Copyright © 1995 Ian Leigh. First Published in Web Journal of Current Legal Issues in association with Blackstone Press Ltd.
* I am grateful to my colleague Simon Cooper for reading an earlier draft. All liability for mistakes or omissions is mine alone. I am grateful also to the Home Secretary's Press Office for a prompt response to my request for materials.
This article makes recommendations for the reform of Public Interest Immunity (PII) against the background of the Matrix Churchill case and of evidence to the Scott Inquiry. Three issues are critically evaluated: the distinction between class and contents claims of immunity, whether ministers should continue to sign PII certificates, and the role of the judiciary in dealing with inspection and disclosure of documents in criminal proceedings. Detailed proposals are made to address shortcomings in each area.
PII has been much-maligned and widely misunderstood in most of the (adverse) media comment arising from the Matrix Churchill trial. In his evidence to the Scott Inquiry, the Attorney-General rightly characterised the use of the expression 'gagging order' as wholly misleading (Lyell, 1994), since what is at issue is not control over free speech nor even over the reporting of legal proceedings, but over the giving of evidence. Moreover, PII does not exist solely for the protection of the government (other bodies and individuals may benefit from it where it is in the public interest) and nor is it the only restriction in the law of evidence on the release of confidences (various professional communications are privileged under other rules). Nevertheless, the use of the such derogatory terms reflects the non-lawyer's bewilderment that there should be any limits to the types of evidence available to a court except those based on their reliability and helpfulness in resolving the disputed facts. The idea of a trial as an elaborate forensic contest played out against arcane rules of proof commands little popular appeal, and none when those rules seem in danger of promoting injustice. Notwithstanding the criticism, the rules do exist to protect some important public interests and any reform will have to take account of these.
As is well known, since the House of Lords' landmark judgment in Conway v Rimmer  AC 910, the law has developed to the point where a court is no longer bound to accept an assertion of PII made by a minister of the Crown, but may exercise independent judgment of what the public interest requires. For this purpose documents for which PII is claimed may be inspected by the court, although a two-stage approach is adopted, with the person challenging the claim of immunity required to demonstrate relevance before judicial inspection will be undertaken (Burmah Oil v Bank of England  AC 1090; Air Canada v Secretary of State for Trade and Industry (No 2)  2 AC 394). Following inspection, the claim may be upheld, or full or partial disclosure ordered. This procedure has been fashioned in civil cases where it dovetails with the process of discovery. However, it should not be forgotten that PII may be claimed by way of oral objection on grounds of public policy to certain evidence in the cut and thrust of advocacy - a long-standing model already exists in criminal trials where claims of informer privilege or its progeny, privilege for the location of visual surveillance points, arise. In the case of oral claims of privilege the notion of ministerial approval is plainly inappropriate and impracticable. Nor does the two-stage approach to inspection and relevance fit tidily into traditional criminal procedure, although it sits more easily within the growing framework of rules concerning advance disclosure of criminal evidence. PII claims may be of two kinds, class claims and contents claims. In the former the claim is based on the notion that the document is one of a class of documents the production of which would be injurious to the public service. In the latter case the claim is based on the actual or potential harm which would result from disclosure of the contents of the document in question. Once again the terminology tends to assume that privilege is being claimed for documentary evidence. Hitherto, the categories of class claims have developed largely as a matter of judicial precedent.
In discussing potential reforms of PII, the emphasis in this article will fall on two areas where shortcomings have been revealed by the Matrix Churchill trial and the evidence to the Scott Inquiry: the class/ contents distinction and the role of ministers in approving claims for immunity. The present two-stage process will then be critically assessed, and further proposals made for procedural innovation. Broadly, the central arguments are that the class/contents distinction should be reformed, at the very least so that it rests upon a legislative basis, and that ministers should either no longer sign PII certificates, or that their signatures should be subject to further independent checks. These proposals would curtail the influence of both the executive and the judiciary over PII.
Further information on the Matrix Churchill prosecution and the Scott Inquiry
Return to Contents.
As every child knows, 'Why?' is the most subversive of questions. The class/contents distinction is such an established piece of orthodoxy about PII that even the criticisms of class claims have become ritualised responses to the British disease - governmental secrecy. Such arguments tend to focus on the failure of the courts to render government accountable by allowing ministers to hide behind imprecise objections to the disclosure of evidence. Although there is plainly much force in this criticism, it relates to the unwanted side-effects rather than the rationale of class claims and so largely misses the central question: 'what practical purpose does the class/contents distinction play in the application of PII?'. Any argument for its retention or modification must be grounded on some useful function which the distinction fulfils.
The evidence given to the Scott Inquiry suggests that the most contentious category of class claims is that protecting advice given to ministers (the so-called Category B in the Matrix Churchill trial). Here the case for either abolishing the class protection or narrowing it appears overwhelming. The main argument advanced for protection of this class is the so-called 'candour' argument: that civil servants would be less candid if their advice was not so protected. There are many criticisms that can be made of the argument, both at the level of factual prediction, and in terms of principle. Lord Justice Scott was plainly sceptical of it and suggested to the Cabinet Secretary, Sir Robin Butler, that even where it was valid the argument could only properly be made for advice opinion, rather than information opinion (Butler, 1994). Malcolm Rifkind differentiated in his use of this class between policy formulation and policy implementation or application (Rifkind, 1994). However, even within the narrower categories of opinion advice and policy formulation, the argument for class protection assumes too much, for it confuses the very interests the Attorney-General helpfully differentiated in explaining that PII certificates are not 'gagging orders': the difference between disclosure for the purpose of the proceedings and publication (Lyell 1994).
