Friday, July 28, 2006

Differences in principle and pragmatism

Appellate judges often disagree over whether an error at trial caused a substantial miscarriage of justice. Finding himself in a minority of one in Bounds v R [2006] HCA 39 (20 July 2006), Kirby J observed, para 85:

"Conclusion: a basic error: I accept that different conclusions are available on this issue. The nature of the controversy and the breadth of the language of the 'proviso' virtually assures the existence of differences of judicial views. Such differences may reflect the diverse values that judges accord to considerations of principle and pragmatism, as they regard them. Those differences appear in many cases in this Court concerned with the 'proviso' … [citing Green v The Queen (1971) 126 CLR 28 at 31; Jones v The Queen (1997) 191 CLR 439; Darkan v The Queen [2006] HCA 34]. They are also reflected in the foregoing differences of opinion in the United States Supreme Court [ie Lane v US 747 US 438 (1986)]."

It seems likely that, on the key point in Bounds, the majority were correct. A second count had been wrongly included in an indictment. The error was that the second offence was purely summary, not triable on indictment. Nevertheless, evidence of its commission would probably have been admissible as similar fact evidence in respect of the only count that should have been in the indictment. The majority (Gleeson CJ, Hayne, Callinan and Crennan JJ) at para 26 held

"… The evidence admitted at the appellant's trial about the downloading of the images the subject of count 2 was very limited. It would have been admissible on the trial of an indictment alleging only count 1."

Kirby J did not think this was so (para 100):

"Nor do I consider that the tender of such material would have been permitted in a trial if that trial had been limited to the indictable offence of possession of child pornography. A judge guarding the fairness of the conduct of such a trial would be properly careful to restrict extraneous, and possibly prejudicial, evidence. By impermissibly charging the two offences in the one indictment, an inter-mixture necessarily occurred. Descriptions or conceptions of the contents of the images became inevitable. It is that inter-mixture, before the jury, that presents the risk of a substantial miscarriage of justice. It is that risk that withholds the application of the 'proviso'."

Apart from the difference of opinion on that issue, Kirby J also differed from the majority on whether the jurisdictional error was, of itself, sufficient to prevent application of the proviso without examination of issues relating to the conduct of the trial. He reasoned (paras 87-88) that the Western Australia legislature had carefully separated the jurisdictions and that the community and the accused are therefore entitled to a trial that conforms to the law and is without jurisdictional flaw.

The majority, however, did focus on what happened at the trial of the count that was properly before the jury (para 13-30). Their conclusion on the trial conduct contrasts with that of Kirby J, who held, para 96:

"… The impermissible inclusion of the second count in the indictment deprived the appellant of a trial that was free from any reference to this extraneous factor [images of bestiality]. It deprived him of the chance of avoiding this added complication in his trial. Specifically, it deprived him of the forensic choices that the separate trial of the offences referred to in the two counts would have entailed."

It is not clear what those "forensic choices" were, even on the basis that the evidence was not (as Kirby J held) admissible as similar fact. Even the mere fact that these images (whatever they might have depicted) were downloaded at a particular time seems to have been relevant to whether the images which were the subject of count 1 were downloaded by the accused (see majority judgment, para 26).

This case is interesting for its references to other cases, including English and United States decisions, where differences of opinion have occurred over how to decide whether an error has caused a trial to be unfair. It is now 99 years since England enacted what has been almost universally adopted as the "proviso", and one would have thought that, by now, appellate judges would have worked out a rational approach to its application, thereby promoting uniformity of result and minimising disagreements. My own view is that the basic problem is that judges have not got to grips with the meaning of the essential requirement of a fair trial. Until judges learn to define the characteristics of trial fairness and to apply them rationally, vagueness over the meaning of "substantial miscarriage of justice" will continue to make the results of appeals difficult to predict.

If one were to apply my own "Fair Trial Analysis", one would identify the following issues in Bounds. Since a fair trial is one where the law is accurately applied to facts determined without bias, the first question is, what errors of law occurred? It was common ground that count 2 should not have been in the indictment. Was the admission of evidence relevant to that count an error? This requires examination of the admissibility of the evidence as similar fact in relation to count 1. The judges in Bounds did not undertake this inquiry in detail, they simply announced their conclusions, and, as noted, Kirby J dissented on this. If it is correct that the evidence was admissible as similar fact, then the question is whether the judge’s direction to the jury on that topic was adequate to ensure that the evidence was not misused. If the evidence was not admissible as similar fact, then the question is whether the jury were adequately warned to ignore it in relation to count 1. The jury direction was not revealed in the judgments in Bounds. Then, in relation to the unbiased determination of the facts, the appellate court should ask whether the similar fact evidence – even if it was admissible – could have, by reason of its content, prejudiced the rational assessment of the other evidence. This point is addressed in Bounds, but again it was a point of difference. Kirby J thought that it could have caused prejudice, simply because of the risk of an averse reaction by any juror to the mere mention of bestiality. The majority, however, considered that the mention of bestiality (the jury were not shown any images) was of peripheral concern and would not have caused illegitimate prejudice. At the heart of this difference of judicial opinion is the extent to which it should be acknowledged that jurors may not approach their task rationally. This was mentioned as recently as the 19th of this month by the House of Lords in R v Coutts (blogged 21 July 2006), where Lord Bingham (with whom the others agreed), citing a High Court of Australia case, Gilbert v R (2000), said at para 20 that the jury room may not be a place of undeviating intellectual and logical rigour. Coutts and Bounds are cases that complement each other: in Coutts the concern was failure to include an offence in the indictment, whereas in Bounds the concern was improper inclusion of a count. Should the court be less sensitive to the risk of jury prejudice in one than in the other?

Friday, July 21, 2006

Fairer than you may want

The right to a fair trial is a right that the accused cannot waive.

This may seem a strange statement: why would the accused ever want to waive the right to a fair trial? Well, trial tactics are complicated things. If charged with one serious offence, the accused may prefer to hope for an outright acquittal on that rather than have the judge tell the jury that if they acquit on the serious charge they may then consider whether to convict on a lesser charge. A conviction on the lesser charge, and the consequence of imprisonment on that, may be something the accused would see as rendering his victory on the more serious charge Pyrrhic.

A fair trial is a trial conducted according to law. This means, as the House of Lords held this week in R v Coutts [2006] UKHL 39 (19 July 2006), that the jury must be fully informed about the alternatives available to it, whether either side want that or not. In Coutts, both sides agreed that the judge should not instruct the jury about the alternative of manslaughter, when the accused faced a solitary charge of murder. The Crown, confident that it would secure a conviction for murder, did not want the jury to instead convict for manslaughter, so, clothing its stance in the guise of a fairness objection, it argued that the accused was entitled to an acquittal if the jury had a doubt about the way the Crown had sought to prove its case for murder. The defence, having argued that the killing was accidental, did not want the risk of a conviction for manslaughter, because the sentence for that would, in the circumstances, be a lengthy term of imprisonment. Following the conviction for murder, the defence appealed, arguing that the judge should have directed the jury about manslaughter. The House of Lords upheld this argument and quashed the conviction, remitting the case so that a retrial could be considered.

Coutts establishes that fairness requires that the jury be fully informed about the law and the alternatives open for consideration in the particular case. Further, failure to inform the jury of the legal position amounts to a substantial miscarriage of justice. The appellate court does not enter into an inquiry about whether the jury only convicted the accused of the serious charge because it did not want him to walk free. To do that would be to engage in speculation. While a foundation of the system of trial by jury is the assumption that juries apply the directions on the law that judges give them (Lord Rodger, para 87), it is proper to recognise that the jury may be affected in its approach by the choices that it perceives as being available to it (an observation by Callinan J in the High Court of Australia case Gilbert v The Queen (2000) 201 CLR 414 para 101, quoted by Lord Bingham at para 20, Lord Hutton at 54, Lord Rodger at 88 and Lord Mance at 99).

Accordingly, the public interest in the administration of justice is best served if the judge leaves any obvious alternative offence to the jury irrespective of the wishes of counsel. Lord Bingham, with whom the other law lords agreed, put the requirement as follows (para 23):

"The public interest in the administration of justice is, in my opinion, best served if in any trial on indictment the trial judge leaves to the jury, subject to any appropriate caution or warning, but irrespective of the wishes of trial counsel, any obvious alternative offence which there is evidence to support. I would not extend the rule to summary proceedings since, for all their potential importance to individuals, they do not engage the public interest to the same degree. I would also confine the rule to alternative verdicts obviously raised by the evidence: by that I refer to alternatives which should suggest themselves to the mind of any ordinarily knowledgeable and alert criminal judge, excluding alternatives which ingenious counsel may identify through diligent research after the trial. Application of this rule may in some cases benefit the defendant, protecting him against an excessive conviction. In other cases it may benefit the public, by providing for the conviction of a lawbreaker who deserves punishment. A defendant may, quite reasonably from his point of view, choose to roll the dice. But the interests of society should not depend on such a contingency.

"24. It is of course fundamental that the duty to leave lesser verdicts to the jury should not be exercised so as to infringe a defendant's right to a fair trial. This might be so if it were shown that decisions were made at trial which would not have been made had the possibility of such a verdict been envisaged. But no such infringement has ordinarily been found where there is evidence of provocation not relied on by the defence, nor will it ordinarily be unfair to leave an alternative where a defendant who, resisting conviction of a more serious offence, succeeds in throwing doubt on an ingredient of that offence and is as a result convicted of a lesser offence lacking that ingredient. There may be unfairness if the jury first learn of the alternative from the judge's summing-up, when counsel have not had the opportunity to address it in their closing speeches. But that risk is met if the proposed direction is indicated to counsel at some stage before they make their closing speeches. They can continue to discount the alternative in their closing speeches, but they can address the jury with knowledge of what the judge will direct. Had this course been followed in the present case there would have been no unfairness to the appellant, and while taking a contrary view the Court of Appeal did not identify the unfairness which it held would arise. It is not unfair to deprive a defendant, timeously alerted to the possibility, of what may be an adventitious acquittal."

