Publication - Research and analysis

An Evaluation of Football Banning Orders in Scotland

Published: 29 Jul 2011
Part of:

This report presents the findings of an evaluation into the use of football banning order (FBO) legislation in Scotland.

72 page PDF

1.1 MB

72 page PDF

1.1 MB

An Evaluation of Football Banning Orders in Scotland

72 page PDF

1.1 MB


3.1 The latest figures for FBOs on conviction, as provided by the FBO authority on the 7th December 2010, are provided below.

Table 3.1: Criminal Football Banning Orders

British Transport Police 140 107 67 9
Central 21 19 12 5
Dumfries & Galloway 4 4 4 0
Fife 9 8 4 0
Grampian 88 75 41 18
Lothian & Borders 170 140 43 4(2)*
Northern 8 6 3 2
Strathclyde 254 157 126 33 (6)*
Tayside 47 42 22 7 (2)*
TOTAL 741 558 322 88

*Numbers in bracket refer to FBOs that were not request by the police, but which Sheriffs chose to impose on their own initiative

3.2 The police applied for seventy-eight of the granted FBOs, whilst a further 13 FBOs have been granted through civil summary applications by Scottish police forces (predominantly Strathclyde which accounts for 11 of these applications). In total, if one includes these summary applications, as of December 2010, 101 FBOs have been issued in Scotland. Whilst, these figures evidence a steep attrition from initial application through to eventual issue, they nevertheless represent a significant improvement on performance. A little over a year after the introduction of the FBO legislation Scotland, in December 2007, out of 221 completed criminal cases where an FBO application had been made only 8 FBOs were issued (just under 4% of all completed cases). This figure has now risen to just fewer than 16% of all cases, or 28% of all convictions.

3.3 A stronger indication of trends can be gained by looking at in-season figures ( e.g. looking at the proportion of FBOs issued on the back of convictions in a particular football season). Complete, in-season figures were only available for five forces [7]. In the 2007-2008 season, 26% of all applied for FBOs were issued, whilst in the 2009-10 season this figure had risen to 53%

3.4 Superficially, the performance in terms of civil summary applications has been stronger, with all but one civil application being granted (and the one application that wasn't granted was never fully tested in court as the police authority concerned withdrew on cost grounds). However, summary applications were no longer being applied for at the point when this research was being conducted, and the majority of applications all related to one football match (Manchester United versus Rangers in 2008), and were applied for because English FBOs issued on conviction were not enforceable at that point in time in Scotland.

3.5 The most relevant perspective however may be the percentage of secured convictions that lead to FBOs, and here some forces within Scotland clearly significantly outperform others, with - out of the larger volume forces - Strathclyde, Tayside and Grampian converting 31%, 41% and 44% respectively. This compares to a Scottish average of 28%.

3.6 Compared to figures for England and Wales these figures still look very modest, with over 2500 FBO orders having been issued in England and Wales since 2006, and with roughly an average of 1000 new orders being issued every year since the introduction of the amended football banning order legislation in the Football (disorder) (Amendment) Act 2002. Moreover, it has been claimed that nearly 80% of all FBO applications in England and Wales have been successful.

3.7 There are some qualifying points that significantly soften the starkness of this comparison. Aside from differences in the levels and types of football-related violence and disorder, as alluded to in the previous section:

  • The application of FBOs in England appears to have been target-driven and partially pump-primed through Home Office funding. As figures later in this report show, English FBOs are dispersed across clubs and across divisions. Whilst this may to some extent reflect the widespread presence of hooliganism, it also seems at least partially to suggest that FBO applications are driven by police-pursued, club-level quotas (whether these are formal and explicit or not).
  • 36% of FBOs in England and Wales are issued to followers of premier league clubs, compared to 93% in Scotland ( FBOs issued up to the end of the 2009-2010 season). This may reflect some lack of awareness or engagement with FBO legislation by security officials outside the SPL, but the more plausible explanation lies in the comparatively small attendance figures associated with non- SPL clubs. Average attendance figures in the League's first division are typically around two to three thousand, whilst for the second division average attendance figures drop down to the hundreds. In comparison, clubs in England's Championship and first division commonly command attendances that are higher than all but the very largest SPL clubs, with average gates of 20,000 or more for some Championship league clubs. Given the significantly greater number of clubs and higher attendance figures in England and Wales, one would obviously expect a much greater number of FBOs to be issued South of the border.
  • Finally, those aspects of the English legislation that make consideration of an FBO compulsory for any relevant FBO offence, and indeed make the issuing of an FBO a presumption, are also therefore more likely to lead to a higher conversion of FBO applications into ' FBOs issued'. As in Scotland, the majority of FBOs "on conviction" are not targeted at risk supporters engaged in organised violence, but are targeted against ordinary supporters who misbehave perhaps only on one or two occasions. It is precisely in these sorts of FBO cases - that are often considered to be somewhat 'borderline' in terms of meriting an FBO - that the presumptive element of the English legislation may potentially generate significantly greater numbers of issued FBOs.

