February 11, 2025
The Supreme Court of Canada (“SCC”) has dismissed the appeal in R v Stevenson, 2024 SCC 41 [Stevenson], upholding a trial judge’s decision to convict based on the evidence of an unsavoury or Vetrovec witness. This case was heard by the SCC as of right from the Saskatchewan Court of Appeal (“SKCA”) 2024 SKCA 40 [Stevenson SKCA], where the court was divided over whether the trial judge could reasonably find the testimony of the appellant’s former gang accomplice credible and reliable enough to ground a conviction.
Facts
In December 2016, two men robbed a pizza restaurant in Regina, Saskatchewan. One man wielded a gun, while the other held a bag to collect money from the cash register (Stevenson SKCA, para 5). While the robbery was captured on video, both men had their faces covered, so the crime remained unsolved for more than three years (Stevenson SKCA, para 6).
In an unrelated incident, the Regina police arrested a man named C.S. in December 2019. C.S. had a lengthy criminal record and had spent considerable time in a gang called the Indian Mafia. After C.S. was arrested, he offered to help the police by providing information about crimes committed by other Indian Mafia members in exchange for entry into the witness protection program. The police accepted the offer, and showed C.S. surveillance footage from seven unsolved robberies across the city. He was able to identify the appellant as the man holding the gun in the pizza restaurant robbery (Stevenson SKCA, para 7).
Judicial History
Trial
The key issue at trial was identity. To prove identity, the Crown relied heavily on C.S.’s testimony. While the Crown called two other witnesses, neither of them were able to identify the perpetrators of the robbery, and there was no other evidence implicating the appellant in the crime (Stevenson SKCA, para 8).
The trial judge held a voir dire to determine whether C.S. was able to provide recognition evidence, i.e., whether he was able to testify as to his recognition of a person he knew. Importantly, this is distinct from eyewitness identification evidence (Stevenson SKCA, para 62). The trial judge determined that C.S. was sufficiently familiar with the appellant to provide this evidence, and C.S. testified at trial. C.S. testified that he had known the appellant for nearly 20 years, they had both been members of the same gang and interacted almost every day, and that he could identify the appellant’s distinct limp due to a leg injury (Stevenson SKCA, para 10). C.S. also testified that at the time of the robbery, the appellant owed approximately $30 000 to the Indian Mafia gang, and that the appellant had told him about the robbery (Stevenson SKCA, para 11). Furthermore, C.S. recognized the gun the appellant used during the robbery because it belonged to the Indian Mafia gang. He admitted to having used the gun in the past, and was able to describe it in detail (Stevenson SKCA, para 12).
The trial judge accepted C.S.’s evidence and found his testimony implicating the appellant as one of the robbers to be credible and reliable (Stevenson SKCA, para 14). She instructed herself on the principles governing Vetrovec witnesses – witnesses that are unsavoury or have a potential interest in testifying favorably for the Crown (discussed below) – and acknowledged C.S.’s criminal record and gang involvement (Stevenson SKCA, para 17). The trial judge found C.S.’s evidence reliable primarily because of how well C.S. knew the appellant and how much time they had spent working closely together in the Indian Mafia gang (Stevenson SKCA, para 20). She was particularly persuaded by the fact that C.S. was willing to implicate a close friend and accomplice. Thus, on the basis of C.S.’s evidence alone, she was satisfied beyond a reasonable doubt that the appellant had committed the robbery. She accordingly convicted the appellant of robbery and having his face masked while committing an indictable offence (Stevenson SKCA, para 1).
Saskatchewan Court of Appeal
The appellant appealed to the Saskatchewan Court of Appeal, alleging that the trial judge had erred in convicting him on the basis of C.S.’s uncorroborated testimony, and finding C.S. credible and reliable (Stevenson SKCA, para 21). The crux of his concern was that the trial judge failed to exercise the proper caution in assessing C.S.’s credibility given that C.S. was a Vetrovec witness.
A Vetrovec witness is a Crown witness who is particularly untrustworthy, perhaps due to their immoral lifestyle choices or an interest they might have in testifying against the accused. The SCC in R v Vetrovec (1982 CanLII 20) held that when presented with a witness who cannot be trusted to tell the truth under oath, trial judges must instruct themselves or the jury with a “clear and sharp” warning about the risks of adopting their testimony without corroboration. At minimum, the warning should refer to the characteristics of the witness that bring their credibility into question, and emphasize the specific dangers inherent in using that evidence to convict (R v Khela, 2009 SCC 4, para 34).
