Informed advice on consequences of plea: changing perceptions of fairness

Sometimes when judges make new rules they apply retrospectively to cases that have already been decided. Obviously this could cause chaos, so there are limitations that confine the situations in which this retrospectivity can apply. In the United States, as we saw here on 21 February 2008, the rule in Teague v Lane, 489 U.S. 228 (1989) recognises two exceptions to the general rule that decisions do not apply retrospectively: where laws proscribing conduct have been declared unconstitutional, and where new rules concern the fairness of trials. Teague has been applied recently in Chaidez v United States
USSC No 11-820, 20 February 2013.

In Chaidez the appellant sought application to her case of a rule that had been declared by the Supreme Court after the proceedings in her trial had been finalised. The rule is that counsel must advise a non-citizen client of the immigration status implications of a plea of guilty: Padilla v Kentucky, 559 U.S. 356 (2010). Ms Chaidez had not been advised that her permanent residency would be revoked and that she faced mandatory removal as a result of her guilty pleas and convictions for mail fraud charges that were classified as aggravated felonies.

The Court held that the Padilla rule did not apply retrospectively.

This rather surprising result arose from what the majority regarded as the first issue which was whether advice about deportation came within the Sixth Amendment right to counsel. Does the right to counsel include the right to advice about a collateral matter – the consequences of a conviction on immigration status? Before Padilla the states had almost unanimously held that it does not. Padilla created a new rule for most jurisdictions. This new rule did not apply to Ms Chaidez's case because (the opinion of the Court is not clear on this but one can infer the reason) it did not, in terms of Teague, concern the fairness of the criminal proceedings against her.

Sotomayor J, joined by Ginsburg J, dissented. They reasoned that Padilla did not create a new rule but it merely applied the established rule that legal representation must be to a standard consistent with prevailing professional norms (Strickland v Washington, 466 U.S. 668 (1984)). Those norms had developed to require advice about the immigration implications of a guilty plea. Padilla did not apply a direct-or-collateral consequence approach to the Sixth Amendment, and Ms Chaidez's legal representation had fallen short of the required standard.

Outside the United States a retrospectivity issue of the kind that arose in Chaidez would probably be resolved by asking whether the defendant should have been advised of the consequences of conviction, including deportation. The rules of professional conduct should indicate the extent of a lawyer's obligation to advise a client of the consequences of a decision. For example, the Lawyers and Conveyancers Act (Lawyers: Conduct and Client Care) Rules 2008 [NZ], rule 13.13.1:

"When taking instructions from a client, including instructions on a plea and whether or not to give evidence, a defence lawyer must ensure that his or her client is fully informed on all relevant implications of his or her decision and the defence lawyer must then act in accordance with the client's instructions."

A retrospectivity issue would arise if standards had changed so that although advice on immigration status might not reasonably have been required to have been given at the time the client entered a guilty plea, it would be required now. This is an artificial issue here, as it has long been recognised that such advice must be given. But if there had been a change in this, so that "all relevant implications" now include the effect of a conviction on immigration status, would all defendants who had previously entered guilty pleas without that advice now be able to have their convictions quashed and their pleas re-taken?

While it is true that standards of fairness change over time, defendants who are still around to complain of treatment that by current standards was unfair are likely to have a remedy. See Krishna v The State (Trinidad and Tobago) [2011] UKPC 18, discussed here on 12 July 2011.

Jeopardy from silly acquittals?

"To permit a second trial after an acquittal, however mistaken the acquittal may have been, would present an unacceptably high risk that the Government, with its vastly superior resources, might wear down the defendant so that 'even though innocent he may be found guilty"

This from Green v United States, 355 U.S. 184, 188 (1957) was quoted in Evans v Michigan
USSC No 11-1327, 20 February 2013. Here an acquittal had been directed at trial although on grounds that were plainly wrong. The acquittal was nevertheless valid and effective to invoke the rule against double jeopardy.

