R v Leon R

The provisions of the Sexual Offences (Amendment) Act 1992 apply to this offence. Under those provisions, where a sexual offence has been committed against any person, no matter relating to that person shall, during that person's lifetime, be included in any publication if it is likely to lead members of the public to identify that person as the victim of that offence. This prohibition applies unless waived or lifted in accordance with section 3 of the Act.

On 30 November 2022 in the Crown Court at Snaresbrook before His Honour Judge Southern, the appellant, who was then aged 36, was convicted of five counts of sexual assault on a child under the age of 13 and four counts of rape of a child under 13. The complainant in each case was the appellant's stepdaughter, who we shall call Z. The offending was alleged to have happened between 2010 and 2013 when the appellant was aged 23 to 26 and Z was aged between seven and 10.

He now appeals against his conviction. The basis of his appeal is that the judge was wrong to admit as evidence in this case the fact that in October 2020 he had been convicted of raping a 27-year-old woman in the circumstances we describe more fully below. Although the appellant was convicted of two counts, and we refer to them as the 2020 convictions, they arose out of one incident that gave rise to separate counts of oral and vaginal rape. The admission of the evidence is said to have been wrong on two grounds: first, the appellant submits that the evidence of his 2020 convictions did not show a propensity to commit sexual assaults and rape against a young child and was therefore not admissible; second, if the evidence was admissible it would have such an adverse effect on the fairness of the proceedings that the court ought not to have admitted it.

It is not necessary to outline the facts that are the subject of the present proceedings in detail. Z alleged that the appellant started by rubbing her "down there" and then progressed over time to licking her vagina, making her suck his penis and to penetrative sexual intercourse which was the subject of the four counts of rape.

All of the offending occurred in the home or the equivalent of a home. One offence was alleged to have occurred when the appellant was in bed with Z and another woman, not her mother.

In addition to Z's evidence, the prosecution relied upon evidence that Z had made disclosures to three people, including her mother.

The appellant denied all of Z's allegations. He said that there had never been any sexual activity between them and that he had no sexual interest in children. The issue for the jury was therefore whether they believed Z and were sure that the offending had occurred.

During the course of the trial the prosecution applied to adduce the evidence of the appellant's 2020 convictions. The application gave as the reason why the convictions were relevant as: "The previous convictions demonstrate a propensity on the defendant's part to sexually offend and rape females". The prosecution skeleton gave as the relevant details that:

The Crown's submission was that:

The essence of the prosecution's submission was repeated and recorded in paragraph 4 of the judge's ruling where the judge said:

The prosecution developed that submission by suggesting similarities between the conduct the subject of the 2020 convictions and the alleged behaviour with Z, namely:

The judge explained the prosecution's case on "c" above as being that the secluded area (as it was described in the prosecution's skeleton) where the appellant took his victim was an alleyway overlooked by houses and the incident came to an end when the occupier of an adjacent house became aware of what was happening and shouted out, which gave the victim a chance to escape.

The judge set out the defendant's submissions at paragraph 7 of the ruling. They were that the circumstances of the 2020 convictions were not sufficiently similar to engage the gateway under section 101(1)(d) or to open that gateway because they do not show any propensity to commit the offence with which the current trial was concerned. In the alternative, it was submitted that the evidence should be excluded under section 101(3).

Having set out the relevant provisions of sections 101 and 103, the judge noted the need for special caution when dealing with old convictions, while also noting that in the present case the situation was unusual because the convictions upon which the prosecution wished to rely were relatively recent. The judge then identified the "matter in issue":

The judge then prefaced his analysis by reference to the three well-known questions from Hanson [2005] EWCA Crim 824, namely:

The judge accepted that there were differences between the index offences and the facts underlying the 2020 convictions:

Having done so, he referred to the facts and judgments in the case of Miller [2010] EWCA Crim 1578, where both sets of offending had involved relatively young girls. In doing so he drew attention to two passages from the judgment of Leveson LJ in that case. The first was:

The second passage was the observation at paragraph 18 of Miller that:

The judge then came to the kernel of his reasoning where he identified similarities he considered to be relevant and concluded his analysis at paragraphs 18 to 22 as follows:

It is to be noted that the judge did not either in paragraph 18 or elsewhere in his ruling indicate that he accepted or relied upon the asserted feature of acute risk of detection.

The judge dealt with the 2020 convictions in part one of his summing-up where he said:

There is no ground of appeal on the basis of the summing-up. Despite that, it is material to this appeal that although the judge said that he would remind the jury of the evidence about the 2020 convictions, he did not in fact do so. The only reference to the 2020 convictions in the summing-up was therefore in the passage that we have set out above.

