Yam v R.

This applicant was convicted of murder and associated counts of fraudulent misuse of the deceased’s identity and bank accounts. His application for leave to appeal against his conviction on a number of different grounds has been referred to the court by the Registrar.

The deceased was a reclusive gentleman in his eighties who lived alone in a large house in Hampstead which he had permitted to decay and in which he lived in conditions of some squalor. He was found dead there on 14 June 2006, but had plainly been dead for some weeks. The last dealing which anyone had had with him was a telephone call on 8 May, and from his condition and other evidence it was distinctly possible, although not certain, that his death had followed soon after that. He had been beaten to death, probably with some blunt instrument and/or the shod foot. There was burning, which might have been inflicted before death but also possibly afterwards, whether in an abandoned attempt to destroy the body or not was unclear. His body had then been buried under a roughly three foot high pile of large printing proofs which had been present in the room in very large quantities. Something like half a ton of these concealed the body.

The deceased had been away on holiday in the USA for the month of April. On his return he became concerned that his post had been interfered with, and that some cheques which he was expecting had been stolen. Those were for some dividends and a refund from the Revenue of overpaid income tax. He reported this likelihood to the police on 2 May. His last known telephone call on 8 May had been to the Revenue in an attempt to find out whether the refund cheque had been sent or not.

There was no doubt that between about mid May and mid June there was quite extensive misuse of the identity of the deceased and his bank accounts. Cheques payable to him were paid into an account with which he had no connection. Persistent efforts were made to extract a substantial sum from his accounts. Telephone calls were made to several of his banks in his name, but not by him. A new E-mail account was created in his name, but not by him, and used from time to time in communication with his banks. Attempts were made to apply for three new credit cards in his name. His existing credit card was used fraudulently. Furthermore, letters had been sent to the deceased’s home which were not found when the murder was discovered. Telephone calls made to the senders, in the name of the deceased but not by him, after such letters would have arrived showed that they had been intercepted. Thus there had clearly been repetitive thefts from the post, some fairly clearly after the murder had been committed.

The defendant lived two or three streets away. The Crown case against him was that he was the thief and fraudster and that the murder had been committed because at some point he encountered the deceased. The Crown case depended upon proving that he was the thief and further satisfying the jury that that fact justified the inference that he must in consequence have been the killer. In due course the case was left to the jury by the Judge, with crystal clarity, on this basis. First, unless the jurors were sure that the defendant had been the sole person making the fraudulent use of the deceased’s identity and bank accounts, there was no basis for the conclusion that he was the killer. Second, if they were sure of that, guilt of murder did not necessarily follow; they must go on to ask whether the only reasonable conclusion was that he must also be the murderer; en route to that decision, unless they could also exclude the possibility of anyone else being involved at the house of the deceased at the time of the murder, acquittal must follow.

The defendant was proved by CCTV camera evidence to have been the person who paid in the missing cheques to an account which had no connection with the deceased. He was proved to have used one of the deceased’s credit cards in a shop, in a restaurant, at a cashpoint and for an online transaction. Those transactions he admitted. But he denied all other fraudulent use of the identity or bank accounts of the deceased. As to the admitted transactions, he contended that he had been handed the cheques and credit card by one or more gangsters whom he had come to know, and that he was playing them along as a means of assembling evidence against them and reporting them. He named them as “Gaz”, Zhao Dong and Ah Ming and he gave descriptions of them, of the places they frequented and in the case of Ah Ming where that man said that he worked. They had wanted him to take part in illegal immigration and other crimes. Gaz had handed him the cheques and asked him to pay them into the bank.

That bald summary does scant justice to the very detailed evidence which was assembled on both sides, but more detail is not required except where it bears on the distinct grounds of appeal advanced.

The indictment contained six counts in all. Count 1 charged murder. Counts 2 and 3 charged burglary and theft, charging in each case the stealing of the cheques and a quantity of mail, and framed as alternatives to cater for the possible view that the cheques etc had been stolen when protruding from the letterbox rather than after entry into the house. Count 4 charged handling the cheques, the mail and the mobile telephone of the deceased. Then counts 5 and 6 charged two specific acts of fraud in the use of the identity of the deceased.

There were two trials. At the first, the jury first convicted upon counts 5 and 6, returning at that stage no other verdicts. After some little time of further retirement, it indicated by a note to the judge that no agreement could be reached on counts 1 -3 and asked whether he wished them to consider count 4. After some discussion with counsel, which included consideration of whether to do so would have any likely effect on a re-trial, the judge resolved to take any verdict which the jury could reach on count 4. In due course the jury returned a verdict of guilty on that count. At the re-trial the defence submitted that the verdict of guilty of handling on count 4 precluded any further prosecution on a basis inconsistent with that verdict, so that there could be no further prosecution for burglary alternatively theft. Since this case was conducted on the basis that to be guilty of murder the defendant had to be shown to be the thief, that meant, said the defence, that the verdict on count 4 also precluded any further prosecution upon count 1 for murder. The judge rejected that argument. It is renewed before us.

