Home | Databases | WorldLII | Search | Feedback Court of Appeal of New Zealand

Mitchell v Attorney-General [2023] NZCA 605 (29 November 2023)

Last Updated: 4 December 2023

JUDGMENT OF THE COURT

1. Leave to adduce further evidence is granted.
2. The appeal is dismissed.

____________________________________________________________________

REASONS OF THE COURT

(Given by Katz J)

Introduction

• [1] Ms Mitchell was sentenced to imprisonment for a term of two years and three months. The Corrections Act 2004 (the Act) requires that all prisoners serving a term of more than three months’ imprisonment be assigned a security classification by the Chief Executive of the Department of Corrections that reflects the level of risk posed by that prisoner while inside or outside prison.^[1] Corrections assigned Ms Mitchell a security classification of “low”. Ms Mitchell applied to the High Court for judicial review of that decision on the basis that Corrections had made a number of errors when assigning her security classification. She submitted that she should have been assigned a security classification of “minimum” rather than “low”. Gendall J dismissed the judicial review application.^[2] Ms Mitchell now appeals.
• [2] Following the High Court decision, Ms Mitchell’s security classification was reviewed twice by Corrections, at six-monthly intervals, as required by the Act.^[3] Both reviews resulted in Ms Mitchell’s “low” security classification being maintained. Ms Mitchell did not challenge those decisions. She was subsequently released into the community on parole, although she is presumably no longer subject to any standard or special release conditions, given the lapse of time.
• [3] Corrections’ view is that the issues raised by this appeal are now moot, and it would therefore be appropriate for the Court to exercise its discretion to decline to determine this appeal. Corrections further submitted that the appeal is unmeritorious in any event, and the Judge’s decision to dismiss Ms Mitchell’s application for judicial review was correct.

Application for leave to adduce further evidence

• [4] Corrections applied for leave to adduce updating evidence on appeal.^[4] The evidence sought to be adduced concerned the two reviews of Ms Mitchell’s security classification which occurred after the High Court decision. The evidence consists of affidavits from Ms Trask and Mr Murphy, who were the approving officers for the reviews of Ms Mitchell’s security classification. As far as we are aware, Ms Mitchell does not oppose the application. In any event, we are satisfied that the evidence is fresh, credible and cogent,^[5] and grant leave to adduce it accordingly.

The process for assigning security classifications to prisoners

• [5] The statutory and administrative scheme for assigning security classifications to prisoners has been considered in a number of previous cases.^[6] As the Judge explained in the judgment under appeal:^[7]

[7] ... The classification is intended to reflect the level of risk posed by the prisoner while inside or outside prison (for example while on a release to work programme) including the risks that would pose to the public. This is also to enable Corrections to appropriately manage prisoners within a prison environment. That security classification of each prisoner must be undertaken and reviewed “in the prescribed manner”.

[8] The Act and the Corrections Regulations 2004 (the Regulations) provide some guidance as to the correct process. Further guidance is contained in internal Corrections guidelines and other documents.

• [6] The security classification process was comprehensively summarised in the judgment under appeal.^[8] For present purposes, we will focus on the key features of the scheme that are relevant to Ms Mitchell’s appeal grounds.
• [7] The process for assigning and reviewing security classifications is set out in various source documents including the Act, the Corrections Regulations 2005 (the Regulations) and internal Corrections guidelines such as Prison Operations Manuals. There are five security classification levels: maximum, high, low-medium, low, and minimum. The Regulations state that prisoners should be assigned the lowest classification at which they can “safely and securely be managed given the assessment of the level of risk posed by a prisoner”.^[9]
• [8] An electronic points-based system (the IOMS), based on a range of internal and external risk factors, is used to provide a preliminary or indicative classification (Preliminary Score). This Preliminary Score can be manually overridden, however, and is not determinative of a prisoner’s final security classification.
• [9] A prisoner’s security classification must be reviewed at least every six months, or if the prisoner’s circumstances change significantly.^[10] A prisoner who is dissatisfied with their assigned classification is entitled to apply to the Chief Executive of Corrections for a reconsideration, following which their classification will either be confirmed or changed.^[11]

