Sexual Assault, Starting Points, and Court of Appeal Panel Composition: A Chilling Effect on Individualized Sentencing?

By: Jennifer Koshan

PDF Version: Sexual Assault, Starting Points, and Court of Appeal Panel Composition: A Chilling Effect on Individualized Sentencing?

Case Commented On: R v Gashikanyi, 2017 ABCA 194 (CanLII)

On the first day of summer, the Alberta Court of Appeal released a decision that has turned up the heat on the approach to sentencing in this province. R v Gashikanyi, 2017 ABCA 194 (CanLII), was the hottest case on CanLII this past week, the Court of Appeal Decision of the Week in Eugene Meehan’s Supreme Advocacy newsletter, and the subject of several media stories (see e.g. Alberta Court of Appeal justice issues scathing critique of his own court; Judge slams Alberta Court of Appeal for potential appearance of bias; Alberta court of appeal judge calls for random assignment of judges to panels). Gashikanyi deals with the propriety of a starting point approach to sentencing, an approach that Justice Ronald Berger has previously critiqued and further critiques here, receiving some support from Justice Brian O’Ferrall. But Justice Berger did not stop there — he called into question whether Court of Appeal justices are bound by horizontal precedent (i.e. decisions of their own court), and criticized the way that Alberta judges are assigned to appellate hearings, suggesting a possible lack of impartiality that Justice O’Ferrall and Justice Patricia Rowbotham (dissenting) clearly distanced themselves from. The fact that this discussion took place in the context of a sentence appeal for sexual interference contributes to the heat caused by this decision in light of the intense public scrutiny surrounding sexual assault law recently.

In this post, I review the Court of Appeal’s approach to sentencing starting points generally and in the area of sexual offences as background to the decision in Gashikanyi. I also explore the ramifications of Justice Berger’s statements about horizontal precedent and appellate panel assignments, bringing into the discussion the recent calls for judicial education on sexual assault law and social context.

A Brief History of Starting Point Sentences at the Alberta Court of Appeal

The Alberta Court of Appeal has used a starting point approach to sentencing for some time now. Following this approach in the sexual assault context in R v Sandercock, 1985 ABCA 218 (CanLII), the Court described it as follows:

The sentencing process now adopted by this Court is to state typical categories with precision, and to acknowledge at the same time that each actual case presents differences from the archetypical case. These differences might mitigate or aggravate. Nevertheless, the idea of a typical case affords a starting-point for sentencing because one can state a precise sentence for that precise category. An actual sentence in a real case will vary upwards or downwards from that depending upon the balance of the factors present in the actual case. (at para 6)

The Court explained the rationale for the starting point approach this way:

This Court has a duty to offer guidance in the form of a statement of typical cases and starting-points. Sentencing courts, in turn, are asked to acknowledge the starting-point and then summarize the relevant factors before passing sentence. We thus have not the injustice of uniform sentences but the justice of a uniform approach. Dangerous rigidity is avoided because there are no arbitrary end-points. Nor is there real disparity, because all sentences of the same genre start at the same point and differences are rationally explained. (at para 8)

The Court of Appeal has also used the starting point approach to sentencing for other offences, for example robbery (see R v Johnas, 1982 ABCA 331 (CanLII) at para 19, establishing a starting point of three years’ imprisonment for the category of “unsophisticated armed robbery of unprotected commercial outlets in the absence of actual physical harm to the victim and with modest or no success”) and cocaine trafficking (see R v Maskell, 1981 ABCA 50 (CanLII), establishing a starting point of three years’ imprisonment; reaffirmed in R v Rahime, 2001 ABCA 203 (CanLII)). However, it has rejected the use of starting points for offences that are too varied in nature, such as manslaughter (see e.g. R v Fengstad, 1994 ABCA 167 (CanLII); R v Laberge, 1995 ABCA 196 (CanLII); R v Cooper, 2003 ABCA 228 (CanLII)) and infant child abuse (see R v Nickel2012 ABCA 158 (CanLII), cited in R v Hajar, 2016 ABCA 222 (CanLII) at para 263).