The rule of autrefois convict did not of course apply, since the second trial proceeded on counts different from those on which verdicts had been returned. But there is no doubt that quite apart from autrefois convict/acquit, English law recognises that it may on occasion be unfair to proceed with a prosecution which is inconsistent with a properly arrived at previous acquittal or conviction. The explanation of this principle as a matter of res judicata by Lord Morris in Connelly v DPP [1964] AC 1254 cannot survive DPP v Humphrys [1977] AC 1 where it was disapproved and the suggestion that issue estoppel could apply in criminal proceedings was roundly rejected. But Lord Devlin’s alternative explanation of the principle as an example of the power to prevent an abuse of the process of the court where a further trial would be unfair or oppressive we assume for the purposes of this case does survive and is consistent with the majority view in Humphrys. A slightly different aspect of the principle is derived from R v Elrington (1861) 1 B & S 688; 121 ER 870 and prevents unfair successive prosecutions of the same man on the same facts upon charges of ascending gravity. Sambasivam v Public Prosecutor, Malaya [1950] AC 1340 is now no doubt best explained as an example of the principle, where the defendant had been charged with carrying a firearm and possession of ammunition, some of the latter contained in the gun. Once he had been acquitted of possession of the ammunition it was held improper to prosecute again on the firearm charge. The key to the principle is that one must look to the substance of the case rather than its form and decide whether the further prosecution is an unfair or oppressive abuse of the process of the court. This jurisdiction differs from autrefois acquit/convict simpliciter. It involves a judgment on fairness in the light of the individual facts of the case: see Connelly per Lord Devlin at 1358 and Lord Pearce at 1364.

Count 4 was plainly an alternative to counts 2/3 and was distinctly presented to the jury as such. As the case was presented to the jury there can be no doubt, and it is not disputed, that count 4 only arose if there was an acquittal on both counts 2 and 3. In summing up the Judge correctly so directed the jury. He said this:

“This is an alternative to count 3. If you are not satisfied that D is guilty of burglary, count 2, and you are also not satisfied that he is guilty on count 3, so that the verdict on each of those 2 counts is not guilty, then consider count 4.”

Those words were also included in written directions supplied to the jury. At that stage there was no emphasis of the need for acquittal on counts 2 and 3, and given the number of counts on which directions had to be given and the relative unimportance of count 4, that is not surprising.

After the jury had been deliberating for two days it delivered verdicts of guilty on counts 5 and 6 but remained undecided on counts 1 – 4 despite a majority direction. On the third day of retirement it sent a note which indicated that it was satisfied, contrary to the defendant’s case, that it was he who had been the person responsible for the misuse of the deceased’s identity and his bank accounts. At that stage the judge very properly reminded the jury of his very clear earlier direction that whilst that decision was a necessary step to conviction of murder, it was not a sufficient one. He reminded them in some considerable detail of the arguments relevant to the remaining step, namely whether that decision did or did not leave as the only reasonable further conclusion that it was the defendant who had killed the deceased. This was the important thing to emphasise at that stage. However, in the course of this revisitation of his comprehensive summing up there was passing reference to count 4. It was in these terms:

“If you are not sure about guilty in relation to counts 2 and 3, and you may be in the light of what you have said, but it is a matter for you, and you are satisfied in relation count 4 [and] you return a verdict in relation to count 4, and reminded (sic) yourselves of the ingredients of that offence which you will find set out in the written directions.”

The transcript is plainly corrupt in small parts. The word [and] which we have placed in brackets must be an erroneous addition. At all events, what has happened, no doubt accidentally, is that this further direction omits the reference to the condition that there has been acquittal on counts 2 and 3. It leaves it open to the jury to convict on count 4 not if they have acquitted on counts 2 and 3 but if they are unable to reach a verdict upon those counts.

It had not been the defendant’s primary case that he was guilty of count 4 as charged. He had given evidence that he had not handled the deceased’s telephone and that whilst he had handled the cheques he had not done so dishonestly. But as Mr Robertson QC explained to us, he had, on the defendant’s behalf, invited the jury to say, if it rejected that primary case, that all he was guilty of was counts 4 – 6. It is therefore entirely unsurprising that the jury, finding itself unable to agree about counts 1 – 3, indicated that it had a verdict upon count 4.

After some discussion with counsel the judge resolved to take a verdict on count 4 even though the jury appeared to be undecided on counts 1 – 3. The jury, as was anticipated, said when asked that it could return neither a verdict of guilty nor of not guilty on counts 1 – 3. But when it was asked about count 4, it convicted.

