King, R (On the Application Of) v The Secretary of State for the Home Department

This appeal concerns the lawfulness of the guidance given by the Parole Board (“the Board”) to its panels in December 2013 (“the December 2013 Guidance”) as to the test to be applied by panels of the Board when considering whether to direct the release after recall to custody of a prisoner serving a determinate sentence of imprisonment. The December 2013 Guidance was issued one year after the coming into force of the Legal Aid, Sentencing and Punishment of Offenders Act 2012 (“LASPO”) which made major changes to the law on recall and release of such prisoners.

The relevant part of the Guidance stated:

“In order to direct release, the Board should be satisfied that it is no longer necessary for the prisoner to be detained in order to protect the public from serious harm (to life and limb). It is not a requirement to balance the risk against the benefits to the public or the prisoner of release.”

The appellant’s case is that the Guidance contains two flaws: (i) the first sentence wrongly directs panels to apply the “public protection test” set out in section 255B(3) and section 255C(3) of the Criminal Justice Act 2003 (“the 2003 Act”) as amended by LASPO; and (ii) the advice contained in the second sentence is wrong: in an appropriate case, the Board is entitled or obliged to conclude that a significant risk of harm to the public by reason of a prisoner’s early release is outweighed by the benefits of such release.

The Divisional Court (Aikens LJ and Mitting J) rejected both contentions in a comprehensive and detailed judgment. Mr Southey QC submits that they were wrong to do so.

The judgment below contains a detailed description of the legislative and case history relating to the release of prisoners: see paras 22 to 49. The following summary should suffice for the purposes of this appeal.

Prior to 1991, there was no statutory test to be applied by the Board in considering the release of prisoners of any kind. The position with regard to prisoners serving discretionary life sentences was as follows. In R v Parole Board, ex parte Bradley [1991] 1 WLR 134, the Divisional Court propounded a test that has become known as the “life and limb” test: is there more than a minimal risk that the prisoner will commit further offences of a violent or sexual nature which will cause serious harm? This test was later approved by this court in R v Parole Board, ex parte Wilson [1992] QB 740.

The Criminal Justice Act 1991 (“the 1991 Act”) introduced a statutory test in section 34(4)(b) for prisoners serving discretionary life sentences. This was reproduced in identical language in section 28(6) of the Crime (Sentences) Act 1997 (“the 1997 Act”). It provided:

“(6)

the Parole Board shall not give a direction [for release] under subsection (5) above with respect to a life prisoner to whom this section applies unless---

…

(b)

the Board is satisfied that it is no longer necessary for the protection of the public that the prisoner should be confined.”

Section 39 of the 1991 Act, which governed release following recall, was silent as to the test to be applied. In R v Parole Board, ex parte Watson [1996] 1 WLR 906, this court decided that the test was the same for release of recalled prisoners as for their initial release.

Sentences of imprisonment for public protection (“IPPs”) were introduced by the Criminal Justice Act 2003 (“the 2003 Act”). They are treated as life sentences for the purposes of release decisions by the Board, so that the test in section 28(6)(b) of the 1997 Act applies. In R (Sturnham) (No 2) v Parole Board [2013] UKSC 47, [2013] 2 AC 254, it was held that the Bradley test should be applied to IPP prisoners.

So far as determinate sentence prisoners are concerned, the position was different. Guidance was issued by the Secretary of State (pursuant to section 32(6) of the 1991 Act and section 239(6) of the 2003 Act) in directions issued in May 2004 (“the May 2004 Directions”) as to the test to be applied. These provided:

“In deciding whether or not to recommend release on licence, the Parole Board shall consider primarily the risk to the public of a further offence being committed at a time when the prisoner would otherwise be in prison and whether any such risk is acceptable.”

Section 29(2) of the Criminal Justice and Immigration Act 2008 (“the 2008 Act”) introduced major changes to the provisions of the 2003 Act relating to the recall of prisoners serving determinate sentences. These were contained in sections 255A, 255B and 255C. They were re-enacted by LASPO with amendments which are not material to the issues arising on this appeal. Like the Divisional Court, I find it convenient to set these out in full as an annex to this judgment.

The December 2013 Guidance addresses separately the test for initial release and release following recall of determinate prisoners. It says that the statutory test is the same for the release of all determinate prisoners, viz: the Board must not give a direction for release unless it is satisfied that it is no longer necessary for the protection of the public that the person should be confined (“the public protection test”). It states that the Board will now apply the public protection test in all determinate sentence cases at first release. It states that LASPO is silent as to the test to be applied for the release of recalled determinate sentence prisoners. It goes on to give reasons for saying that the public protection test should also be applied to recalls: see para 2 above.

With that introduction, I can now address the two issues that arise on this appeal.