3.8 However, even if one puts aside the issue of greater size and attendance associated with football in England and Wales, the rate at which FBOs are issued, even for the larger SPL clubs, is still some way below the rates found for English premier division clubs. Whilst rates of banning orders per 1000 supporters for most of these English clubs typically fall within the range of one to two banning orders per 1000 supporters, the rate for SPL clubs is typically lower, though not nearly as low as one might extrapolate based on a quick reading of FBO headline figures. For instance, two clubs, Aberdeen and Dundee United, come close to matching English clubs in terms of their banning order rates.

Table 3.2: Football Club Attendance and FBOs

Unofficial Average Attendance 09-10 FBOs active 09-10 BO rate per 1000 attendance
Rangers 47,564 26 0.5
Celtic 45,582 21 0.5
Hibernian 11,806 3 0.3
Aberdeen 10,461 9 0.9
Dundee United 7,821 6 0.8

3.9 In contrast, the proportion of orders that are issued through civil summary application in Scotland are almost identical to the proportion of banning orders that are based on civil applications in England and Wales (roughly 13% of all FBOs issued). Thus, it might be argued that the component of the FBO regime that is most tightly targeted on risk supporters, works as well in Scotland as it does in England and Wales.

3.10 This might be partially true, particularly in the Strathclyde police force area. However, these figures somewhat flatter, as the research evidenced a steep tailing off in the propensity to use summary orders in Scotland, primarily due to budgetary constraints. Bluntly put.

'there is no money for civil applications in this force"
(Police Respondent 2)

3.11 The research also showed that in key SPL clubs the size of the risk support groups was often quite comparable to the risk groups in the English premier division. For instance, most of Manchester United's 74 active FBOs are civil applications targeting some of its 30 to 40 core risk supporters (with roughly a further 150 'hangers on' or 'peripherals'). Several key SPL clubs have risk groups that can readily match this both in terms of core group and peripheral numbers. The likely adequacy of Scotland's 13 orders (eleven of which fall within one force) on this basis becomes more questionable. However, without the support of individual forces in terms of committing resources, there is nothing that the FBO authority, or individual Football Intelligence Officers ( FIOs), can be expected to do to increase this number.

The quality of FBO targeting and decision making
3.12 A comparative review of numbers can only tell us so much, particularly - as is touched on in the English section of this report - because the FBO numbers in England and Wales appear partially to be driven by a 'numbers game' which equates more FBOs with more appropriate or more effective FBO practices. To assess this more qualitative aspect of performance we need to look in more detail at who is targeted and how. To do this, the evaluation looked at:

  • The criminal records of a sample of individuals, all of whom had been convicted of a football related offence and considered for an FBO, though only some of whom had actually been issued with an FBO.
  • A similar paired sample of FBO, and non- FBO football related offenders, this time looking at electronic case records held by Crown Office. Our primary interest was the police report section and the police remarks section as well the court instructions given to the depute handling the case in court.

Comparing FBO and non- FBO convictions
3.13 In interviews, a number of cases were raised as examples of instances where Sheriffs had not issued FBOs to potential recipients. Complaints were also made more generally about inconsistencies in outcomes, including cases where trivial incidents led to an FBO. Undoubtedly there have been cases where individuals arguably deserving of FBOs have, for some reason, not been issued them. One case that perplexed most respondents, including Fiscals and Sheriffs, was that of a fan with a previous football-related conviction who set light to a rival supporter's fancy dress costume on a crowded train on the way back from a match, seriously injuring him and endangering the lives of others (many of whom were dressed in similar flammable costumes). Regardless of intent, it seemed questionable to many that a man who clearly has a record of misbehaviour within the context of football matches, escaped an FBO when as FBO was applied for. This however, is an extreme case, and to examine whether FBOs are, or are not, applied appropriately, we looked at Crown Office records (see paragraph 5.7 for further discussion of this case).