We sympathise with the judge, who had to deal with this problem unexpectedly and in the context of a long jury retirement. But this verdict should never have been taken. Count 4 only arose if there was an acquittal on counts 2 and 3, but there had been neither conviction nor acquittal on those counts. The position is much the same as obtained in Fernandez [1997] 1 Cr App R 123, where the judge had omitted to direct the jury that the handling count was an alternative to the robbery, took a verdict on it when the jury had yet to return a decision on the robbery, and before they did so corrected himself by refusing to accept the verdict of guilty of handling. That course was approved by this court. The error in this case was subsequently recognised by the judge. At one stage we understood Mr Robertson to be disposed to agree also that it was an error, but his final position was that the jury, having reached a verdict, was entitled to deliver it. We disagree. On its face a conviction on count 4 involved acceptance that the defendant had received the cheques and mobile telephone from someone else. That was inconsistent with the lack of agreement on counts 2 and 3 which charged him with stealing them himself. Whether the defendant was the thief was a point on which the jury transparently had not agreed. Plainly the jury had understandably lost sight of the earlier direction that it should proceed to count 4 only if acquitting on counts 2 and 3. We have no doubt at all that it took the view that whatever else he was or was not guilty of, as to which there was disagreement, he was at least guilty of count 4, as his own counsel had suggested. It could not, however, come to that opinion consistently with the law whilst it remained undecided on counts 2 and 3.

At the re-trial it was contended for the defendant that the verdict on count 4 necessarily involved acquittal on counts 2 and 3, because it was an alternative to them and because the definition of handling involves the proposition that the receiving is “otherwise than in the course of theft”. That was an impossible contention, given the fact that the jury had transparently refused to acquit on either count, but it emphasises the fact that the verdict taken was inconsistent with the disagreement on counts 2 and 3. Before us, Mr Robertson advanced the slightly different proposition that the verdict on count 4 carried the finding that the defendant had received the cheques and telephone from someone else, and that it was an abuse of process for the Crown to seek to prove the opposite at the re-trial. That might (depending on the circumstances) have been the position if the verdict on count 4 had not been the defective verdict that it plainly was, inconsistent with the disagreement on 2 and 3. But since the verdict on count 4 was plainly inconsistent with the position on counts 2 and 3, and since it should for that reason never have been taken, it is quite impossible to say that it was unfair, or oppressive, or an abuse of process, for the defendant to be re-tried on counts 1 – 3. Rather, it would be contrary to all principle to refuse to try him when there was a clear prima facie case and the first jury had been unable to agree.

For these reasons the judge was right to reject the abuse of process application made at the outset of the second trial and the ground of appeal which asserts that he was wrong fails. There is an extant appeal against conviction on count 4. Although the defendant’s counsel have subsequently indicated that they do not pursue it, it has never been abandoned. Accordingly we quash the conviction on count 4.

Part of this trial had to be conducted in camera. The reasons for that were fully considered by the judge, who looked carefully at whether such procedure could be avoided and concluded that it could not. His decision on the point was appealed by the defendant, although not by the media, to this court which gave independent consideration to the question but concluded that the judge was right. The reasons were very fully set out in careful judgments, both open and closed, delivered by both the judge and this court.

It is of course a matter of first principles that a trial ought to take place in public unless there is very powerful reason to the contrary: Scott v Scott [1913] AC 417 and many subsequent cases. We agree that among the many reasons for that are included the public interest in the proceedings of the court being subjected to scrutiny, the maintenance of public confidence and the possible emergence of evidence which has hitherto been unavailable: see R v Legal Aid Board ex p Kaim Todner [1999] QB 966 at 977E. Exceptionally however, it may be unavoidable to hear some part of the evidence in camera, for example if that is the only way of ensuring a fair trial, if Article 2 rights would otherwise be infringed, or sometimes if State security is involved.

It is accepted that since there has already been an appeal to this court on the point, the defendant cannot now have a second attempt to challenge the decision to hear some evidence in camera. We agree, however, that it is open to the defendant to contend that, assuming the correctness of that decision, as things have turned out it has resulted in a conviction which is not safe.

The central plank of Mr Robertson’s submission that this is indeed what has occurred is this. He contends that if the evidence which was taken in private, which consisted of four witnesses plus that of the defendant, had been heard in public, there would have been likely to be significantly greater media coverage of the trial, and that there is a real possibility that additional witnesses supporting the defendant in his case would have come forward on seeing it. In particular, he suggests that there is a real possibility that witnesses would have come forward to confirm the existence and gangster characteristics of those whom the defendant blamed for the supply to him of the deceased’s cheques, credit card and banking information. Secondly, he says, there may well have been further evidence of the essentially good and non-violent past character of the defendant.

This possibility was considered carefully at the time of the decision to conduct part of the case in camera. We are unable to see that it can be more than the merest speculation. Most of the trial was conducted in public. The defendant was able to name the three persons who he said were responsible for the supply of the cheques and to give a good deal of circumstantial identifying material. The order for the taking of evidence in private had excluded that part of his evidence, expressly so that it could be heard by anyone who chose to be in court, but the defendant when he came to give evidence was unable to confine himself even for a brief period to this kind of material and so it was in the end necessary for all his evidence to be taken in private. Nevertheless, the information about the alleged gangsters was available to be put to several Crown witnesses who gave evidence in open court, including the officer in the case who was cross examined about them and about what efforts had been made to trace them. Moreover, at the first trial counsel for the defendant had made an opening statement after the Crown opening – in public - and had had the opportunity, taken as we understand it, to identify the persons on whom, on the defendant’s case, the defence turned. At the second trial a similar statement could no doubt have been made, but as a matter of trial strategy no request to do so was made. The existence of Aming was confirmed by at least one witness and other information about him was elicited. The defendant was also able to advance, in open court, a number of allegations against a prosecution witness, He Jia Jin, and to put before the jury material which suggested, perhaps without much in the way of proof but advantageously so to the defendant, that that man similarly participated in nefarious activities. This all happened twice, in two trials a year or so apart. We are unable to accept that there is a real possibility that other evidence would have emerged given further publicity and that such would have been exculpatory. In reaching that conclusion we have taken into account the enormously strong evidence, summarised below, that the defendant’s account of being involved only in very limited use of the deceased’s identity and bank accounts at the behest of others, was simply not true.

