Quillan & Ors, R. v

On 11 December 2014 we heard argument on this appeal by the prosecution against a ruling handed down on 5 December 2014 when His Honour Judge Simon Drew QC, sitting at the Birmingham Crown Court, upheld submissions of no case made by all nine defendants indicted in this case on all nine counts contained in the indictment. We announced our decision that the appeal would be dismissed and that reasons would be given in writing. These are those reasons.

The appeal is brought by the prosecution, by leave of the trial judge, under s.58 of the Criminal Justice Act 2003 (CJA 2003). It relates only to the six defendants Gary Quillan, Christopher Hoole, Peter Garrett, Gregory Garrett, Andrew Edmondson, and Neal Thompson; they are the respondents to this application, but for convenience we shall refer to them as the defendants. The prosecution does not appeal against fact specific decisions that there was no case to answer in respect of three other defendants who were tried with the six with whom we are concerned. If the appeal had succeeded in respect of any defendant and any count, it was intended that the trial would continue before the same jury.

The rulings are dealt with in more detail at paragraphs 44 and following below, but involved one pure point of law in relation to counts 2 and 8, and a rather more mixed question of fact and law on counts 1 and 7. The trial started on 1 September but did not involve the jury until the end of September. The jury therefore spent two months hearing evidence on which it was not ultimately required to make a decision.

This judgment contains:

General procedural guidance about the management of legal issues in complex cases;

Our reasons for deciding that the court had jurisdiction to hear this appeal, given the events on the day when the ruling was given in the Crown Court;

Our reasons for dismissing the appeal on the substantive issues.

The procedure under s.58 of the CJA 2003 requires this court, in circumstances of this kind, to reach decisions on issues in complex cases very quickly. Inevitably, the appeal court knows far less about the facts of the case than the trial judge who has listened to evidence over many weeks and who has had an opportunity throughout that time to reflect on the indictment and the sufficiency of the evidence which is said to support it.

The regime for serious fraud cases allows for preparatory hearings under s.7 of the Criminal Justice Act 1987. There is a similar regime for other long complex or serious cases under Part III (ss.28-38) of the Criminal Procedure and Investigations Act 1996 (the 1996 Act). Each of these regimes allows the trial judge to make a ruling as to any question of law relating to the case (among other things), and provides for a right of appeal against such rulings with leave to this court.

One of the purposes of making an order for a preparatory hearing is to assist the judge’s management of the trial. Deciding issues such as those which arise on this appeal in advance of the trial, using this procedure, is often better than the course which has been adopted in this case. It allows the trial to proceed uninterrupted, if that is the result of the determination of the legal issue. If there is an appeal, it allows this court to determine issues without the pressure of time created by the need to ensure that the jury which has spent weeks dealing with the case is not further inconvenienced. Even more importantly, if the decision is that the judge was right to stop the case, then no trial is required at all and enormous public expenditure is saved.

Criminal proceedings are burdensome for all involved, particularly witnesses, jurors and other members of the public who become embroiled in them. In addition such prosecutions are a very considerable burden on public finances; a misconceived prosecution is therefore a very serious matter. If counts in the indictment are misconceived in law it is essential that this is determined at the earliest stage. It is sometimes the case that submissions of the kind made by the defendants in this case are made as late as possible in the hope that by the time the prosecution appreciates its difficulties it will be too late to amend the indictment or to take other steps to remedy them.

This court has given guidance in a number of cases particularly in R v I [2010] 1 WLR 1125, [2010] 1 Cr App R 10 as to the circumstances in which a preparatory hearing under Part III of the 1996 Act should be conducted. In giving the judgment of the court in R v I , the then Vice-President, Hughes LJ, said at paragraph 21:

Whilst that is almost invariably the position, there may be special circumstances where a trial will be very long and very costly and where a ruling on a point of law in relation to the legal basis on which a count in the indictment is founded may determine whether or not a trial is required at all. In such a case such a point of law should be determined well before any trial starts. That is not the same thing as saying that it must be resolved in a preparatory hearing. There is a power in any case under s.40 of the 1996 Act to hold a pre-trial hearing and to decide any question of law relating to the case concerned. This procedure does not involve any of the technicalities which have caused some difficulty in relation to preparatory hearings and there is no interlocutory right of appeal (except where the prosecution treats any ruling as a terminating ruling).

A further benefit of the preparatory hearing procedure is that the prosecution is not required to undertake that a defendant is entitled to be acquitted if an appeal it brings against a ruling in a preparatory hearing fails. Cases of real complexity are crucially affected by the charges the prosecution chooses to prefer and the way in which it formulates its case. In this case, it is most important to emphasise our decision does not mean that there is no offence disclosed by the conduct of the defendants as alleged by the prosecution. That is not an issue which this court can or should decide. It simply means that the conduct as set out in the indictment does not disclose the offences charged. It is obviously of benefit for such issues to be determined in advance of the trial so that, if it can be fairly done, the prosecution has an opportunity to reflect on any adverse ruling and to consider whether an amendment to the indictment should be sought.

In a case of complexity where such issues arise, the trial judge should first consider what rulings to make and when, and then decide whether to do so in a preparatory hearing. It is axiomatic that important rulings should be made as early as they properly can be. The early identification of such issues will require vigorous case management by the judge and the assistance of the parties. It is important to stress that those representing defendants in criminal proceedings owe a duty to the court which has always existed as a matter of law, and which is now to be found in Crim PR 1 and the corresponding provisions of the Consolidated Criminal Practice Direction. It is unnecessary for us to set those provisions out as:

In general, as the court observed at paragraph 37 of R v C [2010] EWCA Crim 2578, [2011] 3 All E.R. 509, [2011] Crim. L.R. 396, the court should make an order for a preparatory hearing well in advance of the anticipated hearing in accordance with Crim PR 15 so that the advocates’ minds are concentrated on the issues. Nonetheless, the judge can always, subject to the constraints set out in R v C and R v YDG [2013] 1 WLR 2014, exceptionally at any stage during that hearing, decide to designate some or all of it as a preparatory hearing and give a ruling within that regime which can come to this court at the suit of the losing party.