Even the most ardent advocates of Open Government admit that there may be a case for embargoing access to information about policy formulation, at least until the policy is announced. But disclosure in court is not and need not be publication to the world at large. While the candour argument may have some (limited) force in relation to publication and access to information under freedom of information legislation, it surely has none in relation to the possibility of disclosure in the limited context of legal proceedings. Are not civil servants already enjoined to watch for 'The Judge Over Your Shoulder'? Put another way: even if it could be shown that civil servants would be inhibited in the advice they give by the prospect of future legal testimony, this would be argument against the candour objection rather than for it. For, a concern for legality is a highly desirable attitude for a public official to have under the rule of law, and should be nurtured and encouraged, not stifled, by the rules of evidence. If necessary, concerns about the broader public interests about publication can be met by further judicial protections such as in camera hearings and publication bans, although these should also be justified according to strict criteria.
Another contentious area of class claims concerns security and intelligence information. For example, the certificate signed by Kenneth Baker in the Matrix Churchill case asserted (to paraphrase) that the operations, sources and methods of the security and intelligence services required blanket protection, and that to answer questions about these matters in court could lead to the most serious consequences for the national interest, possible threats (including death) to security informers, and the reluctance in future of such sources to come forward for fear that they might be publicly named. Mr Baker asserted that in making such a claim he was protecting a recognised class of information, doing no more than previous Home Secretaries had done (and that the courts had approved by upholding such certificates), and that his duty was to protect the integrity and secrecy of the Security Service. The certificate had been widely drawn because the nature of the defence argument was unknown (to the Home Office, anyway) and to guard against the risk of damaging details being 'blurted out' in oral testimony at the committal proceedings (Baker 1994).
Several aspects of this are troubling. Sir Richard Scott made clear his unease about the breadth of the certificate, which could cover much innocuous information. Recommendations can be expected in his report that in future such certificates should be restricted to more narrow categories realistically reflecting damage which might occur in security cases. However, the claim that this was a class of information requiring protection also appears inconsistent with the practice revealed to the Scott Inquiry. Taken literally it would have prevented SIS (MI6) evidence being led for the prosecution, as in fact occurred at the Matrix Churchill trial. Likewise, in the Ordtech case the defence had planned to call a Special Branch officer with the agreement of the government: here a similar certificate signed by Mr. Baker seems to have been intended to provide the peg on which prosecution counsel could object to particular questions as and when necessary, but the terms were wide enough in principle to prevent the witness giving evidence at all. The unavoidable conclusion is that a class claim of this kind was being used as a tactical weapon for litigious advantage and that where it was convenient evidence plainly falling within the purported class would be given by the government itself or allowed - an impossibility if the class were taken literally and were in truth unwaivable. The growing trend since Matrix Churchill for Security Service officers to give testimony in terrorist trials (see Rimington 1994) is a further reason for doubting whether the assertion of grave security consequences arising from any form of security disclosure was anything more than a flag of convenience. The reality seems to be rather that it is a question of judgement over whether the public interest is better served by some disclosure and a successful prosecution than by no disclosure and an acquittal; seen in this way, the question about the width of the class turns on whether the executive should be allowed to make that judgment by itself or whether the court should play a role.
What role then does the class/contents distinction play? The distinction itself is an old one and dates from Duncan v Cammel Laird  1 All ER 587, 592 per Lord Simon. It is significant that the distinction pre-dates the enhanced scrutiny that the courts now conduct in PII claims. Zuckerman argues that class claims must be understood in this changed context, with the result that the distinction is of less importance than formerly (Zuckerman 1994, 707). The logic of class claims is that some documents must not be divulged not because their disclosure is harmful (in which case a contents claim could be justified), but because they would weaken future class claims. The courts themselves have implicitly recognised this conundrum in recently establishing a procedure for approving cases in which potential class claims may be waived. In R v Horseferry Magistrates' Court ex p Bennett (No 2)  1 All ER 289, QBD it was held that the CPS should seek written permission from the Treasury Solicitor before divulging material falling within a class and that the Treasury Solicitor should keep a permanent, central, record of all such approvals. The need for central co-ordination and control was based on the recognition that voluntary disclosures cumulatively weaken future class claims. Accordingly, when a class claim is upheld the effect is to subordinate the administration of justice in the present to the administration of justice in the future. In that sense it represents an indefinite deferral of justice. The only supporting arguments are that it saves judicial and governmental time in considering whether actual harm would arise from disclosure, but these are outweighed by the dangers of abuse, and, increasingly, contradicted by the practice both of the government and of the courts.
It has been argued that the responsibility for the growth of class claims lies primarily at the door of the judges themselves (Zuckerman 1994, 709 ff). After Matrix Churchill it may be expected that judges will be very sceptical of class claims. One prominent public law judge and former Treasury Counsel (Simon Brown LJ) has expressed the view judicially (in ex p Bennett (No 2), 293) that class claims have no more than a presumptive effect until displaced by some weightier public interest, and followed this up with similar comments extra-judicially (Brown 1994, 588). Likewise, in the recent House of Lords' decision of R v Chief Constable of West Midlands ex p Wiley  3 WLR 433 class protection was removed from documents in police complaints proceedings. Lord Templeman expressed the view that if the question was approached on the basis that material and relevant documents should be voluntarily disclosed unless disclosure would cause substantial harm to the public interest, the class/contents distinction 'loses much of its significance' (ibid, 437). However, his was not a unanimous view and Lord Lloyd of Berwick saw a continuing place for class claims (ibid, 461). Despite this trend of recent judicial scepticism, it is argued that the function and constitutional ramifications of this branch of the law of evidence are too important for further development to be left to the judiciary alone. There is a need for legislative reform.
