Friday, October 18, 2013


Of peripheral interest to criminal lawyers are a couple of recent decisions of the United Kingdom Supreme Court.

Prisoners and voting rights

Chester, R (on the application of) v Secretary of State for Justice [2013] UKSC 63 (16 October 2013) illustrates how an issue that should be resolved in favour of the applicant may not require a remedy. Previous decisions [18] of the ECtHR, including an appeal from the United Kingdom, had held that denying prisoners the right to vote is a breach of the Convention. The UK legislature is looking at this [19], and the Supreme Court therefore did not see that a declaration of incompatibility was necessary on the appeals in this case.

The relationship between national courts and the Strasbourg court was considered [27], and the Supreme Court rejected the respondent's submission that the difference here was over "some fundamental substantive or procedural aspect of our law" sufficient to justify departure from Strasbourg jurisprudence.

Indeed, the issue of removing prisoners' voting rights was [35] not "fundamental to a stable democracy and legal system such as the United Kingdom enjoys."

Ah yes, thank you Joni Mitchell ...
"Don't it always seem to go
That you don't know what you've got
Till it's gone
They paved paradise
And put up a parking lot"

That is, is the moral value of having the right to vote equivalent to the value of not having the right to vote?

It is difficult to stir up much public discussion about whether prisoners should be deprived of the right to vote. In New Zealand the Electoral Act was recently amended to further strengthen existing restrictions on prisoners' voting rights, so that now anyone detained in prison pursuant to a sentence of imprisonment imposed after 16 December 2010 does not have the right to vote. For an outline of the reasons this might not be appropriate, see the Report of the Law and Order Committee on the Electoral (Disqualification of Convicted Prisoners) Amendment Bill, particularly the New Zealand Labour Party minority view.

Update: The High Court, in Taylor v Attorney-General [2015] NZHC 1706 (24 July 2015) has declared that:

Section 80(1)(d) of the Electoral Act 1993 (as amended by the Electoral (Disqualification of Sentenced Prisoners) Amendment Act 2010) is inconsistent with the right to vote affirmed and guaranteed in s 12(a) of the New Zealand Bill of Rights Act 1990, and cannot be justified under s 5 of that Act.”

Further update: On 9 November 2018 the Supreme Court confirmed the existence of the High Court's power to make a declaration of inconsistency with the New Zealand Bill of Rights Act 1990, and has dismissed the Attorney-General's argument to the contrary: Attorney-General v Taylor [2018] NZSC 104.

Another further update: On 25 February 2020 the government introduced a Bill to give prisoners serving less than three years' imprisonment the right to vote.

Oral hearings

Sometimes judicial decisions may be made "on the papers" filed by the parties, without the need for an oral hearing of argument. In Osborn v The Parole Board [2013] UKSC 61 (9 October 2013) the Supreme Court considered when an oral hearing would be required by common law procedural fairness.

The Court's press summary sets out the essential points, and they are also summarised in the judgment at [2].

See also R v Parole Board, ex parte Smith and West [2005] UKHL 1 (27 January 2005), discussed here on 31 January 2005 and see para [14] of Osborne for administrative developments, R (on application of Hammond) v Secretary of State for the Home Department [2005] UKHL 69 (1 December 2005), discussed here on 5 December 2005, and Ebanks v R (Cayman Islands) [2006] UKPC 16 (27 March 2006), discussed here on 28 March 2006.

Significantly, the tribunal must "guard against any temptation to refuse oral hearings as a means of saving time, trouble and expense" [2(viii)].

Appellate judges who have the power to deal on the papers with applications for leave to appeal will, no doubt, not need to be reminded of these considerations.

Thursday, October 10, 2013

Unnecessarily attacking the fundamentals

Even the most robust of the fundamentals of the criminal law can be modified by statute. When that happens, the fundamental should retain its strength in all cases to which the statutory modification does not apply.

But sometimes a statute is not explicit on whether it modifies a fundamental of the criminal law, while its purpose seems to require such a modification. If a court accepts that this sort of statute does indeed modify a fundamental of the criminal law, there is a danger that it will support its conclusion by pointing to weaknesses in the fundamental. Those weaknesses may later be used in support of interpretations of other statutes to override the now weakened fundamental.