Ms Mitchell’s security classifications

• [10] Ms Mitchell was held as a remand prisoner from 18 March 2020 until 31 March 2021, when she was sentenced to two years and three months’ imprisonment.
• [11] While Ms Mitchell was in prison on remand, nine incident reports were completed by prison staff regarding her behaviour.^[12] A description of each of the reported incidents, based on the evidence of Ms Roberts, the Senior Advisor to the Regional Commissioner of Corrections for the Lower North region, is set out in the judgment under appeal.^[13] The Judge summarised the relevant evidence as follows:^[14]

Ms Roberts described Ms Mitchell’s behaviour towards and her interactions with prison staff and other prisoners as demonstrably negative and often aggressive given too that they included multiple allegations of abuse and threats to staff, unlawful activating of the sprinkler system, property damage, abusing the prison intercom system, assaulting staff, smearing faeces on walls and refusing to engage with health staff for a Covid-19 assessment.

• [12] Remand prisoners are not assigned a security classification. Ms Mitchell was therefore first assigned a security classification following her sentencing. During her time as a sentenced prisoner, Ms Mitchell had her security classification assessed on three occasions. The relevant chronology is as follows:

(a) First security classification. In May 2021 Ms Mitchell was assigned her first security classification. The ultimate classification was “low”. There were three stages to the assignment of Ms Mitchell’s first classification: the Preliminary Score, the Override Decision, and the Reconsideration Decision. Her Preliminary Score, generated by the IOMS, suggested a security classification of “minimum”. This was manually overridden to “low-medium” to take into account Ms Mitchell’s conduct while she was a remand prisoner, based on the incident reports referred to above (the Override Decision). Ms Mitchell then applied for an internal reconsideration of the Override Decision by Corrections. This application was determined by Ms Roberts, and resulted in her classification being reduced from “low-medium” to “low” (the Reconsideration Decision).

(b) Judgment under appeal. Ms Mitchell applied for judicial review of the Reconsideration Decision in the High Court, on the basis that her classification should have been reduced to “minimum” rather than “low”. Gendall J dismissed Ms Mitchell’s application.^[15]

(c) Second security classification. In November 2021, the first six‑monthly review of Ms Mitchell’s security classification took place. This time, her Preliminary Score generated a recommended security classification of “low-medium”. On 15 November 2021 that was overridden to “low”, on the basis that Ms Mitchell could be appropriately managed at that level.

(d) Third security classification. In May 2022 the second six-monthly review of Ms Mitchell’s security classification took place. Again, her Preliminary Score generated a recommended security classification of “low-medium”. That decision was again overridden to “low” on the basis that Ms Mitchell could be appropriately managed at that level.

(e) Release from prison. On 16 June 2022 Ms Mitchell was released from prison on parole.

The High Court decision

• [13] Ms Mitchell alleged various errors by Corrections in her judicial review proceeding. Although the arguments she advanced were wide-ranging, her key argument appears to have been that her behaviour on remand was wrongly taken into account in both the Override Decision and the Reconsideration Decision. Ms Mitchell submitted that her final security classification should have simply been that generated by her Preliminary Score — “minimum”.
• [14] The Judge dismissed Ms Mitchell’s application for judicial review on the basis that:

(a) The Preliminary Score is “only a preliminary or indicative security classification”. It “is not determinative of the ultimate outcome and can be overridden when appropriate”.^[16] This recognises that the inputs into the Preliminary Score “may not capture all matters that are relevant to the risk assessment of a particular prisoner”.^[17]

(b) Prisoners do not have a legitimate expectation that no override will be applied or that the Preliminary Score will be the final outcome of any security classification decision.^[18]

(c) The relevant guidelines used by Corrections include specific factors which might provide the basis for a manual override, but those factors are not in any way restrictive. Rather, the guidelines “provide a broad discretion to manually override a Preliminary Score so long as that override is not based on a factor that has already been incorporated in the initial assessment itself”.^[19]