Thursday, July 20, 2006

Public privacy

Courts may disagree over whether language is “grossly” offensive. In DPP v Collins [2006] UKHL 40 (19 July 2006) the House of Lords unanimously disagreed with two lower courts on this point. The defendant had left messages by telephone at his MP’s office, employing terms of racial abuse that were unquestionably offensive. The lower courts held that the messages did not cross the line between offensive and grossly offensive. The embarrassment of disagreeing with the lower courts was exacerbated because essentially the question whether the messages were “grossly” offensive was one of fact, to be determined by applying the standards of an open and just multi-racial society taking into account the context and all relevant circumstances (Lord Bingham, para 9) – standards which the lower courts should be particularly well placed to apply.

How had the lower courts gone wrong? Lord Bingham at para 13 concluded:

“Differing from the courts below with reluctance, but ultimately without hesitation, I conclude that the respondent's messages were grossly offensive and would be found by a reasonable person to be so.”

He agreed with Lord Carswell, who was a little more explicit about this (para 22):

“I felt quite considerable doubt during the argument of this appeal whether the House would be justified in reversing the decision of the magistrates' court that the reasonable person would not find the terms of the messages to be grossly offensive, bearing in mind that the principle to which I have referred, that a tribunal of fact must be left to exercise its judgment on such matters without undue interference. Two factors have, however, persuaded me that your Lordships would be right to reverse its decision. First, it appears that the justices may have placed some weight on the reaction of the actual listeners to the messages, rather than considering the reactions of reasonable members of society in general. Secondly, it was conceded by the respondent's counsel in the Divisional Court that a member of a relevant ethnic minority who heard the messages would have found them grossly offensive. If one accepts the correctness of that concession, as I believe one should, then one cannot easily escape the conclusion that the messages would be regarded as grossly offensive by reasonable persons in general, judged by the standards of an open and just multiracial society. The terms used were opprobrious and insulting, and not accidentally so. I am satisfied that reasonable citizens, not only members of the ethnic minorities referred to by the terms, would find them grossly offensive. "

The offence in question was held to require mens rea – an intention that the words be grossly offensive to those to whom they related, or an awareness that they may be taken to be so (Lord Bingham, para 11). But on the approach taken here, that state of mind need not be followed by actual offence taken by the recipient of the message. Indeed, it was held not to be necessary than anyone actually receive the message (para 8).

Consequently, the offence of sending a grossly offensive message could be committed without actually offending anyone. What makes it an offence is the fact that the social standards have been infringed. The offence is, on this interpretation, sending a message that would grossly offend a reasonable person if such a person became aware of it. That seems to be a bit of a stretch from the wording of the relevant legislation: s 127 of the Communications Act 2003[UK]:

127. Improper use of public electronic communications network
(1) A person is guilty of an offence if he—
(a) sends by means of a public electronic communications network a message or other matter that is grossly offensive or of an indecent, obscene or menacing character; or
(b) causes any such message or matter to be so sent.”


The justification for the imposition of community standards is the public nature of the communication network, which would be fine if everyone had ready access to everyone else’s communications. In reality, of course, the network is no more public than words exchanged in a conversation between two people on the street.

Friday, June 30, 2006

Our case against you is secret ...

No one could be surprised that included in the guarantees recognised as indispensable by civilised peoples are the right, when on trial, to be present, and to be privy to the prosecution evidence. In Hamdan v Rumsfeld, 548 US 557 (2006), 29 June 2006 the United States Supreme Court held that these rights applied to detainees that the Government proposed to prosecute in special military commissions.

Stevens J, joined by Souter, Ginsburg and Breyer JJ, wrote, in an important footnote (p 71, n67):

"The Government offers no defense of these procedures other than to observe that the defendant may not be barred from access to evidence if such action would deprive him of a "full and fair trial." Commission Order No. 1, ß6(D)(5)(b). But the Government suggests no circumstances in which it would be "fair" to convict the accused based on evidence he has not seen or heard. Cf. Crawford v. Washington, 541 U. S. 36, 49 (2004) (" ‘It is a rule of the common law, founded on natural justice, that no man shall be prejudiced by evidence which he had not the liberty to cross examine’ ") (quoting State v. Webb, 2 N. C. 103, 104(Super. L. & Eq. 1794) (per curiam)); Diaz v. United States, 223 U. S. 442, 455 (1912) (describing the right to be present as "scarcely less important to the accused than the right of trial itself"); Lewis v. United States, 146 U. S. 370, 372 (1892) (exclusion of defendant from part of proceedings is "contrary to the dictates of humanity" (internal quotation marks omitted)); Joint Anti-Fascist Refugee Comm. v. McGrath,341 U. S. 123, 170, n. 17, 171 (1951) (Frankfurter, J., concurring) ("[t]he plea that evidence of guilt must be secret is abhorrent to free men" (internal quotation marks omitted)). More fundamentally, the legality of a tribunal under Common Article 3 cannot be established by bare assurances that, whatever the character of the court or the procedures it follows, individual adjudicators will act fairly."

There is here an obvious relevance to the law concerning the use of special advocates. Such advocates are used, for example, in England and Wales in relation to some immigration matters: see the Special Immigration Commission Act 1997[UK], and Parts 3, 4 and 7 of the Special Immigration Appeals Commission (Procedure) Rules 2003 SI 2003/1034. These procedures have been described as "an unfortunate legacy from someone who rode roughshod over liberties in this country in a breathtaking manner" (per Dominic Grieve, Conservative Member for Beaconsfield, Commons Hansard, 23 February 2005).

A special advocate procedure is planned in New Zealand in relation to whether a person is a risk to the security of the nation for the purposes of certification by the SIS and consequent expulsion.

Wednesday, June 28, 2006

The Ellis case and trial fairness

For a draft paper on the trial of Peter Ellis for sexual offending at the Christchurch Civic Creche, see my web site, and follow the links to more about me, and articles available on this site.

Thursday, June 15, 2006

A look at torts ...

During this strangely extended lacuna between interesting criminal cases around the world, we have a moment to glance at yesterday’s decision of the House of Lords in Jones v Saudi Arabia [2006] UKHL 26 (14 June 2006).

This case concerns the civil jurisdiction of domestic courts to adjudicate on claims in tort for damages for torture inflicted in another country by officials of that foreign country. In contrast to the universal criminal jurisdiction provided for by the UN Convention against Torture and other Cruel, Inhuman or Degrading Treatment or Punishment 1984 (the "Torture Convention"), there is no corresponding universal civil jurisdiction. The claimants in Jones sought to establish that the State Immunity Act 1978[UK] should be interpreted, contrary to its ordinary and natural meaning, so as to permit refusal of immunity in respect of torture claims, because such an interpretation was required by s 3 of the Human Rights Act 1998[UK] to give effect to the Art 6 of the ECHR right of access to courts.

That claim and argument was rejected unanimously. State immunity is a procedural matter determining jurisdiction, and does not have substantive content. There is no international consensus recognising universal civil jurisdiction, and there is no such exception in the UN Immunity Convention 2004. And, there is no evidence that States have recognised an international law obligation to exercise universal jurisdiction over alleged breaches of peremptory norms of international law. Therefore, the (assumed) restriction on access to the domestic courts was directed at a legitimate objective and was not disproportionate, and the interpretation sought by the claimants could not prevail.

See also, blog entry for 17 March 2006.

Thursday, May 04, 2006

Inchoate offences and the scope of mens rea

We have already noted the Supreme Court’s decision in L v R (see blog for 3 April 2006), in which the mental elements of an attempt to commit an offence that could be committed recklessly were held to include the same recklessness. Yesterday, the House of Lords reached a different conclusion in R v Saik [2006] UKHL 18 (3 May 2006), which concerned the statutory definition of "conspiracy". Both attempts and conspiracies are what is called inchoate offences: they are, in this sense, uncompleted substantive offences.

It is interesting to observe that legislatures can differ on what sort of behaviour, preliminary to the commission of a substantive offence, they consider appropriate to criminalise, and they can also differ on how precisely they wish to indicate the scope of proscribed behaviour.

Under English law, money laundering (the relevant substantive offence in Saik) can be committed knowingly, or by having reasonable grounds to suspect, in relation to the fact that the money or property is the proceeds of criminal conduct. The New Zealand definition of money laundering is broadly similar in these respects; both the Crimes Act 1961 and the Misuse of Drugs Act 1975 contain laundering offences, and the relevant state of mind is knowing, believing, or being reckless as to the money or property being proceeds of a serious offence.

There is, obviously, a difference, in that the English provision encompasses having reasonable grounds to suspect, whereas in New Zealand the scope does not extend beyond recklessness.

The material difference for present purposes is in the definitions of the inchoate offences. The English statutory conspiracy is defined in s 1(1) and (2) of the Criminal Law Act 1977[UK], and, significantly, s 1(2), as interpreted in Saik, requires that, in relation to essential circumstances that need not be known for liability for the full offence, the accused must intend or know of them for liability to arise for the conspiracy. Thus, for conspiring to launder property, where the full offence does not require knowledge that the property is proceeds of criminal conduct (reasonable grounds to suspect being sufficient), the consequence of the Saik interpretation of s 1(2) is that conspiring to launder property requires proof that the accused knew or believed that the property was proceeds of criminal conduct.

In New Zealand, the crime of attempting to commit an offence is defined in s 72 of the Crimes Act 1961, which requires an "intent to commit an offence". In L v R this was interpreted to mean, for attempted sexual violation, intention to penetrate (or, in the peculiar circumstances of that case, an intention that penetration should occur), and recklessness as to whether the victim consented. The policy applied in L has thus, from this point of view, resulted in expansion of the meaning of "intent" in s 72.