Section 101 of the Criminal Justice Act 2003 provides so far as relevant:

Section 103 provides (again so far as relevant):

Certain principles are well established:

Miss Brickel, who represents the appellant before us as she did in the court below, essentially repeats the submissions she made to the judge. She submits that the weight to be attached to the 2020 convictions is less than it would have been had that offending pre-dated the offending that is the subject of the present proceedings. She criticises the judge's treatment of the facts in Miller, submitting that it is not analogous to the present case. In Miller the historic conviction was for the gang rape of a girl of 15 which was said to be evidence of propensity to commit the offences of rape on the defendant's young niece. Here Miss Brickel submits the 2020 convictions were against an adult female with whom he had no prior relationship, which she submits to be a major point of difference, not similarity, when compared with the offending against Z. She also criticizes the nexus relied upon by the judge that the 27-year-old and Z were both "vulnerable", the nature of their respective vulnerabilities being quite different. She does not accept that there is any real similarity based upon acute risk of discovery, again contrasting the circumstances of the 2020 convictions where the rape of a complete stranger occurred in a public place overlooked by occupied houses, with the circumstances of the offending against Z which almost exclusively occurred in what amounted to the home environment, where the appellant was in a position of trust in relation to Z which reduced the likelihood of detection, whether or not there were others at home at the time. Finally, she submits that the fact that the 2020 convictions included the infliction of oral sex on his victim is not a sufficiently distinctive feature to demonstrate propensity in relation to his abuse of Z.

Turning to section 101(3), she submits that the admission of the 2020 convictions had such an adverse effect on the fairness of the proceedings that the judge should have excluded them.

For the respondent, Mr Bearman, who represented the prosecution at trial and has represented the Crown before us, submits that the judge's decision was open to him and discloses no error of principle and no basis upon which it could be said that the judge was plainly wrong on admissibility. He relies upon four features of the 2020 convictions: namely (a) he submits that they demonstrate the appellant's appetite for unlawful sex and an inherent abuse of power; (b) he submits that they demonstrate a tendency to force his victims to give him oral sex; (c) he submits that they demonstrate a willingness to embark on non-consensual sexual activity in circumstances where there was an acute risk of discovery; and (d) he submits that the 2020 convictions and the behaviour alleged against Z demonstrate the appellant preying on extremely vulnerable complainants.

Turning to section 101(3), he adopts the approach suggested by the Court of Appeal in Hanson where the following factors were said to be relevant but not exhaustive:

Overall he submits that the judge did not err in law either in his ruling on admissibility or in refusing to exclude the evidence of the 2020 convictions.

It is essential to bear in mind that the matter in issue between the prosecution and the defence was not simply whether the appellant had a propensity to commit sexual offences against females but whether he had a propensity to commit offences of the kind with which he was charged, namely the sexual abuse and rapes committed against a young child. Had the charges that the appellant faced in these proceedings been sexual abuse or rape of an adult female, we consider that it would have been plain beyond argument that the 2020 convictions were admissible. However the situation here is different.

We start with the judge's reasons which are encapsulated in paragraph 18 and which we have set out above. The first feature relied upon by the judge was that both occasions involved a "young female" drawing a comparison with Miller. We have two major and connected reservations about this feature. First, it seems to us to be stretching language too far to describe Z and the victim of the 2020 convictions as "young females" as if that were potentially probative of a propensity to offend sexually against a seven to nine-year-old child. It is not in our judgment self-evident that a propensity to offend sexually against a 27-year-old woman is probative of a propensity to engage in paedophile abuse against a seven to nine-year-old girl. In this context a 27-year-old woman is mature, both physically and generally, whereas the seven to nine-year-old child is immature. We express no concluded rule because there is no evidence in the case to support it but we do not consider that there was any basis on which the judge could draw support on the basis of "youth" from the fact of the 2020 convictions. What may be said with confidence on the basis of the experience of the courts on an almost daily basis is that the psychology that drives people to paedophilia is not necessarily or even probably the same as the psychology that drives people to commit sexual assaults on mature victims. There appears to have been no consideration of these differences in the judge's reasoning.

Second, we do not consider that the analogy that the judge drew with the facts of Miller was sustainable. In Miller the age of the victims, 11 and 15, were more closely aligned than in the present case. It is in that context that the Court of Appeal recognised in Miller that the previous conviction, though founded on different circumstances, demonstrated a propensity for "sex with a young girl". Furthermore, the passage from paragraph 18 of Miller relied upon by the judge, which we have set out above, concentrated more on the underlying abuse of power as evidence of propensity. Reminding ourselves of the fact-sensitive nature of the decision to be taken in each case, we accept that many serious sexual offences may involve an abuse of power, since the will of the victim is overborne. But that feature is only one factor to be weighed in the balance when considering whether and to what extent a previous or subsequent conviction is probative of a propensity to commit the offence with which a defendant is charged.