(d) Convictions or formal misconduct charges are not required for a behavioural incident to be considered as relevant to a security classification decision.^[20]

(e) Here, Ms Mitchell’s Preliminary Score failed to take into account her behaviour while on remand. It was appropriate for such behaviour to be considered, however, given that at least one incident was the subject of police consideration at the time Ms Mitchell’s judicial review application was heard and there was sufficient evidence of the other incidents, which the Judge considered had involved “what might be seen as lower level violence by a prisoner in custody”. The safety of both prison staff and other inmates is relevant to classification questions.^[21]

(f) Ms Mitchell’s right to natural justice was appropriately addressed through the reconsideration process, which was partially successful and resulted in Ms Mitchell’s classification being reduced from “low‑medium” to “low”.^[22]

• [15] The Judge concluded that:

[48] I accept that [the Reconsideration Decision] was a proper one in line with the statutory and regulatory regime and the requirements for an override as set out in the Guidelines. [The Reconsideration Decision] as I see it was one to assign to Ms Mitchell the lowest level of security classification at which she could be safely and securely managed given the level of risk she posed. In making [the Reconsideration Decision] Ms Roberts took into account all relevant information readily available, she recorded her decision in writing and she provided this with reasons to Ms Mitchell on 13 May 2021. The incident reports which she had taken into account were not information that was previously considered during the automatic points-based calculation of Ms Mitchell’s security classification and therefore no double counting occurred here.

Is the appeal moot?

• [16] Corrections submitted that the appeal is moot. As set out above, the first security classification was superseded by the second and third security classifications, both of which also resulted in a “low” security classification, following manual override decisions. Ms Mitchell did not seek reconsideration of either of the subsequent security classification decisions. She was subsequently released on 16 June 2022. Accordingly, even if Ms Mitchell were to succeed in reviewing the first security classification decision, that would not give rise to any practical consequences.
• [17] Ms Mitchell accepted that the first security classification is no longer operative. She submitted, however, that her appeal raises broader issues of more general application, and that it is in the interests of justice for the appeal to be determined on its merits.

23. Proceedings will be “moot” or “academic” if there is no live dispute between the parties or, as the Supreme Court put it in R v Gordon-Smith, “where the substratum of the ... litigation between the parties has gone and there is no matter remaining in actual controversy and requiring decision”.^[23]

• [19] Here, we accept that the appeal is moot, given that the first security classification is no longer in effect. That is not the end of the matter, however, as we explain below.

Should this Court determine the appeal, despite it being moot?

Approach to moot appeals — legal principles

• [20] The Supreme Court observed in Gordon-Smith that mootness is not a matter that deprives a court of jurisdiction to hear an appeal.^[24] Rather, the issue is one of judicial policy:^[25]

In general, appellate courts do not decide appeals where the decision will have no practical effect on the rights of parties before the court, in relation to what has been at issue between them in lower courts. ... But in circumstances warranting an exception to that policy, provided the court has jurisdiction, it may exercise its discretion and hear an appeal on a moot question.

The Court referred to three primary reasons for the appellate courts’ “general policy of restraint ... in addressing moot questions”, as identified by the Supreme Court of Canada in Borowski v Canada (Attorney General):^[26]

... first, the importance of the adversarial nature of the appellate process in the determination of appeals, secondly, the need for economy in the use of limited resources of the appellate courts and, thirdly, the responsibility of the courts to show proper sensitivity to their role in our system of government. In general advisory opinions are not appropriate.