In summary, the UK Parliament defined conspiracy narrowly, and the Court declined to interpret the definition in a way that would have extended it, whereas the New Zealand Parliament defined attempt without specifying whether it should be construed widely or narrowly, and the Court chose a wide interpretation.

Wednesday, May 03, 2006

Thou art far more fair than she ...

Is Australian criminal law really fairer than New Zealand’s? In Moloney v New Zealand [2006] FCA 438 (21 April 2006) (blogged here on 1 May 2006), Madgwick J thought it is. It is arguable that his approach to this matter was wrong. He should have used as fundamental a point that he did note, namely that since New Zealand, like Australia, is a party to the International Convention on Civil and Political Rights, and, further to what he noted, that since New Zealand, unlike Australia, has the accused’s right to a fair trial included in a Bill of Rights, courts in New Zealand are unlikely to tolerate an unfair trial.

The question would then have been, would New Zealand judges be likely to interpret their rules of evidence in a way that results in unfairness to the accused?

It is not a matter of comparing the rules in Australia with those in New Zealand, as Madgwick J did, because it is the result of the application of the rules that is critical. The real question in Moloney was, does the right to a fair trial in New Zealand have overriding importance?

I have argued that it does: see "The Accused’s Right to a Fair Trial: Absolute or Limitable?" [2005] New Zealand Law Review 217. I must acknowledge, however, that the mere fact that the eponymous question has to be asked indicates that in New Zealand there has been some doubt over whether the accused’s right to a fair trial overrides other rights and interests. My conclusion in that article was that the majority of senior New Zealand judges appear to accept the absolute view of the accused’s right to a fair trial. Concern over the status of trial fairness in New Zealand should focus on the grounds on which judges disagree over that.

In Moloney, Madgwick J gave two main reasons for concluding that trials of the relevant charges, allegations of historical sexual misconduct going back 22 to 31 years, on balance would be unlikely to be conducted fairly in New Zealand, if fairness is assessed by Australian standards. Each reason has its weaknesses.

The first reason concerned judicial warnings to the jury about the reasons that evidence by complainants about events that are allegedly remembered after so long a time may be unreliable, and reasons why the accused may be disadvantaged in challenging such allegations. In Australia such warnings, in cases of delay of this length, are apparently mandatory. That, at least, was the view of Madgwick J, and we may for present purposes assume that he was correct on this point. In New Zealand they are discretionary. What, then, is the significance of this difference?

As suggested above, the question should have been, would New Zealand judges be likely to decline to warn the jury about those matters in the circumstances of these cases? And, if they did not give the warning, would convictions be likely to be upheld on appeal; that is, would the Court of Appeal recognise that failure to give the warning amounted to a substantial miscarriage of justice?

The current state of development of the law in New Zealand is such that, while we may be reasonably sure that a trial judge would warn the jury, if he did not, we cannot be sure that an appellate court would regard that omission as a substantial miscarriage of justice. The problem of appellate recognition of substantial miscarriage of justice is not confined to New Zealand: the Privy Council not infrequently differs among its members on this. There is, therefore, reason for some unease on this score.

The second reason that Madgwick J gave was that in Australia the charges would be heard at separate trials, because the evidence of each complainant was only relevant to the allegations made by that complainant, whereas it seemed that in New Zealand a court would be likely to regard the evidence of some complainants as corroborative of the evidence of others, so that some joinder of trials may occur. This area of the law of evidence concerns what is usually called "similar fact" evidence. Essentially, where different complainants make similar allegations, one tends to increase the likelihood that another is true, unless there was evidence that they had colluded to concoct falsely similar stories. The proper focus of similar fact evidence, where it is admissible, is on the weight to be given to the evidence of the complainant whose allegations are being decided. The great danger of this evidence is that juries will go straight to a conclusion that the accused is the kind of person who does this sort of thing, so he must be guilty. In other words, the error would be to convict the accused for what is now alleged because of what he is said to have done before, rather than because this complainant is believable.

The law about similar facts has been problematic everywhere. This has often been because its admissibility has been linked to the criterion of whether the probative value of the evidence outweighs its illegitimately prejudicial effect. I have discussed this in "Probative value, illegitimate prejudice and the accused’s right to a fair trial" (2005) 29 Crim LJ 8. In Moloney, Madgwick J considered that in Australia the similar fact evidence would not be admissible. Assuming that to be so, would it be admissible in New Zealand? A leading case on this is R v Holtz. I have discussed this case in Misuse of Drugs, para 306, as follows:

"There could appear to be some withdrawal from the requirement of hallmark or striking similarity in R v Holtz [2003] 1 NZLR 667; (2002) 20 CRNZ 14 (CA). But whether that is so, and if so, whether it is to be taken as a generally applicable modification of the law, may be doubted. The Court observed at para 35 that it is wrong to look for principles of admissibility applicable to all evidence of past conduct in all circumstances. Identity was disputed in relation to some of the allegedly similar facts, and at para 43 the Court merely required a credible strand of circumstantial evidence pointing to the accused as the offender. However, a more rigorous requirement appeared at para 47, where it was stated that the evidence must be truly probative and cogent. In full, this crucial paragraph reads as follows:

"[47] The care with which evidence of similar acts is scrutinised is justified because of the prejudice that inevitably arises from the risk of guilt being improperly inferred from mere propensity or disposition evidenced by previous bad conduct. But, where the evidence is truly probative and cogent, admission is appropriate so long as the circumstances are such that, while allowing the probative value of the evidence to be availed of, the risk of improper use can be avoided by appropriate directions to the jury."

"It should be noted here that the requirement is that the risk of improper use of the evidence can be avoided, not just reduced to a level where it is outweighed by the probative value of the evidence."

If this criterion is applied strictly the "trial according to law" aspect of the right to a fair trial would be protected. Unfortunately, it cannot be said for certain, at present, that the "avoiding" of improper use of the evidence would be required, instead of merely the risk of improper use being "outweighed" by the probative value of the evidence. Again, on this point, one cannot be sure that judges in New Zealand would interpret this rule of evidence in a way that would ensure the dominance of the accused’s right to a fair trial.

Monday, May 01, 2006

A difference of reasonable minds

Fairness can be a very difficult matter to agree on, as is demonstrated in cases where judges have differed among themselves: see, for illustrations, blog entries for 10 October 2004, 10 July 2005, 28 August 2005, 9 December 2005 and 7 March 2006. Applications to extradite suspects to other jurisdictions where fairness of trial may be questioned give rise to the issue of what is an acceptable risk of unfair trial in the foreign court. In Bagdanavicius [2005] UKHL 38 (blogged 26 May 2005) it was held that extradition must be refused if there are "substantial grounds" for believing there to be "a real risk" of mistreatment in the foreign jurisdiction.

In Moloney v New Zealand [2006] FCA 438 (21 April 2006) the Federal Court of Australia (Madgwick J) held (para 120) that the burden, on appeal, was on the defendants (those resisting extradition) to show there would probably be an injustice in extradition. This might (it is difficult to say, as the point is not discussed by Madgwick J) be an easier standard to satisfy than that required by the House of Lords in Bagdanavicius.

That aside, Moloney points to some areas in New Zealand criminal law that are not self-evidently fair. Diplomatically, Madgwick J acknowledges (para 108) that "reasonable minds may and do differ on what constitute the incidents of a fair trial", so that, in effect, this is a difference between friends. No insult being intended, the Federal Court held that, judged as it must be by the standards of Australian law (Bannister v New Zealand (1999) 86 FCR 417), trial of the charges in New Zealand would probably be unfair.

Moloney holds that New Zealand criminal law falls short of the Australian standard of fairness in the following respects:

(1) In cases of historical allegations, in Australia judges must warn the jury about the dangers inherent in accepting the evidence of complainants (this is called the direction in Longman v R (1989) 168 CLR 79 (HCA)), whereas the New Zealand Court of Appeal has rejected the mandatory requirement of a warning, preferring to leave the need for a warning as a matter for the judge to decide: R v M 13/11/95, CA187/95. This was held to be the factor that made the difference between Australian and New Zealand law sufficiently serious to prevent extradition. Australian courts regard the Longman direction as being "a vital requirement for a just trial in a case of long delay" (para 109). But there were other factors too.

(2) Whereas in Australian law it is clear that the charges would have to be heard separately, because similar fact evidence would be inadmissible, this was not so clear in New Zealand law (citing R v Holtz [2003] 1 NZLR 667, 675 (CA)). While this uncertainty made this factor less than decisive, it was, nevertheless, "a circumstance exacerbating the disabilities" caused to the defence and arising from the delay (paras 117, 123).

(3) Australia has rejected the use of representative charges (S v R (1989) 168 CLR 266 and KBT v R (1997) 191 CLR 417 (HCA)), whereas they are acceptable in New Zealand: para 110, quoting R v Accused [1993] 1 NZLR 385, 389 (CA). The objections to representative charges concern vagueness as to when the offence for which the accused is convicted occurred, and what facts were accepted as proof of it.

There was, in summary, between Australia and New Zealand "a fundamental difference as to the content of an effective right to a fair hearing, such right being recognised … as a basic human right" (para 113).

Apologists for New Zealand law might argue that, vague though the law may be, everything would turn out alright in the end, as appellate courts can take an overview and correct unfairness. Such wooliness is, indeed, behind the development of the law to this unsatisfactory state. Disturbingly, New Zealand has been led into this difference of "reasonable minds" by some of our foremost judges: participants in New Zealand Court of Appeal decisions cited in Moloney include Cooke P (now, Lord Cooke), Gault P, Richardson J (subsequently P), Casey, Hardie Boys, and Keith JJ. In this area, rules are preferable to discretions, and precision must be pushed as far as it will go.

