Lord Justice Gross :
- This is the judgment of the Court to which each member has contributed.
- Procedural history: Two considerations loom large in these proceedings. First, the duty of an expert to the Court, in particular to stay within the area of his expertise. Secondly, the sole test for this Court when deciding whether to allow or dismiss an appeal against conviction: namely, whether that conviction is unsafe.
- The Appellant, Alex Pabon, now aged 39, together with a number of co-defendants, all employees of Barclays Bank PLC ("Barclays"), Peter Charles Johnson ("Johnson"), Jonathan James Mathew ("Mathew"), Stylianos Contogoulas ("Contogoulas"), Jay Vijay Merchant ("Merchant") and Ryan Michael Reich ("Reich"), faced a single Count of conspiracy to defraud, alleging that they dishonestly rigged LIBOR (as defined below).
- The Particulars of the Offence on the Indictment alleged that the Appellant and co-defendants:
" ….between 1st June 2005 and 1st September 2007 conspired together and with other employees of Barclays PLC and its associated entities (Barclays) to defraud in that:
1) knowing or believing that Barclays was a party to trading referenced to the London Interbank Offered Rate for US dollar (Dollar Libor);
2) they dishonestly agreed to procure or make submissions of rates by Barclays, a panel bank, into the Dollar Libor setting process which were false or misleading in that they:
a. were intended to create an advantage to the trading positions of employees of Barclays; and
b. deliberately disregarded the proper basis for the submission of those rates
thereby intending to prejudice the economic interests of others"
- On the 29th June 2016, in the Crown Court at Southwark, before HHJ Leonard QC, the Appellant was convicted (by a majority of 10 to 2) on the Count in question. On the 7th July 2016, before the same Court, the Appellant was sentenced to 2 years and 9 months' imprisonment.
- As to the co-defendants: Mathew and Merchant were convicted on the same Count. Mathew was sentenced to 4 years' imprisonment and Merchant to 6 ½ years' imprisonment. Both sought leave to appeal against conviction and Merchant against his sentence as well. These applications were dealt with by a different constitution of this Court (Lord Thomas of Cwmgiedd CJ, Dingemans and William Davis JJ): R v Merchant  EWCA Crim 60;  1 Cr App R 11. In the event, Mathew's application for leave to appeal against conviction was refused, as was Merchant's appeal against conviction. Merchant, however, succeeded in his appeal against sentence, which was reduced to 5 ½ years' imprisonment. Johnson had earlier pleaded guilty to the same Count and was sentenced to 4 years' imprisonment. The jury were unable to agree verdicts in respect of Contogoulas and Reich; they were re-tried in April 2017 ("the retrial") and were acquitted – a matter to which we shall return.
- The Appellant's application for leave to appeal conviction was refused by the Single Judge and was not renewed before the full Court. No formal notice of abandonment was served, pursuant to the Criminal Procedure Rules ("Crim PR"), Part 36, r.13. Before us, the Appellant seeks to renew his application for leave to appeal, some seven months out of time, relying essentially on fresh evidence arising out of the retrial. The sole focus of the intended appeal concerns the conduct of an expert witness, called by the Prosecution (i.e., the SFO) at the Appellant's trial and the retrial, a Mr Saul Haydon Rowe ("Rowe"). At the retrial and following cross-examination on new material, not available at the Appellant's trial, Rowe fared disastrously. As already noted, both Contogoulas and Reich were acquitted.
- LIBOR: LIBOR is the shorthand for the "London Inter-Bank Offered Rate". It is a global benchmark interest rate for many types of financial transactions. LIBOR is set in London and is based on the rate of interest banks charge one another for loans of funds or, put another way, the interest rate at which banks could borrow money from each other on a particular day. The rate is published daily shortly before noon. The US$ LIBOR rate was calculated on the basis of daily submissions from a panel of 16 contributing banks; Barclays was on that panel. At the material times, the LIBOR setting process was run by the British Bankers' Association ("BBA"), a trade association, not a regulator. The 16 panel banks were asked "the LIBOR question", namely:
" At what rate could you borrow funds, were you to do so by asking for and then accepting inter-bank offers in a reasonable market size just prior to 11 am?"
- The responses to the LIBOR question were collected by Thomson Reuters and the LIBOR rate for the day was calculated by "trimming" off the highest 4 and lowest 4 submissions and then working out an average of the remaining middle 8 submissions. Panel banks were to make their submissions by 11.00 and those submissions could not be seen by other banks during the submission window. LIBOR submissions would be made in respect of 15 separate time periods: overnight, 1 week, 2 weeks, 1 month, 2 months and so on to 12 months. The present case was mostly concerned with 1 month ("1m") and 3 month ("3m") periods or "tenors".
- It may be noted that LIBOR was not based on actual transactions on the inter-bank loan market. Instead, as explained by Professor Ronald Anderson, Emeritus Professor of Finance at the London School of Economics, the other expert witness called by the prosecution and to whose evidence no controversy attaches, it is an estimate of the rate that a large, well-established bank would be charged if it were to borrow funds from a similar bank.
- The present case is concerned with derivative (or swaps) trading, where the amount to be paid by a party to such a contract was derived from the published LIBOR rate. In every derivative trade referenced to LIBOR, there would be two parties. Movement in LIBOR would result in detriment to one of those parties to an equal and opposite extent as it profited the other party.
- LIBOR has been widely adopted, worldwide, as a benchmark for wholesale money market transactions. Over time, however, it has also been adopted for use in retail financial contracts as well, for example, variable rate mortgages where the borrower's interest rate is referenced to LIBOR. According to Professor Anderson, one of the most widely quoted estimates of the size of LIBOR based financial contracts is US$300 trillion.