Insofar as Mr Robertson suggested that further material might also have emerged on which to cross-examine the few witnesses who gave evidence in private this was not made out. The kind of material to which he referred was available at the time and no attempt was made to deploy it.

The suggestion that additional good character evidence might also have emerged is similarly unarguable. There was a great deal of evidence of the defendant’s character, both praiseworthy and non-violent on the one hand and less good, involving a history of forgery and dishonesty, on the other. The judge summed it up very favourably to the defendant.

We do not agree that the fact that four witnesses were heard in camera gave their evidence additional standing. It is important to record that, although the evidence of these four witnesses was taken in the absence of the public, the jury heard and saw them in person, heard and saw their response to cross examination on behalf of the defendant, and heard and saw the defendant’s evidence in response. The judge’s careful summing up of this evidence, and of what the defendant said in response, demonstrates that the jury was left to assess these witnesses in exactly the same way as any other.

Insofar as it was suggested at one stage that the preparation of the defence had been unreasonably handicapped by the inability to justify to the Legal Aid authorities requests for investigation of matters bearing on the in camera evidence, that turned out to be wrong as a matter of fact. When dealing, with considerable care, with the original decision whether part of the trial needed to be in camera or not, the judge had explicitly directed that no such disadvantage to the defendant must be permitted to ensue, and that if it was thought that it did, he would himself give a suitable indication to the Legal Aid authorities that sanction for preparatory enquiries or other work should be given without the detail which would ordinarily be expected.

Lastly Mr Robertson told us that there had initially been a reluctance on the part of the Crown to disclose material relevant to the closed issues, and that as a result there had been generated for a time a tension in the relationship between the defendant and his own advisers. The disclosure seems to have come much later than it should have done. But it came in November 2007, just under a year before the second trial began. Whatever reservations we might have had about the initial delay, it is impossible to contend that it caused the conviction at that second trial to be unsafe.

The second ground of appeal is, we are satisfied, unarguable.

Mr Sullman was a postman who delivered to the home of the deceased. His evidence was that on a Friday in mid-May, thus around the 12^th or 19^th , he had been approached in the street nearby by a man who asked whether he had any post for that house. The man had said that he was the nephew of the deceased and had offered to clear away some branches which were then obstructing the pathway to the front door. It would appear, we observe in passing, that those branches seem likely to have been arranged across the path as a deliberate obstruction and thus to have been likely to have been put there by the killer as part of the steps taken to discourage finding of the body. Mr Sullman had been suspicious of the man. The man appeared half Chinese.

Mr Sullman came forward on 15 June when the body was found and made a statement that day. Two or three weeks later when Police enquiries had focused on the defendant, Mr Sullman attended a video identification procedure. There could not be a parade of physical persons including the defendant because he had left the country on 15 June, but the police compiled from computer records a collection of nine middle aged men of partly Chinese appearance and that was shown to Mr Sullman. After viewing the collection twice Mr Sullman was asked if he wished to see any again. He said that the only one he wished to see was number 1, which was, by random selection, where the defendant’s picture appeared. Then he asked to see all nine side by side, and did so. He asked the conducting policeman whether he could say if he was not 100% sure. The policeman replied that he had to decide whether he could make a positive identification or not, and that he needed to be fairly certain in his own mind if he did. As the policeman gave that response, Mr Sullman interposed to say “I’m not positive but number 1 is the only one who is like the vision in my head.” The policeman recorded that, correctly, as less than a positive identification.

The first question is whether this evidence, amounting to less than a positive identification, should have been excluded as unfair pursuant to section 78 Police and Criminal Evidence Act 1974. This kind of evidence cannot alone support a prosecution case on identification, for that would be to transform it into the positive identification which it is not, but it may, depending on the judge’s decision as to fairness, be admitted to support other evidence: R v George [2002] EWCA Crim 1923 at paragraphs 34-35. Here there was a great deal of other evidence linking the defendant to the possession of the deceased’s mail. The judge held that this was a qualified identification. Mr Robertson submits that it was no more than a comparison of the nine pictures and a conclusion that the one least unlike the person seen was the picture of the defendant. Of course he is right that the witness’ process of thought involved comparing the pictures, but that is probably true of most if not all such identification procedures. It seems to us that the judge was quite entitled to his view that Mr Sullman’s assertion that the relevant picture was “the only one like the vision in my head” was indeed a qualified identification. It was not an assertion that it was simply the least unlike.

