On a couple of occasions in this blog, I’ve noted what I see as an important difference between the UK’s interpretation mandate:
3(1) So far as it is possible to do so, primary legislation and subordinate legislation must be read and given effect in a way which is compatible with the Convention rights.
32(1) So far as it is possible to do so consistently with their purpose, all statutory provisions must be interpreted in a way that is compatible with human rights.
I think the words ‘read and given effect’ are broader than ‘interpreted’, with the result that Victoria’s (and the ACT’s) interpretation mandates are (for that reason, in addition to others) narrower than the UK’s. In particular, I’m dubious about Australian courts adopting the House of Lords’s view in Ghaidan v. Godin-Mendoza  UKHL 3 that the language of the statute provision isn’t a particularly big barrier to rights-compatible interpretation.
A commenter on a recent post questions me on this:
In a nutshell, I don’t get it. How does a requirement to ‘give effect’ amount to a license to rewrite the law? Surely to ‘read and give effect’ simply means to ‘interpret’: if a law is interpreted to mean X, doesn’t it follow that the law must then be given effect as X?
The commenter rightly points out that most of the attention in debates on the limits of the UK and Victorian interpretation mandates has been directed to the shared ‘possible’ language in both. Neither of the two major texts on the Charter, nor Julie Debeljak in her major essay on Charter interpretation, pay the slightest attention to the words ‘read and given effect’ in the UK mandate, despite all relying heavily on comparison with the UK provision. The implication is that the only relevant difference between the Australian and UK provisions is the ‘purpose’ qualification (a qualification have argued elsewhere was a bad idea when it comes to protecting rights.) But I think the contrast between ‘read and given effect’ and ‘interpreted’ matters. Here’s why: [EDIT: I've updated the argument now to take account of further information mentioned in the comments.]
First, there’s a significant lineage for the ‘read and given effect’ terminology of the UK interpretation mandate in older rights statutes:
- Canadian Bill of Rights Act 1960, s2: Every law of Canada shall, unless it is expressly declared by an Act of the Parliament of Canada that it shall operate notwithstanding the Canadian Bill of Rights, be so construed and applied as not to abrogate, abridge or infringe or to authorize the abrogation, abridgment or infringement of any of the rights or freedoms herein recognized and declare…
- European Convention Act 1972, s2(4): The provision that may be made under subsection (2) above includes, subject to Schedule 2 to this Act, any such provision (of any such extent) as might be made by Act of Parliament, and any enactment passed or to be passed, other than one contained in this part of this Act, shall be construed and have effect subject to the foregoing provisions of this section
Both of these provisions have been held to require far more than interpretation: they both require courts to decline to apply contrary legislation! By contrast, another statute in this line, the Hong Kong Bill of Rights Ordinance, only demands that statutes be ‘constructed’ consistently with the ICCPR (with a separate provision providing for the repeal of inconsistent pre-HKBoR ordinances.)
I think it’s the enormous breadth of the words ‘read and given effect’ (in light of these earlier HR statutes) that is the reason why British judges and commentators typically ignore them and instead concentrate on the ‘possible’ language. Basically, without the ‘So far as it is possible to do so’ proviso, the UK HRA’s interpretation mandate could well have been read the same way as the earlier acts. As noted in Ghaidan, the possible language seems to be borrowed from a classic European Court of Justice judgment on domestic obligations under the EC treaty, which (unlike the ECtHR, which ignores internal distinctions between different branches of member nations) set out the separate obligations of the judicial branch in relation to interpreting local legislation using the ‘possible’ proviso. So, the UK judges interpreting the ‘possible’ language focus on ‘separation of powers’ concerns as the sole significant limit to the UK interpretation mandate.
