 12 17 13 5 18 131 350

Other proceedings % RPKP 9.5 0.6 8.8 0.8 14.4 3.9 11.3 2.4 0.0 0.0 15.8 0.5 15.3 1.3 10.3 0.9 10.0 0.4 18.4 n/a 10.4 0.7 11.6 1.0

Ratio:white 0.9 1.2 5.9 3.6 0.0 0.7 2.0 1.4 0.7 n/a 1.0 1.5

Total 84 285 534 388 12 76 111 126 50 98 1255 3019

RPKP 6.3 9.1 27.1 21.1 3.2 3.1 8.5 9.1 4.3 n/a 6.3 8.7

Ratio:white 1.0 1.4 4.3 3.3 0.5 0.5 1.3 1.4 0.7 n/a 1.0 1.4

NB: The high rate of charging Bangladeshi males will largely be accounted for by practices in two boroughs: 55 per cent of Bangladeshis accused of cannabis possession were recorded in Tower Hamlets and 13 per cent in Newham. Table 12: Number and rate charged with cannabis possession, by 4-year age group and Self-Defined Ethnicity (males aged 10-69 years, 2006) Age Group Self-defined ethnicity Bangladeshi Black African Black Caribbean Black Other Chinese Indian Other Other Asian Pakistani Unknown White Total

Number 0 1 4 2 0 0 1 3 0 0 17 28

10 to 13 RPKP Ratio:white 0.0 0.0 0.1 0.3 0.4 2.4 0.2 1.2 0.0 0.0 0.0 0.0 0.1 0.8 0.4 2.4 0.0 0.0 n/a n/a 0.2 1.0 0.2 0.9

Number 31 146 212 158 7 20 33 41 12 19 381 1060


Number 64 188 314 165 2 27 54 36 29 46 340 1265


Number 38 88 171 89 0 22 31 30 23 28 225 745


Number 14 41 93 44 1 15 9 12 15 9 115 368

28


Number 5 13 56 33 1 7 12 2 9 11 88 237


Number 6 17 40 10 1 1 4 3 2 9 72 165


Number 0 9 47 10 3 2 5 1 1 9 65 152


Number 1 13 78 23 5 5 3 2 0 11 68 209



3.4 In-depth examination of a sample of 144 CRIS records •

Introduction: the supervision of police decision-making

•

Methodology

•

Data Summary

•

Reviewing decisions

• 3.4.1

-

Initial suspicion: police / offender contact

-

Arrest decisions

-

Proceedings decisions

Conclusions and recommendations

Introduction

Having identified that there are disproportionate outcomes by ethnicity in relation to the policing of cannabis, a detailed review of a sample of post-arrest cases was undertaken. The purpose of this review was to explore the degree to which the differential outcomes observed were the result of police decisions rather than, for example, circumstances that differed systematically by ethnicity. To this end, this review has examined both CRIS records and associated sources of evidence (e.g. the MPS Custody Database and Police National Computer) to determine whether the decisions taken by the police to stop, arrest and proceed against individuals (where relevant) were justified by the full facts of each case, and if not, whether this differed by the ethnicity of the offenders. The supervision of police decision-making An essential component of ensuring that the Metropolitan Police’s policies and procedures are applied equitably (that is consistently and based on the full, objective and relevant facts of each case) is case supervision. Supervision may take place at a number of levels, including (in each case, where relevant): •

Decision to stop – where officers are being supervised on the street (e.g. during training)

•

Stop and account/search – by a Sergeant via the ‘Stops and Searches’ form (form 5090).

•

Arrest decision – made by individual officers and therefore generally unsupervised at the time.

•

Detention – by a Custody Sergeant: who will decide whether to authorise detention (and therefore in effect retrospectively supervise the arrest decision).

•

Proceedings decision – by a Custody Sergeant and/or CPS in the case of an arrested individual, by the Crime Management Unit/Designated Decision Maker (DDM) in the case of individuals not arrested (e.g. given a cannabis warning).

•

Supervision of the CRIS record – by a CRIS supervisor in the Crime Management Unit: to ensure that all relevant data and facts are recorded in order to, amongst other things, satisfy the requirements of case integrity – decisions are lawful, reasonable and auditable – and the Home Office crime counting and sanctioned detections rules.

Importantly, cannabis possession offences are typically supervised to a much lesser degree than most other crime types because the majority of cases are dealt with by an individual officer (or pair of officers) using his/her/their discretion to decide ‘in the field’ whether to issue a cannabis warning or pursue an alternative case disposal such as caution or charge. In 2006, for example, 80 per cent of adult males found in possession of cannabis were issued with a cannabis warning (n=26,521), and the vast majority of these cases will have taken place away from the more controlled environment of a police custody facility 23 . As a result, it is essential that the supervising officer can make an informed decision about each case to ensure that procedures and policies are adhered to and applied equitably. 23 In the five boroughs whose records are being reviewed here a total of 266 cannabis warnings were issued to black Caribbean or white offenders during the four weeks under review. Only 22 (8.3 per cent) of those were issued post-arrest (assuming none were issued post arrest where no ‘how arrested’ code or custody number were recorded in CRIS).

29

NOT PROTECTIVELY MARKED Overall, there are three key decision making points: •

Initial suspicion – ‘reasonable grounds’ must exist

•

Whether to arrest – subject to the ‘necessity test’ (i.e. an arrest must be necessary)

•

How to ‘proceed’ with the case – whether to issue a cannabis warning, caution, charge etc.