- Dramatis Personae: The defendants at trial fell into two categories: traders and LIBOR submitters. The derivatives (or swaps) traders in the Barclays New York office were the Appellant, Merchant (to whom the Appellant reported) and Reich. There was, of course, a reporting structure or chain extending upwards from Merchant; more senior Barclays personnel included Messrs. Harrison, Bommensath and Bagguley. Contogoulas was a US$ fixed income swaps trader, based in London.
- The Appellant had been employed by BNP Paribas between 2002 and 2004 as a trader in US$ short-term interest rate products. When he joined Barclays in March 2004, his first task was to develop a model to price the US$ short-term interest rate market. It comprised a set of mathematical curves in graph form that could be used to calculate and price where the LIBOR should be. The model could be used to predict where the LIBOR would be on any given future date and was to enable traders to price the market more accurately.
- Turning to the LIBOR submitters in the Barclays London office, Johnson was the senior LIBOR submitter and was the Director of US$ Money Markets. Matthew was a LIBOR submitter and a junior trader on the Money Markets desk. Johnson was Matthew's line manager and had introduced him to the LIBOR submission process; there was no formal training.
- The Money Markets or cash desk had a "Treasury function" within the bank and was responsible for balancing the bank's books each night. The desk's task was to ensure that the bank had sufficient cash to do its business; this involved lending money to departments within the bank and sometimes borrowing from or lending to other banks. Those on the cash desk were ideally placed to answer the LIBOR question each day as they dealt in cash and borrowed funds from other banks. There was evidence that in some banks, trading desks also acted as LIBOR submitters and, in other banks, the swaps traders and LIBOR setters sat on the same desk.
- The rival cases and the issue for the jury at the trial: The SFO case was that the defendants defrauded their counterparties to their LIBOR referenced trades; they dishonestly agreed to procure or make false or misleading LIBOR submissions. The traders requested the LIBOR submitters to submit false rates and the submitters in turn provided rates that did not honestly and genuinely answer the LIBOR question. The manipulation of the rates was undertaken in order to increase the traders' profits or decrease their losses. Acceding to the traders' requests gave the LIBOR submitters status and standing within the bank. The traders were able to make larger profits for their desk, ultimately increasing their bonuses, their prospects of career advancement and their own status within the bank. The smallest movement in the published LIBOR rate was capable of directly affecting the profit or loss of the bank.
- The SFO identified some 120 requests from traders to the LIBOR submitters, Johnson and Matthew, in London. By way of examples relating to the Appellant, there were the following:
i) The Appellant sent Johnson an e-mail on the 28th November 2005, concerning the 1 month fixings, stating "1350 contracts. We need high one month, we need to get kicked out, 1350 eurodollar contracts";
ii) The Appellant sent Contogoulas e-mails, saying, "Tell PJ [Johnson] to keep LIBOR low" and, on the 29th December 2005, "Need low three month and high one month. Tell him that is the carry on the OIS [i.e., Overnight Index Swap]";
iii) The Appellant sent (the wrong) Johnson at the bank an e-mail, saying "We see three month LIBOR at 5.2225. Anything above that would be great".
- The Appellant's defence was that he only made requests that were in the commercial interest of his trading book and that would benefit his team and the bank. Indeed, in his original Grounds of Appeal, the Appellant's very first submission involved averring that he "sought to move rates in a way that favoured his book". However, he did not agree to procure or make submissions of rates to the LIBOR setting process that were false and misleading. The requests he made were in line with the derivatives market and the calculated rates on the curves of the model he had created. He denied acting dishonestly.
- The Appellant contended that the practice of asking the submitters to put in a rate which suited the traders was so widespread throughout the trading floor that senior management must have been aware of it or condoned it. It was reasonable to believe that this was acceptable practice and not thought to be improper. The Appellant had learnt the practice from his mentors, Merchant and others, who were themselves manipulating the rates. Further, the Appellant would not have used e-mail and telephone as communication channels for dishonest purposes as it was understood that such communications were monitored for compliance purposes.
- The central issue for the jury was, accordingly, dishonesty, essentially posed by the Judge as follows:
i) Whether the defendant in question agreed with one or more employees of Barclays to procure or make submissions of a LIBOR rate which was not the bank's genuine perception of its borrowing rate but was at a rate which was intended to advantage the trading position of an employee or employees of Barclays; and, if the jury was sure of this
ii) Whether or not that defendant was acting dishonestly and appreciated that what he had agreed to do was dishonest.
- Two points may be noted as to the terms of the Judge's direction. First, the requirement in the first paragraph that the estimate be genuine was plainly correct in law, a matter since confirmed by the decision in R v Merchant (supra), at  – . As observed in Merchant (at  - ), it was difficult to see how a benchmark could be set if the submission was made on any other basis.
- Secondly, as to the second paragraph, it may be seen that the jury were properly directed, as the law was then understood, in accordance with R v Ghosh  QB 1053, including the second, "subjective" leg of the test for dishonesty. By contrast, today, in the light of Ivey v Genting Casinos (UK) Ltd  UKSC 67;  3 WLR 1212, that second leg of the Ghosh test has been disapproved as not correctly representing the law. As Lord Hughes JSC observed (at ), the "principal objection" to the second leg of the Ghosh test was that "…the less the defendant's standards conform to what society in general expects, the less likely he is to be held criminally responsible for his behaviour". Instead, Lord Hughes summarised the correct approach as follows (at ):
" ….When dishonesty is in question the fact-finding tribunal must first ascertain (subjectively) the actual state of the individual's knowledge or belief as to the facts. The reasonableness or otherwise of his belief is a matter of evidence (often in practice determinative) going to whether he held the belief, but it is not an additional requirement that his belief must be reasonable; the question is whether it is genuinely held. When once his actual state of mind as to knowledge or belief as to facts is established, the question whether his conduct was honest or dishonest is to be determined by the fact-finder by applying the (objective) standards of ordinary decent people. There is no requirement that the defendant must appreciate that what he has done is, by those standards, dishonest. "
It is therefore apparent that the jury were directed, on the key issue of dishonesty, on a basis more favourable to the Appellant than if he was tried today.