The most radical option would be the simple abolition of 'class claims'. Abolition has the advantages of surgical neatness and doctrinal integrity. It would take the form of a provision that immunity could only be claimed where actual or probable harm could be established by reference to the documents or information in contention. It seems unlikely that the courts would find themselves swamped with time-consuming consideration of contents claims as a result. However, a simple change of this kind may well appear too radical for the government and their advisers.
Codification of the categories of class claims would be an inferior alternative, but nevertheless a considerable improvement on the present position. If particular classes of documents require blanket protection in the public interest, the case should be made at the political level and the protection conferred explicitly on a case by case basis in legislation. This would enable reasoned public debate about the need for and purpose of each such protection, together with the consideration of alternatives and safeguards. There exists a precedent for statutory categories of this kind in the prohibition in the Interception of Communications Act 1985 s 9 on evidence obtained by or about the process of telephone tapping (on which see R v Preston  3 WLR 891, HL). Further precedents exist elsewhere for the statutory protection of high-level policy documents: Canada, for instance, has a (much-criticised) blanket protection for Cabinet papers (Canada Evidence Act, RSC 1985 C-5, s 39). These two examples of statutory class protection are both absolute, that is they leave no room for judicial inspection or disclosure, although the first operates as a simple statutory bar and the second provides for a conclusive ministerial certificate. Neither is appealing as an exact template, but there is no reason why any reformed statutory classes adopted following the Scott report should not be more flexible and specify varying standards appropriate to the subject matter of the class, for the exercise of judicial discretions to inspect and disclose.
Any narrowing of class claims, whether by judicial activism as in ex p Wiley or by legislation, must take account of the probable effect on all actors in the process. Narrower protected classes may have the effect of making the minister's task of sifting documents more rigorous, but conversely it may result in greater presumptive force being given to the class claims by the judiciary, because of the narrower (and, presumably, more pressing) public interest at stake. This process can be seen at work in relation to class claims where national security is asserted. Judge Smedley QC's inspection of and ruling for partial disclosure on the Category C documents in the Matrix Churchill trial is against the trend, both before and since the case. In 1994, in a claim arising from industrial tribunal proceedings, the Court of Appeal held that once there was there was an actual or potential risk to national security demonstrated by an appropriate certificate, the judicial power to inspect the documents should not be used (Balfour v Foreign and Commonwealth Office  2 All ER 588). This decision was reached after the Matrix Churchill case and presumably in full awareness of it (judicial sensitivity to the controversy surrounding PII is obvious both in these judgments and in those of the House in ex p Wiley). For this reason it is important that any reform provisions should not merely specify when PII can be claimed for a class, but also the precise factors which the claimant must take into account before doing so. This brings us to the more general question of who should be responsible for approving a claim of PII.
Return to Contents.
The origin of ministerial signatures lay in the desire of the courts to be assured that the decision to claim immunity had been taken at the highest level in government and (in view of its presumptive legal force) was subject at least to political accountability. When the courts appropriated the potential for independent scrutiny of certificates, they continued to expect ministers to sign. However, the evidence to the Scott inquiry shows ministers being used as rubber stamps - not reading the documents for which they were claiming exemption in either the case of some departmental ministers nor of the Attorney-General in advising other ministers.
Perhaps the clearest illustration was the case of Kenneth Baker who, as Home Secretary from November 1990 to April 1992, signed two PII certificates - one before the committal hearings in the Matrix Churchill case (this was not in the event used although one in identical terms by his successor, Kenneth Clarke, was) and one used in the Ordtech case. The way in which these certificates came to be signed suggests careless disregard for the judicial process. The security information concerned the Secret Intelligence Service (MI6), which at that time was not publicly acknowledged to exist. It was, therefore, thought to be inappropriate for the responsible minister (the Foreign Secretary) to sign the certificate - since, in a display of the relentless syllogistic analysis and intellectual rigour which are the hallmark of an education in Greats, if an agency does not officially exist a minister can hardly be responsible for it. Mr Baker was asked to sign as Home Secretary, since he was publicly identifiable as the minister with responsibility for the Security Service. Presumably, this was close enough and no one would notice the difference. No information was placed before Mr Baker or his Home Office advisers about the detail of the security involvement in Matrix Churchill before the certificate was signed, nor did they think it necessary to obtain any details of this kind (Baker, 1994). Mr Baker's self-imposed vow of ministerial ignorance of the facts involved a dangerous circularity of thought. The belief that blanket immunity from disclosure was necessary seems to have acted as an unquestionable dogma with the consequence that it was unnecessary to delve further to establish what the public interest required to be protected. Mr Baker apparently thought to do so would have been interference with the prosecution process- a startling misapprehension in one assuming the public duty of signing a certificate proclaiming the genuineness of the public interest. To argue, as he did, that the court would act as a safeguard is wholly unrealistic for two reasons. Firstly, the judge would reasonably expect the minister to know what he was signing for. Secondly, the dire consequences which would allegedly follow from over-ruling a certificate in the terms signed by Mr Baker were anything but an invitation to tip the judicial scales in favour of inspection or disclosure. For the signing minister and the officials who advise him to be so wholly unconcerned fatally undermines the argument for ministerial signature.
Nor is this failure redressed by scrutiny by the Law Officers. The Attorney- General gave evidence to Scott that he had not read the documents relating to the various PII certificates signed in the Matrix Churchill case before (or after) advising ministers on them. He took the view that it was not the Law Officers' function to check that the claim was correctly made out: it would be usurping the position of ministers to check that documents were correctly within the class of documents which it was claimed gave them immunity. The Law Officers' role was to ensure consistency in the types of class claims which were made; however, he had no convincing answer to the obvious rejoinder that if this was to be more than an exercise in ensuring consistent wording between certificates (hardly something which required the participation of the Attorney-General) it required sight of the documents. He was forced to concede that the ministers concerned may well have had a different impression of the task being performed by the Law Officers (Lyell, 1994). For example, when Kenneth Baker signed a PII certificate in the Ordtech prosecution he did so after inquiring and being assured that the Law Officers thought it was in order to claim class immunity.