To bring these considerations into focus, consider Lee v New South Wales Crime Commission [2013] HCA 39 (9 October 2013). The relevant fundamental of the criminal law was the principle that the prosecution must discharge the onus of proof and cannot compel the defendant to give evidence to help discharge that onus: [176] per Kiefel J dissenting.

Lee concerns the civil procedure, under the Criminal Assets Recovery Act 1990 (NSW) (the "CAR Act") of compulsory examination of a person to establish whether assets were probably obtained through serious crime. The examinee was also subject to criminal proceedings, and this gave rise to the issue whether the examination should be delayed until the trial had been concluded, so as not to give the prosecution an unfair advantage.

This sort of issue has arisen before: X7 v Australian Crime Commission [2013] HCA 29 (26 June 2013), discussed here on 27 June 2013. In that case the conclusion reached by the majority (Hayne, Bell and Kiefel JJ) was the opposite of that reached by the majority in Lee (French CJ, Crennan, Gageler and Keane JJ). The new players are Gageler and Keane JJ, who delivered a joint judgment in Lee).

A focus on the judgment of Gageler and Keane JJ should therefore reveal the points that carried the day in Lee. The appellants' argument, as refocused in oral submissions, is summarised at [304]-[305]. The inherent prejudice in allowing an examination while criminal charges are pending is, according to this analysis [305]:

" ... the answers given and documents produced by the person in the examination would inevitably constrain the instructions on which the legal representatives of the person could act in the criminal proceedings: the legal representatives would be ethically bound not to lead evidence or cross-examine or make submissions to suggest a version of the facts which contradicted that given by their client on oath in the examination."

The assumption here is that the legal advisers know what the examinee had said during the examination. Normally, a client does not give instructions on oath, and a client's prior statements are not given on oath. A difficulty would only arise if answers given on oath at examination became admissible against the examinee as defendant at trial. So the constraint on instructions referred to at [305] can be avoided if answers at examination are not disclosed to the legal representative in the criminal proceedings, and if those answers are not admissible to rebut the defendant's defence at trial.

However instead of adopting this sort of harm-containment approach, Gageler and Keane JJ took a swipe at the principle of construction (that the legislature does not intend to alter the law beyond the immediate scope and object of a statute [308]) that fundamental rights are not altered by a statute unless that is expressly done. They adopted Gleeson CJ's view that in modern times the strength of that principle will vary with context [312]. But it is not necessary to qualify the strength of this principle of construction, to make the point that the clear intention of a statute may be to alter fundamental rights.

In addition to weakening the principle of construction, Gageler and Keane JJ weakened the fundamental principle of the criminal law that a defendant cannot be compelled by process of law to admit the offence, by saying it is "not monolithic: it is neither singular nor immutable" [318]. Some statutory inroads on the right to silence do not mean that the right is weakened where it does apply. It was unnecessary for the judges to suggest this weakening.

The more conventional part of this judgment addresses directly the interpretation of the CAR Act [326]-[335] and reaches the, no doubt sound, conclusion that [335]:

"The power conferred by s 31D(1)(a) does not authorise the making or implementation of an examination order where to do so would give rise to a real risk of interference with the administration of justice including by interfering with the right of the person to be examined (or any other person) to a fair trial. For reasons already given, however, the making of such an order does not give rise to a real risk of interference with the administration of justice by reason only that the subject-matter of the examination will overlap with the subject-matter of criminal proceedings that have commenced but that have not been completed."

And importantly, as to the way the discretion to order an examination should be exercised, [337]:

"The reasons for judgment of the Court of Appeal do not suggest that the CAR Act indicates a legislative intention that the Supreme Court should allow any proceedings under that Act to proceed if the circumstances of the case, other than the mere pendency of criminal proceedings against the examinee, were such as to reveal a real, as opposed to a speculative or theoretical, risk that the administration of justice would be adversely affected. The exigencies of criminal proceedings might well afford a ground for a refusal to make an order under s 31D(1)(a). For example, the timing of an application may be such as to prejudice the fair trial of a criminal charge because of the likely disruption of the preparation for, or conduct of, a trial which is imminent. As Beazley JA specifically noted [606], that possibility was not raised before the Court of Appeal as a consideration having a claim upon the discretion in the circumstances of this case. Had it been raised, it would obviously be a consideration which might properly be taken into account in exercising the discretion."

This conclusion could have been reached without suggesting that the rule of construction or the right to silence are in any general sense weakened these days. Lee required merely a conventional exercise in statutory interpretation.