There had been two technical errors in the conduct of the parade. Mr Sullman, because he had not made a positive identification, was not asked the question which would have followed if he had, namely whether he had previously seen any photograph of the man he picked out. We agree that in event of a qualified identification that question should be asked. But the issue was explored and Mr Sullman was clear that he had not seen any such photograph, nor were there any in general circulation at that time, so the error was of no significance for the reliability or otherwise of what the witness said. Secondly Mr Sullman and other potential identification witnesses had awaited the procedure together in one room. That should not have happened. Code D10 requires the conducting officer to make sure that witnesses are not able, before they see the pictures, to communicate with one another about the case. In this case the group was chaperoned by a policeman and there was clear evidence that no communication about the case took place, but this is a risk which ought not to have been taken. Nevertheless, this too was an error without any adverse consequences for the reliability of Mr Sullman and it afforded no ground for excluding the evidence.

Next, it is contended that the selection of pictures was unfair. The picture of the defendant showed him wearing what appeared to be a light brown leather jacket. He had a curl or lock of hair disengaged from the rest of his hair and dangling over his forehead. And he was visibly unshaven. It is said that those three features made him stand out and were likely to invalidate the identification process. Mr Sullman had, however, made no reference in his statement to any curl of hair, merely describing the hair of the man he saw as “wispy”, which on the face of it might well mean something quite different. Moreover one of the other pictures was of a man with some, albeit less, hair hanging over the forehead. More significant is the fact that Mr Sullman had described the man he saw as wearing a beige jacket and as unshaven.

We agree with Mr Ellison QC, for the Crown, that if the evidence of what occurred at the video procedure had been excluded the jury would still have heard the description which Mr Sullman had given when making his original witness statement, which included features consistent with the defendant (around 50-year old Chinese, wispy hair falling to the left) although also other descriptions which arguably were not so consistent (black hair with a fringe). We do not, however, agree that the selection of the defendants’ photograph added nothing. It was adduced in order to give some support to the Crown case that the man seen by Mr Sullman was the defendant and in ordinary understanding partial recognition of a face does add to what is otherwise the difficult business of describing features. We also agree with Mr Robertson that the judge had to assess cumulatively, or collectively, the three features alleged to make the defendant stand out, but there is no indication that he did not do so. We agree that the admissibility of this evidence needed careful thought, but the judge plainly gave it such. Like him, we have viewed the nine pictures. We are satisfied that he was entitled to take the view that the defendant does not stand out; indeed the range of facial features is a good mix of broadly similar men. There is a cream or light brown jacket on one other participant (plus some darker brown jackets on others) and one further participant is not clean shaven but has less stubble than in the picture of the defendant.

The defendant complains that when the judge came to sum up this evidence he misdescribed it as “resemblance” evidence and, in doing so, elevated it into a form of identification. That elevation was compounded, it is said, by the judge adopting the conventional Turnbull approach of separating out factors affecting reliability, such as time together, lighting and the like. We are quite satisfied that the judge fell into no error in summing up. The factors affecting reliability needed to be addressed. The use of the expression ‘resemblance’ was adopted precisely to underline the point that this was less than a positive identification; Mr Sullman was saying only that the face of number 1 resembled the man he had seen. The judge was at pains to identify the features of Mr Sullman’s evidence that might be taken to point away from the defendant, in particular his recollection of an English accent, which the defendant does not have, of shoulder length hair, and of a height less than himself at 5’8” to 5’9”. Moreover he specifically asked the jury to decide for itself whether the picture of the defendant stood out in the way suggested. The contention which appeared in the defendant’s written argument that he had erred in referring to the opinion of the inspector who selected the pictures was based upon a misunderstanding; that was necessary because the inspector had been challenged in cross examination about his selection, and the judge went on to make clear that the jury must decide the fairness of the selection for itself.

We conclude that there is on close inspection no substance in the criticism of the admission of this evidence or of its treatment in summing up.

A number of fraudulent telephone calls made to three banks in the name of or in connection with the affairs of the deceased had been recorded. Both the Crown and the defence called voice recognition experts to analyse the speakers and to give opinion as to whether they were or were not capable of being the defendant. The evidence was highly complex and took a long time. In due course the judge summed it up meticulously. The net effect of the evidence was this. All three experts were agreed that all the questioned calls were made by the same person. Dr Holmes, for the Crown, and Mr Hirson, for the defence, were experienced forensic voice experts who used a conventional scale to express their conclusions. The scale ran upwards and downwards from a centre point of “no judgment possible” via “quite likely”, likely” and “highly likely” in one direction and “quite unlikely”, “unlikely” and “highly unlikely” in the other. Dr Holmes was of the opinion that it was “quite likely” that the disputed voice was the defendant. Mr Hirson and Professor Xu, called for the defence, had produced a joint report in which they concluded that it was “quite unlikely” that it was him. Mr Hirson gave evidence in accordance with this report. Professor Xu, however, made it quite clear that he was now certain that the voice could not be that of the defendant. Rather than relying on various vocal features which had formed the principal subject-matter of the opinions of the other two experts, he relied on the level of pitch of the voice.