Having said that, it should be acknowledged that a face-saving appeal against Madgwick J’s decision in Moloney could result in a different view of New Zealand’s approach to fairness. It could be held that Madgwick J was wrong to consider that in Australia a warning was inevitable in the circumstances of the case(s), citing Doggett v R (2001) 182 ALR 1 (HCA), where McHugh J dissented in holding that no warning was required in that case, as an illustration of how the need for a warning can be controversial. That would make the position on warnings as uncertain in Australia as it seems to be in New Zealand. Further, while the risk of evidence of other complainants being admissible on a similar fact basis seems to be high in New Zealand, it cannot be discounted in Australia either, because the merits of the case (correctly recognised as not relevant to the extradition decision) may give other complaints high probative value as corroboration. Such corroboration would, in turn, reduce the need for a warning in Australian law. The result could be that New Zealand’s law on fairness is not significantly different from Australia’s.

That, however, could be said to mean that the law on fair trials is equally unsatisfactory in each country.

Wednesday, April 26, 2006

Public policy and degrees of misconduct

Public policy exclusion of evidence is sometimes still spoken of by courts as if it was a means of protecting the accused’s right to a fair trial. This was done yesterday by the Privy Council in Williams v R (Jamaica) [2006] UKPC 21 (25 April 2006). A statement had been obtained from the accused when he was aged 12, in breach of the procedures laid down for the interviewing of young suspects (para 27). The Board concluded, at para 28:

"…the circumstances of the appellant's detention and of the taking of the statement were such as to create a significant amount of unfairness to him. Their Lordships cannot conclude that in all the circumstances of the case it was fair to admit the statement."

While the Board was not expressly referring to trial fairness, the expressions "unfairness to him" and it not being "fair to admit" the statement do suggest that trial fairness is the object of concern. Another point in this extract that is misleading is the expression "a significant amount of unfairness".

My view is that in these situations the courts are not excluding the evidence for reasons that have anything to do with trial fairness. The probative value of the evidence is not relevant to the decision on admissibility in these cases (see, for example, my blog entry for 4 April 2006 concerning "cogency"). It is not the accused’s guilt that matters, it is the objectionable way in which the evidence was obtained that is critical. This is why the discretion, in this area, is best called the public policy discretion. It is true that, historically, this discretion evolved from – and remains part of - the court’s inherent power to prevent an abuse of process. In turn, that power has been referred to as giving rise to a discretion to exclude evidence in the interests of "fairness", or, sometimes, "fairness to the accused". But these are not references to trial fairness.

One of the reasons for making this distinction between the public policy exclusion of evidence, and the exclusion of evidence to ensure trial fairness, is to preserve the concept of the absolute nature of the accused’s right to a fair trial. In the above quotation from para 28 of Williams, the expression "a significant amount of unfairness" could, wrongly, suggest that fairness of trial exists in gradations, and that some forms of trial unfairness are acceptable. I have given examples of misuse of this terminology in "The Duty to Prevent an Abuse of Process by Staying Criminal Proceedings" in Essays on Criminal Law – A Tribute to Professor Gerald Orchard (Brookers Ltd, 2004), 133, 146.

What, it is respectfully suggested, the Privy Council should have said in Williams, is that the breaches of the Directions on the conduct of interviews of young persons that occurred in this case were sufficiently serious that admission of the statement obtained thereby would be an abuse of process. The evidence was excluded to prevent the administration of justice being brought into disrepute, as would occur if the courts appeared to endorse the police misconduct. In that context it is appropriate, if one must use the "fairness" terminology, to speak of degrees of unfairness, because official misconduct comes in degrees.

There may be cases where trial fairness could be relevant to the admission into evidence of a statement that had been obtained wrongfully. But it must be remembered that, at the stage when the ruling on admissibility has to be made, the effects of admission on the defence will not be known: the Judge will not know whether the defendant intends to give or call evidence, let alone what any such evidence would be. There is a difference between using fairness as grounds for excluding evidence, and using fairness in considering, as an appellate court, whether to apply the proviso. It is the appellate court that is best placed to evaluate the effect of an erroneous admission of evidence, and, when the public policy discretion has been exercised in favour of admitting the evidence, it may be only retrospectively that the trial can be said to have been unfair.

Seeing and believing

Eyewitness identification evidence may need to be treated with some circumspection, and juries are usually given a direction on the special need for caution before relying on such evidence. In some jurisdictions, these warnings are required by statute, although, as is the case in New Zealand, the points required to be covered are not spelt out in great detail. The common law antecedent of these directions is known as the Turnbull direction, originating in the English Court of Appeal’s decision R v Turnbull [1977] QB 224.

In Edwards v R (Jamaica) [2006] UKPC 23 (25 April 2006) the Privy Council indicated that there are some practices that should not be permitted at trials where identification is an issue and the prosecution relies on evidence of an eyewitness to the offence. These are:

(1) The eyewitness should not be permitted to identify the accused in the dock as the offender. The prosecution should, in general, adopt other means for establishing that the accused is the person who was arrested (para 22):

"… it is only in the most exceptional circumstances that any form of dock identification is permissible: cf the discussion in the Scottish devolution appeal Holland v HM Advocate [2005] UKPC D1, 2005 SLT 563. It may be borne in mind that this was far from being a first identification and it can fairly be said that the dock identification may have had little impact on the minds of the jury. It is, however, an undesirable practice in general and other means should be adopted of establishing that the defendant in the dock is the man who was arrested for the offence charged."

(2) A police officer should not be permitted to give in evidence his opinion on why an ID parade was not considered to be necessary (para 23).

(3) The police should not give in evidence the fact that a warrant was obtained for the arrest of the accused, or of the information on which the police acted, as this is hearsay and potentially highly prejudicial (para 23).

(4) The police should not give in evidence the fact that a potential witness was unwilling to come forward (para 23).

(5) The police should avoid confronting the eyewitness with the suspect (para 25).

There is, at this point – para 25 – a possibly unintentional suggestion by the Board that hearsay evidence might be given to establish the link between the person described to the police as the offender, and the suspect:

"The arresting officer would have been quite capable of establishing that the appellant was the person pointed out to him by Bailey [the eyewitness] near the Mango Tree Bar, so it was unnecessary to ask Bailey to come to the station to confirm that."

This, however, should be read as referring to "establishing" in an investigatory, pre-trial, sense, and not as "establishing" in evidence at trial.

In this case, the eyewitness to the killing had been standing next to the victim, and the bullet that killed the victim had passed through the eyewitness. The offender had been trying to rob the eyewitness, who suddenly and unsuccessfully tried to grab the gun. The eyewitness was hospitalised for 4 weeks, and it was 2 months after the killing before he saw the accused near the same bar. He claimed that the accused was the offender. In his first description of the offender, given 5 days after the incident while he was in hospital, the eyewitness failed to mention a prominent birthmark on the accused’s face, he was unable to say what sort of trousers the offender was wearing, and he claimed that the time he had to observe the offender was a couple of minutes although it must have been shorter than that. The circumstances in which the offender was observed were good: inside a bar in the morning with good lighting.

The Privy Council was concerned that there could have been an erroneous association of ideas arising from the location of the offence and the subsequent identification being similar, and that the judge had not adequately warned the jury of the dangers in accepting the evidence. The conviction was therefore unsafe.

In New Zealand, the Evidence Bill 2006, clause 122, almost exactly repeats the current provision on the need for judicial warning: Crimes Act 1961, s 344D. The slight difference is that instead of requiring the judge to "include the reason for the warning", the Bill requires the judge to "warn the jury that a mistaken identification can result in a serious miscarriage of justice". The need for a warning arises "In a criminal proceeding tried with a jury in which the case against the defendant depends wholly or substantially on the correctness of 1 or more visual or voice identifications of the defendant or any other person …". The inclusion of voice identification is new to the Bill.

The Bill contains other provisions relating to the admissibility of visual identification evidence. It is important to note that here the concern is with admissibility, not with the way admissible evidence is treated at trial. These provisions, for visual identification, are in clause 41, and they concern the implications of whether or not a formal identification procedure was used at the investigatory stage. The criterion for admissibility is proof, on the balance of probabilities, that the evidence is reliable. The Bill does not say to what extent, if any, this reliability should be assessed by reference to the other evidence in the case. It seems plain that the other evidence should not be included in the assessment of the reliability of the visual identification evidence, and that the focus should be on the circumstances in which the identification was made.

Tuesday, April 04, 2006

Balancing "cogency" of wrongfully obtained evidence

Simmons v R (Bahamas) [2006] UKPC 19 (3 April 2006) gives us an opportunity to highlight the distinction between "fairness" when that term is used in the context of the exercise of the public policy discretion to exclude evidence, and "fairness" in the separate sense of trial fairness for the accused.

The two appellants were convicted of murder. Included in the evidence against them were statements they had made to the police. These statements, referred to as confessions, although they were partly exculpatory (and, one accused who gave evidence adopted what he had said to the police), were obtained in breach of the appellants’ constitutional right to be informed of the availability of legal advice before they spoke to the police.

Breach of that right gives rise to a judicial balancing exercise to determine whether to admit or exclude the statement. This balancing exercise is a public policy discretion, having nothing to do with the fairness of the trial. It arises because of the conflict between, on the one hand, the public’s right to have suspects prosecuted and offenders brought to justice, and, on the other hand, the public’s right to have officials comply with the law in the investigation of offences. The only sanction that courts can impose on officials who act in breach of the law in this context is to exclude evidence that they obtain thereby.

