This will be a fairly long post (approximately 8 pages). I’ll begin with the background to the Supreme Court of Canada’s latest pronouncement on starting point sentences in Friesen, using Alberta cases such as Hajar, Arcand, Gashikanyi, Godfrey and Parranto, as well as the SCC case of Lacasse. I’ll then discuss the effects of the case – and why (and how) the starting point and sentencing debate will continue to rage.
While R. v. Arcand, 2010 ABCA 363 (CanLII) is cited over 500 times over the past decade in reported decisions alone – it’s even mentioned 5 years ago in Lacasse dissent (para. 130 – parallel citation as R. v. J.L.M.A., 2010 ABCA 363) – one would have hoped the SCC wouldn’t have needed a decade to overrule it (explicitly – it was overruled somewhat less explicitly in Lacasse). Note the strong “concurring” opinion by 2 of the Justices in Arcand (Hunt and O’Brien JJ.) – basically dissenting – pointing out inconsistencies between the binding decision of Proulx and the majority’s view (see e.g. para. 361), etc.
Note that R v Hajar, 2016 ABCA 222 (CanLII) is cited favourably in Friesen in the context of harm done to victims of sexual abuse (for examples, paras. 64 and 78).
The Friesen case mostly serves to increase sentences in the context of sexual assaults, and also seeks to clarify a lot of potential issues relating to child sexual assault (e.g. “de facto” consent, categories of sexual assault, ensuring that assaults against children are sentenced more severely than those against adult, aggravating and mitigating factors etc.). Note that it overrules the lower court (MBCA)’s decision (R v Friesen, 2018 MBCA 69 (CanLII)) to reduce the sentence imposed by the judge. The way I see it – the case is a “win” for defence and sentencing judges in that it strongly denounces the approach of the ABCA in Arcand and establishes that the judge has all of the power to make the sentencing decision. At the same time, it both increases sentences and clarifies the law regarding sexual assaults against children, again, giving considerable power to the sentencing judge and more explicitly wresting it away from the appellate courts.
Arcand suggests that mere departure from a starting point can demonstrate an error in principle (italics in original, at para. 106):
This being so, deviations from the starting point in service of proportionality are an inseparable aspect of the starting point approach. Therefore, mere departure from the starting point does not, by itself, demonstrate error in principle. But a significant degree of departure may do so. A word of caution is in order. Facts relied on to deviate from the starting point should be relevant to sentence and reasonably justify deviation. It is not enough that there be differences from other cases. There always will be. And in dozens of irrelevant ways. The difference said to justify the deviation from the starting point should be a relevant difference. This is consistent with the parity principle which requires similarity of outcome for cases that are relevantly similar.
Also in Arcand (para 273, emphasis added):
Presumably because of the improper characterization of the unlawful act here, the sentencing judge effectively ignored binding authority, Sandercock, and the starting point it set for a major sexual assault, that is three years. In light of the sentence imposed and the absence of any rational justification for it, this too constitutes reviewable error. The corollary of a court of appeal’s authority to set properly defined starting points is that sentencing judges will give due consideration to those starting points and the process that starting point sentencing entails.
R. v. Lacasse, 2015 SCC 64 (CanLII) would appear to overrule Arcand on this point (paras. 60-61, emphasis added):
In other words, sentencing ranges are primarily guidelines, and not hard and fast rules: Nasogaluak, at para. 44. As a result, a deviation from a sentencing range is not synonymous with an error of law or an error in principle. Sopinka J. stated this clearly in McDonnell, although he was referring in that case to categories of assault:
… in my view it can never be an error in principle in itself to fail to place a particular offence within a judicially created category of assault for the purposes of sentencing… . If the categories are defined narrowly, and deviations from the categorization are generally reversed, the discretion that should be left in the hands of the trial and sentencing judges is shifted considerably to the appellate courts. [para. 32]
Any other conclusion would have the effect of authorizing appellate courts to create categories of offences with no real justification and accordingly intervene without deference to substitute a sentence on appeal. But the power to create categories of offences lies with Parliament, not the courts: McDonnell, at para. 33.
