Public Prosecution Service v McKee (Northern Ireland)

The issue in this appeal is: what are the statutory consequences if the fingerprints of a defendant have been taken in a police station in Northern Ireland by an electronic device for which the legislation required approval from the Secretary of State, when such approval has never been given? In particular, is any evidence which makes use of the control fingerprints thus taken inadmissible in any subsequent court proceedings?

Article 61 of the Police and Criminal Evidence (Northern Ireland) Order 1989 (“the Northern Ireland Order”) sets out the powers of the police to take fingerprints without consent. Similar (but not in every respect identical) provision is made for England and Wales by section 61 of the Police and Criminal Evidence Act 1984. These powers are exhaustively defined; otherwise prints may only be taken with consent - see article 61(1) and, in England and Wales, section 61(1). The cases where prints may be taken without consent have been varied a little from time to time and do not need to be set out seriatim here. One of the principal cases, however, was and is where a person is in police detention having either been arrested for a reportable offence, or charged with, or informed that he will be reported for, such. Another is where he has been convicted of such an offence. Generally, fingerprints may be taken once only in the course of any single investigation, although there are now provisions permitting replacement prints to be required if the first ones were of insufficient quality to allow satisfactory analysis, comparison or matching.

Between 1 March 2007 and 12 January 2010 article 61(8B) of the Northern Ireland Order provided:

“Where a person’s fingerprints are taken electronically, they must be taken only in such manner, and using such devices, as the Secretary of State has approved for the purposes of electronic fingerprinting.”

That provision matched an intended section 61(8A) of the Police and Criminal Evidence Act 1984, which latter provision was inserted into the 1984 Act by the Criminal Justice and Police Act 2001 but was never brought into force. Both article 61(8B) and section 61(8A) were later repealed as redundant by section 112 and schedule 8 of the Policing and Crime Act 2009, with effect from 12 January 2010, and with that repeal there disappeared from England and Wales and from Northern Ireland all requirement for statutory approval of fingerprinting devices. However, although the requirement for approval existed in Northern Ireland (but not in England and Wales) from 1 March 2007, such approval was, by oversight, not given to any device until it was belatedly provided on 29 March 2009 after the omission had been noticed. It follows that for the two years from March 2007 to March 2009 article 61(8B) was in force but no device had been approved as contemplated by it.

The two appellants were defendants charged with theft in Northern Ireland. The offence was alleged to have taken place on 6 October 2007, during the two-year period mentioned. A stack of building materials had been found removed from the owners’ depot and placed apparently ready for collection by the thieves. The appellants were found nearby in a van but said that they were there innocently and had not been near the stolen materials. Their fingerprints were taken when they were detained in the police station after their arrest. A fingerprint matching Elliott’s left thumb was found on the packaging of the stolen materials. The match of fingerprints was relied upon by the Crown and proved in the magistrates’ court. The defendants were convicted.

The device used in the police station to collect the control sample of the fingerprints of each appellant was a combination of camera, scanner and computer, known as ‘Livescan’. No-one noticed that no type approval had been given for its use as required by article 61(8B). When this was appreciated, the appellants appealed to the County Court, where the appeal proceeded by way of fresh hearing ab initio. The preliminary point was taken that the evidence of comparison was inadmissible because of the absence of approval. That argument succeeded before the County Court judge but on further appeal by the Crown, by way of case stated, the Court of Appeal ruled against it.

The Livescan process was and is generally used by the police throughout Northern Ireland, as well as throughout England and Wales and Scotland and, indeed, worldwide. It has very largely superseded the traditional process of ink pad and paper. It is possible to have mobile devices as well as those located in police stations. Both are linked directly to computerised storage and searching equipment located centrally. Amongst the advantages is the ease of electronic transmission, storage and sorting of the prints taken. One aspect of that is that a set of fingerprints given at a scene or in the street can now often almost instantaneously verify or refute the identity of the person tested. Another is that international exchange of data is made much easier. Livescan devices were in general use in Northern Ireland from 2006 and throughout the two-year period 2007-2009 when type approval was required by article 61(8B).

For the appellants, the first and principal submission of Mr McMahon QC is that the language of article 61(8B) unequivocally renders a nullity any fingerprints taken by a device which has not been approved. Therefore, no legal use can be made of them. For this reason, there is, he submits, no occasion to investigate what consequences Parliament must have intended should follow from a failure to use an approved device. That would be necessary only if there were an ambiguity in the wording. There is none, and it necessarily follows that the product of an unapproved fingerprinting process is inadmissible. Any other conclusion would, he submits, leave article 61(8B) a dead letter.

