ANDY CAPP, the cartoon character in the Daily Mirror, was once shown as the referee in a football match. Not only was his lower lip adorned, as always, by the ever-present half-smoked cigarette: his referee's shirt was ornamented by a big rosette, as worn by team supporters. To a player, who had noticed it, he was saying “Sumthin' botherin' yer?” Translated into legal terms, this was Hanif and Khan v UK [2012] E.C.H.R. 2247.

On the second day of the defendants' trial in 2007 for conspiracy to supply heroin, a juror warned the judge that he was a police officer and knew one of the police witnesses, whose evidence was crucial against Hanif. The judge, astonishingly, ruled this did not matter and in due course his fellow jurors made him foreman. The jury, readers will be unsurprised to learn, convicted – and long prison sentences were imposed.

The defendants appealed, arguing that the presence of the police officer meant that the tribunal which had tried them was not independent. In 2008 a strong Court of Appeal, containing both the Lord Chief Justice and the President of the Queen's Bench Division, upheld their convictions.

Three years later the Fourth Division of the Strasbourg Court took a very different view. In its judgment, given on 20 December 2011, it said:

It followed, said the Court, that Hanif's right to a fair trial under Article 6(1) of the Convention had been violated because the tribunal that tried him was not “independent”; and as the two defendants' cases were so closely intertwined, Khan's right to a fair trial had been violated too. The UK was ordered to pay compensation to the defendants; and at some point the Court of Appeal will presumably be hearing the appeal again, this time on a reference from the Criminal Cases Review Commission.

To readers unfamiliar with the twists and turns of criminal justice legislation in the last ten years the puzzle will not be the outcome of the case at Strasbourg, but how it ever got there in the first place. How come the Court of Appeal upheld a conviction in a case where one of the jurors was a policeman – let alone a policeman who knew another policeman who was a key prosecution witness in the case?

Traditionally, police officers were ineligible to serve on juries. In their inability to do so they were not alone: their ineligibility was part of a complicated maze of disqualifications, exemptions and potential excusals, the overall effect of which was to relieve a wide range of professionals of the need to serve on juries and to ensure that, in the main, juries were composed of persons who had nothing else to do. This was criticised, for two reasons. First, it spread the burden of jury service unfairly among the social classes, and secondly, it meant the intellectual level of juries tended to be low. (The last point was summed up in a graffito that once allegedly appeared in the urinal at the Crown Court at Snaresbrook: “Here your fate is in the hands of 12 good men too stupid to get out of jury service”.)

This issue was drawn to public attention by Sir Robin Auld in his Review of the Criminal Courts in 2001. In the course of his researches he had visited New York, where jury service was truly universal, and was favourably impressed. On the basis of what he had seen there, he recommended that most of the existing disqualifications, exemptions and excusals should be scrapped: most spectacularly, those that kept out lawyers, judges and policemen. Like most of his other recommendations that could be sold as “tough on crime”, the previous government endorsed it. By Schedule 33 of the Criminal Justice Act 2003, jury service became (in essence) obligatory for all adults between 18 and 70 other than mental patients and convicted criminals.

But how far, if at all, were juries with policemen and CPS employees sitting on them compatible with the basic requirement of a tribunal that is independent? In Abdroikov [2007] UKHL 37, [2007] 1 W.L.R. 2679 the House of Lords held that employees of the CPS – a single body operating nationwide – must not normally serve on juries in cases prosecuted by the CPS. With police officers, who are not members of one single force and who do not prosecute, the answer, they said, must depend on the facts of the individual case. As the court was split 3–2 and the majority judges delivered separate speeches saying subtly different things, the decision left it unclear as to exactly what the facts are that the answer should depend upon. However, one thing the majority all said was that a policeman should not be a juror if he shared the “same service background” as a police witness on whose contested evidence the issue of guilt or innocence would turn.

In Hanif and Khan the Court of Appeal interpreted these words as legitimating the presence of a policeman on a jury who knows another policeman who is a key witness for the Crown, provided (as in Hanif's case) the defence are saying that he is “mistaken” in his evidence rather than that he is telling lies: and on that basis they upheld the conviction. They did so in a composite judgment in which appeals in five separate cases, all turning on related points, were all dismissed. The aim, presumably, was to signal that appeals in this type of case would normally be ill received; and in the following years this policy was carried through. In seven later cases, all cited and discussed in the Strasbourg judgment, convictions were upheld in all but one. This was R v L [2011] EWCA 65, where the jury contained not only a serving police officer but, for good measure, a retired one and an employee of the CPS.

If the Court of Appeal had qualms in taking this hard line, it suppressed them in deference to what it perceived to be the democratic will. “Parliament, in its wisdom, has decided that police officers can be jurors, and so …” The Strasbourg Court, by contrast, did not feel obliged to start its analysis of the issue from the premise that Parliament, like God, must be considered to be wise. In its judgment it surveyed 14 jurisdictions, including Scotland and both parts of Ireland, in Europe and elsewhere, where jury trial is used, and found that police officers are banned from serving in all but two of them. And in Belgium and New York, the two exceptions to the rule, defendants have a right of “peremptory challenge”: that is, the right to object to any juror without giving reason – a right they formerly enjoyed in England too, but lost in 1988 in an earlier “rebalancing of justice”. In consequence, England seems to be the only jurisdiction where a police officer is eligible to serve, and if he does so, the defendant is forced to put up with it. The Strasbourg Court limited itself to condemning the UK on the narrow ground that the independence of the tribunal was compromised by a police officer on the jury who knew a prosecution witness. But the tenor of the judgment suggests that, if it had been obliged to decide the point, it would have ruled that the mere presence of any police officer on a jury is enough to violate the defendant's right to a fair trial.

If it was a good idea in general terms to make a bonfire of the tangled mass of disqualifications and exemptions from jury service, in retrospect it surely was a bad idea to make police officers eligible to serve. In practical terms, the need for judges to question police jurors about their relationships (if any) with witnesses adds a new and needless complication to the trial, and where one does serve and the defendant is convicted, a new and needless ground for possible appeal. And in theoretical terms, however honest the individual officer, in public perception a policeman is a member of the opposing team; and so like Andy Capp with his rosette, he should not be acting as a referee.

The government should now make a virtue out of necessity and, taking the initiative, reverse the change its predecessor made before another Strasbourg condemnation forces it to do so.