The UK sex offender register: Has the net been cast too wide?

Abstract

The UK sex offender register is integral to the public protection system, albeit an elusive tool to evaluate. This project analysed qualitative data from 64 professionals from criminal justice agencies and examined how the register effectively protects the public, and, more pertinently for the purposes of this article, how it could be made more productive. A recurring theme was its incremental growth – an inevitable trend given the length of registration. Participants explained the difficulties in monitoring an ever-increasing population with static resources and expressed doubts that it was purposeful or proportionate to cast the net so wide. Various solutions were suggested, including relaxation of current rules governing indefinite registration, and reducing or removing registration for certain cases, particularly those convicted for first time minor sexual assaults or possession of indecent images of children.

Introduction

The UK sex offender register (SOR) has existed for over a quarter of a century and is a well-established part of the public protection machinery. This project examined the views of professionals involved with managing the register or who work with individuals who are subject to its requirements. It asked what aspects they considered to work well, and what aspects seemed to be less productive, or even counter-productive. To set the scene, it is useful to recall what drove the creation of the UK SOR, to review how it works and then to consider existing research on the impact of registry systems and the views of professionals who work in this area.

The creation of the UK SOR

The UK’s SOR was planned by the Conservative government in the 1990s and implemented when a Labour government came to power in 1997. America already had registers, with California and Arizona creating theirs in the 1940s, and all US states being legally required to keep one since 1994. Among its European neighbours, however, the United Kingdom was an early adopter.

The impetus for the creation of the UK SOR came from various directions. British social workers had been calling for a register of child abusers, passing a motion at the 1988 annual conference of the British Association of Social Workers to that effect (Thomas, 2004). They argued that registers of children at risk of abuse had been kept since the death of Maria Colwell in 1974, so the same should be done for the abusers of children. There had also been intense media focus on certain sexual murders of children during the 1990s; Rose and Fred West tortured, sexually abused and murdered at least 12 women including two of their own children (they received life sentences in 1995), 9-year-old Daniel Handley had been abducted, abused and killed (his murderers received life sentences in 1996) and 7-year-old Sophie Hook was abducted from the tent she was sleeping in, raped and murdered (her killer was sentenced to life in 1996). While such horrific cases had always rarely occurred, the 1990s saw a particularly febrile reaction, including demonstrations and an attack on a paediatrician’s home (the culprits failed to distinguish between the word ‘paediatrician’ and ‘paedophile’), all of which led Nash (1999) to refer the 1990s as ‘the decade of the predatory sex offender’ (p. 45).

The Home Office consultation on the proposed register received 238 responses of which 87% were in favour (Thomas, 2004) and all major political parties were in support. When the Sex Offenders Bill was debated in parliament (Hansard, 1997), most speakers sought reassurance that the register would be robust enough, querying whether it would include those who offended against adults as well as children, or deal with offences committed abroad as well as domestically. There were dissenting voices, but they were fewer and less noisy. The civil liberties campaign group Liberty disputed that it would achieve what it set out to do:

No evidence is given, however, to support any of the supposed benefits of a register, nor is there any suggestion of how the proposed register would achieve the claimed objectives. (Liberty, 1996: paragraph 27)

Matthew Parris (1997) (a former Member of Parliament and now a journalist) objected on the grounds of stigma and proportionality, and questioned whether:

. . . a great swathe of offenders who have served their sentences and returned – disgraced at the workplace, disgraced with landlords, disgraced among their families and their friends, and almost certainly unemployed – to try to begin their lives again, should now be bound to the 20th-century equivalent of the leper’s bell. (The Times, 24 January 1997)

The criminological literature tended to view registry systems as emblematic of the dominant paradigms of the day. Garland (2001) saw them as just one aspect of a ‘culture of control’, a concept he had developed to describe the late 20th-century shift away from the penal welfarism that had dominated criminal justice policy after the second world war, to a more punitive era in which risk was expected to be controlled. Garland’s idea fitted well with sociologist Ulrich Beck’s (1992) notion of a ‘risk society’, which proposed that the mid-20th-century promise of safety, progress and expert knowledge had failed and resulted in a late modernity beset by amorphous anxiety, where the elimination of life’s inevitable hazards was demanded. Pratt (2007) expanded on a similar theme to Garland, using the term ‘penal populism’ to stress the public appetite for more punitive, controlling measures. Seen through a more pragmatic prism, the drivers behind the creation of the SOR could be seen as a complex mix of serious planning to protect the vulnerable, a gesture to assuage a vengeful public, and political signalling that neither major party was soft on crime – a general election was imminent.