In the majority of cannabis possession cases, which are dealt with away from a police station by the issuing of a cannabis warning, supervision will concentrate on (i) ensuring that reasonable grounds existed for the initial suspicion that led to police/suspect contact, (ii) ensuring that the correct procedures were followed, and (iii) ensuring that the necessary evidential requirements were met and recorded in the CRIS record. The present review focuses more narrowly on cases recorded in CRIS as having involved an arrest, on the basis that these should have been supervised more rigorously and should therefore be characterised to a greater degree by the equal treatment of different offenders. It will be seen that this review highlights two issues of fundamental importance to the Metropolitan Police’s ability to monitor the impact of its policies and procedures relating to cannabis possession: the first being data accuracy and the second relating to effective supervision. A number of recommendations are made to address these issues.

3.4.2

Methodology

As stated, in order to examine post-arrest proceedings decisions a sample of CRIS records has been examined in detail. The 144 records are comprised of all white (n=67) and black Caribbean (n=77) adult males (18-69 yrs old) accused of cannabis possession and recorded in CRIS as having been ‘arrested’ (according to the methodology described in section 2.2 above) in five boroughs during two weeks in each of March and September 2006 24 . The single comparison between white and black Caribbean ethnic groups was selected to simplify the analysis. The boroughs selected were those with the greatest number of adult black Caribbean accused in 2006, namely Brent, Hackney, Haringey, Lambeth and Southwark (see Table 13). Table 13: Ethnicity of accused by borough (5 borough sample) Owning Borough Name Brent Hackney Haringay Lambeth Southwark Total

Black Caribbean 18 13 5 25 16 77

White 16 7 20 11 13 67

Total 34 20 25 36 29 144

The 144 records were reviewed by way of a two-stage process: (i)

Initially, each CRIS record was examined in detail, a summary of the account was recorded, and a determination made as to whether the CRIS record provided a reasonable justification for the (a) initial decision to stop, (b) decision to arrest and (c) proceedings decision (all where appropriate). 25

(ii)

Where gaps were found in the CRIS record, further evidence was collected from the following computer systems: (a) the MPS Integrated Intelligence Platform (IIP) (for a consideration of the Stops and Custody database records), (b) the Police National Computer (PNC) (for a consideration of the accused’s offending history), and (c) MCRAC, the MPS-wide version of CrimInt, the Criminal Intelligence database. 26 It should be noted that hard-copy records (e.g. paper custody records, stops slips and officers’ notebooks) were not referred to in the interests of expediency.

24

CRIS records with a recorded date of 13-26 March 2006 or 11-24 September 2006 (inclusive). The two two-week periods were chosen to control for any seasonal effects and limit the impact of any police operations that may have been under way. Furthermore, both were during normal school term time. 25 That is, the decision appeared reasonable on the basis of the entirety of the information provided. 26 The author is grateful to Redacted – 5 for her assistance in conducting this stage of the research.

30

NOT PROTECTIVELY MARKED This process was intended to build up as complete a picture as reasonably possible of the circumstances of each offence, including such factors as the circumstances of the offence, the offenders’ behaviour, and any previous criminal record. Overall, then, a decision was made as to whether, taken as a whole, the details recorded across these various systems provided a reasonable justification for the stop, arrest and proceedings decisions (again, where relevant).

3.4.3

Data summary

Table 14 presents a summary of the findings. It can be seen that, reflecting the MPS-wide strategic analysis presented above, the white offenders were more likely than their black Caribbean counterparts to have been given a cannabis warning (67 per cent as compared to 43 per cent) and less likely to have been charged (12 per cent as compared to 26 per cent). Table 14: Whether arrested and how proceeded against by ethnicity (5 borough sample) How arrested Arrested for cannabis possession offence only Arrested in context of/for other offence(s) Not arrested Total % of ethnic group

Caution

Black Caribbean Cannabis No Further Warning Action

Charged

Total

Caution

Charged

White Cannabis No Further Warning Action

Total

Total

% of total

7

6

1

0

14

6

2

1

0

9

23

16.0

10 3 20 26.0

14 0 20 26.0

12 20 33 42.9

3 1 4 5.2

39 24 77 100.0

7 1 14 20.9

6 0 8 11.9

8 36 45 67.2

0 0 0 0.0

21 37 67 100.0

60 61 144

41.7 42.4 100.0

However, it is also clear that there are significant problems with the data as no arrest took place in 61 cases (42 per cent) where an arrest was nevertheless indicated by the CRIS record 27 ; all related to the issuing of cannabis warnings, although four were incorrectly recorded in the proceedings field in CRIS as cautions and one case was subsequently changed to ‘no further action’ due to an incomplete CRIS record. This appears therefore to be a systematic data-recording problem with serious implications for the ability of the MPS to routinely monitor at a corporate level the impact of its policies and procedures in relation to arrests. 28 This also suggests that the sample being examined here may not be representative, as it is not a sample of all cases, but rather those where an arrest was indicated. In those cases where an arrest was confirmed, it can be seen that the black Caribbean offenders were more likely to have been charged than their white counterparts and less likely to have been cautioned or given a cannabis warning (Table 15). Table 15: Whether arrested and how proceeded against by ethnicity – confirmed arrests (5 borough sample) Black Caribbean How arrested Arrested for cannabis possession offence only Arrested in context of/for other offence(s) Total % of ethnic group

Caution

Charged

White

Cannabis No Further Warning Action

7

6

1

10 17 32.1

14 20 37.7

12 13 24.5

3 3 5.7

Cannabis No Further Warning Action

Total

% of total

Total

Caution

Charged

14

6

2

1

0

9

23

27.7

39 53 100.0

7 13 43.3

6 8 26.7

8 9 30.0

0 0 0.0

21 30 100.0

60 83

72.3 100.0

Total

The 83 cases that were confirmed as having involved an arrest were comprised of 23 (27.7 per cent) in which the arrest was for possession of cannabis alone and 60 (72.3 per cent) in which the accused was arrested in the context of (or for) other offences. The proportion of offenders arrested in the context of or for other offences was slightly higher for black Caribbean offenders (73.6 per cent) than their white counterparts (70.0 per cent). The 23 arrests solely for cannabis possession (14 black Caribbean, 9 white) were based on the circumstances shown below in Table 16.