In R (Clift) v Secretary of State [2007] 1 AC 484, Lord Bingham considered the justification for schemes of early release on licence of prisoners serving long-term determinate sentences. At para 18, he referred to the “recognition that neither the public interest nor the interest of the offender is well served by continuing to detain a prisoner until the end of his publicly pronounced sentence”. He made a similar point in R (Smith and West) v Parole Board [2005] 1 WLR 350. At para 25, he referred to the desirability of the process of transition from prison to life in the community being professionally supervised “to maximise the chances of the ex-prisoner’s successful reintegration into the community and minimise the chances of his relapse into criminal activity”. Mr Southey relies on these observations in support of his submission that there can be a powerful public interest in favour of the early release of a determinate prisoner even where the prisoner poses a high risk to the public. Determinate prisoners, however dangerous they may be, will always be released when they have served their sentence. If they are detained until the expiry of their sentence, they will not on their release be supervised and supported in the community. There is, therefore, a real public interest (and the protection of the public will be enhanced) if a determinate prisoner who has been recalled is released before the expiry of his sentence with the support of probation officers and other professionals.

Mr Southey submits that the Divisional Court was wrong to conclude that the Board is precluded from considering the public interest in the benefit of early release to the prisoner and the public. It would be surprising if Parliament had required a public body such as the Board to act contrary to the public interest.

He relies on the fact that, although, since the passing of the 2008 Act, the Secretary of State has been subject to the statutory public protection test in relation to the release of a determinate prisoner who has been recalled, the Board has never been subject to such a statutory test. He submits that this shows that Parliament did not intend to impose on the Board any test for the release of determinate prisoners who have been recalled. It must be taken to have been content with the test that the Board was required to apply before the enactment of the 2008 Act (as well as LASPO) which was clearly set out in the May 2004 Directions. Mr Southey submits that Parliament is taken to know the state of the law at the time when it legislates: see, for example, R v Immigration Officer at Prague Airport, ex parte European Roma Rights Centre [2005] 2 AC 1 at para 42. Accordingly, if Parliament had intended to amend the test to be applied by the Board in relation to recalled determinate prisoners, it would have made this intent clear by expressly so providing. It did not do so.

The May 2004 Directions could not amend the statutory release test: see R (Girling) v Parole Board [2007] QB 783. It must, therefore be assumed that they reflected the statutory test. They were not withdrawn until July 2013. Just as the issuing of directions by the Secretary of State could not amend the release test, the withdrawal of those directions could not do so either.

There may be a good reason why Parliament intended the Board to apply a different test to the initial release of a determinate prisoner from the test to be applied to the release of a recalled prisoner. Mr Southey points out that, when the Board considers the initial release of a determinate prisoner, it will know that there will be supervision on licence upon release. That will not be the case as regards the release of recalled prisoners.

I do not accept these submissions substantially for the reasons advanced by Mr Grodzinski QC (and accepted by the Divisional Court).

Section 255B is concerned with prisoners who are suitable for automatic release. These are prisoners who the Secretary of State is satisfied will not “present a risk of serious harm to members of the public”. Subsection (3) provides that they may, following recall to prison, be released again by the Secretary of State if he is satisfied that the public protection test is met. Subsection (5) provides that, following representations made to the Secretary of State by the prisoner under section 254(2), the Board may direct the Secretary of State to release the prisoner on licence. Section 255B(5) does not expressly replicate the subsection (3) test. The trigger for the intervention of the Board is a reference by the Secretary of State following the making of representations by the prisoner to the Secretary of State for his recall to be revoked.

Section 255C is concerned with extended sentence prisoners and those not suitable for automatic release. Subsection (2) gives the Secretary of State the power to release a recalled prisoner. Subsection (3) provides that he must not exercise this power unless he is satisfied that the public protection test is met. Like section 255B(5), section 255C(5) does not expressly replicate the test to be applied by the Secretary of State under subsection (3). As in section 255B, the trigger for the intervention of the Board is a reference by the Secretary of State, inter alia, following representations by the prisoner under section 254(2).

Although section 255C does not expressly incorporate a test for the Board to apply when exercising its subsection (5) power, for reasons that I shall explain, in my view it is implicit in section 255C (and section 255B) that the test to be applied by the Board is the same as that to be applied by the Secretary of State if he exercises the power to release without a direction from the Board.

First, I would hold simply on the basis of the language used in section 255B and 255C that Parliament must have intended the Board to apply the same test as that which it had expressly provided was to be applied by the Secretary of State under subsection (3). Only the Secretary of State has the power to release a recalled determinate prisoner. He can do this on his own initiative or at the direction of the Board. The Board has no power to bring about the release of a recalled prisoner other than by means of a direction to the Secretary of State if the Secretary of State has not released the prisoner. The function of the Board is to review the decision of the Secretary of State. In my view, it is implicit in such a scheme that the Board should apply the same test as the Secretary of State whose decision is being reviewed.

