No Case to Answer in a Domestic Homicide: “Reasonable Jury Could Infer Guilt” and Bad Character as Context, Not Substitute

1. Introduction

In R. v Morgan [2026] EWCA Crim 472, the Court of Appeal (Criminal Division) (Lord Justice Coulson, Mr Justice Bennethan, and the Recorder of Bristol (HHJ Blair KC)) refused renewed applications for permission to appeal against conviction (murder and two counts of assault occasioning actual bodily harm) and sentence (life imprisonment with a 21-year minimum term, less remand time).

The case concerned the death of the applicant’s partner (the deceased), who was found in a river late at night after the couple had been drinking. The prosecution case was that the applicant drowned the deceased; the defence position was that she entered the water accidentally (or otherwise without murderous agency) and that the applicant’s actions were consistent with chaotic rescue attempts in darkness while intoxicated.

The central appellate issue on conviction was narrow but important: whether the trial judge was wrong to reject a submission of no case to answer, particularly given the defence complaint that the prosecution could only succeed by impermissibly relying “wholly or mainly” on bad character evidence of prior domestic violence. The sentencing challenge alleged excessive uplift from the Schedule 21 starting point and “double counting” of aggravating features.

2. Summary of the Judgment

  • Conviction: Permission to appeal was refused. The Court held there was direct evidence capable of supporting guilt (a witness described the applicant holding the deceased in a position leaving her head fully submerged), and in any event there was strong circumstantial evidence from which a properly directed jury could infer murder. The bad character material supported (but did not replace) an already strong factual case.
  • Sentence: Permission to appeal was refused. The 21-year minimum term was neither wrong in principle nor manifestly excessive. There was no impermissible double counting: the contemporaneous assaults and the applicant’s prior history/disregard of orders were distinct aggravating features. Intoxication and alcohol misuse did not carry significant mitigation where the applicant knew alcohol made him violent.

3. Analysis

3.1 Precedents Cited

The Court anchored its approach to the no-case submission in a line of authority on the Galbraith test and the treatment of circumstantial cases:

  1. Galbraith [1981] 2 All ER 1060
    The “classic” test: whether there is evidence on which a jury, properly directed, could convict. It distinguishes between (i) cases with no evidence and (ii) cases where the evidence is tenuous/depends on credibility—in the latter, issues are ordinarily for the jury.
  2. R v Bokkum (March 2000, unreported)
    In circumstantial cases, the prosecution need not exclude every alternative inference as a precondition of leaving the case; the question is whether the inference contended for is properly open to the jury.
  3. R v Jabber [2006] EWCA Crim 2694, R v Goring [2011] EWCA Crim 24
    These decisions approved the Bokkum approach and emphasised the limited gatekeeping role of the judge where inferences are available.
  4. R v Goddard [2012] EWCA Crim 1756
    The Court quoted the key summary at [36], stressing that even where the judge’s decision “does involve the rejection of all realistic possibilities consistent with innocence”, the controlling question is whether a reasonable jury (not all reasonable juries) could draw the adverse inference on one possible view of the evidence, taking the prosecution case at its highest.

The present decision is best understood as an orthodox but pointed application of Goddard: inferences, timing, and the overall evaluation of a key witness’s account were matters for the jury, not grounds to stop the case.

3.2 Legal Reasoning

A. “Direct evidence” can be enough to defeat a no-case submission even where surrounding circumstances are disputed

The Court accepted that there was a time gap between the deceased being last seen on the bank and a later witness’s observation in the water. Nonetheless, the trial judge was entitled to treat the account—of the applicant holding the deceased such that her head was fully submerged—as direct evidence of the fatal act. Questions about:

  • the precise timing of submersion,
  • what the applicant was trying to do (rescue vs restraint), and
  • the reliability/interpretation of the witness’s description,

were for the jury. The Court’s treatment reinforces a practical point: once there is evidence capable of establishing the actus reus of killing, challenges that the account is “really” exculpatory ordinarily go to weight, not to whether the case may be left.

B. Circumstantial inference: the prosecution need not disprove every alternative at the no-case stage

The Court went on to hold that, even absent the “direct evidence” strand, the jury could properly infer guilt from multiple circumstances, including:

  • evidence suggesting the applicant impeded rescue efforts;
  • initial reluctance to have police called;
  • medical evidence tending away from accidental explanation for a serious leg fracture;
  • numerous injuries found on the deceased; and
  • inconsistent and demonstrably false accounts given at the scene (including a “suicide text” narrative contradicted by CCTV showing the couple arriving together).