Those representing the defendants invited the judge to hold a preparatory hearing to consider, among other things, the form of the indictment and “Whether the pension scheme operated by Redswan fell within the terms of section 150(1) of the Finance Act 2004”. That is quite close to the point of law which the judge held on 5 December 2014, some months after the trial had begun, was fatal to counts 2 and 8, as will appear from our decision at paragraph 44 and following.

It is unnecessary to examine the way in which the case was being advanced at the date when the judge decided not to have a preparatory hearing. It is clear from the ruling which he gave when he decided not to do so that he was rightly concerned about the lack of a clear explanation of the indictment from the prosecution. He said

As events were to prove, this invitation most regrettably did not force the prosecution to identify its final case and thereby to identify the real legal issues before a great deal of public money had been expended on the trial. We sympathise with the judge, who was grappling with a case which rightly troubled him. The two rulings which we have seen, namely that on the application for a preparatory hearing and that on the submissions of no case, are both conspicuously clear and very carefully reasoned. It was not in any way the fault of the judge that this case turned out as it did.

Nonetheless it is our view that the argument which took place at the conclusion of the prosecution case should have been conducted within a preparatory hearing under Part III of the 1996 Act. It is clear that the judge was unfortunately faced with a regrettable and avoidable lack of precision as to the way in which the prosecution case was being put. It was not put at the outset in the same way as it was by the conclusion of its case at trial. This lack of precision by the prosecution team made the early definition of the crucial issue difficult; it is something that should not have happened and it would be desirable for lessons to be learnt. Despite the position in which the prosecution had put the judge, we nevertheless consider that the judge should have held a pre-trial hearing to require the prosecution to set out its case in law and to hear argument about whether it was well founded in law. It would therefore probably have served a valuable purpose and saved a considerable amount of public money.

Finally on matters of general significance we would observe that on the hearing of this appeal each defendant appeared by counsel. Mr Enoch QC and Mr Jebb appeared for Quillan, but made the necessary submissions on behalf of all defendants. The points were common to all and no possible conflict of interest arose. In these circumstances the court expects that counsel who do not wish to address the court will not attend. Arrangements must for the future in such cases be made for the appeal to be conducted by the preparation and presentation of one single set of submissions by one counsel, or, if the complexity of the case justifies it, by a team of two.

Decisions of this court have established that in an appeal under s.58 of the CJA 2003 strict compliance with the requirements of that section by the prosecution is necessary before this court has jurisdiction to hear it.

In this case it is contended that the prosecution did not comply with those requirements because they did not, immediately after the ruling was made, either apply for an adjournment or inform the court that they intended to appeal under s.58(4) of the CJA 2003. It is not suggested that they failed to give what has become known as the “acquittal undertaking” at or before the time when they informed the court that they intended to appeal as required by s.58(8). It is therefore necessary to examine what happened.

After hearing submissions and taking some time to consider his ruling, the judge sent a draft to the parties by email on 4 December 2014. This was not subject to any embargo expressly, but was accompanied by a message from him which said that he might correct typographical errors and insert further citations from the Finance Act 2004 overnight, but the substance would not change.

The prosecution quite properly treated it as a draft and distributed it between counsel and the CPS lawyer with conduct of the case, but no further. The next morning a further corrected version was sent out by the judge. The prosecution properly decided not to give a copy of either draft to the representatives of Her Majesty’s Revenue and Customs (HMRC) before the hearing, in accordance with the way in which the court usually expects draft judgments to be treated. HMRC were not the prosecuting authority but had investigated this case and were initially interested in it. The effect of the ruling was to bring the case to an end.

When the court convened at about 10:44 the recording equipment did not work, but we have the benefit of an agreed note of what happened. It is as follows:-

By the time the court next sat, the recording equipment had started to work; we have a transcript of what followed. The prosecution informed the court that it intended to appeal against the dismissal of counts 1, 2, 3, 7, 8 and 9 against the present six defendants but not against the ruling in the respect of the other three defendants in the trial and not in respect of counts 4, 5 and 6. After argument, the judge granted leave and ordered that the appeal be expedited.

Nothing at all happened between the first hearing and the second except that the prosecution team consulted with the HMRC lawyer and came to its decision. Mr Sutton told us that his team and the CPS lawyer had already decided what they were going to do. They needed only to acquaint HMRC with the terms of the ruling and to explain what they proposed to do about it. As HMRC was the investigating body in respect of this case, it stood in relation to the CPS and counsel in part in much the same position as the police do in other cases and in part as a complainant or victim as it was the body that had been the subject of the alleged fraud.

It is necessary only to quote the first part of the transcript.

Section 58(4) provides:

This provision is supplemented by Part 67 of Crim PR (Crim PR 67):-

Crim PR 67 which governs the process, subject to the terms of the Act, reflects the way in which s.58(4) had been construed by this court on a number of occasions. In R. v T(N) [2010] EWCA Crim 711; [2010] 2 Cr App R 12 (p.84) Lord Judge CJ presiding over a 5 judge court, said (at [13]):

In R v Mian[2010] 1 WLR 2655, [2012] EWCA 792, [2012] 2 Cr App R 9, the court gave further consideration to the application of the provision in the CJA 2003 and Crim PR 67 to the particular facts in that case. There had been a delay of 10 minutes between the giving of the ruling and the prosecutor complying with s.58(4). During that time the judge had informed the defendant that as soon as the jury could be brought back into court he would be acquitted.

The court decided on the particular facts of the case that this was too long. Rix LJ referred at paragraph 9 to the decision of Lord Judge CJ in R v T(N) and to his own previous decision in R v C, M and H [2009] EWCA Crim 2614 about the need for an “acquittal agreement” to be given at the time that notice of intention to appeal was given. At paragraph 29, he went on to consider what the term “immediately after the ruling” meant. He said at paragraphs 28 and 30:

Unsurprisingly the Crown Prosecution Service has published guidance about the use of this right of appeal on its website. It explains the requirements very clearly and says under “Points to Consider”, the following:-

It is clear, in our view, that what constitutes immediate notice under s.58 and Crim PR 67 will depend upon the circumstances. The day after the ruling was sufficient compliance in O [2008] EWCA Crim 463,because of the extraordinary events of the previous day. Ten minutes was, however, too long in Mian. Three observations about Mian are necessary.

