Whole system approach to young people who offend: evaluation

2. Trends in Crime and Youth Referrals

This section of the report examines overall trends in crime within the three local authority areas and explores how this relates to patterns of deprivation and population size within the areas. It then examines patterns of youth crime as measured through administrative data and looks at how these trends fit broadly within periods of youth justice policy development. Trends in referral to the Children's Reporter are reported and compared pre- and post- implementation of the WSA, as are trends in joint referrals to the Procurator Fiscal and Children's Reporter. The section concludes with an overview of how crime trends and youth referral patterns correspond to the implementation of EEI and WSA. The findings of this section suggest that local offending patterns may be affected by both national and local level factors, and that flexibility in implenting WSA within and across local authority areas may be necessary to addess different contextual conditions.

Two key structural changes have taken place since the roll-out of EEI in 2008 and WSA in 2011. In 2012, the Crown Office and Procurator Fiscal Service ( COPFS) moved from an area-based structure, to a Federation structure, comprised of three regional Federations, each led by a Procurator Fiscal, and one National Federation. On 1 April 2013, the Police and Fire Reform (Scotland) Act, 2012 established the Police Service of Scotland (known as Police Scotland), amalgamating the eight legacy police forces. The single service is structured into three regions (East, West, and North), and 14 local policing divisions, each covering one or more local authorities. Local police Divisions are broken down further into the 353 multi-member wards, which are aligned with local government electoral wards. It was not possible from this small-scale study to determine whether these changes had impacted widely on the operation of EEI or the WSA; however, some study participants noted potential local impact during interviews.

2.1 Crime, deprivation and population in the three local authorities

The three local authorities which constitute the three case study areas vary considerably in terms of geography, population size and crime density. Authority A is a predominantly rural local authority, with a sparse population and comparatively low crime levels. Authority B is city-based, with a large population and more moderate crime levels. Authority C is semi-rural, with a moderate sized population and slightly higher crime levels on average. In terms of population demands, there is negligible variation in the proportion of young people in each authority. In Authorities A and C, 12-15 year olds accounted for five per cent of the total population. In Authority B, this figure was slightly lower, at four per cent. In each authority, 16 and 17 year olds respectively accounted for only one per cent of the overall population.

Overall, trends in recorded crime in Scotland show a significant fall since the early 1990s (Scottish Government, 2014) and this tends to have been reflected across most of the local authorities in Scotland. Table 2.1 compares the three local authorities using crime indicator data from the Scottish Index of Multiple Deprivation ( SIMD). The SIMD ^[2] is a multi-dimensional indicator of overall relative deprivation, which is measured in terms of seven different domains (employment, income, health, education, geographic access, police recorded crime and housing). Data for each individual domain are also available, as shown below. The SIMD crime domain is constructed from a subset of recorded crimes that are most likely to impact on the local neighbourhood. These are violent crime, sexual offences, domestic house breaking, vandalism, drug offences and minor assault. Whilst partial, this provides a useful comparator between the three local authorities.

Table 2.1 Scottish Index of Multiple Deprivation, crime domain data, 2012

	Local Authority
Crime domain	A	B	C
No. of zones in the 15% most deprived datazones	18	85	40
National share of 15% most deprived datazones	1.8%	8.7%	4.1%
Local share of 15% most deprived datazones	9.3%	15.5%	18.7%
Total number of datazones	193	549	214

Table 2.1 shows that Authority B had both the largest number of zones in the 15 per cent most deprived datazones and the highest percentage of the national share, whereas Authority A had the lowest number and percentage share of deprived datazones. However, the key measure in Table 2.1 is the 'local share' of deprived datazones, which show the proportion of datazones within each authority which fall into the most deprived 15 per cent datazones in Scotland. These data show that Authority A has a relatively low proportion of zones in the 15 per cent most deprived datazones in terms of crime (9.3%) and a low population.

Authorities B and C are broadly similar, with a 15.5 per cent and 18.7 per cent share of datazones respectively, athough the population of Authority B is more than double that of Authority C.

If we look more broadly at overall SIMD deprivation (local share) in the three authorities, 6.7 per cent of datazones in Authority A were found in the 15 per cent most deprived data zones in Scotland. In Authority B, 9.8 per cent of data zones fell into the 15 per cent most deprived data zones, and in Authority C, 22.4 per cent of data zones were found in the 15 per cent most deprived data zones. To put these data in context, the (mainland) authority with the largest local share of the 15 per cent most deprived datazones in 2012 was Glasgow City (29.5%), and the authority with the lowest local share of datazones in the 15 per cent most deprived was Aberdeenshire (5%). This places Authorities A and B towards the lowest end of the deprivation scale, and Authority C towards the upper end.