Thus, the register was enacted by the 1997 Sex Offenders Act and later incorporated into the 2003 Sexual Offences Act. In fact, the phrase ‘sex offender register’ is not mentioned in the legislation; it is routinely referred to as such but the legal terminology is the ‘requirement to notify’. The main aim of the UK register was knowing where those convicted of a sexual offence were living, and how to find them. Three ancillary aims were suggested, but tentatively, namely that it might prevent crime, help to identify suspects after a crime was committed and deter re-offending (Home Office, 1996). It was explicit that a register should not be punishment, but ‘a measure aimed at helping to protect the community from sex offenders not an additional penalty for the offender’ (Home Office/Scottish Executive, 2001: 11).

Initially, registered sexual offenders (RSOs) were required to do little more than annually provide their name, aliases, date of birth and address and report any changes. Regulations have been modified many times since then, for instance to increase sanctions for non-compliance (from 6 months to 5 years in the Criminal Justice and Court Services Act, 2000), to decrease the window for initial registration (reduced from 14 days to 3 days in the same Act) and to increase the information RSOs are required to disclose (the 2012 amendment to the 2003 Sexual Offences Act added information such as national insurance number, bank/credit card/passport details). The principle that the SOR does not impose an extra layer of punishment has been questioned over the years, with some arguing that mission creep has taken it beyond a tipping point, and made it a distinctly punitive tool (Kingston and Thomas, 2018; Thomas, 2008). It has equally been argued that measures that strengthen public protection are inevitably also punitive, and that it is unrealistic to expect otherwise (Fitzgibbon and Lea, 2020).

How the UK SOR operates

A conviction for an offence on Schedule 3 of the 2003 Sexual Offences Act (a comprehensive list of sexual offences) makes an individual an RSO. There are some thresholds in Schedule 3, so lesser sexual offences must receive a sentence of a certain gravity to require notification (just a few for adults, more for under 18-year-olds). Schedule 5 of the same Act introduced a list of offences that are not sexual, but which can make the offender eligible for a civil order called a sexual harm prevention order (SHPO) if there is a sexual aspect to their commission (e.g. a burglary with sexual elements). The existence of a SHPO triggers notification requirements, so in a roundabout way it is possible to become a RSO through the commission of a non-sexual crime.

RSOs must attend a police station within 3 days of sentence or release from custody and provide the necessary information. They must then do this annually and notify changes that occur during the year (e.g. a new bank card or passport), any intention to travel abroad, or an intention to spend more than 12 hours at an address where there is someone under 18.

RSOs are also visited at home by their allocated Public Protection Officer (PPO), a police officer who works in a specialist Public Protection Unit (PPU). Frequency of visits is determined by the individual’s risk and may be very frequent for high-risk RSOs, and not at all for a minority of low-risk offenders. RSOs are not legally required to accept visits and PPOs have no automatic right of entry, but S.96b of the 2003 Sexual Offences Act allows warrants to be applied for if two attempts to visit have been unsuccessful.

The duration of registration is determined by sentence, ranging from 2 years for a caution to indefinite registration for prison sentences of 30 months or more. Other durations are staggered between these two points. For under 18s registration periods are halved, with the exception of indefinite registration requirements. Since 2011 those required to notify indefinitely can apply for removal after 15 years, a change brought about after the Supreme Court ruled that indefinite registration was incompatible with Article 8 of the 1998 Human Rights Act (Supreme Court, 2010).

The impact of registry systems

It is straightforward to explain how the UK SOR works, but less so to establish the impact registry systems have on offending, as comparisons of sexual offending rates before and after the introduction of a register are beset by confounding variables. First, the range of codified sexual offences generally increases over time (e.g. in the United Kingdom an offence of grooming was introduced in 2003 and in 2019 the Voyeurism Act made ‘upskirting’ an offence). Second, the last two decades have seen the Internet expand opportunities for sexual offending, and convictions for online offending have increased many times over (Skidmore et al., 2022). Third, attitudes towards consent and disclosure have changed markedly, something which is evident when the Crime Survey for England and Wales (an annual anonymous survey of 50,000 participants) is contrasted with measures of recorded crime (i.e. incidents that are reported to and recorded by the police). Whereas the former has hardly changed since 2005, recorded crimes have increased by over 300% over the same period (ONS, 2022). A further caveat relates specifically to the impact of registration on recidivism; reconviction rates for sexual offences have always been low compared to other types of offences, so changes do not stand out statistically. It should also be noted that most empirical research is American, and so conclusions do not generalise easily; all American states are required to allow some form of public access to registry information (not so in the United Kingdom), and many states impose residence restrictions that lead to levels of homelessness that are more extreme than in the United Kingdom (Savage and Windsor, 2018).

Nevertheless, attempts to establish the differences in sexual offending before and after the introduction of registry systems are available, and in sum, the evidence is mixed. The most recent meta-analysis of 18 studies concluded that registration had no impact on recidivism (Zgoba and Mitchell, 2021).