- By their verdict, the jury decided the issue of dishonesty against the Appellant.
- Procedural skirmishing: Prior to the hearing before this Court, some time and effort was consumed in procedural skirmishing. First, the Respondent (the SFO) resisted disclosure of certain material arising from or relating to the retrial. That material was plainly disclosable (subject to suitable case management measures) and formed the subject of our order dated 20th November 2017. To reiterate a point emphasised elsewhere, the test for disclosure does not turn on ultimate success in the proceedings: see, R v Gohil  EWCA Crim 140, at ; yet that erroneous touchstone appeared to be the only conceivable ground for the SFO resisting disclosure. No more need be said of that dispute.
- Secondly, the SFO sought to resist the Appellant obtaining permission to advance new grounds of appeal, adducing fresh evidence in support of those new grounds and an extension of time to do so. As we understood it, the SFO additionally contended (in writing, if much more faintly orally) that the correct procedural course was for the Appellant to seek recourse from the Criminal Cases Review Commission ("CCRC").
- We can dispose of the CCRC contention at once. Though the Appellant had not sought to renew his application for leave to appeal, following initial refusal by the Single Judge, the appeal had never been abandoned. Nor had it been finally determined by this Court. In the circumstances, it could only be in an exceptional case that recourse to the CCRC would be appropriate: see, s.13 of the Criminal Appeal Act 1995. This is not such a case.
- The remaining matters (new grounds of appeal, adducing fresh evidence and an extension of time) can never be taken for granted. However, the startling manner in which the concerns as to Rowe came starkly to the forefront of the retrial made it plain that the justice of the matter required the granting of the Appellant's various applications. We did so and also granted leave to appeal.
- The rival cases before this Court: For the Appellant, Mr Allen QC submitted that it was, or ought to be, common ground that Rowe's conduct fell far below the standards expected of an expert witness in many different ways. It was wrong for the SFO now to seek to downplay the significance of Rowe's evidence; indeed, before the trial, the prosecution had successfully resisted a defence application to exclude it. Rowe had given expert evidence in two previous LIBOR trials and the prosecution had (overall) spent over £400,000 in payments to him. At the trial, in the absence of the fresh (disclosure) material, the Appellant's counsel had been inhibited in his ability to cross-examine Rowe as to his credibility or expertise; he had understandably limited any such cross-examination, not having the necessary "ammunition" to do so. That picture had been radically altered by the fresh material. It was now clear that Rowe's failings as an expert were extensive; thus:
"….he had signed documents stating that he had complied with his duties when he knew he hadn't; he had failed to report with any detail or accuracy as to how he reached his opinions; he secretly consulted with a number of undisclosed advisors; he blatantly disregarded the directions of a trial judge during the course of a criminal trial; and he knowingly gave evidence about matters outside his area of competence. These are deeply troubling failings that bring the system of justice into disrepute… "
The fresh material would have permitted devastating cross-examination. It did do just that at the retrial, where both defendants were acquitted. The duty resting on an expert witness is so fundamental that where it is abused, the entire process is affected:
"….It leads to the peculiarity of a trial in which the prosecution seeks to prove the dishonesty of a defendant and in order to do so calls a dishonest expert as an essential building block of their case.
…There may be circumstances in which that state of affairs does not impinge upon the safety of the conviction. But this was not one of them."
The appeal ought, accordingly, to be allowed.
- For the SFO, Mr Hines QC, after (if we may say so) a somewhat grudging concession as to Rowe's failings in his skeleton argument, accepted that Rowe had not complied with his duties as an expert. The SFO had itself had no inkling of Rowe's want of expertise (other than his lack of trading experience, known to all at the trial); this was his third "outing" in LIBOR trials. That said, Rowe's evidence went to largely agreed background matters, in particular providing a human face to talk to his 54 page slide presentation dealing with general banking concepts and some of the vocabulary which featured in the defendants' e-mails. The single issue at trial concerned the Appellant's dishonesty; even now, the Appellant had not identified any area of Rowe's evidence which was fundamentally wrong or of any significance to the case advanced by the Appellant at trial. Other witnesses were available at trial who were in a position to assist the jury as to more detailed aspects of trading and LIBOR setting. The Appellant had selected Mr Bagguley (a more senior Barclays executive, referred to above) as having sufficient expertise and insight to be cross-examined about the proper interpretation of the terms used in the Appellant's e-mails. The only potentially controversial evidence proffered by Rowe was a statement in his supplementary report which was excluded by the Judge (see below). Accordingly, and simply put:
"…there was no live issue pursued by the Prosecution arising from Rowe's evidence. Rowe's evidence was the uncontroversial backdrop to the case."
The position at the retrial was very different. Neither the outcome of the retrial nor Rowe's failings cast any doubt on the safety of the Appellant's conviction. The appeal should be dismissed.
- As foreshadowed at the outset, this appeal hinges on two principal Issues:
i) Issue I: The Rowe evidence;
ii) Issue II: The safety of the conviction.
We turn directly to these Issues and take them in turn.
ISSUE I: THE ROWE EVIDENCE
- (1) History: The trial: On the 13th February 2014, Rowe was instructed by the SFO ("the letter of instructions") to provide an expert report "explaining the workings of an investment bank, inter-dealing brokerage and related financial instruments and trading terms used by individuals within these institutions". As background, Rowe was informed that the SFO was investigating allegations that "between 2005 and 2010, LIBOR was dishonestly manipulated across a number of different currencies and tenors". The instructions were thus generic, to the forthcoming series of anticipated LIBOR trials. Included amongst the areas Rowe was instructed to cover were an overview of the trading floor and an explanation of the different types of traders within an investment bank and their functions. The letter of instructions drew specific attention to the duty of an expert "to give objective unbiased opinion on matters within their expertise", together with the relevant provisions of the Crim PR.
