[2023] UKSC 36
On appeal from: [2020] NICA 26
JUDGMENT
In the matter of an application by Rosaleen Dalton for Judicial Review (Northern Ireland)
before
Lord Reed, President
Lord Hodge, Deputy President
Lord Sales
Lord Leggatt
Lord Burrows
Lady Rose
Dame Siobhan Keegan
18 October 2023
Heard on 26 and 27 October 2022
Appellant
Tony McGleenan KC
Nicholas Compton
(Instructed by Office of The Attorney General for Northern Ireland)
Respondent
Fiona Doherty KC
Malachy McGowan
(Instructed by KRW Law Advocates Ltd)
LORD REED:
In these proceedings, Ms Dalton challenges the decision of the Attorney General for Northern Ireland, not to order a further inquest into the death of her father, who died on 31 August 1988 when he unknowingly detonated a bomb which had been placed in his neighbour’s house by the IRA with the intention of killing members of the security forces. The Attorney General’s decision is challenged on the basis that it is incompatible with the state’s procedural obligation to investigate deaths that have occurred in circumstances which potentially engage the state’s responsibility, under article 2 of the European Convention on Human Rights (“the Convention”), as implemented in our domestic law by the Human Rights Act 1998. In response, the Attorney General argues in the first place that the Act does not impose any procedural obligation to investigate deaths which occurred more than 12 years before it came into force on 2 October 2000 (“the commencement date”), absent exceptional circumstances which are not present in this case.
Applying the decisions of this court in In re Finucane [2019] UKSC 7; [2019] NI 292 (“Finucane”) and In re McQuillan [2021] UKSC 55; [2022] AC 1063 (“McQuillan”), it is clear that the Attorney General’s contention is correct. As explained below, it was held in Finucane that the procedural obligation to investigate deaths under article 2, as given effect in our domestic law by the Human Rights Act, does not apply to deaths which occurred before the commencement date unless either there was a “genuine connection” between the death and the commencement date, or the “Convention values” test was satisfied (both the genuine connection test and the Convention values test are explained below). In McQuillan, the court held that the genuine connection test could not normally be met where the death occurred more than ten years before the commencement date, but that a period of up to 12 years was permissible in specified circumstances. Mr Dalton’s death not only occurred more than ten years before the commencement date: it also falls outside the maximum period of 12 years permissible in the circumstances described in McQuillan. It is (rightly) not suggested that the case is one where the Convention values test is met. There is therefore no procedural obligation under article 2, as given effect in our domestic law, to investigate Mr Dalton’s death. It follows that the Attorney General’s appeal against the contrary decision of the Court of Appeal of Northern Ireland (which pre-dated the judgment in McQuillan) must be allowed.
That, however, is not the end of the appeal in relation to this issue. Lord Hodge, Lord Sales and Lady Rose consider that the reasoning in Finucane is erroneous, since it failed to apply the strict and absolute ten year limit which, they say, was adopted by the Grand Chamber of the European Court of Human Rights in Janowiec v Russia (2013) 58 EHRR 30 (“Janowiec”). In their view, that absolute limit should have been adopted in Finucane in accordance with the “mirror principle”, explained below, which they consider to apply in this context. By the same token, Lord Hodge, Lord Sales and Lady Rose are implicitly critical of the reasoning in McQuillan, in so far as it accommodated the result of Finucane and failed to apply a strict ten year limit.
As explained below, I do not consider that the approach adopted by the European court is as inflexible as Lord Hodge, Lord Sales and Lady Rose believe. Nor do I consider that the “mirror principle” applies in this context. Accordingly, I am not persuaded that the decision in Finucane was wrong. Nor am I persuaded that the decision in McQuillan to accept claims after a lapse of more than ten years in the specified circumstances was wrong, notwithstanding that it was influenced by the desire to accommodate the precedent established by the court’s previous decision in Finucane, rather than having a specific basis in the case law of the European court.
Furthermore, in order to adopt the approach which Lord Hodge, Lord Sales and Lady Rose would in principle favour, the court would have to invoke the Practice Statement issued by the House of Lords in 1966 (Practice Statement ( Judicial Precedent) [1966] 1 WLR 1234), and depart from its decision in Finucane as qualified in McQuillan, the latter being a unanimous decision made by an enlarged court of seven justices less than two years ago. Lord Hodge, Lord Sales and Lady Rose have ultimately drawn back from advocating that step. They are right to have done so. As a general rule the court will be “very circumspect” before accepting an invitation to invoke the Practice Statement: Knauer v Ministry of Justice [2016] UKSC 9; [2016] AC 908, para 23. That is because it is “important not to undermine the role of precedent and the certainty which it promotes”: Henderson v Dorset Healthcare University NHS Foundation Trust [2020] UKSC 43; [2021] AC 563, para 87. In the circumstances of the present case, a number of factors give powerful support to that approach. They can be summarised in three propositions: (1) it is unnecessary to decide the point in order to decide the present case; (2) the earlier decisions are in my view correctly decided and, to say the least, not clearly wrong; and (3) departing from the earlier decisions would have a damaging impact on legal certainty in circumstances where maintaining a clear and consistent approach is particularly important.
In this judgment, I shall explain first why I consider that Finucane was correctly decided, and why I also agree with the explanation given in McQuillan of the circumstances in which a period of ten years might be exceeded. In the course of doing so, I shall also explain why I am not persuaded that the approach adopted by the European court is as inflexible as Lord Hodge, Lord Sales and Lady Rose believe, and why I consider that domestic courts need not in any event follow the approach adopted by the European court in this context. I shall then explain why, in any event, it would be inappropriate for this court to depart from the earlier decisions.
Why the decisions in Finucane and McQuillan were correct
The critical question concerns the temporal application of the procedural obligation to investigate deaths that have occurred in circumstances which potentially engage the responsibility of the state, imposed by article 2 of the Convention, as implemented in our domestic legal system by the Human Rights Act. In considering that issue, it is convenient to examine first the development of the case law of the European court in relation to a different question, namely the extent of its temporal jurisdiction under the Convention, before turning to the issue which arises in our domestic law and the criticisms made of Finucane and McQuillan.
The European case law
Where a death occurs before the date when the state in question acceded to the Convention (or, if later, the date when it recognised the right of individual petition: Chong v United Kingdom (2018) 68 EHRR SE2), complaints that the state has failed to comply with its substantive obligations under article 2 (either its negative obligation not to take life, or its positive obligation to safeguard life) will fall outside the European court’s temporal jurisdiction, or jurisdiction ratione temporis, as the court describes it. That much has always been clear.
What has been less clear is the scope of the European court’s temporal jurisdiction in respect of complaints that the state has failed to comply with its procedural obligation under article 2 to investigate deaths that have occurred in circumstances which potentially engage its responsibility. That is a subject on which the European court’s approach has developed over time.
In the early cases, the court proceeded on the basis that its temporal jurisdiction in respect of complaints of a breach of the procedural obligation only extended to deaths which occurred after the critical date (ie the date of accession to the Convention, or of recognition of the right of individual petition, as the case might be): see, for example, Moldovan v Romania (Application Nos 41138/98 and 64320/01) (unreported) 13 March 2001. The court’s reasoning was that the procedural obligation was derived from the substantive obligation, so that its temporal jurisdiction was the same in both cases.
The Grand Chamber departed from that approach in Šilih v Slovenia (2009) 49 EHRR 37 (“Šilih”), where it held that the procedural obligation had evolved into a separate and autonomous duty which was “detachable” from the substantive obligation and was capable of binding the state even where the death took place before the critical date (para 159). However, the court held at para 162 that “where the death occurred before the critical date, only procedural acts and/or omissions occurring after that date can fall within the court’s temporal jurisdiction”. In other words, although the death might have occurred before the critical date, the court had jurisdiction only to examine whether the state had complied with its procedural obligations in respect of its acts or omissions after that date.
The court also made it clear that its jurisdiction over complaints of breaches of the procedural obligation was not open-ended and could not extend to all deaths which had occurred before the critical date, no matter how far in the past. In a passage of central importance, the court stated at para 163:
“… there must exist a genuine connection between the death and the entry into force of the Convention in respect of the respondent state for the procedural obligations imposed by article 2 to come into effect.
Thus a significant proportion of the procedural steps required by this provision – which include not only an effective investigation into the death of the person concerned but also the institution of appropriate proceedings for the purpose of determining the cause of the death and holding those responsible to account – will have been or ought to have been carried out after the critical date.
However, the Court would not exclude that in certain circumstances the connection could also be based on the need to ensure that the guarantees and the underlying values of the Convention are protected in a real and effective manner.”
In Šilih itself, the death had occurred about a year before the critical date, and the genuine connection test was held to be met. In some subsequent cases the genuine connection test was held to be satisfied although the gap was much greater, on the basis that investigative steps had continued after the critical date. For example, in the case of Mladenović v Serbia (Application No 1099/08) (unreported) 22 May 2012 (“Mladenović”) the genuine connection was held to be satisfied notwithstanding a 13 year gap between the death and the critical date. The court referred to the fact that investigative proceedings had been under way for eight years as at the critical date but had been held to be seriously deficient, and that they were still continuing at the time of the proceedings before the court (paras 38-39).
The judgment in Šilih did not provide a clear explanation of the genuine connection test or the Convention values test. Some clarification was provided by the Grand Chamber’s subsequent judgment in the case of Janowiec, which concerned the Katyn Forest massacres of 1940. The court summarised at para 141 the three elements of the approach laid down in Šilih:
“First, where the death occurred before the critical date, the court’s temporal jurisdiction will extend only to the procedural acts or omissions in the period subsequent to that date. Secondly, the procedural obligation will come into effect only if there was a ‘genuine connection’ between the death as the triggering event and the entry into force of the Convention. Thirdly, a connection which is not ‘genuine’ may nonetheless be sufficient to establish the court’s jurisdiction if it is needed to ensure that the guarantees and the underlying values of the Convention are protected in a real and effective way.”
It is relevant to note that, according to that explanation, the genuine connection test and the Convention values test are distinct, and the Convention values test is material where the genuine connection test is not met.
The court then analysed each of those elements in turn. In relation to the genuine connection test, it stated at paras 146-148:
The court considers that the time factor is the first and most crucial indicator of the ‘genuine’ nature of the connection. It notes … that the lapse of time between the triggering event and the critical date must remain reasonably short if it is to comply with the ‘genuine connection’ standard. Although there are no apparent legal criteria by which the absolute limit on the duration of that period may be defined, it should not exceed ten years. Even if, in exceptional circumstances, it may be justified to extend the time-limit further into the past, it should be done on condition that the requirements of the ‘Convention values’ test have been met.
The duration of the time period between the triggering event and the critical date is however not decisive, in itself, for determining whether the connection was a ‘genuine’ one. As the second sentence of para 163 of the Šilih judgment indicates, the connection will be established if much of the investigation into the death took place or ought to have taken place in the period following the entry into force of the Convention. This includes the conduct of proceedings for determining the cause of the death and holding those responsible to account, as well as the undertaking of a significant proportion of the procedural steps that were decisive for the course of the investigation. This is a corollary of the principle that the court’s jurisdiction extends only to the procedural acts and omissions occurring after the entry into force. If, however, a major part of the proceedings or the most important procedural steps took place before the entry into force, this may irretrievably undermine the court’s ability to make a global assessment of the effectiveness of the investigation from the standpoint of the procedural requirements of article 2 of the Convention.
Having regard to the above, the court finds that, for a ‘genuine connection’ to be established, both criteria must be satisfied: the period of time between the death as the triggering event and the entry into force of the Convention must have been reasonably short, and a major part of the investigation must have been carried out, or ought to have been carried out, after the entry into force.”
Some points should be noted in relation to that explanation of the genuine connection test. First, para 146 states that the lapse of time between the triggering event and the critical date “must” remain reasonably short. It then states that the period “should” not exceed ten years. “Must” indicates an absolute obligation, whereas “should” indicates an obligation which is less peremptory. The same distinction appears in other language versions of the judgment (eg in the French, “doit” and “devrait”; in the German, “muss” and “sollte”). In that context, “should” can be read as meaning “should normally”, or “should in principle”.
That reading is supported by the court’s statement, in the third sentence of para 146, that there are no apparent legal criteria by which the absolute limit on the duration of the period between the triggering event and the critical date can be defined. That being so, it is most unlikely that in the next few words the court nevertheless defined an absolute limit: it had just acknowledged that there were no legal criteria which would enable it to do so. Indeed, it would be surprising in any event if the European court set a fixed time limit for which there was no authority in the Convention. Like other courts, it is an adjudicative body, not a legislature. It is therefore unlikely to regard itself as having the competence to lay down a fixed time limit of a specified number of years. On the contrary, its reference to the absence of any legal criteria which would enable an absolute time limit to be defined indicates that it was perfectly aware of the limits of its role.
I therefore do not interpret the judgment in Janowiec, as Lord Hodge, Lord Sales and Lady Rose do, as fixing an absolute limit of ten years for the genuine connection test to be satisfied. An alternative reading, which seems to me to be much more likely, is that while it is not possible to define an absolute limit to what constitutes a reasonably short lapse of time, it “should” not (scilicet, “should not normally”) exceed ten years. The last sentence of para 146 can then be understood as meaning that the time limit can be extended further into the past – further, that is to say, than the “reasonably short” period which the genuine connection test permits – only where the Convention values test is satisfied.
It is also important to note that although the temporal connection between the triggering event and the critical date is said at para 146 to be the first and most crucial indicator of the genuine nature of the connection, it is made clear by paras 147 and 148 that it is not the only relevant factor. The court explains at para 148 that two criteria must be satisfied: the period of time between the death and the entry into force of the Convention must be reasonably short, and a major part of the investigation must have been carried out, or ought to have been carried out, after the entry into force. The latter element, we are told by para 147, includes the conduct of proceedings for determining the cause of the death and holding those responsible to account, as well as the undertaking of a significant proportion of the procedural steps that were decisive for the course of the investigation. Accordingly, although the passage of time is the most important factor in deciding whether the genuine connection test is met, it is not the only one.
The conclusion that Janowiec does not impose a strict or absolute time limit is supported by the fact that the summary of the European court’s decision in relation to the genuine connection test in para 148 repeats that the period “must have been reasonably short”, without mentioning a ten year limit. It gains further support from the fact that, in Janowiec, the court cited Mladenović without any suggestion of disapproval. At para 138, Mladenović was explained as a case where the 13 year period separating the death from the entry into force of the Convention in respect of the respondent state was not seen as outweighing the importance of the procedural acts that were accomplished after the critical date.
In Janowiec the European court also clarified the Convention values test. It accepted that there could be “extraordinary situations” which did not satisfy the genuine connection test, but where the need to ensure the real and effective protection of the guarantees and the underlying values of the Convention would constitute a sufficient basis for recognising the existence of a connection (para 149). It stated at paras 150-151:
“the Grand Chamber considers the reference to the underlying values of the Convention to mean that the required connection may be found to exist if the triggering event was of a larger dimension than an ordinary criminal offence and amounted to the negation of the very foundations of the Convention. This would be the case with serious crimes under international law, such as war crimes, genocide or crimes against humanity … The heinous nature and gravity of such crimes prompted the contracting parties to the Convention on the Non-Applicability of Statutory Limitations to War Crimes and Crimes Against Humanity to agree that they must be imprescriptible and not subject to any statutory limitation in the domestic legal order.”
The court also explained that the Convention values test could not apply to events which occurred before the Convention came into being on 4 November 1950 (para 151), and therefore could not apply to the facts of Janowiec itself.
The Grand Chamber returned to the subject of its temporal jurisdiction in relation to the procedural aspect of article 2 in Mocanu v Romania (2014) 60 EHRR 19 (“Mocanu”). It summarised its ruling on the temporal component of the genuine connection test in Janowiec as follows (para 206):
“It found, in essence, that this temporal jurisdiction was strictly limited to procedural acts which were or ought to have been implemented after the entry into force of the Convention in respect of the respondent state, and that it was subject to the existence of a genuine connection between the event giving rise to the procedural obligation under articles 2 and 3 and the entry into force of the Convention. It added that such a connection was primarily defined by the temporal proximity between the triggering event and the critical date, which could be separated only by a reasonably short lapse of time that should not normally exceed ten years; at the same time, the court specified that this time period was not in itself decisive. In this regard, it indicated that this connection could be established only if much of the investigation - that is, the undertaking of a significant proportion of the procedural steps to determine the cause of death and hold those responsible to account - took place or ought to have taken place in the period following the entry into force of the Convention.” (emphasis added)
It is to be noted that this entire passage is concerned solely with the genuine connection test. The word “normally”, in the second sentence, is therefore not, in my view, alluding to the Convention values test, as Lord Hodge, Lord Sales and Lady Rose consider. It implies that ten years is not an absolute limit to the genuine connection test, but rather the outer limit which will normally apply. That supports the reading of para 146 of Janowiec suggested at paras 16-18 above.
The matter was considered again in Mučibabić v Serbia (2016) 65 EHRR 35 (“Mučibabić”), where the death in question occurred nine years before the respondent state ratified the Convention. The court set out at para 97 of its judgment “the principles concerning the temporal limitations of the court’s jurisdiction”, and stated, in sub para (ii), under reference to Mocanu and Janowiec:
“in order for a ‘genuine connection’ to be established, the period of time between the death as the triggering event and the entry into force of the Convention in respect of that State must have been reasonably short (in principle, not exceeding ten years)”.
The words “in principle” are again expressive of a rule which applies generally but not absolutely. That is consistent with the interpretation of Mocanu and Janowiec suggested above. More recent decisions and judgments of the court have continued to use the expressions “normally” or “in principle” in relation to the ten year period.
The domestic case law
The issue arising under domestic law is different from the issue at the European level. A court such as the High Court, or this court, does not exercise a jurisdiction which is subject to a temporal limitation. The question which arises in relation to the procedural obligation under article 2, as given effect in our domestic law by the Human Rights Act, is whether the obligation applies in the particular circumstances before the court.
Initially, as in the case law of the European court, it was thought that the procedural obligation imposed by article 2 was consequential upon the substantive obligation to protect life. The latter obligation could only apply domestically in respect of deaths which occurred on or after the commencement date (2 October 2000), since the Human Rights Act did not in general have retrospective effect: section 22(4). Accordingly, it was held that the procedural obligation could also apply only in respect of deaths occurring on or after that date: In re McKerr [2004] UKHL 12; [2004] 1 WLR 807, followed in R (Hurst) v London North District Coroner [2007] UKHL 13; [2007] 2 AC 189 and Jordan v Lord Chancellor [2007] UKHL 14; [2007] 2 AC 226.
That position was reassessed, following the European court’s judgment in Šilih, in In re McCaughey [2011] UKSC 20; [2012] 1 AC 725 (“McCaughey”). That case concerned deaths which occurred on 9 October 1990, slightly less than ten years before the commencement date. An inquest was pending on the latter date, and remained pending at the date of the hearing. The issue was whether the inquest was subject to the procedural obligation imposed by article 2. The court, by a majority, held that it was.
All the members of the majority recognised that Šilih had made it clear that the procedural obligation imposed by article 2 was distinct from the substantive obligation to protect life imposed by the same article. Beyond that, they gave differing reasons for their decision. It is unnecessary for present purposes to consider the views expressed. It was, however, generally accepted that inquests held after the Human Rights Act came into force should comply with the relatives’ article 2 rights, even if the death occurred before the commencement date.
This court first considered Janowiec in the case of R (Keyu) v Secretary of State for Foreign and Commonwealth Affairs [2015] UKSC 69; [2016] AC 1355 (“Keyu”), which concerned deaths that occurred in Malaya in 1948, before the Convention entered into force. Lord Neuberger of Abbotsbury, in a judgment with which the majority of the court agreed, began by considering whether any obligation had arisen under the Convention. In relation to the genuine connection test, he interpreted Janowiec as establishing “that one cannot, at least normally, go back more than ten years”: para 87. As I have explained, that statement is consistent with the language used in Janowiec, and follows the statement in the Grand Chamber’s judgment in Mocanu: para 22 above. In relation to the Convention values test, Lord Neuberger noted that Janowiec established that the test did not apply to events which occurred before the Convention came into existence: para 88. There was therefore no claim open to the claimants in Keyu based on article 2. Lord Neuberger also observed that “this is a topic on which clarity and consistency is highly desirable”: para 90.
The court returned to this topic in Finucane, which concerned the investigation of a death which occurred on 12 February 1989, 11 years and 8 months before the commencement date. The circumstances of the death were exceptionally serious from the perspective of the rule of law. The deceased was a solicitor who had been murdered, allegedly with the connivance of members of the police and the armed forces, because he acted in cases brought against the police and the government, and defended republican suspects in criminal cases. The murder had been the subject of multiple inquiries, the bulk of which, and the most important of which, took place after the entry into force of the Human Rights Act. In 2003, the inquiries up to that date were held by the European court not to meet the requirements of article 2 (Finucane v United Kingdom (2003) 37 EHRR 29). Following that decision, and the consideration of the case by the Committee of Ministers, the government established a further inquiry. One of the issues before this court was whether that inquiry was compliant with article 2, as given domestic effect by the Human Rights Act. In that regard, the respondent argued that the genuine connection test was not met, having regard to para 146 of Janowiec, since the death occurred more than ten years before the commencement date. Lord Hodge, Lord Sales and Lady Rose are critical of counsel for the Secretary of State for having submitted that the ten year period was a “pretty serious line”, but one which was “not entirely absolute” and “not entirely unporous”. However, that submission seems to me to be a reasonable description of the position after Janowiec.
The leading judgment was given by Lord Kerr of Tonaghmore, with whom the other members of the court agreed on this issue. Lord Kerr rejected the contention that the genuine connection test was subject to a strict ten year limit, pointing out that in Mocanu the Grand Chamber had referred to “a reasonably short lapse of time that should not normally exceed ten years” (Lord Kerr’s emphasis): para 107. Lord Kerr also referred to Mladenović, where, as explained earlier, the European court had found a violation of the procedural obligation in relation to a death which occurred 13 years before the critical date. He stated at para 108 that a period of ten years or less between the triggering event and the coming into force of the Human Rights Act was not an immutable requirement. The time which elapsed between the two dates was a factor of importance but, when taken into account with the circumstance that the vast bulk of noteworthy inquiry into the death had taken place since the Human Rights Act came into force, the significance of the time lapse diminished. In a sentence which is the subject of particular criticism by Lord Hodge, Lord Sales and Lady Rose, he stated (ibid) that “the decision as to whether there is a genuine connection involves a multi-factorial exercise and the weight to be attached to each factor will vary according to the circumstances of the case”. He also referred to passages in the judgments in McCaughey in which it had been accepted that an inflexible ten year limit was not essential and that the consideration that most of the investigation took place after the entry into force of the Human Rights Act was also a relevant factor, which could compensate for the length of the time lapse.
Both the decision in Finucane and the reasoning of Lord Kerr appear to me to be consistent with the approach adopted by the European court, and in any event a sensible approach to the application of the Human Rights Act. Lord Kerr was correct in proceeding on the basis that a period of time of ten years or less is not an immutable requirement of the European jurisprudence: see paras 16-18 and 22-24 above. He was also correct in his view that the lapse of time, although an important factor, is not in itself decisive: see para 147 of Janowiec (para 15 above). He was also correct in his view that the fact that the vast bulk of noteworthy inquiry had taken place subsequent to the commencement date was also important: see para 19 above. That is what I understand Lord Kerr to have meant when he stated that the decision as to whether there is a genuine connection involves a multi-factorial exercise. (If, as other members of the court consider, he is to be taken as meaning that the decision should be based on an evaluation not only of the factors to which he referred – the lapse of time between the death and the commencement date, and the extent of investigation before and after that date - but also of a potentially unlimited number of other unspecified factors, then I agree that that would be an approach which went beyond the European case law and could not be supported in the light of the subsequent reasoning in McQuillan. Lord Kerr would also, on that interpretation, have expressed himself more broadly than was necessary for his decision, since the only factors which he took into account were the two to which he referred.) On the facts, the decision in Finucane is consistent with Mladenović (para 13 above), and with the explanation of that case in Janowiec (para 20 above).
In the subsequent case of McQuillan, this court, sitting in an enlarged constitution of seven judges, considered two appeals. One concerned a death which occurred in 1972, and the other (in which the procedural obligation under article 3 was in issue) concerned the alleged torture of a number of suspected terrorists during 1971. On behalf of the Secretary of State, it was argued that the genuine connection test involved two requirements. First, there must be a temporal connection between the triggering event (the death or other act) and the critical date (domestically, the date when the Human Rights Act entered into force), which must remain reasonably short and should not normally exceed ten years. Secondly, a major part of the investigation must have been carried out, or ought to have been carried out, after the critical date. On behalf of the claimants, it was argued that the ten year limit was not inflexible and, in relation to the allegations of torture, that the Convention values test was met.
In a judgment given by Lord Hodge, Lord Lloyd-Jones, Lord Sales and Lord Leggatt with which the other members of the court agreed, the court did not depart from the decision in Finucane but observed at para 144:
“With respect to Lord Kerr JSC, he did not identify any clear principle by which one could tell when and to what extent it might be appropriate to water down a strict ten-year requirement as the Grand Chamber of the Strasbourg court had appeared to lay down in Janowiec, para 146. We have reservations as to whether Lord Kerr JSC was right to interpret Janowiec as he did.”
It continued (ibid):
“This court has not been invited to depart from its decision in In re Finucane but we note that the extension beyond ten years allowed in In re Finucane involved less than two more years. It would significantly undermine the legal certainty which the Grand Chamber sought to achieve in Janowiec if longer extensions than this were to be contemplated or permitted. Moreover, in Janowiec, para 146, the Grand Chamber emphasised that the time factor is the ‘most crucial indicator’ in relation to the ‘genuine connection’ test and that the test requires that ‘the lapse of time between the triggering event and the critical date must remain reasonably short’.”
Against that background, the court explained the circumstances in which the genuine connection test might be satisfied notwithstanding a lapse of time of more than ten years (ibid):
“In our judgment, an extension beyond the normal ten year limit of up to two years is permissible where there are compelling reasons to allow such an adjustment constituted by circumstances that (a) any original investigation into the triggering death can be seen to have been seriously deficient and (b) the bulk of such investigative effort which has taken place post-dates the relevant critical date. If in these circumstances there is an extension of no more than two years beyond the ten-year limit mentioned in Janowiec, it remains possible to describe the lapse of time as ‘reasonably short’ in accordance with the guidance in that judgment at paras 146 and 148.”
The two circumstances identified by the court in that passage reflected the circumstances in Finucane (and the reasoning of Lord Kerr in that case), and also those in Mladenović, as explained in Janowiec.
The court went on to decide that the critical date for domestic purposes was the date of the commencement of the Human Rights Act, ie 2 October 2000 (para 168). It was therefore by reference to that date that the genuine connection test, and consequently the normal time limit of ten years, with a possible extension of up to two years in the circumstances described, fell to be applied.
The criticisms of Finucane and McQuillan
Lord Hodge, Lord Sales and Lady Rose consider that the judgment in Janowiec laid down a “strict”, “absolute” or “bright line” time limit of ten years for the genuine connection test to be satisfied. They recognise that such a rule is inconsistent with the approach adopted in Finucane and McQuillan. Applying the “mirror principle”, which seeks to align the application of the Human Rights Act by domestic courts with the application of the Convention by the European court, on the basis that the purpose of the Human Rights Act is to implement the Convention in domestic law, they conclude that the reasoning in Finucane was incorrect. The implication of their reasoning is that the approach adopted in McQuillan also failed properly to reflect the European case law.
I am not persuaded by this argument. In the first place, as I have explained above, it does not appear to me that the European court has adopted an absolute rule that the genuine connection test can only be satisfied where the death occurred within ten years of the critical date. On the contrary, both in Janowiec and in the subsequent case of Mocanu the Grand Chamber expressed its reasoning in terms which allow for a degree of flexibility, and the same is true of the more recent Section judgments such as that in Mučibabić. The correctness of the decision in Mladenović has also remained unquestioned. As the case law of the European court presently stands, it appears to me that the period between the death and the critical date “must” be “reasonably short”. It “should not normally” exceed ten years. “Should” does not mean “must”. “Not normally” does not mean “never”.
Secondly, this is not a situation to which the mirror principle applies. To begin with, that principle requires domestic courts to follow any “clear and constant jurisprudence of the Strasbourg court” (R (Ullah) v Special Adjudicator [2004] 2 AC 323, para 20; see also Manchester City Council v Pinnock [2010] UKSC 45; [2011] 2 AC 104, para 48), unless there is a good reason not to do so. There is no “clear and constant jurisprudence” establishing an absolute ten year limit beyond which the genuine connection test cannot be satisfied.