In his evidence the Attorney-General denied the suggestion that ministers should be regarded as the first of two safeguards where PII was concerned: 'The Minister is not really primarily a safeguard. A Minister is called in because someone has to tell the court that there is a conflicting public interest to the public interest of doing justice.' On the central question of whether ministers were required to sign, he maintained that the common law required them to claim immunity in respect of class documents, subject to two narrow exceptions. These were where the interests of justice clearly outweighed the public interest in disclosure and the minister was so advised by counsel, and a so-called 'conscience' exception. In focusing on the role of prosecution counsel, Sir Nicholas Lyell was emphasising that counsel would be in the best position to determine what the interests of justice would require (in clear cases; most cases would be left to the judge) from a thorough immersion in the probable lines of argument in the case. However, in so arguing he faced the inevitable difficulty of seeming to weaken the constitutional position of the minister. The 'conscience exception' likewise seems difficult to reconcile with the logic of the Attorney's position that ministers were under a Common Law duty to make class claims. Sir Richard Scott suggested that this interpretation implied that Conway v Rimmer had not merely transferred the final word from ministers to the courts on the conclusiveness of certificates, but also effected a corresponding change in the position of ministers considering the claim, so as to no longer require them to balance the relevant interests. However, it is hard to see how a minister who took this task upon him or herself and decided in favour of disclosure could be said to be usurping the court's constitutional function, as Sir Nicholas Lyell claimed.
It is also clear that the government's position is that the wording of a PII certificate allows for nuances of invitation to the judge to inspect; in the Attorney's words the certificate prepared for Michael Heseltine involved 'a light touch' and the wording was intended as a signal to the court and the defence. However, there was some dispute whether Heseltine's preference for disclosure had ever been clearly articulated to the senior prosecution counsel, Alan Moses QC. It was conceded by the Attorney-General that the certificate could have been drafted to make Mr Heseltine's personal position more explicit: for instance, that he was not doing anything more than asserting that the documents fell within the relevant class, and thought that they should be disclosed. It did not do so, merely reciting instead the view that the balance was for the judge. It was assumed by the Attorney-General that counsel would communicate Mr Heseltine's misgivings by way of amplification in oral argument about the certificate. In the event, he did not do so, because he was not aware of them, apparently due to an oversight in drafting the instructions. However, if as Sir Nicholas Lyell conceded, there is no objection to drafting a certificate to reflect a minister's personal preference for disclosure, this raises several difficulties. There may be a conflict of opinion between ministers, even about the same group of documents passing between their departments; in which case counsel will either be put into the anomalous position of receiving divided instructions or separate counsel will have to be instructed to put different departmental viewpoints. In practice, though, an argument from one government department that documents should be withheld would be fatally undermined by a different stance from another department in the same litigation, whether expressed as a failure to oppose disclosure or as a positive argument in favour of it. Alternatively, the Attorney-General might arbitrate, but this would involve more active inspection than the role suggested elsewhere in the evidence.
More fundamentally, should a minister's personal opinion be of any consequence, if the certificate is asserting the public interest? It could only be so if the minister's view is regarded as analogous to expert evidence of the needs of administration in the public interest. But if a minister signs only because somebody has to and he or she is the head of the department, then the idea that the court is being presented with the minister's opinion from the position of a governmental insider with special knowledge can be discarded and the argument for ministerial signature is further weakened.
The argument is also questionable in view of the assistance and advice ministers receive in considering signing PII certificates. In his evidence to the Scott Inquiry the Attorney-General laid particular emphasis on the position of prosecuting counsel. The Attorney's view was that where a minister wished to disclose documents falling within a protected class voluntarily, prosecuting counsel would effectively have a veto, if it was not a case where the court would clearly order disclosure (the so-called Makanjuola 'clear case exception', from the dictum of Bingham LJ in Makanjuola v Commissioner of Police  3 All ER 617, 623). This was in fact what occurred in the case of the certificate signed by Michael Heseltine, who had expressed the view on seeing the documents that they should be disclosed. However, senior prosecuting counsel thought they should not since they did not fall clearly within the exception and it was not for the minister to do the balancing of public interests. On the other hand, where counsel thought a clear case for disclosure applied, the Attorney- General thought the Minister was not bound to follow the advice, although it would be 'very strange' for him or her not to do so. The consequence of this combination is to strengthen executive secrecy. Moreover, the argument advanced against ministerial freedom to disclose - the danger of the appearance that such decisions are taken on political grounds - applies equally to decisions to withhold against the advice of counsel. A further difficulty is that prosecution counsel can only so advise on the basis of what they know about the defence case, which given the stage in the proceedings and the defendant's right not to disclose his defence in advance, may not be much. This strengthens the argument (advanced below) for a judicial buffer in whom the defence can confide before the trial, although it is clear that the government's preferred approach is phased defence disclosure (Howard 1995), in which case the issue would hardly arise.
The argument about the appropriateness of ministerial signature ultimately revolves around two distinguishable issues. The first is a matter of accountability for the exercise of exceptional advantages claimed by the executive in the course of litigation, which if abused have a clear propensity to cause a miscarriage of justice. The second is the issue of who possesses the knowledge and expertise to make a credible assertion that the public interest will be damaged by disclosure of a document. The development of the law of PII, with its progressive weakening of the role played by ministers, has concerned itself mainly with the balance of knowledge and experience between the courts and the executive. However, it cannot remain purely at that level, for the courts are part of the constitutional apparatus by which government is rendered accountable. One advantage of requiring a ministerial signature is a recognition of the seriousness of the action of claiming PII; it should never be routine, even in class cases. Any proposals for reform should be evaluated against these two factors. Three alternatives arise: to strengthen ministerial scrutiny, to recognise that it is a fiction and place responsibility firmly with the civil service, or to introduce an independent safeguard.