The difficulty with this attractively simple submission is that the statute says nothing at all about the consequences of failure to use an approved device. There is ample precedent for such a statutory provision to be accompanied by an express provision that evidence shall only be admissible if obtained in accordance with it. An example is afforded by the statutory rules relating to evidence of speed provided by speed guns. Section 20 of the Road Traffic Offenders Act 1988 provides, for England and Wales:

“(1)

Evidence…of a fact relevant to proceedings for an offence to which this section applies may be given by the production of -

(a)

a record produced by a prescribed device, and

(b)

…

(4)

A record produced or measurement made by a prescribed device shall not be admissible as evidence of a fact relevant to proceedings for an offence to which this section applies unless -

(a)

the device is of a type approved by the Secretary of State, and

(b)

any conditions subject to which the approval was given are satisfied.”

Identical provisions are contained in the equivalent Northern Ireland legislation: article 23(1) and (4) of the Road Traffic Offenders (Northern Ireland) Order 1996. There are provisions to similar effect in section 6 of the Noise Act 1996, and in section 45 of the Antisocial Behaviour etc (Scotland) Act 2004, in respect of noise meters. The absence of this kind of explicit statutory provision from article 61(8B) thus raises the question of what consequence was intended to follow from non-approval.

This legislation was enacted against the background of the well understood general common law rule that evidence which has been unlawfully obtained does not automatically thereby become inadmissible. That has been clear since at least the decision of the Judicial Committee of the Privy Council in Kuruma v The Queen [1955] AC 197, where the defendant was charged with unlawful possession of ammunition which had been found on him as a result of an unlawful search, carried out by a policeman of insufficient seniority to make it. Lord Goddard CJ said this at p 203:

“In their Lordships’ opinion the test to be applied in considering whether evidence is admissible is whether it is relevant to the matters in issue. If it is, it is admissible and the court is not concerned with how the evidence was obtained.”

This proposition was endorsed by the House of Lords in R v Sang [1980] AC 402, which dealt more specifically with the judge’s discretion to exclude evidence which will have the effect of rendering the trial unfair (see now section 78 of the Police and Criminal Evidence Act 1984 and its equivalent, article 76 of the Northern Ireland Order). Likewise in R v Khan [1997] AC 558 evidence obtained by unauthorised surveillance and the secret recording of private conversations was admissible despite the unlawful methods by which it had been obtained. The position was summarised by Lord Fraser, with whom all other members of the House of Lords agreed, in Fox v Chief Constable of Gwent [1986] AC 281, 292A as follows:

“It is a well established rule of English law, which was recognised in R v Sang, that (apart from confessions as to which special considerations apply) any evidence which is relevant is admissible even if it has been obtained illegally.”

It is clear that this inclusive rule of relevant evidence extends equally to evidence created by an unlawful process as it does to existing material uncovered by unlawful process; the recording in Khan is an example of the former. This common law background to the legislation, of which Parliament must be taken to have been well aware, shows that inadmissibility of the fingerprints here under consideration cannot possibly simply follow from the existence of the requirement for device approval. Rather, it is necessary to examine the Parliamentary intention as to consequence.

With great respect to Mr McMahon’s principal argument, it is not correct that article 61(8B) would have no purpose, or would be a dead letter, unless its consequence were that any fingerprints obtained from an unapproved device were inadmissible. Whether or not inadmissibility is the consequence, the article still meant that a requirement by a policeman of a suspect in custody that he provide his fingerprints on an unapproved device would be one which the suspect was entitled to refuse. It might not be very likely that a suspect would be acquainted with the presence or absence of approval, but his solicitor might well be. Such a suspect could therefore refuse to provide his fingerprints on a Livescan device and he would not thereby commit the offence of obstructing a police officer that no doubt he otherwise would. Similarly, if it became known that the police were regularly using an unapproved device, there would be no defence to an application for judicial review in which the unlawfulness of their actions would be declared and, if persisted in, no doubt prohibited. Thus the clear statutory purpose of preventing the use of a device unless it is approved by the Secretary of State would be achieved. There is no need for the additional consequence of inadmissibility of evidence in order to give content to the statute.

It follows that the wording of article 61(8B) does not itself provide the solution to the issue in this appeal. It is necessary to examine the question what Parliament must have intended to be the consequence of non-approval of Livescan. The correct approach to this enquiry was explained by Lord Steyn in R v Soneji [2005] UKHL 49; [2006] 1 AC 340. It had previously been thought that statutory provisions could be classified as either mandatory (carrying the consequence of total invalidity for breach) or directory (carrying lesser consequence). The over-rigidity of that a priori approach had given rise to difficulty. At para 23 Lord Steyn said this:

“Having reviewed the issue in some detail I am in respectful agreement with the Australian High Court [in Project Blue Sky Inc v Australian Broadcasting Authority (1998) 194 CLR 355] that the rigid mandatory and directory distinction, and its many artificial refinements, have outlived their usefulness. Instead, as held in Attorney General’s Reference (No 3 of 1999) [2001] 2 AC 91, the emphasis ought to be on the consequences of non-compliance, and posing the question whether Parliament can fairly be taken to have intended total invalidity. That is how I would approach what is ultimately a question of statutory construction.”