A starting point of three years’ imprisonment for “major sexual assaults” was established by the Court in Sandercock. Major sexual assaults were defined as cases:

where a person, by violence or threat of violence, forces an adult victim to submit to sexual activity of a sort or intensity such that a reasonable person would know beforehand that the victim likely would suffer lasting emotional or psychological injury, whether or not physical injury occurs. … This category, which we would describe as major sexual assault, includes not only … rape, but obviously also many cases of attempted rape, fellatio, cunnilingus, and buggery where the foreseeable major harm which we later describe more fully is present. (at para 13)

This approach to sentencing for major sexual assaults was reaffirmed by a majority of the Court of Appeal in R v Arcand, 2010 ABCA 363 (CanLII). In Arcand, the Court was asked to reconsider a number of its decisions that the Crown claimed had failed to follow the Sandercock starting point approach. The majority of the Court, in reasons written by Chief Justice Catherine Fraser, Justice Jean Côté and Justice Jack Watson, provided a lengthy and detailed judgment justifying the starting point approach in light of the requirements for sentencing in the Criminal Code, RSC 1985, c C-46 and Supreme Court of Canada jurisprudence (at paras 137-146). It also responded to criticisms of the starting point approach, including the arguments that starting points translate into minimum sentences in practice (at para 131); that they compromise an individualized approach to sentencing and consideration of the accused’s good character as a mitigating factor (at paras 132-136); and that use of ranges rather than starting points was a more appropriate approach to sentencing (at paras 147-160). The majority expressed the view that the Court of Appeal’s starting points are binding not just on lower court judges, but on the Court of Appeal as well (at paras 185-208). On this basis, it held that the “Reconsideration Cases”—R v Kain, 2004 ABCA 127 (CanLII), R v Nicholson, 2004 ABCA 310 (CanLII), R v Jefferson, 2008 ABCA 365 (CanLII), and R v White, 2008 ABCA 328 (CanLII) — were “wrongly decided and are overruled as precedent” (at para 250).

Justice Berger wrote judgments in three of the four Reconsideration Cases—Kain, Jefferson and White—expressing a preference for an approach to sentencing that relies on ranges rather than starting points. His discomfort with starting points in sexual assault cases appears to go back as far as 1998, shortly after he was appointed to the Court of Appeal in 1996 (see R v Waldner, 1998 ABCA 423 (CanLII)). This discomfort also extends beyond the sexual assault context (see e.g. R v Ilesic, 2000 ABCA 254 (CanLII), where Justice Berger, in dissent, argued that the starting point for cocaine trafficking was “no longer dispositive” (at para 4)).

It is not only Justice Berger who has expressed concerns about the starting point approach to sentencing and the extent to which appellate courts should readily overturn sentences that stray from starting points. The concurring justices in Arcand — Justices Constance Hunt and Clifton O’Brien — did so as well. They cited Supreme Court of Canada decisions in R v McDonnell1997 CanLII 389 (SCC)R v Stone1999 CanLII 688 (SCC), and R v Wells2000 SCC 10 (CanLII), as authority for their conclusion: “we do not think the Supreme Court’s acceptance of guideline sentencing permits it to be elevated, in effect, to a rule of law. In our view, that is the outcome of the majority judgment in this case” (at para 352). Stated differently, “neither the failure to advert to a starting point sentence nor departure from one is, alone, an error permitting appellate intervention” (at para 440).

The debate about the starting point approach to sentencing surfaced again in R v Hajar, 2016 ABCA 222 (CanLII). In Hajar, a majority of the Court of Appeal (Chief Justice Fraser and Justices Marina Paperny and Jack Watson) wrote of “the imperative need for a starting point for serious acts of sexual interference” (at para 7). The majority imposed a starting point of three years’ imprisonment for cases of “major sexual interference where the offender is an adult” (at para 11). Similar to Sandercock, it defined “major sexual interference” as cases “where the sexual conduct is a serious violation of the physical and sexual integrity of the child and is of a nature or character such that a reasonable person could foresee that it is likely to cause serious psychological or emotional harm, whether or not physical injury occurs” (at para 10). The majority also held that de facto consent on the part of the complainant, and the absence of exploitation by the accused, should not be considered mitigating factors (at paras 84-111). Justice Myra Bielby concurred with the majority’s imposition of a starting point sentence of three years’ imprisonment for major sexual interference, but disagreed with its approach to exploitation, arguing that the starting point should be inapplicable where the sentencing judge finds that the presumption of exploitation has been rebutted (at paras 172-186). Justice Frans Slatter dissented. Although he stated that he agreed with the starting point approach to sentencing generally, he disagreed with imposing a starting point for offences of major sexual interference, finding that this category had “too many combinations and permutations of factual situations” (at para 268). For further analysis of Hajar, see here.