Having read this complex evidence and the detailed summing up, it is plain to us that it was impossible for the Crown to suggest that the evidence of Dr Holmes alone could make the jury sure that the speaker was the defendant. But that was never the suggestion. The principal evidence that the speaker was the defendant was the extremely strong evidence connecting him, and him alone, to the fraudulent use of the deceased’s identity, which we shall summarise below. The question for the jury was, as it seems to us, whether when the expert evidence was added to this, there remained any doubt about the question. However, the assessment of this expert evidence was a matter for the jury and this court, which has not heard it, is in no position to form its own conclusions, even if it were proper to do so. What matters for present purposes is that the defendant contends that the summing up unfairly denigrated Professor Xu.

The pitch of the various calls had been measured and there was no dispute about the figures. Professor Xu relied on a median figure. The usefulness of a median was supported by Mr Hirson, who explained that it eliminates unusual highs or lows. We make the assumption in favour of the defendant that there is no relevant qualification to be made to that proposition. The judge commented to the jury that one needed to be careful to avoid using a median in a way which disguises rather than illuminates the difference. That was a legitimate comment because the disputed calls varied significantly in their medians. Those to the Alliance and Leicester Bank had a median of 114, those to RBS had one of 142 and those to ING had one of 155. The control tapes attributable to the defendant had a median pitch of 98 or 99. What the judge was pointing out was that if one averages the medians one may disguise the extent of the difference between the control tapes and the questioned tapes. It was not a comment made about Professor Xu. The comment in any event favours, rather than disadvantages, the defendant, because it was the defendant who relied on the difference between the questioned and the control tapes.

However, at other points the judge certainly did comment upon parts of Professor Xu’s evidence in a manner which might result in the jury, if it agreed with the comments, treating the force of his evidence as weakened. It was a very significant feature of Professor Xu’s evidence that having agreed to a report written jointly with Mr Hirson in which (a) pitch had no significant part in the conclusions and (b) the eventual opinion was no more than that it was “quite unlikely” that the disputed speaker was the defendant, he now asserted certainty that he was not the speaker, relying on pitch. Even assuming, as we think we should, that Professor Xu was probably initially unfamiliar with the scale of expression of conclusion used by Mr Hirson, that did not alter the step change in his opinion from report to evidence. The original opinions of all three experts had been closely grouped around the mid point of assessment, one saying that the voice was, to the extent of one point above uncertainty ‘quite likely’ to be the defendant and the other two saying that it was one point below uncertainty ‘quite unlikely’ to be him. The pitch figures were not new and had been available all along. Mr Hirson had adhered in evidence to the view that pitch was not a particularly reliable indicator on the facts of this case. The judge did tell the jury that a change of view in an expert was perfectly permissible, if explained, but that it would also be necessary to consider whether a change of opinion was attributable to fresh data, fresh work or a re-think, on one hand, or to becoming more enthusiastic and less objective on the other. But this was exactly what the jury did have to consider. The cross examination of Professor Xu had proceeded largely upon the basis that his change of opinion was unexplained, despite many questions designed to elicit any reason that there might be. The same applies to the comment that whereas the other two experts had adverted to the possibility that the lower median pitch of the Alliance & Leicester calls might be attributable to the speaker attempting to sound older than he was, Professor Xu had at one point appeared to take that possibility as a definite explanation for the difference. The judge’s comment that it was dangerous for an expert to push an inconvenient fact aside too readily in favour of what may be a convenient explanation, but only a possible one, was expressed with some directness, but it was a fair comment and it did no more than raise for the jury the question whether that is what Professor Xu had indeed done. Likewise the judge’s reminder that one had to be careful of Professor Xu’s visual presentation because it only portrayed some of the calls and not those which did not so easily fit his conclusion was a perfectly legitimate, indeed necessary, reminder. This was highly complex evidence of a kind which it is not easy for a jury to evaluate. A jury needs the assistance of the judge where it can fairly, if sometimes pointedly, be given. The comments made by the judge were all based upon the evidence given by Professor Xu, especially in cross examination.

For these reasons, the challenge to the summing up in this respect fails.

On 8 May, in his last known act, the deceased telephoned HM Revenue and Customs to try to find out if his expected tax refund cheque had been sent, and thus had gone astray. By chance, the call was recorded for training purposes and the tape survived. At the second trial, although not at the first, Mr Hirson gave evidence that there were, or appeared to be, some utterances at the end of the call when the conversation between the deceased and the call centre had finished but the line remained open for a few seconds. This was an example of the expert having entirely legitimately undertaken an additional examination of the tape. Notice had been given and the Crown expert could no doubt have responded if she wished, but did not. Some of the last-appearing utterances were attributed to the deceased. One unidentified male voice was, Mr Hirson thought, part of the background noise from the call centre. But he raised the possibility that two words “alright then” and a murmur “Mm” appeared to be of lower volume and different pitch from the rest and thus could originate from someone else in the room with the deceased. Fortified by that possibility, it was suggested for the defendant that the last words spoken by the deceased on the tape, to the effect that he had only just realised that ‘things’ were due to him, might have been spoken to a companion present in the room. This possibility that there was someone else there was of potential significance to the defence because the deceased had a reputation for not permitting people to enter his house. It thus enabled the argument to be advanced that if there was another person present it might be an indication of the presence of the murderer.