Professionals’ views of registry systems

There are several pieces of existing research that pursue a similar research question to our own and explore professionals’ opinions of registry systems. Some of the larger projects (e.g. Harris et al., 2018) study American registers and so will not be considered further here because of the differences from the UK system.

UK research has largely focused on the views of PPOs, and concludes that they see the SOR as a valuable public protection tool but identify a range of challenges. Nash (2016) interviewed and surveyed 21 PPOs and found a consensus that the SOR performed a positive role in public protection, but probably not as great a role as the public imagined or expected. Nash suggests it is ‘regarded as more of a public reassurance exercise . . . under-resourced and under-regarded’ (p. 424). Nash then focused on the way that PPOs were treated negatively by their police colleagues; the term ‘scum-cuddlers’ that features in the title vividly illustrates the labels that were attached to them (it perhaps reflects the passage of time that no such negative views of PPU work were expressed by our participants).

Two recent large-scale UK projects are of particular note. First, McCartan et al. (2018) surveyed 227 police officers and then interviewed 27 of them. Participants were generally supportive of the register but identified logistical problems with its operation, in particular the systems for recording and exchanging information, the operation of the child sexual offender disclosure scheme (whereby police can disclose information regarding sexual convictions upon application from a member of the public who has a legitimate reason to know), and the lack of resources. Second, the ‘Independent Review of Police Management of Registered Sexual Offenders in the Community’ (Creedon, 2023) is significant – not a single piece of research but a wide-ranging review commissioned by the UK Home Secretary in 2022. Creedon found broad support from police for notification requirements, but noted that ‘the international and academic evidence is hardly supportive of registration schemes’ (p. 25). Creedon goes on to argue that the size of the SOR is unsustainable; his proposals are strikingly similar to our own and will be returned to.

Despite the richness and currency of data produced by these two projects, they drew exclusively on police perspectives, and while the police are central to the operation of registry systems, this project differed in that it gathered views from a diverse range of professionals, each of whom had their own opinions and experiences of the SOR.

Methodology

Participants

Ethical approval was granted by Buckinghamshire New University, and a range of criminal justice organisations were approached for access to staff who had expertise and experience in administering the SOR or working with RSOs. Once permission had been granted by various agencies, information about the project and invitations to participate were circulated by managers who acted as gatekeeper to ensure that participation was voluntary. The only exception was the solicitor who was approached directly. A purposive sample of 60 was originally aimed for to ensure depth and breadth of data, but all 64 who came forward were interviewed in the interests of arriving at a point of saturation. The sample comprised:

• 39 police officers from two forces in the south of England, one urban and one partly urban and rural. Police participants were of varying rank, with 17 working in PPUs and the other 22 working in various specialisms (e.g. neighbourhood policing, criminal investigations, child abuse investigation teams).

• 12 probation officers from two areas, one urban and the other partly urban and rural. All probation participants held specialist sexual offending roles or had done in the past.

• 4 youth justice workers from 3 local authorities in the south of England. They were either currently working with under 18-year-olds subject to notification or had done in the past.

• 3 clinicians from two residential treatment centres for young people with harmful sexual behaviours.

• 3 clinicians from the Lucy Faithfull Foundation (a UK wide charity that provides services to sexual offenders, their victims, and the organisations that work with them).

• 2 workers from CirclesUK (a charity that operates across parts of the United Kingdom, known best for its circles of support and accountability, but which also offers services to families and organisations).

• a solicitor specialising in defending individuals charged with indecent images of children offences.

No incentives were offered, but after the initial data analysis, a half day event was held to which all participants were invited, and their views on early findings were sought.

The possibility of including the views of RSOs was considered at the outset but not pursued. Although some researchers have been successful in including individuals with convictions for sexual offences as research participants (e.g. Mann et al., 2021) they are a difficult group to make contact with and recruit. Moreover, such research has already provided in-depth accounts of the experience of living on the register, largely characterised by stigmatisation and alienation from family and society. This project sought the views of those who have an overall view of the system, rather than a singular biographical perspective.

With any volunteer sample there is always the possibility of ‘volunteer bias’, that is, that the views of those who participate might differ from those who choose not to. While not excluding the possibility, participants’ comments suggested that their motivation to take part stemmed from a sense that the SOR was misunderstood and under-examined, and that participation might help to redress this.

Data collection

A qualitative approach was adopted, as the paramount aim was to gather detailed accounts of the functioning of the SOR. Interviews were semi-structured to ensure that all participants were invited to comment on common aspects of the research question, but that there was also freedom for contributions to go in diverse directions. Interview schedules differed slightly between groups of professionals to reflect the variations in their roles. Interviews were conducted remotely, with participants asked at the outset to avoid disclosing any details that could identify particular individuals. Interviews were recorded, ranging from 58 minutes to 2 hours 24 minutes in length, making a total of over 100 hours of interviews.