A similar position exists under comparable New Zealand legislation: Criminal Proceeds (Recovery) Act 2009, s 107. In each case a careful analysis has to be made of the matters which the Commissioner wishes to examine the defendant about, and the likely impact of answering those matters on the subsequent criminal trial: Commissioner of Police v Wei [2012] NZCA 279 at [40].

Tuesday, October 08, 2013

Fresh evidence

Fresh evidence was a central topic in two decisions delivered within hours on opposite sides of the planet yesterday.

In refusing leave to appeal against sentence, the New Zealand Supreme Court addressed a submission that the applicant's cooperation with the authorities was a new fact relevant to penalty: Bland v R [2013] NZSC 93 (7 October 2013). The fact could have been used in submissions to the Court of Appeal, but was not, but after that appeal Mr Bland did give assistance to the authorities. He then asked the Court of Appeal to recall its dismissal of his sentence appeal so that this new fact could be considered. The Court of Appeal dismissed that application.

The Supreme Court said [6(a)]: "This Court does not usually entertain criminal appeals on grounds that, although available, were not raised before the Court of Appeal", citing Mankelow v R [2007] NZSC 57 at [2].

These Supreme Court decisions are dismissals of applications for leave to appeal and are not required to be given in detail: "The reasons may be stated briefly, and may be stated in general terms only": s 16(2) Supreme Court Act 2003.

The risk with these brief reasons is that they may be cited as if they were precedents, as with Mankelow, so that matters of policy are wrongly treated as legal rules. And legal rules may be stated inaccurately, in the effort to be brief, and this creates a risk of misapplication of the law in subsequent cases. This is illustrated in Bland at [6(b)], where the Court diminished the value of Mr Bland's assistance to the authorities, on the basis that it was "self-serving rather than motivated by a genuine desire to cooperate".

Motive for the giving of assistance has never been relevant. Motives for mitigating actions seldom are: for example, credit for early guilty pleas is given without asking whether they were motivated by remorse, and remorse has its own status as an independent mitigating factor: s 9(2)(f) of the Sentencing Act 2002. In R v Stark [2006] NZCA 190 the Court said at [10]:

" ... It is necessary to weigh the assistance given with the type and seriousness of the offending, the sentence that otherwise would be appropriate, the nature and value of the assistance, the situations in which it is given and the consequences. In the end the sentence must be that which is appropriate in light of all of the circumstances including any assistance to the authorities."

There is no mention of the motivation for the assistance in those considerations. Assistance has its own weight, and remorse may be an additional circumstance but its absence should not disentitle an offender from advancing assistance as a mitigating factor.

Still, the Court's real reason for refusing leave seems to be that Mr Bland's assistance to the authorities was of no particular consequence in the prosecution of co-offenders [6(c)]. Any appropriate sentence reduction would have been minimal and the sentence that was imposed was not outside the range of sentences available to the judge. There was not, therefore, a "substantial miscarriage of justice", which is one of the ways a case may come within the qualifying requirement that it should be in "the interests of justice" for the appeal to be heard: Supreme Court Act, s 13(2)(b). The phrase "substantial miscarriage of justice" is not defined.

Several hours later, in London, the Privy Council in Lundy v The Queen (New Zealand) [2013] UKPC 28 (7 October 2013) ordered a retrial because new evidence suggested that the appellant's convictions were unsafe [151], [164].

It is inappropriate for me to say much about this case, pending a retrial. However of legal interest is the Board's clarification of the requirements for the allowing of appeals against convictions on the basis of fresh evidence. There are three tests, to be applied in sequence [120]:

"The Board considers that the proper basis on which admission of fresh evidence should be decided is by the application of a sequential series of tests. If the evidence is not credible, it should not be admitted. If it is credible, the question then arises whether it is fresh in the sense that it is evidence which could not have been obtained for the trial with reasonable diligence. If the evidence is both credible and fresh, it should generally be admitted unless the court is satisfied at that stage that, if admitted, it would have no effect on the safety of the conviction. If the evidence is credible but not fresh, the court should assess its strength and its potential impact on the safety of the conviction. If it considers that there is a risk of a miscarriage of justice if the evidence is excluded, it should be admitted, notwithstanding that the evidence is not fresh."

Here, "a risk of a miscarriage of justice" refers to the safety of the conviction.