I accept that, in theory at least, Parliament might have intended that the Board should apply a different test from that applied by the Secretary of State. I find it difficult to see why Parliament would have wished to do this (see para 25 below). But if Parliament had so intended, it would surely have identified the different test in the statute in explicit terms. In my view, the reason why it did not consider it necessary to set out the test to be applied by the Board was because it was obvious that it was to be the same test as that applied by the Secretary of State. Mr Southey submits that the May 2004 Directions must be presumed to have lawfully reflected the law to be applied by the Board. He says that Parliament must have considered that there was no need to spell out the different test to be applied by the Board because that test had already been set out in the May 2004 Directions. I reject this submission. Directions can be withdrawn by the Secretary of State at any time. They might or might not correctly reflect the law. It is fanciful to suppose that Parliament would have reasoned that it was unnecessary to spell out the different test to be applied by the Board because it was already sufficiently provided in the May 2004 Directions.

Secondly, I do not accept that there is a rational justification for requiring the Secretary of State to apply the public protection test, but requiring the Board to apply the more generous test reflected in the May 2004 Directions. Parliament must have intended that a prisoner either is suitable for release (whether by the Secretary of State or the Board) or is not suitable. It is incoherent to say that, on an application of the test to be applied by the Secretary of State, the prisoner is not suitable for release; but on an application of the test to be applied by the Board, he is suitable for release. Mr Southey points to fact that there is a public interest in releasing a determinate prisoner following recall so that he can benefit from supervision in the community. But I do not see how this can justify imposing different tests on the Secretary of State and the Board: they are both required to address the same question, namely whether a determinate prisoner following recall should be released before the expiry of his sentence.

Thirdly, if Mr Southey is right then the public protection test is not the test to be applied when the Board is considering the release of a determinate sentence prisoner following recall notwithstanding that (i) this is the test to be applied for the initial release of indeterminate sentence prisoners (including IPP prisoners): see Bradley and Sturnham; (ii) it is the test applicable to the release of indeterminate prisoners following recall: see Watson; (iii) it is now the test applicable to the initial release of those categories of determinate sentence prisoners in relation to whom the Board has a role in deciding initial release (namely the various categories of prisoner identified in Schedule 20B to the 2003 Act); and (iv) it is the test which the Secretary of State herself applies to the release of determinate sentence prisoners after recall under section 255B(3) and section 255C(3).

For all these reasons, I would hold that the section 255C(3) test is to be applied by the Board when exercising its power under subsection (5).

This issue raises the question of the proper interpretation of the public protection test: do the words “necessary for the protection of the public” require or permit the Secretary of State (and the Board in the light of my conclusion on the first issue) to balance the risk of offending against the advantages of early release?

Mr Southey submits that the language of the test requires or permits such a balancing exercise to be undertaken. The public protection test is flexible: there are various ways in which the public may be protected from the danger posed by a prisoner. One of these is by releasing him early into the community under supervision. A determinate prisoner who is released after completing his sentence is not subject to supervision. There is, therefore, a real sense in which early release of a recalled prisoner may afford the public greater protection than release at the end of the sentence. It is unlikely that Parliament would have intended to prevent such public interest factors being considered by the Board. Moreover, where Parliament wished to introduce an exclusive test of risk, it did so. Thus, section 255A(4) provides that a person is suitable for automatic release “only if the Secretary of State is satisfied that the person will not present a risk of serious harm to members of the public if released at the end of that period”.

Mr Southey also says that the dicta of Lord Bingham in Smith and West and Clift to which I have referred at para 14 above make clear that early release may positively protect the public by ensuring that the prisoner is supervised in the community. It is not necessary to detain a prisoner if the risk he poses is outweighed by the benefit to the public of early release.

I would reject these submissions. First, as a matter of ordinary language, the words “necessary for the protection of the public” do not entail a balancing exercise in which the risk to the public is to be weighed against the benefits of release to the prisoner or the public. The concept of “protecting the public” does not involve any kind of balancing exercise. It simply involves safeguarding the public from the danger posed by the prisoner. As Mr Grodzinski puts it, the goal to be achieved is clear, namely the protection of the public; and the means by which it is to be achieved, namely by continued confinement of the prisoner, is equally clear. If the Board concludes that confinement is necessary because there will be a (more than minimal) risk of harm if the prisoner is released, then confinement of the prisoner will be required to avoid that risk.

Secondly, when enacting sections 255B and 255C in the 2008 Act (and re-enacting them in LASPO), Parliament must be taken to have been aware of the May 2004 Directions. Those directions were expressed in language which was materially different from the public protection test. They required the Board to consider “primarily the risk to the public of a further offence being committed when the prisoner would otherwise be in prison and whether any such risk is acceptable”. This risk was to be “balanced against the benefit, both to the public and the offender, of early release back into the community under the degree of supervision which might help rehabilitation and so lessen the risk of re-offending in the future”. The difference in language between these words and the words of the public protection test is striking. The omission of the reference to a balancing exercise must have been deliberate.