Police recorded crime data provides a more direct measure of crime in the three local authorities. Figure 2.1 below shows the trends in recorded crime rate per 10,000 people (all ages) in the three authorities between 2004/5 and 2013/14. These data show distinct differences in the rate of crime between the three areas; however, there has been a steady decline in the crime rate in all three authorities over the last 10 years. This fluctuates slightly more in Authority C than the other two, although this is largely due to the much smaller population size in that local authority. Effectively, however, the crime drop in all three authorities pre-dates both the roll-out of EEI Intervention in 2008 and the WSA in 2011. We must be cautious, therefore, about making any claims about the impact of WSA and EEI on overall crime rates within local authorities. Indeed, the most recent year of data shows that the falling crime trend has stalled in Authorities A and C, and reversed in B.

Figure 2.1 Crimes recorded per 10,000 people 2005/6 to 2012/13 (all ages)

Source: Scottish Government (2014a) Recorded Crime in Scotland 2013/14 (Table 8)

These data provide some wider insights into the demands placed on local authorities, and show that the variation in overall SIMD deprivation between the three authorities is more marked than the variation in SIMD crime. Unpacking the data further, deprivation in Authority A was most pronounced in terms of geographic access (cost, time and inconvenience to access basic services). In Authority B, deprivation was most pronounced in terms of housing (overcrowding, lack of central heating). By contrast, in Authority C, deprivation was pronounced across all seven domains (income, employment, health, education, housing, access and crime). This observation is important and suggests that the demands placed on Authority C in terms of crime need to be viewed against a broader, multi-facetted backdrop of welfare and material disadvantage.

2.2 Patterns and trends in youth crime

Patterns and trends in youth crime in Scotland can only be measured by administrative data as there is no objective measure of offending over time. Against the backdrop of a fall in recorded crime nationally, analysis of data from the Scottish Offenders Index shows that the number of young people convicted in Scottish courts has also fallen dramatically (Matthews, 2014). The rate of convictions for men aged 16-20 years fell from 9,500 convictions per 100,000 men in 1989, to just over 2,700 convictions per 100,000 in 2012, whilst the average age of conviction for men rose from 18 to 23). Rates of conviction have also fallen for young women aged 16-20 years, although not to the same extent as for young men.

Recorded crime figures show a particularly significant fall in crime from around the mid-2000s, and Matthews' work indicates that the largest drop in convictions amongst young people occurred from around the same period (Matthews, 2014). Indeed, police data shows that offending by young people aged under 18 years has fallen by almost half (45%) (Scottish Government, 2014) since 2008/09. This apparent fall in youth offending coincides with a significant shift in youth justice ethos and practice (mainly driven by a shift in political administration), away from a performance indicator culture focused on tackling youth crime through targeting 'persistent offenders', towards a less politicised environment that recognised the weaknesses of the previous system and the crude definition of 'persistent offenders' (Scottish Government, 2007).

The new SNP Government heralded a set of policies including Getting it Right for Every Child (Scottish Government 2008a) and Preventing Offending by Young People: A Framework for Action (Scottish Government 2008b) which paved the way for increased diversion from both the youth and adult justice systems for under 18s. Thus, the apparent falls in youth crime since the mid-2000s may be largely reflective of shifts in youth justice policy and practice rather than any underlying change in young people's behaviour.

Nevertheless, the extent of the decline indicates that the gap in recorded offending between young people and older people has significantly narrowed when compared to previous eras, so whatever factors have led to the decline for young people have not had the same impact on older members of the population. In this respect, there may be value in drawing on the WSA, and its theoretcial underpinning, to inform other areas of criminal justice policy.

2.3 Referrals to the Children's Reporter

One of the key aims of the WSA is to reduce the number of referrals to the Scottish Children's Reporter Administration ^[3] ( SCRA) on offence grounds. Diverting less serious cases at an early stage also means that a higher proportion of the young people who are referred to the Reporter should require compulsory measures (i.e. they will represent the more serious cases and, in the case of offence referrals, the most 'persistent offenders'). SCRA data can provide important insights in the direction of youth offending and justice in Scotland; however, it needs to read with caution. For example, changes in referrals to SCRA may reflect shifts in youth justice policy or the actions of practitioners and others who make referrals, rather than necessarily indicating that youth offending is going up or down.

