Applications for leave to appeal before trial: W (CA624/2022) v R [2023] NZCA 397

Today is the nineteenth anniversary of the start of this site. Celebrations have been restrained and dignified, although I do seem to have to correct a lot of typing mistakes just now.

My fear that the day would pass without producing an excuse for me to mention that significant occasion has been allayed by the Court’s publication of its reasons in W (CA624/2022) v R [2023] NZCA 397.

This case deals with the criteria for deciding to grant leave for a pre-trial appeal. Occasions for such appeals are listed in the Criminal Procedure Act 2011, ss 215, 217, 218, which are reproduced at [5], [6], and (for s 218) summarised at [7]. The Court considers the policy matters that the legislation addresses at [16]-[21], and notes the relevant law in comparable jurisdictions at [23]-[24].

New Zealand ’s leading case law is in the Leonard and the Hohipa decisions, considered at [25]-[30] for Leonard, listing the decision criteria at [28]-[29]. Hohipa is referred to at [31]-[33]. The present case (W) does not overrule those criteria, but adds to them in the light of experience. Recent experience and issues arising from it are described at [34]-[41]. The two questions raised by this experience are stated at [41].

The Court does not attempt to answer the first question which is why are so many applications for leave to appeal before trial made? My own guess is that (apart from the misleading heading to s 215, and similarly to 217 and 218 which suggests that there is a “right” to appeal before trial, when there is only a right to seek leave to appeal in the relevant occasions) counsel do not wish to be criticised at a later stage of the proceedings for not having sought leave pre-trial, and do not want any inference adverse against their client to be drawn from the absence of any pre-trial application for leave.

But the other point raised in [41] is more important: the need for counsel to assist in the determination of an application for leave by particularising the issues (see [53]-[55]).

The leave criteria are restated at [52], with the interests of justice being the overall consideration. There are 13 criteria in the list, labeled (a) to (m). In Leonard the Court had identified 5 criteria favouring leave being granted (repeated here at [5]), and 5 against (see [6]). So, what are the 3 that are added in the present case? The criteria in W at [52(a)-(c)] are not in the Leonard list but they do appear in other case law.

My impression is that all the criteria, those in Leonard and those in the present case, are relevant. There is no express departure from the Leonard criteria, but the overlap of particular criteria (for example, Leonard (a) in favour of granting leave and W (d), (e), and (f); Leonard (b) and W (i) and (m)) suggests that they should all be read together. The emphasis in W is on the avoidance of unnecessary delay in the completion of the criminal proceedings [55].

And, as I said in Jeremy Finn and Don Mathias, Criminal Procedure in New Zealand (3rd ed, 2019, Thomson Reuters NZ Ltd, Wellington) at [14.6], it is important to advance at the earliest opportunity strong grounds in an application for leave to appeal, as there is no appeal against a refusal of leave.

Of particular interest to law reformers is the reference at [23] to the law in comparable jurisdictions: Canada, New South Wales, South Australia, Victoria, and England and Wales.

The plea of previous conviction: Mitchell v Police [2023] NZSC 104

Protection against double jeopardy can be given by statute and by common law. Statutory protection can exist in the form of provisions for pleas of previous conviction or previous acquittal, and by a right not to be tried or punished for an offence for which a final conviction or acquittal has been entered. Common law protection can take the form of a stay of proceedings to prevent abuse of process, or in the sentencing process by the avoidance of double punishment, or in the plea of autrefois convict to the extent that it has not been replaced by legislation.

Statutory rules concerning the plea of previous conviction can be difficult to interpret and apply. Such difficulties in New Zealand law are surveyed in Mitchell v Police [2023] NZSC 104 (11 August 2023). The relevant provision here is s 46(1)(b) of the Criminal Procedure Act 2011.

The issue is quite simple, as is the Court’s application of its interpretation of the legislation: [2], [68]-[69]. Arriving at the correct interpretation of the legislation was not entirely straightforward, and the Court noted at [63] that there is “some opacity” in the drafting of s 46(1). The phrase “arising from the same facts” marks a change from the previous legislative focus on the elements of the charges, in favour of placing the facts as central to the inquiry. And here, it is “substantial” similarity that is required: in s 46(1)(b), “those facts” means substantially those facts, and what is required is substantial similarity in the acts that make the charges punishable.

So, Mr Mitchell’s charges referred to common acts (driving, with some alcohol in his breath: [67]) and to facts that made each charge punishable (having a breath alcohol level in excess of the legal limit, and having any alcohol in his system in breach of his zero alcohol licence). Those punishable facts were different for each charge, and there was nothing unlawful in his being convicted on each [69].

