The ‘Bill of Rights Bill’, repealing and replacing the Human Rights Act, has already attracted significant criticism. This post focuses on clause 5, with which the government seeks to give effect to its previously expressed scepticism regarding ‘positive obligations’. These are duties on the authorities to take positive measures to protect individuals from human rights breaches, rather than merely refraining from breaching those rights with their own actions. The obligation to protect those suffering domestic violence is such an obligation.
In the government’s view, while the ‘negative’ obligation preventing authorities from inflicting torture themselves, for example, is within the proper scope of human rights protection, positive obligations are highly suspect. It is asserted that they interfere with (or give the judiciary licence to interfere with) the authorities’ decisions on matters of resource allocation and prioritisation, and on the performance of their duties generally.
The Cambridge Centre for Public Law’s response to the Consultation Paper which preceded the Bill (at pages 91-98, to which this author contributed) made the counter-argument, that positive obligations are important and should not be restricted. Ultimately the case against change remains just as compelling, but the purpose of this post is primarily to engage with the newly published draft provision and understand its effect.
Clause 5 gives very clear expression to the anti-positive obligations sentiment. It seeks to implement a seemingly simple idea: that the UK courts cannot establish or apply any ‘new’ positive obligations, and they must be restrained when they consider claims under those positive obligations which have already been established (by the UK courts or Strasbourg). What its terms fail to do is offer a workable structure for the judges who would need to comply with it in deciding cases; as will be seen, significant interpretative work will be required to articulate such a structure. Even when that is done, judges will be left to either recognise that the flexibility of positive obligations is such that most of the concerns the legislation embodies are misplaced, rendering large parts of clause 5 redundant, or to try and find some means to give effect to the clause’s ill-defined hostility to positive obligations.
What is a ‘positive obligation’?
As defined in clause 5 (supported by sub-clause 12(3)), a positive obligation is one which requires the authorities to take an action in order to protect a Convention right. Although it may not always be easy for the courts to determine whether the focus in a given case is on a requirement to act or to refrain from an action, we can at least identify a number of obligations which the Strasbourg court has clearly labelled as ‘positive’. To illustrate the discussion which follows, this post will focus on such obligations under Article 2 (the right to life) and Article 3 (prohibition of torture). According to the Strasbourg court, the ‘positive obligations’ under Articles 2 and 3 include the following:
- To have and to enforce an adequate legal framework preventing private individuals from killing or inflicting torture or inhuman or degrading treatment or punishment (the ‘Framework Obligation’),
- To take reasonable measures to avert any real and immediate risk of such acts which was or should have been known to the authorities (the ‘Operational Obligation’), and
- To effectively investigate arguable claims that such acts have occurred (the ‘Investigation Obligation’).
The intention is not to under-state the importance of positive obligations under other articles of the Convention, but the Article 2 and 3 obligations provide a reasonably well-defined and manageable basis for exploring the impact of clause 5 in practice.
What is ‘an interpretation’? – Part 1, Old or New?
The fundamental line drawn by clause 5 is between positive obligations which already exist and those which would be new, after clause 5 comes into force. It attempts to achieve that division by focusing on ‘interpretations’ of Convention rights, which may be ‘pre-commencement’, if they have been adopted by Strasbourg or a UK superior court of record before clause 5 comes into force, and have not been resiled from or overruled. Pre-commencement interpretations may be applied by the UK courts, although subject to the constraint discussed further below. Other, ‘post-commencement’, interpretations will not be available to the UK courts at all.
The distinction drawn here raises difficult questions as to the scope of an ‘interpretation’ which has been ‘adopted’. To take an example, in Kotilainen and others v Finland, a 2020 case arising from a school shooting, the Strasbourg court held that the perpetrator’s firearm should previously have been seized by the authorities, since ‘doubts had arisen… as to whether the perpetrator was fit to possess a dangerous firearm’. Judge Eicke’s partial dissent criticised this ‘creation’ of a ‘new’ duty, and it was indeed the first ruling to apply a specific obligation to confiscate a firearm from a private individual, based on a suspicion which fell short of the Operational Obligation threshold of a ‘real and immediate risk’. But the Framework Obligation was well established by that time, and the court situated its ‘new’ specific requirement within that broader general requirement, which in this context required the authorities to maintain and enforce effective firearms controls.