It thus seems safe to conclude that the starting point approach to sentencing has been and continues to be a hot topic within the Court of Appeal. Disagreements exist on whether starting points are appropriate at all, whether they are appropriate for particular categories of offences, and if so, how those categories should be defined, and what the consequences should be when sentencing judges fail to follow specific starting points.

 The Gashikanyi Decision

It is in this context of disagreement that Justice Berger’s critique of starting points was made in Gashikanyi. Gashikanyi involved an offence that could be categorized as one of major sexual interference based on the majority reasons in Hajar. The facts are described at para 55:

On July 7, 2010, the respondent, age 33, was driving when he saw a young girl, age 14, and her female cousin, age 18, at a bus stop. The two girls had run away from home. The respondent offered to take the girls to his home and feed them. All three spent the night in the respondent’s bed. He twice had protected sexual intercourse with each girl and the following morning had unprotected sexual intercourse with the fourteen year old.

Gashikanyi received a sentence of two years less one day imprisonment followed by one year of probation. He was released from prison in January 2015 after spending 16 months in custody.  The Court of Appeal (Justices Berger, Brian O’Ferrall and Patricia Rowbotham) heard the Crown sentence appeal in December 2015, and deferred their judgment to await the outcome of Hajar and to consider the parties’ supplementary written arguments in relation to that decision.

Justice Berger stated at the outset of Gashikanyi that he believes Hajar “was wrongly decided… rendering problematic the precedential value of the majority decision” (at para 2A). He also foregrounded his critique of starting points, arguing that this approach “constrains the “wide latitude” and “broad discretion” accorded to sentencing judges by the Supreme Court of Canada, stifles that sentencing discretion and results in a chilling effect on the ability of sentencing judges to craft individualized dispositions” (at para 2C).

Building on these points, Justice Berger questioned the doctrine of horizontal precedent, and —setting up his critique of appellate panel assignments— stated that “[a] “precedent” may be nothing more than the product of the assignment of a like-minded three or five person panel to hear an appeal” (at para 9). He gave a number of examples, most of them quite dated, of appellate court judges “following their legal conscience and refusing to follow their court’s previous decisions, particularly … in criminal cases” (at para 10). Justice Berger did not refer to any Supreme Court of Canada jurisprudence on horizontal precedent. In contrast, Justice Rowbotham referenced Saskatchewan Federation of Labour v Saskatchewan2015 SCC 4 (CanLII) at para 137, where Justices Rothstein and Wagner (dissenting in part) stated that “To overrule a precedent is to displace community expectations founded on that decision.” Justice O’Ferrall noted that Saskatchewan Federation of Labour did not preclude appellate courts or lower courts from “distinguish[ing] a decision of a prior panel or provid[ing] principled reasons for finding that prior decision inapplicable” (at para 82).

None of the Justices referred to Canada (Attorney General) v Bedford, 2013 SCC 72 (CanLII), where Chief Justice Beverley McLachlin, writing for a unanimous Supreme Court, stated that “Certainty in the law requires that courts follow and apply authoritative precedents. Indeed, this is the foundational principle upon which the common law relies” (at para 38). At the same time, lower courts can revisit vertical precedents (those from higher courts) “if new legal issues are raised as a consequence of significant developments in the law, or if there is a change in the circumstances or evidence that fundamentally shifts the parameters of the debate” (at para 42). Where courts are reconsidering their own precedents horizontally, this is “a balancing exercise, in which the Court must weigh correctness against certainty” (at para 47, citing Canada v Craig2012 SCC 43 (CanLII) at para 27. See also Carter v Canada (Attorney General), 2015 SCC 5 (CanLII)).

Although he concurred with Justice Berger’s reasons on precedent (at para 79), Justice O’Ferrall did not go so far as to say that Hajar was wrongly decided. Instead, he noted that Hajar was released after the sentence in Gashikanyi was rendered, and in any event — similar to the reasons of the concurring justices in Arcand — “while starting points may be helpful to sentencing judges, they are not binding in the sense that it is not an error for a sentencing judge to arrive at a fit sentence without regard to a prescribed starting point” (at para 77).