Furthermore, as was explained in R (AB) v Secretary of State for Justice [2021] UKSC 28; [2022] AC 487, paras 54-59, the mirror principle gives effect to the intended aim of the Human Rights Act: to enable the rights and remedies available in the European court to be asserted and enforced by domestic courts. The mirror principle achieves that objective by aligning the interpretation of the Convention rights under the Human Rights Act with the European court’s interpretation of the rights defined in the Convention. The rationale was explained in R (Elan-Cane) v Secretary of State for the Home Department [2021] UKSC 56; [2023] AC 559, para 87:
“The Act … defines the Convention rights to which it gives effect in domestic law as the rights which are enforceable against the United Kingdom under international law. It follows that the rights given effect in domestic law have the same content as those which are given effect under international law, although they are enforceable before domestic courts rather than the European court, and against public authorities rather than the United Kingdom as a state. Since the rights have the same content at the domestic level as at the international level, it follows that the relevant articles of the Convention should in principle receive the same interpretation in both contexts. That is not to say that domestic courts are bound to follow every decision of the European court, but there should in principle be an alignment between interpretation at the international and domestic levels.”
That reasoning does not apply in the present context because the interpretation of article 2 of the Convention is not in issue. Cases such as Janowiec are instead concerned with the temporal jurisdiction of the European court: something which has no equivalent in our domestic law. The issue in domestic cases such as Finucane, McQuillan and the present case, on the other hand, is whether the defendant public authority was subject to a procedural obligation under article 2, as given domestic effect, in particular circumstances.
Nevertheless, in its decisions in relation to the scope of the domestic investigative obligation, this court has sought to follow the reasoning of the European court in relation to its temporal jurisdiction. The justification for doing so is pragmatic. The court is concerned with the approach which should be adopted by domestic courts when faced with the task of deciding whether the procedural obligation under article 2, as given effect in domestic law, applies in relation to deaths occurring before the Human Rights Act entered into force. An analogy has been drawn with the task faced by the European court when deciding whether it has temporal jurisdiction over complaints alleging a breach of the procedural obligation in relation to deaths occurring before the entry into force of the Convention, or before the right of individual application was recognised. The European court applies a genuine connection test based primarily but not exclusively on temporal proximity, set out in greater detail in cases such as Janowiec, Mocanu and Mučibabić. It also applies a Convention values test in particular circumstances. Domestic courts have followed an analogous approach. This has the pragmatic advantage of giving domestic courts a basis for establishing a temporal limit to the scope of the procedural obligation under article 2, in the absence of any relevant transitional provisions in the Human Rights Act itself. The decision in Finucane, as explained in McQuillan, reflects the element of flexibility which the European court appears to accept.
Following that approach, McQuillan should not be regarded as laying down a judicially legislated time bar. The court’s jurisdiction is to develop and apply legal principles, not to lay down precise time limits. It might find legislative authority for a time bar in the Human Rights Act, if a limitation period had been laid down by the European court and, on a proper interpretation of that Act, also fell to be applied by domestic courts. But, as I have explained, the European court has not adopted such an inflexible rule.
Against that background, the ten year and 12 year periods discussed in McQuillan should in my view be understood as marking points beyond which it is in practice inconceivable that the application of the principles laid down in the case law might result in the imposition of the article 2 procedural obligation in our domestic law. The ten year period fulfils that function in the general run of cases, and the 12 year period does so where the particular factors described in McQuillan are present. So understood, McQuillan provides legal certainty as to the limits of the principles governing the scope of the article 2 procedural obligation under domestic law, in circumstances where the Convention values test is not in issue.
Why the court should not depart from the decisions in Finucane and McQuillan in any event
As I explained earlier, the court should not in my view depart from Finucane and McQuillan in any event, for reasons which can be summarised in three propositions: (1) it is unnecessary to decide the point in order to decide the present case; (2) the earlier decisions are, to say the least, not clearly wrong; and (3) departing from the earlier decisions would have a damaging impact on legal certainty, in circumstances where maintaining a clear and consistent approach is particularly important.
It is unnecessary to decide the point
The first factor which bears on the application of the Practice Statement is that it is unnecessary, in order to decide the present case, to determine whether Finucane and McQuillan were correctly decided. As I have explained, the gap in time between the death of Mr Dalton on 31 August 1988 and the commencement of the Human Rights Act on 2 October 2000 exceeded 12 years, and is therefore beyond the outer limit indicated by McQuillan (the Convention values test not being in issue). This appeal would therefore have to be allowed, whether the court applied the approach adopted in Finucane and McQuillan or were to adopt a stricter ten year rule of the kind favoured by Lord Hodge, Lord Sales and Lady Rose. Although not entirely unknown, it would be unusual for the court to depart from one of its earlier decisions in a case where it was unnecessary to decide the point; especially where the members of the court had not themselves reached a clear and unanimous position on the question in issue: see, for example, Food Corpn of India v Antclizo Shipping Corpn [1988] 1 WLR 603, 607, Channel Tunnel Group Ltd v Balfour Beatty Construction Ltd [1993] AC 334, 343 and Mercedes Benz AG v Leiduck [1996] AC 284, 298.
The decisions in Finucane and McQuillan are not clearly wrong
As I have explained at para 5 above, this court will be very circumspect before accepting an invitation to invoke the 1966 Practice Statement, because it considers it to be important not to undermine the role of precedent and the certainty which it promotes. The court will not overrule a previous decision simply because the justices would decide the case differently today: Peninsula Securities Ltd v Dunnes Stores (Bangor) Ltd [2020] UKSC 36; [2021] AC 1014, para 49, citing Horton v Sadler [2006] UKHL 27; [2007] 1 AC 307, para 29. This principle is vitally important to the operation and reputation of a court which does not sit en banc, and whose composition consequently varies from one case to another. In such circumstances, the principle is essential to counter the risk that the outcome of cases might otherwise depend, or at least might appear to depend, on who happened to be sitting. It is also essential to enable the consistent application of the law, and its coherent development, to take place. As was said in R v National Insurance Comr, Ex parte Hudson [1972] AC 944 (“Ex parte Hudson”), pp 996-997, if a tenable view taken by a majority in the first appeal could be overruled by a majority preferring another tenable view in a second appeal, then the original tenable view could be restored by a majority preferring it in a third appeal, and finality of decision would be utterly lost. For all these reasons, there is great force in the observation made by Lord Hoffmann in relation to the Judicial Committee of the Privy Council in Lewis v Attorney General of Jamaica [2001] 2 AC 50, 90:
“If the Board feels able to depart from a previous decision simply because its members on a given occasion have a ‘doctrinal disposition to come out differently’, the rule of law itself will be damaged and there will be no stability in the administration of justice”.
That observation is equally applicable to this court.
In the present case, it seems to me that the decision in Finucane, as explained in McQuillan, was correctly decided for the reasons given above. At the very least, the two decisions were based upon what Lord Pearson described in Ex parte Hudson, p 996 as “a tenable view, in the absence of any demonstration that it was arrived at per incuriam or is for some other reason clearly unmaintainable”. As was said in that case (ibid), that is a sufficient reason for not overruling the decisions.
Departing from the earlier cases would have an unacceptable impact on legal certainty
There are in addition particular reasons for maintaining the consistency of the court’s approach to the present issue. Virtually all the leading cases in our domestic case law on this issue have concerned deaths occurring in Northern Ireland during the Troubles: In re McKerr, Jordan v Lord Chancellor, McCaughey, Finucane, McQuillan and the present case. Decisions taken now about the investigation of deaths which occurred during the Troubles, whether taken by the legislature, the executive or the judiciary, can affect the long and difficult process of reconciliation between the different sides of that community. In that context, the court has to bear in mind that to hold that the reasoning in McQuillan is wrong and that Finucane was wrongly decided might affect ongoing proceedings in respect of Mr Finucane’s death, as Lord Burrows and Dame Siobhan Keegan explain at para 339 below. It might also affect inquests which are currently outstanding in respect of other Troubles-related deaths, and applications for the holding of an inquest.
Less than two years ago, an enlarged constitution of this court considered the reasoning in Finucane in the case of McQuillan. The correctness of the decision in Finucane was confirmed, and the correct approach to the application of the genuine connection test in our domestic law was authoritatively established. It is difficult to see what has changed since McQuillan. There has been no subsequent change in the case law of the European court. I am not persuaded that there is an adequate justification for departing from the decision in Finucane, as explained in McQuillan, even if this court disagreed with it.
Conclusion
For all these reasons, and those given by Lord Leggatt, Lord Burrows and Dame Siobhan Keegan, with which I am generally in agreement in all important respects, I would hold that this appeal should be allowed. Since Mr Dalton’s death occurred more than 12 years before the Human Rights Act came into force, and the Convention values test is not in issue, the Attorney General’s decision not to order a further inquest into his death cannot be challenged under that Act.
LORD HODGE, LORD SALES AND LADY ROSE:
Introduction
This appeal calls upon this court to look again at the important question of whether families who allege that there has been a failure to investigate the death of their relative in a way which complies with article 2 of the European Convention on Human Rights can bring proceedings before the domestic courts when that death occurred before rights under the Convention were “brought home” by the coming into force of section 6 of the Human Rights Act 1998 (“the HRA”).
Sean Dalton died in an explosion on 31 August 1988 when he accidentally triggered a booby-trap bomb that had been set by the Provisional Irish Republican Army (“the IRA”) in a flat at 38 Kildrum Gardens in Derry/Londonderry. The bomb was intended to kill members of the security forces who the IRA hoped would enter the premises. Mr Dalton and the two other people who were killed by the explosion were the entirely innocent victims of that terrorist attack. The investigations that were carried out in the aftermath of their deaths did not identify the perpetrators.
The daughter of Sean Dalton has brought these judicial review proceedings to challenge the decision taken by the Attorney General for Northern Ireland (“AGNI”) on 2 October 2014 not to order a further inquest into Mr Dalton’s death. She and the other members of Mr Dalton’s family rely on evidence that was investigated by the Police Ombudsman of Northern Ireland that shows that the police were aware that the IRA were trying to lure the security forces to a location at or near the premises in which the bomb had in fact been placed. The police directed their officers and the army to treat the area including Kildrum Gardens as being “out of bounds” to the security forces. But they did not take steps to warn the local residents or the leaders of the community of that danger. Ms Dalton believes that a fresh inquest would help to get to the bottom of what precisely the police knew and when, and why they took the decisions they did. She argues that the investigations so far have not satisfied the United Kingdom’s obligations under article 2.
The AGNI argues that whatever the merits of Ms Dalton’s claim under article 2, it is not a claim that she can bring before the domestic courts under section 7 of the HRA. Mr Dalton’s death occurred just over 12 years before the HRA came into force on 2 October 2000. The AGNI submits that according to the case law of the European Court of Human Rights (“the Strasbourg Court”), if properly applied by the domestic courts to the HRA, there is no right under section 7 to bring an action where the death to be investigated took place more than 10 years before the HRA entered into force.
The main issue in this appeal is how firm that 10 year cut off period is. Ms Dalton relies on the decisions of this court in In re Finucane [2019] UKSC 7; [2019] NI 292 (“Finucane”) and In re McQuillan, in re McGuigan, in re McKenna [2021] UKSC 55; [2022] AC 1063 (“McQuillan”) to argue that there is some flexibility in the application of the cut off date whether it is a 10 year or 12 year cut off . The AGNI argues that this court should depart from those cases because they are inconsistent with the case law of the Strasbourg Court. They are also inconsistent with the earlier decision of this court that established that the HRA was intended to operate in respect of deaths prior to commencement only to the same extent as the Convention has been held by the Strasbourg Court to operate in respect of deaths prior to the respondent State’s adoption of the Convention.
Mr Dalton’s death and the PONI Report
The explosion and its immediate aftermath
The deaths of Sean Dalton, Sheila Lewis and Thomas Curran as the result of the explosion in Derry/Londonderry on 31 August 1988 are among the many sad and terrible deaths that occurred during the Northern Ireland Troubles.
Mr Dalton, Mrs Lewis and Mr Curran all lived in a block of flats at Kildrum Gardens in the residential area of Creggan in Derry/Londonderry. Their neighbour, a vulnerable young man known as Person A, lived at 38 Kildrum Gardens. Person A went missing in mid-August 1988. His neighbours realised that they had not seen him around for a while and, concerned about his welfare, they went to his flat to check up on him. When Person A did not answer their calls, Mr Dalton was able to enter the property through a window. He looked around but could not see Person A. When Mr Dalton went through the flat and opened the front door, there was a large explosion, which was later identified as a bomb. Mr Dalton was killed instantly as was Mrs Lewis. Mr Curran died of his injuries on 31 March 1989.
The IRA later admitted that it had planted the bomb in Person A’s flat. Person A and a friend of his who had been visiting him at the time, a 15 year old boy known as Person B, had been abducted by masked members of the IRA on 25 August 1988. Person A and Person B were held by the IRA at another location. The bomb had been placed in the flat by the IRA some five days before the explosion and the flat had, the IRA said, been monitored on a 24-hour basis by IRA members to target security forces. Their operation had gone wrong. Person A and his friend were released onto a public road after the explosion on 31 August 1988.
A police investigation followed into the murders. It was conducted by a senior investigating officer of Detective Superintendent rank with about 22 officers assisting him. The police gathered about 47 witness statements, carried out forensic enquiries and made several arrests, two on the date of the incident and 11 later on. The deponents included Person A and Person B who had been abducted with him by the IRA. The police investigation did not result in any individual being charged. An inquest into Mr Dalton’s death was held on 7 December 1989. The coroner for Derry/Londonderry found that Mr Dalton died from injuries received when the bomb detonated at around 11:50 am on 31 August 1988. The inquest and the police investigation had run their course by 1990 and the matter rested there until 2005.
The PONI Report
In February 2005, Mr Dalton’s son Martin lodged a complaint with the office of the Police Ombudsman for Northern Ireland (“PONI”). Martin Dalton’s complaints focused on the failure of the police to alert the local community to the threat of a bomb which they knew had been planted in Kildrum Gardens. His concern was that the police had failed to warn in order to protect a police informant and had thereby failed to uphold his father’s right to life. He also complained that the police had failed properly to investigate his father’s death.
The Office of the PONI was established by the Police (Northern Ireland) Act 1998 for the purpose of carrying out independent investigations of complaints relating to, amongst other things, the conduct of the police. The PONI is empowered to investigate historical complaints by the Royal Ulster Constabulary (Complaints etc) Regulations 2001 (SI 2001/184) if the PONI considers that there are grave and exceptional circumstances.
The PONI, Dr Michael Maguire, issued a public statement on 10 July 2013 (“the PONI Report”). He described the work carried out by his investigation team, gathering documents including intelligence from various sources, launching a public appeal for witnesses and interviewing some of the witnesses. At the outset of his Report, the PONI stated that he was clear that the responsibility for the deaths of Mr Dalton, Mrs Lewis and Mr Curran rests with those who planted the bomb. The scope of his investigation was to determine if there was any evidence of police misconduct or criminality in relation to the matters raised by the complaints.
He found that during July and early August 1988, the police were very aware of potential terrorist activity in and around Kildrum Gardens as a result of phone calls and intelligence they had received (para 7.6). On 5 August 1988, Special Branch received intelligence, assessed as reliable, that republican paramilitaries intended to plant a booby-trap type bomb in a house in the Derry/Londonderry area, and then stage an incident designed to prompt police officers to carry out enquiries during which they would be targets of the concealed bomb. Various incidents were then analysed by the PONI, incidents which the Dalton family believe pointed the police to the fact that 38 Kildrum Gardens was that house.
First, there was an incident on 25 August 1988 where a car that might have been involved in an attack on a police barracks at Rosemount Police Station had been found late that evening outside the block of flats in Kildrum Gardens with its doors open. A call to the police told them that a man had been seen running from the car, shouting that there was a bomb in the car. The car then exploded in the middle of the night. The PONI found that the army officer tasked with dealing with the abandoned car viewed the incident as an attempt to lure security personnel into that area and to ambush them. The vehicle was not examined until daylight. The remains of an IRA improvised grenade were recovered from inside the car; probably that had detonated the larger bomb inside the car. The police were not able to determine if the car had been used in the attack on Rosemount Police Station but did record that it was clear that terrorists intended to kill security forces who would have examined the vehicle.
The second incident occurred two days later, just after midnight on 28 August 1988 when there had been a robbery at a fish and chip shop. One of the robbers had apparently dropped a charity event application form bearing the name and address of Person A. The form was handed to the police by one of the shop staff. A Detective Sergeant at the time, now retired, had been advised not to pursue enquiries with Person A.
Thirdly, on 25 August 1988 a member of staff at St Patrick’s Care Home where Person B lived reported to the police that he had absconded. The police were aware that on previous occasions this friend had gone to stay with Person A in Kildrum Gardens but they did not initiate any enquiries about the safety of Person B. There was no evidence to suggest that the police knew, prior to the explosion on 31 August, that Person A and Person B had been abducted. Early on the morning of 30 August, an anonymous call was made to the Care Home claiming that Person B was staying with Person A at 38 Kildrum Gardens and that he might be at risk there. The caller asked the staff member at the Care Home if it would be possible to pass the message on to the police but it appears the message was not passed on to the police.
The PONI found that the police interpreted these incidents as an attempt by the IRA to lure them into a trap. Of particular relevance to these proceedings was the decision of the police force at 4:56 pm on 26 August 1988 to declare that an area including Kildrum Gardens was placed “out of bounds” until further notice. The instruction directed that there were to be no police foot patrols, vehicular movement, stopping or transit in the area. This information was recorded in the local police station’s record and army operations were also informed: para 7.56. The PONI said that an area would be placed “out of bounds” for two reasons. First, if it was suspected or known that an explosive device or ambush was likely within the particular area; or secondly if a covert operation by security forces was taking place or planned to take place in that particular area: para 6.10. He said that in relation to the first reason:
… the decision to designate an area as ‘out of bounds’ was taken by a Sub-Divisional Commander (or their deputy) on receipt of information to the effect that an imminent threat to life existed in a particular geographical area. The designation made it clear that no police patrols were to enter that area without the prior approval of the Sub-Divisional Commander, or where relevant from local Special Branch officers, who had consulted their supervisors before approving the same.
Prior to lifting restrictions on an area that was ‘out of bounds’, the Sub-Divisional Commander had to decide whether a full clearance operation of the whole ‘out of bounds’ area was necessary. The areas were purposely kept as small as possible to contain the threat and to ensure that the scale of policing activity was maintained as far as possible. In respect of areas so designated, records were maintained to include details such as the parameters of the ‘out of bounds’ area imposed and parties notified.”
The PONI was keen to understand more about the decision to place the area including Kildrum Gardens out of bounds. The direction had been given after the suspicious car had exploded and so no longer posed a threat to the security forces. The PONI recorded that Special Branch had received intelligence on 26 August 1988 that the car suspected to have been used in the attack on the Rosemount Police Station had been abandoned “convenient to a house” where a booby-trapped bomb was planted. In other words, that the danger was not only the danger of a possible attack on those going to investigate the car parked in Kildrum Gardens but the danger of a bomb in a house nearby. He concluded (para 9.23) that, although his investigation had not established the exact time at which that piece of intelligence had been received, it appeared likely that the out of bounds instruction at 4.56pm that day was made on the basis of that information.
He said that his investigation found no evidence of a contingency plan by the police to mitigate the threat of the bomb (para 7.60):
“The declaration of ‘out of bounds’ may have been a sound tactical response when, for example, the threat is from a sniper intent on killing a member of the security forces. However, a bomb is less discriminatory and I believe a more diverse tactical response was required.”
He noted that in 1988 it was not the custom or practice to notify the public of an area being placed out of bounds. However, he believed that the police should have instigated further proactive investigations before the explosion to identify the actual location of the bomb. This could have included a review of vacant premises and approaches to community leaders both to inform them of the danger but also to try and identify the location of the bomb from local community intelligence. In his conclusions, the PONI said:
I can find no evidence to suggest that the police put any plan in place to mitigate the real and immediate threat from a bomb in Kildrum Gardens. The use of ‘out of bounds’ was not by itself the right response to protect the community from the potential non-discriminatory attack posed by a bomb. This was even less acceptable considering the community were not told of the ‘out of bounds’ and the threat that lay behind it.
Again, I can see no actions put in place by the police prior to the explosion, to disrupt the terrorist activity nor apparently was there any plan for an evacuation of the area identified as the most likely location.
As such, my only conclusion must be that the police were very aware of the threat of the bomb, its location and their own duty to protect the public and maximise the safety of the police and security staff involved in any response. It is apparent that there was no contingency put in place to protect the public from the bomb, and whilst the responsibility for the murders remains with the bombers, there was a failure by the police to protect the lives of the local community who were in such a real and immediate danger.
The RUC failed to do all that could reasonably be expected of them to avoid a real and immediate risk to life which they knew about. Whether by today’s standards or those of 1988, there was a failure to uphold Mr Dalton’s right to life.”
The PONI also referred to the police investigation of the murders. He described this as “flawed and incomplete”. There had been a failure to carry out comprehensive house-to-house enquiries, to follow up on forensic work and to preserve and manage the investigation documentation. There was little or no communication between the murder investigation and the families of the deceased, leaving them bewildered and frustrated. He said that the investigation had not been left in a fit state to respond to new intelligence or evidence that may have come to the fore.
The PONI did not find evidence to support the contention that police action or inaction was motivated by a desire to protect a police informer.
The Dalton family point to what the PONI said about some lack of cooperation he encountered in the course of his investigation. This is relevant to the question of the advisability of a fresh inquest because a coroner would have powers to compel witnesses to attend and give evidence whereas the PONI has no such powers.
For example, the PONI says that he had not been able to pinpoint the time at which the “convenient to a house” intelligence had been received:
… Therefore, in order to gain greater clarity as to the police assessment of the information received, my investigators made requests to meet with, and interview, a significant number of former police officers who were in various relevant roles and levels of seniority within RUC Special Branch operating in the relevant area of Derry/Londonderry in 1988. No former Special Branch officers co-operated with my investigation. The then Divisional Commander, now retired, also chose not to co-operate with my investigation.”
Further, as regards the steps that had been taken after that intelligence had been received including the decision to place Kildrum Gardens out of bounds, he said:
In the absence of records documenting strategic decisions on the assessment of this information and rationale for actions taken, it is preferable to seek accounts from those involved at the time. A substantial number of retired police officers who were in key positions to assist this investigation were approached but declined to assist. This significantly hampered my investigation and examination of this case.”
He made a similar comment in relation to his analysis of the management of the murder investigation following the deaths. He said that he had sought the cooperation of the Senior Investigating Officer and his deputy to explore the management of the murder investigation “but they did not engage with my investigators” (para 8.7). He expressed serious concern “at the apparent subsequent loss of significant documentation concerning the management of the investigation but also in relation to actions, such as house-to-house enquiries and the results of some significant house searches and forensic recoveries” (para 8.8).
However, the PONI described his investigation as “wide-ranging and thorough”. Witness statements were taken by his team and documents and intelligence material was analysed and assessed. A public appeal for witnesses was made in the area of Derry/Londonderry, as a result of which people came forward and provided information. He also recorded that some retired police officers did provide valuable information and context to the investigation: para 1.5. He said, for example, that police officers had been interviewed by his investigators and confirmed that they had been involved in house-to-house searches over a large area: para 8.31. He accepted that the majority of forensic evidence had been destroyed in the explosion: para 9.50.
The PONI recognised the “serious operational challenges” facing the police in Derry/Londonderry at the time of this bombing and murders. Between July and August 1988 alone 73 areas had been placed “out of bounds” and there were more than 160 recorded security related incidents in the area. The police were aware that their patrols and the movements of both on and off duty officers were being targeted for attack (paras 6.5-6.6). In particular, the police suspected that attempts were being made to draw them into areas where they would be attacked and they were rightly cautious when planning a response to a report of a terrorist incident or other incoming intelligence.
Despite the shortcomings the PONI had identified in his ability to investigate, he set out comprehensive responses to the complaints made to him by the Dalton family and his conclusions as to whether there was evidence of criminality or misconduct by any member of the RUC. He concluded:
There was sufficient intelligence and information available to the police to have identified the location of the bomb in 38 Kildrum Gardens or very close by: para 9.26. Steps could and should have been taken to locate the threat and warn the local community. Failure to do this resulted in the police “not fulfilling their duty to protect the public”. The allegations that the police failed in their responsibilities to advise the local community or its leaders of possible terrorist activities in the area and to uphold Mr Dalton’s life were substantiated: paras 9.34 and 9.47.
He had not been able to find any evidence to support the allegation that the police’s conduct was an attempt to protect an alleged informant: para 9.36. The allegation that the police knowingly allowed an explosive device to remain in a location close to where the public had access in order to protect a police informant was not substantiated: para. 9.39.
There had been failures to collect and process intelligence due to poor systems such that potential opportunities to link intelligence and significant incidents were not fully pursued: para 9.42. The allegation that the police failed in their duty properly to investigate the death of Mr Dalton and Mrs Lewis was substantiated: para. 9.57.
The use of ‘out of bounds’ was not by itself the right response to protect the community from the potential non-discriminatory attack posed by the threat of a bomb in Kildrum Gardens. There was a failure by the police to protect the lives of the local community who were in such a real and immediate danger.
In the concluding paragraphs of his Report, the PONI said:
The obligations of police in respect of protecting life are now properly defined by the PSNI [Police Service of Northern Ireland] and therefore I make no specific recommendations in respect of this matter.
It is important that the Chief Constable reflects on the circumstances surrounding these events, to satisfy himself that the tactical and strategic responses available to his officers continue to be effective in mitigating threats and risk to life.”
The Retired Police Officers’ Association Response
Following the publication of the PONI Report, there was a response from the Retired Police Officers’ Association issued in October 2013 called “The Good Neighbour Bombing”. The Association concluded that the PONI had not discharged the onus of proof that rested on him before making the “extremely grave determination” set out in the Report, and criticised the reliance on hindsight in treating the events as more foreseeable than they were at the time.
The Association agreed that by late evening on 26 August, the police had been alert to the probability that the attack on the Rosemount Police station had been part of an elaborate plan by the IRA to lure the security forces to attend the vehicle, thereby triggering the booby trap device in the car, thereby further triggering a wider search of the homes in the area of Kildrum Gardens leading to the police eventually entering the location where the second booby-trap device was planted to cause further deaths. However, they said that the state of the police’s knowledge was not of sufficient detail to allow for any response to be made, other than that of covertly placing the suspected geographical area out of bounds until further intelligence could be obtained, or the possible house in which the suspected device was located determined by other means: page 14.
The relevant law in outline
The United Kingdom was one of the States which drafted the Convention and was one of the first to ratify it. The Convention came into force for the United Kingdom on 3 September 1953 and the United Kingdom recognised the right of individuals to petition the Strasbourg Court directly on 14 January 1966, pursuant to article 25 of the Convention.
Article 2 of the Convention provides:
“Everyone’s right to life shall be protected by law. No one shall be deprived of his life intentionally save in the execution of a sentence of a court following his conviction of a crime for which this penalty is provided by law.”
It has long been established in the case law of the Strasbourg Court that article 2 imposes not only substantive obligations on the state to protect life in certain circumstances but also a procedural obligation to investigate deaths. The development of this procedural obligation from its first recognition in McCann v United Kingdom (1995) 21 EHRR 97 (“McCann”) and the nature and scope of the duty were described by Lord Bingham of Cornhill in R (Amin) v Secretary of State for the Home Department [2003] UKHL 51, [2004] 1 AC 653 (“Amin”). The obligation to ensure that there is some form of effective official investigation when individuals have been killed as a result of the use of force is not confined to cases where it is apparent that the killing was caused by an agent of the state: see the judgments in Jordan v United Kingdom (2001) 37 EHRR 2 and Edwards v United Kingdom (2002) 35 EHRR 19 and in Amin itself. Lord Bingham also noted in Amin that a properly conducted inquest can discharge the state’s procedural obligation as established in McCann.
Those cases also make clear that the procedural obligation arises even where there is no doubt who was directly responsible for killing the deceased. In Amin, Lord Bingham described the essential purpose of an article 2 compliant investigation in such cases as (para 31):
“… to ensure so far as possible that the full facts are brought to light; that culpable and discreditable conduct is exposed and brought to public notice; that suspicion of deliberate wrongdoing (if unjustified) is allayed; that dangerous practices and procedures are rectified; and that those who have lost their relative may at least have the satisfaction of knowing that lessons learned from his death may save the lives of others.”
It is well established in the case law of the Strasbourg Court that, in general, the Convention, as an international treaty, does not have retrospective effect. This is based on the general rule of international law embodied in article 28 of the Vienna Convention on the Law of Treaties (1969). That article provides:
“Unless a different intention appears from the treaty or is otherwise established, its provisions do not bind a party in relation to any act or fact which took place or any situation which ceased to exist before the date of the entry into force of the treaty with respect to that party.”