27

115 of the 144 records had a ‘how arrested’ code. Of the remaining 29, 20 had a valid custody number and 9 involved cautions being recorded. It should be noted that other enquiries have highlighted the fact that this exact same issue also arises in the context of Stops records: cannabis warnings issued in the absence of an arrest are nevertheless being recorded as ‘arrests’ by some boroughs.

28

31

NOT PROTECTIVELY MARKED Table 16: Reason for arrest by ethnicity: arrests for cannabis possession only (5 borough sample) Reason for arrest Establish particulars Initial denial proved false Not clear Offender behaviour Positive arrest policy Potential PWITS Repeat offender Youth present Grand Total

Black Caribbean 1

White 1 1 2

2 1 2 5 1 2 14

3 2 9

Grand Total 2 1 4 1 5 5 3 2 23

It is only in relation to cases where Possession with Intent to Supply (PWITS) was suspected that any differences by ethnicity are seen, although it should also be said that the numbers involved are very small and therefore potentially unreliable – a larger sample would have produced more robust findings. Those records to which a positive arrest policy applied were in areas where a local decision had been taken to arrest anyone found in possession of cannabis as a means of tackling local crime problems – typically open drug dealing. The ‘repeat offenders’ were those who had previous offences relating to cannabis or other drugs, either referred to in the CRIS record or recorded on PNC. The 60 cases in which the accused was arrested in the context of (or for) other offences (39 black Caribbean, 21 white) related to a wide range of circumstances, as shown in Table 17. These included 11 arrests that related to driving offences, 7 that related to violence, 7 offenders who were believed to be wanted on warrant, 6 that related to public order act offences, 5 that related to offensive weapons and 4 that related to suspected immigration offences. Table 17: Reason for arrest by ethnicity: arrests in context of/for other offences (5 borough sample) Reason for arrest Blackmail Criminal Damage Driving Driving & Motor Crime Driving & Offensive Weapon Drunk and Disorderly Fare Evasion & Identity Firearms Handling Stolen Goods Immigration Money Laundering Motor Crime Motor Crime & Firearms Not clear Offensive Weapon Offensive Weapon & Criminal Damage Other Drugs Public Order Act Soliciting Theft Theft & Violence Violence Violence & Obstruction Wanted on Warrant Wanted on Warrant & Public Order Act Grand Total

Black Caribbean 1 5 1 1

White 2 4

1 1 1 1 2 3 1 2 1 1 3 3 1 1 3 1 5 1 39

1 2 1

2

2 1 2 2 1 21

Grand Total 1 2 9 1 1 1 1 1 2 4 1 3 1 2 3 1 3 5 2 2 1 5 1 6 1 60

At this stage, no clear differences can be seen by ethnicity. The largest difference between the ethnic groups appears to relate to offenders wanted on warrant, circumstances that applied to 6 black Caribbean offenders and only one white offender.

32


3.4.4

Reviewing police decisions

In this section, attention will turn to the audit trail documented in the CRIS records and associated sources. Specifically, the three key decision-making points will be reviewed, namely whether grounds existed for the initial contact between the police and offender, whether an arrest was ‘necessary’, and finally, whether the proceedings decided upon were justified.

Initial suspicion: police/offender contact Table 18 presents the analysis of whether the initial contact between police and offender – typically in the form of a stop – appeared to have been justified, initially on the basis of the CRIS record alone, and then on the basis of all documentary sources taken together. The numbers in the table cells relate to the numbers of records in which the decision – in this case to stop – appears to have been fully justified on the basis of a reasonable assessment of the available evidence; the totals row indicates the overall number of cases being considered. Table 18: Cases where the initial police/offender contact was justified (5 borough sample) Black Caribbean Source CRIS % of total All Sources % of total Total

Caution 15 75.0 16 80.0 20

Cannabis Charged Warning 15 75.0 15 75.0 20

9 27.3 13 39.4 33

White No Further Action 3 75.0 3 75.0 4

Total

Caution

42 54.5 47 61.0 77

7 50.0 9 64.3 14


5 11.1 12 26.7 45

No Further Action 0 0.0 0 0.0 0

Total 19 28.4 28 41.8 67

Total 61 42.4 75 52.1 144

It can be seen that overall only 42.4 per cent of CRIS records presented sufficient information for a judgement to be formed that grounds existed for the initial contact, rising to 52.1 per cent when all documentary sources (CRIS, IIP, MCRAC and PNC) were taken into account. This is principally related to those cases in which cannabis warnings were issued, where only 26.7 per cent of records relating to white offenders and 39.4 per cent of records relating to black Caribbean offenders provided evidence that grounds had existed. The fact that the white offenders were more likely to have received cannabis warnings probably also explains why overall the initial contact was justified in a higher proportion of black Caribbean than white offenders’ cases. A common feature of cannabis warning records in CRIS is that they contain very little detail beyond a standard ‘proforma’ wording that has been cut-and-pasted into the CRIS Details (‘Dets’) pages. In particular this rarely includes any information about the grounds for initial suspicion, simply stating that the suspect was ‘stopped and searched at [the] venue and found in possession of cannabis’. A notable exception is where cannabis was smelt by the police officer and this is referred to in the CRIS record. Notwithstanding the fact that the initial stops will have been supervised to a degree via the Stops slip (form 5090), these findings must raise concerns about the degree to which police powers to stop members of the public either to ask them to account for their behaviour or submit to a search are based on genuine reasonable suspicion. Alternatively, one may ask the question: if reasonable grounds existed, why have they not been routinely recorded alongside the principal offence account in the CRIS Dets? Arrest decisions Table 19 presents the same analysis, but this time in relation to the decision to arrest in the 83 cases where an arrest was confirmed. It can be seen that overall the arrest decision was sufficiently well documented that it appeared justified (that is, necessary) in 84.3 per cent of cases on the basis of the CRIS record alone and 94.0 per cent of cases when all sources were taken into account. Interestingly, fewer of the arrests of white offenders were fully justified by the recorded evidence than is the case for their black Caribbean counterparts, both in relation to the review of CRIS records alone and all sources combined. The differences, however, are not great.