• [21] The Supreme Court noted in Baker v Hodder that for public law cases, questions of mootness “may be less compelling” when deciding whether to entertain an appeal. However, the court’s discretion to hear a moot appeal is not limited to public law cases.^[27] The Court commented that there is no “test” for the exercise of the discretion but, in light of the “policy of restraint” described above, “a decision to hear a moot appeal should be made only in exceptional circumstances”.^[28] Exceptional circumstances could arise from the circumstances of the particular case (the Court gave the example of serious procedural unfairness at first instance) or the broader public interest (the Court gave the example of an appeal which raises an important legal point).^[29]
• [22] Ms Mitchell referred, by analogy, to the matters identified in the recent decision of the Supreme Court in Ellis v R (Continuance).^[30] That case arose in a somewhat different context, involving consideration by the Supreme Court of the factors relevant to determining whether to allow a criminal appeal to continue following the death of the appellant. The Court considered that the overriding consideration was whether continuation is in the interests of justice.^[31] The majority went on to identify a number of other, non‑exhaustive factors which are relevant to the assessment.^[32] We accept that many of those factors have been previously identified as being relevant in a civil context or (if not expressly identified in previous civil cases) may nevertheless be relevant, by analogy.
• [23] With reference to the decisions we have referred to, we summarise the (non‑exhaustive) factors that are potentially relevant to the exercise of our discretion as to whether to hear this appeal, despite it being moot, as follows:

(a) The importance of the adversarial nature of the appellate process in the determination of appeals.^[33]

(b) The strength of the grounds of appeal.^[34]

(c) The need for economy in the use of limited resources of the appellate courts.^[35] This may include consideration of whether the nature of the order which could be made by the court justifies the expenditure of limited judicial (or court) resources to resolve a moot appeal.^[36]

(d) Whether continuing the appeal “would go beyond the judicial function of resolving concrete disputes and involve the court in free-standing, legislative-type pronouncements more properly left to the legislature itself”.^[37] Courts must “show proper sensitivity to their role in our system of government” and “[i]n general advisory opinions are not appropriate.”^[38]

(e) Whether an appeal raises an issue of “significant public importance which is highly likely to come before the court again at some point”.^[39] It is relevant whether the issue is “otherwise evasive of appellate review”.^[40] It is also relevant whether the appeal raises “a systemic issue related to the administration of justice”.^[41]

• [24] The above factors are non-exhaustive. The extent to which these factors, or other additional factors, are engaged and the weight to be accorded to them will be case-specific. In the circumstances of this appeal, there is no concern that the appeal would not be determined in an appropriate adversarial context. Nor would determining the appeal require the Court to go beyond its proper judicial function. The key relevant factors, in our view, are those set out at [23(b)], [23(c)] and [23(e)] above. Of those, we see the strength of the appeal, and whether it raises issues of general or public importance, as being the most critical factors bearing on the exercise of our discretion in this case.

The strength of the appeal

• [25] Ms Mitchell advanced numerous arguments on appeal. Some of them were difficult to follow, others do not appear to have been advanced before the High Court or were or peripheral relevance. We therefore approach our assessment of the merits of the appeal by focussing on the key steps in the Judge’s reasoning process (as summarised at [14] above) and considering any alleged errors made by the Judge in relation to those matters.
• [26] First, the Judge found that the Preliminary Score is only a preliminary or indicative classification, and that prisoners do not have a legitimate expectation that no override will be applied, or that the Preliminary Score will be the final outcome of any security classification decision.^[42]
• [27] It is not reasonably arguable, in our view, that this finding was in error. The Judge’s reasoning is compelling. The initial exercise of allocating points through IOMS is simply a tool to aid in the risk assessment process. It may not, however, capture all relevant risk factors in each case. The Judge’s finding is principled, in accordance with the statutory and administrative scheme, and consistent with previous decisions including Smith v Attorney-General^[43] and Taylor v Chief Executive of the Department of Corrections.^[44] As Ellis J observed in Taylor:^[45]

It is tolerably clear that the override option exists precisely because filling out the review form and applying the points system is intended largely to be a mechanical (and therefore objective and consistent) exercise. The possibility of an override recognises the reality that the guidelines and the form may not capture all matters that are relevant to the risk assessment required by s 44 [of the Act] in relation to a particular prisoner. It contemplates that there may be a departure from policy but ... requires clear reasons for that to be given. The guidelines also make it clear that there cannot be a departure from policy (an override) simply because an application of the points system does not result in adequate weight being given to one of the matters expressly required to be considered.