Justice Rowbotham did consider herself bound by the majority ruling in Hajar, and noted that “the basis upon which [her] colleagues conclude that they are not bound by Hajar, was not argued by the respondent…  [and that] the bulk of the authorities upon which Berger JA relies were never brought to the attention of the parties.” In her view, this was unfair (at para 103). She also noted that Hajar was a “Reserved Judgment” of the Court as opposed to a “Memorandum of Judgment” (at para 101). I will not get into detail on this aspect of her disagreement with the majority (see Justice Berger’s reasons at para 15); suffice it to say that the Court of Appeal has written previously on the differences between reserved judgments (which all members of the Court of Appeal have had an opportunity to comment on) and memoranda of judgment (which are not read by all members of the Court and have less precedential value – see Arcand for further discussion).

I am inclined to agree with Justice Rowbotham’s assessment of the binding nature of Hajar. This approach is consistent with the recent Supreme Court decision in Bedford on precedent. Moreover, as noted and indeed applied in Arcand, the Alberta Court of Appeal has a specific process for the reconsideration of its own precedents, another basis for following horizontal precedent where this process is not invoked. It also seems problematic to reach the conclusion that a case decided only a year ago, where the approach was essentially agreed upon by four of the five members of the same Court, was “wrongly decided.”

Having found that he was not bound by Hajar, Justice Berger went on to critique the majority’s imposition of a sentencing starting point in that case. While much of that critique is similar to his previous arguments against starting points, he found new support for this position in the Supreme Court of Canada’s decision in R v Lacasse2015 SCC 64 (CanLII). Justice Berger argued that although the majority in Hajar purported to follow Lacasse, it “seems not to have appreciated Lacasse’s focus on the increased force of appellate deference” and its reminder “that departure from sentencing ranges or starting points are not only permitted but to be expected” (Lacasse at para 49, cited in Gashikanyi at para 24). He also stated that Lacasse found that “individualization takes priority over parity” in sentencing; the Supreme Court’s point, however, was that “Individualization and parity of sentences must be reconciled for a sentence to be proportionate” (Lacasse at para 53). Justice Berger closed his critique of starting points by noting that none of the leading judgments of the Alberta Court of Appeal in this area — Hajar, Arcand, and Rahime — could have been expected to be appealed to the Supreme Court, given that the Court accepted the argument of the Crown in each case, and the defence had no incentive to appeal as the accused was not re-incarcerated as a result of any of the decisions (at para 36). To say that sentencing starting points are something that the Supreme Court should weigh in on more directly is therefore offset by this practical consideration.

Other aspects of Justice Berger’s critique of Hajar are more specific to the offence of sexual interference. He questioned the need for a sentencing starting point for this category of offences, noting that the cases relied upon by the majority in Hajar to justify its claim that “sentencing has been all over the place” in this category do not bear this claim out (at para 30, citing Hajar at para 72). Justice Berger agreed with Justice Slatter’s dissenting opinion in Hajar that sexual interference is not an appropriate offence for the imposition of a starting point (at para 51). He also argued that the particular starting point adopted in Hajar was arbitrary and akin to “throwing a dart at a board” (at para 33).

In the end, Justice Berger found that the sentence imposed in Gashikanyi “expose[d] no error justifying appellate intervention”, and dismissed the Crown’s sentence appeal (at para 67).

As noted, Justice O’Ferrall also believed that Hajar was not binding on the sentencing judge, but for different reasons. He agreed with Justice Berger that Gashikanyi’s sentence was not demonstrably unfit and that the appeal should be dismissed.

Justice Rowbotham, finding herself bound by Hajar, would have allowed the Crown’s sentence appeal and imposed a sentence of three years. She noted the presence of several aggravating factors — the 20 year age difference between the accused and complainant, the fact that the complainant was “a vulnerable youth in need of food and shelter”, the respondent’s lack of responsibility for the offence and blaming of the victim for his predicament, and the risks of sexual disease and pregnancy (at para 108). She also believed that there were “few if any mitigating factors” and that some of the factors relied on by the sentencing judge as mitigating were simply non-aggravating (at para 109). Her judgment suggests that a sentence of greater than three years may have been appropriate, but certainly supports the view that if Hajar is binding, the sentence in Gashikanyi was demonstrably unfit. Even ignoring Hajar, Gashikanyi could be seen as involving a “major sexual assault” as defined in Sandercock and Arcand, again rendering the sentence of two years less one day demonstrably unfit in all the circumstances.