Substantial sameness does not include single acts that have several victims ([64] and footnote 90). But it probably includes facts which are accompanied by different mental elements ([64] footnote 89, but Ellen France J had some reservations about this [78]).

Section 46(1) has changed the law by reforming the very narrow approach to the availability of the plea of previous conviction that some authorities had applied: [62], [46]. This means that, for example, the following charges (in Hickson, referred to at [46]) would now be held to arise from substantially the same facts: (a) selling liquor in licensed premises when such premises were directed to be closed; (b) exposing liquor for sale in such premises during such a period; and (c) opening such premises for the sale of liquor during such a period.

The Court in Mitchell uses that example to illustrate the former law’s rejection of the plea of previous conviction where a “subtle difference” between the elements of the offences could be discerned [45]. Again, referring at [47] to Brightwell, the offences of presenting a firearm without lawful and sufficient purpose, and by (substantially) the same act threatening to do grievous bodily harm, were a further illustration of the subtle difference that would no longer be recognised. Pursuant to Mitchell, the approach would now be to ask whether the punishable act for each offence is substantially the same, the answer for Hickson and Brightwell being yes it was in those cases.

Ellen France J, concurring but for slightly different reasons, thought it is questionable whether s 46 places a controlling role upon the facts [76]. The point here is that the elements of the offences will “in most cases” (see footnote [94]) assist in defining the necessary facts for the purposes of s 46. She also had reservations about the use of the concept of common punishable act [77].

And, I add for law students to consider: When are the punishable acts “substantially” the same for each offence? If the difference between exposing liquor for sale and selling that liquor is too subtle to exclude the plea of previous conviction (Hickson as reviewed in Mitchell), why isn't the difference between having alcohol in the breath when driving and having excess alcohol in the breath on that occasion too subtle to exclude the plea? Will the rejection of the search for subtle differences in the elements of the offences lead to a search for subtle differences in the punishable acts for each, as grounds for rejecting the plea of previous conviction? And would that be objectionable? Has the precedent value of future cases on this topic been reduced by the mandated emphasis on facts, and, if so, is that consistent with the rule of law?

Fair trials for unrepresented defendants - the use of amici curiae in Canada

In Canada, an amicus curiae may perform a variety of functions, limited by the role of an amicus as a “friend of the court”: R v Kahsai, 2023 SCC 20 (28 July 2023) at [40]. The duty of loyalty to the court must always be upheld, whatever the functions that the judge gives an amicus in an individual case. Therefore, the amicus has no duty of loyalty to a defendant, and is not in a solicitor-client relationship with a defendant. The amicus cannot be dismissed by the defendant. When required to advocate for the defendant, the amicus still does not represent the defendant. The amicus cannot be given functions that undermine the judge’s impartiality, for example by making strategic decisions to assist the defendant ([41]-[42]).

The right of a defendant to self-represent is important, and this may restrict what an amicus can do. Decisions on plea, mode of trial, whether to give or call evidence, and what defence to rely on are decisions that must be left to the defendant. But those decisions must be informed and reality-based, and made within the confines of the law. So, evidence can only be given in accordance with the law of evidence, and only defences that are lawfully available can be advanced by the defendant. Nor can the defendant require an amicus to act unethically. The defendant’s decisions on the conduct of the defence must be respected even though an amicus may have made different decisions. But the judge may require the assistance of the amicus in testing the evidence of prosecution witnesses, although this may require a fine balancing against the defendant’s right to control the defence. Within the limits of the strategic choices made by the defendant, the amicus should always be entitled to test the strength of the prosecution case ([43]-[46], [49]).

Avoiding imbalance in the adversarial process when a defendant is unrepresented by counsel puts a trial judge under a duty to ensure the proceedings respect the defendant’s fundamental rights. However, judicial help must not have the appearance of undermining impartiality [54]. The prosecutor also has duties of ensuring the fairness of trials [55], particularly in relation to ensuring that unrepresented defendants know about their rights to obtain disclosure [56]. The appointing of an amicus may be necessary where the duties of judge and prosecutor in relation to unrepresented defendants cannot go far enough to achieve trial fairness or the appearance of fairness. For example, where a defendant, although fit to stand trial, displays symptoms of mental health challenges, or simply refuses to participate in the trial ([57]-[58]).