Figure 2.2 shows the total number of referrals to the Children's Reporter based on offence and non-offence (care and protection) grounds between 2003/4 and 2013/14 in the three case study areas. It shows that, coinciding with the overall trends in crime described above, referrals to the Reporter fell from around the mid-2000s onwards in all three local authorities.

The local authorities did, however, vary in the number of referrals to the Reporter. Authority B had by far the greatest number overall, while Authorities A and C were much lower, but similar in both number and trends. Nevertheless, the extent of the decrease in referrals between 2003/4 and 2012/13 was broadly similar across the three authorities; over the ten year period, referrals fell by 61 per cent in Authority A; 54 per cent in Authority B; and by 57 per cent in Authority C.

Figure 2.2 Referrals to the Children's Reporter, 2003/4 to 2013/14

Source: Scottish Children's Reporter Administration

Figures 2.3 and 2.4 show the rate of referrals per 1,000 young people in Scotland to the Reporter on offence and non-offence grounds separately between 2003/4 and 2013/14 (i.e. these charts adjust for population size and allow us to compare change over time within the three authorities more directly). Figure 2.3 shows that the rate of referrals on offence grounds dropped sharply from 2004/5 in Authority A; and fell more steadily from 2005/6 in Authorities B and C. The overall fall in the rate of offence referrals over the period was 80 per cent in Authority A, 70 per cent in Authority B and 81 per cent in Authority C. Using standardized data, it is evident that the rate of referrals was much higher in Authority A than in B or C in the early part of the decade, but this difference had all but disappeared by 2011/12 and the three local authorities had almost identical (and much lower) rates of referral on offence grounds by 2013/14.

Figure 2.3 Children referred to the reporter on offence grounds per 1,000 population, 2003/4 to 2013/14

Source: Scottish Children's Reporter Administration

By way of comparison, Figure 2.4 shows that there was no one area with a much larger rate of referral to the Reporter on non-offence grounds, and that the three areas showed a very similar trend and pattern over the 10 year period. The drop in non-offence referrals started slightly later from 2007/8 in Authority A compared to Authorities B and C, which started from 2006/7. However, as with Figure 2.3, by 2013/14 the three areas were barely distinguishable from each other in terms of rate of non-offence referrals, and all were significantly lower than in 2003/4. The overall fall in the rate of referrals on non-offence grounds was much smaller than for offence referrals, at 51 per cent in Authority A, 54 per cent in Authority B and 48 per cent in Authority C. These figures both suggest a shift towards greater consistency in approach to referrals across the three local authorities.

Figure 2.4 Children referred to the reporter on non-offence grounds per 1,000 population, 2003/4 to 2013/14

Source: Scottish Children's Reporter Administration

It is clear looking at Figures 2.3 and 2.4 that the rate of referral on both offence and non-offence grounds stabilised at its lowest point at precisely the time of the roll-out of the WSA in 2011. In other words, the largest declines in referrals had already occurred before WSA was formally implemented, and are most likely to have been influenced by the GIRFEC and Preventing Offending by Young People: A Framework for Action policies which preceded WSA, although this does tie in with the more general approach to EEI which began in 2008.

Coinciding with the fall in referrals to the Children's Reporter over this 10 year period, Figure 2.5 shows an increase in the number of hearings per 1,000 population of young people. So as the number of referrals reduced, the relative number of hearings per person referred increased over the same time period, peaking at around 2008/10 in Authorities B and C, and in 2011/12 in Authority A. Looking across the entire period, the number of hearings per 1,000 population increased from 40 to 48 in Authority A; from 41 to 44 in Authority B; and from 42 to 55 in Authority C.

This suggests that those who are still receiving hearings, although smaller in number, are receiving a higher number of hearings on average than previously. This fits with the hypothesis that those who continue to be referred to the Children's Hearing System are those at the most serious end of the spectrum who are likely to be more in need of compulsory measures of care. Nevertheless, unlike the change in referrals, there appears to be slightly less consistency of approach over time across the three areas.

Figure 2.5 Children's Hearings per 1,000 population, 2003/4 to 2013/14

Source: Scottish Children's Reporter Administration

2.4 Joint referrals to the Procurator Fiscal and Children's Reporter

Figures 2.6 to 2. 8 show the number of joint reports from the police to the Procurator Fiscal and Children's Reporter between 2008/9 and 2013/14 for the three local authority areas. This represents the number of cases that were considered serious enough for potential criminal justice disposal (although in reality, most of these cases prior to WSA would have involved children approaching their 16 ^th birthday). Post 2011, these figures indicate those cases that were referred on offence grounds but not diverted for PRS. These trends appear to paint a less clear picture of the direction of youth justice. Note also that children may be referred more than once.