Sometimes, in carrying out this balancing exercise, judges say that the cogency of the evidence is a factor strongly favouring its admission. An important observation on this point was made by the Privy Council in this case, para 26:

"The Board has one other concern about the judge's balancing of the respective interests of the prosecution and the defence on the issue of fairness: the evident importance which she attached to the confession being "very cogent evidence against Simmons." Their Lordships cannot accept that the potency of such evidence is necessarily a factor in favour of its admission. If, by denying a suspect his constitutional right to see a lawyer and perhaps be advised against making a statement, the state's case is thereby strengthened by a confession, it is by no means self-evident that fairness demands its admission rather than its exclusion."

However, in New Zealand the cogency of the evidence is routinely taken into account in this balancing exercise: R v Shaheed [2002] 2 NZLR 377 (CA), especially at paras 151-152. At para 151 the joint judgment (Richardson P, Blanchard and Tipping JJ) states: "A trial is not to be regarded as potentially unfair by reason of the admission of evidence unless that evidence may lead to an unsafe verdict." That, with respect, needs to be read with some caution. A safe verdict is not a cure-all for trial unfairness. That point was made strongly by Lord Steyn (Sir Swinton Thomas concurring) in Ebanks v R (Cayman Is) [2006] UKPC 16 (27 March 2006), blogged 28.3.06, at para 40. The Supreme Court acknowledged the same point in Sungsuwan v R [2005] NZSC 57 (25 August 2005), blogged 26.8.05, per Elias CJ at para 6 (putting as alternatives trial unfairness and unsafe verdicts), Tipping J at 112 (lack of a fair trial is itself a miscarriage of justice without the need to consider its effect on the verdict).

If I may, I should add that "cogency", which means being convincing or compelling, is always a matter for the jury. It is usually called the "weight" that is to be given to the evidence. Weight is separate from the question of admissibility, except on occasions where it is possible for a judge to conclude that no reasonable jury could give the evidence any weight. The Privy Council is correct to see no reason to link cogency to admissibility. However, this is not the end of the matter. What is being considered is not the admissibility of the evidence, but rather whether, as admissible evidence, it should be excluded because of the objectionable way in which it was obtained. The question whether convincing or compelling admissible evidence should be excluded is, appropriately, part of the weighing of the public interest in bringing suspects to justice. There is, though, a difficulty: the cogency of the evidence is also appropriately considered on the other side of the balance, where weight has to be given to the public interest in prevention of such abuse of process as would bring disrepute to the administration of justice. One might properly object to the inclusion of "cogency" in the weighing process on the basis that it falls on both sides of the scales.

That aside, having said that the exclusion of evidence to prevent the trial being unfair to the accused is separate from the public policy discretion, I should now make the distinction clear. A fair trial for the accused is one where the law is accurately applied and the facts are determined without bias. There may be flow-on effects of a wrongful exercise of the public policy discretion, in the sense that the trial may not be one where the law has been accurately applied. The question, in terms of trial fairness, of the significance of the error, will be determined, not by the strength of the other evidence against the accused, but by whether the error put the accused at a disadvantage in the trial. I made this point in the blog entry on 29 March 2006, discussing Gilbert v R (Grenada) [2006] UKPC 15 (27 March 2006).

Simmons is an example of the error in applying the public policy discretion not affecting the fairness of the trial. The statements in issue were partly exculpatory and the accused who gave evidence adopted what he had said. The judgment does, however, focus on the strength of the other evidence of guilt, concluding, para 31, that acquittals would have defied all reason. The relevance of this point is that the proviso could be applied: the error in admitting the statements did not amount to a "substantial miscarriage of justice." The error caused neither trial unfairness, nor the loss of a real chance of acquittal.

Monday, April 03, 2006

You may not get what you want ...

The crime of attempting to commit a crime has come under scrutiny in L v R [2006] NZSC 18 (30 March 2006), which concerned the state of mind required for liability.

The facts of L were extraordinary. The charge, faced by L, a 49 year old woman, was attempting to sexually violate a 15 year old male. She tried to cause his (presumably erect) penis to penetrate her vagina, being reckless as to whether or not he consented to that. In the circumstances, it seems to have been the case that his lack of consent should have been apparent to her.

The legal question was whether, to be liable for the attempt, L had to intend that he not consent, or was it sufficient that she merely be reckless (ie that she knowingly took a risk) as to his lack of consent. The question arose because the crime of attempting to commit a crime requires the accused to have an intent to commit that crime.

Plainly, L intended that the penis enter her own vagina, and she did acts in an effort to bring that about. Was that intent sufficient? Or, in addition, did she need to be reckless about the male’s lack of consent? Or, did she need to intend both penetration and his lack of consent?

There has been some division among the jurists on this general issue. As Simester and Brookbanks, Principles of Criminal Law, point out at para 6.1.1, the decision is one of public policy. Criminalising attempts is often justified because of the danger posed to society by a person who sets out to cause harm. That is, one whose intention is dangerous, albeit that his (here, her) acts may not constitute actual harm. This sort of deliberation suggests that intention should be given a narrow meaning, so as to exclude people who are reckless about circumstances such as consent. But, that narrow meaning would create anomalies (as has been pointed out by Professor Ashworth, cited ibid): imagine two men who set out to have sexual intercourse with two women. Both men are reckless as to consent, and neither woman consents. One man succeeds in penetration, and is guilty of rape. The other fails to penetrate, and, under this narrow view, is guilty of neither rape nor attempt to rape.

In L the Supreme Court held that recklessness (in the sense of absence of a reasonable belief in consent) as to circumstances is sufficient for liability for the attempt: what has to be intended here is the act of penetration.

All well and good. We are left with a lingering image of the facts of the case. In future, will people charged with sexual violation claim they did not consent, and that the other person should be charged?

Wednesday, March 29, 2006

Good character again

Once again, see Index, the topic of the significance of absence of a good character direction has concerned the Privy Council: Gilbert v R (Grenada) [2006] UKPC 15 (27 March 2006). Their Lordships indicated that a forthcoming case also concerns this subject: Bhola v The State (Trinidad and Tobago) [2006].

Gilbert makes the point that there is a duty on counsel for the defence to ensure that the trial judge understands that a good character direction is sought: para 11, citing Thompson v R [1998] AC 811 (PC). Omission of the direction requires attention to two matters: was the trial thereby rendered unfair to the accused, and did the omission render the verdict unsafe? (citing for these, Teeluck and John v The State [2005] UKPC 14 at para 39, blogged here, 1 April 2005).

Whereas it may have seemed on some readings that the law requires the direction to be given as a matter of course, in Gilbert it was emphasised that the requirement is that the direction should "normally" be given (para 15, citing R v Aziz [1996] 1 AC 41). That is, the issues of fairness and effect on the result are examined in the particular circumstances of each case. That, of course, is obvious for the result point, and Gilbert makes it clear that fairness is also a matter to be determined on the particular circumstances. There are, thus, no hard or inflexible rules about whether fairness requires the direction, rather the nature of the issues in each case must be examined: para 20, citing Lord Bingham in Singh v The State [2005] UKPC 35 at para 14.

Here, then, the circumstantial evidence of guilt was substantial (para 18), and the accused’s good character was wholly outweighed by the nature and coherence of the circumstantial evidence (applying a phrase used by Lord Hope in Balson v The State [2005] UKPC 6, para 37). The absence of the good character direction therefore had no effect on the fairness of the trial or on its result.

One must observe, with respect, that the question of fairness, which is a procedural matter, will be swamped by the question of result, which is an evidential matter, unless the two are kept distinct. Certainly, the Privy Council here, at para 19, noted that the trial was fair in all other respects, even being "unduly fair" [sic] in that the judge refrained from commenting on the accused's failure to give evidence. Nevertheless, the risk remains that the impression will be given that unfair procedure may be cured by strong evidence, which is undoubtedly not what the Privy Council intended to convey.

The difficulty is that a good character direction is aimed at assisting the jury to evaluate the weight to be given to the evidence, and here it was the complainant's evidence. In fairness terms, the direction is concerned with the avoidance of bias. What is relevant to bias is, not the strength of the prosecution evidence, but whether the absence of the direction put the accused at a disadvantage. On the facts of Gilbert, where the accused was a minister of religion, his previous good character was probably taken for granted by the jury. Therefore, it could safely be concluded that there was no unfairness to him occasioned by the absence of a good character direction.

Tuesday, March 28, 2006

Two points of conduct

In Ebanks v R (Cayman Islands) [2006] UKPC 16 (27 March 2006) the Privy Council reminded us of a couple of fundamental points:

- when a client instructs that he will not give evidence, that instruction must be in writing and signed by the client.

- counsel in cross-examining must put his client’s case, including allegations that the witness is lying, even if the client is not going to give, or call, evidence to establish that fact.

In Ebanks, the appeal against conviction for murder was brought on the grounds of failure by counsel to carry out the client’s instructions in the conduct of his defence, and that this resulted in an unfair trial.

Here again, we have the difficult topic of trial fairness, and again the Privy Council divided 3 – 2 on this issue. We saw a similar split in Howse v R (see these blogs for 10 October 2004), in a Board that was differently constituted but with some overlap in membership. In Ebanks, Lord Rodger gave the decision of the majority (Lords Carswell and Mance agreeing) upholding the conviction, and the dissenters were Lord Steyn and Sir Swinton Thomas. In Howse, the majority, upholding the convictions, were Lords Hutton and Carswell, and the dissenters were Lord Rodger and Sir Andrew Leggatt.

In Ebanks, the reason that persuaded the majority that there had not been unfairness emerged from a consideration of counsel’s conduct at trial, when he appeared to be regularly consulting with his client in a way that was consistent with a carefully conducted defence. In those circumstances, it was possible, notwithstanding the absence of a written record, to reject the suggestion that counsel had not followed the client’s instructions.

