CONCURRING OPINION OF JUDGE ROZAKIS

I am in agreement with the conclusion reached by the Grand Chamber that in the circumstances of the case there has been a violation of, inter alia , Article 5 § 1 of the Convention. I also agree with the line followed by the Court which led to a finding of a violation of Article 5 § 1, as it is reflected in paragraphs 81 and 82 of the judgment. I would like, however, to make the following remark in so far as the Court's reasoning in paragraph 81 is concerned.

The Court rightly concludes, in this paragraph, that the continued detention of the applicant, after his sentence for the fraud offence expired on 1 July 1997, was not justified, because there was no causal connection between a possible commission of other, future, non-violent offences and the original sentence for murder in 1967. In reaching this conclusion the Court relied, mainly, on Weeks v. the United Kingdom (judgment of 2 March 1987, Series A no. 114), where the Court said that “the 'detention' must result from, 'follow and depend upon' or occur 'by virtue of' the conviction ... In short, there must be a sufficient causal connection between the conviction and the deprivation of liberty at issue”.

The above excerpt from Weeks , which refers back to Van Droogenbroeck v. Belgium (judgment of 24 June 1982, Series A no. 50), provides a principled ground upon which the Court may build an identical jurisprudential approach on matters of lawfulness of the detention after conviction, but it leaves open, as the present judgment does, the essential question of the notion of causal connection between the original conviction and the continuation of the detention. In other, simple words, the question which has been left unanswered in the present case, in applying the lessons of Weeks and Van Droogenbroeck , is whether a causal link only exists when a person commits or presents a threat of committing a crime identical to the original one (in this case murder), or whether the requirement of the causal connection is satisfied with the commission (also, hence, the threat of future commission) of other offences bearing a resemblance with the original offence (in this case, for example, armed robbery, rape, etc.). I think that the answer to this question is that serious violent offences, other than murder, can satisfy the requirement of a causal connection, and hence, allow the Secretary of State to detain the person concerned. After all, this interpretation of the limit of “causal connection” transpires also from the word “sufficient” which appears in the above-cited excerpt from Weeks .

CONCURRING OPINION OF JUDGE COSTA

(Translation)

I have not hesitated to conclude that there has been a violation of Article 5 § 1 of the Convention. However, like my colleagues Mrs Tulkens and Mr Zagrebelsky, I have difficulty in fully agreeing with the reasoning set out in paragraph 82 of the judgment.

My opinion is similar to theirs, but differs slightly on account of the fact that the applicant never relied on Article 7 of the Convention, which requires that offences and punishments shall be strictly defined by law, and I do not see how his complaint can be reclassified under that provision, albeit not impossible of course.

However, there appears to me to be a hidden converse implication in paragraph 82 of the judgment. According to my colleagues of “the majority unanimous opinion”, a decision-making power by the executive to detain the applicant on the basis of perceived fears of future non-violent criminal conduct (unrelated to his original murder conviction) does not accord with the spirit of the Convention, with its emphasis on the rule of law and protection from arbitrariness.

I agree of course. The fact that it was a Secretary of State who (contrary, moreover, to the Parole Board's recommendation) refused to release Mr Stafford in 1997 (convicted of murder in 1967 and released on licence in 1979 – see paragraphs 10 to 17 of the judgment) reinforces the finding of a violation; the applicant should have been released on 1 July 1997 after serving his sentence imposed in 1994 for conspiracy to forge documents. The Secretary of State decided to keep him in prison for reasons which have quite rightly been censured by our Court for want of a causal link between his initial crime and his supposed dangerousness thirty years later.

The same reasoning would apply, however, even if the impugned decision had been made, not by a member of the executive, but by a judge or a court. I do not doubt that the procedural guarantees would, in theory, have been greater (although the Secretary of State's decision was in this case challenged by the applicant and reviewed by a judge, and subsequently by the Court of Appeal and the House of Lords, in adversarial proceedings). However, in my view, the lack of a causal link would have sufficed to vitiate the decision to keep the applicant in detention from the standpoint of Article 5 § 1 of the Convention, and the rule of law might have been undermined even in that scenario. I therefore consider that paragraph 81 of the judgment was wrong not to envisage it.