It was a case where the acquittal undertaking was not given at or before the time when the court was informed by the Crown that it intended to appeal. S.58(8) is quite specific in requiring compliance with that provision as to timing. That is not a failure which occurred in this case.

The court was not making new law in Mian,butapplying the terms of s.58 and Crim PR 67 and previous decisions to the particular facts of the case.

The court in Mian made it clear that the issue did not turn on “split second timing.” The court in Mian noted that s.58(3) prevents the ruling from having effect while the prosecution is able to take any steps under s.58(4) and would be unnecessary if no such interval could ever occur.

Furthermore, it must be recalled that, although the prosecutor ultimately has to take the decision, a criminal trial also involves complainants or alleged victims whose interests and views are material. In such circumstances it is necessary to allow time for the prosecutor to carry out the necessary consultation before the decision is made. In R v Killick[2011] EWCA Crim 1608 the court held at paragraph 41 that in determining whether to prosecute, the prosecutor had to take into account three interests – that of the state, that of the defendant and that of the alleged victim. As the decision not to appeal is a decision which is final for the complainant or alleged victim, the prosecutor should consult and must be afforded a proper opportunity of doing so.

In each case therefore a careful examination of the facts is required to determine whether the prosecution has acted “immediately” in the context of the case under consideration. Much will depend on the complexity of the case, whether the ruling is oral or handed down and whether the prosecutor has had an opportunity of discussing the position with the alleged victim or other interested parties. In simple cases, such a discussion can well be had, as the CPS guidance suggests, before the ruling. In other cases, where the issues are complex and the ruling complex, time must be afforded for proper consultation; the word “immediately” must therefore allow time for such consultation. A sensible allowance for the requirements of justice and the practicalities of criminal trials must therefore be made. What the prosecution must not do is to “temporise” and cause delay except in accordance with s.58(4) and the provisions of Crim PR 67.

In this case the judge’s ruling was handed down in writing. It was 56 pages long. It contained important rulings of law and also decisions on questions of fact. It would have taken a long time to read in open court, which is the traditional way of ensuring that all relevant parties, as well as the public, are aware of the decision of the court. The prosecution agreed that it did not need to be read. As a consequence HMRC did not know what had happened or why. The prosecutor told us he wanted to tell them. That was not simply his desire; it was his duty to tell them and consult them before he was in a position to give the court notice of his intention to appeal.

We therefore agree with the submission made by the prosecution that where this means of promulgating a ruling is chosen, the court should allow time for it to be read by those interested who had not seen the draft and the views of those interested ascertained. In the present case this required at least an hour, if not significantly longer. The hearing should then be resumed as soon as possible after that proper and reasonable time has elapsed and it is at that time that the prosecution becomes under an obligation under s.58(4) and Crim PR 67 either to inform the court or seek an adjournment. It follows that Mr Sutton did comply with s.58(4).

A good deal of court time would have been saved before us if the prosecution had simply asked for time under s.58(4) and Crim PR 67 to consider the position. A reasonable time would have been until the next business day – the default period set out in Crim PR 67. If they had, such a time would have been granted and no argument such as the present could have been mounted. The decision of Mr Sutton QC to say that he was not applying under s.58(4) has given rise to this argument. It appears likely that he had in mind that he wanted to tell the judge that it would not be necessary to adjourn under s.58(4) with the consequence under Crim PR 67(2)(b) that the judge must in general not require the appellant to decide there and then, but instead must generally allow until the next business day. Mr Sutton was saying, instead, that he could decide “there and then” but needed a short time to explain the ruling to the investigating body, which was also, after all, the alleged victim of counts 1 and 7 on the prosecution case and had to be consulted.

In our judgment the application for “time” which Mr Sutton made during the first hearing could, if necessary, be regarded in substance and effect, as an application under s.58(4) even though he said it was not. The obligation on the prosecution is either to inform the court that it intends to appeal or to request an adjournment to consider whether to appeal. It would be possible to view Mr Sutton QC’s statement as a request for an adjournment to enable a proposed appeal to be discussed with HMRC which had a legitimate interest in being consulted about it. His choice of words was plainly maladroit, but his intention entirely commendable. He was simply saying that he did not need the usual 24 hours but believed he could complete the matter before lunch. On the assumption, contrary to what we have already decided, that there was an obligation under s.58(4) and Crim PR 67 to inform the court before there had been a reasonable opportunity to read the judgment and consult HMRC, we consider that he did make an application to adjourn as required by the CJA 2003 during the first hearing. He set in motion a two stage process involving a short adjournment followed, only if necessary, by the longer adjournment envisaged by the Crim PR 67. That is what he meant when he said

The third ground of our decision is that even if Mr Sutton failed to comply with s.58(4) during the first hearing at 10:44 and even if that was the time the obligation arose under s.58(4), we are in no doubt that his compliance during the second hearing at 11:49 was “immediate”. In those circumstances we consider the whole course of conduct of Mr Sutton QC did amount to immediate compliance with the requirement to inform the court that he was intending to appeal or to seek an adjournment under s.58(4). He did not temporise, rather the reverse. Had he wanted a 24 hour period to take a decision, this would have been available to him. Nothing that the court wished to do was in any way delayed by his chosen course.

The very strict approach to the word “immediate” was prompted by a concern that the prosecution might misuse its new right of appeal and flood this court with work which could not sensibly be accommodated. Trials might routinely be delayed while decisions were taken. In the experience of this court this has not happened. We have set out above the CPS Guidance on the subject. Prosecuting authorities are expected to use this power with restraint and discrimination. Where they do so, the court should be slow to decline to hear appeals by implying into the word “immediate” a wholly unrealistic approach in a complex case. This does not undermine the importance of compliance with the s.58 procedure, but does suggest that an approach which gives effect to the overriding objective of dealing with a case justly should be followed.

In the present case, the prosecution announced its concluded decision in a complex case within 1 hour of the handing down of the judgment which the judge had taken many days to write. As this case shows, careful compliance with the section and Crim PR 67(2) is not difficult to achieve and prevents this kind of problem from arising. It should be achievable and achieved. Any failure which causes any significant difficulty in the proceedings may well result in this court declining to entertain the appeal.

For these reasons we hold that this court did have jurisdiction to hear this appeal. We now move on to determine its merits.