Data analysis

Once transcribed, the material was thematically analysed in accordance with Braun and Clarke’s (2006) steps; after familiarisation with the data, codes were annotated and then organised into coherent themes and sub-themes which were reviewed and named. While Braun and Clarke (2024) acknowledge that thematic analysis is a family of approaches rather than a set of rules, their twin warnings against falling into positivist styles of thinking and arriving at topic headings rather than themes were borne in mind. Coding and identification of themes was conducted jointly by the authors and discussed as the processes were followed, thus avoiding the more quantitative notion of inter-rater reliability and allowing a more reflexive exploration of the data and its alternative meanings. This also supported the authors’ consideration of their positionality which could have pulled them in a variety of directions (e.g. as ex-probation officers who have experienced the sense of responsibility for public protection or as academics with awareness of empirical evidence that doubts the effectiveness of registry systems).

Findings and discussion

It was clear from the outset that the professionals in our sample saw the register as essential. While this might be expected from our police participants, who might be biased towards a process that is effectively owned by them, it was a view overwhelmingly expressed by participants from all agencies. There were numerous examples of PPOs establishing supportive relationships with RSOs and sometimes protecting them against community vigilantism, and of unannounced home visits revealing signs of elevated risk and sometimes evidence that led to prosecutions. There were also accounts of productive working relationship within police forces and between police and other agencies. Equally, some aspects of the SOR’s operation were criticised, for instance, weaknesses in inter-agency communication, the pointlessness of some information that had to be disclosed, the use of trivial technicalities to breach, the potential for RSOs being ‘outed’ while notifying in person at a police station, and the occasional heavy-handed disclosure of RSO status to neighbours and potential employers.

However, one issue stood out above all others as a fundamental, systemic problem, and that was the steady increase in the number of RSOs while resources stayed static:

It has been an extraordinary shift . . . having seen this from day one from ’97 when I first started I had 12 offenders to manage . . . I’ve now got 90 in just my little [area] . . . it’s mushroomed, you’ve got too many offenders to manage and we’ve been pushed into a corner. It feels sometimes it’s a tick box exercise as opposed to actual effective management. (Public Protection Officer)

The register is increasing at a rate beyond the resourcing of the PPU . . . the intimacy of the knowledge is decreasing . . . they will deal with what they can deal with, and that will typically be the top 10%, but when that 10% doubles in size are they then dealing with the top 5% in 2 years’ time? (Criminal Investigations Police Officer specialising in domestic violence)

This issue is indeed a real one – the number of RSOs in England and Wales grows every year by around 2000. In 2002, there were ‘approximately 27,500’ RSOs (National Probation Service, 2004) but by March 2024 there were 70,052 (Ministry of Justice, 2024). The trend has long been recognised, with Thomas (2009) arguing that the register would become silted up with low-risk offenders, and that a review process would be needed to focus resources. The issue is compounded by the numbers of RSOs subject to SHPOs. These are civil orders that can be taken out on individuals convicted of a sexual offence and which contain a list of activities that must or must not be done. After their creation in 1998 (known then as sex offender orders) they were rare, with only 170 issued over their first 4 years (Kingston and Thomas, 2018). More recently however they have become routinely used, to the extent that a total of 6093 fresh SHPOs were issued between March 2023–2024 (Ministry of Justice, 2024). Monitoring SHPO conditions has greatly expanded the demands on the police workforce, and not necessarily in proportion to the risk presented; several police participants noted that an online sexual offence (of any gravity) would almost guarantee the issuing of a SHPO (thus allowing Internet usage to be monitored), whereas a serious contact offence would be less likely to.

The problem of this incremental growth in police workload has been noted elsewhere. McCartan et al (2018) concluded that there were just too many RSOs to monitor given the available resources. The Creedon (2023) Report states frankly that ‘the current . . . process of sex offender management is not sustainable nor efficient’ (p. 13). Like a bus with more passengers getting on than getting off, the number is guaranteed to grow – those who are added remain for substantial periods of time and exceed the number coming off. Cohen’s (1985) notion of ‘netwidening’ described criminal justice initiatives intended to divert and rehabilitate, but which have the perverse and unintended consequence of drawing ever more people into the system. While the SOR does not fit that definition in some ways (it was never intended to rehabilitate or divert) in the sense that it keeps people inside an ever-expanding net, the term fits.

Although a minority of participants believed that all RSOs needed to be registered for current durations, the majority were of the view that not all merited such weighty measures and that the net was being cast too wide, diverting resources away from those who posed significant risk:

I think that there are a group of sex offenders who definitely need more oversight, a huge multi-agency approach. But there’s a group who definitely don’t . . . they’re really unlikely to reoffend. (Probation Officer)

There was considerable discussion around exactly who should be allowed to register for shorter periods of time, or even not to register at all. They broadly fell into three groups.