The Strasbourg Court has on a number of occasions had to grapple with whether and in what circumstances the article 2 procedural obligation can arise in relation to a death which occurred before the date on which the respondent State became bound by the Convention or, if later, before the date on which that State recognised the right of individual petition (“the critical date”). The evolution of that case law and its influence on the allied question of when there is a right to bring a claim before the domestic court under section 7 HRA is discussed in more detail below. To understand how the Dalton family’s claim has been addressed in these proceedings, we need at this stage only to give a brief outline of two strands of authority that have contributed to the current state of the law.
The first strand is the Strasbourg Court’s jurisprudence establishing that a State can be in breach of the article 2 procedural obligation to investigate a death even if the death occurred before the critical date. Despite the Convention not having retrospective effect, the procedural obligation can arise in relation to such a death where much of the investigation, the compliance of which falls to be assessed under article 2, either took place or ought to have taken place after the entry into force of the Convention. The principal authority as to this strand is the judgment of the Grand Chamber in Janowiec v Russia (2013) 58 EHRR 30 (“Janowiec”) which is discussed in detail below.
Clearly if the death occurs very shortly before the critical date, then it is likely that the period during which much of the investigation takes place or during which it ought to have taken place will fall after the critical date. It will therefore be within the temporal scope of the article 2 procedural obligation. If the death occurred some considerable time before the critical date, the second strand of case law becomes relevant. That concerns the revival of the article 2 procedural obligation in a case where, at some time after a death, new material comes to light. There can be such a revival where the death occurs after the critical date, even if the investigation immediately following the death complied fully with the article 2 procedural obligation or even if the time limit during which the family could have brought proceedings before the court in respect of that original investigation has long since expired.
The primary authority on when the article 2 procedural obligation revives is Brecknell v United Kingdom (2007) 46 EHRR 42 (“Brecknell”). In that case the death occurred in County Armagh on 19 December 1975. The investigation at the time, which was regarded as compliant with the UK’s procedural obligation under article 2 operating on the international plane, came to nothing. In 1999 allegations came to light about possible collusion by the security forces with loyalist paramilitaries including in relation to the bombing which led to Mr Brecknell’s death. Before the Strasbourg Court there was no issue of temporal jurisdiction since the death had occurred after the critical date of 14 January 1966. The issue, rather, was whether, given that there had been a timely article 2 compliant investigation in 1975, the UK could be in breach of a revived obligation because of the new material that had since emerged. The Strasbourg Court held that an obligation to take further investigative measures could arise where there is a plausible or credible allegation, piece of evidence or item of information. The precise circumstances in which a revival of the procedural obligation can occur is one of the issues raised in this appeal, in particular whether it is limited to circumstances in which the new information or evidence is relevant to the identification and eventual prosecution or punishment of the perpetrator of an unlawful killing or whether Brecknell applies more broadly than that.
These two strands of case law come together when addressing a situation where material that might revive the article 2 procedural obligation comes to light in respect of a death that occurred and was investigated many years before the critical date and so at a time when the State was under no procedural obligation in its conduct of that original investigation. This was the situation in Janowiec. The Strasbourg Court held in such a case it will only have jurisdiction over such a case ratione temporis if there is a “genuine connection” between the death and the critical date for that State. That connection does not in any sense require some causal link between the death and the adoption of the Convention, but is based on one of two alternative factors. The connection is either a temporal one in that the death must have occurred not more than a “reasonably short time” before the critical date or the case must allege a fundamental breach of “Convention values”, that is to say, a genuine connection may be established, in exceptional circumstances, in order to ensure that the underlying values of the Convention are protected in a real and effective way. The Strasbourg Court in Janowiec went on to confirm that even where there is a genuine connection between the death as the triggering event and the entry into force of the Convention, the procedural obligation under article 2 applies in relation to the procedural acts and omissions which took place or ought to have taken place after the critical date. The kinds of procedural acts covered by article 2 are acts undertaken in the framework of criminal, civil, administrative or disciplinary proceedings which are capable of leading to the identification and punishment of those responsible or to an award of compensation to the injured party. The Court contrasted these kinds of procedural acts with “other types of inquiries that may be carried out for other purposes, such as establishing a historical truth”: para 143.
Turning to domestic law, the HRA does not, in general, apply retrospectively. Section 7(1)(a) of the HRA provides that a person who claims that a public authority has acted or proposes to act in a way which is made unlawful by section 6(1) may bring proceedings against the authority under that Act in the appropriate court. Section 22(4) provides that section 7(1) “does not apply to an act taking place before the coming into force of that section”.
The House of Lords held in Wilson v First County Trust Ltd (No 2) [2003] UKHL 40; [2004] 1 AC 816 that the HRA did not have retrospective effect. But as regards the operation of the procedural obligation under article 2 and deaths occurring before 2 October 2000, this court has also held that the HRA should be construed as having the same degree of retrospective effect in the domestic sphere as the Convention has in the international sphere: In re McCaughey [2011] UKSC 20; [2012] 1 AC 725 (“McCaughey”). The effect of this approach to the interpretation of the HRA has been examined in subsequent authorities in this court, culminating in McQuillan.
The Dalton judicial review proceedings
Mr Dalton’s family have formed the view that a fresh inquest into the death of Mr Dalton is needed to take account of the material considered in the PONI Report which had not been available at the time of the original inquest. Their main concern is that the police knew or should have known about the bomb in or near 38 Kildrum Gardens and that they did not take steps to minimise the threat to the lives and safety of residents in the area. They hope that important witnesses who declined to cooperate with the PONI investigation would be compellable witnesses at a fresh inquest.
Section 14 of the Coroners Act (Northern Ireland) 1959 provides:
“Where the Attorney General has reason to believe that a deceased person has died in circumstances which in his opinion make the holding of an inquest advisable he may direct any coroner (whether or not he is the coroner for the district in which the death has occurred) to conduct an inquest into the death of that person, and that coroner shall proceed to conduct an inquest in accordance with the provisions of this Act (and as if, not being the coroner for the district in which the death occurred, he were such coroner) whether or not he or any other coroner has viewed the body, made any inquiry or investigation, held any inquest into or done any other act in connection with the death.”
On 25 July 2013, the Dalton family’s solicitor wrote to the AGNI enclosing the PONI Report which they said found serious failings with the original RUC investigation. They requested a fresh inquest into Mr Dalton’s death. The AGNI responded on 28 August 2013, asking for more information including as to whether the original inquest papers were available. There was some delay in the family being able to obtain the inquest papers from the Public Record Office of Northern Ireland. But on 14 February 2014 the original inquest papers were sent by the family to the AGNI with a request to reconsider their application. Under cover of a letter dated 30 May 2014, the family set out their detailed submissions as to why a new inquest was “advisable” for the purposes of section 14. They referred in their letter to the decision of this court in McCaughey, arguing that a fresh inquest would take a “significantly wider approach” to its tasks because the question for the coroner or the jury would now encompass the circumstances in which Mr Dalton came by his death. They said:
“The application of the [procedural] aspect of Article 2 to the proposed fresh inquest would have a marked effect on a number of aspects of it, not least upon the duties of disclosure and the scope of the inquest. The original inquest resulted in an open verdict. It is also worth remembering that the fresh inquest would be permitted to deliver the narrative form of the verdict, thereby permitting greater latitude and flexibility.
We take the view that a fresh inquest into this death would now be able to take into account a great deal of pertinent and relevant material, none of which was available at the time of the original inquest. The fact that a fresh inquest would be governed by the procedural aspect of Article 2 is therefore highly significant. It would allow and require a broader scope to be applied, potentially taking into account the relevant circumstances before and after the murder (though the precise scope of the fresh inquest would be a matter for the Coroner’s discretion).”
The submissions sent with the letter cited passages from Brecknell and argued that the holding of a fresh inquest was required by article 2 and in any event was advisable.
The AGNI replied on 2 October 2014 declining to direct a fresh inquest. The AGNI referred to the Strasbourg Court’s decision in Janowiec concerning the application of article 2 to deaths occurring prior to the critical date, which was in this case 2 October 2000. The AGNI said that article 2 does not require proceedings to be held which have as their purpose the establishment of historical truth (drawing there on the exclusion from article 2 referred to in Janowiec, para 143). It appears that the AGNI interpreted Janowiec as limiting any investigation required in respect of a death occurring before the critical date to an investigation capable of leading to the identification and punishment of those responsible for the death in question (“the perpetrators”), particularly in cases of deliberate killing. The AGNI wrote:
“The Attorney has not been provided with any evidence which would suggest that the identification and/or punishment of those responsible could be achieved if a fresh inquest was to be ordered in this application and he is, therefore, of the view that the Article 2 objectives as identified in Janowiec are not likely to be achieved by way of a fresh inquest into Mr Dalton's death. The Attorney further notes that the Police Ombudsman - in a report critical of police - did not refer the matter to the [Public Prosecution Service].
Although the Attorney considers that Article 2 does not require an inquest in this case, he has further considered whether, quite apart from any Article 2 consideration, an inquest could be said to be advisable. Having regard to the investigation by the Police Ombudsman and to the existence of current civil proceedings he does not consider an inquest to be advisable, even if the focus were to be purely on domestic factors.”
These proceedings were launched on 26 June 2015 by Dorothy Johnstone, one of Mr Dalton’s daughters. They are continued, following Mrs Johnstone’s death, by her sister Rosaleen Dalton. It was initially alleged in the proceedings that the refusal to direct a fresh inquest was unlawful in that it was irrational and in breach of article 2. Permission to apply for judicial review was granted by Maguire J on 17 June 2016 on the article 2 ground but not on the irrationality ground. The substantive application for judicial review was heard on 10 and 11 January 2017 and was dismissed by Deeny J on 28 March 2017 ([2017] NIQB 33).
Deeny J noted that the focus of the Strasbourg Court’s observations in Brecknell was on the prosecution and conviction of perpetrators. He stated “It is obviously in the public interest that those guilty of murder be brought to justice even after a long delay”: para 31. Deeny J referred to the considerable number of new inquests that had been ordered in Northern Ireland and noted that they had not been matched by an increase in resources. He said (para 38):
“That leads on to a relevant consideration about this case. If the death had occurred recently an inquest might disclose a systemic flaw in practice which contributed to the deaths. But if, as the Ombudsman found, there was a failure here, it took place 28 years ago at the height of the Troubles. It is very difficult to see how any practical benefit could now be obtained for the public in going over the procedures then being followed by police officers in Derry at that time, when they say that much of the city was out of bounds to them by terrorist activity.”
Deeny J referred briefly to the issue that is the principal issue in this appeal, that is the application of Janowiec and the Strasbourg and domestic case-law concerning article 2 and deaths occurring prior to the adoption of the Convention by the State: para 43. He said that the AGNI had not relied on those cases in his submissions and he did not address the issue. At the end of his judgment, Deeny J returned to the point that the inquest was unlikely to lead to the prosecution of the terrorists responsible for the deaths. He said:
… While there may be cases where an Article 2 obligation is revived even though it is unlikely to advance the goal of successful prosecution of the immediate perpetrators of the unlawful death, such cases are not likely to be common and are not likely to warrant a renewed inquest. If they involve the alleged misconduct on the part of the police then the Police Ombudsman would be, as he was here, appropriate to address the issues.
I am very pessimistic that an inquest held at this time would succeed in securing any significant accession of information compared to that which the Ombudsman obtained. That is always possible, it must be accepted, but it was within the discretion of the Attorney to conclude that it was not likely in this case.
While the existence of civil proceedings brought by the family will not, indeed, necessarily cure a duty of compliance with Article 2, nevertheless it is relevant as giving a further opportunity to seek documents and call witnesses. It is true that the power to subpoena witnesses in a civil action will not lead to a right of the plaintiff to cross-examine the witness under subpoena. In an inquest situation such a witness might be compelled. But the chances of a witness, even a police officer guilty of an error of judgment or negligence or conceivably something worse, let alone a perpetrator, making a concession because he is being cross-examined rather than examined in chief is a very slight one which the State is entitled to conclude does not justify the financial and human cost of a further inquest.”
He therefore held that the AGNI’s decision had been a lawful one.
Mr Dalton’s family appealed Deeny J’s decision to the Court of Appeal. The hearing of that appeal took place on 8 October 2018. After that hearing but before judgment was delivered, this court handed down judgment in Finucane, which we discuss in paras 137-138 below. The Court of Appeal received written submissions from the parties to address the effect of that judgment.
The Court of Appeal (Morgan LCJ, Stephens LJ and Maguire J) handed down its judgment in the Dalton appeal on 4 May 2020 and allowed the appeal: [2020] NICA 26; [2021] NI 405. Submissions were then made by the parties as to the form of the final order. The Court of Appeal’s final order was issued on 27 May 2020.
The Court of Appeal dealt in much more detail than Deeny J with the Brecknell case. The submissions of the AGNI were that the invocation of Brecknell was inappropriate because that case had been concerned with a particular situation where new information had emerged which spoke to the possible identification and punishment of perpetrators of serious crime. Here there was no credible suggestion that the police had been involved in planting the bomb. Further the AGNI argued that even if Brecknell did apply, it did not follow that he was under an obligation to order a fresh inquest. The question of what sort of investigation was needed to satisfy the obligation was dependent on a judgment to be made by him: see para 70.
The Court of Appeal identified the issue as how the article 2 procedural obligation to carry out an effective official investigation operates in the context of cases where the death at issue preceded the introduction of the HRA in October 2000. As to the first point about whether the kind of information which can revive the article 2 procedural obligation is limited to information which might lead to the identification and punishment of the perpetrators, the Court of Appeal held that it was not so limited: paras 101-110. The doctrine of revival could therefore apply in Mr Dalton’s case notwithstanding that it would not be the object of the fresh investigation to identify and punish the direct perpetrators.
They went on to conclude that there had been a revival of the obligation in this case, prompted by the initiation of the complaints of Mr Dalton’s son in 2005 and the material disclosed in support of that complaint: para 113.
The Court of Appeal next considered how Janowiec applied, given that Mr Dalton’s death occurred 12 years and one month before the entry into force of the HRA. Was there still a “genuine connection” as that concept had been explained in Janowiec? The Court of Appeal held that there was, because the 10 year limit set out in Janowiec “should not be viewed inflexibly” (para 115).
The Court held that they inclined to the view that the genuine connection test was satisfied.
The Court of Appeal then concluded (paras 118-124) that the PONI Report did not meet the UK’s obligations under article 2 because it was unclear from the report how the key conclusions had been arrived at. The PONI had felt obliged to say that his work had been “significantly hampered” by the refusal of some police officers to cooperate. However, the Court limited the relief granted to a declaration that there had not been an article 2 compliant investigation into the death of Mr Dalton. They recognised that section 14 of the Coroners (Northern Ireland) Act 1959 made it clear that the question of whether or not to direct an inquest was “very much one involving the personal judgment and assessment of the AGNI” (para 126). Further, they said that it has long been the case that the form of the investigation required by article 2 may vary according to the circumstances: para 138. The matter was not so open and shut in favour of the order of mandamus sought by the applicant: para 141.
Permission to appeal to this court was granted by this court on 15 December 2021.
In parallel with these judicial review proceedings, there is a civil action in progress between the estate of Mr Dalton and the Chief Constable of the PSNI and the Ministry of Defence. The pleadings in that action closed in late 2014 and the disclosure process is still underway.
The detachable procedural obligation under article 2
We observed in para 56 above that a central issue in this appeal is how firm is the time limit on the application of the HRA in relation to an article 2 duty to investigate a death which occurred before the HRA came into force. In order fully to understand this issue it is necessary to outline in some detail how the article 2 procedural obligation has been developed in the case-law of both the Strasbourg Court and the domestic courts.
The initial UK jurisprudence
In this part of the judgment we outline how the judiciary in the United Kingdom have responded to the development of the jurisprudence of the Strasbourg Court in relation to article 2 after that court first recognised the existence of the procedural duty in McCann. Since the HRA came into force on 2 October 2000 and gave domestic effect to certain Convention rights, judges in the United Kingdom have had to work out whether, and if so how, Parliament in enacting the HRA intended that Act to operate in relation to that duty. In so doing, the judges have sought to draw on the jurisprudence of the Strasbourg Court in relation to the non-retrospective nature of the Convention by way of analogy when addressing claims pursued in domestic law which relate to events occurring before the HRA came into force.
The initial analysis of the House of Lords was set out in the test case of In re McKerr [2004] UKHL 12; [2004] 1 WLR 807. That appeal concerned the death of Mr McKerr’s father, one of three men who were shot by members of the Royal Ulster Constabulary in November 1982, almost 18 years before the HRA was brought into force but 16 years after the United Kingdom had recognised a right of individual petition. In 2001 the Strasbourg Court issued its judgment in McKerr v United Kingdom (2001) 34 EHRR 20, in which it found that the failure of the United Kingdom to hold an effective investigation into the death by use of force constituted a violation of article 2 and awarded Mr McKerr £10,000 as just satisfaction. In 2002 Mr McKerr commenced judicial review proceedings against the Secretary of State for Northern Ireland in which he sought (i) declarations that the Secretary of State’s failure to provide an article 2-compliant investigation was unlawful and in breach of section 6 of the HRA and article 2 of the Convention and (ii) a mandatory order compelling him to conduct such an investigation. The question which faced the court was whether the failure to hold a further investigation was in breach of section 6 of the HRA which made it unlawful in domestic law for a public authority, such as the Secretary of State, to “act in a way which was incompatible with a Convention right”. The UK Government’s response was that section 6 of the HRA was not applicable to deaths occurring before that Act came into force.
The House of Lords rejected Mr McKerr’s claim on the basis that the Convention rights had not been part of the domestic law of the United Kingdom until the HRA came into force. The Government’s failure to hold a further investigation in Mr McKerr’s case was not conduct prohibited by section 6 of that Act. Lord Nicholls of Birkenhead stated (para 16) that it was settled as a general proposition that the HRA was not retrospective and that the HRA treated as an exception the retrospective reliance on Convention rights in proceedings brought by a public authority provided for in section 22(4). He held that an unlawful killing which occurred before 2 October 2000 would not be a breach of section 6(1). The investigative obligation is consequential upon the death and Parliament cannot have intended that the HRA would apply differently to the primary obligation to protect life and the consequential obligation to investigate a death. The obligations arising under article 2 were parts of a single whole and Parliament did not intend to create a right to an investigation in respect of a death which occurred before the Act came into force (paras 21-22, 32). Lord Steyn reached a similar conclusion (para 48) as did Lord Hoffmann, who expressed concern that otherwise “there can in principle be no limit to the time one could have to go back into history and carry out investigations” (para 67). Lord Rodger of Earlsferry recognised that the right to an investigation was free-standing to the extent that deaths had to be investigated even though it might turn out that the killing was lawful and not in breach of article 2 (para 78), but in agreement with Lord Nicholls and Lord Brown of Eaton-under-Heywood, he held that Parliament had not made a transitional provision and that the right to an investigation under the HRA was confined to deaths which, having occurred after the commencement of the Act, may be found to be unlawful under the Act (para 81). Lord Brown agreed, holding (para 89) that the duty to investigate is “necessarily linked to the death itself and cannot arise under domestic law save in respect of a death occurring at a time when article 2 rights were enforceable under domestic law, ie on and after 2 October 2000.”
The Šilih case
This approach did not survive the judgment of the Grand Chamber of the Strasbourg Court in Šilih v Slovenia (2009) 49 EHRR 37 (“Šilih”). The case concerned an application by the parents of a young man, whose death was the result of allegedly negligent medical treatment in 1993 about 14 months before Slovenia brought the Convention into force and recognised the jurisdiction of the Strasbourg Court in June 1994. The latter date was the “critical date” for the establishment of the jurisdiction of the Strasbourg Court. The applicants’ complaint was that the state had failed to carry out an effective and prompt investigation of their son’s death contrary to article 2. In the judgment of the majority of the Grand Chamber, which upheld the complaint, it was held that the procedural obligation to carry out an effective investigation under article 2 had evolved into a separate and autonomous duty and the duty “can be considered to be a detachable obligation arising out of article 2 capable of binding the state even when the death took place before the critical date” (para 159).
The majority sought to place some limits on the Strasbourg Court’s temporal jurisdiction in the interests of legal certainty and stated:
“161 … having regard to the principle of legal certainty, the Court’s temporal jurisdiction as regards compliance with the procedural obligation of article 2 in respect of deaths that occur before the critical date is not open-ended.
First, it is clear that, where the death occurred before the critical date, only procedural acts and/or omissions occurring after that date can fall within the Court’s temporal jurisdiction.
Secondly, there must exist a genuine connection between the death and the entry into force of the Convention in respect of the respondent state for the procedural obligations imposed by article 2 to come into effect.
Thus a significant proportion of the procedural steps required by this provision – which include not only an effective investigation into the death of the person concerned but also the institution of appropriate proceedings for the purpose of determining the cause of the death and holding those responsible to account – will have been or ought to have been carried out after the critical date.
However, the Court would not exclude that in certain circumstances the connection could also be based on the need to ensure that the guarantees and the underlying values of the Convention are protected in a real and effective manner.”
It was and is clear that this attempt to impose a limit on the retrospective effect of the procedural duty by reference to a genuine connection and Convention values tests was problematic. Judge Lorenzen in his concurring opinion expressed strong concern that the tests set out in para 163 lacked legal certainty with the result that it was difficult to identify the limits of the Court’s temporal jurisdiction (para O-I 3). Judge Zagrebelsky joined by Judges Rozakis, Cabral Barreto, Spielmann and Sajó in a concurring opinion spoke of the “vague wording” of para 163 and stated that the court would be “forced to carry out complex and questionable assessments on a case-by-case basis that will be difficult to dissociate from the merits of the case” (para O-III 4). Similarly, in a strongly worded joint dissenting opinion Judges Bratza and Türmen expressed concern about (i) the genuine connection test, asking whether it was meant to refer to a close temporal link between the death and the entry into force of the Convention, (ii) how omissions were to be dealt with in the “significant proportion of procedural steps” test, and (iii) the lack of definition in the “Convention values” test (para O-IV 17).
The McCaughey case
Similar concerns were expressed when the Supreme Court came to reconsider the question of the procedural obligation in the context of the temporal jurisdiction of the UK courts under the HRA in McCaughey. That case concerned the killing of the claimants’ relatives by the British Army in October 1990, approximately 10 years before the HRA came into force. The claimants in their application for judicial review sought a direction that the coroner, who was conducting an inquest into the deaths in 2009, was obliged to conduct an inquest in a way which satisfied the state’s procedural obligation under article 2 of the Convention. The courts below the Supreme Court were bound by the decision of the House of Lords in McKerr and the Supreme Court convened a court of seven Justices to reconsider that decision in the light of the judgment of the Grand Chamber in Šilih. By a majority of six to one (Lord Rodger dissenting) the court held that the Strasbourg Court in Šilih had extended the effect of the procedural duty under article 2 of the Convention with the result that in domestic UK law such an obligation could arise in relation to the investigation of a death which occurred before the HRA came into force if a significant proportion of the procedural steps which article 2 would require took place after 2 October 2000. The court therefore made the declaration which the claimants sought.
The President of the court, Lord Phillips of Worth Matravers, analysed the Grand Chamber’s judgment in Šilih and recognised that the United Kingdom was under an international obligation under the Convention to ensure that, if it carries out an inquest into an historic death, that inquest complies with the procedural obligations of article 2. The issue was whether the HRA on its true interpretation applied to that obligation so as to impose an obligation in domestic law. In addressing that issue, he considered two competing principles, “the non-retroactive principle”, which did not permit a claimant to bring a claim for breach of a Convention obligation that occurred before the HRA came into force, and “the mirror principle”, by which the ambit of the HRA should mirror that of the Convention as it was the object of the Act to “bring human rights home” so that claimants could pursue claims within the UK jurisdictions which they would otherwise be permitted to bring before the Strasbourg Court (paras 56-59). Lord Phillips acknowledged that Šilih had removed the spectre that the article 2 procedural obligation would continue indefinitely by requiring that the free-standing obligation would arise by reason of current events. He concluded that the mirror principle should prevail because Parliament was presumed to intend that there was a domestic requirement to comply with the article 2 procedural obligation that currently arose (paras 61-62).
Lord Hope of Craighead considered that the central question was whether section 22(4) of the HRA permitted the Šilih approach to be adopted in domestic law (para 70). He concluded that while there was no domestic law obligation to carry out an article 2-compliant investigation before 2 October 2000, the state had decided to hold an inquest into the deaths with the result that the invocation of the free-standing procedural obligation, which was detached from the substantive obligation under article 2, did not involve a retrospective application of section 6 of the HRA (paras 75-80). Lady Hale (paras 90-93) and Lord Brown (paras 100-101) adopted an essentially similar approach. Lord Kerr of Tonaghmore sought to derive principles from the judgments in Šilih and concluded that because much of the investigation of the deaths occurred after 2 October 2000 there was a genuine connection between the deaths and the commencement of the HRA (para 119). Lord Dyson agreed with Lord Phillips that the mirror principle was relevant to the application of the detachable procedural obligation (para 134). He concluded by expressing the hope that the Strasbourg Court would have an opportunity to clarify the meaning of para 163 of the Šilih judgment which had been the subject of trenchant criticism. Lord Rodger dissented essentially on the basis that Parliament in enacting the HRA would not have wished to bring past deaths within the scope of the Act. The HRA did not have transitional provisions that would accommodate the Šilih judgment and Parliament had not chosen to introduce such provisions (paras 159-162).
The clarification in Janowiec
The opportunity for the Grand Chamber to review and clarify the meaning of the Šilih judgment arose in the case of Janowiec which concerned applications for an article 2 investigation into the infamous killing of many thousands of Polish soldiers, border guards, police officers, prison guards, state officials and other functionaries on the orders of the Politburo of the Central Committee of the Soviet Communist Party in April and May 1940. The Strasbourg Court by 13 votes to four held the court was not competent to examine the complaint as it did not have temporal jurisdiction because (i) the events, which may have met the “Convention values” test, occurred before the adoption of the Convention in 1950 and (ii) the events did not have a genuine connection with the critical date, being the date on which the Russian Federation adopted the Convention in 1998, as they occurred 58 years earlier and the significant procedural steps of the investigation took place before the critical date.
The majority judgment described the decision in Šilih and recorded that recent case law following that judgment had often applied the procedural obligation to deaths which were not the consequence of actions by state agents occurring between one and four years before the critical date. Reference was made to one case when a person died in police custody approximately seven years before Turkey recognised the right of individual petition and to cases where deaths had been caused by insurgents or paramilitary organisations seven or six years before the critical date. The majority also referred to the case of Mladenović v Serbia (Application No 1099/08) (unreported) 22 May 2012 (“Mladenović”) in which a 13-year period separating the death of a person in a brawl and the entry into force of the Convention in respect of Serbia was not “seen as outweighing the importance of the procedural acts that were accomplished after the critical date” (para 138).
The majority then sought to clarify the criteria adopted in Šilih, recognising that they had given rise to uncertainty in their application. As the correct interpretation of Janowiec plays a central role in this appeal it is necessary to inform the discussion below that we set out the court’s reasoning in some detail. The Court first summarised the Šilih judgment in these terms before examining each of the elements of that summary:
“The criteria laid down in paras 162 and 163 of the Šilih judgment can be summarised in the following manner. First, where the death occurred before the critical date, the Court’s temporal jurisdiction will extend only to the procedural acts or omissions in the period subsequent to that date. Secondly, the procedural obligation will come into effect only if there was a ‘genuine connection’ between the death as the triggering event and the entry into force of the Convention. Thirdly, a connection which is not ‘genuine’ may nonetheless be sufficient to establish the Court’s jurisdiction if it is needed to ensure that the guarantees and the underlying values of the Convention are protected in a real and effective way.” (para 141)
Taking the elements of the summary in turn, the Court first addressed the procedural acts and omissions in the period after the critical date. It stated (para 142):
“The Court reiterates at the outset that the procedural obligation to investigate under article 2 is not a procedure of redress in respect of an alleged violation of the right to life that may have occurred before the critical date. The alleged violation of the procedural obligation consists in the lack of an effective investigation; the procedural obligation has its own distinct scope of application and operates independently from the substantive limb of article 2. Accordingly, the Court’s temporal jurisdiction extends to those procedural acts and omissions which took place or ought to have taken place in the period after the entry into force of the Convention in respect of the respondent Government.” (Emphasis added)
The Court then explained the meaning of “procedural acts” in the context of articles 2 or 3 of the Convention (para 143) as:
“acts undertaken in the framework of criminal, civil, administrative or disciplinary proceedings which are capable of leading to the identification and punishment of those responsible or to an award of compensation to the injured party. This definition operates to the exclusion of other types of inquiries that may be carried out for other purposes, such as establishing a historical truth.”
The Court then addressed the concept of “omissions” in its summary in para 141 of its judgment (para 144):
“The mention of ‘omissions’ refers to a situation where no investigation, or only insignificant procedural steps, have been carried out, but where it is alleged that an effective investigation ought to have taken place. Such an obligation on the part of the authorities to take investigative measures may be triggered when a plausible, credible allegation, piece of evidence or item of information comes to light which is relevant to the identification and eventual prosecution or punishment of those responsible.”