33

NOT PROTECTIVELY MARKED Table 19: Cases where the arrest decision was justified (5 borough sample) – only where an arrest was confirmed Black Caribbean Source CRIS % of total All Sources % of total Total

Caution 16 94.1 16 94.1 17


12 92.3 13 100.0 13


Total

Caution

47 88.7 51 96.2 53

9 69.2 9 69.2 13


7 77.8 9 100.0 9


Total 23 76.7 27 90.0 30

Total 70 84.3 78 94.0 83

Perhaps unsurprisingly, all 60 of the arrests in cases where the cannabis offence was in the context of another offence were fully justified by the documented evidence; when only the CRIS records were examined this falls to 93.3 per cent (56 out of 60 records). By contrast, in the 23 cases where the arrest was only for a cannabis offence, the arrest appeared justified in only 14 (60.9 per cent) of the records on the basis of the information recorded in CRIS records alone, and 17 cases (73.9 per cent) when all sources were examined. This difference can largely be accounted for on the basis that where cannabis possession was a secondary offence, either the principal offence was one for which arrest was necessary or the police discovered the cannabis subsequent to an initial arrest. Conversely, in a number of cases of simple cannabis possession cautions were given when it appears a cannabis warning could have been issued (and therefore arrest avoided) and no explanation was offered to explain why the arrest was necessary. Proceedings decisions Tables 20 and 21 present the analysis for proceedings decisions, respectively examining all 144 cases and then only the 83 where an arrest was confirmed. Table 20: Cases where the proceedings decision was justified (5 borough sample) Black Caribbean Source CRIS % of total All Sources % of total Total

Caution

Charged

Cannabis Warning

6 30.0 9 45.0 20

10 50.0 12 60.0 20

26 78.8 25 75.8 33


Total

Caution

Charged

Cannabis Warning

45 58.4 49 63.6 77

4 28.6 4 28.6 14

0 0.0 6 75.0 8

44 97.8 44 97.8 45


Total 48 71.6 54 80.6 67

Total 93 64.6 103 71.5 144

Table 21: Cases where the proceedings decision was justified (5 borough sample) – only where an arrest was confirmed Black Caribbean Source CRIS % of total All Sources % of total Total

Caution

Charged

Cannabis Warning

4 23.5 7 41.2 17

10 50.0 12 60.0 20

6 46.2 7 53.8 13


Total

Caution

Charged

Cannabis Warning

22 41.5 28 52.8 53

3 23.1 3 23.1 13

0 0.0 6 75.0 8

8 88.9 8 88.9 9


Total 11 36.7 17 56.7 30

Total 33 39.8 45 54.2 83

Here it can be seen that the proceedings decisions were fully justified by the documentary evidence in relation to a greater proportion of white offenders than their black Caribbean counterparts, both overall (80.6 per cent vs. 63.6 per cent) and when examining only confirmed arrests (56.7 per cent vs. 52.8 per cent). Interestingly, it can be seen that proceedings decisions were most problematic in relation to the issuing of cautions, for both black Caribbean and white offenders; 10 cautions were not supported by sufficient evidence to make them seem reasonable in relation to each group. Most commonly, this was in cases where it appears that a cannabis warning could have been issued, for example where a full and frank admission was given at the first opportunity in the absence of a history of cannabis possession offences. By way of an example, in a small number of cases this review has highlighted what appear to be inconsistent practices relating to the issuing of cannabis warnings in custody, notably in the context of

34

NOT PROTECTIVELY MARKED other offences. In some of the cases examined individuals were charged on the non-cannabis offences and issued with a cannabis warning; in other cases suspects were charged on all allegations. These findings are quite problematic because they leave open the possibility that proceedings decisions are inconsistently applied, both within ethnic groups, but also between them.

3.4.5

Conclusions and recommendations

This detailed review of a sample of 144 CRIS records drawn from five London boroughs has raised two significant concerns about the ability of the Metropolitan Police Service to routinely monitor the impact of its procedures and policies relating to the policing of the possession of cannabis. (i)

Data recording problems: some boroughs at least appear to be routinely recording offences involving the issuing of cannabis warnings as arrests when no arrest took place.

(ii)

A failure to routinely record the grounds for key decisions: the audit trail is fragmented across multiple records and systems and even then appears often to be incomplete.