Note that starting points are synonymous with sentencing ranges, for the purpose of this analysis; see para. 57 of Lacasse (emphasis added):
Tariffs differ from sentencing ranges in that tariff-based sentencing is theoretically the opposite of sentence individualization, which the ranges allow: Thomas, at p. 8. On the other hand, the principle underlying the two approaches is the same: ensuring that offenders who have committed similar crimes in similar circumstances are given similar sentences. The same is true of the starting-point approach, which is used mainly in Alberta but sometimes also in other Canadian provinces: R. v. McDonnell,  1 S.C.R. 948, at para. 69. Ultimately, whatever mechanism or terminology is used, the principle on which it is based remains the same. Where sentencing ranges are concerned, although they are used mainly to ensure the parity of sentences, they reflect all the principles and objectives of sentencing. Sentencing ranges are nothing more than summaries of the minimum and maximum sentences imposed in the past, which serve in any given case as guides for the application of all the relevant principles and objectives. However, they should not be considered “averages”, let alone straitjackets, but should instead be seen as historical portraits for the use of sentencing judges, who must still exercise their discretion in each case […]
Hajar acknowledges Lacasse’s ratio that failure to follow a starting point is not an error in principle per se – albeit only mentioned in technical dissent, and by Slatter J. (paras. 260, 272) – not the majority.
I’ll cut to the chase here with the new decision of R. v. Friesen, 2020 SCC 9 (CanLII), before getting back to the contextual background. The following quote is most relevant regarding Arcand, where the latter is more explicitly overruled (emphasis added):
37 This Court has repeatedly held that sentencing ranges and starting points are guidelines, not hard and fast rules (R. v. McDonnell, 1997 CanLII 389 (SCC),  1 S.C.R. 948, at para. 33; R. v. Wells, 2000 SCC 10,  1 S.C.R. 207, at para. 45; R. v. Nasogaluak, 2010 SCC 6,  1 S.C.R. 206, at para. 44; Lacasse, at para. 60). Appellate courts cannot treat the departure from or failure to refer to a range of sentence or starting point as an error in principle. Nor can they intervene simply because the sentence is different from the sentence that would have been reached had the range of sentence or starting point been applied (McDonnell, at para. 42). Ranges of sentence and starting points cannot be binding in either theory or practice, and appellate courts cannot interpret or apply the standard of review to enforce them, contrary to R. v. Arcand, 2010 ABCA 363, 40 Alta. L.R. (5th) 199, at paras. 116-18 and 273. As this Court held in Lacasse, to do so would be to usurp the role of Parliament in creating categories of offences (paras. 60-61; see also McDonnell, at paras. 33-34). […]
41 Many practitioners, judges, and academics have consistently expressed these concerns (see, e.g., A. Manson, “McDonnell and the Methodology of Sentencing” (1997), 6 C.R. (5th) 277; J. Rudin, “Eyes Wide Shut: The Alberta Court of Appeal’s Decision in R. v. Arcand and Aboriginal Offenders” (2011), 48 Alta. L. Rev 987; L. Silver, Sentencing to the Starting Point: The Alberta Debate, May 23, 2019 (online) [link]). We realize that the Alberta Court of Appeal has repeatedly defended the utility of the starting point methodology in the face of these concerns (see Arcand, at paras. 130-46; R. v. Parranto, 2019 ABCA 457, 98 Alta. L.R. (6th) 114, at paras. 28-38; see also P. Moreau, “In Defence of Starting Point Sentencing” (2016), 63 Crim. L.Q. 345). However, this Court has not yet addressed these concerns. We make no comment on the merits of these concerns. Nor should anything in these reasons be taken to suggest that starting points are no longer a permissible form of appellate guidance. While we have determined that this case does not provide an appropriate opportunity to assess the merits of these concerns, they raise an issue of importance that should be resolved in an appropriate case.
R v Godfrey, 2018 ABCA 369 (CanLII) is problematic on multiple levels (I’ve addressed it previously here). It likely both misquotes the sentencing judge and misapplies the law, per Lacasse which had already overruled Arcand (Godfrey, at para. 8)
[…] It was an error of principle calling for appellate intervention for the sentencing judge to conclude that she was not bound by the appellate case law in Alberta. The point was summed up in Arcand […]
The damage is partially mitigated in para. 17: “The trial judge noted the starting point in passing, but abandoned it as a meaningful foundation for her analysis, and never explained why the resulting sentences were compatible with it”. As pointed out in O’Ferrall J.’s dissent:
 With respect, I cannot agree. The sentencing judge considered herself bound by the starting point. She did so explicitly, acknowledging the starting point and that she was “of course bound by the directions of the Alberta Court of Appeal”: R v Godfrey, 2018 ABPC 45 at para 31. […]
 The sentencing judge acknowledged the three-year starting point, but all parties were of the view that a sentence less than the starting point was appropriate. The question was, how much less? To answer that question, the sentencing judge sought the guidance of a whole range of authorities. But in the final analysis, as helpful as such guidance might have been, she had to decide what constituted a fit sentence for these offences and these offenders and she did so in a transparent and reasonable way.