That more flexible approach does not necessarily mean that failure to comply with statutory provisions may not have far reaching consequences. It may sometimes yield the conclusion that the inevitable consequence is total invalidity. That was the outcome in R v Clarke and McDaid [2008] UKHL 8; [2008] 1 WLR 338, where the question was whether the failure to sign an indictment nullified the ensuing trial. The statutory provisions there in question were sections 1(1) and 2(1) of the Administration of Justice (Miscellaneous Provisions) Act 1933 which provided for a bill of indictment (which had of itself no legal standing save as a proposal of charges) to become an indictment when signed. It was common ground that a valid indictment was a pre-condition to a valid Crown Court trial. It can be seen from Lord Bingham’s speech at para 18 that he faithfully posed the Soneji question, namely what Parliament had intended, when passing the 1933 Act, should be the consequence of lack of signature. Since at the time of the 1933 Act the signature was taking the place of the previously existing endorsement of the bill by a Grand Jury, the answer was inescapable, if inconvenient: the signature validated the indictment in the same way as the Grand Jury’s decision previously had done. Accordingly the absence of signature did indeed invalidate the subsequent trial, notwithstanding the fact that modern changes in the routes by which criminal cases arrive in the court of trial had in the meantime reduced the signature, in practice, to mere formality. The position had to be put right by amending legislation, in the form of the Coroners and Justice Act 2009.

Should a similar parliamentary intention be deduced from article 61(8B)? Mr McMahon relies upon the well established rule that the product of a ‘breathaliser’ test is inadmissible unless the testing device is an approved one. The cases begin with Scott v Baker [1969] 1 QB 659, decided in the infancy of the Road Safety Act 1967, which had introduced for the first time the offence of driving with blood alcohol beyond a prescribed statutory limit. The power to require a suspect to provide a laboratory blood or urine sample, by which blood alcohol could be tested, was made dependent upon a complex step-by-step procedure. The first step in that procedure was the taking of a preliminary (usually roadside) ‘breath test’. By section 7 a ‘breath test’ was defined as one carried out using a device approved by the Secretary of State. The court held that such approval was essential to the statutory steps leading to a validly required laboratory sample, and that approval must be proved. The details of the blood alcohol driving legislation have been changed from time to time since then, and breath tests of a different kind are nowadays used not simply as a screening test but to determine the blood alcohol level. However, it remains the statutory rule, under section 7(1) of the Road Traffic Act 1988 and, in Northern Ireland, under article 18(1) of the Road Traffic (Northern Ireland) Order 1995, that a specimen of breath may be required in the course of an investigation into the offences of driving with excess alcohol, or of driving when unfit through drink or drugs, or of causing death by careless or dangerous driving when over the limit or under the influence, and that what may be thus required is limited to:

“specimens of breath for analysis by means of a device of a type approved by [the appropriate person.]”

Mr McMahon is therefore right to say that a breath specimen may be adduced in evidence against a defendant not only when the result constitutes the very offence of driving with excess alcohol but also where it is simply some part of the evidence relied on to prove an offence with different components, such as driving when unfit through drink. No one doubts the rule, however, that the product of a breath test will not be admissible unless the device used is an approved one.

The statutory requirement for approval of an electronic fingerprint reader is not, however, analogous to the approval requirements in the cases of breath test or speed gun devices. Both the latter are methods of measuring something which cannot subsequently be re-measured. They capture a snapshot of a suspect’s activity. The snapshot is often itself the offence. It is the speed, as measured by the device, which constitutes the offence of exceeding the speed limit. It is the blood alcohol content, as measured by the device, which constitutes the offence of driving with excess alcohol. In other cases, the snapshot is simply part of the evidence, for example if the offence charged is careless driving, or driving whilst unfit through drink. But in both kinds of situation, the activity measured by the device cannot be reproduced to be re-measured. It is therefore entirely comprehensible that there should be a statutory requirement that the device should be approved, and that the measurements which can be relied upon in evidence should be limited to the products of such devices. That is no doubt why there are the specific statutory provisions in relation to speed guns described at para 8 above, and it is clearly why the courts have held that the requirements for approval in the case of breath tests have the like effect.