The data recording problem should not need further discussion here: clear guidance must be issued to CRIS users that arrests should only be recorded when arrests actually take place, and this fact should be reflected in the Standard Operating Procedures (SOPs) for Cannabis Possession (which they currently are not). CRIS supervisors must ensure the integrity of CRIS data and some form of monitoring process may be required to maximise compliance. It should be said that this issue also applies to both the recording of invalid custody numbers in CRIS, and arrests recorded in the Stops database, when cannabis warnings have been issued and no arrest has taken place. The latter clearly has broader implications for the monitoring of stop and search powers. The question of how the MPS captures a complete audit trail of decisions relating to individual cases and ensures an adequate and consistent degree of supervision is rather more problematic. At present the grounds for key decisions are fragmented across a range of sources, including officers’ notebooks, Operation Emerald ‘cannabis warning’ forms (Form 67), stops slips (Form 5090), custody records (including handwritten sections available only as hard copy), CrimInt and the CRIS database (which in turn may be fragmented further where multiple offences are recorded for the same incident and have to be cross-referenced – which does not always happen). Implicit in this is the assumption that the sum of the parts is the same thing as the whole: in other words, the independent supervision of individual sequential decisions adds up to adequate end-to-end oversight of what is inevitably a highly complex process. The evidence presented in this detailed review of a sample of records suggests that at present the CRIS database does not routinely allow for the end-to-end supervision of these key decisions, and that even when a range of additional sources are taken into account significant gaps often remain in the audit trail. In particular, it is of concern that reasonable grounds for initial police/offender contact could only be established in 52 per cent of records and proceedings decisions appeared fully justified in only 54 per cent of records. This is not to say that grounds definitely did not exist or that proceedings decisions were wrong, but rather that sufficient evidence could not reasonably be obtained to fully support them. This combination of a fragmented and potentially incomplete audit trail leaves open the possibility that the supervision process is not sufficiently effective to ensure the equal treatment of different individual members of the public when they come into contact with and are dealt with by the police. Taken together, these issues give rise to the following question: does supervision need to be rethought in order to reassure both the public and the MPS itself that policies and procedures are applied consistently and in an even-handed manner? At this juncture it should be noted that it is only in the CRIS record that a full account of an individual offence can be recorded, from initial police/offender contact through to case disposal. It is therefore proposed that each CRIS record should include a comprehensive, transparent and unambiguous account of the grounds for the following decisions (where they are relevant): (i)

The initial police/offender contact: e.g. why ‘reasonable grounds’ existed;

35

NOT PROTECTIVELY MARKED (ii)

Arrest: why it was ‘necessary’;

(iii)

Proceedings: why the chosen proceeding was appropriate and the why the alternatives were inappropriate.

It is further proposed that these requirements be incorporated into the ‘required information’ detailed in the CRIS ‘Details’ screen. It is recognised that these proposals raise two important issues. Firstly, they would inevitably introduce a degree of additional bureaucracy, which would probably be resisted. Secondly, some form of monitoring and enforcing of standards would be required to ensure that (a) the additional requirements are actually observed, and (b) that the information recorded is sufficiently detailed, rather than simply paying lipservice to any new guidelines. 29 Exactly how these issues would be dealt with remains open for discussion. In all likelihood, the additional bureaucracy would be minimal and come with the dividend of additional transparency. The question of standards would probably have to be scrutinised via a dip-sample approach, which in turn might incorporate some degree of external review, for example in the form of an independent advisory group (IAG). The bottom line is that current procedures do not appear to be fit for purpose. Monitoring and supervision are both compromised and arguably the Metropolitan Police would struggle to reassure external audiences that, as things presently stand, it is confident that its policies and practices in relation the policing of possession of cannabis are equitable. It is clear that action must now be taken to address these issues.

29 It is suggested that a test of ‘sufficiently detailed’ would be that an independent reader (perhaps even a non-police reader) could be satisfied that he or she could understand what happened and why key decisions were taken, without reference to other sources.

36


4. Conclusions and discussion This review of disproportionality in relation to the policing of cannabis has been frustrated at a number of stages by one fundamental overarching problem: a lack of consistency and accuracy in the recording of data. Moreover, this has been compounded by a systematic failure to use the CRIS database to record the full details of individual cases. The consequences of these issues are that, firstly, in many cases (specifically in relation to cannabis warnings) it is not possible to know without examining individual records in detail whether an individual was arrested; and secondly, it is not possible to know the full circumstances of each incident recorded in CRIS without having to reconstruct them from a range of different sources (e.g. officers’ notebooks, stops records and custody records), which may nevertheless remain incomplete. Two fundamentally important consequences arise: 1. Reliable, systematic and effective corporate-level monitoring of the impact of policies and practices relating to the policing of cannabis is not presently possible around the arrest decision. 2. It is not at all clear that cannabis possession cases are presently subject to an adequate level of supervision, most notably around the initial grounds for suspicion and proceedings decisions. Nevertheless, it is possible to draw some conclusions from the analysis presented here, not least that many ethnic minority groups, most notably black Caribbean and ‘other’ black males, are both heavily overrepresented overall and more likely to receive more punitive sanctions – two factors that are likely to be very closely related. It is also quite evident that a small number of locations have a disproportionate impact on the picture for the MPS as a whole, with offences clustered not only at a borough level, but importantly also at a ward level. This finding in turn highlights the importance of thinking about the policing of places, rather than policing as a uniform activity that should always ‘look the same’ across London. Crime problems differ from place to place, and it is therefore appropriate that so does the policing response, not least in locations such as Camden and Brixton where open drug dealing is being met with a robust set of responses, including pro-arrest policies. One consequence of this may well be the introduction of a degree of disproportionality if the level of policing differs by location and, for whatever reason, is higher in locations with a proportionately larger non-white population than London as a whole. Nevertheless, it is also important to remain alert to the possibility that police powers are being over-used in some locations – that is, to a degree that is not proportionate to the nature and scale of local crime problems relative to the rest of London. These contextual factors lead onto a more discursive level of analysis, reflecting on the full range of factors that influence the policing of cannabis and then seeking to identify those that are of greatest importance. Figure 19 below seeks to illustrate this complexity by grouping these factors into ‘domains’ (see also Appendix D for more details).

37

NOT PROTECTIVELY MARKED Figure 19: Factors impacting on cannabis policing outcomes Profile of cannabis possession accused Organisational culture Legislative domain

Performance domain



Offender domain

Lifestyle domain




Demographic profile












The narrative that appears to tie these ‘domains’ together, based on the data analysed in this report and discussions with a range of individuals from across the MPS, positions the policing of cannabis at the intersection of three issues: (i)

Stop and search, most notably as one of the major planks of the Metropolitan Police’s response to personal robbery, in relation to which (a) levels of frontline police resources and (b) the general profile of robbery suspects are critical, and (c) it is not clear how specific are the grounds used to justify many stops.

(ii)

Other efforts to tackle local crime and disorder problems, which might include issues such as drug dealing and open drug use.