Of course the Kotilainen obligation would now be ‘pre-commencement’, but it neatly encapsulates the difficulty which can arise of identifying whether, in any given case, Strasbourg should be treated as having adopted a new ‘interpretation’ or simply applied an existing ‘interpretation’ to a new factual context. Much depends on the degree of specificity in which the relevant obligations must be framed in the pre-commencement jurisprudence. If clause 5 comes into force as presently drafted, much will then turn upon that somewhat abstract distinction, giving litigants plenty of scope for complex arguments about the nature of ‘interpretation’ and ‘obligation’, and degrees of specificity or generality. It is not, however, the most challenging aspect of the clause.
What is ‘an interpretation’? – Part 2, All or Nothing?
Having cleared the hurdle of being found to reflect a ‘pre-commencement interpretation’, and thus potentiallyapplicable in UK courts, a positive obligation then faces sub-clause 5(2), as follows:
‘(2) In deciding whether to apply a pre-commencement interpretation of a Convention right that would require a public authority to comply with a positive obligation, the court must give great weight to the need to avoid applying an interpretation that would—
(a) have an impact on the ability of the public authority or of any other public authority to perform its functions;
(b) conflict with or otherwise undermine the public interest in allowing public authorities to use their own expertise when deciding how to allocate the financial and other resources available to them, including in particular the professional judgment of those involved in operational matters;
(c) require the police to protect individuals who are involved in criminal activity or otherwise undermine the police’s ability to determine their operational priorities;
(d) require an inquiry or other investigation to be conducted to a standard that is higher than is reasonable in all the circumstances;
(e) affect the operation of primary legislation (including primary legislation relating to supply and appropriation).’
Sub-clause 5(2) thus confers a statutory mandate on the courts to decide whether or not to ‘apply’ something which isan established interpretation of the Convention.
The first question to consider is how the impact of the ‘interpretation’ in question should be assessed. Although sub-clause 5(2) requires that ‘great weight’ be afforded to the need to avoid the outcomes listed at sub-clauses 5(2)(a)-(e), it does not indicate whether the requirement is simply to avoid those outcomes in the specific case under consideration (in which the ‘interpretation’ will or will not actually be applied), or to consider the implications of the ‘interpretation’ more widely.
Presumably it is the latter, given the very general framing of the outcomes themselves (such as ‘impact on the ability of the public authority or of any other public authority to perform its functions’). It is after all an ‘interpretation that would…’ have those consequences which the courts are instructed to avoid applying: in other words, a general rule rather than its specific application. Further, the Explanatory Notes accompanying the Bill suggest that the listed consequences to be avoided ‘are intended to guide courts to consider the wider implications of their decision (rather than just the need to do justice in the particular case).’
Excluding the narrowest ‘case by case’ option still leaves the question of exactly how widely the implications of the interpretation in question need to be considered. The first limb of sub-clause 5(2)(c) appears the simplest in this respect. The established interpretation of the Operational Obligation appears to require the protection of those involved in criminal activities, where the ‘real and immediate risk’ threshold is met in relation to such an individual. Applying that interpretation in cases where the individual requiring protection was not so involved has no implications for sub-clause 5(2)(c), and so presumably it can be applied (subject to the other sub-clauses). Applying it in cases where the individual is so involved does engage clause 5(2)(c), and so ‘great weight’ will be given to the need not to apply it.
The other sub-clauses are harder. From the text it seems that the question to be asked is whether the existence of the Operational Obligation (for example), of itself, overall has ‘an impact on the ability of the public authority or of any other public authority to perform its functions’? Unlike the Framework Obligation, which could plausibly be sub-divided (into, say, the gun control obligation from Kotilainen, the obligation to manage the threats to the public associated with prison release schemes, and so forth), the Operational Obligation is general and seemingly indivisible. If so, does that mean ‘great weight’ should always be given to the need to avoid applying it? Or is there an alternative more sophisticated balance available?