Why Godfrey was not appealed to the Supreme Court is difficult to understand – absent the fact that the majority’s decision disincentivized the defendant to appeal, as it did not directly affect the appellant – only everyone else in Alberta who came after (note the majority’s decision in para. 22 to vary the sentence to time served in that particular case, and see Berger J.’s criticism on this point below from Gashikanyi).
R v Felix, 2019 ABCA 458 (CanLII) and R v Parranto, 2019 ABCA 457 (CanLII) (essentially “sister” decisions – one of which I discuss here) are careful to state correctly – unlike Godfrey – that failure to adhere to a starting point is not an error in principle per se (e.g. Parranto – para. 72), yet it establishes a new starting point for wholesale fentanyl trafficking of 9 years. Although the sentencing judge made reference to a sentencing range, it was simply found to be inappropriate, and Ontario caselaw (paras. 69-70) – surprisingly – was used to decide a 9-year starting point was appropriate. It would be great if someone planned on appealing either case [Feb. 2022 edit: the Supreme Court decided these cases in 2021, a divided court upholding the sentence and finding that starting points are entirely not binding on a sentencing judge – as long as the sentence is not “demonstrably unfit”, while refusing to abolish them – R. v. Parranto, 2021 SCC 46]. Likely, the ratio (or outcome) of Parranto [at the Court of Appeal] is – in retrospect, at least – inconsistent with Supreme Court jurisprudence in Friesen and probably Lacasse – assuming of course, that the original sentence was not “demonstrably unfit” which is a rather difficult determination for anyone to make – yet its decision to set a starting point – likely is not wrong – yet [Feb. 2022 edit: unfortunately, nothing has changed substantively, other than ABCA less likely to overturn sentences only because of lack of compliance with a starting point. I’m quoted by CBC here]. The Supreme Court will need to decide the issue squarely once and for all, now that it’s again decided that failure to follow a starting point/range is not an error in principle per se; per Berger J.’s criticisms of the starting point approach in R v Legerton, 2015 ABCA 79 (CanLII – see e.g. para. 12) and Gashikanyi, etc. and the interveners’ concerns in Friesen: are starting points problematic in that they arguably fetter the discretion of a sentencing judge far more than a sentencing range does? Also – practically speaking, when is a judge’s sentencing decision simply inappropriate and requiring appellate intervention? This is also a crucial problem, and one that is clearly evident in the Hajar/Gashikanyi debate (I’ll get into Gashikanyi shortly: note there is much discussion there, as in Arcand and Hajar about “re-consideration cases” and significant debate about what constitutes an appropriate sentence). I suspect that that debate will rage on – although it will now be termed more in the language of “demonstrably unfit” rather than “error in principle”. Bottom line when it comes to sentencing: people (and courts) want numbers. Lacasse is helpful in providing guidance about what constitutes “demonstrably unfit” in terms of a sentence – I’ll get to this below. I’m hopeful that Lacasse will be sufficient, but I suspect that more clarity will be required – particularly when courts (and individual judges) disagree over the appropriate quantum of sentence for a particular type of offence or offenders.
Lacasse clarifies “demonstrable unfitness”, including:
52 It is possible for a sentence to be demonstrably unfit even if the judge has made no error in imposing it. As Laskin J.A. mentioned, writing for the Ontario Court of Appeal, the courts have used a variety of expressions to describe a sentence that is “demonstrably unfit”: “clearly unreasonable”, “clearly or manifestly excessive”, “clearly excessive or inadequate”, or representing a “substantial and marked departure” (R. v. Rezaie (1996), 1996 CanLII 1241 (ON CA), 31 O.R. (3d) 713 (C.A.), at p. 720).All these expressions reflect the very high threshold that applies to appellate courts when determining whether they should intervene after reviewing the fitness of a sentence.
53 This inquiry must be focused on the fundamental principle of proportionality stated in s. 718.1 of the Criminal Code, which provides that a sentence must be “proportionate to the gravity of the offence and the degree of responsibility of the offender”. A sentence will therefore be demonstrably unfit if it constitutes an unreasonable departure from this principle. Proportionality is determined both on an individual basis, that is, in relation to the accused him or herself and to the offence committed by the accused, and by comparison with sentences imposed for similar offences committed in similar circumstances. Individualization and parity of sentences must be reconciled for a sentence to be proportionate: s. 718.2(a) and (b) of the Criminal Code.