(iii)

Local interpretation of performance culture, not least viewing cannabis offences as a readily available source of easily obtained sanctioned detections; for example, in crude performance management terms one cannabis warning is ‘worth’ the same as a charge brought against a rape suspect.

It should be said that these issues assume different degrees of significance in different locations, and indeed it may be possible to envisage a classification schema that identifies the key drivers on a place-byplace basis. It is also worth stating that these themes corroborate, and are corroborated by, previous research (e.g. May et al., 2007). In terms of the ethnic disproportionality dimension, one of the most consistent themes that has arisen in discussing these issues is the profile of personal robbery suspects in London, a crime type that is widely held to be dominated by young black men. In this respect, it remains to be established how ‘reasonable suspicion’ is interpreted, most notably in areas that have an identified robbery problem; in this regard the quality and specificity of suspect profiles produced by borough analysts is of fundamental importance. The concern is that in some cases, ‘reasonable suspicion’ falls at the margins of what might be referred to as ethnic profiling, for example where ‘suspect profiles’ refer in general terms to ethnicity and age

38

NOT PROTECTIVELY MARKED characteristics, rather than more specific personal identifiers. 30 Confronted with a consistent failure to record the grounds for initial suspicion it is difficult to explore this further. It should also be said that it is not clear how reliable the ‘reason for search’ data recorded on the 5090 stops forms are, as in many cases the ‘reason for search’ appears to have been recorded once the facts had been established; that is, a form may state ‘drugs’, which were discovered, even in cases where it is not at all clear why the individual searched was suspected of drugs possession or indeed whether that was the reason for a search in the first place, rather than a more generalised suspicion that ‘something wasn’t right’. The same issue applies to CRIS records that refer to s23 (Misuse of Drugs Act 1971) searches when no supporting evidence is provided to explain the initial suspicion. Finally, in relation to the impact of the MPS’s ‘performance culture’, the central importance of cannabis warnings to sanctioned detections will in all likelihood cease to be an issue once the next round of Public Service Agreement (PSA) targets have been agreed and implemented in 2008, as it is widely expected that the focus of policing will shift towards policing serious crime and serious harm. This will probably mean that cannabis possession offences ‘brought to justice’ will no longer contribute to overall sanctioned detection performance measures and it will be interesting to see, therefore, whether the number of cannabis possession offences falls accordingly. It is of course possible that stop and search and other similarly intrusive tactics will be used to an increasing degree, and continue to find cannabis as a derivative outcome. Concluding remarks The 2006 cannabis enforcement Monitoring Report identified ‘what could be construed as… a disproportionate impact to the African/Caribbean community’, and called for ‘in-depth analysis’ to explore ‘whether or not there are sound reasons for the disparity in charges and cautions between the communities identified’ (MPS, 2006: 7). This research has examined the Metropolitan Police’s data in considerable detail, and as a result has uncovered a number of areas of concern. Some are process related; concerned with data accuracy and ensuring that the MPS can understand how its own policies work in practice. Others, however, are more problematic and relate to the impact of policing on particular sections of London’s diverse population. Nowhere is this more apparent than in relation to the numbers (and especially the proportion) of black Caribbean men in their late teens and early twenties coming into contact with the police – in this case in relation to the policing of cannabis, but by implication in many other ways as well. When as many as one quarter of all 21 year old black Caribbean men are found in possession of cannabis in a single year, difficult questions need to be asked about whether the public interest is being served or whether in fact significant harm is being done to the legitimacy of policing and police/community relations in some parts of London. The rapid increase in the number of cannabis possession offences recorded since early 2004, when viewed against a wealth of evidence suggesting falling cannabis use, suggests that these issues may be becoming more acute. Moreover, it appears that this is a process that has been encouraged by the prevailing performance regime. It may well now be appropriate – indeed unavoidable – to reconsider the recent direction of travel. Although specific recommendations have not been made about responding to these findings, it seems likely that any response needs to involve local communities in a conversation about the appropriate response to local problems of crime and disorder.

30

Two issues are worth mentioning here. Firstly, National Intelligence Model (NIM) products have a section titled ‘inferences’, in relation to which the author has seen examples of very generalised suspect profiles. For example “The suspects are young IC3 males…” in relation to a crime type in which 71 per cent of suspects were described as African/Caribbean. In other words, the 29 per cent of suspects who were not African/Caribbean were ignored. Secondly, it often appears that analysts have failed to distinguish in suspect profiles between the number of suspects recorded in CRIS records and the number of actual individuals who committed the offences. Notwithstanding the fact that the latter may be unknowable, this does at least demand that the data are treated critically. By way of a hypothetical example, assume 100 personal robbery offences have been recorded in one area over a period of time, in relation to 80 of which the suspect was described as black and 20 where the suspect was described as white. It does not necessarily follow that 80 per cent of suspects are black: if, for example, 20 black offenders each committed four offences and 20 white offenders each committed one, then the true picture would be that 50 per cent of the suspects were white and 50 per cent black. It is entirely possible that a handful of prolific offenders could skew the profile of suspects in an area in this way, but this does not generally appear to be considered in the production of suspect profiles, which ultimately form an important component of the establishment of ‘reasonable suspicion’.

39


5. Appendix A: Proceedings codes Table 22: Grouping proceedings codes Proceedings Code AB AF AS AT BA BB BC BD BJ BK CH D1 D2 D3 D4 D5 D6 D7 D8 DF EA PC XD

Description Summons First time charged Second time charged Third and subsequent time charged Youth reprimand 1st offence Youth warning 1st offence Youth warning 2nd offence Youth warning 2 years + Adult caution Conditional caution Taken Into Consideration Accused dead Accused ill etc Complainant dead Victim etc refuses to testify (1/10/2005 onwards) Child under 10 CPS decline prosecution Police decline prosecution Time expired DDM - Cannabis formal warnings Fixed penalty notice Pending decision No further action

Summary Other Charged Charged Charged Caution Caution Caution Caution Caution Caution Other Other Other Other Other Other Other Other Other Cannabis Warning Other Other Other

It should be noted that these are the proceedings codes that appeared in the data analysed in this report, not all possible proceedings codes.