The majority of the SKCA accepted that the trial judge had followed these requirements and dismissed the appeal. The majority commented that “Vetrovec does not stand for the proposition that a trier of fact necessarily commits a legal error by relying on the testimony of a [Vetrovec] witness.” (Stevenson SKCA, para 32). The trial judge’s Vetrovec caution explicitly recognized that C.S. had received benefits from the witness protection program and had a possible resulting interest in providing evidence favourable to the Crown (Stevenson SKCA, para 40). The appellant argued that because the trial judge found that C.S. did not have a personal animus towards the appellant or motive to testify against him, she did not grasp the essence of the Vetrovec caution. The SKCA rejected this argument and found that the trial judge approached C.S.’s evidence with an understanding of the factors she was required to consider in assessing his evidence (Stevenson SKCA, para 42). The trial judge understood that C.S. had a criminal record, was a gang member, and had a personal interest in testifying, and she also understood that she was still permitted to rely on his evidence nonetheless if she was satisfied that it was true (Stevenson SKCA, para 45). The SKCA emphasized deference to the trial judge’s Vetrovec caution, commenting that “the important question is whether a trier of fact’s confidence in an unsavoury witness is restored, not how.” (Stevenson SKCA, para 43)
In dissent, Drennan J.A. disagreed that the trial judge applied the scrutiny required of her to C.S.’s evidence, and held that it was not open to her to rely on C.S.’s evidence uncorroborated. On the spectrum of possible types of unsavoury witnesses, C.S. “presented as particularly perilous,” and the trial judge’s reasons for accepting his evidence were unresponsive to the Vetrovec concerns (Stevenson SKCA, para 78).
Drennan J.A. took issue with the trial judge’s Vetrovec caution on three points. First, she found that the trial judge erred in using evidence of C.S’s good relationship with the appellant and lack of animus towards him as a factor bolstering his credibility. She argued that a lack of a personal animus was at best a neutral factor as to whether C.S. was telling the truth, and it should not have been elevated to a factor that enhanced C.S.’s credibility (Stevenson SKCA, para 84). Because the trial judge had used this factor as a “key building block” in accepting C.S.’s evidence, it was a material error (Stevenson SKCA, para 87).
Second, Drennan J.A. argued that the trial judge engaged in circular reasoning when she found that C.S. was credible because he had accurately described in his testimony what was depicted in the security footage, despite having seen the security footage prior to testifying (Stevenson SKCA, para 90). For evidence to be confirmatory for a Vetrovec witness, it must be both independent, in that it is unconnected to the Vetrovec witness, and material, in that it supports a relevant aspect of the witness’s testimony (Stevenson SKCA, para 91). Neither of these requirements were met because C.S.’s testimony was directly tainted by his Vetrovec status, and C.S. only identified the appellant as the robber after seeing the video, even though he testified that the appellant had told him he committed the robbery (Stevenson SKCA, para 92).
Lastly, Drennan J.A. found that the trial judge was not appropriately rigorous in assessing C.S.’s evidence. The trial judge was “obliged to apply a level of scrutiny commensurate with the multiple inherent frailties in C.S.’s evidence,” and while she instructed herself on the dangers of convicting solely on the Vetrovec testimony, she did not instruct herself on the dangers inherent in recognition evidence in the Vetrovec context (Stevenson SKCA, para 102). On these bases, Drennan J.A. would have ordered a new trial (Stevenson SKCA, para 110).
Supreme Court of Canada
Both the majority and the dissent at the SCC issued brief reasons. Four members of the court — Martin, Kasirer, Jamal, and O’Bonsawin JJ — dismissed the appeal, while Rowe J in dissent would have allowed it.
The majority accepted that the trial judge applied the requisite scrutiny to C.S.’s evidence. They found that she had properly self-instructed on the Vetrovec dangers, and correctly applied the relevant legal principles bearing on C.S.’s reliability (Stevenson, paras 4-6). The trial judge’s findings of credibility were entitled to significant deference nonetheless, and the SCC saw no reason to interfere (Stevenson, paras 6-7).
Rowe J would have allowed the appeal. In his view, the fact that C.S. implicated the appellant in the robbery and told the police that the appellant had confessed to the robbery only after he had seen the surveillance footage enhanced the unreliability of his testimony (Stevenson, paras 9-10). This was the only evidence linking the accused to the crime, and required a particular degree of scrutiny, which the trial judge failed to apply (Stevenson, para 11).
Analysis
In my view, this case boils down to the appropriate level of appellate deference owed to a trial judge’s findings of credibility. Credibility and reliability assessments are trial judge’s bread and butter — they observe the witnesses, no matter how unsavoury, and are privy to the dynamics of a trial in the way that appellate judges are not. I do not think this was a case which met the high bar for appellate intervention on the trial judge’s findings of credibility. The trial judge was clearly alive to the Vetrovec dangers and considered them in her self-instruction. The SCC in Vetrovec was clear that, despite the risks, it remains open for triers of fact to convict on a Vetrovec witness’s evidence alone. This holding was meant to reflect that trial judges and juries are in closest proximity to the evidence, and are best positioned to make findings of credibility. Thus, the fact that the trial judge exercised her discretion to convict on C.S.’s evidence, when coupled with a comprehensive Vetrovec instruction, was not in itself an error that required appellate intervention. The SCC’s concise reasons reflect the clear-cut nature of this case, and I continue to appreciate their increased emphasis on deference to trial judges when assessing testimonial evidence.