Appointments of amici are reserved for exceptional circumstances and are not routine. Relevant considerations will be the nature and complexity of the charges, the mode of trial (judge alone or jury), the attributes of the defendant, the need for assistance in advancing the defence case or in testing the prosecution evidence, and the extent to which the judge and the prosecutor can provide assistance ([60]-[61], [64]-[66]).

In New Zealand we address the ways in which assistance may be given by counsel appointed by the court by categorising such counsel as special advocates, independent counsel, amici curiae, and standby counsel. Read more about this in Jeremy Finn and Don Mathias, Criminal Procedure in New Zealand (3rd ed, 2019, Thomson Reuters Ltd, Wellington) at [11.5]. Go on, you know you want to.

Stage fright in the CCRC?

In a 2021 review of the operation of the Criminal Cases Review Commission (the Commission) for England, Wales and Northern Ireland, the Westminster Commission on Miscarriages of Justice had a significant concern that the Commission is at risk of being too deferential to the appeal court. Or, as I would say, stage fright might cause it to miss its cue.

A recommendation was made for change to the criterion for referral of cases to the appeal court.

Other concerns raised in that review included that the Commission was under-funded and over-worked.

Glancing at the online statistics published by the New Zealand Criminal Cases Review Commission (CCRC), in the light of the referral criterion set out in s 17 of the Act, there are currently 136 completed cases, including 1 referral to the court. While there is only one statutory ground for referral (the interests of justice), the published grounds for not referring a case are more informative insofar as they separate out some specific reasons in 70 cases, leaving 65 described as “Not in the interests of justice”. On this approach, the interests of justice are separate from the specific reasons (active appeal on same grounds, appeals process available, applicant deceased, application withdrawn, incomplete application and no criminal conviction). Presumably, then, the interest of justice here means the prospects of the court allowing the appeal (s 17(2)(c)).

So, in 65 out of 135 (48 per cent) of applications that do not progress to the court, the CCRC has concluded that the prospects of the court allowing the appeal are insufficient to justify a referral. When combined with the other reasons for refusing to refer a case, the CCRC has referred 1 in 136 (0.74 per cent) of cases it has completed. But, separating out those cases where there are specific statutory reasons for refusal to refer an application, the 65 occasions where the prospects of an appeal being allowed are the criterion have included one referral, so that is approximately 2 per cent of these “prospects” cases.

Plainly, the prospects of the court allowing the appeal is a very significant consideration when assessing the interests of justice. That is quite sensible, but one can ask whether the CCRC may be being too conservative in making that assessment. Another thought would be that appellate courts set reasonable doubt at too high a standard.

Returning to the experience in England, Wales and Northern Ireland (where the rate of referral to the appeal court is about 3 per cent) and remembering the horrific miscarriages of justice recognised in the Post Office appeal cases, in which inappropriate reliance was placed on a computer accounting system, and more recently in the successful appeal in the Malkinson case (see the article in The Guardian on 27 July 2023) in which the Commission had twice refused to make a referral, not having carried out the necessary investigatory work, we can wonder whether the criterion for referral could be improved. The Westminster inquiry recommended (at p 37) that their criterion could be redrafted:

“the 'real possibility' test should be redrafted to expressly enable the CCRC to refer a case where it determines that the conviction may be unsafe, the sentence may be manifestly excessive or wrong in law or where it concludes that it is in the interests of justice to make a referral. By definition this would include all cases where it finds that a miscarriage of justice may have occurred including 'lurking doubt' cases.”

As funding will inevitably become inadequate, and resources inappropriately limited, the only protection that applicants will have against the CCRC’s wrongful refusal to refer cases will be a referral criterion that avoids the danger of what I have here called stage fright.

The New Zealand CCRC was established on 1 July 2020, so it may be too soon for an accurate calculation of the rate of referrals to be made. It is most unlikely that the New Zealand justice system is about four times more accurate at convicting people than are the courts in England and Wales (needing to refer only 0.74 per cent of applications, compared to 3 per cent for England and Wales). It would be more informative to compare the “prospects” statistics, if one could dig them out of the 2021 review for Engalnd and Wales. In any event, the CCRC should accept an error rate - that is, a rate of referrals which do not result in a successful appeal - as something similar to the rate of unsuccessful appeals that occurs in legal practice, which seems to be about 66 per cent. In a sense the CCRC is acting like counsel for the appellant where fresh evidence has been obtained to challenge a conviction, or where new considerations have been unearthed to call into question the appropriateness of a sentence. Counsel will pursue an appeal in those circumstances, being aware of the likelihood of failure from the point of view of statistics, and the CCRC should too.