Evans highlights the distinction between procedural terminations of trials and findings on the merits. A procedural termination is not an acquittal because it is not a finding on the merits. A procedural termination avoids the defendant being in jeopardy of a conviction. A properly granted mistrial does not give rise to an expectation of finality.

In Evans the judge ruled at the close of the prosecution case that the evidence so far adduced was insufficient to convict the defendant. The judge wrongly thought that the relevant law of arson required proof that a particular building was not a dwelling, and directed a verdict of not guilty accordingly. This was held to have been a ruling on the merits and a bar to a retrial.

You can see that this is a bit silly. We saw an analogous sort of silliness in R v Taylor [2008] NZCA 558, discussed here on 18 December 2008. Taylor was applied in Connolly v R [2010] NZCA 129 at [65]-[66].

It is silly because it turns on the stage of a trial at which an objection is taken: has jeopardy of conviction arisen? If the prosecution case has been closed, then yes, there is jeopardy, as in Evans. But what if, after the jury had been selected and the trial begun, the defence had objected that on the basis of the evidence that the prosecutor intended to call there was no proof of an element of the offence, and before any evidence was called the judge ruled that the prosecutor needed that evidence, and the prosecutor confirmed it would not be adduced, and the judge then directed an acquittal? Would jeopardy have arisen? Is this a procedural termination? The position as it would inevitably be if the prosecution case proceeded to closure is being anticipated. Why wait until closure of the case? Yet the authorities suggest that this would be a procedural termination, not a decision on the merits.

Is there a sensible policy behind treating directions that are obviously wrong in law as nevertheless giving rise to legally effective acquittals? If there is, is it consistent with whatever policy might justify distinguishing between procedural dismissals and rulings on the merits?

Sniffer dog reliability and probable cause for search

There were times when Aldo the drug dog was not on his best form. He was trained to detect the presence of particular drugs, but sometimes he would say one or more of them was there when that was wrong, although on the occasion that concerned the United States Supreme Court there were other drugs present: Florida v Harris, USSC No 11-817, 19 February 2013.

What was the relevance of a record of false positives to the existence of probable cause for a search based on Aldo's indication?

Probable cause requires the kind of "fair probability" on which "reasonable and prudent [people,] not legal technicians, act": Illinois v Gates, 462 U.S. 213. The totality of the circumstances must be looked at, and rigid rules, bright line tests and mechanistic enquiries are inappropriate, according to Gates.

The Court, in a unanimous opinion delivered by Kagan J, considered that field-performance records may not capture a dog's false negatives or may markedly overstate a dog's false positives. This is because in the field – that is, in real-world performance – a dog's failure to detect drugs will often not be followed by a search and so the failure will go unrecorded. Similarly, an apparent false positive may in reality be a correct response to a residual odour or a small amount of drug that the officer failed to find. But evidence of the dog's reliability in a controlled setting, where errors are known, can provide sufficient reason to trust the dog's alert in the field.

All the evidence needs to be evaluated, said the Court. Here the facts supported a finding of probable cause because the defence had not contested the State's evidence that Aldo performed reliably in controlled settings. But where the defence did challenge the evidence of the dog's reliability the court should assess all the evidence and ask whether all the facts, viewed through the lens of common sense, would make a reasonably prudent person think evidence of offending would be found. "A sniff is up to snuff when it meets that test." Haw haw.

Here the problem with the Florida Supreme Court's approach was that:

"No matter how much other proof the State offers of the dog's reliability, the absent field performance records will preclude a finding of probable cause. That is the antithesis of a totality-of-the-circumstances analysis."

There may be other indicia of reliability, and absence of one sort may be compensated for by presence of another.

Because the defendant had not raised issues of Aldo's reliability at trial, he could not do so for the first time on this appeal.

"As the case came to the trial court, Aldo had successfully completed two recent drug-detection courses and maintained his proficiency through weekly training exercises. Viewed alone, that training record—with or without the prior certification—sufficed to establish Aldo's reliability."