Second, not all British judges have ignored the ‘read and given effect’ language, even in Ghaidan. Lord Millet certainly read the language as dominated by the word ‘read’:
[T]he exercise which the court is called on to perform is still one of interpretation, not legislation: (legislation must be “read and given effect to”). [Emphasis in original]
But Lord Rodger said the opposite:
The use of the two expressions, “read” and “given effect”, is not to be glossed over as an example of the kind of cautious tautologous drafting that used to be typical of much of the statute book. That would be to ignore the lean elegance which characterises the style of the draftsman of the 1998 Act. Rather, section 3(1) contains not one, but two, obligations: legislation is to be read in a way which is compatible with Convention rights, but it is also to be given effect in a way which is compatible with those rights. Although the obligations are complementary, they are distinct. So there may be a breach of one but not of the other. For instance, suppose that legislation within the ambit of a particular Convention right requires a local authority to provide a service to residents in its area. The proper interpretation of the duty in the legislation may be straightforward. But, even if the local authority interprets the provision correctly and provides the appropriate service, if it provides the service only to those residents who support the governing political party, the local authority will be in breach of article 14 in relation to the other article concerned and, in terms of section 3(1), will have failed to give effect to the legislation in a way which is compatible with Convention rights. So, even though the heading of section 3 is “Interpretation of legislation”, the content of the section actually goes beyond interpretation to cover the way that legislation is given effect.
He goes on to make the interesting observation that the ‘give effect’ aspect of the UK interpretation mandate fits in with the UK’s conduct mandate, which he notes applies to courts (a contrast with the ACT and Victorian conduct mandates, which exempt courts in the their non-administrative functions.) Rodger also notes that the UK’s equivalent to Charter s. 38(2) effectively provides a limit to the ‘give effects’ aspect of the UK interpretation mandate. He’s a smart one, I reckon. But who’s right?
Unsurprisingly, all UK judges are at pains to say that they agree with one another on the basic test, including the ‘fundamental features’ limitation. The trick to working out the difference between them is to see how they apply the test. In Ghaidan, Millet – who emphasised ‘read’ – dissented, refusing to interpret the words ‘as his or her wife or husband’ as covering opposite sex couples. Lord Rodger, in common with the other majority judges, was happy to ‘read in’ the words ‘if he or she was’ between ‘as’ and ‘his or her’. So, I reckon, when we are talking about the approach in Ghaidan, we are talking about Rodger’s approach, which sees the UK interpretation mandate as requiring more than just interpretation.
Third, against this background, surely it’s not irrelevant that the Charter’s drafters opted for ‘interpreted’. [EDIT: According to information supplied in the comments, the Consultative Committee's draft used the UK's 'read and given effect' language, but without any commentary on that in the Report, which appeared to equate that language with 'interpreted'. For whatever reason, this language was replaced with 'interpreted' by the meddlers, the nameless government folks who made numerous annoying changes to the Committee's draft, without any explanation. While the Committee's report might suggest that the Committee thought that the language made no difference, it's hard to see any explanation for the meddling except that the meddlers - and therefore the government - thought that it would make a difference, presumably with 'interpreted' being narrower than 'read and given effect', as I've argued here.]
The EM isn’t especially explanatory, of course:
Sub-clause (1) establishes the requirement that courts and tribunals must interpret all statutory provisions in a way that is compatible with human rights, so far as it is possible to do so consistently with the purpose of the statutory provision. The object of this sub-clause is to ensure that courts and tribunals interpret legislation to give effect to human rights. The reference to statutory purpose is to ensure that in doing so courts do not strain the interpretation of legislation so as to displace Parliament’s intended purpose or interpret legislation in a manner which avoids achieving the object of the legislation.
But it is worth noting that the drafters’ choice to implement the purposive limitation by inserting the words ‘consistently with their purpose’ after ‘So far as it is possible to do so’ could be read as giving a new meaning to the ‘possible’ language as being exclusively concerned with respecting Parliament’s purpose (rather than the separation of powers or whatever.) That would mean that UK precedents on the possible language really aren’t that important here. (I’m no fan of the ‘consistently with their purpose’ insertion, both because I don’t much like the purposive limitation in a rights context, but also because I think it’s a clumsy textual way of limiting the interpretation mandate. Why not draft the purpose requirement as an additional sub-section? That way, it would have been clear that the rest of the language of Charter s. 32(1) was intended to closely follow the UK approach.)
Fourth, linguistically, the odd one out in all these provisions is NZ’s interpretation mandate:
6 Wherever an enactment can be given a meaning that is consistent with the rights and freedoms contained in this Bill of Rights, that meaning shall be preferred to any other meaning.