In England and Wales the rate of successful referrals has been about 70 per cent (see p 35 of the review) and this has been criticised as indicating that more cases should be referred. One might say that it is the courts' error rate, not the CCRC's, that matters. Judicial blindness to the risk of error was a strong motivator for the establishment of the CCRC.

Three quite interesting appeals: live issues, police photography, and the plea of previous conviction

Three quite interesting appeals:

Refraining from irrelevancies

There are times when, no matter how interesting a point is and how helpful it would be to decide it, an appellate court will not address issues that are irrelevant to the appeal before it. A recent example is Moss v R (Bahamas) [2023] UKPC 28 (25 July 2023). Whether duress is an available defence at English common law to a charge of conspiracy to murder where the murder is actually committed remains undecided, because in this appeal, arising from a trial in which the defendant (appellant) had been charged with and convicted of conspiracy to murder, there was insufficient basis in the evidence to establish a foundation for duress. The defence was that there had been no conspiracy. It had been correct for the defence at trial not to seek to rely on duress, and the judge had correctly not left it with the jury as a potential defence. The question was only raised on appeal, and the Court of Appeal of the Commonwealth of the Bahamas had correctly rejected it.

The Board differed from the Court of Appeal on whether there was some evidence to support some of the requirements for a defence of duress ([57]-[58]), but agreed with that Court on one of the requirements, namely that there was no assertion in the defence evidence that there was no reasonable evasive action that the defendant could have taken to avoid the harm threatened ([68]).

The refusal to decide the common law point in this case is at [53], where the Board leaves it to a case in which duress arises on the facts.

Behind this refusal is, firstly, the recognition that counsel should not be made to do work that will be of no relevance to counsel’s case. What client would willingly pay their lawyer to research and argue an issue that could not assist in deciding the case? Similarly for publicly funded litigation. Counsel might be reluctant to take on appeals if there was a risk of having to futilely argue a cluster of complex issues that would only divert their time and energy unproductively. A second reason for the refusal to consider irrelevant points is that a proper factual context will assist in deciding what answer is supported by policy. The common law has developed piecemeal, working out case by case what is the reasonable answer to the issues that have to be decided in reaching a conclusion. But, you may say, couldn’t the Board in Moss simply have imagined that the defendant had had no reasonable opportunity to take evasive action? This might weaken the second reason for refusing to consider the point, but not the first.

But a third reason for refusal to consider points that have no basis in the facts of a case is more fundamental because it is constitutional. The courts refuse to trespass on what is properly the domain of the legislature. Courts decide the issues that arise between the parties to a case, the legislature addresses proactively broad matters of policy. This is not to say that the courts will decline to address hypotheticals in the course of examining the potential consequences of decisions on properly raised issues.

Police photography of people in public

The New Zealand Supreme Court has given leave to appeal, and, in effect, to cross-appeal, in Tamiefuna v R [2023] NZSC 93 (leave granted). The Court of Appeal decision is of great interest: Tamiefuna v R [2023] NZCA 163. I say no more in public (at least, not before the appeal has been determined, as it is sub judice). Save to say, it is about whether there is a right not to be photographed in public by the police when the photography is not in the course of an active investigation or enforcement need, and about the admissibility consequences in this case if the photography was a breach of the appellant’s right not to be photographed.

The plea of previous conviction

Speaking of sub judice, what has happened to the appeal in Mitchell v Police? The appeal was heard in the Supreme Court on 10 May 2022 (SC116/2021). From the Court’s Report on delivery of judgments for 2022 we can infer that this appeal is not considered to be unreasonably delayed, and that therefore the issues it has to resolve are complex. I have commented on the Court of Appeal decision here, on 23 September 2021. Very exciting.

Update: Mitchell has since been decided, as noted here on 11 August 2023.

How to instruct a jury in a criminal trial, and what is required for there to be a criminal organisation: R v Abdullahi, 2023 SCC 19

Judges who have to prepare directions to juries in criminal trials could be assisted by studying R v Abdullahi, 2023 SCC 19 (14 July 2023). The perspective is that of an appellate court, but the majority judgment of Rowe J with concurrence by Wagner CJ, Karakatsanis, Martin, Kasirer, Jamal and O’Bonsawin JJ provides guidance for trial judges.

The essentials are summarised at [72], with the concept of a functional approach explained at [4], the underlying principles reiterated at [35], the requirement that the jury be properly equipped to decide the case explained at [37] (the jury must be both accurately and sufficiently instructed), the difference between mandatory and contingent instructions (at [48], [49]), and the importance of the context of the trial emphasised at [57].