The second feature relied upon by the judge in paragraph 18 of his ruling was the vulnerability of each of his victims. Once again, such a feature can be identified from significant number of cases of serious sexual offending. But the question to be asked is whether and to what extent it is a feature that renders the bad conduct probative of a propensity to commit offences of the type charged. Here as the judge acknowledged the circumstances of the vulnerability of the victim in the 2020 convictions was quite different from the circumstances of the offending alleged against Z. Furthermore, because of the limited information about the circumstances of the 2020 convictions, there was little information before the judge that would have been enabled either him, or later the jury (not least because the judge did not return to the evidence in part two of his summing-up) to make any real assessment of the level of vulnerability of the victim of the 2020 convictions. Once again, we do not consider it to be self-evident that the nature of the victim's vulnerability in the 2020 convictions demonstrates or even evidences a propensity to commit the offences alleged against Z, whose vulnerability was quite different as a young child and where the source of the vulnerability because of the familial relationship was also quite different.

It seems to us to be tenuous, though not completely irrelevant, to bring into account that the other sexual offending was committed against a vulnerable victim. At its highest, the weight to be attributed to it as a feature in the absence of more compelling factors such as a distinctive mode of operation is likely to be slight.

The judge was right to remind himself that propensity can be established by a single conviction. But he did not in paragraph 19 or elsewhere give reasoned consideration to the essential question whether the differences between the circumstances of the 2020 convictions, taken together with the similarities that he had identified viewed overall meant that the 2020 convictions could or could not be regarded as probative of a propensity on the part of the appellant in 2010 to 2013 to carry out the offences of the kind alleged against Z.

In our judgment, given the nature of the facts underlying the 2020 convictions, the differences between those facts and the matters alleged against the appellant in these proceedings can only be described as stark. The 2020 convictions involved a mature adult aged 27. The offences against Z were alleged to have occurred when she was aged seven to ten. For the reasons we have given, we regard this difference as being of the highest importance in the absence of evidence or good reason to suppose that a willingness to carry out the 2020 offences could or did demonstrate a propensity to have committed offences of the kind alleged against Z some seven to 10 years before.

Second, the 2020 convictions involved a stranger. The offences alleged in these proceedings were against a close family member.

Third, the 2020 convictions arose out of a spontaneous and opportunistic encounter with the victim. The allegation in these proceedings was of long term and determined grooming and sexual abuse within the protective environment of the family.

Fourth, and related to the previous point, the 2020 offence could be described as a one-off. That is quite different from the course of conduct alleged in these proceedings.

Fifth, the 2020 offending occurred when the appellant had been drinking. There is no such suggestion in relation to the offending against Z.

Sixth, we consider that the reference to an acute risk of discovery is of limited if any value, given the differences between the facts of the 2020 convictions (where the appellant lured the victim down an alley with no one immediately present) and the facts of the offending alleged against Z (which, as we have said, occurred in what was for the appellant protective of his position of trust in the home environment).

We accept that it can be said of the 2020 convictions that they showed a willingness to have non-consensual and unlawful sex with his victim in that case. But for the reasons we have given, it seems to us that this is of limited value when trying to assess whether the appellant had a propensity to abuse Z as alleged.

Turning to the matters now relied upon by the prosecution, the only additional feature is the fact that the appellant was convicted of oral rape and Z said that the same had happened to her. While that is a feature that is common to the 2020 convictions and the offending alleged against Z, it seems to us to be a feature that should be given little weight, falling far short of being a feature that could properly be regarded as distinctive so as to be significantly probative.

We return to the first question in Hanson. Do the matters relied upon by the prosecution establish a propensity to commit offences of the kind charged in relation to Z? We remind ourselves that we are reviewing the judge's decision and it would not be sufficient, or even particularly relevant, that we think either individually or collectively that we would have come to a different answer than that reached by the judge.

In our judgment, the reasons given by the judge for answering this first question as he did, did not justify the conclusion he reached. What is lacking is an absence of any assessment of the similarities and differences with a view to determining whether the 2020 convictions did in fact establish the relevant propensity. Had the judge engaged in such an exercise, as in our judgment he should have done, he should have reached the conclusion that the probative value of the 2020 convictions was limited. At their highest and at the most general level the 2020 convictions could be said to establish a propensity to commit serious sexual offences against adult females. But that leads directly to the second Hanson question: does that propensity make it more likely that the defendant committed the crimes against Z with which he was charged? The judge did not separately address this question. Had he done so, he should in our judgment have concluded that the entirely general propensity identified in answer to the first question did not make it appreciably more likely that he committed the particular offences alleged against Z.