The minority in Ebanks focused on the conduct of the hearing in the Cayman Islands Court of Appeal. They held that this had involved a failure of due process, because the Court had wrongly refused to allow the appellant to supplement his affidavit with oral evidence. That Court had also been wrong in its understanding of the duty of counsel in cross-examining to put the client’s case that the police witnesses were lying. The Court had inferred that these allegations were not put because Mr Ebanks had instructed that he would not give evidence, and that his post-conviction claim that he wanted to give evidence was wrong (para 42 per Lord Steyn). This, Lord Steyn (with Sir Swinton Thomas agreeing) held, was a material irregularity, potentially prejudicing the appellant. It seems that in his affidavit filed in the Court of Appeal, Mr Ebanks had not addressed the point about what he had instructed counsel about whether he would give evidence. Lord Steyn considered that if he had been allowed to give oral evidence on that, counsel may well have agreed that he had indeed given that instruction. If that had happened, he would have been denied a fair trial. Sir Swinton Thomas agreed, emphasising the significance of the absence of a written record of the client's instructions, despite his waiver of privilege.

Process takes priority over substance. It cannot be said that guilt can be determined reliably at an unfair trial. Lord Steyn applied, mutatis mutandis, a passage from Wade and Forsyth, Administrative Law, 9th Ed (2005) at 506-508 (see para 40 of his judgment - here I have italisised the passage from the textbook):

" "Procedural objections are often raised by unmeritorious parties. Judges may then be temped to refuse relief on the ground that a fair hearing could have made no difference to the result. But in principle it is vital that the procedure and the merits should be kept strictly apart, since otherwise the merits may be prejudged unfairly. Lord Wright once said:
'If the principles of natural justice are violated in respect of any decision it is, indeed, immaterial whether the same decision would have been arrived at in the absence of the departure from the essential principles of justice. The decision must be declared to be no decision.'
The dangers were vividly expressed by Megarry J, criticising the contention that 'the result is obvious from the start':
'As everybody who has anything to do with the law well knows, the path of the law is strewn with examples of open and shut cases which, somehow, were not; of unanswerable charges which, in the event, were completely answered; of inexplicable conduct which was fully explained; of fixed and unalterable determinations that, by discussion, suffered a change.'
. . . Judges are naturally inclined to use their discretion when a plea of breach of natural justice is used as the last refuge of a claimant with a bad case. But that should not be allowed to weaken the basic principle that fair procedure comes first, and that it is only after hearing both sides that the merits can be properly considered."

Although made in the administrative law context, these observations are also germane to the question whether the Court of Appeal could have declined to hear oral evidence on the ground that it would make no difference. To have decided the case on such a basis would have been unfair and contrary to due process. After all, it is entirely possible (and even likely) that Mr St. John Stevens [counsel for Mr Ebanks at trial]would have made the same core concession that Mr McGrath [Attorney for Mr Ebanks]made, viz that the Appellant insisted that he made no confession to the police."

One must, of course, make allowances for the differing circumstances of each case, but it is interesting to observe that a judge (Lord Rodger) who could uphold the trial in Ebanks as being fair, was such a strong dissenter in Howse.

The dicta in this case about counsel’s duty to put the case to witnesses in cross-examination, even where evidence is not being called to establish the points, in effect amplify Rule 10.02 of our Rules of Professional Conduct for Barristers and Solicitors.

Thursday, March 23, 2006

School Rules, OK!

Yesterday, the House of Lords decided Begum v Denbigh High School [2006] UKHL 15 (22 March 2006), a case that has some bearing on how a provision commonly found in Bills of Rights is to be understood.

The relevant provision in New Zealand is s 5 of the Bill of Rights:

5. Justified limitations - Subject to section 4 of this Bill of Rights, the rights and freedoms contained in this Bill of Rights may be subject only to such reasonable limits prescribed by law as can be demonstrably justified in a free and democratic society.

In Begum, the interpretation of article 9 of the ECHR was in issue, and it can be seen that art 9.2 corresponds to s 5 of NZBORA:

"Freedom of thought, conscience and religion
9.1 Everyone has the right to freedom of thought, conscience and religion; this right includes freedom to change his religion or belief and freedom, either alone or in community with others and in public or private, to manifest his religion or belief, in worship, teaching, practice and observance.
9.2. Freedom to manifest one's religion or beliefs shall be subject only to such limitations as are prescribed by law and are necessary in a democratic society . . . for the protection of the rights and freedoms of others."


The difference in wording is immaterial, because neither s 5 nor art 9 seeks to establish a set of procedural steps that courts must apply in deciding whether limitations on rights are justified.

Lord Bingham, at para 31, emphasised that what matters is the practical outcome (ie whether a rule is an unjustified limitation of a right), not the quality of the decision making process that led to it (ie not, on the facts of this case, whether the school authorities reasoned correctly when they decided upon the rule).

Lord Hoffmann agreed, saying at para 68:

" … article 9 is concerned with substance, not procedure. It confers no right to have a decision made in any particular way. What matters is the result: was the right to manifest a religious belief restricted in a way which is not justified under article 9.2? The fact that the decision-maker is allowed an area of judgment in imposing requirements which may have the effect of restricting the right does not entitle a court to say that a justifiable and proportionate restriction should be struck down because the decision-maker did not approach the question in the structured way in which a judge might have done. Head teachers and governors cannot be expected to make such decisions with textbooks on human rights law at their elbows. The most that can be said is that the way in which the school approached the problem may help to persuade a judge that its answer fell within the area of judgment accorded to it by the law."
 
Begum suggests that a formalistic approach to the decision, about whether a right has been subject to unjustified limitation, is inappropriate, and that it is the practical outcome that matters. In Begum the school rule that proscribed the wearing of religious clothing was held, in the particular circumstances of this case (it is important to recognise that the Law Lords were not usurping Parliament’s power by seeking to rule on the validity of legislation that gave the power to make rules to the governing body of each school), not to be an unjustified limitation of the right to express religious belief. The school in question had carefully considered its rule about uniforms, in the context of the composition of its roll and the views of the community it served. Lord Bingham concluded, para 34:

"On the agreed facts, the school was in my opinion fully justified in acting as it did. It had taken immense pains to devise a uniform policy which respected Muslim beliefs but did so in an inclusive, unthreatening and uncompetitive way. The rules laid down were as far from being mindless as uniform rules could ever be. The school had enjoyed a period of harmony and success to which the uniform policy was thought to contribute. On further enquiry it still appeared that the rules were acceptable to mainstream Muslim opinion. It was feared that acceding to the respondent's request would or might have significant adverse repercussions. It would in my opinion be irresponsible of any court, lacking the experience, background and detailed knowledge of the head teacher, staff and governors, to overrule their judgment on a matter as sensitive as this. The power of decision has been given to them for the compelling reason that they are best placed to exercise it, and I see no reason to disturb their decision."

In Hansen, blogged here 19 September 2005, alternative interpretations of a section that has a limiting effect on the presumption of innocence are under consideration. The Begum approach would look at the practical outcome of each. One would require an accused to establish innocence on the balance of probabilities. This is a serious limitation on the presumption of innocence. In the absence of any evidence that Parliament deliberately intended (ie intended after deliberation) that that limitation should be imposed, the alternative interpretation, that the section requires the accused to raise a reasonable doubt about his guilt, should be preferred. In deciding whether Parliament intended the more limiting meaning, it must be relevant to note that when the section was recently re-enacted, there was no Report from the Attorney-General to the effect that the provision appears to be inconsistent with any of the provisions of NZBORA, as is required by s 7. In the absence of an express parliamentary intention to limit the presumption of innocence, the courts should not impose such a limitation through interpretation.

On this approach, s 5 works as a guide to Parliament when it considers legislating to limit rights. This point was made by the Chief Justice in argument during the Supreme Court hearing in Hansen v R. It also reminds courts engaged with provisions that may impinge on rights to look for indications of legislative intent to limit rights.

Friday, March 17, 2006

Rights and Power

This week’s decision of the European Court of Human Rights (Fourth Section) concerning Application no. 23276/04 by Saddam Hussein draws our attention to some characteristics of rights and jurisdiction.

Ideally, human rights should be enjoyed by everyone, no matter where they happen to be. Rights, however, mean nothing unless they can be enforced. Enforcement involves an assertion and exercise of jurisdiction by an authority with power. Rights ultimately depend on power.

It was argued for the Applicant that he had rights under the European Convention on Human Rights, because he was arrested in an area where jurisdiction was exercised by countries, as part of a coalition, that included countries that are bound by the Convention:

"He maintained that he fell within the jurisdiction of all the respondent States because they were the occupying powers in Iraq, because he was under their direct authority and control or because they were responsible for the acts of their agents abroad. He further argued that he remained within their jurisdiction following the transfer of authority, and his transfer, to the Iraqi authorities in June 2004 because the respondent States remained in de facto control in Iraq."

The Court rejected these submissions because they were not substantiated by evidence of the kind of power and control necessary to establish jurisdiction.

The sort of argument advanced in this case calls to mind the current international expansion of criminal jurisdiction without there being a need for territorial control. There are now numerous examples. One in New Zealand law is s 12C of the Misuse of Drugs Act 1975, which provides that every person commits an offence against this Act who, outside New Zealand, does or omits to do any act that would, if done or omitted in New Zealand, constitute an offence against s 6 (dealing with controlled drugs), s 9 (cultivation of prohibited plants), s 12A (manufacturing, producing, supplying equipment or precursor substances for use in manufacturing or cultivating), s 12AB (knowingly importing or exporting precursor substances for unlawful use), or s 12B (laundering proceeds of drug offences). Consequently, courts in New Zealand have jurisdiction over the proscribed behaviour no matter where in the world it occurs. The offender need not be a New Zealander, and the offence need have nothing to do with New Zealand. Once the offender is in New Zealand territory he can be charged. See also s 7A of the Crimes Act 1961 for other examples of this sort of extension of jurisdiction.