This appeal also addresses the essential requirements for proof of the existence of a criminal organisation, with case law explaining that such an organisation (in contrast to, for example, a mere conspiracy [80]) to be “organised” must have both structure and continuity (see [76] ff). While the various circumstances of each case require flexible assessment, it is an error (as it was in this case) to take a flexible approach to the requirements of structure and continuity [81]. Improper reasoning can flow from this error ([84] ff). Here, the trial judge had failed to expand upon the statutory definition of criminal organisation by addressing the requirements of structure and continuity [89] and so had failed to properly equip the jury to decide the case.

It was on the issue of whether the judge had properly equipped the jury to decide the issue of whether the facts established the existence of a criminal organisation that Côté J dissented ([99]-[100]. Points of criticism of the majority’s categorisation of errors in jury instructions (eg at [107]) are rather weakened by the reality that this is a sole dissent in a court of final appeal.

Procedural delay: Morgan v R (Jamaica) [2023] UKPC 25 and Yikar v R [2023] NZCA 296

In law, you can’t gain an advantage from your own inexcusable procedural delay. The word inexcusable makes that obvious, but is there just to acknowledge that sometimes an understandable mistake may be made or circumstances may change.

Two recent cases illustrate this. The prosecution was not (ultimately) successful in opposing an application for leave to appeal out of time against sentence in Morgan v R (Jamaica) [2023] UKPC 25 (11 July 2023), and a late application for name suppression was an abuse of process in Yikar v R [2023] NZCA 296 (13 July 2023).

In Morgan, the delay was the fault of officials. It was significant that the proposed appeal apparently had merit, and it was incorrect on the particular facts to suggest that because the sentence had been served the issue was academic (at [69]-[72]). Nor did it assist the prosecution that through official error no record of the original sentencing remarks had been produced. At [73]:

“ … it would offend the basic principles of fairness that failures by the justice system, for which the appellant can bear no responsibility, should amount to a countervailing criterion in the exercise of discretion under the proviso to hear and determine an appeal. The Board considers that the Court of Appeal was in error when it concluded that a countervailing criterion was the justice system’s own failure to produce a record of the proceedings in the Resident Magistrate’s Court.”

Public interest also strongly favoured ventilating the errors in order to maintain public confidence in the justice system ([74]).

The decision whether to permit an out-of-time appeal against sentence was a discretionary matter, requiring factors for and against to be considered ([66]), including finality, merit, whether the issue is academic, the wider public interest, the period of delay and the extent to which the appellant was in default, and here there had been a serious miscarriage of justice. At [76]:

“… the Board sets aside the Court of Appeal’s exercise of its discretion and exercises the discretion afresh. In doing so the Board takes into account the following criteria, namely, (a) the appellant has a meritorious appeal against sentence; (b) the appellant has done everything reasonably possible to file his grounds for appeal with the Clerk of the Courts; (c) none of the delay can be attributed to the appellant; (d) the appeal is not “academic”, see paras 70-72 above; (e) the failure of the judicial system in not producing the Resident Magistrate’s Court record is not a criterion to be taken into account against hearing and determining the appeal, see para 73 above; and (f) the appeal is of general as well as particular significance, see para 74 above.”

I have suggested, in discussing Pierre v R (Bahamas) here on June 16, 2023, that late applications for leave to appeal against conviction should be addressed by focusing mainly on the question of merit. The wider approach, in which more factors are relevant to the exercise of the discretion to grant leave, seems more appropriate for appeals against sentence.

In Yikar, the late application for interim name suppression was, in context, an abuse of process for three reasons, stated at [25]. The factual background is fairly complex, but the Court usefully reviews the abuse of process doctrine at [21]-[24], including references to decisions of the House of Lords, the Privy Council, the Court of Appeal of England and Wales, as well as local decisions. Failing to take an opportunity to make the application, and attempting to gain a collateral advantage were, in the Court’s opinion, significant here.

Retrial after invalid conviction is not double jeopardy: Smith v United States 21-1576, 15 June 2023

It has long been recognised at common law that a trial held in the wrong location does not on that ground result in a verdict that can subsequently be the basis for a plea of previous acquittal or previous conviction: Smith v United States 21-1576, June 15, 2023. That common law dealt in the United States with the venue Clause, and this case deals with breach of the vicinage Clause.

If a trial is held without jurisdiction, there is no lawful determination of criminal culpability. Therefore, a subsequent trial, held lawfully, does not raise issues of double jeopardy. The usual rule is that a retrial is the remedy.