The first group comprised those whose sentences were 30 months or more in custody, and so are required to notify indefinitely. Currently applications for removal can be made after 15 years, but the case was made for relaxing this process. There was a good deal of support among our participants for reducing the 15-year threshold to 10 years (this is also a recommendation of the Creedon Report). A further suggestion was the presumption of removal on application at 15 years if there was no compelling reason to stay on, rather than the current requirement for a lengthy application with evidence:

I definitely think there needs to be a review of the length of time on the register . . . the review should be about not ‘should we take them off?’ but rather ‘should we keep them on?’. At the moment, the amount of work it takes to consider taking them off is huge, so we are just not doing it. But if it was a case of ‘this review is coming up, what have they done in the last 15 years to keep them on?’ (Public Protection Officer)

A number of our police participants supported this move, both because it would reduce total numbers somewhat, and because lifetime registration was viewed as excessive and disproportionate when applied to all of those sentenced to over 30 months custody. A total of 357 applications for removal were approved in England and Wales between April 2023 and March 24 (Ministry of Justice, 2024); what proportion of indefinite registrants this represents and whether there is scope to increase this figure cannot be known – the number of indefinite registrants among the total complement is not publicly available, and possibly not recorded.

It was further suggested that denial of offence should not be a barrier to removal:

I had one that I had approved and it was rejected because he never admitted his guilt. So, the argument was that he hadn’t done a sex offender treatment programme . . . and he can reapply in eight years’ time – what’s going to change in eight years’ time? He’s still going to deny his offence. (Public Protection Officer)

The second suggestion was that there should be a mechanism that allowed individuals convicted of some sexual assaults or exposure offences to either avoid notification altogether, or to have their period of registration reduced. Sexual assaults when committed drunk during a night out, often in a pub, nightclub or town centre were frequently mentioned:

With some of the sexual assaults . . . you could look at the threshold there, maybe raise that and take a few of them off because you have people that have a great time in a nightclub on a night out . . . and then they end up on the register for five years. And that seems a little bit harsh compared to some of the other people that are on it. (Public Protection Officer)

Similarly, we heard of sexual assault convictions being gathered by individuals with substance misuse and/or mental health problems, who became involved in altercations (often during arrest) and grabbed at breasts or genitals (frequently a police officer’s). A variation of this theme was the drunken public urinator who ended up with a conviction for exposure. The consensus of the police participants was that these types of sexual assault did not necessarily indicate a risk of future sexual harm, and that they could more accurately be defined as public order offences. It was seen as disproportionate to have these individuals on the register, and a poor use of police resources.

Some police participants also mentioned individuals who had been convicted of historic sexual offences, often as younger men, as a group who did not necessarily merit a requirement to notify. Their argument was that their sentences were justified, but if 20 or 30 years had passed since the offence had taken place and there was nothing to indicate subsequent offending, there was little logic in starting to monitor them decades later. Conversely, a small number observed that it would be an affront to victims if historic sexual offenders had different registration requirements, but this suggested a view of registration as marking and punishment (which is not the stated intention of the register) rather than a risk-based public protection measure.

The third suggestion was that some who are convicted for the first time of possessing indecent images of children (IIOC), also known as child sexual abuse material (CSAM) were being required to register for excessive durations. This was the group where opinions were most strongly held, and most divided. PPOs consistently estimated that the number of RSOs convicted of online offences equalled those whose offences were not committed online. However, this is a disparate group, including those who possess, produce and distribute CSAM (including live streamers) and online groomers. There was no doubt in our participants’ minds that most of these offenders needed to register for current durations. However, some participants were unconvinced that first time offenders who had possessed IIOC (and not produced or distributed it) should be required to notify at all, or at least not for so long. They had several reasons for arguing this point.

The first was their belief that first time IIOC offenders presented a very low risk of reconviction and in particular a low risk of escalation to a contact offence. We heard numerous comments to this effect:

a lot of our indecent images offenders I would say confidently wouldn’t go on to reoffend. (Probation Officer)

There was a recognition that views around escalation from non-contact to contact offences has evolved over time:

25 years [ago] I think when the police started dealing with these individuals, their view was very clear. Viewing CSAM is indicative of paedophilia, paedophilia is indicative of child molestation. Try harder and you’ll find that he’s a contact offender . . . in the last 10 years they have begun to deal with so many suicides and so many ordinary men where they absolutely do not recognize the personality type as the archetypal contact child molester type that they have begun to revise their view. (Clinician at Lucy Faithfull Foundation)

These views are congruent with much empirical evidence. Seto et al.’s 2011 meta-analysis concluded that out of a sample of 2620 Internet offenders who were followed up for between 1 and 6 years, 4.6% had committed a further sexual offence. More recent, and for our purposes more pertinent because it distinguished between IIOC only offenders and mixed Internet offenders is Elliott et al.’s 2019 UK study. This compared 584 individuals convicted only of IIOC offences with 106 who had convictions for IIOC and other sexual offences. Follow over a 13-year period found that 2.7% of the IIOC only group were reconvicted of a sexual offence compared to 14.2% of the mixed group. Similarly, the proposal that IIOC offenders are destined to escalate to contact offences does not hold up to examination; an extensive evidence review prepared for the Scottish government concluded that ‘there is little support in the literature for the proposition that viewing IIOC is a gateway to committing contact offences against children’ (Hamilton and Belton, 2022).