3.14 Police case descriptions were coded so that we could identify those cases - based on a close reading of the legislative guidance - that potentially merited an FBO at least in terms of the presenting aspects of the case. We also coded those that appeared so minor or trivial that the appropriateness of an FBO was questionable. These cases were rated as 'marginal':

  • Of the 20 cases examined where FBOs were issued, only one incident was rated as marginal.
  • Of the 39 cases examined where FBOs were not issued, 12 incidents may be regarded as marginal, and in a further two cases the police had not supplied information that in any way evidenced the football-relatedness of the offence.
  • In contrast, 6 of the 20 FBO cases (30%) either involved incidents of group-based, organised violence (4 cases), or serious violence of some other form (2 cases). Of the non- FBO group, 11 out of 39 cases (28%) also involved what appeared to be group-based organised violence (5 cases) or other serious incidents of violence (6 cases).
  • The other key difference between these two samples was that the non- FBO sample had a higher proportion of offences which had a sectarian or racial element (14 sectarian, and 4 racial, when combined 46% of the sample), compared to the FBO sample (6 sectarian and 1 racial, when combined 35% of the sample).

3.15 These findings are mixed. On the one hand there are, as one would hope and expect, more marginal cases in the non- FBO group. Whilst these cases may have received an FBO in England and Wales, it is not particularly troubling, and indeed it may be considered desirable that these cases are selected out during the course of the judicial process. On the other hand, the findings also show serious incidents being missed in terms of the issue of FBOs. Finally, it would appear that offences that have a sectarian or racial element to them are less likely to attract an FBO than those that don't. At one level, it might be argued that these offences are comparatively trivial when compared to some of the serious incidents of violence that attract an FBO. However, given the context of sectarian abuse within Scottish football, and given that the Scottish FBO legislation was framed in order, explicitly to target such abuse, this finding is problematic. The picture becomes even starker if we examine our larger set of FBO and non- FBO convictions. Here convictions for sectarian and racial offences make up 41% of non- FBO cases, but only 19% of FBO cases.

Number of previous convictions
3.16 Another way to assess the appropriateness of FBO targeting is to look at the prior convictions of offenders with, and without, an FBO. To do this we looked at our criminal record sample. The number and nature of the convictions received by members of the two groups was a key question for the analysis.

Table 3.3: Pre-Convictions for FBO and non- FBO Groups

Percentage* of each group previously convicted of:
Any crime Violence (inc. domestic and football related) Football related Breach of the peace (including football-related)
FBO Group 76 (66) 43 (37) 33 (29) 43 (37)
Non- FBO Group 46 (32) 23 (16) 4 (3) 18 (13)
Two sample test of proportions** p<0.001 p<0.05 p<0.001 p<0.01

* Numbers in brackets
** This is a test of statistical significance, e.g. determining whether the observed differences between the FBO and Non- FBO group can be judged as significant and probably not due to chance. All the comparisons here, are significant at the 0.05 level or below (the conventional threshold of significance in social science statistics).

3.17 Table 3.3 shows the pattern of convictions across the two groups prior to the first day of the order. It clearly shows that the FBO group were considerably more likely to have not only any conviction (76 per cent compared with 46 per cent), but were also more likely to have been convicted of a violent offence, a football related offence and/or a breach of the peace. Note that the time periods covered are not the same for both groups - this is because FBOs often did not commence on the day of the trial, so it was necessary to add a correction to take this into account. One implication here is that those given FBOs were more likely than those not given them to have had a previous criminal conviction at the time the decision was made. Methodological caveats aside, this finding is clearly encouraging and suggests that broadly, many of the appropriate individuals are being targeted by the legislation.

Quality of written evidence
3.18 Whilst FBO targeting may broadly look quite appropriate, there clearly are cases where FBOs have not been issued, where both the case details and offender's prior conviction history would strongly suggest that an FBO was appropriate. We examined our sample of Crown Office records to see if there were any differences in the quality of written submissions between FBO and non- FBO cases.