Lundy was an appeal under what we can now call the old law. Now, the criteria for allowing an appeal against conviction are enacted in the Criminal Procedure Act 2011 ("CPA"), s 232. They do not specifically refer to fresh evidence. Cases of fresh evidence must therefore come within the phrase "miscarriage of justice", which is defined in subsection (4) as:

"... miscarriage of justice means any error, irregularity, or occurrence in or in relation to or affecting the trial that—

(a) has created a real risk that the outcome of the trial was affected; or

(b) has resulted in an unfair trial or a trial that was a nullity."

The obtaining of new evidence can hardly be called an error or irregularity, but it may be "an occurrence ... affecting the trial". It might, also at a pinch, be put within the meaning of "an unfair trial" in the sense that the trial that happened involved assessments of the weight of items of evidence that have since been shown to have been inappropriate. The trial, although it seemed to have been fair when it was conducted, is now shown to have been unfair.

These ways of including cases of fresh evidence within the grounds for allowing appeals against conviction are a bit forced, and one might say that it was a legislative oversight to omit specific reference to fresh evidence in s 232. Certainly, the CPA does permit an appellate court to hear evidence (ss 334 and 335), and the Criminal Procedure Rules 2012 make provision for adducing fresh evidence: r 8.8, in terms which make it clear that miscarriage of justice encompasses fresh evidence.

The Board did not say that the trial had been unfair. It was not considering s 232, and Lundy should not be applied as if it were authoritative on the meaning of that section, but it would not be inconsistent to say that the better interpretation is that fresh evidence is an occurrence in relation to the trial.

But there is good reason to conclude that a substantively fair trial is one where the evidence is weighed properly, and that retrospective appreciation of unfairness in the light of fresh evidence is grounds under s 232 for allowing an appeal against conviction.

There are two more interesting aspects of Lundy that I can mention.

First, the Privy Council seized jurisdiction – without creating a precedent – just in the interests of being sensible [11]. Strictly, it is a breach of the rule of law for a court to do something and say it is not creating a precedent: compare John Gardner, Law as a Leap of Faith (OUP, Oxford, 2012) at 210 (reviewed by me here on 6 July 2013).

Secondly, the Board ordered a retrial rather than remit the case to the Court of Appeal. There is nothing particularly unusual in doing that, but the difficulties that have been experienced when appeal judges attempt to reach verdicts perhaps suggest that the Board considers that the approach recently favoured – under what is now the old law – by the New Zealand Supreme Court in Matenga v R [2009] NZSC 18, discussed here on 9 July 2009, is inappropriate. It is likely that under the new law (s 232) – see the digression by me on 19 August 2013 – the Matenga approach will not apply. Unfortunately however, the Board endorsed what it took to be the Matenga approach: [143]-[151].

On this topic the Board appears to be glossing over difficulties. How does an appeal court decide whether a conviction is safe? Does it (1) decide for itself whether there is a reasonable doubt about the appellant's guilt? Does it (2) decide what a jury would have concluded had the error at trial not occurred? Does it (3) do a bit of each – deciding for itself until it gets stuck, in which event asks what a jury would have done? These have all been tried at various times. Delivering the judgment of the Board, Lord Kerr endorsed [146] his own judgment, dissenting on the facts, in Taylor v R [2013] UKPC 8 (discussed here on 19 March 2013) taking the third (called the Pendleton) approach, but which was put by the majority in Taylor [20] as one of asking whether the jury might reasonably have come to a different conclusion as to whether the appellant was guilty; this, confusingly, looks like the second approach.

My own view is that this is rather silly. The focus in these sorts of appeals should always have been on whether the error at trial could have significantly affected the weight given to contested evidence on an issue central to the logic of the prosecution case. Under the new law, s 232(4)(a) is consistent with this, using the phrase "a real risk that the outcome of the trial was affected". Unfortunately, it is arguable that this phrase is also consistent with the other approaches, and it would be sad – so sad – if Lundy were taken to be an endorsement of the survival of Matenga in the new statutory environment.

Thursday, October 03, 2013

Reasonable grounds to suspect

Now that we have search on "reasonable grounds to suspect" the commission of an offence, a lesser threshold than the reasonable belief that had previously been required – and that had been replaced by stealth – it is relevant to consider the definition of reasonable suspicion developed and applied by the Supreme Court of Canada: R v Chehil, 2013 SCC 49 (27 September 2013) and R v MacKenzie, 2013 SCC 50 (27 September 2013).