Participants often described IIOC offenders as somewhat different to other child sexual abusers in ways that made them more profoundly impacted by apprehension:

On the IIOC side of things the arrest, conviction and registration is that much of a shock, that much of a life changer, they are then committed to ‘this will never happen again’. (Public Protection Officer)

Again, this lines up with existing research that identifies key differences between IIOC offenders as a group and other child sexual abusers (e.g. Babchishin et al., 2018; Thanh et al., 2018). IIOC only offenders tend to be lower on indicators of anti-social tendencies (including previous convictions) and more likely to be employed than other child sexual abusers. They are also particularly likely to kill themselves after arrest. All child sexual abusers have a heightened risk of suicide, but IIOC offenders stand out even within this group, with a risk of up to 100 times that of the broader population (Key et al., 2021). Thus, as a group they are arguably more likely to be deterred by arrest and have more to lose by persisting with illegal behaviours.

Reflecting on the scale of IIOC offending, some participants (from all agencies) suggested that a singularly criminal justice approach was pointless and admired other countries’ ability to integrate public health approaches which recognise pornography use (legal and illegal) as a social and mental health issue:

if you talk to Europeans especially Germans they go ‘why are we even dealing with this population – they’re no risk’. So, you have these two extraordinary poles of opinion between, like, the Dunkelfeld, you know . . . the German notion, and the American FBI Model. The UK sits uncomfortably between those two notions. (Clinician at Lucy Faithfull Foundation)

Participants were acutely aware that those who access CSAM are driving demand and therefore the abuse of children. However, some did question the proportionality of current notification periods for first time IIOC offenders – common disposals were community sentences which require 5 years of registration, or suspended custodial sentences which require a minimum of 7 years.

To probation participants, whose practice has been governed by desistance theory and good lives models for some time, the length of registration for IIOC was redolent of a past era:

I can think of one guy and it was historic, child abuse images . . . it had been years, given the length of time it takes for the police to investigate, to download computers and stuff . . . and he’s like, ‘so I’m still a sex offender’. And you know, the likelihood is he’s not going to re-offend, all of his risk assessments were low, but . . . years from now, the police would still be popping up at least once a year. He would still have to declare every time he wanted to go on holiday, get a new car, change his bank account, insure a car, get house insurance. (Probation Officer)

So, for many participants, it made sense that convictions for some IIOC offences did not merit registration for current duration periods. Such recommendations were made cautiously and specifically for first time offenders who download only – any offenders whose online activities sought contact with children were not included. There are many typologies of online sexual offenders available, but this could be seen as aligning with Merdian et al (2018) who distinguished between those who are fantasy driven and those who are contact driven – these recommendations would just apply to the former.

Some participants suggested a lighter touch for first time IIOC offenders which required a psycho-educational course to be completed. These two comments were typical:

If on that first conviction of a low number of images you get some sort of referral order where you have to go and do [a] treatment programme . . . but then possibly to have another seven years of management on the register afterwards just seems somewhat unnecessary. (Public Protection Officer)

I’ve been talking about diversionary schemes for years. A bit like speed awareness courses . . . I would say 70 - 80% of those guys don’t need to be on the register . . . they’re very time-consuming. It all silts up at the RSO end of things and the courts are full of these guys . . . it’s an easy nicking, I suppose, but that’s about it. (Worker at CirclesUK)

Views were divided on this issue however, and there were participants (concentrated within the police) who were in complete disagreement. To them reconviction figures were meaningless and bore little relationship to actual rates of reoffending:

I’ve got very little faith that these people that have shown a complete absence of self-restraint have got the ability to police themselves, to not go out and do something. (Public Protection Officer)

Even if they were to accept that reconviction rates were exceptionally low for IIOC offenders, in their view it was precisely because they were on the register:

The data doesn’t pick up those that do that do commit contact re-offences and we never find out about it. And it also doesn’t encapsulate those that would but don’t, only because we are managing them. (Public Protection Officer)

These participants rejected the notion that there was such a thing as a simple downloader, and cited examples of IIOC only offenders who were later revealed to be sharing images with others, or who were engaged in contact offending (or efforts to do so). For them, any suggestion that this group’s requirement to notify should be reviewed smacked of a disregard for children who had been abused in the creation of the images, and whose image was now distributed internationally, and possibly in perpetuity. They stressed that CSAM would not be created if there were not a huge demand for it, and that IIOC offenders were part of the supply chain, albeit at the far end of it. That meant any risk was too much:

I think some [online] sex offenders may not commit contact offences. But some of them might. And how do you know which ones are and how do you stop them if they’re going to? (Police Officer in linked series unit)

This is a difficult split in opinion to resolve, and the gap between offending and conviction will always remain an unknown. However, it is clear from reconviction data that a great deal of police time is being spent monitoring IIOC offenders who are very unlikely to see the inside of a court again.