• [28] We further note the subsequent overrides in Ms Mitchell’s case, in respect of the two six-monthly classification reviews, reduced her security classification level from the “low-medium” indicated by the Preliminary Score to “low”. Perhaps not surprisingly, Ms Mitchell has not sought to argue that she had a legitimate expectation that she would be assigned the security classification that was generated by the Preliminary Score in relation to either of those decisions.
• [29] We also do not think it is reasonably arguable that the override involved a “double counting” of factors which were already factored into the Preliminary Score. We agree with the Judge that:^[46]

Here, behavioural incident reports were not included as part of the automated matrix of factors Ms Fili [the assessing officer] properly applied in carrying out the initial security classification using the IOMS form. This is precisely what led to [the recommending officer] Ms Reedy’s override of Ms Mitchell’s Preliminary Score, given that it failed to take into account her conduct while on remand, conduct which according to Ms Roberts was repeated and significantly troubling.

• [30] On the evidence before the Court, it is not reasonably arguable that the information in the behaviour incident reports was already incorporated into the IOMS automated matrix of factors, and hence reflected in Ms Mitchell’s Preliminary Score. It was therefore appropriate for this information, which is clearly relevant to assessing risk, to be considered at the override stage, as the Judge found.^[47]
• [31] The Judge was also correct, in our view, to find that convictions or formal misconduct charges are not required for a behavioural incident to be considered as relevant to a security classification decision.^[48] Any submission to the contrary overlooks the fundamentally different nature of the prison disciplinary regime and the security classification process. As Woolford J observed in Bell v Chief Executive of the Department of Corrections:^[49]

The misconduct regime and security classification decisions are separate matters. Charges of misconduct are subject to the prosecution and hearing regime specified in sch 7 of the Corrections Regulations. Misconduct must be proved beyond reasonable doubt. In contrast, security classification decisions are inherently about risk.

• [32] Similarly, in Newton v Superintendent of Rimutaka Prison, which was referred to by the Judge in the judgment under appeal,^[50] Wild J considered an argument that Corrections had erred by considering a prisoner’s drug test results in a security classification decision, after disciplinary charges based on those results had been dismissed by a Visiting Justice for procedural reasons. Wild J observed that:^[51]

This argument equates the prison’s essentially judicial inmates disciplinary offence process ... with the prison’s day-to-day administration, including decisions on security classification ... Taken to its logical extreme, the argument would preclude the prison ... from taking any administrative action unless and until appropriate disciplinary offences were proved against the inmate.

• [33] Ms Mitchell made several submissions to the general effect that her case was distinguishable from previous security classification cases and that the reasoning in those cases did not therefore apply to her situation. These submissions were somewhat difficult to follow at times. In addition, some of these arguments do not appear to have been advanced in the High Court. As a general observation, however, it does not appear to us that any of the differences relied on by Ms Mitchell would lead to a different outcome on appeal. For example, on the information before us, Ms Mitchell has failed to establish that there is (or should be) a materially different approach to assessing the security classification of male and female prisoners. Other matters referred to by Ms Mitchell, including that her offending was less serious than that in some other cases, were considered when assessing the Preliminary Score.
• [34] In conclusion, we assess the strength of Ms Mitchell’s appeal as weak. There is nothing to suggest that the security classification process in her case went fundamentally awry, for any of the reasons she has advanced.