There is no substitute for reading the judgments, but the essentials, set out in Chehil, are:
  • The central question is: Is the totality of the circumstances, including the specific characteristics of the suspect, the contextual factors, and the offence suspected, sufficient to reach the threshold of reasonable suspicion? [39]
  • "Suspicion" is an expectation that the targeted individual is possibly engaged in some criminal activity. A "reasonable" suspicion means something more than a mere suspicion and something less than a belief based upon reasonable and probable grounds. [26, citing Binnie J in R v Kang-Brown, a case discussed here on 28 April 2008]
  • Reasonable suspicion is a lower standard than reasonable belief, as it engages the reasonable possibility, rather than probability, of crime. As a result, when applying the reasonable suspicion standard, reviewing judges must be cautious not to conflate it with the more demanding reasonable and probable grounds standard. [28]
  • The reasonable suspicion standard addresses the possibility of uncovering criminality, and not a probability of doing so. [32]
  • The constellation of facts must be based in the evidence, tied to the individual, and capable of supporting a logical inference of criminal behaviour. If the link between the constellation and criminality cannot be established by way of a logical inference, the Crown must lead evidence to connect the circumstances to criminality. This evidence may be empirical or statistical, or it may be based upon the investigating officer's training and experience. [46]
  • An officer's training and experience may provide an objective experiential, as opposed to empirical, basis for grounding reasonable suspicion. However, this is not to say that hunches or intuition grounded in an officer's experience will suffice, or that deference is owed to a police officer's view of the circumstances based on his training or experience in the field. A police officer's educated guess must not supplant the rigorous and independent scrutiny demanded by the reasonable suspicion standard.[47]
  • While a trial judge is owed deference in relation to his factual findings, whether those factual findings support reasonable suspicion is a question of law, and as such is reviewable on the correctness standard. [60]
And, applying these principles in MacKenzie, the Court emphasised:
  • In assessing whether a case for reasonable suspicion has been made out, the analysis of objective reasonableness should be conducted through the lens of a reasonable person standing in the shoes of the police officer. [63]
  • The hallmark of reasonable suspicion, as distinguished from mere suspicion, is that "a sincerely held subjective belief is insufficient" to support the former (Kang-Brown, at para. 75, per Binnie J., citing P. Sankoff and S. Perrault, "Suspicious Searches: What's so Reasonable About Them?" (1999), 24 C.R. (5th) 123, at p. 125). Rather, as Karakatsanis J. observes in Chehil [26], reasonable suspicion must be grounded in "objectively discernible facts, which can then be subjected to independent judicial scrutiny". [41]
  • Exculpatory, common, neutral, or equivocal information should not be discarded when assessing a constellation of factors. However, the test for reasonable suspicion will not be stymied when the factors which give rise to it are supportive of an innocent explanation. We are looking here at possibilities, not probabilities. Are the facts objectively indicative of the possibility of criminal behaviour in light of the totality of the circumstances? If so, the objective component of the test will have been met. If not, the inquiry is at an end. [72]
Giving "reasonable suspicion" as precise a definition as is possible runs counter to a suggestion that the statutory scheme requiring reasonable grounds to suspect the commission of a qualifying offence and reasonable grounds to believe that evidence will be found in the search creates "a regime of relativity: the differentiation simply means that above a minimum floor, more is required in relation to the location of evidence at the target place than in relation to the commission of a crime … [with] a moderated and relativistic interpretation of the thresholds."

Chehil also discusses the use of profiles [39], rejecting it as a basis for suspicion. Here the judicial reasoning is perhaps politically correct, if too subtle, if profiles are based on experience.

MacKenzie mentions facts such as nervousness, red eyes, and erratic driving. The Court split 5-4 on the facts in this case, perhaps illustrating how, notwithstanding the utmost care in being clear about the criterion of reasonable suspicion, its application to particular facts can be controversial.