The nature of the case and evidence against the defendants

As the indictment in this case was a complex document, we must at the outset encapsulate the allegations it made as briefly as we can in order to simplify what follows. Essentially the allegations against the defendants were that two schemes were set up (one following the other) with the dishonest intention of securing the payment of Income Tax relief at source (RAS) from HMRC by paying the same sum of money into pension schemes over and over again, each time causing the payment of tax relief from HMRC to the scheme. The circular nature of the transactions was set out in a chart which was shown to the jury as Figure 3.1 to the Report of Daniel Ryan of Berkeley Research Group as part of the prosecution case.

The indictment contained 9 counts. Counts 1-3 relate to Scheme 1 and counts 7-9 relate to Scheme 2. Counts 1 and 7 are allegations of Conspiracy to Defraud, count 1 in relation to Scheme 1 and count 7 in relation to Scheme 2. Counts 2 and 8 charge Conspiracy to Cheat, count 2 in relation to Scheme 1 and count 8 in relation to Scheme 2. Counts 3 and 9 are counts of Conspiracy to Launder Money, one in relation to each scheme. Counts 4-6 are charges of False Accounting; which were charged as alternatives to counts 1-3 and related specifically to the Pension Administrators of Scheme 1.

This appeal concerned only counts 1 and 7, and 2 and 8. That is because it was agreed that counts 3 and 9 depend upon the outcome of the submission on these counts, and because the judge ruled that there was no case to answer on counts 4, 5 and 6 on evidential grounds and there is no appeal against that part of his ruling.

After what appears to have been a great deal of debate at trial, the prosecution case eventually became clear during the submissions of no case. It was that the conspiracy to cheat counts (2 and 8) alleged that the schemes were fraudulent from the outset because the sums paid as RAS were never lawfully due. The conspiracy to defraud counts (1 and 7) were intended to cover the possibility that the RAS may have been paid lawfully, and then diverted because it was extracted from the pension schemes and not used to make any investment or provide any benefit to the supposed member of the scheme, but instead paid out to, among others, the defendants who had set up the schemes. The judge set out the history of the way in which the prosecution case achieved this degree of clarity in his ruling, but it is not necessary for us to rehearse it.

Counts 1 and 7 were in the following terms:

Counts 2 and 8 were in the following terms:

We are indebted to the judge for his clear and careful analysis of the evidence which we can summarise as follows.

This case centred on the activities of Gary Quillan (“GQ”) and Christopher Hoole (“CH”). They were both Independent Financial Advisers (“IFAs”). They came together towards the latter part of 2006 in order to market a Self Invested Personal Pension scheme (“SIPP”), which had originally been devised by GQ. The original scheme, Scheme 1, which used a registered pension scheme latterly called Redswan, resulted in counts 1 and 2. That activity continued until June 2009. During the latter part of that period, CH established a second scheme, Scheme 2, which was the subject of counts 7 and 8. This scheme used a registered pension scheme called MW Pensions. The indictment period is a little under 3 years.

The schemes operated in what appears to have been a novel way. They involved borrowing and lending money in a more or less circular manner. The essential structure of the schemes appears to have been as follows:

In order to establish a SIPP an investor borrowed money from an off-shore loan company recommended by the IFA.

That loan money was used by the investor as his contribution into the SIPP.

The administrators of the SIPP then claimed RAS, which was also paid into the SIPP.

The investment strategy was delegated to the IFA, who recommended that all the funds should be invested in purchasing shares in certain unquoted UK companies.

The unquoted UK companies then used the capital generated by the sale of shares to lend to the off-shore loan companies recommended by the IFAs.

That then enabled the cycle to start again.

The role played by the defendants in the schemes was:

GQ and CH promoted the scheme to potential investors through their IFA businesses.

The off-shore loan companies were owned and controlled by associates of GQ and CH, in particular Peter Garrett (“PG”) who was the director of ANZ Direct Ltd, Smoothrock Investments Ltd and BAC Investments Ltd.

The unquoted UK companies were also owned and controlled by associates of GQ and CH, in particular Andrew Edmondson (“AE”) who was the director of Jav Investments Ltd, Pyra Investments Ltd, Squared Associates Ltd, IPPP Ltd and ACP.

The SIPP administrators for Scheme 1 were Purplecircle Consulting Ltd, which became Redswan, whose director was Mark Andrews (“MA”) and employee was Richard Barlow (“RB”); both were trustees of the Redswan pension scheme. MA and RB were indicted on counts 1-6 but as we have said above, the case against them was properly stopped by the judge on the ground that even if those counts disclosed offences there was no evidence against them that they had knowingly participated in any criminal conduct.

Between October 2006 and February 2007 a firm of IFAs, Clear Blue Finance, introduced 22 SIPP clients to Zurich Financial Services Ltd. Their SIPP contributions were funded by the taking out of loans either directly from the defendants or from companies that they controlled, plus what RAS claimed from HMRC. An examination of their gross contributions reveals that most contributed at or near to their declared yearly earnings. By February 2008 all 22 had transferred their funds away from Zurich to another pension provider, The Purplecircle Private Pension Plan. In total some £557,881 of investments was transferred to Purplecircle, which later changed its name to Redswan.

In order to get the scheme going the defendants needed capital with which they could make the initial loans to the SIPP clients. The SIPP funds transferred to Redswan from Zurich were used as that capital. Although the prosecution have only been able to trace what happened to £427,187 of those funds, all of that sum was invested in buying shares in Jav, one of the unquoted companies. The capital generated by the sale of shares was then distributed as follows: £15,000 was paid to AE; £33,000 was paid to Bright Ltd, a company controlled by GQ and CH and which had provided some of the original funding to make loans to the Zurich SIPP clients; £375,000 was lent to the off-shore loan company Smoothrock. Smoothrock then lent that money to new SIPP clients of GQ and CH who then used the funds to contribute to new SIPP schemes administered by Redswan. Redswan then claimed RAS and then both the initial contribution and RAS was invested in the unquoted UK companies controlled by the defendants. And so the cycle went on.

Once the scheme was up and running other off-shore loan companies were used to lend money to new clients. New unquoted UK companies were invested in by the SIPP. The common theme was that all of the companies were controlled by the defendants.