Good doggie.

Secondary participation in conspiracy

For a compellingly logical analysis of secondary participation in conspiracy, see R v JF, 2013 SCC 12 (1 March 2013).

Although the form of secondary liability considered in JF is aiding or abetting under s 21 of the Criminal Code, the same reasoning would apply to the other forms of participation that are usually included as constituting secondary participation: inciting, counselling or procuring.

The focus is on the actus reus of conspiracy, which is an agreement to commit an offence. The actus reus for secondary liability is aiding, abetting, inciting, counselling or procuring the formation of the agreement. Because the agreement may continue up to the attainment of its unlawful object, a person may join the agreement after its initial formation. A person who aids, abets, counsels or procures that person to join at that later stage commits the actus reus of secondary participation in the conspiracy.

A person who aids (etc) a conspirator in the attaining of the object of the conspiracy is not for that fact alone a conspirator, because the attaining of the object of the conspiracy is not the actus reus of conspiracy. But that evidence will usually support an inference of assistance in the entering into of the agreement by the person who is aided (etc), and if that inference is permissible in the circumstances then it is evidence of the actus reus of the aider's (etc) secondary participation.

This decision removes doubts that had existed in Canada over whether liability could arise from the mere assistance of a conspirator in the furtherance of the conspiracy. The approach in R v McNamara (No 1) (1981) 56 CCC(2d) 193 (Ontario CA) was not followed.

Old problems with new evidence

Whether failure to call evidence at trial should be grounds for quashing a conviction is an issue that often has to be decided in appeals, as it was in Taylor v The Queen (Jamaica) [2013] UKPC 8 (14 March 2013).

Here the Board split, Lord Kerr dissenting, on the application of the law to the facts. There was no difference of opinion on what the law is.

It was unnecessary to decide whether counsel were at fault in not calling the relevant witness, because the issue is the impact of the absence of the evidence on the verdict: Lord Hope for the majority at [13], applying Lord Carswell in Teeluck v State of Trinidad and Tobago [2005] UKPC 14 at [39], (discussed here on 1 April 2005). The test is whether, after taking all the circumstances of the trial into account, there is a real possibility of a different outcome: Taylor at [20]. Might the evidence reasonably have affected the jury's decision to convict? The majority held that there was no reasonable possibility in this case that the jury would have arrived at a different verdict.

Lord Kerr said that the requirement of a real possibility of a different verdict signifies no more than an acceptance that one is left in doubt as to the safety of the conviction [39]. This, he added (more obscurely), was not a matter for one side or the other to take on the burden of showing, but instead is a matter that must be decided in the round. The question of the effect of the new evidence on the safety of the verdict is for the appellate court to decide for itself; the court does not decide what effect the evidence might have had on the jury: [43], applying Dial v State of Trinidad and Tobago [2005] UKPC 4 (noted here on 17 February 2005). Further, the question is not whether there was evidence on which the jury might reasonably convict, but rather whether there was evidence on which it might reasonably decline to do so: [45], applying Bain v The Queen (New Zealand) [2007] UKPC 33 (discussed here on 11 May 2007).

Therefore it is in this sense, said Lord Kerr, that the question, where new evidence has come to light, is whether the evidence might reasonably have led to an acquittal [46]. He proceeded to analyse the case and reached a different conclusion to that of the majority.

The application of this law is notoriously difficult for appellate courts, and it is not unusual to find dissenting judgments. One also finds dissents in appeals which have to consider, not the failure to call evidence at trial, but the wrongful adducing of evidence at trial. A spectacular example is Howse v R [2005] UKPC 31, discussed here on 23 July 2005. The wrongful adducing of evidence raises the substantive question of the fairness of the trial, and the inevitability of a guilty verdict is not the touchstone, for even a guilty defendant is entitled to a fair trial.