40


6. Appendix B: Cannabis warnings as a percentage of all Sanctioned Detections – 2006/07 by borough Figure 20 illustrates the contribution of cannabis warnings to overall sanctioned detections, on a boroughby-borough basis. Figure 20: Cannabis warnings as a percentage of Sanctioned Detections – by borough, 2006/07 Cannabis warnings as a percentage of Sanctioned Detections (SDs) By borough 2006/07 Source: PIB (Stephen Oleskiw) Report 20th April 2007 "Monthly Offences and Detections Ordered by % Detections " 30.0

Cannabis warnings as % of SDs

25.7 25.2 25.0 24.0 23.5

25.0

21.6

21.0 19.8

20.0

18.5 16.8 16.8 16.4 16.4

15.6

15.0

14.8 14.4 14.0 13.7 13.5 12.0 12.0 11.5

10.7 10.5 10.2

10.0

8.4

8.0

7.5 6.3

6.2

5.9

5.9 4.9

5.0

1.3

Sutton

Heathrow

Bexley

Merton

Kingston Upon Thames

Barnet

Bromley

Havering

Richmond Upon Thames

Islington

Newham

Harrow

Greenwich

Croydon

Hounslow

Redbridge

Barking & Dagenham

Lewisham

Hammersmith & Fulham

MPS Total

Tower Hamlets

Camden

Hackney

Hillingdon

Kensington & Chelsea

Lambeth

Southwark

Ealing

Haringey

Waltham Forest

Westminster Total

Brent

Enfield

Wandsworth

0.0

Borough

Figure 21 illustrates the fact that as the number of cannabis warnings increases so does the proportion of all sanctioned detections comprised of cannabis warnings – the number of cannabis warnings is not proportionate to the overall number of sanctioned detections. Taken together, these two graphs illustrate the differential contribution of cannabis warnings to sanctioned detection performance. Figure 21: Scatter plot of the relationship between the number of cannabis warnings issued and cannabis warnings as a percentage of sanctioned detections – by borough, 2006/07 Cannabis warnings as a percentage of Sanctioned Detections - Boroughs 2006/07 Source: PIB (Stephen Oleskiw) Report 20th April 2007 "Monthly Offences and Detections Ordered by % Detections "

Cannabis Warnings as % of Sanctioned Detections

40

y = 0.0084x + 6.2616 2 R = 0.7472

35

30

Wandsworth

Bent

25

20

Westminster

15

10

5 Heathrow 0 0

500

1000

1500

2000

2500

3000

3500

Number of Cannabis Warnings Issued

Note: excluding Westminster (a clear outlier) from Figure 20 raises the R2 value to 0.8897.

41

4000


7. Appendix C: Ethnic Groups Table 23 below indicates the relationship between the Aggregated Ethnic Groups (AEGs) used by the GLA to present population projections, and the 2001 Census categories. The latter are used for the Metropolitan Police Service self-defined ethnicity (SDE) codes recorded in CRIS. Table 23: Aggregating ethnic groups GLA Aggregated Ethnic Group

2001 Census Ethnic Category

White

White: British White: Irish White: Other

Black Caribbean

Black or black British: Caribbean

Black African

Black or black British: African

Black Other

Black or black British: Other black Mixed: White and black Caribbean Mixed: White and black African

Indian

Asian or Asian British: Indian

Pakistani

Asian or Asian British: Pakistani

Bangladeshi

Asian or Asian British: Bangladeshi

Other Asian

Mixed: White and Asian Asian or Asian British: Other Asian

Chinese

Chinese or other: Chinese

Other

Mixed: Other mixed Chinese or other: Other

Finally, in analysing the MPS data the following SDE codes were categorised as ethnicity ‘unknown’: •

Officer urgently required elsewhere

•

Person declines to define ethnicity

•

Person does not understand

•

Situation involving public disorder

42


8. Appendix D: Factors impacting on cannabis policing outcomes Profile of cannabis possession accused

Profile of cannabis possession accused

Lifestyle domain

Legislative domain Cannabis reclassified Reasonable grounds for suspicion Legislation used for stop/search

Potential changes to availability and use: frequency, visibility

Related to each other – e.g. s60 does not require reasonable suspicion

Attitudes towards cannabis use differ

Intended prioritisation of Class A drugs

Recreation and social spaces (public vs. private) related to social, cultural and economic factors


Socio-cultural factors Economic factors Peer group effects

Potential tension between personal and organisational performance measures / targets: results vs. ‘fairness’

Organisational culture Police officer domain Professional experience

Peer groups may regulate or encourage cannabis use and visibility

Identification of the ‘usual suspects’


Relationship to ‘reasonable suspicion’


Profile of cannabis possession accused Organisational culture Legislative domain

Performance domain



Offender domain

Lifestyle domain




Demographic profile













Profile of cannabis possession accused Organisational culture


Performance domain PSA Targets: reduce overall crime

Tension with proactive policing of drugs

OBTJs: maximise SD rate

Cannabis warnings: quick and cheap wins

Organisational priorities (robbery)

Influences policing tactics and locations targeted

Scrutiny (e.g. Operation Pennant)

Will modify behaviour, e.g. emphasis on arrest rates


Age and ethnicity profile will differ between locations – e.g. concentration of particular ethnic communities

Demographic profile Local (criminal) economy Crime problems and policing priorities

All closely related to each other and subsequently to the tactical and strategic interventions put in place – e.g. the importance placed on stop and search vs. visibility vs. intelligence


43

Exposure to policing a function of various factors including: types of offences, visibility (time and location of offending), other signals of suspicion and nature of local crime problems Attitude important in relation to officer decisions during interaction with suspect Females problematic as search requires female officer. More generally, the closer an individual is demographically to suspect profiles (e.g. robbery), the greater the likelihood they will be stopped and searched.