In contrast to the steady fall in referrals to the Reporter on offending and non-offending grounds (see Figures 2.3 and 2. 4), the data in Figures 2.6 to 2. 8 are more variable.

Figure 2.6 Number of Joint Reports to PF and Reporter (by age at receipt): Authority A

Source: Scottish Children's Reporter Administration

Figure 2.6 shows that in Authority A, the number of joint reports for under-16s were relatively stable other than a sharp spike in 2011/12; the number of joint reports for 16 year olds was a bit more variable, but this also spiked in 2011/12; whilst the number of joint reports for 17 year olds was very low in 2008/9 but gradually increased across the period, showing a particular increase in 2013/14. Overall, the number of joint reports in Authority A was small, and the trends difficult to detect.

Figure 2.7 shows that in Authority B, the number of joint reports for under-16s grew three-fold between 2008/9 and 2010/11, but fell steadily thereafter; for 16 year olds, there was a similar three-fold increase in joint referrals between 2008/9 and 2010/11, but the pattern then varied year on year to 2013/14; whilst the number of joint reports for 17 year olds was very low and remained largely unchanged over the period. Most joint reports in Authority B were made in relation to under-17s, thus suggesting that 17 year olds were more likely to be referred directly to the Fiscal.

Figure 2.7 Number of Joint Reports to PF and Reporter (by age at receipt): Authority B

Source: Scottish Children's Reporter Administration

Figure 2.8 shows that in Authority C, the number of joint reports for under 16 year olds varied year to year but was largely stable over time and showed no clear change in direction; for 16 year olds, the number of joint reports increased slightly between 2008/9 and 2011/12 but then spiked sharply in 2012/13 and fell thereafter, suggesting either a very short term change of policy or a surge in offence reports; whilst the number of joint reports for 17 year olds was very low and stable until 2011/12, but rose considerably in the two years after that. Overall, the use of joint reports was very low until the most recent two years, when there was evidence of some change for 16 and 17 year olds, but more data would be required to detect whether there was a clear pattern.

Taking an overview of the six year period covered, the number of joint reports in Authority A increased from 71 to 123 (76%); from 101 to 151 (46%) in Authority B; and from 73 to 170 (133%) in Authority C. Clear trends in these data are difficult to detect as the numbers are relatively small; however, there does seem to be some increase across the three areas combined since WSA was implemented in joint referrals for 16 and 17 year olds. This fits well with the ethos of maximum diversion and minimal intervention for under 18 year olds

Figure 2.8 Number of Joint Reports to PF and Reporter (by age at receipt): Authority C

Source: Scottish Children's Reporter Administration

Note that for children under 16 "there is a presumption that the child will be referred to the Children's Reporter in relation to the jointly reported offence", whereas for children over 16 years, "there is a presumption that the PF (Procurator Fiscal) will deal with the jointly reported offence" ( COPFS/ SCRA, 2014). The overarching principle directing COPFS decision-making is 'whether it is in the public interest to prosecute the child'. It has been suggested that if children under the age of 18 were referred to the Children's Reporter, this would constitute a progressive step, ensuring that the majority of young offenders are dealt with in a non-adversarial system ( CYCJ, 2014: 18).

2.5 Summary

Taking an overview, this section has shown that there is marked variation in offending levels, geography and urbanisation across the three local authorities. Patterns of recorded crime show a distinct fall over time in all three areas which pre-date either EEI or WSA; however, there is some evidence to suggest that there have been significant falls in youth offending (or at least recording of youth offending) since the mid-2000s which ties in with the development and early implementation of GIRFEC and the Preventing Offending by Young People: A Framework for Action on youth offending. There have been significant falls in referral to the Children's Reporter on both offence and non-offence grounds in all three areas, with the fall in offence referrals being most pronounced. The impact of EEI or WSA on patterns of joint referral to the Procurator Fiscal and Children's Reporter is not entirely clear; however, there are encouraging signs of some increase for 16 and 17 year olds.

Differences in baseline conditions within local authorities are important, and may help to explain the different methodologies and processes adopted by the authorities in implementing WSA and differences in observed outcomes. These preliminary observations also support the need for flexibility within the WSA, rather than a fixed-framework, which allows authorities to adapt to local conditions and demands. However, it is also clear that there have been some changes at national level that appear to have impacted similarly across local authorities. Further work on the national and local impact of specific factors would be of value in this regard.