Trial fairness is also raised by a failure to adduce evidence and we might wonder whether the criterion for quashing a conviction in such a case should be whether the absence of the evidence might have undermined the fact-finder's impartiality: could inappropriate weight have been given to critical prosecution evidence because of the absence of the evidence?

Attribution and the actus reus

Sometimes where there is more than one defendant alleged to be a principal offender (that is, one who personally does the actus reus with the required mens rea) it is difficult to say who did what. This is particularly so where there are a lot of eye witnesses, for it is normal to find that everyone has a different account of what happened. There may be no doubt that an actus reus occurred, but who did what?

This was considered in Huynh v The Queen [2013] HCA 6 (13 March 2013). "Joint enterprise" liability of this kind (which is not to be confused with the term joint enterprise in the context of an exception to the rule against hearsay, nor with similar terminology that is sometimes used in relation to extended secondary liability) depends on proof that the defendant was a party to an agreement to commit the offence that was committed and in addition - for this is not mere conspiracy - that the defendant participated in the offending. Presence when the offending occurred can be sufficient to prove participation. The circumstances may also, as they were in this case, be such that the defendant's presence supports an inference that it was pursuant to the required agreement.

This common law solution to the determination of liability as a principal offender where the facts are vague has a parallel in the interpretation of the statutory criterion, for liability as a principal, of being a person "who actually commits the offence": see for example s 66(1)(a) of the Crimes Act 1961 [NZ]. There is an appeal case, the name and details of which seem still to be subject to a suppression of publication order, so I give only the citation for people with access to databases on which publication is permitted: "[2010] NZCA 256". In that case there were several steps needed to complete the actus reus and several defendants were allegedly involved, but no one person did all the steps. In relation to any given defendant, the prosecutor had to prove that that defendant did any of the required steps, that one or more other of the defendants did the other necessary steps, and that the defendant under consideration had mens rea. This was called a distributive application of s 66(1)(a).

Is the distributive approach the same as the joint enterprise liability illustrated by Huynh? Probably yes, although particular facts may produce different results. That would depend on what inference could be drawn from a defendant's presence when another defendant completed a step in the actus reus. The common element is participation in the offending in circumstances where the acts of one defendant are attributed to another.

When will this attribution occur? Policy considerations strongly favour public safety, but there should be clear limits on attribution. For a controversial approach see R v Gnango [2011] UKSC 59, [2012] 1 AC 827, discussed here on 18 December 2011, where the majority favoured liability as a principal for murder where the defendant was in a gunfight with the person who, missing him, shot a bystander.

Guidance was given to judges in both Huynh and "[2010] NZCA 256" on how to handle complex cases, where there are multiple defendants and different forms of alleged liability. Huynh stresses the importance of getting to the real issue in the case, so that the judge can give a focused direction on the law. The New Zealand case at [26] invites judges, when faced with disagreement from counsel on the structure and content of a proposed question trail for the jury, to make a minute recording the nature of the dispute and the judge's reasons for adopting the trail ultimately used.

Essential reading for trial lawyers

Thanks to Peter Tillers for bringing attention to a paper on DNA evidence that is essential reading for trial lawyers, particularly defence counsel: William C Thompson, Laurence D Mueller, and Dan E Krane, "Forensic DNA Statistics: Still Controversial in Some Cases".

Amending the Criminal Procedure Act 2011 [NZ]

Some tweaking of the Criminal Procedure Act 2011 has been found to be necessary. My submissions on the Criminal Procedure Legislation Bill are here.

Contempt, imprisonment and appeals

A few interesting remarks on imprisonment for contempt and on sentence appeals generally were made in B (Algeria) v Secretary of State for the Home Department [2013] UKSC 4 (30 January 2013).