Appellate Court Panel Composition

This takes us to Justice Berger’s comments on the assignment of appellate panels. His concern is that “this Court has failed to establish and abide by a protocol that provides for the random assignment of judges to sentencing panels” (at para 70) and that “the presence of individual discretion in a system of assignment poses a risk that some may think that panelists will be selected based on their perceived predispositions” (at para 71). Not only is perception a problem, but “An appellate court that utilizes discretionary non-random methods to assign (or to replace an assigned judge) leaves open the potential for manipulation” (at para 71). The basis for his comments was said to be “publicly accessible records of this Court” which “demonstrate that the failure to implement and adhere to an objective protocol for the random assignment of judges has resulted in significant discrepancies in both the number of sentencing panels on which some judges of the Court sit and a marked difference in the number of sentence appeals heard by certain justices of the Court as compared with their colleagues.” The result, according to Justice Berger, “is a disproportionate opportunity afforded to certain judges to shape the jurisprudence of the Court” (at para 74).

Although Justice Berger framed his critique as being against “the Court”, it is clear that the “individual discretion” in assignments that he assails is that of the Chief Justice.

Justice O’Ferrall referred to the Chief Justice explicitly in his reasons, opining that “there are situations in which it is imperative that a chief justice have the discretion to constitute panels and to assign panels to appeals. Indeed, I would go so far as to say that the proper assignment of judges to panels and panels to appeals is one of the critical responsibilities of a chief justice” (at para 83). As examples of such situations, he cited the need for expertise in particular areas of law such as criminal law, confessing that “many of our judges have limited criminal experience” (at para 85). In his view, “it is critical that those judges who do have criminal expertise are assigned to sit on sentencing panels, not because they dictate the result, but because they provide those of us without that expertise with the tools necessary to independently and impartially decide whether the sentencing judge imposed an unfit sentence in any particular case” (at para 85). Other rationales for non-random assignments include the need to balance uneven workloads and to avoid conflicts and appearances of bias (at para 86). Justice O’Ferrall closed by noting that “the Chief Justice does not have the time to micro-manage the constitution of panels or the assignment of specific cases to particular panels. As a consequence, assignments will mostly be random as Justice Berger quite correctly says they should be” (at para 87).

Justice Rowbotham called Justice Berger’s allegations about the assignment of judges “completely baseless” and argued that his “failure to identify such ‘publicly accessible records’ seriously undermines the foundation” of his contentions (at paras 117, 114). She rejected “in the strongest possible terms” any suggestion that justices of the Court of Appeal are not deciding sentence appeals impartially (at para 117).

I agree with Justice Rowbotham’s response, in that it is very difficult to evaluate Justice Berger’s critique in the absence of hard evidence or examples. Readers may have noticed that Chief Justice Fraser and Justice Jack Watson sat on both the Arcand and Hajar panels — is this the sort of “publicly accessible record” that Justice Berger was referring to as the basis for his allegation of non-random assignments? Even if so, should we critique this overlap, or should we see it as a justifiable one, based on the need for criminal law expertise identified by Justice O’Ferrall?

I would actually go further, and argue that specific expertise in sexual assault law may sometimes be a valid basis for the Chief Justice’s assignment of panel members. An important part of the context going forward may be Bill C-337, the Judicial Accountability through Sexual Assault Law Training (JUST) Act. This Act was introduced in the House of Commons by former Leader of the Opposition Rona Ambrose in February 2017. Although bills introduced by opposition MPs rarely make it very far through Parliament, the JUST Act received broad support from all parties and passed Third Reading in the House of Commons on May 15, 2017. If passed by the Senate in its current form, the Bill will require new federally appointed judges to receive education on sexual assault law and social context (including rape myths) (see section 2(2)) and will require the Canadian Judicial Council to report annually on the number of judges from each court who have attended such education sessions as well as “the number of sexual assault cases heard by judges who have never participated” (see section 4).