The control fingerprints taken from the appellants in the police station were not snapshots. The impressions which their fingers provided could be reproduced at any time afterwards, and would be the same. The accuracy of the Livescan readings, if disputed, could readily be checked independently by the appellants providing more samples, whether by ink and paper or by any other means, for examination by an independent expert. The ease with which this can at any time be done demonstrates that there was no need at all for Parliament to stipulate, or to intend, that the product of unapproved electronic fingerprint readers should be inadmissible. It is the fact that in the present case there was no challenge whatever to the accuracy of the control fingerprints taken from Elliott by the Livescan device; the fingerprint found at the scene matched his control prints in no less than 45 particulars and there was no sign of any reliance on expert opinion either in the magistrates’ court or, after the absence of approval was appreciated, in the County Court. But if there had been a dispute, as in other cases it is at least possible that there might be, it would have been the simplest possible matter for new control prints to be provided so that independent expert opinion could be obtained.

There appeared at first to be some limited support for the appellants’ contentions in an explanatory note which accompanied the proposed insertion into the Police and Criminal Evidence Act 1984 of section 61(8A) requiring type approval of electronic fingerprint readers. That amendment of the 1984 Act would have been achieved through section 78(7) of the Criminal Justice and Police Act 2001, had that subsection ever been brought into force. The explanatory note to that subsection (number 234) read as follows:

“Subsection (7) provides that where fingerprints are taken electronically, the device used must have type approval from the Secretary of State. This is to ensure that the device will produce images of the appropriate quality and integrity to be used for evidential purposes.”

However, the other background material shown to this court demonstrates that the purpose of the proposal for type approval was not principally the protection of the individual against risk of conviction on inaccurate evidence. The concern was much more closely related to the needs for the technology to work properly so that investigations could proceed confidently, for compatibility between police forces, both domestic and foreign, and for uniform machinery for search and comparison. The then Minister of State referred to the aim of facilitating a ‘proper evidential trail’. The House of Lords Select Committee on Science and Technology had emphasised the need for the technology to be robust. It is also clear that there was thought at one time that type approval would curtail any potential for unnecessary dispute in court about the legitimacy of electronically taken control fingerprints. The initial recommendation of the Police Scientific Development Branch had been against any stipulation for type approval. The reasons for that stance included the difficulty of formulating a test standard and the frequency of developments to many of the component parts of the system. The successful operation of Livescan in England and Wales over a decade without any type approval, as well as the experience in Northern Ireland, clearly contributed to the subsequent decision in 2009 not to commence the amendment to the English statute, and to repeal both article 61(8B) and the uncommenced section 61(8A). Overall the legislative history does not suggest any basis for concluding that Parliament intended that the consequence of use of unapproved apparatus should be the exclusion of the evidence.

Such a consequence would, it is clear, be unnecessary and inappropriate. It is unnecessary because a reading of control fingerprints can always be checked subsequently. It is inappropriate because to exclude such evidence would deprive courts of reliable and relevant material. Since the product can be checked, and the evidence it provides is relevant, it ought to be admissible. If it were not, it would not be open to the police to take further control fingerprints without the consent of the subject, because he would no longer be in detention following arrest on suspicion of the offence, nor would he have been convicted of it. If the control fingerprints were to be inadmissible, not only would there be a windfall benefit to those who have committed crimes, perhaps of great gravity, but also defendants would be unable to rely on the evidence of the fingerprints of others when it was necessary for them to do so in order to defend themselves. A defendant who wished to show that a fingerprint found in an incriminating place belonged to another person, whom he contends committed the offence rather than himself, would be unable to adduce the evidence to do so.

Some years after the provisions which we have here to construe, the Protection of Freedoms Act was enacted in 2012. Part 1, Chapter 1 contains, by way of proposed amendments to the Police and Criminal Evidence Act 1984, prospective provisions relating to fingerprints and other biometric data. Equivalent provision for the amendment of the Northern Ireland Order is made by section 9 of and Schedule 2 to the Criminal Justice (Northern Ireland) Act 2013. Neither set of provisions is yet in force but there is a proposed timetable for commencement. If and when these provisions are commenced they will provide for the destruction of fingerprints and other data in certain defined circumstances and/or after prescribed periods. There is express provision in proposed new section 63T(2) of the Police and Criminal Evidence Act 1984 (and in proposed new article 63Q(2) of the Northern Ireland Order) making inadmissible (at least “against the person to whom the material relates”) fingerprints or other data which the police have come under a duty to destroy. This proposed statutory scheme is consistent with the construction of the provisions we are considering in the present case. Where the intention is to make material inadmissible, express provision is made saying so, in the same way as it was in the statutes considered at para 8 above. Moreover, the proposed new scheme for destruction of biometric data is clearly founded on a view of individual rights which was considered to justify the consequence of inadmissibility if there is a duty to destroy the material. Such considerations do not apply to type approval for the machinery of taking fingerprints which there is no requirement to destroy.

For these reasons it is clear that the correct conclusion is that Parliament did not intend, by enacting article 61(8B), that the consequence of an absence of approval should be to render inadmissible any fingerprints produced electronically. The decision of the Court of Appeal that the evidence of Elliot’s control fingerprints was admissible was correct. It follows that this appeal must be dismissed.