At this point the Court referred in a footnote to among others the case of Brecknell, which we discuss in paras 180 – 191 below. The Court continued:
“Should new material emerge in the post-entry into force period and should it be sufficiently weighty and compelling to warrant a new round of proceedings, the Court will have to satisfy itself that the respondent State has discharged its procedural obligation under article 2 in a manner compatible with the principles enunciated in its case-law. However, if the triggering event lies outside the Court’s jurisdiction ratione temporis, the discovery of new material after the critical date may give rise to a fresh obligation to investigate only if either the ‘genuine connection’ test or the ‘Convention values’ test, discussed below, has been met.”
In the following four paragraphs (paras 145-148) the Strasbourg Court set out its analysis of the “genuine connection” test which lies at the heart of this appeal. We therefore set out those paragraphs in full:
The first sentence of para 163 of the Šilih judgment posits that the existence of a ‘genuine connection’ between the triggering event and the entry into force of the Convention in respect of the respondent State is a condition sine qua non for the procedural obligation under article 2 of the Convention to come into effect.
The Court considers that the time factor is the first and most crucial indicator of the ‘genuine’ nature of the connection. It notes … that the lapse of time between the triggering event and the critical date must remain reasonably short if it is to comply with the ‘genuine connection’ standard. Although there are no apparent legal criteria by which the absolute limit on the duration of that period may be defined, it should not exceed ten years. Even if, in exceptional circumstances, it may be justified to extend the time-limit further into the past, it should be done on condition that the requirements of the ‘Convention values’ test have been met. (Emphasis added)
The duration of the time period between the triggering event and the critical date is however not decisive, in itself, for determining whether the connection was a ‘genuine’ one. As the second sentence of para 163 of the Šilih judgment indicates, the connection will be established if much of the investigation into the death took place or ought to have taken place in the period following the entry into force of the Convention. This includes the conduct of proceedings for determining the cause of the death and holding those responsible to account, as well as the undertaking of a significant proportion of the procedural steps that were decisive for the course of the investigation. This is a corollary of the principle that the Court’s jurisdiction extends only to the procedural acts and omissions occurring after the entry into force. If, however, a major part of the proceedings or the most important procedural steps took place before the entry into force, this may irretrievably undermine the Court’s ability to make a global assessment of the effectiveness of the investigation from the standpoint of the procedural requirements of article 2 of the Convention.
Having regard to the above, the Court finds that, for a ‘genuine connection’ to be established, both criteria must be satisfied: the period of time between the death as the triggering event and the entry into force of the Convention must have been reasonably short, and a major part of the investigation must have been carried out, or ought to have been carried out, after the entry into force.” (Emphasis added)
The Strasbourg Court then considered the “Convention values” test which, in para 146, it saw as the basis for extending the time period in exceptional circumstances. It stated:
The Court further accepts that there may be extraordinary situations which do not satisfy the ‘genuine connection’ standard as outline above, but where the need to ensure the real and effective protection of the guarantees and the underlying values of the Convention would constitute a sufficient basis for recognising the existence of a connection. The last sentence of para 163 of the Šilih judgment does not exclude such an eventuality, which would operate as an exception to the general rule of the ‘genuine connection’ test. In all the cases outlined above the Court accepted the existence of a ‘genuine connection’ as the lapse of time between the death and the critical date was reasonably short and a considerable part of the proceedings had taken place after the critical date. Against this background, the present case is the first one which may arguably fall into this other, exceptional, category. Accordingly, the Court must clarify the criteria for the application of the ‘Convention values’ test. (Emphasis added)
Like the Chamber, the Grand Chamber considers the reference to the underlying values of the Convention to mean that the required connection may be found to exist if the triggering event was of a larger dimension than an ordinary criminal offence and amounted to the negation of the very foundations of the Convention. This would be the case with serious crimes under international law, such as war crimes, genocide or crimes against humanity, in accordance with the definitions given to them in the relevant international instruments.”
The Grand Chamber returned to the duty to undertake an effective investigation of a death occurring before the critical date in Mocanu v Romania (2014) 60 EHRR 19 (“Mocanu”), which concerned the death of one man and the ill treatment of another at the hands of the state’s security forces in June 1990, four years before the Convention entered into effect in respect of Romania. In that case the Grand Chamber summarised the ruling on temporal component of the “genuine connection” test in para 146 of Janowiec in these terms (para 206):
“It added that such a connection was primarily defined by the temporal proximity between the triggering event and the critical date, which could be separated only by a reasonably short lapse of time that should not normally exceed ten years.”
In the same paragraph the Grand Chamber went on to explain that the time period was not itself decisive as the “genuine connection” test had also to meet the requirement in para 147 of Janowiec that much of the investigation had taken place or ought to have taken place after the critical date. Counsel for Ms Dalton founds on the Grand Chamber’s use of the word “normally”, a submission to which we return in para 158 below.
UK jurisprudence since the Janowiec judgment
The first case which came to the Supreme Court after the Janowiec judgment concerned the killing of 24 unarmed civilians by members of a patrol of the Scots Guards in the village of Batang Kali in Selangor in December 1948 at a time when Selangor was a British protected State in the Federation of Malaya: R (Keyu) v Secretary of State for Foreign and Commonwealth Affairs and another [2015] UKSC 69; [2016] AC 1355 (“Keyu”). Close relatives of the victims sought to challenge the refusal of the UK Government to hold a public inquiry into the circumstances of the killings. The court unanimously dismissed the applicants’ appeal so far as it rested on article 2 of the Convention. Lord Neuberger, in a majority judgment with which Lord Hughes agreed and, so far as relevant to this appeal, Lord Mance agreed, expressed the view (para 89) that the Strasbourg Court would rule the application inadmissible because the killings occurred more than ten years before UK citizens had the right to petition the Strasbourg Court. Unlike Lady Hale and Lord Kerr, he considered the Strasbourg jurisprudence to be clear that the critical date was the date when the right of individual petition was given and not the date on which the Convention came into force on the international plane in respect of the relevant State. He stated (para 90):
“[T]his is a topic on which clarity and consistency is highly desirable, and, unless the guidance from Strasbourg seemed unclear, incoherent or unworkable, I would be reluctant not to follow and apply it. Having permitted a degree of retroactivity, I believe that the Strasbourg court has rightly imposed some pretty clear rules with a view to ensuring a degree of clarity and consistency in this area. Particularly in the absence of any invitation to do so, I consider that, at least in this case, this is an area on which we should follow, but go no further than Strasbourg jurisprudence.”
In the following paragraphs he left open the question whether, if the Strasbourg Court were to hold that the appellants were entitled after 2 October 2000 to seek an investigation into the killings, the UK courts would be bound to order an inquiry pursuant to the HRA.
Lord Kerr’s analysis of the Strasbourg case-law emphasised the detachable nature of the procedural obligation under article 2, stating (para 208) that there was “no inescapable point of principle … which requires the adoption of a ten-year period as the absolute limit on the period between the death and the critical date”. He considered that practicability of inquiry “must play a part in the evaluation.” Similarly, at para 238, in summarising Janowiec (para 146) he emphasised that the period between the triggering event (ie the deaths) and the critical date must be reasonably short, and while there was no absolute limit, it should not exceed 10 years. When considering the question whether the right under the HRA should be the same as the temporal jurisdiction of the Strasbourg Court he suggested (para 255) that a limit must be set which was “essentially arbitrary but which accords with what is, in most cases, practically possible.” He did not need to decide the limit in that appeal because it could not extend back 52 years from the coming into force of the HRA.
Thereafter in Chong v United Kingdom (Application No 29753/16) (2018) 68 EHRR SE2, the Strasbourg Court considered an application by the relatives of those killed at Batang Kali in 1948 and ruled it to be inadmissible. The Strasbourg Court clarified that the critical date so far as its own temporal jurisdiction was concerned is the date on which the right of individual petition was recognised rather than the date on which the Convention came into force: see para 84.
The Supreme Court again considered the jurisprudence of the Strasbourg Court in relation to the time limit in 2019 in Finucane. This case, as is well-known, concerned the shocking allegations that members of the United Kingdom security forces had colluded in the murder by “loyalist” paramilitaries of a solicitor in his home in the presence of his wife and children in February 1989, 11 years and eight months before the HRA came into force. Mr Finucane’s widow applied for judicial review of the decision by the UK Government not to hold a public inquiry into his murder. The application proceeded on two grounds. First, Mrs Finucane argued that she had a legitimate expectation that a public inquiry would be held based on an earlier undertaking. This ground failed and it is not relevant to this appeal. Secondly, she argued that her rights under article 2 of the Convention had been breached by the failure to establish such an inquiry. In her appeal to the Supreme Court she succeeded on this second ground and the court unanimously upheld the decision of the first instance judge that a declaration be made that there had not been an inquiry which was compatible with article 2 of the Convention. This left it to the state to decide what form of investigation, if any were feasible, was required to meet the article 2 requirement.
Lord Kerr wrote the leading judgment and Lady Hale, Lord Carnwath, Lord Hodge and Lady Black agreed with his reasoning and conclusions. In setting out the court’s view on this question Lord Kerr (paras 85-100) considered, among others, the cases of McKerr, Brecknell, McCaughey and Šilih. At para 101 he turned to the Strasbourg Court’s decision in Janowiec. The essence of this court’s decision on the “genuine connection” test is set out in paras 107-109 of the judgment:
I consider that a genuine connection has been established between the triggering event and the critical date in this case. As Stephens J pointed out in para 34 of his judgment, ECtHR in Mocanu … para 206 referred to ‘a reasonably short lapse of time that should not normally exceed ten years’ (emphasis added). And in Mladenović v Serbia … the court considered it could examine the procedural aspect of article 2 (and found a violation) in relation to a death that had occurred in 1991 when Serbia’s ratification of the Convention took place some 13 years later in 2004.
A period of ten years or less between the triggering event (the murder of Mr Finucane) and the critical date (the coming into force of the HRA) is not an immutable requirement. The time which elapsed between the two dates is a factor of importance but, when taken into account with the circumstance that the vast bulk of noteworthy inquiry into his death has taken place since the HRA came into force (Stevens III, the Cory Inquiry and the de Silva review), the significance of the time lapse diminishes. Nothing in Janowiec detracts from the proposition in Šilih that the decision as to whether there is a genuine connection involves a multi-factorial exercise and the weight to be attached to each factor will vary according to the circumstances of the case.
Moreover, in McCaughey it was made clear that an inflexible ten-year limit was not essential and the consideration that most of the investigation took place after the critical date could compensate for the length of the time lapse – see paras 118, 119 and, in particular, 139 where Lord Dyson said:
‘The deaths were ten years before the HRA came into force. That is a relevant factor to be taken into account when considering whether there is a sufficient connection between the deaths and the coming into force of the Act. But Šilih v Slovenia … shows that it is not the only factor. In particular, of considerable importance is the fact that at that date the investigation had been initiated, but a significant proportion of the procedural steps required to be taken had not yet been taken. In that respect, the facts of the case are similar to the facts in Šilih v Slovenia. This is the feature of Šilih v Slovenia which is emphasised by the majority at para 165 and by Judge Lorenzen at para O-I4 of the EHRR report.’
Significantly, we were not invited to depart from the decision in McCaughey.”
This court’s judgment in Finucane understandably had a considerable influence on the reasoning of the Court of Appeal in Northern Ireland in the next two relevant cases.
In In re McQuillan [2019] NICA 13; [2020] NI 583 the Court of Appeal considered the procedural obligation under article 2 in the light of the Strasbourg Court’s judgment in Brecknell, where military logs had come to light many years after the death of Ms Smyth in June 1972 and constituted plausible and credible pieces of evidence which might point to possible military involvement in the death. In addressing the genuine connection test in this context, the Court of Appeal drew on this court’s judgment in Finucane and treated the test as a multifactorial exercise in which the lapse of time was given very little weight where new evidence which met the Brecknell criteria had come to light (paras 131- 138). The Court held that the genuine connection test was met in this case and generally would be met in any case that satisfied the Brecknell test. This court reversed that ruling in the appeal which we discuss in paras 143 – 145 below.
Similarly in its decision in the current case of Dalton ([2020] NICA 26; [2021] NI 405), the Court of Appeal again relied on Finucane and treated the genuine connection test as a multifactorial exercise in which the passage of time could be outweighed by the coming to light at a later date of a plausible or credible allegation, piece of evidence or item of information: see paras 84-88, 111-116. This analysis is inconsistent with the analysis of this court in the appeal in the McQuillan case to which we now turn.
The Supreme Court again addressed the “genuine connection” test in a seven-Justice appeal in McQuillan. As we explain in para 162 below, counsel appearing in that appeal did not challenge this court’s judgment in Finucane and the court therefore sought to accommodate that judgment while expressing concern as to the correctness of its analysis of the Strasbourg Court’s judgment in Janowiec.
The court (paras 147ff) also addressed and rejected an argument that for the purposes of the HRA the critical date was 14 January 1966, when the United Kingdom accepted a right of individual petition to the Strasbourg Court, rather than 2 October 2000, when the HRA came into force. It recognised that the question was one of statutory interpretation and that it involved a balancing of the non-retroactive principle and the mirror principle which the court discussed in McCaughey. The court analysed the judgments of the Strasbourg Court in Šilih and Janowiec and concluded that one could not separate the questions of genuine connection and critical date from the issue of the non-retrospectivity of the HRA. In para 161 it interpreted the jurisprudence of the Strasbourg Court as establishing that:
“the investigative obligation is detachable not in the sense of being completely free-standing, but in the more limited sense that it is capable of arising notwithstanding that the death which triggers the obligation preceded the Convention coming into effect, provided there is a sufficient connection between that event and the death.”
This court continued (para 162):
“Thus the point of the genuine connection test articulated in Šilih and in Janowiec is to identify when a death before the critical date is capable of triggering an investigative obligation under the article 2 right set out in the Convention which the Strasbourg Court will recognise as falling within its jurisdiction ratione temporis. … Although the article 2 investigative obligation is detachable, it remains the case that it has to be anchored in the circumstances of a death which has occurred in suspicious circumstances and which the court is prepared to treat as relevant. This means that there is inevitably an issue of retroactivity where the death happened before the relevant article 2 obligation was brought into operation.”
This court turned to the regime of domestic Convention rights under the HRA and stated (paras 163-164) that the genuine connection test had to be applied by analogy having regard to the fact that the substantive right under article 2 has no application in domestic law before 2 October 2000. It stated (para 164):
“Under the HRA, the substantive right under article 2 has no application before 2 October 2000, by contrast with the substantive right under article 2 of the Convention. … For the purposes of application of the genuine connection test, the ‘critical date’ is the date of inception of the relevant right under article 2. In the case of the Convention, so far as concerns the United Kingdom, it is 14 January 1966; and different dates are relevant for other contracting states, depending on the date of inception of the right under article 2 for them. In the case of the HRA, applying the same principle, the critical date is 2 October 2000.”
The detachable duty and the HRA after McQuillan
The position in the United Kingdom after the McQuillan judgment can be summarised briefly: (i) the triggering event for both the article 2 substantive duty and the article 2 procedural obligation is the date of the relevant death; (ii) the critical date for the purposes of the HRA is 2 October 2000, when the substantive duty under article 2 first arose in domestic law; and (iii) in relation to the time lapse between (i) and (ii) the court has sought to adopt by analogy the approach in the jurisprudence of the Strasbourg Court in relation to its temporal jurisdiction under the Convention and adapt it to the jurisdiction of the United Kingdom courts in relation to the HRA.
In this appeal, Mr Tony McGleenan KC for the AGNI has mounted a challenge to the position which this court reached on point (iii) above in Finucane and which it preserved in McQuillan, arguing that it is contrary to the guidance of the Grand Chamber of the Strasbourg Court in Janowiec. This court must therefore examine further the judgments in Janowiec to ascertain what that case decided.
What did the Grand Chamber decide in Janowiec about the genuine connection test?
Mr McGleenan’s submission, in short, is that the Strasbourg Court in Janowiec sought to lay down a bright line rule that there should be a maximum period of ten years between the relevant death and the critical date and that the only exception to that rule was if the enormity of the circumstances were such that, exceptionally, a longer period of time lapsed would be accepted as a sufficient connection under the “Convention values” test, where for example war crimes, genocide or crimes against humanity were alleged.
Ms Fiona Doherty KC for Rosaleen Dalton submits that on a proper interpretation the Strasbourg Court established a more flexible test of a “reasonably short period” in Janowiec. Depending on the particular circumstances of a case, that period could be 13 years as in Mladenović or Varnava v Turkey (Application Nos 16064/90 and others) (unreported) 18 September 2009 (“Varnava”). In any event, she submits that this court cannot be sure that the Strasbourg Court has sought to create an absolute limit in Janowiec.
We are persuaded that as part of the genuine connection test the Grand Chamber did lay down a bright line rule of a time limit of ten years in Janowiec in response to widespread criticism of the lack of clarity in the Šilih judgment. We have come to this view for four principal reasons.
First, the wording of the principal judgment: the critical paragraph (146), which we have quoted in para 131 above, makes it clear that the time factor is the “first and most crucial indicator” of the “genuine” nature of the connection. The Grand Chamber recognises that there are no apparent legal criteria with which to define an absolute limit on the duration of the period between the relevant death and the critical date but nonetheless decides that that period “should not exceed ten years”. The Grand Chamber recognises that it may be justified to extend the time limit further into the past in exceptional circumstances, but takes care to define such circumstances as being where the requirements of the “Convention values” test have been met. In the following paragraph (147) the Grand Chamber states that the duration of the lapse of time is not decisive in establishing a genuine connection. But this is not envisaging an extension of that time limit. On the contrary, the Grand Chamber makes the point about the time period not being decisive because it recognises a further requirement, in addition to the temporal limit, that “much of the investigation into the death took place or ought to have taken place in the period following the entry into force of the Convention”. Thus, even if the time lapse is under ten years, the connection will not be “genuine” unless there has been or ought to have been investigation post-critical date. The Grand Chamber makes clear in para 148 that there are two criteria for the genuine connection; they are (i) the time limit and (ii) a major part of the investigation must have been carried out, or ought to have been carried out, after the critical date.
Ms Doherty points out with justified emphasis that in the summary in para 148 the Grand Chamber states that the “period of time between the death as the triggering event and the entry into force of the Convention must have been reasonably short”; it does not state that that time must not exceed ten years. We address this point when we consider below the wider context of the Janowiec judgment. She also points out that the Grand Chamber, after discussing the Varnava and Mladenović cases among others earlier in its judgment, states in para 149 that “In all the cases outlined above the Court accepted the existence of a ‘genuine connection’ as the lapse of time between the death and the critical date was reasonably short and a considerable part of the proceedings had taken place after the critical date” (emphasis added). In our view, however, this statement is not sufficient to support the inference which Ms Doherty seeks to draw from it, namely that the Grand Chamber was accepting that, post-Janowiec, periods of 13 years could be treated as reasonably short, depending on the circumstances of the case, absent an extension under the Convention values test. We reach this view for four reasons.
First, in the earlier discussion of the Varnava judgment the Grand Chamber emphasises the difference between (i) the obligation to investigate a suspicious death, which had the “anchoring factual element” of the loss of life of the victim which was known as a certainty, and (ii) the obligation to investigate a suspicious disappearance, which, as stated in Varnava (para 148), was characterised by an “ongoing situation of uncertainty and unaccountability” during which the procedural obligation persisted: see paras 131-135 of the Janowiec judgment. The Grand Chamber’s discussion of Mladenović took place in its narrative of recent case law after the Šilih judgment (paras 136-139) and before the section of the judgment in which it sought to clarify the Šilih judgment in order to establish a regime with the legal certainty required in this area of the law. In addition, the statement on which Ms Doherty founds is in para 149 of the judgment where the Grand Chamber has turned to discuss the Convention values test. The point being emphasised by the Grand Chamber in that paragraph was that it had moved away from discussing the genuine connection test, which was the subject of consideration in the earlier passages of the judgment which we have discussed above.
Secondly, the wider context: the purpose of the Grand Chamber judgment in Janowiec was to give further clarification of the Šilih criteria as their application in practice had given rise to uncertainty (para 140). This brings to mind the concurring opinion of Judge Zagrebelsky joined by Judges Rozakis, Cabral Barreto, Spielmann and Sajó in Šilih in which (as mentioned above) they state (para O-III 4) that the court will be forced to carry out “complex and questionable assessments on a case-by-case basis that will be difficult to dissociate from the merits of the case”. A criterion of a “reasonably short period” unanchored by a ten-year time limit would involve such assessments and would not achieve the legal certainty to which the Janowiec judgment was intended to contribute.
Thirdly, the authors of the minority opinions in the Janowiec case understood at the time that the majority was laying down a bright line rule. In his partly concurring and partly dissenting opinion, Judge Wojtyczek explained his understanding of the decision of the majority on the genuine connection test (para O-III 8):
“In the instant case, the majority has proposed amending the criteria established in the Šilih judgment by limiting the retroactive effect given to the Convention in that judgment. First, they assert that the ‘genuine connection’ between an event and the ratification of the Convention exists if the lapse of time between the two is relatively short. Secondly, they set the maximum period for this lapse of time at ten years. Thirdly, while they accept that the requirements of protection of the Convention values may require acceptance of a longer time-limit, they set the time limit for retroactive application of the Convention at 4 November 1950.” (Emphasis added)
A similar understanding of the time limit on the genuine connection set by the majority can be found in the joint partly dissenting opinion of Judges Ziemele, De Gaetano, Laffranque and Keller, who discuss the second Šilih principle – the genuine connection test – at para O-IV 20:
“As to the ‘genuine connection’ between the triggering event and the entry into force of the Convention in respect of the respondent State, the majority’s finding emphasises the time element and, by referring to classical disappearance cases, repeats that the period concerned should not exceed ten years. In exceptional circumstances, the majority’s finding allows an extension of the time-limit further into the past, but only ‘on condition that the requirements of the ’Convention values’ test have been met’.” (Emphasis added)
Fourthly, the Registry of the Strasbourg Court in its “Practical Guide on Admissibility Criteria” (updated 28 February 2023) follows this understanding of the judgment in Janowiec and equates the reasonably short period criterion with a 10-year time limit. It states in its discussion of the procedural obligation under article 2 (para 324, omitting references):
“Firstly, only procedural acts and/or omissions occurring after the critical date can fall within the court’s temporal jurisdiction. Secondly, the court emphasises that in order for the procedural obligations to come into effect there must be a genuine connection between the death and the entry into force of the Convention in respect of the respondent State. Thus, for such connection to be established, two criteria must be met: firstly, the lapse of time between the death and the entry into force of the Convention must have been reasonably short (not exceeding ten years) and, secondly, it must be established that a significant proportion of the procedural steps – including not only an effective investigation into the death of the person concerned but also the institution of appropriate proceedings for the purpose of determining the cause of the death and holding those responsible to account – were or ought to have been carried out after the ratification of the Convention by the State concerned.” (Emphasis added)
While this practical guide is not binding on the Strasbourg Court, the equation of a “reasonably short period” with a time limit of not more than ten years is consistent with the words which the majority use in para 146 of Janowiec, with the minority opinions in that case which we have quoted and with the context of the judgment which was designed to bring greater legal certainty into the genuine connection test in light of the widespread criticism of the relevant criteria in the Šilih judgment. The leading work for practitioners interprets Janowiec in the same way: Karen Reid (ed.), A Practitioner’s Guide to the European Convention on Human Rights 7th ed (2023), p 40. In our view the statement by the Strasbourg Court in Mocanu (para 133 above) that the time lapse should not normally exceed ten years is not a softening of the strict time limit but a recognition that the time lapse may exceed ten years if, exceptionally, the Convention values test is met.
We recognise, however, that not all members of the panel who have heard this appeal agree with this analysis. That recognition has influenced our approach to the domestic authorities of Finucane and McQuillan which we discuss further below.
The impact of this analysis on this court’s judgment in Finucane
As was held by this court in McCaughey and as explained in McQuillan (paras 147-168), the proper interpretation of the HRA is that the Convention rights under it have application in accordance with the tests laid down in the case-law of the Strasbourg Court, but with a critical date of 2 October 2000: see para 145 above. In the present case, we are concerned with the application of the genuine connection test laid down by the Strasbourg Court, in particular as explained by the Grand Chamber in Janowiec. So far as concerns the temporal application of the HRA, that was also the position in Finucane.
However, as is clear from the foregoing discussion, it is not possible in our view to reconcile a critical part of the reasoning in Finucane with the strict approach to the long-stop time limit of ten years which forms part of the Strasbourg Court’s specification of the genuine connection test. In Janowiec the Grand Chamber held that the ten-year period was an absolute requirement of the genuine connection test, and that a longer period could only be justified in a case where the much more demanding Convention values test is satisfied (see paras 127 and 131 – 133 above). But in Finucane, at para 108, Lord Kerr said that the period of ten years or less between the triggering event and the critical date “is not an immutable requirement” and described the application of the genuine connection test as “a multi-factorial exercise” in which “the weight to be attached to each factor will vary according to the circumstances of the case”. (This is consistent with Lord Kerr’s approach in Keyu where he considered that the practicability of the enquiry was a factor which might extend the time period: see para 135 above.) On that basis, with the agreement of the other members of the court, he held that the genuine connection test was satisfied in Finucane notwithstanding a gap between the trigger event (Mr Finucane’s death) and the critical date (2 October 2000) of some 11 years eight months.
The decision in Finucane was not challenged in McQuillan, even though in McQuillan this court was concerned with the proper application of the genuine connection test derived from Janowiec. The court noted (para 144) that it had reservations as to whether Lord Kerr was right in his interpretation of Janowiec, but it had not been invited to overrule Finucane. It stated that Finucane:
“did not identify any clear principle by which one could tell when and to what extent it might be appropriate to water down a strict ten-year requirement as the Grand Chamber of the Strasbourg Court had appeared to lay down in Janowiec, para 146. We have reservations as to whether Lord Kerr JSC was right to interpret Janowiec as he did. This court has not been invited to depart from its decision in In re Finucane but we note that the extension beyond ten years allowed in In re Finucane involved less than two more years. It would significantly undermine the legal certainty which the Grand Chamber sought to achieve in Janowiec if longer extensions than this were to be contemplated or permitted. Moreover, in Janowiec, para 146, the Grand Chamber emphasised that the time factor is the ‘most crucial indicator’ in relation to the ‘genuine connection’ test and that the test requires that ‘the lapse of time between the triggering event and the critical date must remain reasonably short’.”
Nonetheless, the court accommodated the judgment in Finucane as it went on to state (also para 144):
“In our judgment, an extension beyond the normal ten year limit of up to two years is permissible where there are compelling reasons to allow such an adjustment constituted by circumstances that (a) any original investigation into the triggering death can be seen to have been seriously deficient and (b) the bulk of such investigative effort which has taken place post-dates the relevant critical date. If in these circumstances there is an extension of no more than two years beyond the ten-year limit mentioned in Janowiec, it remains possible to describe the lapse of time as ‘reasonably short’ in accordance with the guidance in that judgment at paras 146 and 148.”
Moreover, in McQuillan itself (and in the Hooded Men case to which the judgment also relates), the gap in time between the relevant trigger event and the critical date of 28 years or more was so great that it did not appear that the genuine connection test could be satisfied on either the strict approach in Janowiec or the multifactorial approach proposed in Finucane. Nonetheless, the court observed that “[i]t would significantly undermine the legal certainty which the Grand Chamber sought to achieve in Janowiec” if extensions longer than two years beyond the ten years referred to in Janowiec were permitted. Therefore, by way of giving guidance on such an important issue, even though it was not strictly required for the decision in the case, the court sought to re-establish legal certainty by stipulating a strict approach to the time limit aspect of the genuine connection test similar to that in Janowiec.
In the light of the fuller debate about the issue in this appeal, in which the decision and approach in Finucane has now been directly challenged before a seven-Justice panel, it is clear that the interpretation in that case of the judgment in Janowiec cannot be supported. The Grand Chamber laid down a strict ten-year time limit as one limb of the genuine connection test, not a flexible time limit capable of being extended beyond ten years according to a multifactorial approach.
The statement of the test in Finucane, para 108, was the result of an argument for the Secretary of State for Northern Ireland in that case regarding the effect of Janowiec which was materially different from that of Mr McGleenan in the present appeal. The nature and application of the genuine connection test was only one of a range of arguments which the Secretary of State had to address in that case. There had been an extensive investigation into the circumstances of the death (the review by Sir Desmond de Silva) and the Secretary of State was seeking to uphold the compatibility of that investigation with the standards required by article 2 of the Convention on the merits. Although the written case for the Secretary of State was consistent with the interpretation of Janowiec set out above, the oral submissions of counsel for the Secretary of State were less definitive and could be taken to have indicated that the time limb of the genuine connection test was a factor to be weighed against other factors. The ten-year rule laid down in Janowiec was said to be a “pretty serious line”, but one which was “not entirely absolute” and “not entirely unporous”; and reference was made to Mladenović as a valid example of the application of the Janowiec test rather than as an illustration of a trend which the Grand Chamber was concerned to rectify. Counsel did not refer to the minority opinions in Janowiec, reviewed above. The facts of the case were particularly abhorrent, and it is not surprising that counsel may have felt a degree of reticence in pressing an answer based on strict application of a bright-line time limit rule rather than seeking to justify the investigation itself.