- On 26th November 2015, the SFO informed the defence of the instruction of Professor Anderson and Rowe in the forthcoming trial. This letter informed the defence of the materials supplied to the two experts and the sites and materials consulted by the experts. The present trial was the third LIBOR trial in which Rowe had been instructed. In the event, his principal report ran to 121 pages.
- In March 2016, the defendants in the trial sought a pre-trial hearing to exclude or restrict Rowe's evidence, on the ground of lack of expertise. Although it was accepted that Rowe had some general banking experience, it was contended that he had no direct experience of interest rate derivatives trading, cash trading or making LIBOR submissions. His evidence ought to be restricted to the "core" of his report, covering the structure of banks, financial concepts and an overview of the relevant financial instruments in the case. Rowe, it was submitted, should be prevented (inter alia) from giving "inadmissible evidence as to the permissible approach to the LIBOR-setting process, the permissible extent of communications between traders and LIBOR submitters, or whether conduct of any kind is or could be regarded as being dishonest". Rowe was not competent to provide opinion evidence on such matters. He had never worked as an interest rate derivatives trader, a cash desk trader or a LIBOR submitter and appeared to have no direct knowledge of the LIBOR submission process. He had not worked as a trader of any kind since 2000 and, from 2005 onwards, had acted as a professional expert witness on general banking disputes. Particular exception was taken to various Powerpoint slides.
- Resisting the defence application, the SFO response was that, as shown by the declaration in his expert's report, Rowe understood his duty to the Court, including as to expressing his opinion (only) on matters within his expertise. The SFO pointed to the fact that he had given expert evidence in two previous LIBOR trials, apparently without challenge to his expertise. Nor had there been any challenge from the defence to the accuracy or reliability of the evidence proposed to be adduced by Rowe. It would be open to the defence to cross-examine Rowe at the trial on his alleged lack of expertise.
- The matter came before the Court on the 21st March 2016. On the 22nd March, HHJ Leonard QC (the designated trial Judge) gave a written ruling. He indicated that, insofar as certain of Rowe's "graphics" sought to answer the question of whether the defendants were acting dishonestly, the SFO had indicated that it would not lead that evidence. The Judge observed that there was no real dispute as to the principles governing the admissibility of expert evidence. The essential question was whether the witness was skilled, rather than the way he came by his skill. The Judge constructively encouraged the production of an agreed (or non-contentious) glossary of banking terms – and that in fact happened, with the glossary helpfully available to us. Aside from various specific items, the Judge did not accede to the defence application. For present purposes, the key passage in the Judge's ruling was as follows:
"7. I cannot assess his [Rowe's] expert knowledge against that of any expert to be called by the defence, because the defence do not intend to call any expert evidence. Whether, as a result of cross-examination, deficiencies in his knowledge become exposed will have to wait for the trial process to resolve. Whether he is right in what he says, I cannot at this stage say, and in any event I should not usurp the task of the jury in this regard.
8. I do not consider that I am in the area, as Mr Darbishire QC [for Reich] would put it, of hearing an expert in orthopaedics give evidence about cardiology, but rather that I have a witness who has an expertise in banking matters and, no doubt, more so in some areas than others. This is something that can be tested by the defence and the jury can decide what weight they should give to his evidence in respect of any particular point."
On the material available to the Judge at the time, we cannot fault his wholly understandable, clear and careful ruling.
- In an Addendum to his report served on the 31st March 2016, the SFO and Rowe sought to introduce some further material based on an answer he had given in re-examination during the first LIBOR trial, reading as follows:
" It is my opinion as an expert, having been in and around investment banks, that a trader in fixed income, which I used as a catch-all term for a trader trading interest rate sensitive products including cash, derivatives and Libor-sensitive products would know that Libor is an 'independent' rate which needs to be set without influence from traders who are not submitters.
This is because traders in major banks understand their markets, and markets which affect their markets, in great detail and it is my opinion that as part of their training, research and general market knowledge, they would have learned that Libor had to be set independently of their views.
This is my view as an expert…..it would have been to all intents and purposes impossible for a fixed income trader….in a major bank to have been unaware of this fact."
- Almost presciently, the Judge excluded this evidence. His conclusion was expressed in a written ruling dated 13th April 2016:
" ….the lack of any detailed analysis by Rowe in his report as to the basis for his opinion and, without any research being apparent into what was in fact happening at Barclays at the time on which he could have relied, leads me to rule that, as the report is presented, there is no admissible basis for his evidence on this issue."
- At the trial, Rowe gave evidence on the 13th and 14th April 2016. He ran a company providing expert consulting and testimony in banking cases. He had worked in the finance industry between 1989 and 2000. He gave evidence concerning various banking terms and concepts. He was asked to comment upon e-mails sent between the defendants and the terms used in those e-mails. He utilised slides to demonstrate concepts such as the interrelationships and functions of various desks in banks and the categories of trader and the financial instruments within which they were concerned. As summarised in the Appellant's Application to serve additional grounds of appeal against conviction out of time:
" The prosecution had called the evidence of Rowe for a number of reasons. The principal one was to explain banking to the jury at increasing levels of sophistication. The second and more focussed one was to provide assistance with STIR [Short Term Interest Rate] trading and the emails relating to it. In the course of this exercise many topics were covered, such as DV01 [The change in Dollar Value of a contract or book resulting from a 1 basis point move in the interest rate – a measure used by traders to assess risk], curves, stub, resets, risk and the importance of tiny movements in Libor….."
As we understand it, however, Rowe's evidence did not form the basis of any cross-examination of the Appellant.