When dealing with a contemnor and considering the option of committal to prison, the court needs to assess two things, coercive and penal:

"[14] ... Committal is appropriate where it can reasonably be expected that this will induce the contemnor to purge his contempt (the coercive effect). It is also appropriate to punish contempt of a court's order (the penal element). Frequently both elements will underlie a committal order. Where there is reason to believe that committal will secure compliance with a court's order, the fact that the person subject to it has already substantial restrictions on his liberty is immaterial. Where it is required in order to properly punish the contemnor, the loss of residual liberty is unlikely to weigh heavily against the making of the order."

The contemnor must be punished for misbehaviour and also must be induced to behave properly.

The other topic considered here is when an appellate court should correct a sentence itself instead of remitting the case to the lower court for it to impose an appropriate sentence:

"[11] Where an appellate court has concluded that the basis on which the decision of the lower court to sentence someone for contempt is flawed, it does not follow that the sentence chosen by the lower court is inevitably wrong. It may be an entirely correct sentence but for different reasons from those articulated by the original sentencing court. The affirmation of the original sentence does not necessarily entail an endorsement of the reasons for which the decision to sentence was made. Where it has been determined that the basis for the original sentence of imprisonment is wrong, a de novo assessment must indeed occur. A fresh look at the circumstances material to the question of whether imprisonment is the right disposal should take place in light of the correct understanding of those circumstances.

"[12] It is not essential, however, even as a matter of generality, that the fresh look be undertaken by the original sentencing court. If it is sufficiently clear to the appellate court that a sentence of imprisonment is appropriate in light of its revised view of the relevant facts, it is not required as a matter of principle or of practice that the matter be remitted to the court which first imposed the sentence. As Jackson LJ said in JSC BTA Bank v Solodchenko (No 2) [2012] 1 WLR 350 para 60, where an appellate court is seised of the case and in possession of all relevant facts, the proper course is for the appellate court to determine what the proper sentence for contempt should be on the basis of the true facts. Where, of course, a fresh investigation of new facts is required and it is necessary or desirable that this be undertaken by a first instance court, remittal will be suitable. This is not such a case. The Court of Appeal was able to evaluate the medical evidence and reach reliable conclusions on its significance. It could decide what the appropriate sentence should be and it was right to do so."

And on the topic of contempt of court, there is a particularly interesting discussion of the common law powers by Professor ATH Smith, "Reforming the New Zealand Law of Contempt of Court – An Issues/Discussion Paper", 18 April 2011, and also in Siemer v Solicitor-General [2010] NZSC 54, [2010] 3 NZLR 767, (2010) 24 CRNZ 748 (SC).

Withdrawal from offending at common law – evidential and legal burdens

The burden and standard of proof of the affirmative defence of limitation, and some aspects of the affirmative defence of withdrawal, were considered in Smith v United States, USSC No 11-8976, 9 January 2013.

Subject to any statutory definition to the contrary, an affirmative defence is one that affirms that even though the actus reus and the mens rea may have been proved, nevertheless the defendant is not guilty.

Examples of affirmative defences are infancy, insanity, self-defence, coercion, and limitation.
This last one, limitation, amounts to the defendant saying, "Yes, I committed the offence, but that was so long ago that the law prevents me being charged." Not all offences are limited in this way, and limitation periods are established by legislation. Terminology can be a bit confusing, so when I say the offence was inside the limitation period that means the defendant can be prosecuted, and when I say it was outside the limitation period that means the defendant cannot be prosecuted.

In Smith the relevant offences were conspiracies to commit various serious offences. A limitation period applied. The defendant relied on the limitation period, saying his participation in the conspiracies had ceased outside that period because he had withdrawn his participation. At trial the issue of withdrawal only arose as a result of a question from the jury, and the judge gave a direction that the defendant had to prove withdrawal to the standard of the balance of probabilities. The Supreme Court held that this did not violate the Due Process clause.

Placing the burden of proof of withdrawal on the defendant did not violate the right to be presumed innocent because, as the Court had said in Patterson v New York, 432 U.S. 197 (1977),

"This was the rule when the Fifth Amendment was adopted, and it was the American rule when the Fourteenth Amendment was ratified. Commonwealth v. York, 50 Mass. 93 (1845)."