Offender domain Behaviour: visibility, attitude Age, gender, ethnicity Relationship to suspect profiles


9. Bibliography Note: Internet references were correct as at 2nd July 2007 but may be subject to change. Aust, R. and Smith, N. (2003) Ethnicity and drug use: key findings from the 2001/2002 British Crime Survey. Findings 209. London: Home Office. See: http://www.homeoffice.gov.uk/rds/pdfs2/r209.pdf Bains, B. and Klodawski, E. (2006) GLA 2005 Round: Interim Ethnic Group Population Projections. DMAG Briefing 2006/22 (November 2006). London: Greater London Authority Data Management and Analysis Group. See: http://www.london.gov.uk/gla/publications/factsandfigures/dmag-briefing-200622.pdf Bowling, B. and Phillips, C. (2002) Racism, Crime and Justice. London: Longman. FitzGerald, M. (1999) Searches in London: Under s1 of the Police and Criminal Evidence Act 1984 (Final Report). London: Metropolitan Police. Graham, J. and Bowling, B. (1995) Young People and Crime. Home Office Research Study 145. London: Home office. See: http://www.homeoffice.gov.uk/rds/pdfs2/hors145.pdf Greater London Authority (2006) GLA 2006 Round Ethnic Group Population Projection – RLP High (Data CD). London: Greater London Authority. Holloway, K. and Bennett, T. (2004) The results of the first two years of the NEW-ADAM programme. Home Office Online Report 19/04. London: Home Office. See: http://www.homeoffice.gov.uk/rds/pdfs04/rdsolr1904.pdf House of Commons: Home Affairs Committee (2007) Young Black People and the Criminal Justice System. Second Report of Session 2006–07: Volume I. HC 181-I. London: The Stationery Office. See: http://www.publications.parliament.uk/pa/cm200607/cmselect/cmhaff/181/181i.pdf May, T., Duffy, M., Warburton, H. and Hough, M. (2007) Policing Cannabis as a Class C Drug: An arresting change? York: Joseph Rowntree Foundation. See: http://www.jrf.org.uk/bookshop/eBooks/1961-policing-cannabis-classc.pdf May, T., Warburton, H., Turnbull, P.J. and Hough, M. (2002) Times they are a-changing: Policing of cannabis. York: Joseph Rowntree Foundation. See: http://www.jrf.org.uk/bookshop/eBooks/1859353878.pdf Mott, J. and Mirrlees-Black, C. (1995) Self-reported drug misuse in England and Wales: Findings from the 1992 British Crime Survey. Research and Planning Unit Paper 89. London: Home Office. See: http://www.homeoffice.gov.uk/rds/pdfs2/rup089.pdf MPS (2006) Policy on possession of cannabis – enforcement policy following its reclassification from a Class B to a Class C controlled drug – Monitoring Report (Version 1.5). Drugs Directorate (SCD3(3)) May 2006. London: Metropolitan Police Service. MPS (2007a) Metropolitan Police Drugs Strategy 2007-2010. London: Metropolitan Police Service. See: http://www.met.police.uk/drugs/publications/drugsstrategy.pdf MPS (2007b) Standard Operating Procedure (SOP): Policing offences of cannabis possession and cannabis warnings (Version 1.8). London: Metropolitan Police Service. MPS (2007c) Stops and Searches Monitoring Mechanism, April 2007 (Version 1.91). London: Metropolitan Police Service.

44

NOT PROTECTIVELY MARKED MVA and Miller, J. (2000) Profiling Populations Available for Stops and Searches. Police Research Series Paper 131. London: Home Office. See: http://www.homeoffice.gov.uk/rds/prgpdfs/prs131.pdf National Centre for Social Research and National Foundation for Educational Research (2007) Smoking, drinking and drug use among young people in England in 2006: headline figures. Leeds: The Information Centre for health and social care. See: http://www.ic.nhs.uk/webfiles/publications/smokedrinkdrug06/file.pdf Phillips, C. and Bowling, B. (2003) The experiences of crime and criminal justice among minority ethnic groups: a review of the literature (Final Report – June 2003). London: King’s College London. Quinton, P., Bland, N. and Miller, J. (2000) Police Stops, Decision-making and Practice. Police Research Series Paper 130. London: Home Office. See: http://www.homeoffice.gov.uk/rds/prgpdfs/prs130.pdf Ramsay, M., Baker, P., Goulden, C., Sharp, C. and Sondhi, A. (2001) Drug Misuse Declared in 2000: results from the British Crime Survey. Home Office Research Study 224. London: Home Office. See: http://www.homeoffice.gov.uk/rds/pdfs/hors224.pdf Ramsay, M. and Spiller, A. (1997) Drug Misuse Declared in 1996: latest results from the British Crime Survey. Home Office Research Study 172. London: Home Office. See: http://www.homeoffice.gov.uk/rds/pdfs/hors172.pdf Roe, S. and Man, L. (2006) Drug Misuse Declared: Findings from the 2005/06 British Crime Survey. HOSB 15/06. London: Home Office. See: http://www.homeoffice.gov.uk/rds/pdfs06/hosb1506.pdf Sharp, C. and Budd, T. (2005) Minority ethnic groups and crime: findings from the Offending, Crime and Justice Survey 2003 (2nd Edition). Online Report 33/05. London: Home Office. See: http://www.homeoffice.gov.uk/rds/pdfs05/rdsolr3305.pdf Waddington, P.A.J., Stenson, K. and Don, D. (2004) ‘In proportion: Race, and Police Stop and Search’. British Journal of Criminology 44:6, 889–914.

45