The Attorney-General impliedly invited the Scott Inquiry to consider the second option - replacing ministerial certificates with those signed by Permanent Secretaries, who would be equally if not more familiar with the significance of particular classes of documents and the public interest in non-disclosure. Although justifiable on grounds of knowledge and experience, a reform of this kind would constitute an apparent dilution of accountability. A move in the opposite direction would be to strengthen accountability for these decisions by making ministerial responsibility tangible and realistic, rather than notional as at present. However it is doubtful whether individual ministerial responsibility for the decision to sign a PII certificate could ever be more than token, despite the rush of ministers unusually volunteering to resign if criticised or seriously criticised in the Scott report. The problem is with the architecture of political accountability in this country - a subject which the Scott report will undoubtedly address so far as the information given to Parliament about the Arms for Iraq affair is concerned - but one unlikely to be speedily reformed, since (to change metaphors) it would require a near-total constitutional blood transfusion. In the more narrow sense of accountability to the courts, ministerial signatures may have the effect that claims of PII are taken more seriously by departments than they would otherwise be: civil servants can be expected to devote more careful attention to matters personally approved by the minister than to those handled at a lower level.
Even if accountability at the departmental level could be strengthened, there would still be a place for a system for over-arching review of PII. It is apparent from his evidence that the Attorney-General has not in practice taken active overall responsibility for ensuring real consistency in the use of PII. There is a case for introducing an independent person to regularly review and report to the public via Parliament on the use of what should be a wholly exceptional power by the executive. The reports would include annual statistics on the use of PII. Whoever fulfilled this role would also need statutory powers of access to all documents and evidence concerning the making of PII claims.
An independent standpoint might also assist in ensuring that the public interest is truly represented in PII claims. This could be achieved in two ways: by adding an additional layer of approval outside government before PII could be claimed and/or by allowing independent testimony or advice on the public interest in determining PII claims in court. The first would involve an independent person, knowledgeable about public administration, but detached from the executive, reading the documents in question and deciding whether a claim of PII should be advanced to the court. The second would involve intervening in litigation to appear as amicus curiae to assist on the broader public interests at stake in the use of PII in particular instances. One possible candidate for these tasks is the Parliamentary Commissioner for Administration. Although involvement in legal proceedings, might seem a curious task to bestow on an independent official, the task would be quite similar to that entrusted to the Ombudsman under the government's new arrangements for open government (Open Government 1993; Birkinshaw 1993; Code of Practice). However there the Ombudsman acts in an appellate role in substitution for a legal remedy, and here he would be an intermediate quasi-independent point of reference, for the courts would retain the final word on release of a document. If the Ombudsman were added as an additional stage rather than in substitution for departmental or ministerial approval, the independence of the office need not be compromised. The Ombudsman's views could be binding or merely advisory (but available to the court also). Alternatively, reference to the Ombudsman could be used as an adjunct to the court's power of inspection of the documents, at a later stage in the proceedings. Used in this way judges would have an additional resource available to supplement their technical understanding of whether a document was relevant to the case and whether it was capable of falling within the field of alleged immunity, which would assist considerably in providing an alternative assessment to the government's perspective on the question of the damage that the release of a given document might cause. One difficulty would be that the Ombudsman's jurisdiction has exempted areas corresponding to some of the grounds on which PII is claimed (notably protection of foreign relations, prevention and detection of crime, national security and so on: Parliamentary Commissioner Act 1967, schedule 3). This need not prevent amendment since the reason for these exemptions is not lack of competence, rather historical sensitivity about releasing information within these areas. Overall, the Ombudsman's view might be extremely valuable, especially if any concept of a class claim were retained in the revised law (contrary to the recommendation above).
Return to Contents.
The origin of the power of judicial inspection is in Conway v Rimmer, where it appears as an obvious concomitant to the assertion of judicial control over PII claims. However, there is a danger that the individual defects of either ministerial or judicial scrutiny, far from being remedied by adding the one to the other sequentially, will be compounded. Either the minister will undertake only a cursory examination on the basis that the judge will prevent abuse, or the judge will be excessively deferential to ministerial assertions. The possibility of mutual misunderstanding of roles is exacerbated by the two-stage process, since judges have been reluctant to inspect without clear reason to do so, and ministers have been over- reliant on the safeguard of judicial inspection.
In evidence to the Scott Inquiry, the Attorney-General argued that it was always the government's intention that the judge should read the documents in the Matrix Churchill trial, and therefore, there had been no attempted cover-up. Indeed Sir Nicholas Lyell took the position that in criminal cases the law imposed an automatic duty on the judge to inspect because of the seriousness of the interests involved. Although this principle now seems to be established by implication in a recent line of decisions dealing with disclosure in criminal proceedings (R v Ward  1 WLR 619; R v Davis  1 WLR 613; R v Keane  2 All ER 478; and see O'Connor 1992), it was less clear at the time of the Matrix Churchill trial due to the paucity of authority on the application of PII in criminal cases. Indeed, in a similar export licence prosecution of executives from Ordtech at Reading Crown Court in 1992 in which PII was claimed, the judge upheld a PII certificate without inspection and the defendants then changed their pleas to guilty, due to inability to advance their defence without the privileged evidence. Moreover, in Canada where PII has been claimed for comparable security information at criminal trials, such certificates have been upheld without inspection; in these cases it has been noticeable that the courts have refused to inspect unless the defence demonstrates relevance to a high degree: Goguen v Gibson  2 FC, FCA; Kevork v The Queen  2 FC 753. In the Matrix Churchill case itself, leading prosecution counsel argued prior to inspection that some of the documents ultimately disclosed were not relevant to the defence. Overall, it appears that the emerging test of relevance at the inspection stage in Britain is much more generous than that in civil actions (Zuckerman 1994, 721-2). Nevertheless, it would still be preferable if an automatic inspection rule were enacted to cover criminal cases, so there was no risk of the doctrine being eroded by the adoption of a more stringent test of relevance in the future.
