A LONG-STANDING issue as regards admissibility of evidence of an accused’s bad character has been the need to distinguish situations where the prosecution seeks to use it to establish a propensity in them towards conduct of the kind in question in the case at hand from those where it seeks to use it for some other purpose. Indeed, for many years, the leading case, Makin v Attorney General for New South Wales [1894] A.C. 57 at 65 established that reasoning via propensity – also described as disposition – was forbidden. However, it was later decided in D.P.P. v P [1991] 2 A.C. 447 at 460 that that mode of reasoning was permissible as long as the general test for admissibility, namely that the probative value of the evidence exceeded its prejudicial effect, was satisfied.

By contrast, the statutory regime in the Criminal Justice Act 2003 that replaced the common law one specifically embraced propensity evidence, under the section 101(1)(d) “gateway” to admissibility, as potentially relevant to an important matter in issue between the defendant and the prosecution. Thus, section 103(1)(a) states that the matters in issue between the defendant and the prosecution include “the question whether the defendant has a propensity to commit offences of the kind with which he is charged, except where his having such a propensity makes it no more likely that he is guilty of the offence”.

That might have been thought to render propensity reasoning no longer any kind of special issue, but, in two respects, it has proved otherwise. First, in R. v McAllister [2009] 1 Cr. App. R. 129 (followed in a number of later cases), a wide view was taken of what did not count as propensity reasoning, but rather as coincidence reasoning, and, though it was not there suggested that, where taken as the latter, the admissibility test was any less stringent, the rather clear message given was that a finding that propensity reasoning was not being relied upon removed a significant worry as regards admitting bad character evidence.

The other respect in which propensity reasoning has been treated as a special issue is shown by R. v Mitchell [2017] A.C. 571, and it is the way in which that case was distinguished in R. v Kawa and Davies [2023] EWCA Crim 845 that is the subject matter of this note.

In Mitchell, the accused, charged with murdering someone by using a knife, had accepted that she had stabbed the victim, but claimed to have acted in self-defence and under provocation. The prosecution was, without defence opposition, allowed to call evidence of two other occasions on which she had used or threatened to use knives during the course of a dispute with others. Neither of these incidents had led to prosecution. Here, there was no doubt at all that the prosecution were relying on propensity reasoning. As put by Lord Kerr, “[t]his was said to be for the purpose of showing that she had a propensity to use knives in order to threaten and attack others” (at [7]). It was the failure of the trial judge to direct the jury that it was only if they found the propensity in question to have been proved to the criminal standard from the two incidents that they could use it in deciding if the accused had committed the offence charged that led the Supreme Court to uphold the Court of Appeal in Northern Ireland’s quashing of Mitchell’s conviction.

In R. v Kawa and Davies, the two accused had been convicted of the murder of Nsaka by stabbing him in the chest. Kawa’s defence was that it was Nsaka who had had possession of the knife and had used it to attack Kawa. The knife had been knocked to the ground and, during the ensuing struggle, must accidentally have entered Nsaka’s chest. Davies’s defence was that he had not been armed and had played no part in the violent episode. It followed that one important issue at the trial was whether the jury could be sure that it had been Kawa, and not Nsaka, who had brought the knife to the scene (at [9]).

The prosecution sought leave to call evidence of the earlier convictions of both accused for possession of knives, as well as of an occasion on which they had purchased at a supermarket two blocks containing a number of knives. It also sought to call evidence against Kawa of him being found by the police, on another occasion, wearing a coat and gloves and carrying two knives. The trial judge allowed all three items of evidence to be called as being relevant and admissible on the central issues in the case, namely who had the knife in their possession and, if it was Kawa or Davies, whether the other one would have known that to be so (at [18]). Though, in so ruling, he did not say that the prosecution were not relying on propensity reasoning, Holroyde L.J., delivering the judgment, remarked that the prosecution application at trial had made no reference to any propensity on the part of either accused and that their argument on appeal was that the evidence was relevant to “whether a defendant had a knife at the time of the stabbing, not whether he had a propensity to behave in a particular way” (at [15]). He then went on to hold that the evidence in question had been adduced not in order to establish a propensity to carry knives, but rather to prove that one of the two accused, and not Nsaka, had brought the knife to the scene (at [37], [40] and [43]).

Somewhat awkwardly, the trial judge had, at one point in his jury directions, referred to the prosecution saying that “it was the [two accused’s] habit to go armed” (at [20], emphasis added). Holroyde L.J. (at [40]) accepted that that reference did suggest a propensity to behave in a particular way, but stressed that it was not echoed elsewhere in the directions. With respect, that gives the game away. Surely, to adduce evidence of the previous carrying or possession of knives does tend to establish a propensity to carry or possess them, with that propensity then being quite properly capable of being used to decide if one of the accused, rather than the victim, had brought this knife to the scene.

The court found in R. v Okokono and Wilson-Moonie [2014] EWCA Crim 2521 support for its view that propensity reasoning had not been relied upon in the instant case. There, as in Kawa and Davies, the victim had been stabbed to death during a gang fight. However, in other respects, the facts were significantly different. Though the crime charged was also murder by use of a knife and the bad character evidence adduced was of Wilson-Moonie’s conviction for possession of a lock knife, it was not alleged that Wilson-Moonie himself had wielded the knife in question. Moreover, the killing had, at least partly, been one of revenge for the earlier killing of Wilson-Moonie’s friend, Steven Lewis, during a gang fight at which Wilson-Moonie had also been present. Indeed, it was at that very fight that he had been found in possession of the lock knife. The Court of Appeal did indeed rule that evidence properly admissible, and, in doing so, clearly held it not to amount to propensity evidence. Rather, as Holroyde L.J. himself stated in Kawa and Davies (at [41]), the conviction was sought to be adduced “to demonstrate two things: that he was a party to the attack seeking to revenge the death of his friend in a gang–related feud which he may have triggered, and his realisation that a knife or knives might be used with lethal intent if gang violence broke out” (see [2014] EWCA Crim 2521, at [64] (Hallett L.J.)). In fact, Holroyde L.J. left out an important sentence that immediately followed, namely, “It was evidence showing the true nature and extent of [Wilson-Moonie’s] connection to the death of Steven Lewis and his affiliation to his gang”. In other words, it suggested that Wilson-Moonie would be well aware of the potential for use of knives by other gang members, as well as being himself keen to see his friend’s death avenged. Thus, it does indeed seem clear that the prosecution was not seeking to establish any propensity on the part of Wilson-Moonie himself. Therefore, the earlier case is not at all at one with Kawa and Davies itself.

At all events, a further issue arose from the fact that the judge had not directed the jury, in accordance with Mitchell, that it was only if they found the relevant propensity proved to the criminal standard that they could rely upon it. Here, precisely because the Court of Appeal found that propensity reasoning had not, in fact, been employed, it felt able to distinguish Mitchell (see [2023] EWCA Crim 845, at [43]). This seems a most unsatisfactory basis for ruling that the Mitchell direction did not need to be given.