In our view, the adoption in Finucane of the multifactorial approach to extending the time limit under the genuine connection test was an error which left the law in an unsatisfactory state. As explained in McCaughey and McQuillan, proper interpretation of the HRA involves application of a modified version of the mirror principle by applying the genuine connection test set out in the Strasbourg caselaw (in particular as now laid down in Janowiec) with the adaptation necessary to accommodate adjustment of the critical date to 2 October 2000, when the HRA came into effect.
This is an area in which the legal certainty afforded by an absolute time limit is important, as was emphasised by the Grand Chamber in Janowiec. To abandon such a time limit in favour of a multifactorial approach would introduce a high degree of uncertainty for the relevant public authorities, for persons whose Convention rights have arguably been violated and (in the case of a death) for their next-of-kin and family members. This would be unfair for all concerned; particularly so in the context of Northern Ireland, with its history of the Troubles. It would promote litigation around the application of an inherently unclear standard in a context potentially involving a substantial number of cases, where this would be likely to undermine community cohesion and reconciliation, involve a significant diversion of public resources and embroil the courts in having to make difficult and politically sensitive decisions without any readily understandable guidance in law. In our view, for the reasons explained in Janowiec, it is far more appropriate that the genuine connection test should apply with a high degree of certainty as to time limit, so that all parties can know where they stand and the courts have a clear rule to apply.
This is the context in which the court has been invited to apply the 1966 Practice Statement (Practice Statement (Judicial Precedent) [1966] 1 WLR 1234) in this appeal, so as to depart from the reasoning in Finucane at para 108, to apply correctly the genuine connection test laid down in Janowiec and to draw the inference, which must necessarily follow, that the actual decision in Finucane was wrong.
Our initial view, after hearing the arguments of the parties, was that this is a case in which it was right for the court to depart from its previous decision in Finucane in accordance with the approach laid down by the 1966 Practice Statement. We are persuaded that the interpretation of Janowiec in Finucane was wrong. Not only does it depart from a correct understanding of the text of the judgment in Janowiec, but also it undermines the fundamental rationale for the test laid down in that case, which was to promote legal certainty in this difficult area. This has the detrimental effects for individuals and the public interest referred to above.
On further reflection, having considered the judgments of Lord Reed, Lord Leggatt, and the combined judgment of Lord Burrows and Dame Siobhan Keegan, it is clear (i) that the panel is not at one in interpreting Janowiec and (ii) that, as set out in the judgment of Lord Burrows and Dame Siobhan Keegan, this court in McQuillan has devised a genuine connection test for the HRA which, while not a precise application of the mirror principle, achieves a high degree of legal certainty. The test is that, in the absence of acts which meet the Convention values test, the period between the triggering event (the death) and the critical date should normally not exceed ten years but there can be an extension for a further two years to an outer limit of 12 years only for the compelling reasons specified in para 144 of the McQuillan judgment, namely (i) that any initial investigation was seriously deficient and (ii) the bulk of the investigative effort has taken place after the critical date.
For these reasons and because it is not necessary for the determination of this appeal, it would not meet the criteria of the 1966 Practice Statement for this court to depart from its earlier judgment in McQuillan at para 144.
Conclusion on Issue 1
It may be helpful if we summarise our conclusions, which are in line with the reasoning of Lord Burrows and Dame Siobhan Keegan, especially at paras 333 - 334 and 337, on the test to be applied on the question of a genuine connection with the coming into force of the HRA. We consider that the authorities support the following summary:
We would uphold the result in Finucane while not agreeing with the multi-factorial reasoning in that judgment as we treat the judgment in McQuillan as the correct analysis of the domestic test.
That means that there is an outer period of 12 years since the death of the person (unless the Convention values test is met) in order to bring a claim under the HRA.
In other words, if the death occurred more than 12 years before 2 October 2000, those proceedings should be struck out, even if there has been a Brecknell-type revival of the duty more recently.
If the death occurred between 10 and 12 years before 2 October 2000 then in exceptional circumstances there may be a genuine connection with the coming into force of the HRA (even leaving to one side the Convention values test). Those circumstances (as explained in McQuillan at para 144) are: (i) any original investigation into the death can be seen to have been seriously deficient or non-existent and (ii) the bulk of such investigative effort which has taken place, or which ought to have taken place, post-dates 2 October 2000.
If the death occurred less than 10 years before 2 October 2000, then the claim meets the “within a reasonable time” limb of the test but must still overcome the second part of the Janowiec test that a major part of the investigation took place or ought to have taken place after 2 October 2000.
For the reasons given above, as Mr Dalton’s death occurred more than 12 years before the critical date and as 12 years is the outer limit of the temporal element of the genuine connection test, and as there is no question of the Convention values test being satisfied in this case, the appeal must be allowed. The domestic courts do not have jurisdiction under the HRA in respect of Mr Dalton’s death. That being so, it is not appropriate to express any concluded view on the application of article 2 should Mr Dalton’s family seek to bring a claim in the Strasbourg Court.
Issue 2: the ambit of the revival principle in Brecknell
The conclusion on Issue 1 is sufficient to dispose of the appeal in the present case. However, the ambit of the Brecknell principle regarding revival of the duty of investigation was addressed by the courts below. There was a difference of view. The Court of Appeal held that on a proper interpretation of Brecknell the revival of a duty of investigation is not limited to an inquiry to see whether it is possible to bring the perpetrator of a death to trial, but extends more widely to require, in an appropriate case, an inquiry into the responsibility of agents of the state who may have failed in some way to take adequate steps to protect the public. The Attorney General was given permission to appeal on the question whether this interpretation of Brecknell is correct. The court heard full argument on the point. Counsel for the Attorney General invited us to address this issue even if the Attorney General were to succeed on issue 1.
The court is aware that this issue is likely to be live in other cases at first instance in Northern Ireland. Since the present case has been treated as a test case to provide guidance for other cases concerned with deaths which occurred during the Troubles, we are persuaded that it is appropriate to address this issue.
Apart from the debate about jurisdiction under the HRA (issue 1 above), the primary defence advanced by the AGNI in the courts below was that the Brecknell revival principle only requires a new investigation if it is realistically capable of leading to the identification and punishment the perpetrators of a death. In the present case it is common ground that there is no realistic prospect that the perpetrators will be identified and put on trial. Therefore, the AGNI contends that she is not subject to any obligation under article 2 to direct a new inquest to investigate the circumstances of the deaths in this case.
At first instance, this contention of the AGNI was accepted. However, the Court of Appeal allowed Mrs Dalton’s appeal on this point (see paras 108 – 109 above). On this appeal, the AGNI submits that the Court of Appeal erred on this point and that the judge was correct.
This issue turns on a proper understanding of the judgment of the Strasbourg Court in Brecknell. In our view, the Court of Appeal’s reading of that judgment was correct and it was correct to hold that the revival principle is not limited to the aspect of the article 2 procedural obligation which is directed to the identification and punishment of perpetrators.
In the first place, we would point out that there is no sound reason why the revival principle should be limited in the manner proposed by the AGNI, and Mr McGleenan could suggest none. On the contrary, the revival principle is concerned with revival of the article 2 procedural obligation in circumstances where it has since transpired that the objectives of that obligation have not been completely fulfilled by a previous investigation and remain reasonably capable of being fulfilled despite the closure of that investigation and the passage of time. As explained in the Strasbourg and domestic case-law, the objectives of the article 2 procedural obligation are wider than the identification and punishment of the perpetrators. With a view to ensuring that proper standards in relation to the protection of life are upheld by the authorities of the state, they extend to examination of the circumstances in which the relevant authorities, knowing of a threat to life, failed to take adequate measures to diminish that threat and a death occurs as a result: see, among other authorities, Osman v United Kingdom (1998) 29 EHRR 245 and C. Grabenwarter, European Convention on Human Rights: Commentary (2014), p 27, citing Branko Tomašić v Croatia (Application No 46598/06) (unreported) 15 January 2009, paras 52ff; Maiorano v Italy (Application No 28634/06) (unreported) 15 December 2009, paras 127ff; and Dink v Turkey (Application Nos 2668/07 and others) (unreported) 14 September 2010, paras 77ff. If it emerges from new information that those wider objectives of the article 2 procedural duty have not been fulfilled and there remains a reasonable prospect that they can be, the revival principle should apply for reasons similar to those which justify the imposition of the wider aspects of the article 2 procedural obligation in the first place.
Secondly, we consider that, on a proper reading of the Brecknell judgment, it endorses this interpretation of the width of the revival principle. On its facts, Brecknell was concerned with the identification and punishment of perpetrators in circumstances where a murder was carried out by paramilitary terrorists and, some years after the initial police investigation, new information emerged which suggested that members of the state security forces might also have been involved. This explains why part of the judgment of the Strasbourg Court (para 71) had a focus on the identification and punishment of the perpetrators, on which counsel for the AGNI fastened for the purposes of his submission. But the reasoning of the Court was not limited to this; it explained how the revival principle is located in the context of the general article 2 procedural obligation and is capable of application in relation to all aspects of that obligation. The Court did not say that the revival principle is confined to the identification and punishment of perpetrators.
The Strasbourg Court opened its statement of the applicable principles by referring (para 65) to the general procedural obligation under article 2, established in its case-law, to carry out an effective investigation into unlawful or suspicious deaths and emphasised that:
“the essential purpose of such investigation is to secure the effective implementation of the domestic laws which protect the right to life and, in those cases involving state agents or bodies, to ensure their accountability for deaths occurring under their responsibility. Furthermore, even where there may be obstacles or difficulties which prevent progress in an investigation in a particular situation, a prompt response by the authorities is vital in maintaining public confidence in their adherence to the rule of law and in preventing any appearance of collusion in or tolerance of unlawful acts.”
At para 66 the Court referred to a situation in which, some time after an initial investigation, information “purportedly casting new light on the circumstances of the death comes into the public domain” and stated that “[t]he issue then arises as to whether, and in what form, the procedural obligation to investigate is revived.” The Court was plainly referring to the general article 2 procedural obligation.
At para 67 the Court referred to McKerr v United Kingdom, a case in which later information gave rise to concerns about excessive use of force by the RUC and the deliberate concealment of evidence, and quoted from its judgment in that case where it said “there may be circumstances where issues arise that have not, or cannot, be addressed in a criminal trial and that article 2 may require wider examination” since “the aims of reassuring the public and members of the family as to the lawfulness of the killings had not been met adequately by the criminal trial”, and it had found that article 2 required a procedure “whereby these elements could be examined and doubts confirmed, or laid to rest.” Again, the Court was clearly referring to the general article 2 procedural obligation, including aspects of it (such as investigation of the possibility of concealment of evidence) which were not concerned with the identification and punishment of perpetrators.
At para 68 the Strasbourg Court again referred to the general article 2 procedural obligation and the wider aspects of that obligation when it cited its decision in Hackett v United Kingdom (Application No 34698/04) (unreported) 10 May 2005, and said that later events or circumstances may arise which cast doubt on the original investigation or trial “or which raise new or wider issues” so that “an obligation may … arise for further investigations to be pursued.” The Court then observed, “the nature and extent of any subsequent investigation required by the procedural obligation would inevitably depend on the circumstances of each particular case and might well differ from that to be expected immediately after a suspicious or violent death has occurred.” The Court did not suggest that the obligation could only be revived so far as concerned the identification and punishment of perpetrators.
At para 69 the Court referred to the fact that “the public interest in obtaining the prosecution and conviction of perpetrators [many years after a killing] is firmly recognised, particularly in the context of war crimes and crimes against humanity”, but this was simply to support a comment that there is little ground to be overly prescriptive as regards the revival “of an obligation to investigate unlawful killings arising many years after the events”. This latter reference was to the general article 2 procedural obligation, with all its aspects.
The judgment continues at paras 70-72:
“70. The Court would, however, draw attention to the following considerations. It cannot be the case that any assertion or allegation can trigger a fresh investigative obligation under article 2 of the Convention. Nonetheless, given the fundamental importance of this provision, the state authorities must be sensitive to any information or material which has the potential either to undermine the conclusions of an earlier investigation or to allow an earlier inconclusive investigation to be pursued further. Both parties have suggested possible tests. The Court has doubts as to whether it is possible to formulate any detailed test which could usefully apply to the myriad of widely differing situations that might arise. It is also salutary to remember that the Convention provides for minimum standards, not for the best possible practice, it being open to the contracting parties to provide further protection or guarantees. For example, contrary to the applicant’s assertion, if article 2 does not impose the obligation to pursue an investigation into an incident, the fact that the state chooses to pursue some form of inquiry does not thereby have the effect of imposing article 2 standards on the proceedings. Lastly, bearing in mind the difficulties involved in policing modern societies and the choices which must be made in terms of priorities and resources, positive obligations must be interpreted in a way which does not impose an impossible or disproportionate burden on the authorities [footnote reference to Osman v United Kingdom (1998) 29 EHRR 245].
71. With those considerations in mind, the Court takes the view that where there is a plausible, or credible, allegation, piece of evidence or item of information relevant to the identification, and eventual prosecution or punishment of the perpetrator of an unlawful killing, the authorities are under an obligation to take further investigative measures. The steps that it will be reasonable to take will vary considerably with the facts of the situation. The lapse of time will, inevitably, be an obstacle as regards, for example, the location of witnesses and the ability of witnesses to recall events reliably. Such an investigation may in some cases, reasonably, be restricted to verifying the credibility of the source, or of the purported new evidence. The Court would further underline that, in light of the primary purpose of any renewed investigative efforts [footnote reference to para 65], the authorities are entitled to take into account the prospects of success of any prosecution. The importance of the right under article 2 does not justify the lodging, willy-nilly, of proceedings. As it has had occasion to hold previously, the police must discharge their duties in a manner which is compatible with the rights and freedoms of individuals and they cannot be criticised for attaching weight to the presumption of innocence or failing to use powers of arrest, search and seizure having regard to their reasonably held view that they lacked at relevant times the required standard of suspicion to use those powers or that any action taken would not in fact have produced concrete results [footnote reference to Osman v United Kingdom, para 121].
72. The extent to which the requirements of effectiveness, independence, promptness and expedition, accessibility to the family and sufficient public scrutiny apply will again depend on the particular circumstances of the case, and may well be influenced by the passage of time as stated above. Where the assertion or new evidence tends to indicate police or security force collusion in an unlawful death, the criterion of independence will, generally, remain unchanged [footnote reference to Ramsahai v Netherlands [(2007) 46 EHRR 34], paras 325, 333-341]. Promptness will be likely not to come into play in the same way, since, for example, there may be no urgency as regards the securing of a scene of the crime from contamination or in obtaining witness statements while recollections are sharp. Reasonable expedition will remain a requirement, but what is reasonable is likely to be coloured by the investigative prospects and difficulties which exist at such a late stage.”
In these paragraphs the Strasbourg Court was again plainly intending to indicate that the revival principle applies in relation to the general article 2 procedural obligation, with all its aspects: see, in particular, the opening part of para 70 and the references in para 72 to the various requirements which are a feature of that general obligation, which have wider application than simply being concerned with the identification and punishment of perpetrators.
In support of his submission that the revival principle is limited to the identification and punishment of perpetrators, Mr McGleenan emphasised para 71. But that paragraph has to be read in the context of the whole statement by the Strasbourg Court of the applicable principles, at paras 65-72, and in the context of the nature of the complaint being made on the particular facts of the case. The Court did not say that the revival principle was limited to the specific aspect of the article 2 procedural obligation which is concerned with the identification and punishment of perpetrators. Indeed, counsel’s reliance on this paragraph is clearly misplaced, since by the footnote the Court explained the primary purpose of any renewed investigative efforts by reference to the general objective it had identified earlier, at para 65 of the judgment, which gives rise to all the aspects of the ordinary article 2 procedural obligation.
Although it is clear that the revival principle is not confined to the identification and punishment of perpetrators, at paras 68 and 70-72 the Strasbourg Court emphasised that the extent to which the article 2 procedural obligation is revived is dependent on the particular facts of each case, bearing in mind that any revival of the obligation should not impose a disproportionate burden on the authorities and that they are only required to take steps which are reasonable. It is not the case that, where new plausible or credible information emerges, there is a complete revival of the article 2 procedural obligation with the same force and effect as it would have had if the death had only just occurred.
Rather, there is a spectrum of obligation which may apply regarding further inquiry. If new information bearing on the circumstances of a death emerges, there should be some process of evaluation to see if it is plausible or credible. If it is found to be plausible or credible to a sufficient degree, a further evaluation is required to see whether a renewed investigation in the light of that information would be capable of serving any or all of the objectives which an article 2 investigation is directed towards and whether it is reasonable to pursue those objectives. This is likely to require taking into account the resources available, current calls upon those resources and the degree to which such objectives are likely to be achieved.
It is not determinative that it may now be impossible to identify and punish the perpetrators, but that is a relevant factor. The passage of time may also have a bearing on the extent to which other relevant objectives can be achieved. If all relevant police officers have retired, it will not be relevant to seek to commence disciplinary action against them. Also, public confidence in the enforcement of the legal regime designed to ensure the protection of the right to life may be better maintained by dedication of available resources to combatting and investigating current or recent crimes and threats to life than events long in the past, although a great deal will depend upon the extent of alleged involvement of state authorities in the crimes to which the new information relates.
Issue 3: whether the investigations which were carried out were effective investigations which complied with any revived article 2 procedural obligation
As mentioned above, since article 2 has no application in this case for jurisdictional reasons (issue 1), it would not be appropriate for us to give any final ruling on this question. In the light of the analysis in Brecknell and at paras 190 – 191 above, it does not seem to us to be a straightforward matter.
If the deaths in question had just occurred, so that the general article 2 procedural obligation applied with full force and effect, the fact that civil proceedings were being or could be brought to allege negligence against the police would not be sufficient to satisfy that obligation. Nor would the investigation by the PONI have satisfied it, since the non-compellability of relevant witnesses would plainly have been problematic in that regard: see Finucane.
But with the lapse of time the circumstances now are very different and the extent of the revived article 2 procedural obligation is a matter for debate. It is common ground that there is no realistic possibility that further investigation could lead to the identification and punishment of the perpetrators. The deficiencies identified by the PONI in the investigation in the immediate aftermath of the murders and the failure to gather or preserve evidence at that time could not be remedied by any new investigation. After the passage of so many years, even if retired police officers were now compelled to give evidence it is very doubtful that they could add anything to elucidate the facts beyond what appears from the contemporary documentary material already reviewed by the PONI. The PONI has investigated the circumstances of the deaths with considerable care and attention, including involving the families of the deceased. Independent investigation by the PONI is capable of making a substantial contribution in holding agents of the state accountable for their actions and omissions and has done so in this case. It is significant that, in consequence of his thorough investigation, the PONI felt able to make a series of findings, including that certain allegations of inadequate performance of their duty by the police were substantiated. He has reported these findings publicly and in considerable detail. It is also highly significant that the PONI found that the most contentious allegation against the police, that they had deliberately failed to act in order to protect an informant, was not substantiated. Further, the civil proceedings in this case give additional scope for involvement of the families and could potentially lead to a detailed examination of the facts by a judge in a public judgment. These are matters which are relevant to the question whether any revived article 2 procedural obligation is now of such force and potency as to require a new inquest or any other investigatory measures. Although we make no final ruling on the point, it seems to us to be strongly arguable that no new inquest is required. However, that is a question which can be raised before the Strasbourg Court, which would have jurisdiction to consider it.
Conclusion
We would allow the appeal on the grounds that this court has no jurisdiction to consider the application for judicial review under section 7(1)(b) of the HRA because the exception to the general non-retrospectivity of that Act arising from the Janowiec decision of the Strasbourg Court and the McCaughey and McQuillan decisions of this court mean that the claim is outside the temporal scope of the HRA. Accordingly, the application for judicial review must be dismissed.
LORD LEGGATT:
Introduction
On this appeal this court is asked to consider, once again, the scope of the state’s procedural duty under article 2 of the European Convention on Human Rights to investigate deaths for which the state may bear a responsibility. More particularly, the court is asked to consider the temporal scope of this duty imposed on public authorities as a matter of UK law by the Human Rights Act 1998. How far does the duty extend to deaths that occurred before the Act came into force in 2000?
This question has been the subject of no fewer than six decisions of the UK’s highest court, two of them given within the last five years. Most recently, in Re McQuillan, Re McGuigan, Re McKenna [2021] UKSC 55, [2022] AC 1063 a unanimous panel of seven Justices gave guidance on the question which was intended to be definitive. The present appeal was already pending when the judgment in McQuillan was handed down in December 2021, and the guidance given in that judgment is sufficient to dispose of this appeal. It has not, however, prevented each party from seeking to persuade the court to revisit its judgment in McQuillan. The Attorney General for Northern Ireland goes further. Her counsel makes the ambitious submission that we should resile, not only from the guidance given in McQuillan, but also from this court’s decision in In re Finucane (Northern Ireland) [2019] UKSC 7, [2019] NI 292. The Attorney invites us to conclude that the court reached the wrong result in that case and, on the basis of that conclusion, to overrule it.
This invitation should be firmly declined. To overrule the decision in Finucane would be destructive of consistency and certainty in the law in an area where those values are of even more than ordinary importance. Nothing has changed since Finucane was decided in 2019 which might justify reconsidering its correctness. The decision has given rise to no injustice - unless it be an injustice for the Supreme Court of this country to decide that there had been no adequate inquiry into events that were an affront to the rule of law in Northern Ireland rather than requiring the victim’s family to apply to have that question decided by the European Court of Human Rights in Strasbourg. The Finucane family have since brought fresh proceedings in the courts of Northern Ireland in which they are relying - as they are entitled to rely - on the correctness of this court’s decision in 2019. Such justified reliance provides yet further reason, had any further reason been called for, to stand by what this court has decided.
I will need to return to this question. But before doing so, I will outline the facts and address the two other issues raised on the appeal.
The “Good Samaritan” bombing
On 31 August 1988 Sean Eugene Dalton was killed in a bomb explosion at 38 Kildrum Gardens in Derry / Londonderry. Sheila Lewis and Thomas Curran also died from the blast. Those three people had gone to the flat out of concern for the welfare of a neighbour who had not been seen for several days. When they got no response from inside, Mr Dalton managed to gain access through the kitchen window. As he went to open the front door of the flat, the bomb was triggered by a booby trap. Because the victims were killed as a result of going to the aid of a fellow human being, their murders have become known as the “Good Samaritan” bombing.
Soon afterwards the Provisional IRA admitted responsibility for planting the bomb. Apparently, the occupant of the flat had been abducted, the bomb planted and clues pointing to the flat or its vicinity created in an attempt to lure members of the security forces to the flat with the intention that they would be killed upon entering it.
The police investigated the murders, but no one was charged. The perpetrators have never been identified. The investigation was closed down in 1989. In December 1989 an inquest was held which determined how the victims had died but did not examine what the police knew in advance of the explosion.
The Police Ombudsman’s investigation
Some 16 years later, in February 2005, Mr Dalton’s son lodged a complaint with the office of the Police Ombudsman for Northern Ireland. The complaint alleged that before the bomb explosion the police had failed to warn the local community of possible terrorist activities in the area; that the police had knowingly allowed an explosive device to remain in a location close to where the public had access in order to protect a police informant; and that the police had failed properly to investigate the murders.
The office of the Police Ombudsman investigated the allegations. The Ombudsman did not report until July 2013, over eight years later. The report, which ran to some 63 pages, described the investigation undertaken as “wide-ranging and thorough”. Documents and intelligence material were assessed. A public appeal for witnesses had been made, 65 potential witnesses were identified and 42 of these were interviewed. (Others could not be interviewed because they had died, were in poor health or in some cases declined to cooperate.) Witness statements were taken from 23 individuals. The report said that some retired police officers had provided valuable information but that the investigation had been hampered by the refusal of a number of retired police officers, some formerly of senior rank, to co-operate and also by the fact that significant documentation from the original police investigation had been lost.
In his report the Ombudsman rejected the allegation that the police had failed to act in order to protect an informant. But he found the other complaints to be substantiated. In particular, he found that, whilst he could not be certain the police knew there was a bomb specifically at 38 Kildrum Gardens, there was strong evidence that the police had sufficient information and intelligence to identify the location of the bomb, that they ought to have known it was in the vicinity of 38 Kildrum Gardens and that steps could and should have been taken to mitigate the threat and to warn the local community. These steps were not taken and the focus of the police effort appeared to have been the protection of officers from the terrorist threat. The Ombudsman also found that the police investigation of the murders had been inadequate.
These proceedings
After the Ombudsman’s report was issued, Mr Dalton’s family requested the Attorney General for Northern Ireland to exercise the power under section 14 of the Coroners Act (Northern Ireland) 1959 to direct a fresh inquest into the death of Mr Dalton for the purpose of examining, in particular, the extent of any responsibility of the police for his death. The Attorney decided not to do so. The principal reasons given for this decision were that the circumstances had already been the subject of a detailed examination by the Ombudsman and that the Attorney had not been provided with any evidence which would suggest that a fresh inquest could lead to the identification and punishment of those responsible for Mr Dalton’s murder. The final decision was communicated to the family’s solicitor in March 2015.
Leave was given to one of Mr Dalton’s daughters to apply for judicial review of that decision on the ground that a fresh inquest is necessary to comply with article 2 of the European Convention on Human Rights. The application for judicial review was dismissed by the High Court (Deeny J) [2017] NIQB 33; but in May 2020 the Court of Appeal allowed an appeal by Rosaleen Dalton [2020] NICA 26; [2021] NI 405, and granted a declaration that there has not been an enquiry into the death of Sean Eugene Dalton that is compliant with article 2.
Alongside these proceedings, a civil claim for damages has been brought by Mr Dalton’s estate against the Chief Constable of the Police Service of Northern Ireland and the Ministry of Defence.
Article 2 of the Convention
Article 2 of the Convention states: “Everyone's right to life shall be protected by law. No one shall be deprived of his life intentionally …” This most basic and fundamental of the Convention rights imposes on contracting states both a negative obligation not to take life intentionally and unlawfully and also, in the first sentence, a positive obligation to take appropriate steps to safeguard the lives of those within their jurisdiction. In Osman v United Kingdom (1998) 29 EHRR 245, para 116, the European Court of Human Rights held that such a positive obligation arises where:
“the authorities knew or ought to have known at the time of the existence of a real and immediate risk to the life of an identified individual or individuals from the criminal acts of a third party and … failed to take measures within the scope of their powers which, judged reasonably, might have been expected to avoid that risk.”
There is no suggestion in this case that any agent of the state was implicated in planting the bomb which killed Mr Dalton. But the findings of the Ombudsman disclose an arguable breach of a positive duty under article 2 to take steps to safeguard the lives of the inhabitants of Kildrum Gardens. It is not a breach for which a remedy can be sought in the courts of the United Kingdom. It was not until 2 October 2000, over 12 years after Mr Dalton’s death, that the Human Rights Act came into force, which gives domestic effect to the Convention rights by making them enforceable before courts in the United Kingdom against public authorities. Under section 6(1) of the Act it is unlawful for a public authority to act in a way which is incompatible with a Convention right and, by section 6(6), this includes a failure to act. Section 7 enables a person who claims that a public authority has acted in a way which is incompatible with a Convention right to bring proceedings against the authority under the Act in a domestic court. The Act itself, however, in section 22, makes it clear that this does not apply to an act that took place before the Act came into force - a conclusion confirmed by case law: see Wilson v First County Trust Ltd (No 2) [2004] UKHL 40, [2003] 1 AC 816.
The procedural duty
In addition to the substantive obligations already mentioned, article 2 of the Convention, as it has been interpreted, also imposes on the state a procedural duty to investigate deaths wherever life has been lost in circumstances which potentially engage the state’s responsibility: see eg Öneryildiz v Turkey (2005) 41 EHRR 20, paras 91-94; Da Silva v United Kingdom (2016) 63 EHRR 12, para 229. Where such a procedural duty arises, the basic requirement is that the investigation must be effective. This embraces a number of elements: the investigation must be independent; it must be adequate, albeit that this is “not an obligation of result but of means”; it must be carried out with reasonable promptness and expedition; it must involve a sufficient element of public scrutiny to secure accountability; and the victim’s next of kin must be involved to the extent necessary to safeguard their legitimate interests. See eg Jordan v United Kingdom (2001) 37 EHRR 2, paras 106-109; Tunç v Turkey [2016] Inquest LR 1, paras 169-179; Da Silva v United Kingdom 63 EHRR 12, paras 229-235.