Another way of introducing a counter-balance to the government's perspective within the proceedings would be to devise procedures to open up argument as part of the inspection process. Plainly this could not involve the other party to the litigation unless some kind on Chinese wall were to operate between the client and counsel, something of which the courts have disapproved in similar contexts on the grounds that it undermines the professional relationship of trust which should subsist between them ( R v Preston  4 All ER 638, 671 per Lord Mustill). However, an independent person acting as amicus curiae might be able to argue before the court the reasons favouring disclosure, having had sight of the documents; the arguments in favour of this have already been advanced in Section 3.
Inspection is only a possibility where a certificate is used. However, the service of a certificate is not a pre-condition to a successful claim of PII, which may also be made by oral argument. No doubt this facility is necessary to cope with unpredictable developments in the line argument and evidence may take in the cut and thrust of advocacy. The uncertainty is exacerbated because under the present regime there is no general duty in criminal cases to disclose the lines of the defence. However, the Crown view appears to be that the choice of a certificate or an oral claim of PII does not turn merely on whether the evidential point has been anticipated before the trial. Rather, the Attorney-General seemed to suggest that the certificate was something which was determined by the category of the case - for instance immunity for security and intelligence documents has traditionally been claimed by certificate (Lyell 1994). The danger is that oral submissions deprive ministers of a clear role and the court of the knowledge that a claim is ministerially approved. This is too important to be left as a question of adversarial discretion and it is hoped that the Scott report will recommend that where the Crown could reasonably foresee the point involving immunity, a certificate should always be used. If it is argued that this would overwhelm ministers in routine cases, one solution would be that such cases should be formalised by a (rebuttable) presumption, expressed in legislation against disclosure of certain recognisable categories of document or information (see the discussion of class claims in Section 2). This would have the advantage of inviting Parliamentary endorsement of PII and at the same time the provision could also deal with the situations where the presumption could be overridden and which factors should be taken into account in so doing.
The court's function in weighing the competing public interests in a criminal case is problematic. Mann LJ has acknowledged that the nature of the exercise will necessarily be different to a civil action, with a very great weight attached to the defendant's interests: R v Governor of Brixton Prison, ex p Osman (No 1)  1 All ER 108, 116. In practice most criminal cases are likely to turn on the question of relevance, because if the evidence is relevant the public interest will mean that it is too important to be excluded. For example, in an Australian terrorist case security records were inspected but then not disclosed, on grounds that they were irrelevant to the defence: Alister v R (1984) 50 ALR 41. In practice it is hard to imagine circumstances in which a trial judge would determine that material is relevant to the defence after inspection but, nevertheless, proceed to uphold a PII certificate covering it. A course of action of this kind would be unthinkable on a class claim certainly, because the public interest in not convicting the innocent must necessarily weigh heavier than the arguments for convenience in upholding such claims. If this point is accepted, then the role of PII in relation to class claims in criminal trials is effectively reduced to a procedure for reviewing a prosecutorial decision not to disclose material on grounds that it is not relevant. However, in this case it is hard to justify a separate function for PII at all since if the evidence is not relevant it ought not to be admissible and will not fall within the rules on disclosure of prosecution material. As a matter of strict logic there would be no need to serve a certificate in respect of these, still less put them before the judge for inspection.
It is arguable that the weak weight attached to class claims in criminal cases means that the prosecution should be put to a straight choice between asserting a contents claim, disclosing the material, or abandoning the prosecution. This would be an argument for the abolition of class claims in criminal cases, even were they to be retained in civil cases. This seems a powerful argument but perhaps assumes that relevance can be too easily determined at an early stage in the proceedings by the prosecution.
Requiring a defendant at a criminal trial to satisfy the relevance test has more place in contents claims. However, once again it must be rare that a judge would accept relevance but uphold the claim. Cases, especially those from Canada and Australia, in which PII has been upheld in criminal context can be explained as judicial applications of the relevance test, although, sometimes these seem to set an impossibly high hurdle for the defendant. The alternative interpretation requires an acceptance of an unpalatable possibility: that the public interest may occasionally compel the conviction of the innocent. One Canadian judgment upholding a claim of PII appears to tolerate this possibility in suggesting that probable sentence is one of several factors to be weighed in deciding whether or not to order disclosure in a criminal trial (Goguen v Gibson  1 FC 872, 904-7 per Thurlow CJ). Although it was not ultimately the sole decisive factor, the mere suggestion that non-disclosure might be tolerated on grounds of a probable light sentence in the event of conviction is sufficiently sobering to bring home the logical implications of allowing PII in criminal trials at all.
The courts have fashioned new procedures to reconcile a supposed conflict between PII and the duties of disclosure under the Attorney-General's Guidelines and under common law. The Court of Appeal has created an exceptional procedure whereby the prosecution may apply ex parte to the court for an order exempting material from disclosure in advance of the trial, and, in cases of extreme sensitivity, the defence will be unaware that such an application has been made (see R v Davis  2 All ER 643; R v Keane  2 All ER 478). These new processes have been created without regard to the class/contents distinction (Smith 1993b, 359). However, it is hard to see that there could ever be a case for ex parte hearing of the Davis kind in relation to class claims, since the issue is a general one of principle on which the defence should be entitled to express a view without compromising sensitive details about the prosecution case.