Implications for policy and practice

As participants explicated how the constant expansion of RSO numbers militated against its effectiveness, their strategies for reducing the numbers of low-risk RSOs often started with a wish for more discretion, either at the point of sentence, or once the RSO had been notifying for some time:

If you’ve got a drunk man – and I’ve got a few who touched a woman’s breast as they’ve walked past, a sexual assault, I just think there needs to be a process of professional discretion, so they still get their conviction, but do they need to go on the sex offenders register? (Clinician at Lucy Faithfull Foundation)

However, as this line of thinking progressed, there was usually a conclusion that introducing professional judgement would demand time-consuming risk assessments. It would also place the burden of responsibility and blame on the agencies who had decided that early removal was appropriate, or that notification was not merited, or merited only for a short duration, if the individual then reoffended:

In my fantasy world you’d have that, but as soon as you open up to judgment and discretion, then one, there are huge resources and costs, and two you open it up to subjective judgments . . . it’s not an exact science. (Clinician at Lucy Faithfull Foundation)

So, a bespoke approach was not favoured, but the wish remained that the register should be slimmed down in the interests of proportionality and targeting resources. Participants alluded to amendments to guidance, changes in thresholds, or the creation of referral systems as possible means to this end, and so the practical application of these suggestions needs to be considered alongside the legislation and regulation in this area.

The first of our participants’ suggestion was that it should be easier for indefinite registrants to successfully apply for removal at 15 years, and on this point, a possible solution already exists within the United Kingdom. Current guidance for police in England and Wales is set out by the Home Office (2012), and as our participants indicated, the bar is set high and considerable evidence is needed. Guidance for police in Scotland (Scottish Government, 2022) in contrast sets out a presumption of removal unless there are reasons not to, offering a workable model that could be drawn in England and Wales. Reducing the time before an application could be made would however be a more complex prospect and would require S91b of the 2003 Act to be amended.

Turning to minor sexual assaults and exposure cases of the type described earlier, it could be argued that the 2003 Act is designed to keep such offences away from the SOR, but that the Sentencing Council (an advisory body sponsored by the Ministry of Justice and whose guidance is closely observed by sentencers) ensures that they are put on it. Schedule 3, sections 18(b)(ii) a b and c already contains thresholds to keep the lowest level of sexual assaults and exposure away from registration; a conviction for either offence committed by an adult against another adult needs to attract a sentence of imprisonment, hospital detention or a community sentence of at least 12 months before registration is required. A sentence falling below that does not. However, taking sexual assault as an example, Sentencing Council guidance advises that disposals should start at ‘medium level’ community orders or 26 weeks custody. The level of community order is defined in terms of the conditions that are attached rather than the length of the order, but drawing on the comments of our participants, it is unusual to see a community order passed of less than 12 months, which would trigger registration for 5 years. A custodial sentence (suspended or otherwise) of 26 weeks would incur 7 years of registration and anything over 7 years would incur 10 years. If the Sentencing Council endorsed 6-month community orders, and steered sentencers away from suspended sentences it would be possible to eliminate certain low-risk cases from registration Alternatively, the community order threshold in Schedule 3 could be raised to 24 months, though this would require a legislative change. It is telling that S130 of the 2003 Act explicitly gives the Secretary of State the power to amend Schedule 3, but in one direction only by adding an offence or removing a threshold – the power does not extend to adjustments that would move individuals outside the registration net.

The issue of historic offences was also mentioned by some participants, and this is more difficult; their gravity may indeed justify a substantial sentence, yet the time elapsed may demonstrate that the risk is now so low as to make registration pointless. This situation does not lend itself to an easy solution such as keeping the sentence low or adjusting a threshold – but it is a clear example where registration represents additional punishment, rather than public protection.

Regarding IIOC offences, there are no thresholds in Schedule 3, so any sentence incurs a notification requirement. Sentencing is determined by the category of images (A, B or C, with C being the least serious) and volume of images, but at the lowest end Sentencing Council guidance is identical to sexual assault – a starting point of a medium level community order, or 26 weeks custody. So, again, a community order of a year or more would incur 5 years registration, and any custodial sentence, whether suspended or not, 7 years and upwards. Indeed, suspended sentences have become increasingly popular disposals for IIOC offences in recent years, with 43% of IIOC offenders received one in 2020 (Skidmore et al., 2022).