54 The determination of whether a sentence is fit also requires that the sentencing objectives set out in s. 718 of the Criminal Code and the other sentencing principles set out in s. 718.2 be taken into account. Once again, however, it is up to the trial judge to properly weigh these various principles and objectives, whose relative importance will necessarily vary with the nature of the crime and the circumstances in which it was committed. The principle of parity of sentences, on which the Court of Appeal relied, is secondary to the fundamental principle of proportionality. This Court explained this as follows in M. (C.A.):
It has been repeatedly stressed that there is no such thing as a uniform sentence for a particular crime. . . . Sentencing is an inherently individualized process, and the search for a single appropriate sentence for a similar offender and a similar crime will frequently be a fruitless exercise of academic abstraction. [para. 92]
See Berger J.’s prescient comments – criticizing Arcand, and writing for the majority – in R. v. Lee, 2012 ABCA 17 (CanLII) (emphasis in original):
53 The danger is that a judgment arising from a particular “doctrinal disposition” may be perceived as a response to the agenda of those who would, by way of example, “get tough on crime”. Such a pronouncement purporting to vindicate certain promises of the law but excluding others, may be seen as an echo of political or public sentiment. It is an inescapable truth that to call upon the judiciary to adopt the selective approach of the majority of an appellate sentencing panel, failing which “Parliament will act”, erodes the independence of the Court and threatens public confidence in the administration of justice.
54 In remarks to the Council of the Canadian Bar Association (the “CBA”) at the Canadian Legal Conference in Halifax on August 13, 2011, Chief Justice McLachlin spoke about public confidence in the judiciary. She referenced the response of the CBA to a speech given by a Minister of the Crown suggesting that some judges were insufficiently solicitous to government policy. She thanked the CBA for reminding the Minister in question “of the importance of public confidence in an independent and impartial judiciary that bases its decisions on the law and not on government policy.” That admonition, it seems to me, also applies to judges who may run the risk of being perceived as overly-solicitous to government policy.
Berger J. in R. v. Gashikanyi, 2017 ABCA 194 (CanLII) notes in paras. 16-36 an excellent critique of starting points. I’ll quote extensively, emphasis in original:
19 It is in this latter regard that the majority reasons in Hajar fail to provide meaningful guidance to sentencing judges in Alberta. The abandonment of sentencing ranges and substitution by this Court of starting points without first determining the “acceptable range of sentence” before fixing the starting point, operates in practical terms as a constraint on the discretion afforded to sentencing judges to impose individualized sentences. After all, sentencing ranges are, as the Supreme Court of Canada has made clear, “historical portraits” for the use of sentencing judges whose discretion should not be interfered with absent demonstrable unfitness:
“Sentencing ranges are nothing more than summaries of the minimum and maximum sentences imposed in the past, which serve in any given case as guides for the application of all the relevant principles and objectives. However, they should not be considered “averages”, let alone straitjackets, but should instead be seen as historical portraits for the use of sentencing judges, who must still exercise their discretion in each case.”
20 The majority in Hajar not only failed to conduct a detailed, comprehensive review of the minimum and maximum sentences for sexual interference imposed in the past by both trial and appellate judges, but also summarily resiled from the established sentencing range with the admonition that “…the Reasons [of the sentencing judge] overlook the fact that sentencing precedents from this Court are of limited precedential value.” (at para. 150)
21 Both the majority in Hajar and the Crown in the case at bar relied upon the majority decision in Arcand which, it will be remembered, suggested that starting point sentencing is “effectively neutered” or rendered “meaningless in practice” unless elevated to a rule of law (at paras. 99 and 110 of the Arcand majority judgment).
 The failure of the majority opinion in Arcand to recognize that starting points are guidelines and are not to be elevated to the status of legal imperatives is, with respect, a serious error, as is its suggestion that deviation from a starting point constitutes reversible error. The result is that the majority in Arcand relegates to insignificance the rich trove of trial and appellate pronouncements that deviate from starting points but inform the range of sentence that may properly be imposed for various crimes. (emphasis added)
23 In the result, as did the majority in Arcand, the collection of trial and appellate pronouncements that deviate from the starting point was again relegated to insignificance by the majority in Hajar.