A requirement that a criminal defendant demonstrates the relevance of the information sought to be disclosed can be seen at worst as a reversal of the burden of proof, or if it requires advance disclosure of the defence case, as an incursion on the right of silence. However, it has been argued convincingly that it does not fall within the scope of the immunities traditionally recognised by the common law as comprising the right of silence (Tomkins 1993, 661-2). A possible solution to the prejudice caused, would be to use the trial the judge as a buffer in the sense of allowing ex parte disclosure by the defence so that in arguing relevance the defence is not forced to disclose its case in advance to the prosecution (in effect a reversal of the Davis-type procedure). The judge could then weigh the protection claimed (whether class or contents based) against the need asserted by the defence. Alternatively, it has been suggested that initial sifting of material for disclosure should be undertaken by independent counsel (Glynn 1993, 847-8); a procedure of this kind could equally easily accommodate a defence buffer of the type proposed. Attractive though the ex parte defence disclosure proposal is, it flies in the face of current government policy, which, following the Royal Commission's recommendations (Royal Commission 1993; Glynn 1993), tends increasingly towards defence disclosure routinely in criminal cases.
In a speech in April 1995, the Home Secretary argued that the current duties of disclosure in criminal cases were one-sided in falling solely on the prosecution. Although he cited Davis-type hearings as a means for protecting sensitive material, such as the names of informers or the location of visual surveillance points, he nevertheless found it unsatisfactory that an adverse court ruling could result in a choice between disclosure and abandoning the prosecution. Accordingly, the Home Secretary intends to bring forward proposals for legislation and a Code of Practice governing disclosure by both the prosecution and defence. This is likely to provide for staged disclosure by each side (with defence disclosure prior to prosecution further disclosure). One aim which has been singled out is the 'better protection for sensitive material, at least where the disclosure of this material is not relevant to the issues in the case' (Howard 1995). This is not the place to enter these wider arguments. Nevertheless, it should be remembered that the new prosecution duties of disclosure were created as a direct result of notorious miscarriages of justice (see especially R v Ward  1 WLR 619) and it may yet be concluded that they succeeded in averting a further injustice in the Matrix Churchill case. At the very least, the burden is on the government to provide evidence of tangible harm ie cases abandoned because of the failure to protect sensitive material. Even then the judgment may be that this is the justifiable cost of a system of open justice. However, if defence disclosure is to be introduced, it is preferable that the initial duty should be on the prosecution, with a staged response from the defence, and prosecution reply as the government proposes. Even within a system of this kind ex parte defence disclosure could play a part, reciprocal to ex parte prosecution disclosure, which will presumably be retained.
Return to Contents.
One major problem is the invidious choice that may arise between damaging one public interest (open justice) by favouring the other (justifiable secrecy), and vice versa. The cost of 'zero-sum' choices of this kind is too high: criminal defendants may successfully employ 'Graymail' to evade liability, but conversely the government may be allowed to win a civil action for reasons of secrecy unrelated to the merits. A possible way of resolving the dilemma would be to introduce a statutory power for the judge to presume facts to the disadvantage of government where PII is upheld. This would be a departure from the present position, where if evidence is subject to PII it cannot be used for any purpose. A facility of this kind could be useful in both criminal and civil cases. Somewhat similar, is a neglected provision which already exists in criminal procedure allowing for the voluntary concession of facts by the prosecution or defence (Criminal Justice Act 1967 s 10). More widespread use of this provision could reduce the need for PII in criminal cases ie if the document is relevant it should either be disclosed, the facts conceded, or the case dropped. However, a power to presume or concede facts is unlikely to cope with the hard cases where there are powerful opposing arguments for disclosure and suppression - the classic case where the defendant's argument for disclosure goes to the heart of the question of guilt or innocence (as in the Matrix Churchill case).
One last procedural question concerns who is to hear PII claims. Currently determination of PII falls to the judge trying the case in which the issue arises. However, judges determining PII claims lack the specialist knowledge which would enable an assessment of the harm which would be caused by disclosure to be made. This is one reason why they may be driven to rely on ministerial opinion. Is there a case for a specialist group of judges to handle these claims? Specialism might have the advantage of ensuring consistent decisions about disclosure. One consequence might be that specialist judges were sceptical of inflated claims of probable harm arising from disclosure, because they would have reference points for comparison. However, equally, it could result in a too-ready familiarity with and accommodation to the executive arguments in favour of non-disclosure. Since the mid-1980s, Canada has operated a system whereby the Chief Justice of the Federal Court (Trial Division) or one of a handful of other designated judges deal with PII claims involving alleged prejudice to defence, security or international relations. (Canada Evidence Act, RSC 1985 C-5, s 38; and see Cooper 1992, 134-138). There is no evidence from the Canadian experience that these designated judges adopt a less deferential attitude to governmental claims than their British counterparts. One clear disadvantage apparent from the Canadian experiment is that the claim to immunity is artificially detached from the other issues in the trial, to be resolved by a judge unfamiliar with the full background, with the danger that issues may be taken out of their proper context in the proceedings. In any event, the idea of a judicial specialism in this area is questionable: while the idea of judicial specialism in an area of law makes sense (and, for instance, underlies the arrangement of business in the High Court), it is hard to see how a judge could ever acquire specialised experience related to administration, except by doing it. Judging or advising in cases involving governmental questions does not make one an administrator, nor would it be helpful if it did, since the administrator's viewpoint on the consequences of disclosure on the public interest is no more determinative than the judge's. The specialism argument is, therefore, an unconvincing one.