- Before leaving Court on the 13th April, the Judge gave Rowe the usual admonition, "You know not to talk about this case to anyone whilst your evidence is in progress". Rowe would of course have known that he would face cross-examination on the 14th April and, on that day, he was cross-examined by Mr Davies, for Merchant, by Mr Allen QC for the Appellant and, thereafter, by Mr Darbishire QC for Reich. Mr Allen QC explained to us that he curtailed his cross-examination; that was a decision of prudence, based on the material then available to him.
- By the time of the summing-up, Rowe's evidence was clearly perceived by the Judge to be of relatively little moment. Thus, he was mentioned by the Judge early on, in the course of giving the "standard" direction as to expert evidence:
" ….you have heard from two experts, Anderson and Rowe, who gave evidence on behalf of the prosecution about banking concepts and practice relevant to this case. Expert evidence is permitted to assist you with matters which are likely to be outside your experience and knowledge. As with any evidence, it is for you to decide what you accept and what you reject….
You should take account of their qualifications and to what extent their practical experience equipped them to give evidence on the topics they were asked about.
You should remember that this evidence relates only to part of the case and that whilst it may be of assistance to you in reaching your verdicts you must reach your verdicts having considered all the evidence.
You have also heard about banking practices from a number of witnesses, including the defendants, who were not called as expert witnesses but because they have worked in the areas of banking relevant to this case [and] can be expected to know about the particular area in which they operated….."
- As far as we can ascertain, there was only one other reference to Rowe in the summing-up, days later, when the Judge said this:
"…Rowe, another expert called by the prosecution, showed how much a difference of one basis point on LIBOR would affect a loan of $1 billion. If a trader has ten such swaps, which is a realistic amount for one trader, then it would create a cumulative profit or loss of $500,000."
- The retrial: The retrial saw dramatic developments concerning Rowe's evidence. The matter is most helpfully introduced through the witness statement of Mr Kuhn, dated 24th May, 2017 (forming part of the fresh material), a legal representative acting on behalf of Reich at the retrial.
- At the retrial, Rowe began his evidence on 9th March 2017. Pressed on the 9th and 10th March for disclosure by the defence, Rowe provided e-mail correspondence revealing that Mr Dominic O'Kane, a partner at Rowe's firm and a part-time Professor of Pricing and Risk Financial Derivatives, had been responsible for drafting sections of Rowe's report. This was not previously known.
- Furthermore, as Mr Kuhn recorded, the newly disclosed material revealed:
"5. ….that prior to April 2016 Mr Rowe had sent excerpts of the case papers to Ms Signe Biddle an interest rate derivatives trader and financial consultant at RBS and Mr Michael Zapties, Head of Rates Trading at HSBC and sought their assistance.
6. On Friday 10 March 2017 I contacted both Ms Biddle and Mr Zapties. I spoke with each by phone on 15 March 2017. They both said that they had been contacted by Mr Rowe and that in the course of their respective conversations he had neither told them that he was acting as an expert witness in a criminal trial, nor had he explained to them the caution which they should exercise in expressing an opinion.
7. On the weekend of the 11 March 2017 Mr Rowe provided further material to the SFO, in the form of text messages between Mr Rowe, Ms Biddle and Mr Zapties…..
8. From the additional disclosure it appeared that in the month prior to his giving evidence in the 2016 trial, Mr Rowe exchanged around 60 text messages with Ms Biddle and 27 text messages with Mr Zapties, as well as numerous emails with both them and Mr Nick Van Overstraeten, a third expert."
- It may be noted that these exchanges included reference to STIR trading, a technical area which Rowe purported to cover when giving evidence.
- So far as concerns the exchanges between Rowe, Ms Biddle, Mr Zapties and Mr Van Overstraeten, matters do not end there. A remarkable feature of the newly disclosed material was the revelation that at the conclusion of the first day of his evidence at the trial – and having been expressly warned by the Judge not to discuss his evidence until it was concluded (see above) – Rowe went on to do just that.
- Within an hour, he sent a text to Ms Biddle, asking her to do "30 minutes (paid!) work tonight", explaining that he needed an "interpretation of a STIR/OIS [Overnight Index Swap] email". In the course of further exchanges with Ms Biddle, Rowe said "I don't know the usual trades STIR people put on but I am learning". Ms Biddle emphasised the importance of context, to which Mr Rowe agreed but added "…it doesn't help when I have to explain a few emails and look knowledgeable". On the same night, he remarked to Mr O' Kane that he would do no more STIR cases; they would engage a STIR specialist, "as the mission has crept beyond where it was meant to for me". In cross-examination at the retrial, he explained this by saying that he was "at the edge" of his expertise.
- Overall, between the conclusion of his evidence on the 13th April 2016 and resuming his evidence on the 14th April, Rowe exchanged some 26 texts or e-mails with Ms Biddle and Mr Zapties. When questioned on the 14th April, he made no mention of his contacts with Ms Biddle, Mr Zapties and Mr Van Overstraeten, notwithstanding, put at its lowest, their important contribution to the answers he gave.
- Inevitably, Rowe was subjected to damaging cross-examination (by Mr Darbishire QC) at the retrial on this rich seam of material, including his duties as an expert and the declaration in his report, required by the Crim PR. His answers were such as to prompt the following exchange with the Judge:
" Q: Are you really saying that when you signed off the declaration which I suspect is in standard form, you hadn't in fact read either the CPR or the booklet?
A: I don't think I could have read them fully….
Q: Did you read them at all?
A: I'm pretty sure that I glanced at something."
- There was, likewise, equally damaging cross-examination as to Rowe's expertise, including this passage:
" Q: What you did in 2016 was to start pinging out emails and texts to people, passing on the material you had been provided with by the SFO and saying to people: can you help me to understand it because I don't understand it? That's what you did isn't it?
A: So what else am I supposed to do as an expert?
Q: Say it is not my field; I cannot give you an expert opinion; you the SFO should go and speak to someone else.