  • One key difference is that FBO cases were almost uniformly supported by well-written, and often quite extensive police reports, which summarised the offence for the Procurator Fiscal depute. They were also supported by material in the follow-on 'police remarks' section that often went beyond making a standard request for an FBO, but which reinforced the impact or gravity of the offence: sometimes describing how an FBO would help prevent future offending; sometimes directing the Sheriff to guidance on FBO legislation enclosed in submitted papers; and often directly alluding to the offenders' prior football-related convictions, or their status as a known risk supporter. There were a small number of high-profile cases where no remarks were entered on the electronic record, but in each of these cases it is known that the police had face-to-face liaison with Procurator Fiscal deputes to prepare these cases in detail.
  • In contrast, 14 (36%) of the non- FBO cases were poorly supported either by police reports, or by police remarks. Typically, the football-related nature of the offence wasn't emphasised, or the remarks section failed to reinforce the impact or context of the offence ( e.g. why a sectarian offence may be more serious than it appears to the untutored eye within a particular context).
  • Finally, there were clear differences between the two samples in terms of written instructions for Procurator Fiscal deputes. In only 6 of the FBO cases (30%) were no clear instructions provided. Many of the instructions that were provided also went beyond the basic re-iteration that an FBO should be asked for, but also again re-emphasised the impact of the offence. In contrast, in 29 out of 39 non- FBO cases (74%) the FBO was not supported by written court instructions. In some instances, bail was supported but an FBO was not. In six cases there were no court instructions at all.
  • It is important to note though that this is only written material from court papers, and the absence of an instruction does not mean that an FBO wasn't in fact asked for by the Procurator Fiscal depute. However, one must presume, and interview evidence would back this up, that in a number of cases Fiscal deputes aren't asking for FBOs even though the police have requested them. Indeed in one case here an FBO for a sectarian offence was explicitly discounted in the written instructions on the grounds that the FBO legislation did not cover such offences, clearly indicating that the knowledge and awareness of FBO legislation is not always what it might be.

The effectiveness of issued FBOs
3.19 This leads to the question of whether issued FBOs are effective in preventing future football-related offending. Technically, this is a difficult question to answer definitively, though we will present what figures we can below. There was also ample anecdotal evidence from respondents on the effectiveness of orders, both in terms of impacting on those subject to them, and in terms of deterring others. For example as one respondent observed when talking to a banned supporter:

"...he basically says, 'that's my life, take anything away from me but don't take away my ability to go to a football match'. And that individual he had bail conditions that lasted for about six months, and he openly admitted that was like the hardest six months of his life"

(Police Respondent 5)

3.20 Taking again the FBO and non- FBO conviction samples, we looked at convictions after the first day of the FBO/day of the trial. Because the motivation behind the analysis was to examine whether receiving an FBO had an influence on subsequent offending, a six month delay was added in each case, making it more likely that the offence tried was committed after the FBO was granted (or the 'non- FBO' trial was over - note that missing data reduced the sample size to 69 in this case). In the event, it appeared that the FBO group were more likely to have been found guilty after receipt of an order than were the non- FBO group after their trial - but only in relation to football related offences (none of the non- FBO group were convicted of a football related offence after the original trial). For 'all offences', violence and breach of the peace there was little difference between the groups. This finding is perhaps not surprising given the much higher levels of football related offending among the FBO group prior to receiving an order (see Table 3.3). A higher rate of re-offending might have been expected on the basis of their prior conviction history, but this does also suggest that orders are not fully effective at preventing or deterring recipients from re-offending.

Table 3.4: Post-Convictions for FBO and non- FBO Groups

Percentage* of each group convicted
6 months or more
after initial conviction
Any crime Violence (inc. domestic and football related) Football related Breach of the peace (including football related)
FBO Group 28 (24) 7 (6) 9 (8) 9 (8)
Non- FBO Group 18 (13) 7 (5) 0 4 (3)
Two sample test of proportions p=0.19 p=0.94 p<0.05 p=0.23

* Numbers in brackets

3.21 Tables 3.3 and 3.4 show the proportion of each group found guilty of at least one offence of each type. Another way to slice the data is to look at the average number of guilty verdicts received for each group, before and after the FBO commenced compared with before and after the non- FBO trial. Table 3.5 shows that, the FBO group appeared to be the more prolific offenders before receiving an order, with an average of 6.1 offences per person compared with 2.0 for the non- FBO group. By contrast, the difference between the groups was much smaller (and non-significant) after the granting of the order/day of the trial. This finding needs to be treated very cautiously, but it may indicate that the FBOs have some influence in decreasing the intensity of offending among those who receive them, even if these individuals are still more likely to go on to commit offences than those considered for but not given an order [8].