This is not to say that there are no difficult issues concerning whether a trial has been a nullity. There are dangers in making generalisations in the absence of factual contexts, as was observed in AJS v The Queen, noted here on 15 June 2007, and see Bounds v The Queen, noted here on 28 July 2006. Decisions of appellate courts grappling with the consequences of errors are often not unanimous (although Smith was a unanimous judgment, but Blueford v Arkansas, noted here on 26 May 2012, was not). The intention of the legislature in making a rule that has been broken should be determinative of the consequences: DPP v Penn (Virgin Islands), noted here on 13 May 2008. In Smith, breaches of the venue and vicinage clauses of the Constitution clearly pointed to nullity (although the Court did not use that word, instead putting the question as whether such breaches were exceptions to the rule that retrial is the appropriate remedy, and holding that, like the venue clause, the vicinage clause was not such an exception).

It may be possible for a defendant to waive a fundamental error at trial by knowingly acquiescing expressly or tacitly in the continuation of the trial: Ruddy v Procurator Fiscal [2006] UKPC D2 (Scotland), noted here 20 February 2006, while failure by a court official to sign an indictment has been held to result in a nullity: R v Clarke, noted here on 7 February 2008.

In Smith the result was plainly right: the appellant was asking the Court to hold that a conviction entered without jurisdiction could nevertheless support a plea of previous conviction.

Now, if you haven’t already done so, look up vicinage.

The interests of justice and time limits for filing conviction appeals: Pierre v R (Bahamas) [2023] UKPC 15

Legislation that leaves judges to decide an issue “in the interests of justice”, without any indication of what that means, can leave us wondering if there is any law involved in the decision process.

An example is a decision whether to extent the period in which a person may file an appeal against conviction, so as to override a statutory limitation (often 20 working days) on that period.

In the absence of statutory guidance, beyond referring to “the interests of justice”, it is for judges to work out in case law what that means. To the extent that such case law actually assists in the decision process, the decision is a matter of law. Otherwise, it is just a policy decision, at risk of being arbitrary and not law at all, even though made by judges.

The Privy Council has recently addressed this decision: Pierre v R (Bahamas) [2023] UKPC 15 (11 May 2023).

The decision whether to grant an extension of the period within which an appeal against conviction may be filed is a balancing exercise [28]. It is characterised by flexibility [29]. The ultimate question, whether extension is in the interests of justice [26], requires consideration of the overall justice of the case, taking into account the important public interest in the finality of legal proceedings, the efficient use of judicial resources, good administration, the interests of other litigants, the interests of victims and their families, and the interests of witnesses [27]. An extension of time should be granted if the appeal may expose an injustice, and each application will turn on its own facts and circumstances as the discretion is unfettered and extremely wide, and the interests of justice may require departure from legislated time limits [30].

All these matters are nebulous and don’t help to clarify what the threshold is for granting an extension of time to file an appeal against conviction. It is trite and unhelpful to say each case turns on its own facts and circumstances. In adding the various considerations to the interests of justice, the Board seems to have redefined what the interests of justice means, but without setting out that new meaning.

The expression “in the interests of justice” usually means in the interests of correctly deciding an issue according to law. The issue here is the conviction and its correctness in law. All that should matter when deciding whether an appeal should be heard is whether there is a reasonable prospect that the appeal will be successful. There are questions subsidiary to that, such as whether the arguments proposed to be advanced have previously been determined correctly, whether a legal point was overlooked or wrongly decided in the earlier proceedings, and whether any new evidence could cast doubt on the conviction.

It might be useful to look at legislation governing when a criminal cases review commission can refer a conviction to an appeal court. If such a reference is made, the appeal court must hear the appeal. Legislation may state that the governing criterion for the commission’s decision is the interests of justice, and specify some matters that must be taken into account.

An illustration is New Zealand’s Criminal Cases Review Commission Act 2019, s 17. The matters listed in s 17(2)(a)-(c) are helpful, but the addition of (d) “any other matter that the Commission considers relevant”, while serving as a sort of safety net to avoid excluding meritorious references arising from circumstances that can’t be predicted, could potentially have the unfortunate effect of allowing the Commission to refuse to refer a case on grounds that extend the meaning of “the interests of justice” to include the dreadful inertia of the justice system (the so-called interest in the finality of judicial proceedings), an apologetic reference to the inconvenience of diverting resources to the issue, and a simpering solicitousness for the supposed interests of victims and prosecution witnesses.