Other relevant factors

• [35] We now turn to consider the other factors relevant to the exercise of our discretion as to whether to determine this appeal, as summarised above, to the extent they are engaged in this case.
• [36] Of those factors, the most relevant is whether the appeal raises an issue of “significant public importance which is highly likely to come before the court again at some point”.^[52] In our view it does not. The issues raised are largely fact-specific and, to the extent that issues of wider significance are raised, they largely relate to principles that have been considered in previous cases.
• [37] It necessarily follows that the need for economy in the use of the limited resources of the appellate courts weighs against determination of the appeal.^[53] This is not a case where the nature of the order which could be made by this Court would justify the expenditure of limited judicial and court resources to resolve a moot appeal.^[54]

Conclusion

• [38] The security classification decision that is the subject of the appeal is now moot. It has since been superseded by two subsequent classification decisions (resulting in the same “low” security classification) that were not challenged, and Ms Mitchell’s release on probation. The grounds of appeal advanced by Ms Mitchell relate to factual circumstances that no longer exist. For the reasons outlined above, it is our view that the merits of the appeal are weak. This is a factor that weighs heavily against exercising the Court’s discretion to determine a moot appeal. The Judge’s reasoning (as summarised at [14] above) is compelling. It is consistent with previous case law and does not disclose any apparent error.
• [39] The statutory and administrative scheme for allocating a security classification to prisoners has been comprehensively considered in several cases.^[55] Ms Mitchell’s appeal does not raise any new, and seriously arguable, issues of broader public importance relating to the implementation of the security classification process by Corrections. On the contrary, the issues raised by this appeal are largely fact-specific and limited to the circumstances of Ms Mitchell’s case.
• [40] For the reasons outlined, it is our view that the Court should not exercise its discretion to hear the appeal.

Result

• [41] Corrections’ application for leave to adduce further evidence on appeal is granted.
• [42] We are satisfied that the appeal is moot and that this is not a case where the Court should exercise its discretion to hear the appeal. The appeal is dismissed accordingly.

Solicitors:
Luke Cunningham Clere, Wellington for Respondent

^[1] Corrections Act 2004, s 47.

^[2] Mitchell v Attorney-General [2021] NZHC 2946 [Judgment under appeal].

^[3] Corrections Act, s 47(3).

^[4] Court of Appeal (Civil) Rules 2005, r 45.

^[5] Rae v International Insurance Brokers (Nelson Marlborough) Ltd [1998] 3 NZLR 190 (CA) at 192; and Paper Reclaim Ltd v Aotearoa International Ltd (Further Evidence) (No 1) [2006] NZSC 59, [2007] 2 NZLR 1 at [6].

^[6] See for example Gorges v Chief Executive of the Department of Corrections [2017] NZHC 2518; Taylor v Chief Executive of the Department of Corrections [2015] NZHC 2196; Genge v Chief Executive of the Department of Corrections [2018] NZHC 1302; Smith v Attorney‑General [2017] NZHC 136, [2017] NZAR 331; and Bell v Chief Executive of the Department of Corrections [2021] NZHC 413.

^[7] Judgment under appeal, above n 2 (footnotes omitted).

^[8] See [7]–[24].

^[9] Corrections Regulations 2005, reg 44(1).

^[10] Corrections Act, s 47(3)(b).

^[11] Section 48(2); and Corrections Regulations, reg 51.

^[12] Judgment under appeal, above n 2, at [27] and [43].

^[13] At [27].

^[14] At [43].

^[15] Judgment under appeal, above n 2, at [69]–[71].

^[16] At [16].

^[17] At [17], citing Smith v Attorney-General, above n 6, and Taylor v Chief Executive of the Department of Corrections, above n 6.

^[18] At [19].

^[19] At [35].

^[20] At [44]–[46], citing Newton v Superintendent of Rimutaka Prison HC Wellington CIV‑2003‑485‑1778, 5 September 2003 at [35].

^[21] At [37]–[38].

^[22] At [41] and [52].

23. ^[23] R v Gordon-Smith [2008] NZSC 56, [2009] 1 NZLR 721 at [14] (footnote omitted), quoting Finnigan v New Zealand Rugby Football Union Inc (No 3) [1985] NZCA 111; [1985] 2 NZLR 190 (CA) at 199 per Richardson J. The Court also cited the principle referred to by Viscount Simon LC in Sun Life Assurance Co of Canada v Jervis [1944] AC 111 (HL) at 114 that “it is an essential quality of an appeal fit to be disposed of by this House that there should exist between the parties a matter in actual controversy which the House undertakes to decide as a living issue”.