The purpose of a SIPP is to invest the clients’ contributions, including the RAS, in such a way so as to generate a fund which can then be used to finance a pension. One of the most obvious “curiosities” of the GQ scheme was that it did not obviously do that. This was because the SIPP clients borrowed money to make their contributions at a rate of interest that was higher than the rate of interest at which the unquoted UK companies, into which they had invested, lent. Thus, on the face of it, they were never going to generate enough money to repay their initial loans plus interest let alone generate a fund for a pension.

According to what GQ said in interview, that was to be overcome by investing the RAS in land/property developments and bridging finance, both of which he expected to yield high returns. Whether or not that was really his intention, and if so whether or not the return on those investments was ever likely to generate sufficient funds to repay the loans, plus interest, and generate a pension fund, were factual issues in the case. There was evidence from an actuary who gave expert evidence as to the likely pension outcome from the scheme that the projected rates of return (15-18%) involved “a brave assumption” but were not impossible. However, it would need a rate of return of this order over many years if there was ever to be a fund sufficient to provide a pension.

There were several other “curious” features of the scheme:

Not only did the clients invest close to their annual income in the scheme (which was the maximum permissible), but they did so for two and sometimes three years running. That involved heavy borrowing for people who were, in the main, not particularly high earners.

The loans they took out from the off-shore companies were unsecured.

The loans made by the unquoted UK companies to the off-shore loan companies, into which the SIPP clients had invested their large pension contributions, were also unsecured.

However, their positions were protected by a further “curious” feature of the scheme. The SIPP holders were informed that they would never have to put any money into the scheme, whatever happened to the investments. In essence it was risk free.

The loans taken out by the SIPP holders were renewable annually. When they entered into the initial loan agreements, and when they renewed those agreements, they were also given what was called a “conversion option”. That was the option to enter into a further agreement with the loan company. That agreement allowed the SIPP holder, at the expense of paying a higher rate of interest, to cap the amount he had to repay on the loan. That cap, another “curious” feature of the scheme was, in essence, the value of the SIPP fund on maturity.

Therefore the effect of entering into the conversion option agreement was to extinguish the SIPP holder’s liability to repay any sum by which the amount of the unpaid loan and any unpaid interest exceeded the value of the fund. Thus he was absolved from making any contribution to the SIPP at any time. If the conversion option was exercised, as it always was, the chance of a pension ultimately being paid was reduced still further by the higher rate of interest now charged.

A final “curious” feature of the scheme was that, as a further inducement to enter into the scheme, the clients were offered the payment of “cashback”, ranging from a few hundred to a few thousand pounds, for entering into the scheme. As a result about 100 people entered into the Redswan scheme. Some of them gave evidence. They had various degrees of knowledge and understanding of the scheme and financial matters in general. The common theme was that they all said they thought that the scheme was legitimate.

Towards the end of 2007 GQ and CH appear to have had a difference of opinion as to how the scheme should be operated. As a result CH set up his own scheme, using MW Pensions, although he appears to have paid GQ some sort of commission or franchise payment in order to base his scheme on the scheme GQ had devised. The Redswan scheme was referred to at trial as Scheme 1 and the MW Pensions scheme as Scheme 2.

Scheme 2 operated in essentially the same way as Scheme 1. Over the first four months of Scheme 2 £322,750 was introduced into the scheme from various sources: £156,000 from SPO, a company CH owned; £26,500 from CH himself; £90,000 from Andrew Sweeney’s business Premier Financial; and £50,000 from Andrew Sweeney himself. CH promoted the scheme to potential clients, and he received some help in doing so in the North East of England from Luke Thompson. His principal assistant was Neal Thompson. He was the director of the off-shore loan company Prospect Investments, and indeed prior to Prospect setting up a bank account he allowed his First Reserve bank account to be used in order to provide investors with loans using the money that came in from SPO etc, on what he called an agency basis. He was also the director of the unquoted UK companies Atrium Investments Ltd and Axiom Investments Ltd. The same recycling of funds took place in Scheme 2 as in Scheme 1, and the same extracting of RAS. As a result all of the money that had originally been introduced into the scheme was then repaid over time from the RAS. There were however two differences between the schemes. The Pension Administrators were MW Pensions; and the only use to which the share sale proceeds was put was in lending to new SIPP holders, thus there was no investing in any other enterprise.

Scheme 2 also attracted clients, in excess of thirty, and some of them gave evidence. They too had varying levels of understanding of the scheme, but once again they all said that they had thought that the scheme was a legitimate one.

In 2008 the FSA started an investigation into the operation of both schemes. On 22 June 2009 these defendants were all arrested. Their homes and business premises were searched and a large quantity of documents and computers was seized. They were bailed and then re-interviewed in January and February 2010. In interview all denied any wrong doing.

The law relating to SIPPS

During the indictment period the operation of SIPPs was principally governed by Part 4 of the Finance Act 2004. The key features of the legislation as it applied to SIPPs were as follows:

In any tax year an individual could invest a sum up to the equivalent of 100% of his annual income (subject to a cap) in a personal pension scheme (which included a SIPP). The funds could come from any source, including his income, his savings, or a loan.

That money was held by the scheme in a trust and it was administered by HMRC registered Pension Administrators, who were independent of the IFAs or other financial advisers who promoted the scheme.

Provided the pension scheme was registered with HMRC the investor was entitled to claim basic rate tax relief on his investment “at source” (“RAS”). What that meant was that the investor need only pay in 78/80% of his proposed contribution, depending on the level of the basic rate of tax. The balance was then claimed from HMRC by the Pension Administrators as RAS. HMRC paid those claims directly to the Pension Administrators who then paid the funds into the individual SIPP holder’s account.

From 2006 onwards SIPP funds could be invested in very many different ways and, provided the investor chose to invest his funds in a permitted (“authorised”) way, he was free to invest all of the funds in whichever way he chose. In practice it was often the IFA who directed the investment strategy, for a fee, and the SIPP clients often delegated their investment powers to them. Any income derived from those investments was not subject to income tax, and any increase in the value of an investment was not the subject of capital gains tax.

The role of the Pension Administrator was to oversee the administration of the SIPP; in particular claiming RAS and ensuring that the funds were invested in permitted investments. However, the Pension Administrator was not permitted to give financial advice to the SIPP holder.