Jurisdiction to punish does not, as these examples show, depend on the exercise of territorial control. Are rights limited by territorial control?

If an Iraqi committed in Iraq what we in New Zealand call a drug dealing offence, and then came to New Zealand, it is likely that he would enjoy the protection of our Bill of Rights, and of the International Covenant on Civil and Political Rights, to which New Zealand is a signatory. So, if evidence had been obtained against him in Iraq by methods that would have contravened his rights if he were in New Zealand, a court in New Zealand would apply New Zealand law to determine the admissibility of that evidence on a charge for the offending against s 12C.

For an analogous illustration, see the rejection by English courts of evidence obtained by torture, including torture in a foreign country, discussed in these blogs on 11 August 2004.

In this sense, then, rights do extend beyond the limits of territorial control of the State in which they may be recognised. Was the European Court wrong to require a demonstration of power and control by the respondent States?

The difficulty faced by the Applicant is that the power to try him is asserted by the Iraqi Government, the existence of which is endorsed by the UN Security Council, and that Government is not a party to the ECHR which establishes an obligation to enforce the rights on which he sought to rely. To enforce those rights, the Applicant would have to be tried by a court in one of the respondent States.

Thursday, March 16, 2006

Legality v Common sense

Should one feel sorry for a person who, having served a sentence of imprisonment, has his convictions quashed on appeal because the evidence against him should not have been ruled admissible?

A concurrent sentence of two years imprisonment for two aggravated robberies was served by the appellant "S", a 15 year-old. Then, on appeal, two High Court judges sitting together held that the evidence of his guilt, which was entirely contained in his statement to the police, should have been excluded, and his convictions were quashed: S v Police 14/3/06, Baragwanath and Heath JJ, HC Auckland CRI 2004-404-515.

In reality, unless his confessions were false, S was guilty and, no doubt, he deserved his sentence. But that is to ignore the rule of law, which in this context means that punishment can only be imposed following due process of law. If by "due process" we mean upon lawful conviction, then S did not receive due process.

The problem arises from practicalities. Sentences take effect before rights of appeal against convictions have been exercised. Trials proceed before rights of appeal against evidential rulings have been exercised. Evidential rulings during a trial must be left for appeal after completion of the trial, otherwise trials would have to adjourn to allow appeals to be heard.

The obvious course, to minimise "false" imprisonment, would be to defer sentencing hearings until appeals had been determined, but that is often impractical: many convicted people would run away, or would have to be denied bail.

If there is no viable alternative to the present procedure, should people like S be compensated? How would compensation be measured? Our assessment of the value of due process may have to be balanced against the harm caused by the offender. The courts may be forced to say that the real remedy is the vindication that follows from the quashing of the conviction: see further, my blogs of 6 and 30 March 2005.

In S v Police, a police officer had told S that he wanted him to point out where burglaries had occurred, so that files could be cleared up, and that he would not be charged with committing them. Unexpectedly, from the officer’s point of view, S said he had committed a couple of aggravated robberies. He was interviewed about those, he confessed to them, and was charged. If his initial mention of them had been a "spontaneous" utterance, his consequential dealings with the police would not have been flawed, as the correct procedures for dealing with young people had been followed thereafter. However, if the initial reference to the robberies was not spontaneous, in the sense that it was made in reliance on the officer’s assurance that he would not be charged, then it was made in circumstances that were in breach of the statutory procedures, and what followed, although correct in itself, could be tainted.

The High Court held that the utterance was not spontaneous, but rather it had been made in reliance on the officer’s assurance. There was an error of procedure surrounding that, so the question became whether the following procedures, which led to the making of the confession, were indeed tainted.

The Court found it necessary, at this point, to consider the nature of causation (para 56 – 67). This came down to asking whether, in the circumstances, the non-spontaneous utterance had been the "effective cause" of the subsequent confession. It had, and because the use of an inadmissible statement to secure a subsequent statement was contrary to the policy of the legislation (para 59), both were inadmissible.

A lot turned, in this case, on how the judges assessed the circumstances. It could easily have been decided that S had, on receiving the advice to which he was entitled, decided to make a clean breast of everything, and that that decision, rather than the earlier indication that he would not be charged, was the effective cause of his confession. In cases like this, where the assessment of the circumstances seems to be finely balanced, it is tempting (but, in law, wrong) to think that the court was influenced by the common sense solution: S had served his sentence and may as well be relieved of the convictions to give him some incentive to obey the law.

Did the absence of any compensation for the "wrongful" imprisonment facilitate this balancing of the circumstances?

Tuesday, March 14, 2006

Inferences of innocence

A commonly advanced ground of appeal against conviction is that the evidence supported an inference of innocence and the judge did not correctly tell the jury how they should deal with inferences.

We have passed through a period when much attention was given to what inferences are and how they should be handled. Things got rather complex as efforts were made to be precise about this. The relationship between the process of drawing inferences, and the process of deciding whether something has been proved, was at the centre of this complexity.

Some facts are proved by direct evidence: for example, a witness says that he saw something happen. Other facts are inferred from direct evidence: the facts that are directly proved suggest that something else is true. In turn, these inferred facts may combine, with other inferred facts, or with directly proved facts, or with a mixture of inferred facts and direct facts, to enable a further inference that something else is true. Wherever an inference occurs, it is usually described as an ordinary process of logical thought.

As to proof, it is relatively easy to see that evidence of directly observed facts may prove those facts, to the necessary standard. If the fact is an element of the alleged offence, then, when the jury considers all the evidence in the case, it must decide whether that element has been proved beyond reasonable doubt. But, in relation to other facts, (and with some exceptions), no particular standard of proof is required. The jury does not have to be instructed by the judge about the standard to which they must be satisfied that inferred facts are proved before they can use them to support other inferences, such as an inference of an element of the offence, as long as, ultimately, the jury tests the proof of the elements of the offence against the standard of beyond reasonable doubt.

So, what about the inferences supporting innocence? In what has been regarded as the leading case on inferences in New Zealand, R v Puttick (1985) 1 CRNZ 644 (CA) the Court summarised the position:

"Inference is simply one of the mental processes which may be used by a jury in carrying out its primary task of assessing the evidence and deciding whether or not it establishes the guilt of the accused beyond reasonable doubt. Where the charge has several essential elements, proof of guilt necessarily involves proof of each of those elements to the same standard. It does not, however, require proof beyond reasonable doubt of every fact which may be relevant to proof of each essential element.
… It must be equally unhelpful to tell jurors that, if proven facts support two inferences of equal weight, they should accept one and reject the other. To draw an inference either way from such facts would be pure speculation. Jurors should not be directed to accept or reject inferences when they have no logical basis for either step."


This could suggest (although we can be sure the Court did not intend this reading) that an inference supporting innocence can neither be accepted nor rejected if it is equal in weight to an inference of guilt.

A clearer account has been given in R v Seekamut 10/7/03, CA82/03:

"If on an objective basis that has regard to all the circumstances, a rational alternative to guilt is not excluded, there must for that reason be a reasonable doubt. But the mere fact that some of the circumstances might arguably permit an inference inconsistent with guilt is not enough. The jury’s function is to assess the whole of the evidence and in so doing may conclude that a suggested alternative is not reasonably tenable."

Similarly, yesterday the Privy Council in Taylor v R (Jamaica) [2006] UKPC 12 (13 March 2006), para 18, held:

"Their Lordships agree with the submission made on behalf of the appellant that in the circumstances of this case it was essential that the judge … spell out the possible inferences to be drawn … and instruct them that they must rule out all inferences consistent with innocence before they could be satisfied that the inference of guilt has been proved correct."

This could be read as suggesting a slightly different approach to that in Seekamut: instead of treating the evidence as all being in a big pool, to be considered in the round to see what ends up being proved, the process alluded to in Taylor seems to be one of deciding first whether inferences of innocence can be ruled out, then, if they are, turning to see whether guilt has been proved. We might, however, reasonably wonder whether this explication is correct, insofar as it seems to place a burden of proof on the defence, and it also seems to make an artificial distinction between stages of the reasoning process. The correct approach would be to tell the jury that, if after considering the evidence, they are left with a reasonable doubt about the accused's guilt, they must find him not guilty.

Because the judge in Taylor failed properly to direct the jury in this regard, the appeal against conviction was allowed and the case remitted to determine whether there should be a retrial.

Monday, March 13, 2006

Metaphysical laws

Can the law have been other than it was understood to have been at the time?

If it can, are we currently operating under a misunderstanding of the law?

Are there points of law that we are mistakenly taking for granted?

If we all think that the law says one thing, can we be wrong?

Can some future court say we were wrong, and that the law said something else?

The first of these questions concerned the Privy Council last week in Bowe v R (Bahamas) [2006] UKPC 9 (8 March 2006). The issue involved detailed analysis of legislation, but in essence the question came down to whether the death penalty was mandatory in the Bahamas in 1973, notwithstanding that cases had been conducted on the basis that it was.

The Privy Council answered the conundrum this way (para 42):

"… it took some time for the legal effect of entrenched human rights guarantees to be appreciated, not because the meaning of the rights changed but because the jurisprudence on human rights and constitutional adjudication was unfamiliar and, by some courts, resisted."

Further, it would be wrong to answer the question by now trying to assess what chance the argument would have had of being accepted in 1973.

Instead,

"The task is to ascertain what the law, correctly understood, was at the relevant time, unaffected by later legal developments, since that is plainly the law which should have been declared had the challenge been presented then." (ibid)

So,

"It matters little what lawyers and judges might have thought in their own minds: in the context of a codified Constitution, what matters is what the Constitution says and what it has been interpreted to mean. In 1973 there was no good authority contrary to the appellants' argument, and much to support it. In the final resort, the most important consideration is that those who are entitled to the protection of human rights guarantees should enjoy that protection. The appellants should not be denied such protection because, a quarter century before they were condemned to death, the law was not fully understood." (ibid)

At this point, it seems to be only a subsidiary consideration that there was no authority to the contrary as at 1973, as the governing matter ("the final resort") is that those who are entitled to human rights protection should be entitled to enjoy that protection. Question-begging, of course, as the issue was whether the people were (not "are") entitled to that protection.