I’m totally with the House of Lords, who think that the drafters of the UKHRA deliberately rejected the NZ language as too weak (and, incidentally, I also assume that the drafters of the NZBoRA deliberately rejected the wording of the CBoR as too strong.) It’s true that the NZ Supreme Court in Hansen rejected the view that their interpretation mandate was worded in a weaker way than the UK’s. It’s an amusing dispute: My mandate is stronger than yours! No it’s not! Maybe they should arm wrestle.
But I think there’s three reasons to think that the NZSC’s view isn’t good evidence that the words ‘read and given effect’ mean the same thing as NZ’s ‘given a meaning’ or Victoria’s ‘interpreted’. First, the UK and NZ judges – when they discussed eachother’s statute – were concerned with the ‘possible’/’can’ and ‘must’/’shall’ distinctions; they weren’t talking about the ‘read and given effect’/’given a meaning’ language. Second, Tipping J in the NZSC, in arguing that both were mere interpretation provisions, cited Lord Millet, the judge in Ghaidan who emphasised the word ‘read’. As argued above, Millet’s view was not the majority’s. And, third, also as argued above, the real test of the relative strictness of the two mandates isn’t what the judges say about them, but what the judges do with them. A comparison of the outcome of Hansen (where the NZSC refused to read ‘proved’ as ‘tested’ in a reverse onus provision and instead issued a declaration of incompatibility) and the earlier UK decision in Sheldrake (where the HoL made just that reading) is quite telling.
So, I think the NZ interpretation mandate is weaker than the UK’s. Indeed, I think there’s good reason to think that the NZ mandate is pretty similar to Victoria’s. The commenter is right to point out, though, that the NZBoRA doesn’t have a built-in purpose limitation, though NZ does have a fairly pushy requirement that the meaning of a statute ‘must be ascertained… in light of its purpose’ in its Interpetation Act. The ACT’s requirement is pushier, of course, and that’s something the Canberrans will have to work out on their own. (Does the ACT interpretation mandate trump the ACT’s purpose rule?) And Victoria’s is weaker (so the interaction between the two rules doesn’t matter.) But I’d say that, apart from these different purpose rules, the NZ, Victorian and (now) ACT (and HK) interpretation mandates are close enough to make them much more important precedents for eachother, than the UK mandate is for any of them.
Fifth, [EDIT: interestingly, in the ACT, their consultation committee originally proposed this as their jurisdiction's interpretation mandate:
A court or tribunal must interpret a law of the Territory to be compatible with human rights and must ensure that the law is given effect to in a way that is compatible with human rights, as far as it is possible to do so.
This is the clearest example of an approach that treats 'given effect' as something different to 'interpret'. The Report uses the wording 'read and given effect' and 'interpreted' interchangeably and refers to quite traditional limits on interpretation. It's hard to know what to make of this draft, but I can't help but think that if it had been enacted, then the ACTHRA's operation might have been argued to be much more far reaching than it currently is, as the second part of the above clause might have been regarded as a conduct mandate but without the statutory defence and (unlike the current Victorian and ACT versions) applicable to courts in all functions. A sovereignty-destroying Bill of Rights by stealth? Naughty consultation committee! Alas, the ACT's meddlers, for whatever reason (but perhaps because they realised these potential arguments) replaced this draft with a NZ-like 'meaning' mandate.]
The ACT has of course recently adopted Victoria’s language (a poor choice, I think, but better than the ACTHRA’s dreadful previous provision, with its bizarre deference to the ‘best achieves’ purpose rule.) Now, it’s true that the ACT’s EM expressly states that the ACTLA’s intention was to follow Ghaidan. If that’s the case, I wonder, why on earth didn’t they use the UK wording (perhaps with Victoria’s clumsy ‘purpose’ limitation?) I imagine that this is yet another example of the bad effects of the drive to uniformity and harmonisation. It’ll be up to ACT courts to puzzle out whether the words (favouring the Victorian approach) or the EM (favouring the British approach) win out in the ACT. But the ACT’s EM scarcely provides any credible guidance on what Victoria’s drafters meant to (or did) achieve.