If SIPP funds were invested in investments that were not permitted then they were likely to be subject to tax charges and to tax penalties.

Payments out of the scheme could only be made in certain limited circumstances. They included the payment of a pension, transferring to another SIPP, and administration and management fees. However, it should be noted that there were strict rules preventing the making of payments to the SIPP holders themselves, including by way of loan or benefit in kind.

On retirement, death, reaching a particular age or on the onset of serious ill health or incapacity the SIPP holder or his estate was then entitled to draw on the pension fund.

The parts of the Finance Act 2004 which are central to the argument are as follows:

The judge had to grapple with a number of different documents which contained somewhat different explanations of how the prosecution put its case. We are not concerned with that history, because we are only concerned to determine whether he was right to say that the case which the prosecution wished to advance at the conclusion of its case was a case which disclosed the offences alleged in counts 1 and 7 and 2 and 8 on the indictment.

It was the prosecution case that both schemes were a sham.

The defendants knew full well that the schemes would never generate a fund large enough to pay off both the loan and interest, and then also have sufficient left over to provide a pension.

Such investments as there were were small scale, and took place only towards the end of the running of the schemes, primarily when they were under investigation.

The defendants’ aim was to establish schemes which recycled the same capital and which generated large amounts of RAS.

The defendants were principally after the RAS money, which once the initial funding was repaid, was the only real money in the scheme.

Almost all of the RAS was extracted from the scheme by way of payments to the defendants disguised as administration fees, costs and other payments or through repayments to all of those who had provided the initial funding for the scheme.

That this was taking place was of no concern to the SIPP holders. That was because they had not provided any capital investment themselves and, by reason of the conversion option, they took no risk. They were essentially unconcerned whether or not the scheme actually provided them with a pension or not. In essence, they received payment for lending their names to a scheme designed to extract RAS for the defendants and were not likely to obtain any other benefit.

Whatever may have been the position at other stages of the case, by the close of the prosecution case, the prosecution firmly told the judge that counts 1 and 7 related to the extraction of RAS after it had been legitimately paid to the schemes, if that was the finding of the jury. He proceeded on that basis, as do we.

This means that the logical order in which to consider the submissions is first to decide whether there was evidence on which a jury could properly conclude that the RAS was obtained unlawfully in the first place (counts 2 and 8) and then whether there was evidence that, even if it was lawfully obtained, the extraction of the money from the pension schemes constituted a fraud on either HMRC or the SIPP investors (counts 1 and 7).

By the time of the argument before the judge, the prosecution put its case on the lawfulness of the obtaining of RAS (counts 2 and 8) in two ways:

No contribution

Contribution – but non compliance with the Act

The general defence submission that the claims for RAS were lawful can be summarised as follows:

Pension schemes, and SIPPs in particular, were at the relevant time governed by the provisions of the Finance Act 2004.

Prior to the commencement of either Scheme 1 or 2, Redswan and MW Pensions established “pension schemes” within the meaning of s.150(1) of the 2004 Act. They then registered those schemes in accordance with s.150(2) and Chapter 2 of the Act. Those were agreed facts.

The Scheme 1 and Scheme 2 SIPP holders became members of the Redswan and MW Pensions registered pension schemes when their applications were accepted by the pension schemes. They had entered into an arrangement, or agreement, with those pension schemes pursuant to ss.151(1) and (2) and s.152 of the Act.

Their membership status was that each was “an active member of a pension scheme”: see s.151(2). That was because there was an arrangement in place under the pension scheme “for the accrual of benefits to or in respect of the person”.

As a result they were entitled to the benefits set out in Chapter 4 of the Act. In particular when the money they had borrowed from the off-shore companies in order to fund their pension was paid into the SIPP account that was a “relievable pension contribution” pursuant to s.188(2) of the Act, and as an active member of a registered pension scheme they were entitled to tax relief pursuant to s.188(1).

Accordingly they were entitled to make the deduction set out at s.192(1), that is they were entitled “to deduct and retain a sum equal to income tax on the contribution at the basic rate for the tax year in which the payment is made”. Furthermore the scheme administrators had to allow that deduction in accordance with s.192(2), and they were entitled to recover the amount deducted from HMRC in accordance with s.192(3).

That was what was intended to happen in relation to the members of each of these two pension schemes and it was in fact what did happen on each occasion.

Thus the RAS payments were all lawfully obtained from HMRC.

Accordingly there was no evidence from which it could be inferred that the defendants conspired to cheat HMRC.

“Relievable pension contributions” are defined in s.188(2) of the Act as “contributions by or on behalf of the individual under the pension scheme other than contributions to which subsection (3) applies”. It was not suggested that subsection (3) applied here. The judge decided that the word “contribution” simply means a payment. Thus in order to satisfy s.188(2) of the Act there must be:

a payment (which could be money or money’s worth)

by or on behalf of the individual (who must be an active member of a pension scheme; see s.188(1))

under (into) the pension scheme.

The judge said that it was not in dispute that all of the SIPP holders were enrolled as members of a registered SIPP. It was also not in dispute that payments were made into the SIPP clients’ pension scheme accounts from the loan companies via the SIPP schemes’ bank accounts, by or on the SIPP clients’ behalf. The judge held that whatever the underlying value of the investments in the pension scheme were, those transfers or payments were of real money. The Bath Building Society bank statements showed those transfers.

The second prosecution submission concerned the correct interpretation of “active member” in s.151(2). The prosecution submitted that according to s.188(1) before an individual can claim RAS he must be an active member of a registered SIPP scheme. The prosecution’s argument is that the requirement that “there are presently arrangements made under the pension scheme for the accrual of benefits to or in respect of the person” in s.151(2) was a requirement that there should be an examination of the particular arrangements that were in place between the SIPP holder and the pension scheme to see if benefits were in fact accruing under the SIPP scheme at the time when the contribution was made. Only if there were could the claimant be “an active member” of the pension scheme. On this appeal, the prosecution disavow that formulation of its case, contending that the prosecution argument is that the entitlement to tax relief has to be examined on an individual basis and by reference to the nature of the pension scheme to which he makes his contribution. Whether or not the method by which tax relief (where appropriate) is given is efficient or costly is not to the point. The point is whether, on a proper construction of the individual’s position, he is entitled to the relief. If he is not entitled to the relief and he has been given it either in error or as a result of fraud, the method by which he obtained it is immaterial.