A: I think I have had conversations with the SFO to check that they know that I am not a STIR expert. "
The difficulty with this last answer (and others similar to it) was that the SFO's principal investigator, Mr McLaughlin, gave evidence at the retrial saying that he had not been present at any discussion with Rowe in which Rowe had told the SFO that he was not qualified to express an opinion on any issue in the case.
- The Judge's summing-up at the retrial was telling as to Rowe's evidence, in an extended passage early in his summing-up:
"Despite that catalogue of experience, you may have formed a judgment that he knew very little about the duties of being an expert……he seems to have been perfectly content to sign a standard declaration in which he declared that he had read the Criminal Procedure Rules which govern his conduct as an expert, both before trial and in giving evidence, and the booklet on his duties of disclosure without doing anything really to familiarise himself with either of those documents.
It will be for you to judge whether he has in fact given expert opinion which falls outside his true expertise.
Any expert is entitled to research a topic on which he is to give evidence and obtain the views of others, including work colleagues, about it to enhance his opinion, so long as he records where he went for that advice and so long as it is to enhance an expertise he already has, rather than to become an expert on a subject where he has no knowledge whatsoever.
There seems to be no dispute that he has a general expertise in banking and finance and that many of the issues he dealt with involved basic matters which are not in dispute.
……But there are other areas of his evidence where you would be entitled to conclude that he has gone beyond his general knowledge of banking into very specific areas, which were at the very edge of or beyond his knowledge. One such areas….is his knowledge of the short-term interest rate trades, the STIR trades…
He seemed to suggest that the SFO were aware of the limits on his expertise, but when McLaughlin gave evidence, he said that he had not been present at any conversation when Rowe said that he was not qualified to give an opinion, or that he refused to look at documents because they were outside his expertise….
….when he was cross-examined by Mr Darbishire, he accepted what perhaps he should always have accepted, that some of the things he was being asked about were beyond his expertise, in particular in relation to STIR traders.
It follows that you ought to be very careful indeed before relying on his evidence on the topic, and you may think it safer to ignore it…..
On the other hand, there are other areas in relation to general banking, of which he has given evidence, and which are not really in dispute where you might find his evidence of use, and that will be for you to determine."
- (2) Legal framework: There was no dispute as to an expert witness's duties to the Court, now enshrined in Part 19 of the Crim PR (unchanged from Part 33, in force at the times with which these proceedings are concerned). Thus, as provided by Part 19.2 of the Crim PR:
"(1) An expert must help the court to achieve the overriding objective –
(a) by giving opinion which is –
(i) objective and unbiased, and
(ii) within the expert's area or areas of expertise…..
(2) This duty overrides any obligation to the person from whom the expert receives instructions or by whom the expert is paid.
(3) This duty includes obligations –
(a) to define the expert's area or areas of expertise –
(i) in the expert's report, and
(ii) when giving evidence in person
(b) when giving evidence in person, to draw the court's attention to any question to which the answer would be outside the expert's area or areas of expertise…."
See too, R v Harris  EWCA Crim 1980;  1 Cr App R 5, esp. at ; R v B (T)  EWCA Crim 417;  2 Cr App R 3. Further detail as to recording, reporting and disclosing to the prosecution the material in the expert's possession, is contained in the Guidance Booklet for Expert's Disclosure: Experts' Evidence, Case Management and Unused Material.
- The essence of the matter is straightforward. As explained in the "standard" direction given to juries in respect of expert evidence (see above, for the direction given by the Judge at the trial), expert evidence is adduced to assist with matters likely to be outside their experience and knowledge. A partisan expert is quite incapable of furnishing such assistance, quite apart from the breach of ethical and legal duties thus entailed. So too, to state the obvious, expert evidence must be expert; it can only be such if it is within the expert's area/s of expertise; if the so-called expert witness gives evidence outside of his area/s of expertise it is both of no use to the jury and corrosive of the trust placed in such witnesses.
- English law is "characteristically pragmatic" as to the test for establishing expertise: Bingham LJ (as he then was), in R v Robb  93 Cr App R 161, at p.164, immediately before citing Lord Russell of Killowen CJ's observations in Silverlock  2 QB 766, at 771:
" ….It is true that the witness who is called upon to give evidence founded on a comparison of handwritings must be peritus; he must be skilled in doing so; but we cannot say that he must have become peritus in the way of his business or in any definite way. The question is, is he peritus? Is he skilled? Has he an adequate knowledge? Looking at the matter practically, if a witness is not skilled the judge will tell the jury to disregard his evidence. There is no decision which requires that the evidence of a man who is skilled in comparing handwriting, and who has formed a reliable opinion from past experience, should be excluded because his experience has not been gained in the way of his business……"
- That said, however the expertise is acquired, the expert must be confined to matters within his area/s of expertise. In Robb, Bingham LJ went on to express the risk otherwise (at p.166):
" …We are alive to the risk that if, in a criminal case, the Crown are permitted to call an expert witness of some but tenuous qualifications the burden of proof may imperceptibly shift and a burden be cast on the defendant to rebut a case which should never have been before the jury at all. A defendant cannot fairly be asked to meet evidence of opinion given by a quack, a charlatan or an enthusiastic amateur…."
- (3) Conclusions: Our conclusions on this Issue can be briefly stated. They are, in like measure, inevitable and deeply regrettable, albeit that a sense of perspective must be retained.
- Put bluntly, Rowe signally failed to comply with his basic duties as an expert. As will already be apparent, he signed declarations of truth and of understanding his disclosure duties, knowing that he had failed to comply with these obligations alternatively, at best, recklessly. He obscured the role Mr O'Kane had played in preparing his report. On the material available to us, he did not inform the SFO, or the Court, of the limits of his expertise. He strayed into areas in his evidence (in particular, STIR trading) when it was beyond his expertise (or, most charitably, at the outer edge of his expertise) – a matter glaringly revealed by his need to consult Ms Biddle, Mr Zapties and Mr Van Overstraeten. In this regard, he was no more than (in Bingham LJ's words) an "enthusiastic amateur". He flouted the Judge's admonition not to discuss his evidence while he was still in the witness box. We take a grave view of Rowe's conduct; questions of sanction are not for us, so we say no more of sanction but highlight his failings here for the consideration of others.