There are cases in which, years after a death occurred and any original investigations were concluded, further information emerges which potentially casts new light on the circumstances of the death. In Brecknell v United Kingdom (2007) 46 EHRR 42 the European court held that in such cases a duty may arise to take further investigative steps. The applicant in the present case relies on this principle to argue that, since the Human Rights Act came into force in 2000, information suggesting that the police bear a responsibility for Mr Dalton’s death has come to light which has given rise to a procedural duty to take further investigative steps; that this duty has domestic effect under the Act; and that the duty has not been satisfied.
The issues
In the order in which I will address them, the issues raised on the appeal are accordingly these:
Did the allegations of police responsibility for Mr Dalton’s death give rise to a fresh duty under article 2 of the Convention to carry out some form of effective official investigation?
If so, has that duty been satisfied by the Police Ombudsman’s investigation or is it still unfulfilled?
If there is such a duty which has not been fulfilled, is it enforceable in the courts of the United Kingdom under the Human Rights Act, or only on the international plane by an application to the European court in Strasbourg? (This issue turns on the temporal scope of the Act.)
Did a new duty to investigate arise?
The facts of the Brecknell case were that, more than 23 years after a sectarian killing in Northern Ireland and more than 17 years after the original investigation into the killing had ended, a witness came forward who made apparently plausible allegations that the security forces had colluded in the killing. In considering whether, and in what form, the procedural duty to investigate the circumstances of a death may be “revived”, the Fourth Section of the European court expressed the view, at para 71, that:
“where there is a plausible, or credible, allegation, piece of evidence or item of information relevant to the identification, and eventual prosecution or punishment of the perpetrator of an unlawful killing, the authorities are under an obligation to take further investigative measures.”
The Attorney relies on this statement to argue that the principle illustrated by the Brecknell case is limited to allegations, evidence or information “relevant to the identification, and eventual prosecution or punishment of the perpetrator of an unlawful killing”. Here the allegations made by Mr Dalton’s son in 2005 and investigated by the Ombudsman were not directed at identifying and prosecuting the perpetrators of Mr Dalton’s unlawful killing. Rather, they were directed at alleged failures of the police to take steps to protect the lives of local residents. The Attorney submits that it cannot in these circumstances be said there was a duty to investigate the allegations, applying the Brecknell test.
The Court of Appeal was, in my opinion, right to reject this submission, which takes too narrow a view of the applicable principle. The statement in Brecknell quoted above did not purport to be a comprehensive test. Indeed, the European court had doubts “whether it is possible to formulate any detailed test which could usefully apply to the myriad of widely differing situations that might arise” (para 70). The reference to “the identification, and eventual prosecution or punishment of the perpetrator of an unlawful killing” is explained by the fact that this was the relevance of the new allegations in the Brecknell case itself. But the logic of the decision is wider. The procedural duty imposed by article 2 is not limited to the identification, prosecution and punishment of individuals who have perpetrated an unlawful killing. It unquestionably extends to investigating potential breaches of the state’s positive obligation to protect life. Examples (among many that could be given) of such cases are: Edwards v United Kingdom (2002) 35 EHRR 19, para 74 (death in custody of a person killed by another detainee); Öneryildiz v Turkey 41 EHRR 20, paras 91-94 (lethal methane explosion at a rubbish tip for which municipal authorities were responsible); and Mastromatteo v Italy (Application No 37703/97) (unreported) 24 October 2002, para 92 (murder committed by a prisoner released on prison leave). The rationale for investigating new credible information applies equally where the information may disclose culpability on the part of the state authorities as where it could lead to the prosecution of the perpetrator of an unlawful killing. An example of a case in which a fresh obligation to investigate alleged state responsibility arose is Harrison v United Kingdom (2014) 59 EHRR SE1, para 58, where the European court held that findings of a report published in 2012 had “revived the positive obligation of the British state to carry out adequate investigations into the cause and circumstances of the Hillsborough tragedy” in 1989.
Counsel for the Attorney also argues that the Brecknell case can be distinguished on the ground that, unlike the new material in Brecknell, the allegations of police responsibility made here, even if substantiated, were not capable of undermining the conclusions of the original police investigation, namely that the bomb that resulted in Mr Dalton’s death was planted by the IRA; nor were they capable of taking the earlier investigation any further, in that they could not assist in identifying the individuals responsible for planting the bomb. Hence the “revival” doctrine does not apply. Again, however, this takes too restricted a view of the legal principle. Although the court in Brecknell, at para 66, described the issue as whether the procedural obligation was “revived”, I do not think that this expression is strictly apt. What is triggered by a new, credible allegation, piece of evidence or item of information is not the revival of an old obligation which had ceased but the creation of a new obligation to investigate the allegation or material (in conjunction with any relevant information previously available). The fact that the subject of the new allegation or material has not been investigated before cannot be a good reason why no investigation should take place. To the contrary, the absence of any earlier investigation of the matter is part of what triggers an investigative duty.
I think it plain that in this case a procedural duty under article 2 was triggered by the complaint made to the Police Ombudsman in 2005, as the Court of Appeal held.
Has the duty been satisfied?
The Attorney next contends that, if a duty to investigate the allegations of police responsibility did arise, it has been satisfied by the Ombudsman’s investigation, together with the availability of civil proceedings. I consider this contention to be well founded.
It is important to recognise that, where a duty to take further investigative steps is triggered by a later allegation or information, the nature and extent of the investigation required may well differ from that to be expected immediately after the death occurred: Harrison v United Kingdom [2014] ECHR 511, para 51. In its judgment in Brecknell, para 71, the European court emphasised that in a such case “the steps that it will be reasonable to take will vary considerably according to the facts of the situation” and that “the lapse of time will, inevitably, be an obstacle as regards, for example, the location of witnesses and the ability of witnesses to recall events reliably”. Furthermore, the extent to which the requirements of effectiveness, independence, promptness, public scrutiny and accessibility to the family apply will depend on the particular circumstances of the case and “may well be influenced by the passage of time” (para 72). In some cases the investigation may, “reasonably, be restricted to verifying the credibility … of the purported new evidence” (para 71).
The judgment in Brecknell further underlined that the Convention provides for minimum standards, not for the best possible practice (para 70); that the authorities are entitled to take into account the prospects of success of any further measures (paras 71 and 72); and that, “bearing in mind the difficulties involved in policing modern societies and the choices which must be made in terms of priorities and resources, positive obligations must be interpreted in a way which does not impose an impossible or disproportionate burden on the authorities” (para 70, citing Osman).
In this case the allegations of police culpability made in 2005 have been extensively investigated by the Ombudsman. The independence of the Ombudsman’s investigation has not been questioned in these proceedings and his description of the investigation as “wide-ranging and thorough” is borne out by the account in his report of the steps taken (summarised at para 204 above).
The Court of Appeal nevertheless did not think that the investigation carried out was adequate: [2020] NICA 26; [2021] NI 405, paras 118-124. They found two deficiencies. First, they criticised the Ombudsman’s report for not specifying in detail the evidence underlying key conclusions. That criticism in my view expects more than the procedural duty under article 2 requires. The report sets out the investigative steps taken, facts established and conclusions drawn in sufficient detail to show that the conclusions were arrived at on the basis of a thorough and objective review of the available evidence. The level of detail provided is enough to allow for adequate public scrutiny of the investigation and explain the basis of its conclusions to the victims’ families and to the public at large. There is no requirement that the investigator should describe the underlying evidence in the degree of detail that might be expected, for example, in a court judgment.
The second and more substantial point made by the Court of Appeal is that the Ombudsman felt obliged to say that his investigation had been “significantly hampered” by two matters: the fact that documentation from the original police investigation had been lost and the fact that a “substantial number” of retired police officers had declined to assist. In particular, no former Special Branch officers had co-operated with the investigation nor did the local Divisional Commander at the time of the incident. I agree with the Court of Appeal that this limitation inevitably detracts from the level of public confidence a report of this type will enjoy. I do not, however, agree with their view that it “points towards the need for still further probing of the facts”: [2020] NICA 26; [2021] NI 405, para 121.
The relevant question is not whether the Ombudsman’s investigation was as complete or comprehensive as any inquiry could have been at the time when it took place: it is whether in March 2015, when the Attorney refused to direct a fresh inquest, there were further investigative steps that it was reasonable to take, having regard to the passage of time since the relevant events occurred. In considering that question, the following six points are relevant.
First, the two matters which the Ombudsman said had hampered his investigation did not prevent him from reaching definite conclusions on the allegations that he investigated. As the Court of Appeal observed, he did not express any lack of confidence in his findings or propose that any further investigation was needed. It is difficult to see how in these circumstances a further investigation could be warranted unless there was reason to anticipate that it would obtain significant further evidence which might justify different conclusions.
Second, in that regard there is no reason to suppose that further searches might locate the documentation missing from the records of the original police investigation carried out after the bombing in 1988. If, as appears likely, the lost documentation no longer exists, this limitation on the ability to probe the facts is insurmountable and will equally apply to any further inquiry.
Third, the main reason advanced for holding another inquest is that the coroner could compel the retired police officers who declined to co-operate with the Ombudsman to attend to give evidence (provided they are still living and medically fit to do so). The potential value of their testimony must, however, be assessed having regard to the passage of time. By 2015, when the Attorney decided not to direct a fresh inquest, more than 26 years had elapsed since Mr Dalton was killed. Given that memory fades with time, the likelihood that, if a new inquest were held, witnesses who declined to co-operate with the Ombudsman would or could from their recollection add significantly to the evidence collected and reviewed in his investigation must, realistically, be regarded as remote.
Fourth, whilst the Ombudsman in his report criticised the police for failure to warn the local community of a terrorist threat and failing to carry out an adequate investigation of the murders, neither the Ombudsman nor the applicant has suggested that his findings warrant a criminal investigation of any police conduct. This is an important consideration. Although the identification, prosecution and punishment of criminal conduct is not the only aim of the procedural duty under article 2, it is undoubtedly a central purpose. The case for pursuing further investigative steps many years after a death has occurred is on any view substantially stronger if there is a prospect that the steps could lead to criminal proceedings. It cannot be said that there is any such prospect here.
Fifth, if the question then is asked “what purpose might a further investigation usefully serve?”, all that can, I think, fairly be said is that it could (i) subject the alleged shortcomings in the police handling of the incident to further public scrutiny and/or (ii) enable the victims’ families to obtain compensation if it can be established that any such shortcomings give rise to civil liability. The latter purpose can be fulfilled by the civil proceedings brought by Mr Dalton’s estate (see para 208 above). While it is clear that the availability of a civil claim cannot be a substitute for a criminal investigation, it may suffice to discharge an article 2 procedural duty where no, or no further, criminal investigation is required: see Calvelli and Ciglio v Italy (Application No 32967/96) Reports of Judgments and Decisions 2002-I, p 25, para 51; Mastromatteo v Italy (Application No 37703/97), para 90. As for shining further light on the conduct of the police, I agree with what Deeny J said in his judgment in the High Court at [2017] NIQB 33, para 38:
“If the death had occurred recently an inquest might disclose a systemic flaw in practice which contributed to the deaths. But if, as the Ombudsman found, there was a failure here, it took place 28 years ago at the height of the Troubles. It is very difficult to see how any practical benefit could now be obtained for the public in going over the procedures then being followed by police officers in Derry at that time, when they say that much of the city was out of bounds to them by terrorist activity.”
Sixth, given the very limited practical benefit that any further inquest or other public inquiry could reasonably be expected to achieve, Deeny J was also right to take into account, at para 46 of his judgment, the costs involved (both financial and human) of such an inquiry and “the choices which must be made in terms of priorities and resources”, as recognised in the Brecknell case (see para 221 above). This is plainly an important factor in the context where, as noted by Deeny J at para 37, there have in recent years in Northern Ireland been many calls to hold fresh inquests in historic cases. Such proceedings, when undertaken, are inevitably made more difficult, time-consuming and costly by the passage of time and, in allocating resources, it is necessary to decide on priorities. The Attorney is well placed to make those decisions and was entitled to conclude that in the present case it would be disproportionate to hold a new inquest.
In my view, these factors together lead inexorably to the conclusion that, when in 2015 the Attorney was asked to direct a fresh inquest into Mr Dalton’s death, there was no obligation under article 2 to do so or to take any other investigative steps. Deeny J was therefore right to hold that the Attorney’s decision was lawful and, for the reasons I have given, the Court of Appeal was wrong to take a contrary view. It follows that the Attorney’s appeal should be allowed and Ms Dalton’s application for judicial review of his decision dismissed.
Does the human rights act apply?
In light of this conclusion, the third issue raised on this appeal does not need to be decided. This is the issue whether, if there was an obligation under article 2 to take further investigative steps, failure to comply with this obligation would be actionable in the courts of the United Kingdom under the Human Rights Act or only by seeking redress from the European court. Nevertheless, in view of the arguments directed to this question, the disagreement among the members of this court about the correct legal analysis, and the important point of principle raised about whether it is appropriate to depart from two recent previous decisions of this court, I will give my opinion on these matters.
How the case law has developed
After the Human Rights Act came into force in 2000, the question soon arose of whether or to what extent the domestic obligation imposed on public authorities by the Act to comply with the procedural requirements of article 2 applies to the investigation of deaths which occurred before the Act came into force. In the 20 odd years since then, this question has been repeatedly litigated and the ground is now very well-trodden. Although I will try to do so succinctly, it is necessary to retrace the path taken in the key cases in order to understand and evaluate the Attorney’s argument that the Supreme Court should now depart from conclusions on this question which it has only recently reached.
In re McKerr
The question was first addressed by the UK’s highest court (then the appellate committee of the House of Lords) in In re McKerr [2004] UKHL 12, [2004] 1 WLR 807. The applicant in that case sought a declaration that the UK government was in breach of a procedural duty to carry out an investigation into his father’s killing by members of the Royal Ulster Constabulary in 1982 which complied with article 2. The House of Lords held that the procedural duty under the Act to investigate deaths does not apply to deaths which occurred before the Act came into force. The logic was straightforward. The object of the procedural duty is to make the substantive right to life protected by article 2 effective by requiring possible violations of that right to be properly investigated. It follows that the procedural duty arises only where there is reason to believe that there has been, or may have been, a violation of the substantive right. As Lord Bingham later put it in R (Gentle) v Prime Minister [2008] UKHL 20; [2008] AC 1356, para 6, the procedural duty is “parasitic on the existence of the substantive right, and cannot exist independently.” If, therefore, at the time when a death occurred there was no substantive article 2 right in domestic law which could have been violated, no procedural duty to investigate a possible violation of that right could arise under the Human Rights Act when the Act came into force.
Šilih v Slovenia
This logic was undone by the decision of the European court in Šilih v Slovenia (2009) 49 EHRR 37. In that case the court, sitting as a Grand Chamber, considered whether it had temporal jurisdiction to deal with a claim that Slovenia was in breach of a procedural duty to investigate a death which had occurred some 14 months before the Convention entered into force for Slovenia. The court decided that it did have such jurisdiction. A majority held that the procedural obligation under article 2 had “evolved into a separate and autonomous duty” and “can be considered to be a detachable obligation … capable of binding the state even when the death took place before the critical date” (para 159). Thus, in the court’s view, although the procedural obligation is triggered by “the acts concerning the substantive aspects” of article 2 (para 159), it is not necessary that those acts should have occurred after the entry into force of the Convention for the state concerned. A procedural duty may be imposed by article 2 even though the acts which triggered it could not amount to breaches of the substantive obligations in article 2 because the state concerned was not bound by the Convention when the acts took place and the Convention is not retrospective.
The court in Šilih was nevertheless concerned, in the interests of legal certainty, to set a limit on its jurisdiction as regards compliance with the procedural obligation in article 2 in respect of deaths that occurred before what it called the “critical date” (para 161). The judgment set out a test, albeit an obscure one, for defining such a limit. The main requirement stated was that there must exist a “genuine connection” between the death and the entry into force of the Convention for the state concerned (para 163). This was further glossed as meaning that “a significant proportion of the procedural steps” required by article 2 “have been or ought to have been carried out after the critical date”. A qualification was then added that “the Court would not exclude that in certain circumstances the connection could also be based on the need to ensure that the guarantees and the underlying values of the Convention are protected in a real and effective manner” (para 163).
The phrase “genuine connection” is potentially misleading, as it is naturally understood to denote a causal connection. Plainly, however, the court cannot have envisaged any causal connection between a death to be investigated and the entry into force of the Convention for a contracting state. The only kind of connection between the two events that could exist is proximity in time. The judgment did not, however, set out any requirement that the death must have occurred within any particular time before the “critical date”. Rather, the court focused on the extent to which the procedural steps required by article 2 have been or ought to have been carried out after the critical date. As well as being unclear, this criterion is problematic. It would be perverse if the fact that steps have been taken after the critical date to investigate a prior death were to have the effect of imposing an obligation to comply with article 2 which would have been avoided by failing to take such steps. Hence the alternative “or ought to have been carried out”. Yet this alternative is in danger of being circular. It seems hard to know what procedural steps the state ought to have carried out after the critical date without already knowing the extent to which the procedural obligation extends to prior deaths. Added to these difficulties, the court left it entirely unclear in what circumstances the “connection” might also be based on the need to protect “the underlying values of the Convention”.
In re McCaughey
It was against this background that in In re McCaughey [2011] UKSC 20, [2012] 1 AC 725 the Supreme Court revisited the question whether the domestic procedural duty to investigate deaths applies to deaths which occurred before the Human Rights Act came into force. The specific issue in McCaughey was whether an ongoing inquest into the deaths of two individuals shot and killed by British soldiers in Northern Ireland in October 1990 had to comply with the article 2 procedural duty. The Supreme Court (by a majority of 6-1) held that it did. The majority concluded that, in the light of the decision of the European court in Šilih, the reasoning of McKerr could no longer be supported. Given the holding in Šilih that the procedural duty is “autonomous” and “detachable”, it could no longer be said that the procedural duty is parasitical on the existence of the substantive right under article 2 so that, if the substantive right did not exist at the time of the death, a procedural duty could not exist either.
All the Justices struggled to make sense of the “genuine connection” test articulated in para 163 of the judgment in Šilih. Lord Hope of Craighead observed that only the most starry-eyed admirer of the European court could describe the guidance offered in that passage as clear (para 73). Lord Phillips of Worth Matravers described it as “difficult” and in part “totally Delphic” (paras 49, 53), Lord Brown of Eaton-under-Heywood as “deeply unsatisfactory” (para 100), and Lord Dyson as “extremely obscure” (para 130). Rather than attempt any wider exegesis, the majority thought it sufficient to decide that, in circumstances where an inquest or other investigation was already taking place into a death which had occurred before the Human Rights Act came into force, the investigation had to comply with the procedural requirements of article 2.
Janowiec v Russia
In Janowiec v Russia (2013) 58 EHRR 30, the Grand Chamber of the European court endeavoured to clarify the criteria adopted in Šilih. Three key points of clarification were provided. First, the court held that, for a “genuine connection” to be established, two criteria must both be satisfied: (i) the period of time between the death and the critical date must have been “reasonably short”; and (ii) a major part of the investigation must have been carried out, or ought to have been carried out, after the critical date (para 148). Second, the court held that the circumstances in which a procedural obligation will arise even though these criteria are not satisfied, in order to protect “the underlying values of the Convention”, are truly exceptional and limited to events that “amounted to the negation of the very foundations of the Convention”. The court said that this would be the case with “serious crimes under international law, such as war crimes, genocide or crimes against humanity, in accordance with the definitions given to them in the relevant international instruments” (para 150). Third, the court noted that, as had been recognised in Brecknell, an obligation to take investigative measures may be triggered by a new allegation, evidence or information coming to light. However, the court made it clear that where such new material relates to a death that occurred before the critical date, its discovery “may give rise to a fresh obligation to investigate only if either the ‘genuine connection’ test or the ‘Convention values’ test … has been met” (para 144).
Argument on the present appeal has focused on one sentence of the judgment in Janowiec which has, regrettably, still left ambiguous one limb of the so-called “genuine connection” test: the requirement that the period of time between the death and the critical date must have been “reasonably short”. On this point the European court offered, at para 146 of the judgment, the following further guidance: “Although there are no apparent legal criteria by which the absolute limit on the duration of that period may be defined, it should not exceed ten years”. The question to which I will return is whether this should be understood to mean that the relevant period should not normally exceed ten years or that it should never exceed ten years (or perhaps that the court deliberately left some ambiguity around this question).
Keyu
The Supreme Court first considered the significance of the judgment in Janowiec in R (Keyu) v Secretary of State for Foreign and Commonwealth Affairs [2015] UKSC 69, [2016] AC 1355. This case involved a challenge to a decision not to hold a public inquiry into the deaths of civilians killed by British soldiers in Malaya in 1948. A majority of the Supreme Court held that, for the purpose of the “genuine connection” test as explained in Janowiec, the “critical date” was not the date when the Convention came into force for the United Kingdom or was then extended to Malaya in 1953 but was the date in 1966 when the United Kingdom recognised the right of individual citizens to petition the European court.
In reaching this conclusion, Lord Neuberger of Abbotsbury, who gave the lead judgment on this issue, pointed out that, although in Janowiec the European court had referred to “the date of the entry into force of the Convention” for the state concerned, that statement was made in a case where the state concerned, Russia, had recognised the right of individual petition on the same date as it acceded to the Convention. Lord Neuberger cited other decisions of the European court, including Šilih at para 140, which indicate that, where the two dates differ, the critical date is that on which the right of individual petition was recognised. As the deaths in Keyu had occurred considerably more than ten years before the United Kingdom accepted the right of individual petition to the European court, the claim based on article 2 necessarily failed. The European court itself subsequently endorsed that reasoning and conclusion in Chong v United Kingdom (2018) 68 EHRR SE2.
In re Finucane
I come next to a decision of the Supreme Court which is important not only in the development of the law in this difficult area but also because of its subject matter: In re Finucane [2019] UKSC 7, [2019] NI 292. Patrick Finucane was a solicitor who on 12 February 1989 was murdered in his home in front of his wife and children by loyalist paramilitaries. Over the following years evidence accumulated that members of the Royal Ulster Constabulary and of an army unit operating under cover had colluded in his murder. In addition, independent investigations carried out by a senior British police officer, Sir John Stevens, were (as Sir John found) deliberately obstructed.
A retired Justice of the Supreme Court of Canada was asked to consider the matter by the Irish and United Kingdom governments and recommended a public inquiry. Between 2004 and 2008 numerous statements were made by UK government ministers, including the Prime Minister, that such an inquiry would be held. However, the government then decided instead to order a more limited, paper-based review (the de Silva review). Mr Finucane’s widow applied for judicial review of that decision. The application succeeded before Stephens J (as he then was) [2015] NIQB 57 in the High Court but the Court of Appeal reversed his decision on the question of relief [2017] NICA 7.
On the further appeal to the Supreme Court, this court unanimously held that the relevant statements made by government ministers amounted, individually and collectively, to an unequivocal undertaking to hold a public inquiry into Mr Finucane’s death (para 68) from which the government had resiled; that this undertaking did not, however, give rise to a legally enforceable legitimate expectation (para 81); but that there was a procedural obligation on the UK authorities under the Human Rights Act to carry out an investigation into the death of Mr Finucane that complied with article 2 of the Convention and such an investigation had not taken place.
Lord Kerr gave the lead judgment with which all the other Justices (Lady Hale, Lord Carnwath, Lord Hodge and Lady Black) agreed. Having reviewed all the cases to which I have referred above, including Janowiec, Lord Kerr held that a “genuine connection” had been established between the murder of Mr Finucane and the critical date (which he took to be the date of entry into force of the Human Rights Act). The period of time between these dates was 11 years and eight months. Lord Kerr interpreted the statement in Janowiec that the period of time between the death and the critical date should not exceed ten years as meaning that it should not normally exceed ten years (para 107). He considered that a period of ten years or less was not an “immutable” or “inflexible” requirement and relied on McCaughey as authority for the proposition that the test was “multi-factorial” in the sense that “the consideration that most of the investigation took place after the critical date could compensate for the length of the time lapse” (paras 108-109). Lord Kerr also rejected an argument made on behalf of the Secretary of State that the criteria established by the European court in Šilih and Janowiec were not applicable to UK domestic law on the ground that that question had already been decided in McCaughey (paras 110-111).
After analysing the steps taken to investigate Mr Finucane’s death, including the de Silva review, Lord Kerr found that a domestic article 2 procedural duty had been triggered by new credible evidence that emerged after the Human Rights Act came into force in 2000 and that the constraints on the de Silva review meant that that review was incapable of satisfying that duty (para 134). In the result, the Supreme Court made a declaration that there had not been an article 2 compliant inquiry into the death of Patrick Finucane.
The subsequent Finucane proceedings
It is an unhappy sequel to this decision that, since it was given, little progress seems to have been made towards complying with the state’s procedural duty under article 2 to carry out an effective investigation into the death of Patrick Finucane, and further proceedings have ensued. On 30 November 2020 the UK government again decided not to establish a public inquiry into Mr Finucane’s death, this time to await the outcome of certain further investigations. Mrs Finucane again applied for judicial review of this decision. For reasons given in a judgment delivered on 21 December 2022, the High Court (Scoffield J) granted the application and quashed the decision not to establish a public inquiry on the grounds of error of law and breach of article 2 (in failing to comply with the requirement of reasonable expedition): see In re Finucane (No 2) [2022] NIKB 37. Scoffield J rejected arguments made by the Secretary of State as “an inappropriate attempt to re-litigate or circumvent the clear findings of the Supreme Court, which formed part of the reasoning leading to the declaration it made” (para 100). He ended his judgment by stating, at para 127, that:
“… in light of the significant delay there has been from the time of Mr Finucane’s death until now - and perhaps more importantly in the present context from … the grant of the Supreme Court’s declaration - the applicant is entitled to expect a clear indication of how, if at all, the Secretary of State now proposes to proceed. … If the Secretary of State … is … not prepared to establish a public inquiry or some other mechanism in order to remedy the article 2 deficiencies identified by the Supreme Court, notwithstanding that that will result in continuing breach of article 2, he should state that clearly and give his reasons.”
The Supreme Court’s decision in Finucane therefore remains directly relevant, not only for other cases which remain unresolved relating to deaths that occurred a decade or more before the Human Rights Act came into force, but also to ongoing attempts to require the Secretary of State to establish an effective inquiry into Mr Finucane’s death itself.
In re McQuillan
The most recent decision at the highest level is In re McQuillan, In re McGuigan, In re McKenna [2021] UKSC 55, [2022] AC 1063, another unanimous decision of the Supreme Court (this time sitting as a panel of seven Justices). McQuillan concerned the death in Belfast in 1972 of a woman who was killed by a bullet whilst sitting as a passenger in a stationary car. Investigations at the time were inconclusive but in 2014 military logs were discovered which recorded that a covert army unit operating in the area had opened fire at the relevant time. On the appeal to the Supreme Court it was common ground that, applying the Brecknell test, this new evidence was sufficient to warrant a further investigation but there was a dispute about whether the death in 1972 came within the temporal scope of the Human Rights Act. A similar issue (in relation to the equivalent procedural duty under article 3 of the Convention) was raised by the claims of McGuigan and McKenna seeking a further investigation into the allegation that in 1971 the UK Government had authorised and used torture in what has become known as the case of the Hooded Men. On that appeal, however, it was held that there was no new evidence which would warrant a fresh investigation.
The argument based on the “mirror principle”
The claimants’ principal argument on those appeals that the cases came within the temporal scope of the Human Rights Act was based on what is sometimes called the “mirror principle”. This principle is that the scope of the Convention rights given effect in the domestic law of the United Kingdom by the Human Rights Act should mirror the scope of the Convention rights enforceable against the United Kingdom in international law, of which the final arbiter is the European Court of Human Rights. As Lord Reed explained in R (Elan-Cane) v Secretary of State for the Home Department [2021] UKSC 56, [2023] AC 559, para 87 (a judgment which, coincidentally, was handed down on the same day as the judgment in McQuillan):
“The Act … defines the Convention rights to which it gives effect in domestic law as the rights which are enforceable against the United Kingdom under international law. It follows that the rights given effect in domestic law have the same content as those which are given effect under international law, although they are enforceable before domestic courts rather than the European court, and against public authorities rather than the United Kingdom as a state. Since the rights have the same content at the domestic level as at the international level, it follows that the relevant articles of the Convention should in principle receive the same interpretation in both contexts. That is not to say that domestic courts are bound to follow every decision of the European court, but there should in principle be an alignment between interpretation at the international and domestic levels.”
The starting point for the argument made in McQuillan was that article 2 of the Convention has been interpreted by the European court in Šilih and Janowiec as including a procedural duty to investigate deaths which occurred before the “critical date” where either the “genuine connection” test or the “Convention values” test is met. The “critical date” was identified by the European court (see para 244 above) as the date when the Convention entered into force or, if later, when the right of individual petition was recognised by the state concerned. It follows - so it was argued - that the right under article 2 to an investigation of a death given effect in domestic law by the Human Rights Act is a right applicable to deaths which occurred within this time frame. The United Kingdom recognised the right of individual petition with effect from 14 January 1966. Where, therefore, material new evidence comes to light which satisfies the Brecknell test, provided the death in question occurred after 14 January 1966 there is no need to satisfy the “genuine connection” or “Convention values” test. Before courts in the United Kingdom as well as in proceedings before the European court, those tests are relevant only in relation to deaths that occurred before that date.