The first proposition advanced to us is that this evidence destroyed the Crown case. That is no small overstatement. Mr Hirson made clear that he was doing no more than raising what he regarded as a real possibility, albeit one that he himself thought was likely. The jury may or may not have accepted that possibility. If it did, it still had to evaluate the evidence alongside the overwhelming evidence that the person misusing the identity of the deceased was the defendant. There is no basis for inviting this court to form its own view about the possibility that other parts of the taped conversation can be read as consistent with the presence of someone with the deceased, which suggestions might in any event be described as highly speculative.

The second challenge to the conviction related to this evidence arises from the summing up. Complaint is made that the judge:

The judge’s summing up of this part of Mr Hirson’s evidence occupies 9 pages of transcript. It begins with the reminder that the Crown expert could have been called to deal with it but was not. It ends with the argument which the defence derived from the evidence, clearly articulated. Overall we have no doubt that it is a careful and entirely fair summary of what Mr Hirson said and what each side had said about his evidence.

It is correct that in the course of this summary the judge observed that the jury had to be cautious about the extent of Mr Hirson’s knowledge of “whether this 86 year old man tended or did not tend to talk to himself.” That was, however, not a new possibility raised by the judge in summing up. Mr Hirson had raised it himself, albeit dismissively, in the course of cross examination, when he had answered a question by saying that he had considered all possibilities “including him being a mad individual who has conversations with himself.” The judge’s mild observation was fully justified. The scraps of words could not on any view be described as a ‘conversation with himself’ and the possibility of an elderly man (or for that matter any other) muttering briefly to himself is not a wholly fanciful one.

The judge’s other comment related to the cross examination of Mr Hirson, because it had been suggested to him that he had reached the conclusion he did on the basis of material far more exiguous than he insisted on having when advising on the attribution of disputed voice material. He had responded that the exercise of identifying a disputed voice as that of X or Y was, in his view, different from merely saying that there was a voice on the tape which was not that of the deceased. The judge reminded the jury of that response. He reminded the jury that there is a ‘big difference’ between saying that a disputed voice ‘is X’ and saying that it ‘is not Y’. But he went on to leave it to the jury to consider whether or not Mr Hirson had in fact been engaged in what was in substance the same exercise as when comparing the disputed voice tapes with the control tapes of the defendant, namely advising whether the two speakers were the same. That had been the Crown’s suggestion, and it was properly left to the jury.

We are accordingly unable to see anything in these passages in the summing up which could render the conviction unsafe.

These two grounds can be taken together.

Mr Robertson conceded before us that the evidence linking the defendant to the fraudulent use of the deceased’s identity and bank accounts was ‘compelling’. His submission is that nevertheless there was no evidence fit to be left to the jury from which it could make the step from thief and fraudsman to murderer.

Given the inevitable concession, it is not necessary to do more than summarise the evidence of the fraud, but its shape and extent is important. The deceased was last known to have been alive on 8 May, when he had made the Revenue telephone call. From 14 May determined efforts had been made to extract money from five different financial institutions with whom he had accounts. This had begun with the creation of a new E-mail account on 14 May, which was done at an internet café at 138 Charing Cross Road, near to Tottenham Court Road station, which café the defendant frequented. Thereafter contact was made, in the name of the deceased, with the Alliance and Leicester Bank, HSBC, the Royal Bank of Scotland, ING direct and Sainsbury’s, with whom the deceased had a credit card. An effort was made to pay three cheques worth about £14,000, sent by post to the deceased, into first Barclays Bank, then the Bank of China via HSBC. Instructions were given for £20,000 to be remitted from the deceased’s ING account into his current account at the Royal Bank of Scotland; that succeeded, and it was followed by instructions to pay £10,000 of the proceeds to an account at ‘Money TT’, although this second step failed. Several of the banks sent further documents to the address of the deceased as a result of being contacted in his name, and it could be seen that the fraudsman had come into possession of these documents, because he was aware of the information contained in them. The deceased’s Sainsbury’s credit card was used online, at an ATM, in a shop and at a restaurant. Applications were made in his name to American Express, Morgan Stanley and Capital One for further credit cards

There was indeed compelling evidence connecting the defendant to this fraud. It included, but was by no means limited to the following.

The defendant admitted only a very small part of this evidence, in effect where it was not practicably disputable. He admitted trying to pay in the stolen cheques at HSBC, but said that they had been handed to him by one of the gangsters with a request to pay them in. He admitted using the Sainsbury’s credit card, but said that the gangsters had given it to him. He admitted the connection to the Money TT account but said that it was all part of the efforts of the gangsters to cash the cheques. He said that the telephone call to Barclays had been on an entirely innocent topic. He suggested that his daughter’s date of birth must have been used deliberately by the gangsters, having spotted it on a photograph in his flat. Otherwise he denied any part in the fraud.