24 The Hajar majority reasons, while citing Lacasse for other purposes, seems not to have appreciated Lacasse’s focus on the increased force of appellate deference and Lacasse’s reminder that reliance on the proposition “that an appellate court is to defer to a sentencing decision absent an error in principle, failure to consider a relevant factor, [or] an overemphasis of the appropriate factors” needed to be clarified and qualified (Lacasse at para. 47) With that in mind, Justice Wagner reminded both trial and appellate courts that departure from sentencing ranges or starting points are not only permitted but to be expected. He cautioned against “trivialization” of the term “error in principle”, emphasizing that an appellate court may not intervene simply because it would have weighed the relevant factors differently. (Lacasse at para. 49)
25 The majority reasons in Hajar also run afoul of Lacasse by failing to appreciate that the choice of a sentencing range [or a starting point] or of a category within a range falls within the trial judge’s discretion and cannot in itself constitute a reversible error. (Lacasse at para. 51)
26 The majority judgment in Lacasse took pains to emphasize (and the majority in Hajar failed to recognize) that the principle of parity of sentences is secondary to the principle of proportionality and must not be given priority over the principle of deference to the trial judge’s exercise of discretion. The majority failed to appreciate Lacasse’s lesson that “individualization and parity of sentences must be reconciled for a sentence to be proportionate” (Lacasse at para. 53) and that individualization takes priority over parity.
27 The foregoing clarifications and qualifications seem to have eluded the Hajar majority. That is not surprising given its view that the role of an appellate court is to stamp out “caprice” and to “act as a buffer against outliers of sentencing outcomes.” (Hajar at paras. 47, 50)
35 Notwithstanding this Court’s disclaimers to the contrary, sentencing judges in Alberta know that departure from a starting point facilitates appellate intervention in this jurisdiction and constrains their discretion. Simply put, they know it is much easier to upset a sentence of, say, 15 months’ imprisonment when the starting point is 3 years than it is when the range is 12 months to 3 years. The departure from the low end of the range may be seen on appellate review as acceptable while the departure from the starting point is more readily adjudged to be reversible error.
At para. 36 of Gashikanyi, Berger J. discusses the practical immunity to appeal of Hajar, Arcand and Rahime (I’ll add the more recent Godfrey decision to the mix); emphasis added at para. 36:
I would add only that in a number of leading judgments of this Court including Hajar, Arcand and R. v. Rahime, 2001 ABCA 203 (all five person panel appeals), the ultimate disposition of the Court was viewed favourably by both Crown and defence – the Crown, because the Court endorsed the legal argument advanced by the Crown in fixing the starting point for future cases; the defence, because in each case, although the sentence below was adjudged inadequate, the accused was not re-incarcerated. Neither side would have any interest whatsoever in appealing the matter to the Supreme Court of Canada. In each case the Crown had its precedent and the accused had his freedom; in each case the Court’s judgment was thereby rendered immune from further review by the Supreme Court.
Note the ABCA’s majority approach in important, 5-panel cases like Hajar and Arcand, the makeup of which, as Berger J. observes at paras. 68-75 of Gashikanyi, suggests something other than a random assignment of justices. I note that all 3 of the cases Berger J. cites (Hajar, Arcand, and Rahime) all include a 5-person panel for the ABCA, all include Fraser CJ, and none of them include Berger J. Further, both Hajar and Arcand include Fraser CJ and Watson J. – both of whom are necessary for the 3-person majority opinions (not that the “dissenters” were particularly opposed in Hajar). Berger J. is quite likely making a rather obvious point – although it’s one that’s denied perhaps too vociferously by commentators and other ABCA jurists in Gashikanyi. Clearly, it’s not random assignment or expertise preventing Berger J. from sitting on – say – Arcand. Whether or not that’s appropriate or ethical is perhaps “above my pay grade” – but let’s not pretend that his point isn’t obvious. Note also Watson J.’s presence in both R v Godfrey, 2018 ABCA 369 (CanLII) and R v Sumner, 2019 ABCA 399 (CanLII) – recent ABCA decisions I have criticized.
These points makes the comments of Fraser CJ in Arcand especially ironic (para. 8, emphasis added):
The third truth. Judges are not the only ones who know truths one and two, and thus judge shopping is alive and well in Canada – and fighting hard to stay that way. All lead inescapably to the fourth truth. Without reasonable uniformity of approach to sentencing amongst trial and appellate judges in Canada, many of the sentencing objectives and principles prescribed in the Code are not attainable. This makes the search for just sanctions at best a lottery, and at worst a myth. Pretending otherwise obscures the need for Canadian courts to do what Parliament has asked: minimize unjustified disparity in sentencing while maintaining flexibility.