The judge decided that the term “arrangement” in s.151(2) means agreement. He relied on the Explanatory Notes to the Finance Bill 2004 (as it then was). This contains the following:-

Thus the judge held that a person will be an active member of a pension scheme if there is an agreement in place between the SIPP holder and the pension scheme which is for the accrual of benefits. He found that there were such agreements in place in this case. In relation to Scheme 1 the terms of those agreements are contained within and evidenced by such documents as the SIPP Application forms, the letters of acceptance, the Appointment of a non-regulated Intermediary forms, and the instructions given to the trustees as to where to invest the funds. In relation to Scheme 2 the terms of the agreements are contained within and evidenced by such documents as the SIPP Application forms, the CHF SIPP proposal letters and forms, and the Unquoted Investments Purchase Application forms.

The judge also held that s.152(2) did not impose a requirement that the current investment activity under the pension scheme needs to be examined on an individual basis each time a person seeks to claim RAS, as he held must be the case if the prosecution submission were correct. Accordingly he found that the RAS payments were all lawfully obtained from HMRC and therefore that there is no evidence from which it could be inferred that the defendants conspired to cheat HMRC and that there is no case for any of the defendants to answer on counts 2 and 8.

The fundamental problem faced by the prosecution in relation to these two counts may be simply stated. The claims for RAS were made, in accordance with the specified statutory procedure, by the duly appointed scheme administrators of the registered pension schemes (Redswan and MW Pensions) of which the clients of Scheme 1 and Scheme 2 were admittedly members. The claims were made in respect of contributions paid into the registered pension schemes by the clients, funded by the loans which they had taken out with the relevant offshore loan companies. The movements of the money can be traced in the documentation, and are reflected (for example) in the bank statements of the accounts held by Bath Building Society with the NatWest Bank, which the scheme administrators of the Redswan SIPP had specified to HMRC as the bank account to which payments of RAS should be made. How, then, could it plausibly be argued that the RAS was not in fact due in respect of the contributions thus made by the clients to the registered pension schemes of which they were at all material times members?

The fact that the contributions were made with borrowed money cannot in itself be an objection, because everybody agrees that an individual active member of a registered pension scheme may fund his relievable pension contributions (up to the yearly ceiling laid down in the legislation) by borrowing for the purpose. Nor, on the face of it, are the terms on which the money was borrowed of any relevance, because that is a matter between the lender and the client, outside the purview of the pension scheme. Once the money was paid into the pension scheme, the purposes for which it could then be paid or applied were in many respects closely regulated, although the whole point of a SIPP was that it allowed a wide degree of freedom in how the member’s fund was invested. But again, none of that appears to have any obvious bearing on the entitlement of the member to tax relief on his contribution under s.188 of the Act, or on the right of the scheme administrators to claim RAS pursuant to s.192.

The first way in which the prosecution sought to confront these difficulties was by arguing that the arrangements made under Scheme 1 and Scheme 2 were sham. Unfortunately, however, this contention was throughout advanced at a very high level of generality, without any attempt being made to analyse the precise respects in which it is alleged that the arrangements were sham, or to identify the real underlying transaction which it is said that the parties were in fact undertaking.

The concept of sham has a well-established legal meaning, the classic statement of which is to be found in the judgment of Diplock LJ in Snook v London and West Riding Investments Ltd[1967] 2 QB 786 at 802:

This definition has been repeatedly applied by courts and tribunals in the context of tax cases: see, for example, the judgment of the Court of Appeal (which was not cited to us, but is well known to practitioners in the field) in Hitch v Stone[2001] EWCA Civ 63, [2001] STC 214. Delivering the only reasoned judgment (with which Sir Martin Nourse and Kay LJ agreed), Arden LJ encapsulated the relevant law as follows:

In the light of these principles, it is in our judgment clear that a high-level and unparticularised invocation by the prosecution of the concept of sham is not good enough, and could never have provided safe grounds for a conviction on counts 2 or 8. If the prosecution wished to establish that the arrangements which the promoters of Schemes 1 and 2 made with the clients did not generate relievable pension contributions, it would have been necessary for them to satisfy the jury, to the criminal standard of proof, that the contributions lacked any true legal substance, notwithstanding their payment into the registered pension schemes of which the clients were members.

In our judgment it would have been impossible to make good any contention of this nature without also establishing that the clients were conscious participators in the sham transactions, with the intention (for example) that their apparent contributions should be no more than movements of money in a fraudulent plan designed to extract RAS from the Revenue. If there is one thing which the authorities insist upon, it is that a sham must reflect the common intention of all the parties to the impugned transaction. Yet the prosecution had throughout studiously refrained from alleging that the clients were necessarily implicated in the alleged fraud. As the judge aptly remarked in paragraph 64 of his ruling, albeit in the context of counts 1 and 7, the prosecution’s stance in relation to the SIPP clients “has remained throughout this trial as one of studied neutrality if not masterly inactivity”. For this reason, if no other, the sham allegation is in our judgment a hopeless one which could never have been permitted to go before the jury. For the purposes of counts 2 and 8 the prosecution, in our judgment, needed to show that the clients were knowing parties to a sham transaction, but as will appear below for the purposes of counts 1 and 7 they sought to show that the clients were the victims of the fraud. This illustrates, in a nutshell, a central difficulty in this case.

The prosecution’s second argument focused on the definition of “active member” in s.151(2) of the Act. If the client was not an active member of the registered pension scheme which he joined, he could not have been entitled to tax relief for his contributions under s.188(1) because relief is granted only to “[a]n individual who is an active member of a registered pension scheme”. So much, as the judge said, was common ground. But since each client was admittedly a member of the pension scheme in question, what basis could there be for denying that he was an active member? S.151(1) defines “member”, in relation to a pension scheme, as meaning “any active member, pensioner member, deferred member or pension credit member of the pension scheme”. It is clear that the clients did not fall into the second, third or fourth of these categories, so by a process of elimination alone it would seem to follow that their membership must have been as active members.

