In the wake of the judiciary's decision not to pursue a second Maxwell trial because of the level of public prejudice, Michael Levi asks if it is time to replace trial by jury in cases of fraud. The decision of Mr Justice Buckley that a second Maxwell trial on charges of defrauding the banks of Pounds 100 million - charges of defrauding pensioners having failed before a jury - would be too oppressive and that prejudice against Kevin Maxwell remained unreasonably high has re-awakened the periodic debate over "rational" methods of trial for complex frauds.

The Maxwell jury verdict and judicial ruling have made some prosecutors and politicians question whether these cases can ever produce convictions. I see no grounds for supposing that this result was an establishment "fix": the Maxwells were too disliked and their father's behaviour too egregiously painful to pensioners to generate any elite sympathy. But like previous panics about fraud trials, the decision has once again evoked the spectre of jury reform.

The prejudice arguments in this case were highly unusual for fraud suspects. More than 150,000 people were affected directly by the Maxwell collapse and, fuelled by shame at their obeisance to him in his lifetime and by the inability of the dead Robert to sue them for libel, the media carried out an awesome long-term campaign of vilification embodied in The Sun's poster headline: "Robert the Hood: Prince of Thieves". The impact of media and other social prejudice on jury decision-making in English courts remains the focus of ongoing research.

What evidence is there that juries have an inadequate understanding of fraud trials? The research carried out for Roskill's Fraud Trials Committee in 1986 demonstrated little of value to the debate about jury competence, since playing a short tape summarising a summing up to a mock jury is almost bound to under-represent levels of jury understanding at the end of a six to 17-month trial. If the jury becomes more confused as the case goes on, perhaps this is not incompetence but an entirely proper reflection of the difficulty in coming to a certain judgment about complex events which are impossible to reconstruct objectively. What evidence there is from unsystematic interviews in non-random cases (in which there was far less pre-trial prejudice than in Maxwell) suggests that jurors do approach their task very seriously, most absorb a good deal of information and change their attitude during the course of the trial. Many jurors search for apparent inconsistencies as well as reaching for a "storyline" that makes sense of the events, but this does not mean that those trained to examine consistencies of evidence carefully or those with inside knowledge of how those sorts of businesses worked would have come to the same conclusions.

³ÉÈËVRÊÓƵ

Options for reform include the prospect of trial by judge(s) alone. Judges already sit alone in long civil trials, though the stress is enormous, and is greater still in criminal cases: it is helpful to have others against whom to bounce interpretations. The advantage of judicial trial is to cut down on emotive oratory and time-wasting, the hearsay rule can be abandoned, and a much more businesslike atmosphere engendered. The disadvantage is that if the moral standards of ordinary people are the baseline for definition as "crime", how are judges to know what those standards are? Trial by judge alone exposes the judiciary to accusations of political bias.

Another option is trial by judge with assessors. This would take some of the political heat off the judiciary should there be an acquittal, and would inject some market expertise into the judgment process. However, how are the assessors to be chosen? From a "fit and proper" general panel, as they are at regulators' disciplinary hearings? Is the defence going to be allowed to propose persons or if not, to object for cause, and on what criteria? Trials may be just as long if the experts become entangled in technical debates. How will the defendants and the general public view the objectivity of the verdicts, irrespective of how fairly the assessors have actually approached the issues?

³ÉÈËVRÊÓƵ

A third option is trial by special jury. This is consistent with the original concept of the jury as persons who were familiar with the accused and the general environment in which the offence occurred. But special jurors may be no more inclined than present jurors to follow the judge's legal directions. In some respects, the legitimacy problems are even worse. Let us imagine that a jury of merchant bankers or of City folk, however distinguished, had acquitted in Blue Arrow or Guinness: would the public have believed it was anything other than a "fix" by "the City looking after its own kind"?

One possible mixed system would be to have a tribunal composed of a judge, two assessors, and two lay persons, with a requirement that at least one of each would have to agree before someone could be convicted. (Perhaps a conviction might require the agreement of the judge.) In France, the Cour d'Assises, which deals with major crimes which have a maximum penalty of five years or more people, has a jury composed of three judges and nine people drawn from the electoral register. An alternative, more modest, sociological approach would not violate the essential principle of jurors being a representative sample of the population: a quota system which would ensure a more statistically proportionate number of middle-class people than one often observes. Given their numbers in the general population, there ought to be some five middle-class people on a jury, and there could be a special jury panel to make up that proportion of the whole. This would ensure some spread of opinion and debate, and allay the fears of the political conspiracy theorists.

A further possibility is to give defendants the right to elect trial without a jury, as in North America and New Zealand and some parts of Australia. This also offers a bulwark against jury prejudice, particularly if the public has been inflamed by the media.

All of these choices involve balancing conceptions of rational decision-making following systematic processing of data, on the one hand, against both tradition and the legitimacy of the criminal justice process in the eyes of the public on the other. A distinguished colleague once argued against examining prejudice levels among potential Maxwell jurors scientifically lest this undermine confidence in the criminal justice system: ironically, the decision of a jury that had undergone a vetting process conducted by the brilliant and fair-minded Lord Justice Phillips may have achieved a similar result.

³ÉÈËVRÊÓƵ

Michael Levi is professor of criminology at the University of Wales, Cardiff.