Not to worry, for in Bowe the death sentences were quashed because the law as at 1973 was that the sentence was discretionary, not mandatory.

This interpretation was therefore carried through to the present pursuant to the legislative provisions. The cases of the two appellants were remitted for consideration of the appropriate penalty.

It seems, consequently, that we are governed by laws that we have not yet discovered, and may not be discovered until we are all dead.

Thursday, March 09, 2006

What a "fair trial" means

Today, the High Court of Australia decided Nudd v R [2006] HCA 9 (9 March 2006), a case with helpful dicta on the meaning of "fair trial". I have recently noted cases concerning what is a substantial miscarriage of justice (see Index), and have described the two ways in which miscarriages of justice may be considered "substantial" for the purposes of the proviso: by depriving the accused of a real chance of an acquittal, or by causing the trial to be unfair.

Nudd recognises that, even though the evidence of guilt may be overwhelming, there may be rare cases where there has nevertheless been a miscarriage of justice because the trial was rendered unfair to the accused: see Gleeson CJ at para 6, Kirby J at 87 and 100, Callinan and Heydon JJ at 158. The other judges, Gummow and Hayne JJ jointly decided that no miscarriage of justice had been demonstrated to have occurred (para 25). Gleeson CJ agreed with that conclusion.

Kirby J treated the outcome aspect (the depriving of the accused of a real chance of acquittal) in the same way as the other judges, holding that on any approach, there was no prospect of acquittal as the prosecution case was overwhelming, but he also gave the most detailed consideration to the other aspect, that of the fairness of the trial. He found this the more difficult point to decide (para 106), because the defects at trial (counsel’s serious incompetence, especially in not knowing the law and in failing to take detailed instructions and consequently to advise the client) were so serious. He held that this case was "borderline" on the question of fairness (para 109), but that the decisive point was that the evidence against the accused was conclusive of guilt.

This might be thought to muddy the distinction between the outcome aspect and the fairness aspect of substantial miscarriage of justice. However, I suggest that Kirby J’s reasoning is essentially that, from the point of view of fairness, the errors were inconsequential, because the defence in the trial had (and could only have) been to put the prosecution to proof, and the judge in her summing up had properly directed the jury on the law.

I have previously suggested that a fair trial is one where the law is accurately applied to the facts determined without bias. Applying this to Nudd, we can say that the errors at trial (counsel’s incompetence) did not cause the law to be applied inaccurately, nor did it cause the jury to be biased against the accused.

Thursday, March 02, 2006

Bad technique!

Sometimes, evidence that the accused has been convicted for committing other offences is admissible to prove that he committed the offence with which he is now charged. Or again, allegations that he committed other offences may be used to prove that he committed this one. Or, yet again, allegations of offences for which he has been acquitted may be used as evidence that he is guilty of this one.

This sort of evidence, called similar fact evidence, should only be allowed to be used in exceptional circumstances, for the obvious reason that it will have a tendency to bias the tribunal of fact (usually the jury) against the accused.

It must be said that the current law on similar fact evidence in New Zealand is in need of clarification. This is largely due to the decision of the Court of Appeal in R v Holtz [2003] 1 NZLR 667; (2002) 20 CRNZ 14 (CA), where the criterion for admissibility was held to be whether the probative value of the similar fact evidence outweighs its illegitimately prejudicial effect:

"One of the problems with evidence of past conduct, whether labelled evidence of bad character, propensity evidence or similar fact evidence, is the tendency to look for principles of admissibility applicable to all such evidence in all circumstances. At a very general level there is the broad principle that to be admissible the evidence must be such that its probative value outweighs illegitimate prejudice to the accused in having adduced evidence of past conduct that might be given undue weight or used improperly in reasoning towards guilt of the crime charged. It is in the application of that general principle in the circumstances of particular cases that difficulties arise. Often this is because the approach adopted in one set of circumstances is taken as the rule to be applied in a quite different case. Because of the wide variation in circumstances in which the issue calls for consideration, the preferable approach in a particular case is that which determines and weighs probative value and potential prejudice in the circumstances. It is necessary to consider the nature of the evidence, what it is sought to prove, what other evidence there is and its relationship and whether the evidence relates to prior proved offending or to concurrent charges, all against the underlying dangers inherent in propensity or bad character evidence." (para 35)

The so-called weighing of probative value against prejudicial effect is problematic conceptually, and has been criticised in the leading Australian case on similar fact evidence, Pfennig v R (1995) 182 CLR 461 per McHugh J on the basis that these were "incommensurables", and the Supreme Court of Canada has agreed with his criticism, in R v Handy (2002) 164 CCC (3d) 481 (SCC); 213 DLR (4th) 385, at paras 148-149. And I have written an article on it: "Probative value, illegitimate prejudice and the accused's right to a fair trial" (2006) 29 Crim LJ 8.

In Phillips v R [2006] HCA 4 (1 March 2006) the High Court of Australia has confirmed that Pfennig continues to be the authority on similar fact evidence, binding on Australian courts. It issued a thinly-veiled rebuke to courts that sought to find reasons for departing from Pfennig, at para 60:

"It must be said at once that it is for this Court alone to determine whether one of its previous decisions is to be departed from or overruled …. Of course, in criminal cases it is often necessary for trial judges and Courts of Criminal Appeal to elaborate upon rulings of this Court; to gather together rules expressed in several cases; to apply rules to different facts; and sometimes to reconsider rules affected by later legislation. Within spaces left by the binding determinations of this Court, trial judges and intermediate courts retain their proper functions …. However, these do not extend to varying, qualifying or ignoring a rule established by a decision of this Court. Such a rule is binding on all courts and judges in the Australian Judicature."

One misinterpretation of Pfennig, alluded to briefly in Phillips at para 62, is that it is appropriate and sufficient to regard similar fact evidence as a form of circumstantial evidence. This is what had occurred in Holtz. See Adams on Criminal Law, Ch2.8.19(5), and my blog entry for 14 October 2004. A danger of that view is that it leaves the process of drawing inferences unelaborated by particular directions to the jury, as authorities on inferences tend to avoid such assistance. It also tempts the court to admit the evidence, as circumstantial evidence is just a category of admissible evidence.

Phillips and Pfennig hold that similar fact evidence must pass a high threshold before it can be admitted (para 79):

"Criminal trials in this country are ordinarily focused with high particularity upon specified offences. They are not, as such, a trial of the accused's character or propensity towards criminal conduct. That is why, in order to permit the admission of evidence relevant to several different offences, the common law requires a high threshold to be passed. The evidence must possess particular probative qualities; a strong degree of probative force; a really material bearing on the issues to be decided. That threshold was not met in this case. It was therefore necessary that the allegations, formulated in the charges brought against the appellant, be separately considered by different juries, uncontaminated by knowledge of other complaints. This is what Pfennig and other decisions of this Court require. To the extent that O'Keefe [a Queensland Court of Appeal decision critical of the High Court of Australia's judgments in Pfennig] or other authority suggests otherwise, it does not represent the law. No other outcome would be compatible with the fair trial of the appellant."

Phillips had been charged with 8 counts of sexual offending, and had unsuccessfully sought severance of those. Joinder was permitted on the basis that the evidence on each charge was admissible on the others, as evidence of similar facts. The High Court summarised the evidence and its conclusion in this way (para 56):

"The similarities relied on were not merely not "striking", they were entirely unremarkable. That a male teenager might seek sexual activity with girls about his own age with most of whom he was acquainted, and seek it consensually in the first instance, is not particularly probative. Nor is the appellant's desire for oral sex, his approaches to the complainants on social occasions and after some of them had ingested alcohol or other drugs, his engineering of opportunities for them to be alone with him, and the different degrees of violence he employed in some instances. His recklessness in persisting with this conduct near other people who might be attracted by vocal protests is also unremarkable and not uncommon."

Bad technique, indeed, as it got Phillips to court.

The trial judge, and the Queensland Court of Appeal, had reasoned that so many similar complaints could not have arisen by chance, so together they had credibility. The High Court was not so slap-dash. It identified the issue on which the challenged evidence was relevant as being consent, and it pointed out that the non-consent of one complainant could not possibly prove that another had not consented.

One thing this case illustrates is the way judges can lose their reasoning powers in cases of alleged sexual misconduct. There is, in other words, a grave danger of judicial bias (probably unconscious bias) in such cases. In those circumstances, clear rules, rather than discretions, are needed concerning the admissibility of evidence of other alleged misconduct. The Pfennig rule is that similar fact evidence is not admissible unless there is no reasonable view of it that is consistent with the accused’s innocence. It must have strong probative force, a really material bearing on the issues to be decided; its probative force must clearly transcend its merely prejudicial effect; it must have a sufficient nexus, some specific connection, with the issues; its probative force must be such as to make it just to admit it (Phillips, para 54). None of those criteria were met in Phillips (para 55).

It is sad to note that in New Zealand, the Evidence Bill proposes an approach to similar fact evidence (or, in its terminology, "propensity evidence") that applies the probative value/prejudicial effect weighing, criticism of which is noted above. Clause 39(1) of the Bill provides:

"The prosecution may offer propensity evidence about a defendant in a criminal proceeding only if the evidence has a probative value in relation to an issue in dispute in the proceeding which outweighs the risk that the evidence may have an unfairly prejudicial effect on the defendant."

This proposal does nothing to improve our law on this point.