S.151(2) says that for the purposes of Part 4 of the Act:

A sufficient indication of the relevant contractual arrangements can, we think, be gained from the application forms filled in by the clients, several of which were included in the bundle. Thus, the application form for membership of the Purplecircle/Redswan private pension plan contained the following declarations to be made by the applicant:

We would add that, in reaching this conclusion, we do not consider it necessary to have regard to the Explanatory Notes to clause 142 of the Finance Bill 2004, which was enacted as s.152 of the Act. On the other hand, the terms of s.152 itself make it clear that the “arrangement” in relation to a member of the scheme means “an arrangement relating to the member under the pension scheme”. This therefore reinforces the point that one has to look at the terms of the pension scheme itself in order to identify the relevant arrangements. Although nothing turns on the point, we think that the arrangements relating to the client members must have been “money purchase arrangements” as defined in ss.152(2), (4) and (5).

For these reasons, which are essentially the same as those given by the judge, we consider that he was clearly correct to uphold the submissions of no case to answer on counts 2 and 8.

Counts 1 and 7 allege that the fraudulent agreement was to divert the RAS money away from the pension scheme once it had been lawfully obtained. The only point of counts 1 and 7 was to seek to set up that case if necessary.

The judge dealt with the diversion of RAS first. He decided that on the evidence as it currently stood it was open to the jury to infer that there was an agreement that the defendants could put the RAS, or at least some of it, “to their own use”. He also decided that it was open to the jury to conclude that the agreement to divert funds in this way was dishonest.

He then turned to the question of who the victim of this conspiracy was. The prosecution, by that stage, was contending that it was both HMRC and the SIPP clients. The defence submitted that neither HMRC nor the SIPP clients could have been defrauded and thus there was no case to answer on these counts.

The judge was referred to the analysis of the offence of conspiracy to defraud by Hickinbottom J in the case of R v Evans and others[2014] 1 WLR 2817. At paragraphs 28-48 of that judgment there is an examination of the scope of the offence of conspiracy to defraud in the light of the analysis of the late Professor Sir John Smith, in his Law of Theft 2007 Edition at paragraph 5.12:

The defence submitted that once RAS was paid into the SIPP holder’s bank account HMRC ceased to have any proprietary or economic interest in it. Thus a conspiracy to divert RAS from the pension was not one to defraud HMRC.

The prosecution submitted that HMRC retained an economic interest in the RAS after it was paid to the SIPP holder. They submitted that there was an obligation to invest the RAS for the purpose of a pension and that if it was not so invested and was used for some other purpose then HMRC’s economic interests were or might be put at risk. They did not contend that HMRC had a proprietary interest in the RAS (which they obviously did not). Instead they submitted that HMRC had an economic interest in the RAS which was analogous to those deemed to exist pursuant to s.5(3) of the Theft Act 1968 and section 4 of the Fraud Act 2006.

The prosecution cited a number of authorities both to the judge and to us which they submitted supported this contention. They were R v Hall (1972) 56 Cr App R 547, Wakeman v Farrar [1974] CLR 136, R v Meech & Ors(1974) 58 Cr App R74, R v McHugh[1993] 97 Cr App R 335, R v Wain [1995] 2 Cr App R 660, Klineberg v Marsden[1999] 1 Cr App R 427, R v Floyd (2000) CLR 411, and Re Holmes [2005] 1 Cr App R 16. It is unnecessary to burden this judgment with an analysis of these decisions. None of them comes close to establishing any principle of the kind contended for.

S.5 of the Theft Act 1968 provides

Section 4 of the Fraud Act 2006 provides:

The judge rejected the suggestion that these provisions were capable of establishing that HMRC had any economic interest in the destination of RAS which they had lawfully paid into a registered pension scheme, as do we. Those are provisions which create or define an element of statutory offences which were not charged in this case. They are simply irrelevant to the definition of the kind of economic interest which must exist before the offence of conspiracy to defraud can be committed.

Accordingly the judge was right to hold that they were of no value in the current exercise. HMRC were under an obligation in this respect to make a payment of RAS in circumstances where it was due. Thereafter HMRC had no further continuing interest in that money at all. In those circumstances it follows that in this case any agreement to extract RAS from the SIPP clients’ pension schemes was not one to prejudice the economic interests of HMRC and could not, as a matter of law, be a conspiracy to defraud HMRC.

The prosecution conceded that the jury might find that all of the SIPP clients were parties to the conspiracy. A number of the clients of both schemes were called by the prosecution to give evidence. They all said that so far as they were concerned they were involved in a genuine scheme. The judge remarked that “they would say that, wouldn’t they”. He said that the jury might accept what they said, but they would be equally entitled to conclude that the SIPP clients knew full well what was going on, that they joined in the conspiracy willingly and that the “cashback” payments amounted to their “cut”. If that was so, the judge held that the jury could not be sure that some or all of the SIPP clients were victims and not co-conspirators. We have referred to this passage of the judge’s reasoning at paragraph 89 above.

In reality, the jury could not exclude the possibility that the SIPP clients had simply received a payment for doing nothing. They never invested any money in any pension and were not exposed to any financial risk because of the “conversion option”. It is highly likely that none of them will ever receive a pension because the money has been diverted and not invested on their behalf. The jury may have found that there were some who thought that a scheme in which they had never invested a penny would provide them with a pension. The prosecution could not, however, prove that. If they were conspirators, knowing exactly what was going on, it cannot be said that their economic interests were injured or put at risk in any way. Similarly, if they thought that the schemes were legitimate methods of enriching the promoters of the schemes at the taxpayers’ expense but did not expect to get anything except their initial cashback payment out of it, then they could not be described as “victims” in that sense. Conspiracy to defraud of the first variety cannot be committed without a victim with, in this case, an interest in the money which was diverted.

It is not for us, as we have said at paragraph 11, to determine anything more than whether the case as eventually put by the prosecution gave rise to a case to answer. We are not concerned in any way with any other issue about the schemes nor with any regulatory matter arising from the schemes. We have concluded simply that the judge was correct in his ruling and that the way in which the prosecution eventually put their case was misconceived. Although that is the narrow ambit of our decision, it followed that the appeal was dismissed and the undertaking given by the prosecution (which was required in this form of appeal) took effect. The prosecution were not entitled to amend their case or put it any other way.