Decoy “Child” Incitement Attempts: Category 1A Classification and Immediate Custody as the Norm (Underwood)

1. Introduction

R v Underwood concerned an Attorney General/Solicitor General reference for undue leniency under section 36 of the Criminal Justice Act 1988. The respondent (born 1964) pleaded guilty in the Crown Court at Truro to attempting to incite a child to engage in sexual activity, contrary to section 1(1) of the Criminal Attempts Act 1981, arising from online communications with what he believed was a 13-year-old child, but which was in fact a decoy profile operated by “paedophile hunters”.

The central issues were:

  • the correct application of the Sentencing Council Definitive Guideline for causing/inciting sexual activity with a child (harm and culpability categorisation);
  • the appropriate reduction (if any) where the “child” is a decoy and no sexual activity occurs;
  • whether a sentence of imprisonment could properly be suspended and, more fundamentally, whether the custodial term should have exceeded the legal limit for suspension;
  • the role of delay in proceedings as mitigation.

2. Summary of the Judgment

The Court of Appeal (Criminal Division) held the original sentence (18 months’ imprisonment suspended for 2 years with a rehabilitation activity requirement) to be unduly lenient. It concluded that the offending should have been sentenced as a Category 1A case under the relevant Definitive Guideline, producing a sentence which had to be well above 2 years (the maximum that can be suspended as a matter of law).

The court quashed the suspended sentence and substituted 3 years’ imprisonment, to be served as immediate custody. It reasoned that:

  • the correct guideline category was Category 1A (not 1B);
  • any reduction for the “no real child/no completion” feature could only be very small, since the offence was only not completed due to external intervention;
  • even allowing for good character and systemic delay, the notional post-trial term could not fall below 4 years, yielding 3 years after a 25% guilty plea credit.

3. Analysis

3.1 Precedents Cited

R v Timpson [2023] EWCA Crim 453

Timpson was relied upon in relation to delay as potential mitigation. The Court of Appeal in Underwood drew from Timpson (especially the passage quoted at [23]) that where there is delay between offence and charge not attributable to the offender, some reduction may be justified if it has had a detrimental effect, but reductions of the magnitude of 25% will be “most unlikely”, particularly for serious offending.

In Underwood, delay was accepted as relevant (offending 2022/23; charge May 2025) but treated as incapable of justifying a move into a substantially lower custody bracket where the guideline classification and offence seriousness demanded otherwise.

R v Stewart [2016] EWCA Crim 2238

Stewart was cited for the principle that the Court of Appeal and the Law Officers are not bound by prosecution concessions. In Underwood, this mattered because the prosecution’s approach at first instance had been erroneous and incomplete as to guideline factors. The appellate court emphasised that it could (and should) sentence on the correct legal basis notwithstanding concessions below.

R v Reed & Ors [2021] EWCA Crim 572; [2022] 1 Cr App R(S) 6

Reed provided the key framework on suspending sentences in serious sexual offending involving online conduct and, critically, cases where the “victim” is not a real child. The court in Underwood relied on Reed for two connected propositions:

  • Even where a custodial term is capable of being suspended “as a matter of law”, immediate imprisonment will usually be inevitable for offences of this type.
  • This remains so “even when the victim is not a real child” (echoing the quotation from Reed at [70] in the judgment text).

Underwood operationalised Reed by treating “decoy victim” as, at most, a limited harm-completion feature (warranting only a small reduction), not a gateway to suspension or a wholesale recharacterisation of seriousness.

R v BRJ [2023] EWCA Crim 351

BRJ was invoked by the respondent to emphasise the well-known restraint of the Court of Appeal in undue leniency references: the question is not whether the appellate court might have sentenced differently, but whether the sentence fell outside the range reasonably open to the judge.

The Court of Appeal accepted that framework but concluded that the Crown Court’s sentence, being premised on an incorrect guideline category and resulting in a suspended term where the proper custodial term had to exceed 2 years, was outside the permissible range and therefore unduly lenient.

3.2 Legal Reasoning

(a) Correct guideline classification: Harm Category 1 and Culpability Category A

The Court of Appeal endorsed the parties’ common ground that harm was properly Category 1, assessed by reference to the sexual activity the offender intended to incite (digital penetration). The principal appellate correction lay in culpability:

  • The sentencing judge treated the case as Category 1B, focusing largely on age disparity but treating it as insufficient for Category A.
  • The appellate court held the case properly fell within Category 1A, because, in addition to the significant age disparity, sexual images were solicited (the request for photographs “in the bath” was treated as plainly sexual in nature).

The doctrinal importance is that the court treated “solicitation of sexual images” as a serious culpability escalator even where images are not ultimately sent, reinforcing that such conduct forms part of the exploitation dynamic the guideline addresses.

(b) Attempts and “decoy victims”: only a small reduction where completion was prevented externally

The court applied the Definitive Guideline’s instruction for “section 10 cases” (as set out in the judgment text) by analogy to the attempt context: identify harm by the intended sexual activity and then apply a downward adjustment for lack of completion. Critically, it stressed that where the offender would have carried out the offence but for intervention, or where no child exists, only a very small reduction will “usually be appropriate”, and no additional reduction should be made merely because it is an attempt.