This is a textbook Goddard application: the judge’s task is not to decide the best interpretation, but to decide whether a reasonable jury could, on one possible view, draw the inference of guilt.

C. Bad character: permissible as supporting context, impermissible as sole or main proof

The defence argued that any conviction would be “wholly or mainly” founded on bad character (a history of violence towards partners, including prior convictions and similar fact evidence). The Court rejected that characterisation. It stressed that:

  • the prosecution case on the night’s facts was strong (direct and circumstantial); and
  • the bad character evidence was therefore not a “prop” for a weak case, but evidence consistent with and supportive of the prosecution narrative.

The reasoning draws an implicit boundary: the risk that a jury reasons “once violent, always violent” is at its highest where the contemporaneous evidence is thin; where the contemporaneous evidence is substantial, the bad character evidence more readily operates as context (subject to proper directions), rather than as a substitute for proof.

3.3 Sentencing Reasoning

The sentencing framework was orthodox: the judge adopted the 15-year starting point under paragraph 5 of Schedule 21 to the Sentencing Act 2020 and uplifted to a 21-year minimum term for aggravating factors, with limited mitigation.

The Court endorsed three practical propositions:

  • No double counting: the contemporaneous offences (counts 2 and 3) could aggravate the minimum term, and the applicant’s broader record/disregard of orders could separately aggravate it.
  • Domestic context and prior orders matter: describing breaches of domestic protection orders as “relatively minor” was inconsistent with the demonstrated pattern of coercive and violent behaviour.
  • Intoxication is not mitigation where foreseeably violence-inducing: where the offender knows alcohol triggers violence but drinks anyway, intoxication may aggravate rather than mitigate.

Ultimately, the Court treated the quantum of uplift (from 15 to 21 years) as a matter within the trial judge’s evaluative discretion, especially given the judge’s oversight of the trial evidence.

4. Impact

  • No-case applications in homicide: The decision is a reminder that where there is evidence of the fatal mechanism (here, the victim’s head held submerged), the Court will be slow to reframe disputes about interpretation as “no evidence” points.
  • Circumstantial proof and lies: The Court’s emphasis on demonstrably false early accounts (and “carefully trailing” a narrative at the scene) underlines the continuing forensic importance of CCTV/telephony evidence in testing whether an account is merely inconsistent due to intoxication or intentionally false.
  • Bad character in domestic abuse homicide: While not changing the law, the judgment illustrates how prior domestic violence may legitimately strengthen inference-based reasoning when there is already substantial incident-specific evidence—reducing the traction of the “wholly or mainly” objection in factually strong cases.
  • Sentencing messaging: The Court reinforced that alcohol dependence does not readily mitigate serious domestic violence, particularly where the offender has insight into alcohol-fuelled risk yet persists.

5. Complex Concepts Simplified

Submission of no case to answer
A mid-trial request to stop the case going to the jury because the prosecution evidence is legally insufficient. The judge does not decide guilt; the judge asks whether a properly directed jury could convict on the evidence taken at its highest.
Direct vs circumstantial evidence
Direct evidence describes the act itself (e.g., a witness seeing conduct that could cause death). Circumstantial evidence requires inference from surrounding facts (e.g., inconsistent accounts, injuries, conduct at the scene).
“A reasonable jury” (not “all reasonable juries”)
From R v Goddard [2012] EWCA Crim 1756: the question is whether at least one reasonable jury could convict on one possible view of the evidence—not whether conviction is inevitable.
Bad character evidence
Evidence of prior misconduct used for specific purposes (such as propensity). It must not become the main basis for conviction if the incident evidence is otherwise inadequate; however, it can legitimately support a strong factual case.
Minimum term (life sentence)
In murder cases, life imprisonment is mandatory. The court sets a “minimum term” to be served before parole eligibility, using Schedule 21 starting points and then adjusting for aggravation/mitigation.
Totality
A sentencing principle ensuring the overall sentence is just and proportionate when multiple offences are involved—avoiding an overall result that is either crushing or unduly lenient.

6. Conclusion

R. v Morgan [2026] EWCA Crim 472 reaffirms the limited role of the judge on a no-case submission: where a reasonable jury could convict on one possible view of the evidence—whether direct, circumstantial, or (appropriately) supported by bad character—the case must be left to the jury. The decision also confirms a robust sentencing stance in domestic homicide: prior domestic violence, breach of protective orders, and knowingly violence-inducing intoxication can justify substantial uplift from the Schedule 21 starting point, without amounting to double counting.