Of course legislatures could change that rule, as indeed most have, so that it is now usual for defences, whether they be affirmative or not, to be for the defendant to raise but for the prosecutor to disprove. The defendant under this usual approach has only the "evidential burden" of pointing to some evidence sufficient to put the existence of the defence into issue.

But in the eighteenth century, as Blackstone described in his Commentaries, the defendant had to prove the defence. In Smith the Court, in its unanimous opinion delivered by Scalia J (who did not on this occasion have the opportunity to indulge his appetite for the purple prose of vigorous dissent), pointed out that the burden on the defence to establish withdrawal on the balance of probabilities was justified because it was the defendant who would have the relevant evidence on the issue. The prosecutor could not be expected to prove a negative. The defence here did not seek to negative any element of the offence that the prosecutor had to prove.

To negative his participation in the conspiracies the defendant had to prove that he took steps "to dissociate from his confederates". The conspiracy was a continuing offence, and the prosecutor had proved that it continued inside the limitation period, but the prosecutor could not be expected to prove that the defendant did not withdraw before then. Participation in the conspiracy continues even if the defendant is inactive after joining it, until he withdraws or until the conspiracy terminates.

I should emphasise that the law is not necessarily the same outside the United States. The common law has developed since the eighteenth century and legislation usually reflects those developments: a defence carries an evidential burden for the defendant and then a legal burden of disproof on the prosecutor. The usual exception is insanity, where the legal burden is on the defendant to the standard of the balance of probabilities. But where the defence of withdrawal is still a common law defence the policy of placing the legal burden on the defendant, on the issue of the taking of sufficient steps to withdraw, may need some consideration.

Withdrawal cannot be effective if the offending cannot be undone. A conspiracy is complete upon agreement, and it continues while that agreement exists even if the parties to it change. Once the defendant has entered the conspiracy it is too late for him to withdraw. And in relation to attempts, it is too late to withdraw once a sufficiently proximate act has been committed. And incitement is committed when the incitement occurs – it is too late to withdraw after then. But liability can arise from assistance or encouragement, and it may be possible for the defendant to withdraw his assistance and encouragement. Liability can also arise from participation in a common purpose, and a defendant may be able to withdraw by negativing his participation. More than verbal withdrawal may be required, for example where the defendant has provided others with equipment needed to commit an offence. Even verbal withdrawal may require more than mere dissociation, and the defendant may have to try to persuade the principal offender not to continue.

In Ngawaka v R [2004] NZCA 249 (6 October 2004) the Court of Appeal approved a statement of the law on withdrawal by Hammond J in R v Pink [2001] 2 NZLR 860:

"As a matter of legal doctrine, it seems to me that the following conditions must be met:

• First, there must in fact be a notice of withdrawal, whether by words or actions.

• Secondly, that withdrawal must be unequivocal.

• Thirdly, that withdrawal must be communicated to the principal offenders. There is some debate as to whether the communication must be to all the principal offenders, but here all were told.

• Fourthly, the withdrawal may only be effected by taking all reasonable steps to undo the effect of the party's previous actions. (See R v Menniti [1985] 1 Qd R 520.) As with any test of "reasonableness", it is impossible to divorce that consideration from the facts of a given case. The accused's actions may have been so overt and influential that positive steps must be taken by him to intercede, and prevent the crime occurring. There is at least one authority (R v Grundy [1977] Crim LR 534 (CA)) which suggests that where the accused's participation was in the form of counselling, attempts by the accused to dissuade the principal offenders from proceeding with the crime are sufficient."

Hammond J continued:

"[15] That is the law, as I understand it, in all the British Commonwealth jurisdictions. It follows that because the onus is on the Crown, where "withdrawal" is raised by a party the onus is on the Crown to negative any such "defence". The difficulty which has arisen in the cases appears to be as to the precise conditions of withdrawal which have to be in place for this doctrine to apply.