On the facts, the failure to complete was not due to voluntary desistment or a genuine internal restraint; it was due to detection and intervention. That placed the case at the “small reduction only” end of the spectrum.

(c) Delay as mitigation: relevant but limited

The judge below relied in part on delay “endemic to the system”. The Court of Appeal accepted delay could be considered but (consistent with R v Timpson [2023] EWCA Crim 453) treated it as incapable of producing a sentence that fell below the guideline-driven seriousness. Delay may reduce sentence at the margins; it cannot justify an outcome inconsistent with correct categorisation and the custody imperative for serious sexual incitement of children.

(d) Suspension: the legal ceiling and the “usual” need for immediate custody

Two distinct problems underpinned the appeal outcome:

  1. Legal capacity to suspend: the appellate court held the sentence “needed to be well above” 2 years (the maximum custodial term capable of suspension). Once the correct guideline approach was applied, the appropriate term exceeded that ceiling, removing suspension from the menu altogether.
  2. Merits of suspension even if available: drawing on R v Reed & Ors [2021] EWCA Crim 572; [2022] 1 Cr App R(S) 6, the court reinforced that immediate custody is usually required even where there is no real child victim. The Crown Court’s optimism about non-repetition and its view that the offender lacked a sexual interest in children did not displace the sentencing aims of punishment, deterrence, and public protection demanded by the offence type.

(e) The recalibrated sentence

Having fixed the case as Category 1A (starting point 5 years; range 4–10), the court reasoned:

  • only a small reduction for the “decoy/no completion” feature was justified;
  • good character and delay were relevant but not transformative;
  • the notional post-trial sentence could not fall below 4 years;
  • after a 25% guilty plea reduction, the correct outcome was 3 years immediate custody.

The court also acknowledged fairness consequences: the respondent had been “given the impression” he would not go to prison and had complied with rehabilitation requirements. Even so, the public interest in correcting an unduly lenient sentence, in a serious safeguarding context, required substitution.

3.3 Impact

  • Stricter insistence on Category 1A where image-solicitation is present: The decision signals that requesting “bath” photographs (and similar content) may be treated as solicitation of sexual images for guideline culpability purposes, pushing cases into higher sentencing brackets.
  • Decoy cases are not “discount cases”: The judgment reinforces that the absence of a real child is not a broad mitigating factor; it primarily affects completion-based harm, and often only marginally where the offender would otherwise have proceeded.
  • Suspension is exceptional in serious online child sexual incitement: Building on Reed, the case strengthens the working assumption that immediate custody will ordinarily be required even when suspension is technically available—still more so when correct guideline application produces a term above 2 years.
  • Delay mitigation is constrained: Courts may acknowledge systemic delay, but Underwood illustrates that delay cannot be used to engineer a suspended sentence where the guideline category points to a substantially longer term.
  • Attorney General/Solicitor General references remain a corrective mechanism for guideline misapplication: The case exemplifies appellate intervention where categorisation errors and improper reliance on suspension produce an outcome outside the permissible range.

4. Complex Concepts Simplified

  • “Unduly lenient” (section 36 references): A sentence is not changed merely because the Court of Appeal would have passed a different one. It is changed only if it falls outside the range of sentences that a judge, applying the law correctly, could reasonably impose.
  • Guideline “harm” and “culpability” categories: Harm reflects the nature/seriousness of the sexual activity intended or caused (e.g., penetration is highest). Culpability reflects aggravating features (e.g., exploitation indicators such as soliciting sexual images).
  • “Decoy victim”: The “child” is not real, but sentencing focuses on the offender’s intent and actions. The key question is whether the offender would have proceeded but for intervention; if yes, the reduction for non-completion is usually small.
  • Suspended sentence: A custodial term imposed but not immediately served, provided the offender complies with conditions during an operational period. In England and Wales, a sentence can only be suspended if it is 2 years or less.
  • Sex offender notification / SHPO: Separate from imprisonment, conviction triggers police notification requirements under Part 2 of the Sexual Offences Act 2003, and the court may impose a Sexual Harm Prevention Order to manage future risk.

5. Conclusion

Underwood clarifies and reinforces three practical sentencing rules for attempted online incitement of children where the “victim” is a decoy:

  1. Where the offender incites penetrative activity and solicits sexual images from a purported child, the case will commonly fall within Category 1A, driving multi-year custody.
  2. The “no real child/no completion” feature generally justifies only a small reduction where the offender was prevented by external intervention rather than self-restraint.
  3. Consistent with R v Reed & Ors [2021] EWCA Crim 572; [2022] 1 Cr App R(S) 6, immediate custody will usually be the appropriate punishment in this offence type; and where correct guideline application yields a term above 2 years, suspension is legally unavailable in any event.

The decision therefore strengthens sentencing consistency in “paedophile hunter”/decoy prosecutions, ensuring that seriousness is measured by intent, exploitation markers (including image-solicitation), and the limited role of non-completion and systemic delay.