Sub-clause 5(2) does not say what the courts should do if they decide not to apply the interpretation in question. Sub-clause 5(1), as outlined earlier, apparently precludes adopting a new, narrower one, and applying that. However, an ‘all or nothing’ approach would seem to yield bizarre outcomes; for instance, if the Investigation Obligation were considered to generally impose higher than reasonable requirements and thus trigger sub-clause 5(2)(d), reading down the obligation to one which is reasonable is more logical than refusing to apply the obligation entirely. Much the same would be true if the extent of the Operational Obligation were held generally to intrude on police priorities (sub-clause 5(2)(c)) or affect the ability of the police to perform their functions (sub-clause 5(2)(a)). It seems necessarily implicit that, in requiring the application of an over-broad interpretation be avoided, without saying how it should be avoided, sub-clause 5(2) enables the courts to substitute their own narrower interpretation.
De facto, if so, the courts would be providing a new interpretation, as indeed they need to if public authorities are to have rules to rely on in determining what is required of them under positive obligations. But formally speaking (for purposes of sub-clause 5(1)) these must be regarded not as new interpretations of the Convention, but as rules about how great weight will be given to avoiding the consequences listed in sub-clauses 5(2)(a)-(e) in the courts’ decisions on whether or not to apply pre-commencement interpretations. It should be apparent from the discussion thus far that clarification in the wording of sub-clause 5(2) would be advisable.
Such clarification could also address the apparent disconnect between sub-clause 5(2) and sub-clause 12(1) (which renders it ‘unlawful for a public authority to act in a way which is incompatible with a Convention right’, subject to exceptions at sub-clause 12(2)). Sub-clause 5(2) means a court may decide not to ‘apply’ a pre-commencement interpretation of a Convention right which requires a public authority to comply with a positive obligation. But it does not purport to invalidate that interpretation. Accordingly, if a public authority does not comply with a pre-commencement positive obligation, its failure to do so would appear to be ‘unlawful’ under sub-clause 12(1) even though a court might then decide not to ‘apply’ the interpretation giving rise to that positive obligation under sub-clause 5(2). Although a public authority might not be so concerned about its failure to act being unlawful if sub-clause 5(2) insulates it from having a court actually find that failure unlawful, from the perspective of the rule of law such a situation should be troubling to the authority, as well as to the judiciary and the public.
The question which remains is how a court should decide whether one of the outcomes listed at sub-clauses 5(2)(a)-(e) actually requires that an ‘interpretation’ should not be applied.
Most of the outcomes are expressed in broad terms, with little qualification (as to whether the anticipated effects which would favour avoiding an interpretation must be material, significant or indeed adverse). If the avoidance of such effects is plausibly to carry ‘great weight’, logic suggests that some threshold must be imposed in applying the vague terminology of the relevant sub-clauses.
Even then, as the Cambridge Centre for Public Law Response noted, in fact the Convention’s positive obligations have been carefully crafted to avoid imposing unreasonable burdens or preventing the authorities from using their expertise. Further, the protection of at least the ‘fundamental rights’ protected by Articles 2 or 3 (or 4) should also attract ‘great weight’. In that light, much of sub-clause 5(2) appears redundant. Most sensibly, sub-clause 5(2) could be construed as a ‘safety valve’ to be relied on in the (remote) eventuality of clear evidence that the wrong balance has been struck by the court which formulated the relevant ‘pre-commencement interpretation’. However, the government’s hope may be that this provision will simply tilt the scales further in favour of the authorities, with little or no detailed enquiry into the underlying assumptions that positive obligations are detrimental to decision-making. As Lock has argued, ‘it is the Government who will claim to have the authoritative perspective on this and will inevitably call for deference from judges while pressuring them to avoid making a finding of a rights violation’.
In each of the provisions of clause 5, the sentiment is clear. How the courts are actually intended to give effect to them is not, and significant judicial work would be necessary to articulate a workable scheme which takes proper account of the legislative text and the context in which it will operate. The danger which seems most likely to crystallise is that the courts look through the text to the underlying sentiment and simply adopt a more deferential pose, in favour of the executive, when they consider the essential field of positive obligations. It is to be hoped that, if the clause is not clarified or rejected, the courts will at least take a suitably rigorous approach to its interpretation and application.
The author is very grateful to Professor Michael Gordon and Professor Alison Young for their valuable comments on an earlier draft. Any errors remain his own.
Dr Edmund Robinson is a Bye-Fellow in Law at Downing College, University of Cambridge.
(Suggested citation: E. Robinson, ‘Fumbling with interpretation – Clause 5 of the Bill of Rights and the positive obligations challenge’, U.K. Const. L. Blog (27th July 2022) (available at https://ukconstitutionallaw.org/))