- All this said, it remains the case – as HHJ Leonard QC aptly observed in the summing-up in the retrial – that Rowe did have "a general expertise in banking and finance and that many of the issues he dealt with involved basic matters which are not in dispute". With both Rowe's significant failings and a careful assessment of his true significance to the trial in mind, we turn to address the safety of the Appellant's conviction.
ISSUE II: THE SAFETY OF THE CONVICTION
- (1) The test: This is well-travelled ground. Ultimately, the sole question for this Court is whether the Appellant's conviction was safe. As provided by s.2(1) of the Criminal Appeal Act 1968 ("the 1968 Act"), as amended, this Court:
"(a) shall allow an appeal against conviction if they think that the conviction is unsafe;
(b) shall dismiss such an appeal in any other case."
- The approach to this question in the case of fresh evidence, whether by way of late disclosure or otherwise, has been authoritatively considered. In R v Pendleton  UKHL 66;  1 WLR 72, the House of Lords approved and explained the observation of Viscount Dilhorne, in Stafford v Director of Public Prosecutions  AC 878, at p.906:
"While….the Court of Appeal and this House may find it a convenient approach to consider what a jury might have done if they had heard the fresh evidence, the ultimate responsibility rests with them and them alone for deciding the question."
As Lord Bingham of Cornhill said (at ):
"…I am not persuaded that the House laid down any incorrect principle in Stafford, so long as the Court of Appeal bears very clearly in mind that the question for its consideration is whether the conviction is safe and not whether the accused is guilty. But the test advocated by counsel …..does have a dual virtue… First, it reminds the Court of Appeal that it is not and should never become the primary decision-maker. Secondly, it reminds the Court of Appeal that it has an imperfect and incomplete understanding of the full processes which led the jury to convict. The Court of Appeal can make its assessment of the fresh evidence it has heard, but save in a clear case it is at a disadvantage in seeking to relate that evidence to the rest of the evidence which the jury heard. For these reasons it will usually be wise for the Court of Appeal, in a case of any difficulty, to test their own provisional view by asking whether the evidence, if given at the trial, might reasonably have affected the decision of the trial jury to convict. If it might, the conviction must be thought to be unsafe."
- Though less often cited, the speech of Lord Hobhouse of Woodborough contains a valuable summary of the applicable principles. Lord Hobhouse (at ) took as his starting point two "fundamental considerations of policy" that "no one should be convicted of an indictable crime save on his own plea or on the verdict of a jury; that the verdict of the jury should be final." The latter principle, originally absolute, was qualified by the creation of the Court of Criminal Appeal but, from the outset, its jurisdiction was strictly limited. In the light of s.2(1) of the 1968 Act, "….the sole criterion which the Court of Appeal is entitled to apply is that of what it thinks is the safety of the conviction. It has to make the assessment..…..appeals are not to be allowed unless the Court of Appeal has itself made the requisite assessment and has itself concluded that the conviction is unsafe…..". He went on to say (at ),
"…..Unless and until the Court of Appeal has been persuaded that the verdict of the jury is unsafe, the verdict must stand. Nothing less will suffice to displace it. A mere risk that it is unsafe does not suffice: the appellant has to discharge a burden of persuasion and persuade the Court of Appeal that the conviction is unsafe…."
"Unsafe" (at ) was an ordinary English word, "connoting a risk of error or mistake or irregularity which exceeds a certain margin so as to justify the description 'unsafe'. It involves a risk assessment." It was not right to attempt to look into the minds of the members of the jury. For an appellate court to speculate "whether hypothetically or actually" was not appropriate. It was for the Court of Appeal "to answer the direct and simply stated question: Do we think the conviction was unsafe?"
- In Dial v Trinidad and Tobago  UKPC 4;  1 WLR 1660, Lord Brown of Eaton-under-Heywood (giving the judgment of the majority) put the matter this way:
" 31. ….the law is now clearly established and can be simply stated as follows. Where fresh evidence is adduced on a criminal appeal it is for the Court of Appeal, assuming always that it accepts it, to evaluate its importance in the context of the remainder of the evidence in the case. If the court concludes that the fresh evidence raises no reasonable doubt as to the guilt of the accused it will dismiss the appeal. The primary question is for the court itself and is not what effect the fresh evidence would have on the mind of the jury. That said, if the court regards the case as a difficult one, it may find it helpful to test its view by asking whether the evidence, if given at the trial, might reasonably have affected the decision of the jury to convict…."
Having referred to the observation of Viscount Dilhorne in Stafford, approved in Pendleton, Lord Brown continued as follows:
"32. That is the principle correctly and consistently applied nowadays by the Criminal Division of the Court of Appeal in England….. It was neatly expressed by Judge LJ in R v Hakala [i.e.,  EWCA Crim 730], at para. 11, thus:
'However the safety of the appellant's conviction is examined, the essential question, and ultimately the only question for this court, is whether, in the light of the fresh evidence, the convictions are unsafe.'"
- In Burridge v R  EWCA Crim 2847, at , Leveson LJ (as he then was), saw no difference between the analyses in Pendleton and Dial. As he observed, the "jury impact test" was not treated as determinative; it was only a mechanism in a difficult case for the Court of Appeal to "test its view" as to the safety of a conviction.
- (2) The application of the test: In the light of the fresh evidence as to Rowe, emerging from the retrial, all of which we accept, was the Appellant's conviction unsafe?
