Indefinite restraining orders for stalking: proportionality where risk is enduring and contact serves no legitimate purpose

Introduction

In R v Taylor [2026] EWCA Crim 727, the Court of Appeal (Criminal Division) (Popplewell LJ, Goose J, and the Recorder of Sheffield sitting as a judge of the CACD) considered a narrow but important sentencing question: when is it lawful and proportionate to make a restraining order “until further order” (i.e., indefinite) following conviction for stalking?

The appellant had pleaded guilty to stalking involving serious alarm or distress contrary to Protection from Harassment Act 1997, s 4A and received 14 months’ imprisonment (unchallenged). The appeal concerned only the duration of a restraining order protecting the complainant (a surgeon who had previously treated the appellant). The terms of the order were agreed; the issue was whether an indefinite order was disproportionate.

Summary of the Judgment

The Court of Appeal dismissed the appeal. It held that, on the facts and the risk evidence, an order until further order was:

  • Necessary to protect the complainant from further harassment/fear of violence; and
  • Proportionate in both scope and duration, given the absence of any legitimate reason for future contact and the evidence of long-term, relapse-prone risk.

The court accepted the sentencing judge’s assessment that indefinite restraining orders are “exceptional or rare”, but concluded that this was one of those cases.

Analysis

Statutory framework: restraining orders on conviction

The power derives from Sentencing Act 2020, ss 359–360. In summary:

  • A restraining order can prohibit conduct to protect a victim (or others) from harassment or fear of violence.
  • It may last for a specified period or until further order.
  • It must be made in addition to dealing with the offender for the offence (i.e., it is ancillary to sentence).

The appeal therefore turned on the correct exercise of discretion within a statutory scheme that expressly permits an indefinite duration.

Precedents cited

The court treated R v Khellaf [2016] EWCA Crim 1297 as the principal guide. The judgment extracted and applied the following core principles:

  • Victim’s views: the protected person’s views should be sought.
  • Necessity: an order should not be made unless necessary to protect the victim.
  • Proportionality: terms must be proportionate to the harm sought to be prevented.
  • Children: special care where children are involved (not central here).

Khellaf did not prescribe when an order should be indefinite; rather, it anchored the analysis in necessity and proportionality. In [2026] EWCA Crim 727, those principles were used to justify indefinite duration where the evidence supported an enduring, reactivatable risk and the restrictions were modest relative to the harm prevented.

Legal reasoning

(1) Necessity and risk: why the order was needed at all

Necessity was effectively common ground. The complainant’s evidence (including a victim personal statement) described serious alarm and distress and protective lifestyle changes. The appellant’s conduct included renewed contact attempts after a long hiatus and later explicit threats of violence. The court accepted that the risk of future harm was “ongoing”.

(2) Proportionality: why “until further order” was not excessive

The key proportionality reasoning was practical and fact-sensitive:

  • No legitimate basis for contact: the judge found (and the CACD agreed) there was “absolutely no reason” for contact. The appellant was not expected to resume a clinical relationship with the complainant, and there was no other relationship.
  • Scope of restrictions was targeted: the order focused on contact, family contact, and attending addresses where the complainant lived/worked, with limited exceptions for emergency healthcare and pre-booked appointments with proof.
  • Duration matched the risk profile: professional risk assessments indicated a high risk of persistence/recurrence, including recurrence after a substantial non-contact period.

The court placed weight on the fact that the appellant had previously resumed stalking after over three years without any identified trigger, supporting the sentencing judge’s conclusion that the risk might recur unpredictably over a long timescale.

(3) Mental health and the “gap” in offending: why they did not require a time limit

The appellant relied on: previous good character, no prior breaches of court orders, mental health difficulties, and the long gap between episodes (2021–2024). The CACD treated those points as legitimate mitigation, but concluded they were outweighed by the risk evidence:

  • The psychiatric and pre-sentence materials supported a view of fluctuating insight and a stalking motivation that could be resentment-based (long-lasting) and intimacy-seeking (linked to instability), making future relapse plausible.
  • The gap was not evidence of safety; it was evidence that risk can remain latent and then re-emerge.

(4) “Until further order” and later variation: a safety valve, not the justification

The appellant argued the judge should not have relied on the possibility of a future application to vary/discharge. The CACD’s approach was that the judge’s primary decision was based on present necessity/proportionality; the possibility of variation simply reflected the statutory design (an indefinite order remains subject to “further order”) and did not make an otherwise disproportionate order lawful.

(5) Public-facing victims and ongoing accessibility

The Crown’s response (accepted as relevant) emphasised that the complainant’s public-facing professional role meant workplace contact details and other information would remain accessible. This mattered because it reduced the likelihood that the risk would naturally extinguish through loss of access.

Impact

The decision reinforces several points likely to influence future restraining order appeals:

  • Indefinite can be proportionate even where it is “exceptional or rare”, if the evidence supports an enduring/recurring risk and the offender has no legitimate reason for future contact.
  • A long hiatus in stalking behaviour may support, rather than undermine, the case for an indefinite order where relapse occurred after a dormant period.
  • Courts may place substantial reliance on structured risk assessments (psychiatric and probation) when selecting duration.
  • Where the protected person is professionally visible and contactable, the “opportunity structure” for renewed harassment may persist, supporting longer orders.

Complex concepts simplified

  • Restraining order (on conviction): an order made after conviction to prevent future harassment or fear of violence. Breach is a separate criminal offence.
  • “Until further order”: the order has no fixed end date but can be varied or discharged by a later court order if circumstances materially change.
  • Necessity vs proportionality: “necessary” asks whether protection is needed at all; “proportionate” asks whether the restrictions (including duration) go no further than required to prevent the future harm.
  • Victim Personal Statement: a statement describing the impact of offending, relevant to assessing harm and protective need.

Conclusion

R v Taylor [2026] EWCA Crim 727 affirms that an indefinite restraining order may be justified where the offender presents a long-term, relapse-prone risk, the victim faces serious psychological harm, and the offender has no legitimate reason for any future contact. The case shows that a substantial gap in offending does not necessarily support a time-limited order; it may instead demonstrate that risk can remain dormant and return, making “until further order” a proportionate protective measure.