Finally, my view that the UK’s words matter comes from my own policy view on statutory laws: that all statutory words matter! The Charter’s and every other statute. I’m suspicious of arguments that we should just assume that variations in wording are meaningless, especially when the variations were drafted in full knowledge of the alternatives. It smacks of a ‘cases, not statutes’ approach so beloved by law students. Indeed, I think a lot of the poor wording of our statutes (including the Charter) comes down to a similarly lazy view held by some drafters. Interestingly, probably the only Australian academic other than me to note the particular language of the UK’s interpretation mandate is James Allan. He, like Lord Millet, emphasises the ‘read’ bit of ‘read and given effect’. Allan’s agenda, of course, was to play down the differences between the UKHRA and the Charter, so he could play up his argument that the Australian statutes are just as bad for democracy as the UK statute (and the Canadian Charter and the US Bill of Rights and all that.) We shouldn’t be helping him out in that particular game.
The House of Lords in Ghaidan certainly make a beautiful argument for why words shouldn’t matter in human rights cases:
[I]t becomes impossible to suppose Parliament intended that the operation of section 3 should depend critically upon the particular form of words adopted by the parliamentary draftsman in the statutory provision under consideration. That would make the application of section 3 something of a semantic lottery. If the draftsman chose to express the concept being enacted in one form of words, section 3 would be available to achieve Convention-compliance. If he chose a different form of words, section 3 would be impotent. (Lord Nicholls)
Nowhere in our legal system is a literalistic approach more inappropriate than when considering whether a breach of a Convention right may be removed by interpretation under section 3. (Lord Steyn)
Attaching decisive importance to the precise adjustments required to the language of any particular provision would reduce the exercise envisaged by section 3(1) to a game where the outcome would depend in part on the particular turn of phrase chosen by the draftsman and in part on the skill of the court in devising brief formulae to make the provision compatible with Convention rights. The statute book is the work of many different hands in different parliaments over hundreds of years and, even today, two different draftsmen might choose different language to express the same proposition. In enacting section 3(1), it cannot have been the intention of Parliament to place those asserting their rights at the mercy of the linguistic choices of the individual who happened to draft the provision in question. What matters is not so much the particular phraseology chosen by the draftsman as the substance of the measure which Parliament has enacted in those words. Equally, it cannot have been the intention of Parliament to place a premium on the skill of those called on to think up a neat way round the draftsman’s language. Parliament was not out to devise an entertaining parlour game for lawyers, but, so far as possible, to make legislation operate compatibly with Convention rights. This means concentrating on matters of substance, rather than on matters of mere language. (Lord Rodger)
These are compelling arguments. But I think an equally, if not more, powerful argument is that fidelity to the words of all statutes also serves important human rights ends: making it possible for people who are affected by statutes to understand what they do (rather than waiting for a court to read in or read out words); and forcing parliament to take ownership of its work. Otherwise, if we leave it up to courts to make our statutes work, then we are exposing people to a different lottery: the one that occurs whenever a courts decides (or doesn’t decide) to read in or read out words from statutes. So long as the courts stick to interpretation, rather than more than that, I’m sanguine about the meanings of statutes altering as the Charter takes hold; but if the courts go beyond that, down the UK line, then we’ll have to start taking precautions to ensure that people aren’t disadvantaged by retrospective changes in the rules that apply to them.
A limited interpretation mandate doesn’t mean limited rights protection, but rather the different rights protection offered by declarations of inconsistent interpretation. One of the major bodies with a ‘responsibility’ for human rights in the Charter is Parliament, which is meant to fix its own statutes. We don’t have the ECtHR breathing down our necks. Instead,our statutes are meant to be as much about a human rights dialogue as they are about a human rights culture. I’m especially concerned that if Victorian courts are too keen to fix Parliament’s errors using the interpretation mandate, then the drafters (and politicians) will get even lazier than they already are about trying to draft and pass good legislation in the first place. So, I’m especially alarmed at the prospect of Victorian judges ‘reading in’ or ‘reading out’ words in statutes, along UK lines. Just read the words.