- We take as our starting point the issue for the jury at the trial. As will be recollected, in his original Grounds of Appeal, the Appellant himself admitted and averred seeking to move the LIBOR rate to suit his book and to favour Barclays – and went on to say this:
"In order to do so, the ….[Appellant],,,, had to conspire with a submitter who would input rates in accordance with his wishes;
Equally, the submitter who was inputting artificial rates had to be conspiring with a trader who stood to benefit from his actions…."
Accordingly, he faced insuperable difficulty with the initial question of the genuineness of the LIBOR submissions – both in terms of the (unchallenged) direction framed by the Judge and the law as explained in R v Merchant. All that remained was the key issue, concerning the Appellant's alleged dishonesty.
- Plainly, the SFO had a strong case. The admissions in the Appellant's original Grounds of Appeal were extremely damaging here too. Further, we have already made reference to certain of the Appellant's e-mail communications, highlighted by the prosecution. In the course of a careful cross-examination by Ms Deacon QC, for the SFO, those and other communications were highlighted, including the following:
i) "We have to have a low LIBOR. Please remind Peter Johnson."
ii) "PJ [i.e., Johnson] said that going for 4.48 is too low. He will look ridiculous if he does that."
iii) "PJ's gotta jam that shit tomorrow".
- Equally plainly, the Appellant's defence did not require or involve delving into the technical details of STIR trading, or related matters.
- Against this background, we return to Rowe's evidence. We accept that the principal reason for the SFO calling him was to provide a "human face" to introduce to the jury essentially uncontroversial banking and trading concepts. Thus far, as the Judge observed in the summing-up in the retrial, he had sufficient general expertise to act as an expert. Where Rowe went wrong – gravely wrong, as we have concluded – was to go further and enter into debate on topics, beyond or at the very outer edge of his expertise, principally STIR trading. It was in this regard that Mr Allen made complaint, submitting that cross-examination was inhibited by not having the fresh material. Pressed by this Court in argument, he specifically complained of an inability to cross-examine on the "Stub" – i.e., the risk attributed to the points on the curve out to three months.
- We struggle with this submission. Mr Allen made an entirely prudent decision, on the material available, to curtail cross-examination of Rowe. No doubt too, the cross-examination on behalf of the co-defendants was itself limited. But, on the crucial issue for the jury, the Appellant's position would not conceivably have been assisted by a more detailed technical examination of STIR or the Stub. The issue was far more fundamental: it was a basic question of dishonesty, as already outlined. Further, there were other witnesses available to whom more detailed questions could have been put about derivatives trading and who had the knowledge to answer – Messrs. Bommensath, Bagguley and Harrison, all from Barclays, each of whom had been STIR traders. There was also the agreed glossary of terms available to the jury. Still further, by the conclusion of the trial, as is apparent from the paucity of references in the summing-up, the importance of Rowe's evidence could only have been of the most limited kind.
- The position might well have been very different had Rowe been permitted to adduce by way of his Addendum Report his opinion that it was impossible for a trader in a major bank to have been unaware that LIBOR was to be set independently of the trader's book. As already remarked, however, the Judge wisely excluded this item of evidence, so that it never came before the jury.
- In all these circumstances, notwithstanding the firm conclusions we have reached as to Rowe, we are unable to conclude that the Appellant's conviction was unsafe. In a nutshell and conducting the "risk assessment" suggested by Lord Hobhouse in Pendleton, we do not think that Rowe impacted at all or sufficiently on the key issue in the trial so as to affect the safety of the Appellant's conviction. We should underline that this conclusion is naturally fact sensitive and turns on a consideration of Rowe's evidence in the round, evaluated in the context of the trial as a whole.
- In reaching this conclusion, we have not overlooked the outcome of the retrial, following the devastating cross-examination of Rowe, once the fresh material had come to light. Mr Allen, understandably, made much and attractive play with this point. It does not, however, dissuade us from the conclusion to which we have come. The manner in which the new material emerged at the retrial was, of course, very damaging but there is no or no proper basis for assuming that that would have been replicated at the trial. If the fresh evidence had been available at the trial, the most likely outcome is that Rowe would not have been called at all, or that his evidence would have been tightly circumscribed. It would thus be pure speculation (even if permissible) to transpose the outcome of the retrial to the trial and to conclude that the Appellant's conviction was unsafe. In any event, regardless of the events at the retrial, we are wholly unable to make the causal link between Rowe's failings and the issue of the Appellant's dishonesty, which was the key focus of the trial. In our judgment, the issue of the Appellant's dishonesty was wholly unaffected by Rowe's evidence, even considering Rowe's presentation in the round. Accordingly, our conclusion as to the safety of the Appellant's conviction stands.
- Although the question is not determinative (Burridge, supra) and whether or not strictly necessary to do so, we have additionally asked ourselves the "jury impact" question, to test our view. Suffice to say, we are satisfied that, if the new material had been available at the trial, we do not think that it might reasonably have affected the jury's decision to convict.
- Accordingly, we conclude that the Appellant's conviction was safe and we dismiss the appeal.
- The instruction of Rowe turned into an embarrassing debacle for the SFO, all the more so, given the high-profile nature of these cases and notwithstanding that, in the event, it has had no impact on the outcome in this case. We pressed Mr Hines as to whether there was an internal report, dealing with lessons learnt. We subsequently received a helpful letter from the SFO's General Counsel, dated 27th November 2017, stating that there was no such document but that there had been extensive internal discussions resulting in the conclusion "…that Rowe's conduct resulted from a failure of integrity on his part rather than a failure of SFO policies or procedures". The SFO undertook to look again at the matter to see whether there was any way in which it could reinforce expert witnesses' awareness of their obligations under the Crim PR.
- In fairness to the SFO, this was the third time that Rowe had given evidence in LIBOR trials and the first time any questions concerning his expertise had apparently arisen. Nonetheless, there is no room for complacency and this case stands as a stark reminder of the need for those instructing expert witnesses to satisfy themselves as to the witness' expertise and to engage (difficult though it sometimes may be) an expert of a suitable calibre.