In 1986 Lord Roskill's Fraud Trials Committee did not find trial by a random jury a satisfactory way of achieving justice in cases of long and complex fraud. Lord Runciman's Royal Commission on Criminal Justice, which reported in 1993, was more cautious. It pointed out that no-one really knew how juries reached their verdicts because it was against the law to ask them. It recommended a change in the law to enable research to be carried out. Lord Justice Auld was quite clear, however, in his Review of the Criminal Courts in 2001. He said, in line with Roskill's favoured option, that in serious and complex frauds the nominated trial judge should have the power to direct a trial by himself and two lay members drawn from a panel established by the Lord Chancellor for the purpose, or, if the defendant so requests, by himself alone.

After considerable parliamentary difficulty the government succeeded in including a provision in the Criminal Justice Act 2003 enabling the prosecution in cases of serious or complex fraud to apply to a judge of the Crown Court for a trial on indictment to take place before a judge without a jury. The judge would have to be satisfied that the complexity or length of the trial (or both) was likely to make the trial so burdensome to the jury that justice required that serious consideration be given to trial without a jury. The provision is not yet in force, although the Labour Party's manifesto says: "We will overhaul laws on fraud and the way that fraud trials are conducted to update them for the 21st Century and make them quicker and more effective."

The reluctance of Parliament to abolish juries in complex fraud cases is understandable. They have been with us in criminal trials since the reign of Henry II. In the 17th and 18th Centuries, with considerable determination, the jury won its independence from the trial judge in reaching its verdict. Every law student learns of the trial of William Penn, when the Recorder of London repeatedly directed the jury, keeping them in increasing discomfort, to find the defendant guilty of riotous assembly. Yet, repeatedly, the jury refused to do so, until in the end their will prevailed.

Professor Radzinowitz, in his treatise on the history of the English criminal law, quotes a foreign source on the subject of the English jury.

"Englishmen regard trial by jury as the greatest bulwark of their civil rights; and most assuredly, the advantages they derive from it are incalculable; nor is it an extravagant position to assert that this is the cornerstone of their whole constitution. This is the cement by which the political fabric is held together; the assimilating principle which unites the constituent members of the commonwealth and constrains them to guard with increasing vigilance the impartial distribution of justice".*

The retentionists have no difficulty in finding an abundance of such passages to support their point of view. It must also be said, that the existence of the jury exerts an essential discipline on the prosecution to present the case in a form that can readily be understood by untrained lay people.

Another retentionist argument is that, more often than not, the dispute is about whether the defendant behaved dishonestly. A jury, it is said, can make up its mind about this just as well as can a judge and two lay members who are experienced in the field in which the fraud has occurred. However, to decide upon someone's honesty, it is necessary to understand the context in which the dishonesty is being alleged. That involves a grasp of the complex financial transactions which so often are the vehicle for the fraud.

The abolitionists, on the other hand, claim that society has an attachment to juries which is emotional rather than logical. That was the majority view of the Roskill Committee. They also argued that for the vast proportion of our criminal business in this country juries are not used. Over 90 per cent of criminal cases are dealt with by magistrates. In a number of overseas jurisdictions trial by jury is not the norm, and it does not appear that in those jurisdictions people are clamouring for jury trial.

In England specially qualified tribunals are common. As Roskill found,

"In the vast majority of legal cases in England and Wales today the persons selected to hear them are skilled people with particular knowledge and attributes, as opposed to a random selection drawn from the public at large, which is the principle upon which jury trials are at present founded."

It is something of a paradox that for this small, but important, class of heavy fraud cases we use a tribunal consisting of 12 untrained lay people, whereas for many less important cases, or in any event many which do not involve the loss of liberty, we use highly trained adjudicators.

The abolitionists also say that complex fraud trials are not unlike cases dealt with in the Commercial Court. They are commercial cases in the criminal context. There the judges sit alone, never with juries, and the reputation of the Court is second to none. At the lower end of the system specialist tribunals deal with hundreds of thousands of cases a year, ranging from immigration to social, employment and tax issues.

Nothing appears to have changed in the 20 years since Roskill reported to alter the persuasive case he made based on these arguments for ending jury trials in complex fraud cases. Indeed, if anything has changed, it is that fraud trials have become even longer and more complex. So, if the abolitionists' case has been made out for at least the most complex frauds - and the Government may well have further difficulty in bringing into force the relevant sections of the 2003 Act - where is the line to be drawn? Which are to have juries and which are not?

The recent protocol issued by the Lord Chief Justice on "Control and Management of Heavy Fraud and Other Complex Criminal Cases" holds out the prospect of shorter trials. Better case management and disclosure which is strictly limited to what is likely to assist the defence or undermine the prosecution may reduce the number of cases that are inappropriate for jury trial. But, there will undoubtedly still be some which are too long and too complex to be tried before juries.

The Protocol reflects in part the directions of the Court of Appeal in the 1992 case of R v Cohen and others, where the Court encouraged the robust and early use of the power of severance of the indictment in order to secure a manageable and fair trial before a jury. The court also commented that trial judges would seldom have occasion to sever, if, when performing their difficult task, the prosecuting authorities framed indictments which had due regard to the limitations of jury trial. As a result, the prosecuting authorities have not in my experience overloaded indictments, although they are often accused of doing so. They have, however, endeavoured to reflect in the charges sufficient criminality to enable juries, on the evidence, to convict with confidence.

Notwithstanding adherence to these directions there have been (and I am confident will continue to be) a small number of cases with several principal defendants which have been so complex that it has been impossible to reduce them to a point where the whole indictment against all defendants can be tried in a single trial. In these cases the judge has to sever the indictment, splitting it up to be dealt with in two or more trials. By this means, the theory goes, each trial will be fair and manageable before a separate jury.

The consequence, however, is that each trial will proceed on only part of the evidence with only some of the defendants before the court, and a single jury will never see the full weight of the case against all. The trials will be more manageable, but will justice be achieved, if the jury does not see the whole picture? It is, therefore, in this small class of very complex case that we must ask whether the Court of Appeal's directions in Cohen are workable.

Is severance the answer or should we be readdressing the issue of whether there is a means by which the entire indictment can be tried in one trial?

If it is unreasonably burdensome for a single jury of twelve untrained lay people to be asked to try the whole case, the only suitable alternative is the Roskill solution of a judge sitting alone with two lay members who have skill and expertise in business generally and experience of complex financial transactions in particular.