We use the phrase "appreciably more likely" for two reasons. First, because we recognise that a propensity to commit sexual offences generally could filter through to make it at least theoretically more likely that the appellant would commit sexual offences against anyone, including Z. But second, the question whether the propensity evidenced by the 2020 offences makes it more likely that the appellant committed the offences against Z, must also take into account not only the differences to which we have referred but also the time difference between 2020 and the time of the offences alleged against Z.

In our judgment, the facts of the 2020 convictions provided only limited evidence that is capable of going to support a conclusion that the general propensity identified in answer to question 1 made it more likely that the appellant committed the offences alleged against Z. Once all the features to which we have referred are taken into account, we consider that the second Hanson question should be answered by saying that any increased likelihood was at best slight.

Turning then to the third Hanson question, we are in no doubt that the impact of admitting the convictions would be very potent and that their prejudicial effect would substantially outweigh their limited probative value. In his measured and realistic submissions, Mr Bearman recognised the potency of the impact that the admission of the 2020 convictions would be likely to have, though he did not accept that their probative value was limited. Once the limited probative value of the 2020 convictions is weighed against the likely prejudice to the appellant, we are satisfied that the decision not to exclude them is insupportable.

The judge in paragraph 21 of his ruling was correct that merely damaging the defence case was not a good reason on its own to exclude the evidence; and he was also right to say that the prosecution case without the 2020 convictions was not a weak one. We are not sure we understand his reference to the chronology of events not being an obstacle to admission. If he was referring to the fact that the 2020 offending post-dated the offending alleged against Z that appears to us to be a matter that went to questions 1 and 2, rather than to the question the judge was addressing in paragraph 21 of his ruling.

What the judge did not do in his ruling was to engage in any exercise that sought to assess the balance between the probative and the prejudicial effects of admitting the 2020 convictions. Had he done so, as we have indicated, we consider that he should have concluded that the 2020 convictions should be excluded. Their admission could only distract from the important questions under sections 101(d) and 103 to which we have referred.

We note in passing that the approach we have adopted is materially similar to the approach adopted by a very strong constitution of this court in Benabbou [2012] EWCA Crim 1256 at [16] to [24], which it is not necessary to set out in detail here.

We have given intense attention to the summing-up to see whether the potential prejudice to the appellant was either obviated or ameliorated by the directions to the jury about how they should approach the facts of the 2020 convictions and the principles pursuant to which they had been admitted. Regrettably we are unable to find anything in the summing-up that might have cured the potential injustice. In the passage we have set out above, the judge did not identify what, if any, features were said to be probative of the appellant's alleged propensity to commit offences of the kind with which he was charged and why it was suggested that they might have probative value. His summary of the defence case was brief and did nothing to explain how, for example, the jury should approach the fact that the 2020 offending post-dated the offending alleged in the current proceedings as they did. It may be that the judge intended to address such topics when he returned to the convictions in part two of his summing-up. But in event he did not do so.

Finally we stand back to assess whether the appellant's conviction of the index offences is safe, despite the wrongful admission of the 2020 convictions. We accept, as did the judge, that the case against the appellant was not weak in the absence of evidence about the 2020 convictions. But in a case that was substantially the appellant's word against Z's, we are unable to accept that the outcome would necessarily have been the same if the 2020 convictions had not been admitted. In other words, we are unable to conclude that the convictions are safe in any event.

For these reasons, we are driven to conclude that this appeal must be allowed. Mr Bearman, are you going to seek a retrial?

My Lord, we do make an application for a retrial.

We have allowed the appeal. We quash the convictions. We order a retrial on all the counts in respect of which we have quashed the convictions. We direct that the appellant be re-arraigned on the fresh indictment within two months. You will be conscious of the decision of the Lady Chief Justice yesterday about the significance of the two months?

Very.

Are there any other directions that we should give, for example should we give a direction that the trial be conducted by a Presiding Judge or by the Resident Judge or by a judge other than the judge who conducted the last trial? I would expect that as a matter of normal listing practice it will be listed in front of another judge.

My Lord, yes. I would not have thought it necessary to make any direction as to specify the Resident Judge or a Presiding Judge, only that it be retried by a different judge.

And he, as we know, is in custody and not liable to be released in the period leading to -- is that right?

I understand his earliest date of release is some time in 2026.

Yes.

So we do not need to be concerned about that. If there was any likelihood of him being released the matter should be brought back to court for consideration.

I have already given a warning about -- the sexual offences warning. I should probably also make a direction under section 4(2) of the Contempt of Court Act restricting reporting of the proceedings until after the retrial.

Yes, thank you.

Is there anything else?

No, my Lord. My Lord, is he on technical bail therefore on this matter since he is serving on the other matter?

Yes. Do we need to say anything about representation orders?

My Lord, no.

Let us leave it like this, if we do need to say something about representation orders we make representation orders limited to junior counsel only. But that can always be brought back if need be.