^[24] R v Gordon-Smith, above n 23, at [16], citing R v Secretary of State for the Home Department, ex parte Salem [1999] UKHL 8; [1999] 1 AC 450 (HL) at 456–457.

^[25] At [16] (footnote omitted), citing Borowski v Canada (Attorney General) [1989] 1 SCR 342 at 353.

^[26] At [18], citing Borowski v Canada (Attorney General), above n 25, at 358–363. The Court also referred to R v Smith [2004] 1 SCR 385.

^[27] Baker v Hodder [2018] NZSC 78, [2019] 1 NZLR 94 at [33].

^[28] At [32]–[33].

^[29] At [33].

^[30] Ellis v R (Continuance) [2022] NZSC 114, [2022] 1 NZLR 239.

^[31] At [48] and [57] per Glazebrook J, [152] per Winkelmann CJ, [233] per Williams J and [294] per O’Regan and Arnold JJ.

^[32] At [57]–[58] and [144] per Glazebrook J and [278] and [292]–[293] per O’Regan and Arnold JJ.

^[33] R v Gordon-Smith, above n 23, at [18] and [20]. The Supreme Court elaborated in R v Gordon‑Smith that “having a stake in the outcome fosters full argument on the questions before the court”; and that “[t]he need for legal principles to be applied to particular facts is ... a valuable discipline for the courts in determining those principles.” See also Ellis v R, above n 30, at [57(a)] per Glazebrook J and [278] and [292] per O’Regan and Arnold JJ.

^[34] Ellis v R, above n 30, at [57(b)] per Glazebrook J and [278] and [292] per O’Regan and Arnold JJ.

^[35] R v Gordon-Smith, above n 23, at [18], citing Borowski v Canada (Attorney General), above n 25, at 358–363.

^[36] Ellis v R, above n 30, at [57(f)] per Glazebrook J and [278] and [292] per O’Regan and Arnold JJ.

^[37] At [57(g)] per Glazebrook J and [278] and [292] per O’Regan and Arnold JJ.

^[38] R v Gordon-Smith, above n 23, at [18].

^[39] At [24].

^[40] Ellis v R, above n 30, at [57(e)(i)] per Glazebrook J and [278] and [292] per O’Regan and Arnold JJ.

^[41] At [57(e)(ii)] per Glazebrook J and [278] and [292] per O’Regan and Arnold JJ.

^[42] Judgment under appeal, above n 2, at [19], [21] and [52].

^[43] Smith v Attorney‑General, above n 6.

^[44] Taylor v Chief Executive of the Department of Corrections, above n 6.

^[45] At [88] (footnote omitted). See also Smith v Attorney‑General, above n 6, at [47].

^[46] Judgment under appeal, above n 2, at [37], and see also [48] in the context of the Reconsideration Decision.

^[47] At [37]–[39], and see also [48]–[51].

^[48] See [38], [44]–[45] and [48].

^[49] Bell v Chief Executive of the Department of Corrections, above n 6, at [77] (footnote omitted), citing Genge v Chief Executive of the Department of Corrections, above n 6, at [12].

^[50] Judgment under appeal, above n 2, at [45].

^[51] Newton v Superintendent of Rimutaka Prison, above n 20, at [35].

^[52] R v Gordon-Smith, above n 23, at [24].

^[53] At [18].

^[54] Ellis v R, above n 30, at [57(f)] per Glazebrook J and [278] and [292] per O’Regan and Arnold JJ.

^[55] See for example Gorges v Chief Executive of the Department of Corrections, above n 6; Taylor v Chief Executive of the Department of Corrections, above n 6; Genge v Chief Executive of the Department of Corrections, above n 6; Smith v Attorney-General, above n 6; and Bell v Chief Executive of the Department of Corrections, above n 6.