The Supreme Court in McQuillan did not accept this argument. The court held that the “mirror principle” does not apply to the question whether a claim that a public authority is in breach of a procedural duty to investigate a death falls within the temporal scope of the Human Rights Act. The essential reason why the mirror principle does not apply is that the answer to the question depends, not on the content of the obligation/ right, but upon the extent to which the Act is retrospective. This is purely a matter of UK domestic law.
Although one element of the reasoning in Šilih and Janowiec is that the procedural duty is separate and detachable from the substantive aspects of article 2, it is clear from the judgments in those cases that the question with which the European court was concerned was whether it had temporal jurisdiction - or “competence ratione temporis”- to deal with the merits of the applicants’ complaints. The justification given for the “genuine connection” test was that, “having regard to the principle of legal certainty, the court’s temporal jurisdiction as regards compliance with the procedural obligation of article 2 in respect of deaths that occur before the critical date” should not be “open-ended” and that limits of this jurisdiction need to be defined: see Šilih, para 161, and Janowiec, para 133. The identification of the “critical date” as the date when the state concerned recognised the right of individual petition, where this was later than the date of the entry into force of the Convention, is consistent with this. It reflects the fact that the European court is concerned with the right to bring a claim rather than with the existence or otherwise of an obligation.
Properly understood, therefore, the “genuine connection” test and “Convention values” test propounded by the European court are not part of the content of the procedural duty under article 2 of the Convention. Rather, they are tests which that court applies to decide whether it has jurisdiction to examine a complaint. As such, the “mirror principle” has no application. There is nothing in Šilih or Janowiec or any other case law of the European court which determines or bears directly on whether the facts of any complaint fall within the temporal scope of the Human Rights Act.
Nevertheless, the majority of the Supreme Court in McCaughey and all the members of the court in Finucane treated the “genuine connection” test as applicable to the latter question. In McQuillan this was explained on the basis that the test should be applied by analogy with the approach taken by the European court in defining the scope of its own temporal jurisdiction: see paras 154, 163 and 167. The same need for legal certainty which requires limits to be placed on the jurisdiction of the European court to examine claims asserting a procedural duty to investigate deaths that occurred long before any enforceable procedural duty existed also applies to claims made in courts of the United Kingdom under the Act. In McQuillan the Supreme Court endorsed the approach taken in Finucane, which borrowed the “genuine connection” test from the criteria developed by the European court to decide whether complaints fall within its jurisdiction but adapted it for the purpose of UK domestic law by substituting as the “critical date” 1 October 2000 when the Human Rights Act came into force.
Guidance on the time factor
The question also arose in McQuillan, as it had in Finucane, of what was meant by the statements in Janowiec, in seeking to elucidate the “genuine connection” test, that the period between the death and the critical date “must remain reasonably short” and “should not exceed ten years”. The Secretary of State for Northern Ireland and the other public authorities represented on the appeals in McQuillan (and McGuigan and McKenna) did not submit that a period of ten years was intended to be laid down by the European court as an absolute limit. Their position was that the court had held that the relevant period should not normally exceed ten years and that the decision in Finucane, where the period in question was 11 years and eight months, illustrates the outer limits of the temporal connection requirement.
The Supreme Court accepted that submission. On any view the period of some 28 years which had elapsed in McQuillan between the death and the entry into force of the Human Rights Act could not be described as “reasonably short”. (The same was true in McGuigan and McKenna, where the period was 29 years.) But the Supreme Court went further and, at para 144 of the judgment, expressed the opinion that it would significantly undermine the legal certainty which the European court sought to achieve in Janowiec if the period between the death and the entry into force of the Act were stretched any further than the period of nearly 12 years allowed in Finucane. The following guidance was given:
“In our judgment, an extension beyond the normal ten year limit of up to two years is permissible where there are compelling reasons to allow such an adjustment constituted by circumstances that (a) any original investigation into the triggering death can be seen to have been seriously deficient and (b) the bulk of such investigative effort which has taken place post-dates the relevant critical date. If in these circumstances there is an extension of no more than two years beyond the ten-year limit mentioned in Janowiec, it remains possible to describe the lapse of time as ‘reasonably short’ in accordance with the guidance in that judgment at paras 146 and 148.”
I agree with what Lord Burrows and Dame Siobhan Keegan say about this guidance at paras 332-337 of their judgment. In particular, I agree that what underpinned this guidance was the need to articulate in terms that can be applied in other cases the circumstances which justified going back beyond the normal ten-year limit in Finucane. It is important to appreciate what Stephens J in his judgment at first instance in Finucane described as the “larger dimension” of that case. That “larger dimension” was the adoption by members of the police and the Army of a regime of “murder by proxy” whereby loyalist terrorists were used as proxies to murder suspected republican terrorists. Mr Finucane was not connected with terrorism but was allegedly targeted because, as a solicitor, he often acted against the police and government and defended republican suspects in criminal cases. It was this shocking attack on the rule of law by those whose duty was to defend it which led Stephens J to characterise the murder and the state’s complicity in it as negating the very foundations of the Convention and of a democratic society such that the “Convention values” test was met: see Finucane’s (Geraldine) Application [2015] NIQB 57, para 35.
The Court of Appeal did not regard that finding as “necessarily unreasonable”: see [2017] NICA 7, para 167. The Supreme Court saw no need to address this question after concluding that the “genuine connection” test was satisfied: see [2019] UKSC 7, para 113. Given the apparent restriction of the “Convention values” test in Janowiec to serious crimes under international law, such as war crimes, genocide and crimes against humanity, I would not feel able to say that the facts of Finucane came within this category. But in my view Stephens J identified with complete clarity the circumstances which made the evidence of state involvement in Mr Finucane’s murder a matter of such gravity. He also described in detail in his judgment how the original police investigation lacked independence and should have resulted, but did not result, in the identification and prosecution of Mr Finucane’s killers; how the inquest failed to examine the allegations of collusion; and how the later investigation by Sir John Stevens met with “active and significant obstruction” by both the Army and the police. In Finucane v United Kingdom (2003) 37 EHRR 29 the European court found that there had been a violation of the procedural obligation in article 2.
These were the matters which, in the Finucane case, made the original investigation seriously deficient and created a compelling need, recognised by the unequivocal undertaking given by government ministers to hold a public inquiry, to carry out a further investigation. Furthermore, the failure of the original investigation properly to investigate the very serious allegations of state complicity in the murder had the result that the bulk of the investigative effort into those allegations had taken place (or remained to take place) after the Act had come into force.
The lapse of time in this case
In this case Mr Dalton’s death on 31 August 1988 occurred a little over 12 years before the entry into force of the Human Rights Act on 2 October 2000. Following the guidance given in McQuillan, there are two reasons why his death is outside the temporal scope of the Act. The first is that it occurred more than ten years before the Act came into force and there are no compelling reasons which could justify any extension beyond this normal limit. In particular, the allegations of police responsibility which are said to require further investigation do not involve any suggestion of collusion with the perpetrators of Mr Dalton’s unlawful killing, let alone the “larger dimension” present in Finucane of state complicity in the targeted assassination of a solicitor seeking to defend legal rights. Nor is it suggested that the initial police investigation was vitiated by lack of independence or obstruction. Judged by these measures, it cannot be said that the failure of the original investigation into Mr Dalton’s death to examine what the police knew before the bomb exploded renders it seriously deficient. By the same token, it cannot be said that that the bulk of the relevant investigative effort post-dates the Act’s entry into force.
Second, the lapse of time in this case between the death and the entry into force of the Act in any event exceeds the period of 12 years which this court went out of its way in McQuillan to identify as the outer limit. This is a second and additional reason why the death is outside the temporal scope of the Act. Although that outer limit is only just exceeded, it is in the nature of a limit that it will result in some cases falling foul of it by only a narrow margin. I agree with the submission made in the Attorney’s written case on this appeal (at para 77) that, although “this case falls only a month or so outside the maximum 12 year outer limit, re-amending the guidance [given in McQuillan] would undermine the legal certainty which both the Grand Chamber in Janowiec and this court in McQuillan sought to achieve.”
The conclusion that the death in this case falls outside the temporal scope of the Act is a further reason why the appeal must be allowed.
The invitation to re-write McQuillan and overrule Finucane
That should be the end of the matter. Yet although, as noted, the Attorney in her written case for this appeal submitted that re-amending the guidance given in McQuillan would undermine the legal certainty which this court sought to achieve in that case, in oral argument leading counsel for the Attorney, Mr Tony McGleenan KC, invited the court to do exactly that. He has invited the court to say that the law is the precise opposite of what it was said to be in McQuillan in the passage quoted at para 260 above. If this invitation were to be accepted, the guidance which this court gave in McQuillan would now be reversed so as to say this:
“In our judgment, an extension beyond the normal ten-year limit of up to two years is not permissible even where there are compelling reasons to allow such an adjustment … If … there is an extension of no more than two years beyond the ten-year limit mentioned in Janowiec, it does not remains possible to describe the lapse of time as ‘reasonably short’ in accordance with the guidance in that judgment at paras 146 and 148.”
The further consequence of such an about-turn, from which Mr McGleenan KC did not balk, would be that Finucane was wrongly decided and that this court ought to have held that Mr Finucane’s death fell outside the temporal scope of the Human Rights Act. The court should therefore have declined to grant a declaration that there had not been an investigation into the death of Mr Finucane complying with article 2 of the Convention and instead have left his family to seek redress by petitioning the European court in Strasbourg.
To justify this radical course, the Attorney does not rely on anything that has happened since Finucane (or McQuillan) was decided. It is not suggested, for example, that there has been any material further development in the case law of the European court which this court should follow. Nor is it suggested that there was any relevant matter not drawn to the attention of this court in Finucane, such that it might be said that the decision in that case was reached in ignorance of a point which, had it been known to the court, would have led to a different outcome. Exactly the same relevant legal materials are before the court on this appeal as were considered in Finucane, including the judgment of the European court in Janowiec. What is submitted is that the Justices who decided Finucane misread that judgment. They erroneously understood it to say that the period of time between the death and the critical date should not normally exceed ten years, when in fact it laid down a “bright line” rule that the period should never exceed ten years, except where the “Convention values” test is met.
Although three members of the court have been persuaded that the panel which decided Finucane did misunderstand the judgment in Janowiec, I disagree for the following reasons.
The argument that Janowiec decided on a strict 10-year time limit
If it were indeed clear, as the Attorney submits, that the European court in Janowiec laid down a “bright line” rule that ten years between the death and the critical date is an absolute upper limit, then it would be disappointing - to put it no higher - that the very experienced Treasury counsel who represented the Secretary of State in Finucane and all five Justices who heard the appeal, not to mention Stephens J and all three members of the Court of Appeal of Northern Ireland in the courts below, all failed to perceive this. The relevant part of the judgment in Janowiec comprises only a few paragraphs. The key passages were quoted and discussed in the judgments given in the Finucane case at all three levels. The question whether the judgment in Janowiec does or does not say that ten years should be an absolute limit is not a complicated one. All nine judges who considered the judgment concluded that it does not. The suggestion that they all misunderstood the judgment is far-fetched.
Here is the key passage, at para 146 of the judgment in Janowiec:
“The Court considers that the time factor is the first and most crucial indicator of the ‘genuine’ nature of the connection. It notes … that the lapse of time between the triggering event and the critical date must remain reasonably short if it is to comply with the ‘genuine connection’ standard. Although there are no apparent legal criteria by which the absolute limit on the duration of that period may be defined, it should not exceed ten years. Even if, in exceptional circumstances, it may be justified to extend the time-limit further into the past, it should be done on condition that the requirements of the ‘Convention values’ test have been met.” (emphasis added)
The interpretation of this passage contended for by the Secretary of State in the High Court, Court of Appeal and (at least in his written case) in the Supreme Court in Finucane, and again by the Attorney on this appeal, is that, although in the sentence highlighted the court acknowledged that there are “no apparent legal criteria” by which an absolute limit on the duration of the relevant period may be defined, the court nevertheless proceeded to specify such an absolute limit, of ten years. The only exception allowed is if the requirements of the “Convention values” test have been met. In favour of this interpretation is that it is how Judge Wojtyczek, in his partly dissenting opinion at para 0-III 8, evidently understood the majority judgment. (The partly dissenting opinion of Judges Ziemele, De Gaetano, Laffranque and Keller, at para 0-IV 20, on which Mr McGleenan KC also relied seems to me equivocal, as it essentially just repeats what the majority judgment says in the passage quoted above.) Mr McGleenan KC also prayed in aid the summary of what Janowiec decided in the “Practical Guide on Admissibility Criteria” issued by Registry of the Strasbourg Court (as updated 31 August 2022, now reissued 28 February 2023). However, this guide has no legal force.
The other way of reading the passage quoted above is as making these two points. First, in order to satisfy the “genuine connection” test, the lapse of time between the death as the triggering event and the critical date must be reasonably short: if that requirement is not satisfied, it may only be justified to extend the time limit further into the past if the requirements of the “Convention values” test have been met. Second, although it is not possible to define an absolute limit on the duration of the period which will count as “reasonably short”, that period should not (as a general rule) exceed ten years.
The following four points support this less rigid interpretation.
The first is that the footnote to the sentence (highlighted in the above quotation) stating that the period “should not exceed ten years” says this:
“See, by analogy, Varnava [(Application Nos 16064/90 and others) (unreported) 18 September 2009, para 166], and Er v Turkey (2012) 56 EHRR 13, paras 59-60.”
Article 35.1 of the Convention established a six-month period for making an application to the court after all domestic remedies have been exhausted (since 1 August 2021, four months). The two cases cited in this footnote were concerned with identifying when this period starts to run in a situation where, following the disappearance of a relative of the applicant, there has been a sporadic investigation of the disappearance by the state authorities. In Varnava, para 166, which was followed in Er v Turkey, the European court said that, in such a situation:
“Where more than ten years have elapsed, the applicants would generally have to show convincingly that there was some ongoing, and concrete, advance being achieved to justify further delay in coming to Strasbourg.”
It is clear that this period of ten years, which was relied on as an analogy in Janowiec, is a general rule only and not an absolute limit.
Second, at para 138 of the Janowiec judgment, in summarising its recent case law, the court refers to its decision in Mladenović v Serbia (Application No 1099/08) (unreported) 22 May 2012 as follows:
“Nor was the 13-year period separating the death of the applicant’s son in a brawl and the entry into force of the Convention in respect of Serbia seen as outweighing the importance of the procedural acts that were accomplished after the critical date.”
Nothing is said anywhere in the judgment to suggest that the decision in the Mladenović case was inconsistent with the clarification of the “genuine connection” test given in the judgment.
A third relevant consideration is that to stipulate a period of time that is intended to operate as an absolute and inflexible time limit would be a very unusual thing for a court to do. Courts are not legislators and are not normally regarded as having authority when deciding a case to enact a rule, such as a precise time limit, which must be applied in all future cases irrespective of the particular circumstances of the case. Indeed, I cannot call to mind any example of a case in any jurisdiction where a court has sought to do this. If this was the intention in Janowiec, one would therefore expect to see it very clearly expressed.
The fourth point, and in my view the most telling one of all, is that the European court did not include a time limit of ten years as part of its holding on what the “genuine connection” test requires. The conclusion stated at para 148 of the judgment is as follows:
“Having regard to the above, the Court finds that, for a ‘genuine connection’ to be established, both criteria must be satisfied: the period of time between the death as the triggering event and the entry into force of the Convention must have been reasonably short, and a major part of the investigation must have been carried out, or ought to have been carried out, after the entry into force.”
This paragraph is evidently intended to encapsulate the court’s decision on the requirements of the “genuine connection” test, and it is notable that it stipulates only that the period of time “must have been reasonably short” and makes no mention of any ten-year time limit. It is hard to suppose in these circumstances that the court intended to lay down an absolute rule that the period should not exceed ten years.
Taken together, it seems to me that these points present a strong case for reading the judgment as saying that, although there is no absolute rule or limit on the duration of the period that can be regarded as “reasonably short”, it should not normally exceed ten years.
This is also consistent with how the court itself, again sitting as a Grand Chamber, summarised the effect of Janowiec in its later judgment in Mocanu v Romania (2014) 60 EHRR 19, para 206. In explaining the “genuine connection” test, the court said that in Janowiec it had held that:
“such a connection was primarily defined by the temporal proximity between the triggering event and the critical date, which could be separated only by a reasonably short lapse of time that should not normally exceed ten years; at the same time, the Court specified that this time period was not in itself decisive.” (emphasis added)
Mr McGleenan KC sought to explain the word “normally” in this passage as alluding to the possibility, in exceptional circumstances, of applying the “Convention values” test. But the words “that should not normally exceed ten years” are part of the description of “a reasonably short lapse of time”. If the intention had been to refer obliquely to the “Convention values” test - even though it is nowhere in the judgment mentioned expressly - rather than indicate that ten years is not an absolute limit, then the judgment would have said: “… which normally could be separated only by a reasonably short lapse of time that should not normally exceed ten years.” The natural and grammatical reading of the Grand Chamber’s judgment in Mocanu v Romania is as confirming that a “reasonably short lapse of time” is one that “should not normally exceed ten years”. Similarly, in Mučibabić v Serbia (2016) 65 EHRR 35, para 97(b)(ii), the court summarised the test as being that the lapse of time “must have been reasonably short (in principle, not exceeding ten years)”.
Accordingly, it is very far from clear that the judgment of the European court in Janowiec should be taken to have prescribed ten years as an absolute time limit. Indeed, the better interpretation, in my view, is that it should not.
It is therefore unsurprising that on the appeal to this court in Finucane leading counsel for the Secretary of State, Sir James Eadie QC, in addressing the effect of Janowiec, put his case in a more nuanced way than Mr McGleenan KC for the Attorney has done on this appeal. As quoted by Lord Hodge, Lord Sales and Lady Rose at para 165 of their judgment, in his oral submissions Sir James described the ten-year period as a “pretty serious line”, but one which was “not entirely absolute” and “not entirely unporous”. I think that was a fair submission. It is equally unsurprising that the Justices who heard the appeal in Finucane, in common with Stephens J and the Court of Appeal, did not read the judgment in Janowiec as laying down a bright line rule that (subject only to the “Convention values” test) ten years is an absolute time limit.
Thus, while I accept that the point is one on which there are two possible views, if this had been the first occasion on which the Supreme Court had been asked to consider how para 146 of the judgment in Janowiec should be read, I would have reached the same conclusion as all nine judges, including all five members of this court, did in Finucane: that is, I would have concluded that what was meant is that, if it is to comply with the “genuine connection” test, the lapse of time between the death and the critical date must be reasonably short and should not normally exceed ten years - albeit that a limit of ten years is not an immutable or inflexible requirement.
Adherence to precedent
My view has only to be stated, however, for it to be evident that it is beside the point. What view any member of the court sitting on this appeal would have taken if the question were being raised for the first time is irrelevant; for the question is not being raised for the first time. It has already been decided by the Supreme Court. A principal reason for having a final court of appeal which hears second appeals raising arguable questions of law of general public importance is to enable such questions to be finally settled, at least until circumstances materially change. It would make a mockery of this system if the Justices hearing an appeal felt able to overrule what the court has previously decided just because they take a different view on a question from that taken by the Justices who decided the earlier case. If such an approach were to be sanctioned, there would be nothing to prevent litigants from rearguing any question of law on any appeal. Because the United Kingdom Supreme Court, unlike some courts of final appeal, does not sit as a full court but usually in a panel of five and very rarely more than seven of the 12 Justices, it is all the more vital that any particular constitution of the court should respect earlier decisions on points on which different reasonable views are possible, especially when the decision in question is recent and unanimous.
In Fitzleet Estates Ltd v Cherry [1977] 1 WLR 1345, 1349, Lord Wilberforce said:
“Nothing could be more undesirable … than to permit litigants, after a decision has been given by this House with all appearance of finality, to return to this House in the hope that a differently constituted committee might be persuaded to take the view, which its predecessors rejected.”
Lord Wilberforce returned to this theme in Vestey v Inland Revenue Comrs [1980] AC 1148, 1176, stating that:
“… this House ought not to sanction attempts to obtain reversals of decisions deliberately reached however attractive to their successors another view may appear to be.”
See also R v Knuller (Publishing, Printing and Promotions) Ltd [1973] AC 435, 455 (Lord Reid); Paal Wilson & Co A/S v Partenreederei Hannah Blumenthal [1983] 1 AC 854, 917 (Lord Diplock), 922 (Lord Roskill); R v Kansal (No 2) [2001] UKHL 62, [2002] 2 AC 69, para 10 (Lord Slynn), para 21 (Lord Lloyd) and para 27 (Lord Steyn); Rees v Darlington Memorial Hospital NHS Trust [2003] UKHL 52, [2004] 1 AC 309, para 32 (Lord Steyn); Horton v Sadler [2006] UKHL 27, [2007] 1 AC 307, para 29 (Lord Bingham).
These strictures apply with equal importance to the Supreme Court as they did to the appellate committee of the House of Lords. If a panel of Justices was willing to overrule a decision reached by their predecessors for no better reason than that they (or a majority of them) were persuaded by a differently presented argument on a point previously decided, the stability of the legal system would be undermined and the legitimacy of the court imperilled.
Accordingly, the fact that some members of the court on this appeal take a different view on a point that was critical to the decision reached in Finucane is not a valid reason to overrule that decision. That is so in principle, even before account is taken of the disruptive consequences that overruling that decision would have in practice, not only for legal certainty regarding the temporal scope of the Human Rights Act, but for the fair disposition of litigation still pending in Northern Ireland.
Regarding that last matter, it is said that at least overruling Finucane would not affect the rights of the Finucane family, as those rights are res judicata. I am not so sanguine. The Finucane case decided that there had not, at the time of the Supreme Court’s decision in 2019, been an investigation into Patrick Finucane’s death which complied with article 2. That point is clearly res judicata. But it is by no means clear that, if the decision in Finucane were now to be overruled, the Secretary of State would be unable to maintain that subsequent proceedings brought by Mr Finucane’s family seeking a public inquiry into his death are outside the temporal scope of the Human Rights Act. Moreover, even if the Secretary of State were precluded from so contending in relation to the Finucane family, proceedings brought by another family in a similar position would be treated as outside the temporal scope of the Act. Treating like cases differently would create injustice of the kind which the doctrine of precedent is designed to avoid.
It would be all the more inappropriate for the court to take this step when it is divided on the issue and when in any event the question whether Finucane was correctly decided does not affect the outcome of this appeal.
Revisiting McQuillan
What of the Attorney’s submission that the court should review the guidance which it gave in McQuillan (quoted at para 260 above) and now decide that this guidance should be disregarded?
It is said that the guidance was obiter, in that whether the relevant period was ten years or 12 years did not affect the outcome of the particular cases before the court. Precisely the same is true in this case, as Mr Dalton died over 12 years before the Human Rights Act came into force. Whether the maximum period is ten years or 12 years therefore makes no difference to the outcome of the present appeal. Obiter or not, the guidance which the court gave in McQuillan was given only recently, was carefully considered and plainly intended to be authoritative. Furthermore, it was unanimously agreed by a panel of seven Justices. Yet it is suggested that it should now be retracted and replaced with guidance to exactly the opposite effect. I do not see how it could be justifiable for the Supreme Court to play fast and loose with the law of the United Kingdom in this way.
The reason put forward is that the guidance given in McQuillan does not mirror what the European court said in Janowiec about the temporal limb of the “genuine connection” test. There is no inconsistency, however, between what was said in Janowiec, as it was interpreted by this court in Finucane, and the guidance given in McQuillan. It is only if the court were (obiter) to substitute a different interpretation of the judgment in Janowiec for the one on which the decision in Finucane was based that there would be a discrepancy. I have already explained why that would not be legitimate.
In any event, as discussed at paras 253 – 258 above, the court in McQuillan held that the “mirror principle” does not apply in this context and no one is suggesting that the court should revisit that conclusion. Whether there is a procedural duty enforceable in the courts of the United Kingdom to investigate the circumstances of a death that occurred more than ten years, or more than 12 years, before the Human Rights Act came into force does not depend on the content of a Convention right. It depends on the temporal scope of the Act. That is a matter purely of domestic law on which the European court has not expressed, and is not competent to express, any opinion. In identifying the temporal scope of the Act, the Supreme Court has thought it appropriate to adopt a test formulated for a different purpose by the European court by way of analogy. But it is only an analogy. The common principle is the principle of legal certainty. If, in identifying the temporal scope of the Act, this court thinks it right to define what period of time should count as “reasonably short” in slightly different terms from the European court, neither that principle nor any other legal principle is violated.
I would add that this is a yet further reason why there is no justification for overruling Finucane. Even if the decision of the Supreme Court in that case about the temporal scope of the Human Rights Act did not precisely align with the view of the European court about the temporal scope of its own jurisdiction, that would not demonstrate that Finucane was wrongly decided, as there was no imperative for it to do so. Nor, whatever the approach taken by the European court, would it be appropriate for this court to lay down an absolute time limit. I agree with what Lord Reed says in this regard at paras 43 – 44 of his judgment.
Conclusion
For these reasons I would allow the appeal and dismiss the application for judicial review. I would also respect the doctrine of precedent and stand by what the court has decided in Finucane and McQuillan.
LORD BURROWS AND DAME SIOBHAN KEEGAN:
Introduction
Sean Dalton was instantly killed in a bomb explosion at 38 Kildrum Gardens, Londonderry, on 31 August 1988. The explosion also resulted in the deaths of Sheila Lewis, who died on the same day, and Thomas Curran, who died from his injuries several months later. The incident has become known as the “Good Samaritan bombing”. This is because, when the bomb went off, Mr Dalton was entering a neighbour’s house in order to check on the welfare of his neighbour. The bomb was planted by the IRA with the apparent purpose of luring members of the security forces into a trap.
Shortly after the death, there was a police investigation, followed by an inquest on 7 December 1989. No-one was charged with the murders and the inquest simply recorded the facts of Mr Dalton’s death.
Nothing else of note happened until February 2005 when Mr Dalton’s son lodged a complaint with the Police Ombudsman of Northern Ireland (PONI). In broad terms, this complaint alleged police misconduct in failing to warn those in the neighbourhood of the bomb and alleged that this conduct was to protect a police informant.
The PONI investigated the claims made by the Dalton family and reported eight years later on 10 July 2013. The PONI report contained criticisms of the police after what was described as a “wide-ranging and thorough” investigation. These criticisms centred upon a failure to advise the local community of the bomb. The claim of collusion with a police informant was not upheld. The PONI also commented that a substantial number of retired police officers refused to co-operate and that documents were missing.
As a result of the PONI report, Mr Dalton’s family asked the Attorney General of Northern Ireland (AGNI) to direct a new inquest. The AGNI’s power to direct an inquest derives from section 14 of the Coroners Act (Northern Ireland) 1959 which provides:
“Where the Attorney General has reason to believe that a deceased person has died in circumstances which in his opinion make the holding of an inquest advisable, he may direct any coroner (whether or not he is the coroner for the district in which the death has occurred) to conduct an inquest into the death of that person …”
The AGNI ultimately declined the request for a new inquest for reasons contained in a letter of 27 March 2015. In that letter, the AGNI summarised his reasoning as follows:
The circumstances surrounding your application have already been the subject of a detailed examination by the PONI. The Attorney does not consider that your submission or the material provided [make] the holding of an inquest advisable.
The Attorney, having taken into account the case of Janowiec, remains of the opinion that Article 2 of the Convention does not require proceedings to be held for the purposes of establishing historical truth.
The Attorney has not been provided with any evidence which would suggest that the identification and/or punishment of those responsible could be achieved if a fresh inquest was directed.”
The present claim, now pursued by Rosaleen Dalton, the daughter of Mr Dalton, is for judicial review of the refusal of the AGNI to direct a fresh inquest. It is alleged that that decision infringes her rights under Article 2 of the European Convention of Human Rights (“ECHR”) and is therefore in breach of the Human Rights Act 1998 (“HRA”). This challenge failed at first instance before Deeny J [2017] NIQB 33. The Court of Appeal (Morgan LCJ, Stephens LJ, Maguire J) [2020] NICA 26; [2021] NI 405 reversed the decision made at first instance and declared that an Article 2 compliant investigation had not taken place. The Court of Appeal then effectively remitted the case to the AGNI for further consideration. The AGNI now appeals to the Supreme Court.
What has been set out above is a summary of the relevant facts and of the proceedings so far. For greater detail, we are grateful to rely on paras 57 – 83 and 96 – 114 of the joint judgment of Lord Hodge, Lord Sales and Lady Rose.