[16] A classic statement of the defence is that by Plowden in his commentary on R v Saunders and Archer (1576) 2 Plowd 473, 476; (1576) 75 ER 706, 710:

"If I command one to kill JS and before the fact done I go to him and tell him that I have repented, and expressly charge him not to kill JS and he afterwards kills him, there I shall not be accessory to this murder, because I have countermanded my first command, which in all reason shall discharge me, for the malicious mind of the accessory ought to continue to do ill until the time of the act done, or else he shall not be charged; but if he had killed JS before the time of my discharge or countermand given, I should have been accessory to the death, notwithstanding my private repentance."

[17] A more modern statement of the doctrine is to be found in the judgment of Slane JA in the Canadian appellate decision of R v Whitehouse [1941] 1 DLR 683, 685: the doctrine of withdrawal only obtains on the footing that, "where practicable and reasonable, there must be timely communication of the intention to abandon the common purpose".

[18] That decision gave rise to some debate as to what is meant by "timely" and "effective"? What, for instance, is the position to be where timely communication is not practicable? Some jurists suggested that withdrawal by means of countermand would not then be available to the accomplice at all. On that point, the High Court of Australia in White v Ridley (1978) 52 ALJR 724 has held that the withdrawal must be sufficiently timely to be capable of being effective.

[19] That there is a doctrine of withdrawal in New Zealand law is at least implicitly accepted by the Court of Appeal in a decision cited to me by Mr Ellis – R v Wilcox [1982] 1 NZLR 191. In that case, at p.196, Sir Owen Woodhouse accepted that there was a material misdirection on the part of the trial Judge in that case, insofar as the jury had been left under the misapprehension "that any subsequent change of mind following the first step taken (the purchase of the weapons) could not be used by Wilcox as a defence".

[20] It is not possible in a trial ruling of this kind to consider at length, even if it were appropriate, the very considerable academic literature and interest which has been generated on this difficult topic. The literature includes Lanham "Accomplices and Withdrawal" (1981) 97 LQR 575; Smith, "Withdrawal from Criminal Liability for Complicity and Inchoate Offences" (1984) 12 Anglo-American Law Review 200; O'Regan, "Complicity and the Defence of Timely Counterman or Withdrawal Under the Griffith Code" (1986) 10 Criminal Law Journal 236; Marcus, "Joint Criminal Participation: Establishing Responsibility, Abandonment" (1986) 34 American Journal of Comparative Law 479. See also in New Zealand, Simester and Brookbanks, Principles of Criminal Law (1998) at para 5.1.4.4, and R v Malcolm [1951] NZLR 470 (CA).

[21] Without attempting to resolve the jurisprudential debate as to the basis of this "defence" (if a defence it be, properly so called) it seems to me that, as a matter of public policy, there should be a plea of this kind which is open to an accused. It is surely in the public interest that somebody who has contemplated criminal endeavour and changed their mind, should be able to do so. At the same time, attempting to withdraw from a crime about to happen may not in itself be sufficient, since the accomplice's prior act may have some very distinct impact on what in fact occurs. It may be that the law should look for some kind of abrogation of the influence of that act."
 

The question of who should have to prove this abrogation is also a policy matter. Hammond J has, naturally, assumed at [15] that where withdrawal is raised it is for the prosecutor to negative it, and this was applied in Pink in reaching the conclusion that the prosecutor could not establish beyond reasonable doubt that the defendant had not withdrawn. But, especially where there are technical questions about whether withdrawal is really a defence, it is not inevitable that the common law must put the legal burden of negativing it on the prosecutor.

Update: the New Zealand Supreme Court has held that withdrawal is a common law defence and that Hammond J's points in Pink need modification: Ahsin v R [2014] NZSC 153, especially at [134] per McGrath, Glazebrook and Tipping JJ.