Return to Contents.
1. Class claims should be abolished by legislation preventing PII from being claimed except on grounds of actual or probable harm resulting from the document or information in contention in the case.
2. Failing implementation of (1) class claims should rest upon legislatively designated categories of documents and information. Further judicial accretion to these classes should be forbidden. The statutory categories should differentiate between classes in the factors to be taken into account in exercising discretion to inspect or disclose documents or information falling within the class.
3. There is a case for either abolishing the ministerial signature requirement or strengthening it by laying down procedural safeguards in legislation. The respective roles of departmental ministers and the Law Officers should be clarified to prevent future misunderstandings.
4. Whether or not ministerial signatures are retained, the consent of an independent person (possibly the Parliamentary Ombudsman) should be required before a claim of PII can be made to a court in cases falling outside any statutory classes. The independent person should be well-versed in public administration and have full access to all documents for which a claim is made. He or she should also submit an annual report to Parliament on the operation of PII, and have power to intervene as amicus curiae in any proceedings involving contested PII questions.
5. There should be a duty to use a certificate to claim PII wherever the immunity point could reasonably be foreseen before the trial.
6. A mandatory statutory requirement of judicial inspection should apply to PII claims in criminal proceedings.
7. If class claims are retained contrary to (1), they should not apply in criminal proceedings, where PII should rest solely on actual or probable harm.
8. (Alternatively) the scope of PII in criminal proceedings should be limited to testing the relevance (according to a statutory formula) to the defence of the disputed material. Procedures should allow for ex parte representations by the defence to the judge on the issue of relevance prior to trial. The procedures established by the Court of Appeal in R v Davis allowing exceptional ex parte prosecution representations should be amended to allow for independent argument on the case for disclosure.
9. A limited power to presume facts contrary to the party whose PII certificate is upheld should be introduced.
10. PII claims should continue to be dealt with by the trial judge, without recourse to a specialist panel of designated judges.
Return to Contents.
Allan, T R S (1993) "Public Interest Immunity and Ministers' Responsibilities"  Criminal Law Review 613.
Baker, K (1994) Evidence to Inquiry into Exports of Defence Equipment and Dual Use Goods To Iraq.
Birkinshaw, P (1993) "'I only Ask for Information'- the White Paper on Open Government"  Public Law 557.
Butler, R (1994) Evidence to Inquiry into Exports of Defence Equipment and Dual Use Goods To Iraq.
Bradley, A (1992) "Justice, Good Government and Public Interest Immunity"  Public Law 514.
Brown, S (1994) "Public Interest Immunity"  Public Law 579.
Code of Practice on Access to Government Information
Cooper, T (1992) Crown Privilege (Aurora, Ontario: Canada Law Book Inc).
Ganz, G (1993) "Matrix Churchill and Public Interest Immunity" 56 Modern Law Review 564.
Glynn, J (1993) "The Royal Commission on Criminal Justice (4): Disclosure"  Criminal Law Review 841.
Howard, M (1995) Newsam Memorial Lecture, 19 April 1995.
Jacob, J (1993) "From Privileged Crown to Interested Public"  Public Law 121.
Leigh, D (1993) Betrayed: The Real Story of the Matrix Churchill Trial (London: Bloomsbury).
Leigh, I (1993) "Matrix Churchill, Supergun and the Scott Inquiry"  Public Law 630.
Lustgarten, L, and Leigh, I (1994) In From the Cold: National Security and Parliamentary Democracy (Oxford: Clarendon).
Lyell, N (1994) Evidence to Inquiry into Exports of Defence Equipment and Dual Use Goods To Iraq.
O'Connor, P (1992) "Prosecution Disclosure: Principles, Practice and Justice"  Criminal Law Review 464.
Open Government (1993) White Paper, Cm 2290 (London: HMSO).
Rifkind, M (1994) Evidence to Inquiry into Exports of Defence Equipment and Dual Use Goods To Iraq.
Rimington, S (1994) "James Smart Lecture: Intelligence Security and the Law" 3 November 1994.
Royal Commission (1993), Report of the Royal Commission on Criminal Justice, Cm 2263 (London: HMSO).
Scott, Sir Richard (1995) "Procedures at Inquiries: the Duty to be Fair", Chancery Bar Association Spring Lecture 1995, 2 May 1995.
Smith, A T H (1993a) "Public Interest Immunity in Criminal Cases"  Cambridge Law Journal 1.
Smith, A T H (1993b) "Public Interest Immunity and Sensitive Material"  Cambridge Law Journal 357.
Tomkins, A (1993) "Public Interest Immunity After Matrix Churchill"  Public Law 650.
Woodhouse, D (1995) "Matrix Churchill and Judicial Inquiries" 48 Parliamentary Affairs 24.
Zuckerman, A (1994), "Public Interest Immunity- A Matter of Prime Judicial Responsibility", 57 Modern Law Review 703.
(For further background see: Leigh, D. 1993; Tomkins 1993.)
'Having examined the facts in relation to the export from the United Kingdom of defence equipment and dual use goods to Iraq between December 1984 and August 1990 and the decisions reached on the export licence applications for such goods and the basis for them, to report on whether the relevant Departments, Agencies, and responsible Ministers operated in accordance with the policies of Her Majesty's Government; to examine and report on decisions taken by the prosecuting authority and those signing public interest immunity certificates in R v Henderson and any other similar cases that he considers relevant to the inquiry; and to make recommendations.' (HC Debs, Vol 214, col 74 (w) (November 16, 1992).
(For further discussion of the inquiry scope, procedure and format see: Leigh, I. 1993, Woodhouse 1995, Scott 1995.)