The Bill was studied by the Status of Women Committee — and in the interests of disclosure, I was a witness at the Committee’s hearings. I spoke in favour of a requirement for education on sexual assault law not only for new judges, but for all judges. Having participated in several judicial education sessions on sexual assault law, my view is that these sessions do not undermine judicial independence; they are aimed at avoiding errors of law, including those that occur when judges rely on rape myths in their reasoning. My experience has been that judges take this education very seriously. However, if judges decline to take whatever training on sexual assault law is made mandatory, that may also be a valid basis for a chief justice to use a non-random assignment of panel members — i.e. to decline to assign judges to sexual assault cases in such circumstances. Judges must jealously guard their independence, and they should not allow themselves to be swayed by popular opinion, but the public’s lack of confidence in the courts’ handling of sexual assault cases is not something that courts or their chief justices can ignore. One also has to wonder what impact Justice Berger’s critiques of horizontal precedent and panel assignments will have on public confidence in the judiciary, even though he claimed to be raising his critiques in order to foster such confidence.

This post may be cited as: Jennifer Koshan “Sexual Assault, Starting Points, and Court of Appeal Panel Composition: A Chilling Effect on Individualized Sentencing?” (30 June, 2017), online: ABlawg,

To subscribe to ABlawg by email or RSS feed, please go to

Follow us on Twitter @ABlawg


About Jennifer Koshan

B.Sc., LL.B (Calgary), LL.M. (British Columbia). Professor. Member of the Alberta Bar. Please click here for more information.
This entry was posted in Criminal. Bookmark the permalink.

2 Responses to Sexual Assault, Starting Points, and Court of Appeal Panel Composition: A Chilling Effect on Individualized Sentencing?

  1. This comment relates primarily to the distinction drawn by ABCA between “Reserved” judgments and “Memorandum” decisions. Prof. Koshan does not really address this, but it underlies so many of the issues raised by Gashikanyi. That a decision should acquire some higher precedential force merely because other members of the Court who did not sit on the appeal, did not hear any of the argument, had no opportunity to question counsel, and most likely have not even read the factums have nevertheless had the opportunity to “comment” on the decision surely strikes at the very core of the rule of law. Further, we do not know how many, if any, judges availed themselves of this opportunity, what their comments were, or how they were dealt with or addressed. We also do not know how and why some decisions are selected for this dubious honour. I hope that the SCC gets the opportunity to address this and all the other issues in Gashikanyi. For the record, I tend to agree with Justices Berger and O’Ferrall that mandatory “starting points” suffer from many of the same defects as do mandatory “minimums” – it seems to me that the real difference is that judges strike down legislative mandatory minimums as unconstitutional in many cases but have difficulty conceiving of the reality that their own behaviour with “starting points” may well suffer from distinctly analogous defects in the minds of many of us.

  2. Jennifer Koshan says:

    Thanks for your comment, Chris.

    As noted in my post I decided to leave aside the issue of the precedential value of memoranda versus reserved reasons for judgment. To provide more context to ABlawg readers, this distinction is from Court of Appeal Practice Direction A4. I believe this Practice Direction is still in effect but could not find it on the ABCA website (including in the consolidated practice directions); if ABlawg readers know where to find it please weigh in. Justice Rowbotham’s reasons in Gashikanyi suggest that the Direction is still being applied (see para 101, where she states that “Hajar was a reserved judgment of the court which means each member of this court had an opportunity to comment on it”, suggesting reserved reasons have more weight than memoranda for decision).

    This issue of differential weight is one that Justice Berger has raised in previous judgments – in fact, in two of the reconsideration cases from Gashikanyi. In R v Jefferson, 2008 ABCA 365, Justice Berger noted (at paras 10-11) that the Crown and defence bar do not follow the Practice Direction and it is not legally binding. He also noted that the Supreme Court of Canada does not follow it, citing R v McDonnell, [1997] 1 SCR 948, where the Court considered memoranda of judgment and judgments delivered from the bench when assessing the range of sentences for sexual assault in Alberta. See also his reasons in R v White, 2008 ABCA 328 at paras 25-6 and R v Lee, 2012 ABCA 17 at paras 62-73).

    As I said in a post I wrote on Jefferson (, I tend to agree that the distinction in weight given to different categories of judgment is problematic. At the same time, I find Justice Rowbotham’s other reasons to treat the ABCA’s decision in Hajar as binding to be persuasive.

Leave a Reply

Your email address will not be published. Required fields are marked *