The difficulties are apparent from this extract from my Misuse of Drugs text, para 1406(a) (citing cases not available online):

For an illustration of analysis of testimony asserting indicia of drug use, see R v Herlund 28/5/08, Duffy J, HC Auckland CRI-2006-004-21413. Here, a suspect's nervousness in police presence lacked significance when it was explained by his being on active charges concerning drug dealing (at [81]), and her Honour emphasised that it is necessary to tie grounds for search to the present occasion, so as to avoid subjecting suspicious looking people, or those known to have drug histories, to a lower standard for search (at [83], applying R v Anderson [2005] 21 CRNZ 393 at [33]). This is not to say that police knowledge that a suspect has active drugs related charges is to be ignored, for in the same case Duffy J held that such knowledge, coupled with the suspect's withdrawal of his consent to a search — upon the discovery of a P pipe in his pocket — did provide proper grounds for a further search: at [58]. This withdrawal of consent was likened to Mr Carroll's [R v Carroll 21/5/04, Rodney Hansen J, HC Auckland CRI–2003-004-41192] effort to conceal the cigarette packet: at [61].

But assertion of rights is not properly a basis for reasonable suspicion, as was stated in Chehil at [44]:

"Nor should the exercise of Charter rights, such as the right to remain silent or to walk away from questioning made outside the context of a detention, provide grounds for reasonable suspicion. These rights become meaningless to the extent that they are capable of forming the basis of reasonable suspicion. Individuals should not have to sacrifice privacy to exercise Charter rights."

[Update:] As this is an oft-visited posting, I should add that the Canadian definition of belief is not universally accepted. Instead of being a high level of confidence, it can be defined as thinking something is the case. If you are thinking of what belief means, without looking for a distinction from suspicion, then indeed you might think that a belief is a high level of confidence in something. But once you have to distinguish between belief and suspicion it usually makes more sense to apply the "suspect ... may, believe ... is" construction. That will be so unless an enactment itself has a "believe ... may" usage, in which case believe must mean thinking something is highly likely. The terms must always be construed in their context, subject to enacted definitions, if there are any. The "suspect ... may, believe ... is" interpretation makes for law that is more readily predictable in its application, than saying belief is a high level of likelihood, and it works against an arbitrary application of the law.

Ethnicity, deprivation, and manifest inadequacy of sentence

In the absence of legislation requiring special consideration at sentencing for the ethnicity of the offender, the relevance of social deprivation is the same for all offenders: Bugmy v The Queen [2013] HCA 37 (2 October 2013) at [37].

Where an offender's abuse of alcohol is relevant to the commission of the offence and is a reflection of the environment in which the offender was raised, it may be taken into account as a mitigating factor [38], as also it may be if the offender's background would make imprisonment particularly burdensome [39]. This is because all material facts must be taken into account in all sentencing decisions.

However the weight to be given to circumstances arising from deprivation may vary according to the purpose of punishment that is being considered at each stage of the exercise of determining the appropriate sentence [44]. Sentence reduction arising from circumstances of deprivation, like that arising from mental illness, is not inevitable [47].

A court of final appeal is not a sentencing court, and if it identifies an error in principle it is likely to remit the case for reconsideration by the appellate court that is responsible for oversight of sentencing decisions, cf [49] and see [60] of Munda, below.

Munda v Western Australia [2013] HCA 38 (2 October 2013) addresses the relevance of previous sentencing decisions in determining whether a present sentence is manifestly inadequate, referring at [39] to Hili v The Queen [2010] HCA 45 (discussed here on 10 December 2010), and holding that reference to comparable cases may be an indication of inadequacy but, because the demonstration of an existing range of sentences does not establish that that range is correct, it is not determinative. Also relevant are the maximum penalty and the seriousness of the particular offence. (Bell J dissented on whether the sentence here had been manifestly inadequate.)

Also in Munda the Court approved [43] the principle that "[it is wrong] to reduce the weight to be given to general deterrence in circumstances where alcohol-fuelled violence is endemic in the community generally, even if not sufficiently deterred in fact by the prospect of imprisonment" (quoting McLure P in the Court of Appeal in Munda).

The prospect of retribution being exacted in the community was raised and although the point didn't have to be decided in Munda the Court strongly hinted that this should be irrelevant [61]-[63], because vendettas must be discouraged, punishment is meted out by the state, and offenders should not be given a choice as to the mode of their punishment.

There is also some discussion in Mundy of the residual discretion of an appellate court to decline to increase a sentence that is manifestly inadequate, particularly to avoid double punishment or interference with rehabilitation, but subject to the need to avoid the manifest injustice of upholding an inadequate sentence [64]-[78].