Some of our participants argued that a more appropriate response would be to deal with first time, low-risk IIOC cases by requiring some psycho-educational treatment, but keeping registration to a minimum. The most obvious way of achieving this would be by the use of conditional cautions; although a caution is an ‘out of court’ disposal, imposed with the agreement of the individual after arrest without further court procedures, any caution requires registration for two years, and conditions can be attached. There would certainly be no problem in devising such programmes as they exist already (e.g. Maps for Change run by Probation or InformPlus run by Lucy Faithfull Foundation). However, on this point it is not only the Sentencing Council placing obstacles in the way; current Crown Prosecution Service (CPS) guidance specifically advises that ‘an out of court disposal is unlikely to be a suitable method of disposal where indecent images of children are found on the suspect’s device’ (CPS, 2024).

It is not only our participants’ comments that question the wisdom of the Sentencing Council and CPS on this point. A series of voices from law and policing have advocated the reduction or removal of notification requirements for first time IIOC offenders and a move towards diversion plus treatment:

• • In 2017, Chief Constable Simon Bailey (2017), the NPCC lead for Child Protection declared in a press release that the scale of arrests for online abuse was at ‘saturation point’, and in order to give sufficient attention to those who presented serious risk, ‘alternatives, such as rehabilitation and treatment rather than prosecution’ should be found for low-risk IIOC using ‘a conditional caution that enforces rehabilitation with an organisation like the Lucy Faithful Foundation’.

• • In 2019 a report by Peter Rook (2019) KC for the organisation Justice proposed that a ‘conditional diversion scheme’ should be tried for first time IIOC offenders, so that prosecution is avoided if a course is completed covering education around the CSAM industry, safe use of the Internet and personal safety planning. It noted the low reconviction rate for such offences, the high suicide rate, and the attraction of taking a ‘substantial number of cases out of the CJS’. Moreover, a diversionary scheme could be implemented speedily, obviating the lengthy current delays between arrest and court case.

• • In 2022, a report by Michael Skidmore for The Police Foundation argued that the massive police resources needed to investigate, convict and manage online sexual offenders made the current practice of prosecuting high volumes of low-risk online offenders untenable. The report recommended conditional cautions for such cases, the condition being that the individual pays for a psycho-educational course.

• • The executive summary of Creedon’s (2023) Independent Review of Police Management of Registered Sexual Offenders includes exactly the same suggestion – to pursue the idea of ‘non-judicial diversion options’ for some online offenders. Creedon’s full report has never been published, and so it may be that his conclusions were not politically palatable.

Thus, when the dots are joined between our findings and the recommendations of other commentators, the result is forceful support for amendments that reduce the requirements placed on some low-risk offenders.

To say definitively how much police time would be saved would however require a breakdown of the RSO population. The total number is published in annual MAPPA reports, but key questions remain – the number of indefinite registrants, how many RSOs are convicted for first time sexual assaults or IIOC offences of the type described. Existing evidence confirms that rates of recorded IIOC offences have increased several times over – by over 800% between 2012–2013 and 2020–2021 according to Skidmore et al (2022) – but a recorded crime is not the same as a conviction, and a conviction can lead to a variety of sentences and notification requirements. The same authors make the point that a substantial proportion of reported IIOC offences are committed by those under 18 (often involving self-generated imagery) who are exempt from registration because of a threshold in Schedule 3. So, in a practical sense this should be the first step – a scoping exercise to build a picture of this heterogenous population.

Limitations

All participants worked in the south and southeast of England, and this restricted the geographical range of voices. However, efforts to gain access to agencies in other parts of the country (including the midlands and the northeast) were unsuccessful. The absence of RSOs in the sample could be viewed as a weakness, but a decision was taken to focus exclusively on professionals’ views in this study. Other work is available that examines the experiences of those who are on the register (e.g. Mann et al., 2021).

Conclusion

Our participants brought decades of experience detecting, managing and treating sexual offending to this research, and identified many strengths in the SOR. However, when their data is considered alongside recent reports, the conclusion is clear; work with high- risk RSOs is diminished by the resources needed to manage those who present little risk. The lengthy periods of registration may have been congruent with the criminal justice climate when the SOR was created, but treatment approaches have evolved since, recognising that for most individuals convicted of sexual offences the prospect of a liveable life is a key element of rehabilitation and risk management. Moreover, data confirming low reconviction rates for IIOC offenders is now considerable. It would be naive to think the public would approve of such modifications, but if the wishes of public and media are to be the only influence, then the management of those convicted of sexual offences will go in one direction only, that of ‘more, longer, tougher’ (Bartels et al., 2019), and result in a SOR that so bloated that it defies management. Workable solutions are available without a radical re-write of the 2003 Act – changes in prosecution and sentencing guidance, amendments to schedule 3 – and the case for such changes is compelling.