It is well-established that Article 2 of the ECHR imposes both a substantive and a procedural duty on the State and that, although the HRA came into force on 2 October 2000 and is non-retrospective, the procedural duty to investigate deaths can apply to deaths prior to the coming into force of the HRA. This is so where, first, new information has come to light which satisfies the test for reviving an investigation laid down in Brecknell v United Kingdom (2007) 46 EHRR 42 (“Brecknell”); and, secondly, there is a “genuine connection” between the triggering death and the “critical date” of 2 October 2000. There is an exception to the need for a genuine connection if the death undermines the values of the Convention (as, for example, with genocide) so as to satisfy the “convention values” test.
This case is primarily concerned with the “genuine connection” test (although that was not in issue before either the High Court or the Court of Appeal). That is Issue 1: whether the Court of Appeal erred in concluding that the genuine connection test was met on the facts of this case. Issues 2 and 3 are closely connected. Issue 2 is whether the Court of Appeal erred in finding that the Article 2 investigative obligation had been revived by the complaint to the PONI made in 2005 by Mr Dalton’s son which drew attention to various incidents that preceded his father’s death; and Issue 3 is whether the Court of Appeal erred in concluding that the Article 2 investigative obligation was not satisfied by the PONI report.
Although Issue 1 is the primary issue, it is convenient to look first at the Brecknell revival test and thereby to answer Issues 2 and 3.
The Brecknell revival test
Introduction to investigations required by Article 2, including the Northern Irish background
The procedural obligations of the State under Article 2 were defined by a series of cases from Northern Ireland starting with McCann v United Kingdom (1995) 21 EHRR 97. This case concerned the use of lethal force by State agents. In its decision the European court of Human Rights (the “Strasbourg Court”) reiterated the need for an effective official investigation. The purpose of such an investigation is to secure the effective implementation of the domestic laws safeguarding the right to life and to ensure accountability in circumstances which potentially engage the State’s responsibility for deaths (see Jordan v United Kingdom (2001) 37 EHRR 2).
What form the investigation takes may vary depending on the circumstances of a particular case. The proper standards of investigation, now well embedded in Convention jurisprudence, are independence, adequacy, promptness, reasonable expedition, and public scrutiny/participation of next of kin.
In Northern Ireland, such investigations have often taken place by way of inquest. For an inquest to be effective it must be capable of leading to a determination of whether the force used in such cases was justified in the circumstances and to the identification and punishment of offenders. Inquests have also dealt with the planning and control of State actions to answer the broad question of how the deceased came to die. This is “not an obligation of result but of means”: see Jordan v United Kingdom (2003) 37 EHRR 2 at para 107; Nachova v Bulgaria (2005) 42 EHRR 43; In re McCaughey [2011] UKSC 20, [2012] 1 AC 725; In re Jordan’s Application for Judicial Review [2014] NIQB 11; [2014] Inquest LR 1.
The law has also developed as the Strasbourg Court has accepted that the procedural obligation arises in a wide range of circumstances. This expansion in approach stems from the need to secure the right to life by ensuring that appropriate mechanisms are in place to deter the commission of offences or life-threatening behaviour. Circumstances where the obligation arises extend beyond the actions of State actors and include alleged negligence in healthcare and road traffic accident cases – see Lopes de Sousa Fernandes v Portugal (2017) 66 EHRR 28 and Todorova v Bulgaria (Application No 23302/03) (unreported) 24 May 2011. The underlying objective in these circumstances is accountability.
The subject matter surrounding a death will obviously dictate the form of investigation. For instance, in cases of lethal force committed in collusion with State agents, civil proceedings by the next of kin which do not involve identification and punishment of any alleged perpetrator, cannot fulfil the Article 2 obligation because in such cases damages are an inadequate remedy: see McKerr v United Kingdom (2001) 34 EHRR 20, Bazorkina v Russia (2006) 46 EHRR 15, Al-Skeini v United Kingdom (2011) 53 EHRR 18. On the flip side, in medical negligence or road traffic accident cases, where Article 2 is engaged, compliance may be achieved by the civil proceedings: see Mastromatteo v Italy (Application No 37703/97) (unreported) 24 October 2002 (Application No 37703/97). The choice of investigative method is firmly within the State’s margin of appreciation.
The Brecknell case
In the Brecknell case the Fourth Chamber of the Strasbourg Court considered whether an obligation to investigate which had gone dormant could revive with the receipt of new information. The death in question in Brecknell occurred in 1975. The new evidence emerged in 1999 regarding possible police collusion in the murder. This was a perpetrator case: that is, the new evidence might assist in identifying the perpetrator.
The Strasbourg Court reviewed the Article 2 requirements in such a circumstance. The court said that new information may revive the obligation and that “the issue then arises as to whether, and in what form, the procedural obligation is revived” (para 66, Brecknell). In addition, the Strasbourg Court referred to the fact that if “an investigation ends without concrete, or with only limited, results [it] is not indicative of any failings as such” (ibid).
Paras 70 -72 of the judgment in Brecknell refer to the scope of any revived obligation and the utility of further investigation in the following terms:
The Court would, however, draw attention to the following considerations. It cannot be the case that any assertion or allegation can trigger a fresh investigative obligation under Article 2 of the Convention. Nonetheless, given the fundamental importance of this provision, the state authorities must be sensitive to any information or material which has the potential either to undermine the conclusions of an earlier investigation or to allow an earlier inconclusive investigation to be pursued further. Both parties have suggested possible tests. The Court has doubts as to whether it is possible to formulate any detailed test which could usefully apply to the myriad of widely differing situations that might arise. It is also salutary to remember that the Convention provides for minimum standards, not for the best possible practice, it being open to the contracting parties to provide further protection or guarantees… Lastly, bearing in mind the difficulties involved in policing modern societies and the choices which must be made in terms of priorities and resources, positive obligations must be interpreted in a way which does not impose an impossible or disproportionate burden on the authorities.
With those considerations in mind, the Court takes the view that where there is a plausible, or credible, allegation, piece of evidence or item of information relevant to the identification, and eventual prosecution or punishment of the perpetrator of an unlawful killing, the authorities are under an obligation to take further investigative measures. The steps that it will be reasonable to take will vary considerably with the facts of the situation. The lapse of time will, inevitably, be an obstacle as regards, for example, the location of witnesses and the ability of witnesses to recall events reliably. Such an investigation may in some cases reasonably be restricted to verifying the credibility of the source, or of the purported new evidence. The Court would further underline that, in light of the primary purpose of any renewed investigative efforts, the authorities are entitled to take into account the prospects of success of any prosecution. The importance of the right under Article 2 does not justify the lodging, willy-nilly, of proceedings. As it has had occasion to hold previously, the police must discharge their duties in a manner which is compatible with the rights and freedoms of individuals and they cannot be criticised for attaching weight to the presumption of innocence or failing to use powers of arrest, search and seizure having regard to their reasonably held view that they lacked at relevant times the required standard of suspicion to use those powers or that any action taken would not in fact have produced concrete results.
The extent to which the requirements of effectiveness, independence, promptness and expedition, accessibility to the family and sufficient public scrutiny apply will again depend on the particular circumstances of the case, and may well be influenced by the passage of time as stated above…” (Emphasis added)
Our answer to Issue 2
The passages that we have just cited from Brecknell refer to evidence as to the perpetrator as that was the issue in Brecknell. However, given the breadth of the procedural obligation discussed above and developed by Strasbourg jurisprudence, we do not think that the obligation can be confined to perpetrator only cases. We therefore reject the principal submission on Issue 2 of Tony McGleenan KC, counsel for the AGNI, that the Brecknell revival test is confined to such perpetrator cases and that the facts of this case therefore automatically fell outside the Brecknell test as there was no realistic prospect of identifying the perpetrators. In our view, there may be good reasons for the investigative duty to be revived by fresh evidence even if identification of the perpetrators is not the primary purpose. On these facts, we therefore consider that, applying Brecknell, the investigative duty was revived by the complaint made in 2005.
That said, the question becomes – and this is Issue 3 – whether there has been any remaining investigative duty after the PONI report. Put another way, was the Article 2 obligation satisfied by the PONI report?
Our answer to Issue 3
In the present case it is known that the IRA claimed responsibility for the bomb and no further identification of the perpetrators can realistically be achieved. Any further investigation is directed against the police. This is in the context of an independent PONI report which reached conclusions in relation to police misconduct (albeit with some caveats, due to non-co-operation and missing documentation).
We question what more can truly be achieved in this case. Another inquest could go over the same ground as the PONI investigation but that is not the purpose of Article 2. Another inquest could try to secure the missing documentation. But PONI had eight years of searching. Another inquest could summon witnesses. However, whether witnesses would yield further answers after the passage of time is open to question.
When an overall dispassionate view is taken, the AGNI’s position that there would not be any utility in a further investigation is attractive. An investigation by PONI, whilst not a perfect process, is a recognised way in which the State can fulfil its obligations. A thorough report has been compiled after eight years’ work. In our view, this satisfies the Article 2 procedural obligation upon the State in circumstances where police misconduct is the only remaining issue.
Lest our view on utility is too narrow, there is another avenue which avails the family in this case. That is the civil proceedings which are at an advanced stage of readiness. These proceedings, whilst not at the instigation of the State, are directed at the remaining issue of police liability. In circumstances where prosecution or accountability of the perpetrator is no longer live, and where State collusion has been ruled out, we think that such civil claims can potentially meet the Article 2 obligation.
Maguire J, in the Court of Appeal, was alive to these arguments. At para 132, he referenced utility. At paras 134-139, he referenced the civil claims. Maguire J also cited Dumpe v Latvia (Application No 71506/13) (unreported) 16 October 2018, para 74, which refers to the State’s margin of appreciation in choosing the means by which the Article 2 obligation is satisfied:
“The court reiterates that the choice of means for ensuring the positive obligations under Article 2 of the Convention is in principle a matter that falls within the Contracting State’s margin of appreciation. There are different avenues for ensuring Convention rights and even if the State has failed to apply one particular measure provided by domestic law, it may still fulfil its positive duty by other means… the Court’s task, having regard to the proceedings as a whole, is to review whether and to what extent the domestic authorities submitted the case to the careful scrutiny required by Article 2 of the Convention.”
Maguire J ultimately found as follows:
The court should not be interpreted as saying that in this case the AGNI should not order a fresh inquest or should regard civil proceedings as the means of taking the matter forward.
All the court is saying [is] that the matter is not so open and shut in favour of the remedy which the applicant seeks as to cause the court to resort to an order of mandamus.”
In our view, if the Court of Appeal had assessed the AGNI’s decision for itself it could not on the facts of this case have found that decision unlawful or irrational applying the utility argument allied with the ongoing civil proceedings. We therefore conclude that the AGNI succeeds on Issue 3.
The genuine connection test (Issue 1)
Introduction
We now turn to the principal issue in this case which is the application of the “genuine connection” test. In our view, the “genuine connection” test is not satisfied in this case (so that the appeal should be allowed). We do not regard it as appropriate, in applying the “genuine connection” test, to depart from the decision in In re Finucane [2019] UKSC 7, [2019] NI 292 (“Finucane”), or the obiter dicta in In re McQuillan’s Application [2021] UKSC 55, [2022] AC 1063 (“McQuillan”), at para 144. The dictates of precedent, and the stability and certainty which it is designed to ensure, mean that caution should be exercised before deciding to apply the Practice Statement (Judicial Precedent) [1966] 1 WLR 1234. Similar considerations may apply where the question is whether to reject carefully considered obiter dicta of a panel of the Supreme Court. Particular caution is required where, as here, one is dealing with two cases that are very recent, where the second involved a seven-person panel, where the facts in question raise particularly sensitive issues, where there has been no subsequent legal development (eg a new case from the Strasbourg court) that casts doubt on those decisions, and where there is relevant ongoing litigation. We have also not been referred to any academic or other criticism of those two cases. It is therefore our view that this appeal should be allowed because, applying the rationalisation of the authorities (especially of the decision in Finucane) given by this court (sitting as a panel of seven) only a year ago in McQuillan, the “genuine connection” test is not satisfied. There is no justification or need to overrule Finucane.
Domestic law: the “by analogy” approach to the Strasbourg jurisprudence
One possible approach to the State’s Article 2 procedural duty to investigate deaths, given effect in domestic law by the HRA, would have been to say that that duty arose only in respect of deaths that occurred after the HRA came into force on 2 October 2000. That would have been to apply a rigid non-retrospectivity analysis (and was the approach taken by Lord Rodger in his dissent in In re McCaughey [2011] UKSC 20, [2012] 1 AC 725). But the Strasbourg Court has not itself applied a rigid non-retrospectivity approach in respect of Article 2. Instead, the jurisprudence of the Strasbourg Court has been that there is a free-standing (or autonomous) procedural duty to investigate deaths even though the relevant death occurred before the “accession/petition date” ie the date when citizens of a particular State have been given the right to bring cases to Strasbourg (in respect of the UK the relevant date was 14 January 1966). That means that a claimant, applying to the Strasbourg court, can go back to before the accession/petition date provided there was a “genuine connection” between the triggering death and the accession/petition date.
Although the House of Lords and Supreme Court were free to take a different approach to the coming into effect of the HRA, the majority (6-1) in Re McCaughey chose to apply the Strasbourg jurisprudence by analogy. This is not a direct application of the so-called “mirror principle” (the classic case on the “mirror principle” being R (Ullah) v Special Adjudicator [2004] UKHL 26, [2004] 2 AC 323) because it is not being suggested that, under the HRA, a claimant can go back to deaths prior to 14 January 1966. The critical date for the purposes of the HRA is 2 October 2000 not 14 January 1966. Therefore, there will be a gap (and hence no precise mirroring) between a person’s right to go to Strasbourg and “bringing home” the right under Article 2 by reason of the HRA.
The Strasbourg jurisprudence on the “genuine connection” test
What, then, is the Strasbourg jurisprudence on the “genuine connection” test that is to be applied by the domestic courts by analogy? There are two leading Strasbourg cases: Šilih v Slovenia (2009) 49 EHRR 37 and Janowiec v Russia (2013) 58 EHRR 30 (“Janowiec”). Of the two, Janowiec is the later and more important decision. Paras 142- 148 of Janowiec are crucial and will now be set out almost in full:
The Court reiterates at the outset that the procedural obligation to investigate under article 2 is not a procedure of redress in respect of an alleged violation of the right to life that may have occurred before the critical date. The alleged violation of the procedural obligation consists in the lack of an effective investigation; the procedural obligation has its own distinct scope of application and operates independently from the substantive limb of article 2. Accordingly, the Court’s temporal jurisdiction extends to those procedural acts and omissions which took place or ought to have taken place in the period after the entry into force of the Convention in respect of the respondent Government.
The Court further considers that the reference to ‘procedural acts’ must be understood in the sense inherent in the procedural obligation under article 2 or, as the case may be, article 3 of the Convention, namely acts undertaken in the framework of criminal, civil, administrative or disciplinary proceedings which are capable of leading to the identification and punishment of those responsible or to an award of compensation to the injured party. This definition operates to the exclusion of other types of inquiries that may be carried out for other purposes, such as establishing a historical truth.
The mention of ‘omissions’ refers to a situation where no investigation, or only insignificant procedural steps, have been carried out, but where it is alleged that an effective investigation ought to have taken place. Such an obligation on the part of the authorities to take investigative measures may be triggered when a plausible, credible allegation, piece of evidence or item of information comes to light which is relevant to the identification and eventual prosecution or punishment of those responsible. Should new material emerge in the post-entry into force period and should it be sufficiently weighty and compelling to warrant a new round of proceedings, the Court will have to satisfy itself that the respondent State has discharged its procedural obligation under article 2 in a manner compatible with the principles enunciated in its case law. However, if the triggering event lies outside the Court’s jurisdiction ratione temporis, the discovery of new material after the critical date may give rise to a fresh obligation to investigate only if either the ‘genuine connection’ test or the ‘Convention values’ test, discussed below, has been met.’
…
The Court considers that the time factor is the first and most crucial indicator of the ‘genuine’ nature of the connection. It notes …that the lapse of time between the triggering event and the critical date must remain reasonably short if it is to comply with the ‘genuine connection’ standard. Although there are no apparent legal criteria by which the absolute limit on the duration of that period may be defined, it should not exceed ten years. Even if, in exceptional circumstances, it may be justified to extend the time-limit further into the past, it should be done on condition that the requirements of the ‘Convention values’ test have been met.
The duration of the time period between the triggering event and the critical date is however not decisive, in itself, for determining whether the connection was a ‘genuine’ one. As the second sentence of para 163 of the Šilih judgment indicates, the connection will be established if much of the investigation into the death took place or ought to have taken place in the period following the entry into force of the Convention. This includes the conduct of proceedings for determining the cause of the death and holding those responsible to account, as well as the undertaking of a significant proportion of the procedural steps that were decisive for the course of the investigation. This is a corollary of the principle that the Court’s jurisdiction extends only to the procedural acts and omissions occurring after the entry into force. If, however, a major part of the proceedings or the most important procedural steps took place before the entry into force, this may irretrievably undermine the Court’s ability to make a global assessment of the effectiveness of the investigation from the standpoint of the procedural requirements of article 2 of the Convention.
Having regard to the above, the Court finds that, for a ‘genuine connection’ to be established, both criteria must be satisfied: the period of time between the death as the triggering event and the entry into force of the Convention must have been reasonably short, and a major part of the investigation must have been carried out, or ought to have been carried out, after the entry into force.”
It can be seen from this that Janowiec laid down the following in respect of the “genuine connection” test:
The time period between the relevant death, which is the “triggering event”, and the “critical date” (which as regards the ECHR in the UK is 14 January 1966 as we have explained above) is the most important aspect of the “genuine connection” test.
That time period should be “reasonably short”. Leaving aside where the “convention values” test applies, it was indicated that that time period should not exceed ten years (see para 146). But it is not entirely clear from the overall context (see the summary at para 148) whether that was an indication of what is meant by “reasonably short” or the laying down of a fixed rule of no more than ten years. Although the laying down of a fixed rule may be thought to be contrary to the normal approach of the Strasbourg court, which tends to favour flexibility and a “margin of appreciation” to individual States, we accept that it is possible to interpret what was said as laying down a fixed rule.
Apart from the time period, the other relevant factor to consider in relation to the “genuine connection” test is whether a major part of the investigation has been carried out or ought to have been carried out after the critical date.
The overall duty is to carry out an effective investigation.
The domestic law’s application by analogy of the Strasbourg jurisprudence: McQuillan and Finucane
In McQuillan the Supreme Court accepted that the correct approach in domestic law is to apply Janowiec by analogy where the relevant critical date, for the purposes of the HRA, is 2 October 2000. The Court then went on to set out its rationalisation of the authorities in para 144.
Para 144 in McQuillan reads as follows:
“With respect to Lord Kerr JSC [In re Finucane], he did not identify any clear principle by which one could tell when and to what extent it might be appropriate to water down a strict ten-year requirement as the Grand Chamber of the Strasbourg court had appeared to lay down in Janowiec, para 146. We have reservations as to whether Lord Kerr JSC was right to interpret Janowiec as he did. This court has not been invited to depart from its decision in In re Finucane but we note that the extension beyond ten years allowed in In re Finucane involved less than two more years. It would significantly undermine the legal certainty which the Grand Chamber sought to achieve in Janowiec if longer extensions than this were to be contemplated or permitted. Moreover, in Janowiec, para 146, the Grand Chamber emphasised that the time factor is the ‘most crucial indicator’ in relation to the ‘genuine connection’ test and that the test requires that ‘the lapse of time between the triggering event and the critical date must remain reasonably short’. In our judgment, an extension beyond the normal ten year limit of up to two years is permissible where there are compelling reasons to allow such an adjustment constituted by circumstances that (a) any original investigation into the triggering death can be seen to have been seriously deficient and (b) the bulk of such investigative effort which has taken place post-dates the relevant critical date. If in these circumstances there is an extension of no more than two years beyond the ten-year limit mentioned in Janowiec, it remains possible to describe the lapse of time as ‘reasonably short’ in accordance with the guidance in that judgment at paras 146 and 148.”
In essence, therefore, the Supreme Court’s rationalisation of the authorities (putting to one side the “convention values” test) was as follows:
The “genuine connection” test requires that there is a “reasonably short” period between the triggering death and the critical date.
Normally, that time period should not exceed ten years.
For compelling reasons (and leaving aside the “convention values” test), there can be an extension for a further two years to an outer period of 12 years. The compelling reasons spelt out as justifying that extension are where the original investigation was “seriously deficient” and where “the bulk of [the] investigative effort” has taken place after the critical date.
It can be seen that that rationalisation is closely analogous to the reasoning in Janowiec but differs by allowing an extension of two years (from 10 to 12 years) provided two circumstances – which are slightly differently expressed than in Janowiec - are shown to exist: first, an original investigation into the triggering death can be seen to have been seriously deficient (or, one should insert, non-existent); and, secondly, the bulk of such investigative effort which has taken place (or, one should insert, ought to have taken place) post-dates the relevant critical date. It should be noted that the second compelling reason goes beyond the standard requirement in Janowiec that “much” or “a major part” of the investigative effort post-dated (or should have post-dated) the relevant critical date. Therefore, according to the McQuillan rationalisation, with which we agree, the normal rule is that there is no genuine connection if the time period between the relevant death and 2 October 2000 is more than 10 years (and even within that period, applying para 148 of Janowiec, it must be shown that a major part of the investigative effort which has, or ought to have, taken place was after the critical date). But there is a genuine connection if the time period between the relevant death and 2 October 2000 is no more than 12 years provided that the original investigation was seriously defective or non-existent and that the bulk of the investigative effort which has, or ought to have taken place, was after 2 October 2000. That rationalisation combines the certainty of two fixed periods (10 years and 12 years) with the flexibility, for the compelling reasons explained in para 144 of McQuillan , to extend the primary period of 10 years to the outer period of 12 years.
An important feature of that rationalisation – and one might say the driving force behind allowing an extension to 12 years in the circumstances explained in McQuillan – was that it accommodated the decision (albeit not the multi-factorial reasoning of Lord Kerr: see especially at para 108 of his judgment) in the notorious case of Finucane. Applying that rationalisation, although the facts of Finucane fell outside the primary ten-year period required for there to be a genuine connection, they fell within the outer period of 12 years: the period between the triggering event and the critical date was 11 years 8 months. The facts of Finucane were compelling because the original investigation was “seriously deficient” and “the bulk of the investigative effort” had taken place after the critical date.
One aspect of the facts in Finucane marks it out as a particularly appalling case. The victim was targeted and killed because he was a solicitor who had been involved in defending the rights of those charged with terrorist offences and there was alleged collusion in that murder by State agencies. There is therefore a very troubling rule of law aspect to the case. At first instance in Finucane, [2015] NIQB 57, at para 35, Stephens J said the “convention values” test was met. The Court of Appeal, [2017] NICA 7, said the following about this at para 167:
“Stephens J concluded that the obligation of a State not to kill but to protect its citizens and ensure the rule of law was a value at the foundation of the Convention and the murder of a solicitor involving collusion by State agencies negated that foundation. Clearly the facts of this case are unique in this regard and self-evidently would not cover every case where rogue elements in the police or security forces colluded in the murder of a victim. Each case would be fact specific and we do not go so far as to say that the finding of the learned trial judge was necessarily unreasonable in the instant case albeit different conclusions might equally reasonably be reached by other courts.”
Lord Kerr in the Supreme Court in Finucane, at para 113, said that the “convention values” test “…did not occupy much of the oral submissions that were made in this case. In light of that and of my conclusion in relation to the existence of a genuine connection, I propose to say nothing more about it.”
We would not go so far as to suggest that the facts of Finucane met the “convention values” test and to that extent we disagree with Stephens J. That test imposes an extremely high hurdle. What is principally in mind are serious crimes under international law, such as war crimes, genocide or crimes against humanity. In McQuillan, while not necessary for the decision, the Supreme Court considered it likely that acts of torture by the state would also satisfy the test. But, while not falling within the “convention values” test, it is our view that, not least because of the “rule of law” aspect of the facts that we have identified, the decision in Finucane is one that this court ought to be very reluctant indeed to overrule.
The rationalisation given in para 144 of McQuillan, which is consistent with the decision in Finucane, remains valid and should not be departed from. Although strictly speaking obiter dicta (because the time periods in McQuillan were well outside the 12-year period), that rationalisation was carefully considered by the panel of seven justices of the Supreme Court. To depart from that now by insisting on a rigid ten-year period (leaving aside where the facts fall within the “convention values” test) would not only undermine the rationalisation put forward by the Supreme Court but would have the unfortunate consequence of overruling Finucane. There is no justification or need to do so. As has been explained (see para 327 above), this is a matter of domestic law where Strasbourg law is being applied by analogy and there is no precise mirroring of the position in Strasbourg; even with a 12-year period, rights will not be fully brought home.
Mr McGleenan KC submitted that it would be appropriate to overrule Finucane albeit that the appeal would succeed in this case even if the 12 year time limit were applied. In addition to the reasons for rejecting this submission that we have set out at paras 325 and 334 – 336 above, we would make four further observations.
The first is that the Finucane family have clearly relied upon the decision. To that end litigation is ongoing in Northern Ireland given delays in relation to implementation of the decision. A brief summary of the subsequent litigation illustrates the point and highlights the critical importance of this decision in human rights terms. This court, on 27 February 2019 made a declaration that there has not been an Article 2 compliant inquiry into the death of Patrick Finucane, leaving it for the state to decide upon the form of any future investigation if feasible. That case and all subsequent litigation has proceeded on the basis that there has not been an Article 2 compliant inquiry in this case. On 10 October 2020, following a subsequent further judicial review, the Secretary of State provided an apology and damages of £7,500 were paid. Most recently, on 21 December 2022, the High Court in Northern Ireland made a declaration following a second set of judicial review proceedings in terms that at the date of that judgment, there had still not been an Article 2 compliant inquiry into the death of Patrick Finucane: see In re Finucane (No 2) [2022] NIKB 37. Following from this declaration an award of £5,000 damages was also made: see In re Finucane (No 3) [2023] NIKB 42. The litigation remains ongoing in Northern Ireland and will inevitably have to react to the judgment of the court in this case.
The second and linked observation is that, prior to the domestic litigation we have referred to, the Strasbourg Court also considered this case. On 1 July 2003 it held that there had not been an inquiry which complied with Article 2 of the Convention. As a result, the Committee of Ministers, the decision-making body of the Council of Europe, commenced supervision of the Strasbourg Court’s judgment. Whilst this supervision ceased in 2009, in 2021 the Committee of Ministers decided to reopen the execution of the Strasbourg Court’s judgment. That remains the current position. Again, an overturning of the Supreme Court’s decision in In re Finucane would be likely to have a serious impact on the ongoing consideration of the case by the Committee of Ministers.
The third observation we make is that the outcome we favour could not realistically be said to result in a flood of claims which would otherwise be time barred. The extra two years is not going to capture many cases even if a claimant could make out the two compelling reasons specified in McQuillan (ie the original investigation was seriously defective or non-existent and the bulk of the investigative effort which has, or ought to have taken place, was after 2 October 2000).
The final observation is that the rationalisation put forward in McQuillan, which we are here following, achieves the legal certainty that was thought important by the Strasbourg Court in Janowiec. There is a fixed outer limit of 12 years; and the normal ten year period can be departed from only where the two specified compelling reasons exist.
Applying domestic law to the facts of this case
Applying the rationalisation in McQuillan to the facts of this case, the “genuine connection” test is not met. This is because the time period between the triggering death and the critical date (2 October 2000) was outside the outer period of 12 years (albeit by only one month). But in any event, these facts are not such as to take the case outside the primary time period of 10 years. This is because, even accepting that the bulk of the investigative effort has (or ought to have) taken place after the critical date, the original investigation (by which we here mean the PONI investigation) was not “seriously deficient” (see para 320 above dealing with Issue 3).
Applying Finucane as modified by McQuillan, this case falls outside the outer reaches of the 12 year temporal limit and so (given that there is no question of the “convention values” test being satisfied in this case) the court does not have (temporal) jurisdiction. If a further reason is needed not to overturn past precedents or to depart from obiter dicta, it is that overruling the decision in Finucane and departing from the rationalisation in McQuillan would make no difference to the decision on whether there was a genuine connection in this case (and, moreover, the appeal would be allowed in any event given our decision – see para 324 above - on Issue 3).
The AGNI therefore succeeds on Issue 1 albeit that we firmly reject the submissions put forward by Mr McGleenan, on behalf of the AGNI, that Finucane should be overruled and that the obiter dicta in McQuillan should be rejected.
It can be seen, especially from paras 333 – 334 and 337 above, that we agree with the summary of their conclusions on the “genuine connection” test set out by Lord Hodge, Lord Sales and Lady Rose, at para 172 of their judgment.
Conclusion
For the reasons we have given, we would allow the appeal.