Table 3.5: Average number of convictions for FBO and non- FBO Groups

Prior to first date of FBO/date of trial at which FBO could have been issued After first date of FBO/date of trial at which FBO could have been issued
FBO group 6.1 0.6
Non- FBO group 2.0 0.5
T-test, independent samples* p<0.001 p=0.42

* A test of statistical significance.
Note: Averages calculated including those with zero convictions. 'After' figures calculated excluding offences tried within 6 months of FBO commencing/non- FBO trial date

The squeezed middle?
3.22 One final aspect of orders, that hasn't been examined in detail to date, is the conditions associated with those orders. Beyond a basic ban on attending football matches, or individual stadia, or the standard additional ban on travelling abroad during certain match periods, what other measures are put in place to control the behaviour of misbehaving supporters? Here the contrast with England and Wales is at its most stark, because the answer to date for Scotland is that with literally two or three exceptions, additional conditions have not been imposed on any convicted FBO-recipient.

3.23 If an individual is passionate about attending football, or only misbehaves within football stadia, then a ban on attending matches may be effective either in terms direct preventative control, or in terms of deterring that individual from misbehaving away from the stadia in the future. But, if the individual concerned, as is characteristic of a significant number of risk supporters (though by no means all), has no interest in attending the match, but rather uses match days and bonding over club-allegiances as a way of accessing opportunities for recreational violence well away from the stadia, then FBOs without conditions (for instance banning them from city centre pubs on match days, or stopping them from visiting towns, and going on trains to towns, where 'their team' are playing away) are going to have no plausible impact.

"no impact on them whatsoever, because they can't go to the football match. So what! They can still go to the city centre, they can still meet with their chums in the bar, he can still travel"

(Police Respondent 4)

"For me the banning order legislation is fine in its own right, however, the risk supporters won't fight at the stadium they will fight outwith the stadium, so me for the legislation's got to go further and to ban them from a town or city where that club is playing"

(Police Respondent 3)

3.24 Interviews with Sheriffs and Procurator Fiscal deputes have evidenced why gaining such conditions may be difficult. As it stands these professional groups are demanding of clear evidence that incidents away from stadia are football related. It may be that they require an appropriate level of linkage or that there is a lack of awareness of the dynamics of this sort of violence. As one respondent noted:

"If these people really wanted to engage in violent disorder, in those sorts of numbers why not do it this afternoon? Why not go up to ____ Park this afternoon and, er, one group approach from the West and one from the East at twelve o'clock today? Why does that not happen? Why does it have to be a football day?"

(Police Respondent 5)

3.25 However, the case for conditions has rarely been tested, as officers themselves have admitted that they very rarely try and ask for conditions. In their view getting a basic FBO is hard enough without trying to argue for more restrictions. This is supported by our sample of Crown Office data, where only one out of fifty nine cases included a request for additional conditions. But these restrictions are enshrined in the legislation, and their inclusion was based precisely on the anticipated need to deal with organised violence away from the ground.

3.26 In England and Wales conditions have been pursued and pushed - often at a club level based on well-evidenced knowledge of out-of-stadia trouble spots associated with match days - through the pursuit of civil banning orders. In Scotland the data analysed here would suggest that it is neither the sectarian offender nor risk groups - the two primary targets for the Scottish legislation - who are most likely to be affected by that legislation. Rather, it is in fact ordinary fans in the middle who commit acts of violence and disorder within or immediately around stadia, who are arguably most likely to feel its full force. This is not to put forward the view that this group don't fully deserve banning orders when committing often quite serious criminal acts, but it is worth considering whether more needs to be done to improve the 'reach' and effectiveness of the legislation in terms of these other groups.