Mr Robertson’s contention that none of this entitled the jury to conclude that the defendant was not only the fraudster but also the murderer proceeds on the basis of attacking various of the more theoretical aspects of the Crown’s suggested scenarios for the mechanics of the murder. We accept that there were arguments to be advanced against some parts of them, such as the theory that the murder must have been committed in the porch, even if the deceased had indeed surprised the thief interfering with his post. But such hypothetical scenarios are often ventured, and they may or may or may not end up convincing when the whole of the evidence is examined. If some details turn out not to be, that does not mean that there is no case at all to answer. Similarly, there were certainly arguments to be advanced for the defendant. There were cigarette ends in the room next door to the one where the body was found, which enabled there to be the suggestion that there had been a number of people there, although they could not of course be dated. We agree also that the defendant was able to say that there was no scientific evidence linking him to the murder, and that a killing could be said to be out of character with what was known of him, and quite a lot was. All those arguments were, however, for the jury. The evidence that he was the fraudster, the fact that he denied what could be proved, and, most of all, the evidence that he had regularly used (in three different handsets) the sim card which must have been stolen at the time of the murder, all created the plainest possible case to answer. The contrary is, we are satisfied, unarguable.

We have asked ourselves whether the state of the evidence in this case leaves us with a lurking doubt about the safety of the conviction. We have reminded ourselves of Mr Robertson’s complaint that the police approached the investigation after a certain point with the assumption that the defendant was responsible, and that there were other enquiries which might have been made but were not. In an investigation of this size, it is almost always possible to find some such possible enquiry. The evidence was long, detailed and thorough. Its assessment was for the jury and not for us, but we have no doubt that the jury was entitled to conclude that it demonstrated that the defendant alone had been the fraudsman, that his suggestions that his actions had been motivated only by trying to acquire evidence against gangsters was untrue, and that he must have been the killer. We are not left with any doubt about the safety of its conclusion.

The judge’s direction as to circumstantial evidence was impeccable. We do not agree, for the reasons already given, that the judge ought to have directed the jury that it could not convict of murder unless sure that the Crown’s hypothetical scenario was accurate. What it had to be sure of was that the defendant was the killer, not exactly how the murder had come about. That the defendant had to assert that he must have been shadowed by, and framed by, the gangsters does not alter that position.

Mr Robertson asks us to say that the summing up did not sufficiently get across to the jury the defence which the defendant was advancing, namely that whatever he had done, if proved contrary to his case to have been done, had been done because he was using his initiative in an effort to assemble evidence against the gangsters and report them. We have re-read the summing up with Mr Robertson’s stance very much in mind, but we do not agree. The ‘gangster’ explanation was of course not an arguable defence to murder; it was a case advanced to explain the apparent associations with the fraud, and if it might have been true it went to draw the sting of the Crown’s assertion that the fraud explained the reason for the murder. But all of that was made completely clear to the jury in a summing up which was long and thorough.

For the reasons set out we entertain no doubts about the safety of this conviction. We give leave to appeal, albeit the grounds which persuade us to do so are limited to grounds 1, 3 and the arguable combination of grounds 5 and 6, but we dismiss the appeal against conviction.

The sentence of life imprisonment was mandatory on conviction of murder. The relevant statutory starting point for the minimum term under Schedule 21 of the Criminal Justice Act 2003 was 30 years because this was a murder committed for gain. It was aggravated by a plain intention to kill, and by the age of the deceased, although it was not a case of his being targeted by reason of his age. There was substantial mitigation. First there was a lack of premeditation, since the judge accepted that the defendant had not expected the deceased to be present. Second, although there was some limited evidence that the defendant had recently resorted on occasion to the forgery of financial documents, this had been more for amour propre or perhaps satisfaction of fantasy than with intent to cause significant loss to others. It was far outweighed by the defendant’s positive good character, about which the judge heard clear evidence. The judge reduced the minimum term to 20 years. That was a very substantial reduction, and was merited, but it is impossible to say that it was wrong because the reduction should have been yet greater.

The judge passed a concurrent sentence of four and a half years for the burglary. That fell to be assessed on the hypothetical basis that the defendant had been the fraudsman but not the killer. The theft was of mail, from the owner’s home. It was done with a view to wholesale manipulation of the victim’s identity and bank accounts, which was the carried out over a period of three weeks or so. There is nothing arguably wrong with four and a half years, after trial, for such a burglary.

For those reasons, the application for leave to appeal against sentence must be refused.

There has been no occasion for this court to give any part of its judgment in closed form. This judgment may be reported without restriction. But in order to confirm the position, we reiterate the order which we made at the outset of the hearing. There were orders made by Ouseley J at the trial, which prevented any publication which could reveal the grounds or reasons for excluding the public from part of the proceedings, or any evidence, submissions, judicial decisions or other matter heard or dealt with in camera. Those orders are to remain in force. They are attached hereto, together with an informative explanation. Similarly there must be no publication of anything which could reveal the grounds on which a limited part of the proceedings before us was conducted in camera, or anything which occurred during that part of the hearing before us.

This order to remain until further order.”

“Informative: Media organisations are reminded that purporting to reveal or speculating as to what was or may have been said in camera may be an attempted contempt of court, punishable as a contempt: A [2006] 1 WLR 1361.”