Thirdly, I need to refer to Watson. That was a case concerned with a prisoner serving a life sentence. He was released on licence and then recalled to prison under section 39(2) of the 1991 Act. His recall was confirmed by the Board. On a review of his case, applying the public protection test stated in section 34(4)(b), the Board concluded that it was not satisfied that it was no longer necessary for the protection of the public that the prisoner be confined.

The public protection test set out in section 34(4)(b) was the test to be applied by the Board when considering the initial release of a discretionary life sentence prisoner. Section 39(4) prescribed no test which the Board was to apply following a recall. Sir Thomas Bingham MR said at p 916H that it was to be assumed that the same test was applicable at both stages. He added:

“In exercising its practical judgment the board is bound to approach its task under the two sections in the same way, balancing the hardship and injustice of continuing to imprison a man who is unlikely to cause serious injury to the public against the need to protect the public against a man who is not unlikely to cause such injury. In other than a clear case this is bound to be a difficult and very anxious judgment. But in the final balance the board is bound to give preponderant weight to the need to protect innocent members of the public against any significant risk of serious injury. This is the test which section 34(4)(b) prescribes, and I think it is equally appropriate under section 39(4).”

At p 918A, Rose LJ said: “In both cases the need to protect the public is paramount”. Roch LJ said at p 919F:

“The test that the board has to apply in both situations is the same; that the board has to be satisfied that further confinement of the prisoner is no longer necessary for the protection of the public. That this is the test in all cases is consistent with the intention of Parliament as manifested in the language of section 32(6). The test is justified if the protection of innocent people is to be placed above the personal liberty of one who has been guilty of grave offending and who may still represent a danger to others. In my opinion, it is incontrovertible that Parliament in the Act has manifested an intention to put the protection of the public as the over-riding consideration. It follows that in my judgment the board directed itself correctly as to the test it had to apply. ”

Mr Southey submits that the reference by Sir Thomas Bingham to a balancing exercise supports his argument. But Sir Thomas went on to say that the Board was bound to give “preponderant” weight to the protection of the public. Perhaps even more importantly, neither of the two other members of the court referred to a balancing exercise at all: both emphasised the protection of the public. In my view, far from assisting Mr Southey, these dicta provide further support for Mr Grodzinski’s submissions.

Fourthly, Mr Southey’s submission does not reflect the public protection test that had been developed by the courts even before Parliament intervened in the 1991 Act to impose any kind of statutory test. The “life and limb” test which the Divisional Court laid down in Bradley was: is there a more than minimal risk that the prisoner will commit further offences of a violent or sexual nature which will cause serious harm? At p 146F, the court said:

“What it all comes to is this. The Parole Board have to carry out a balancing exercise between the legitimate conflicting interests of both prisoner and public. They must clearly recognise the price which the prisoner personally is paying in order to give proper effect to the interests of public safety. They should recognise too that it is a progressively higher price. Accordingly, the longer the prisoner serves beyond the tariff period, the clearer should be the Parole Board’s perception of public risk to justify continued the deprivation of liberty involved.”

Bradley was approved and followed by this court in R v Parole Board, ex p Wilson [1992] QB 740. Having quoted the passage from Bradley which I have set out, Taylor LJ said at p 747E:

“If that passage means only that the longer a prisoner has remained in custody the more anxiously the Parole Board should scrutinise whether the risk of releasing him is at the unacceptable level, then I would entirely agree with it. The use of the work “clearer” rather than “greater” suggests that is what the court intended. However, Mr. Fitzgerald argued that passage meant that the longer a prisoner has been detained the greater would the risk to the public need to be in order to justify his continued detention. This would import a sliding scale whereby the protection of the public would diminish and taper off in proportion to the length of the prisoner’s detention. In my judgment that cannot be right. The level of risk to the public which the Parole Board regards as unacceptable cannot properly be varied to accommodate the release of a prisoner even if he was jailed in his twenties and has been detained for many years.”

In other words, the public protection test did not permit the Board to balance a risk of harm to the public against other matters. As we have seen, in Watson it was held that the Bradley approach continued to apply following the adoption of a statutory test in the 1991 Act.

This is the background against which Parliament enacted the 1997 Act and then the provisions of the 2003 Act governing IPP prisoners. As the Supreme Court held in Sturnham, Parliament can be taken to have been aware of the approach that the courts had taken to the construction of the public protection test in the 1991 and 1997 Acts and must, by the adoption of the identical statutory language to IPP prisoners, have intended that this approach would also apply to the release of IPP prisoners. LASPO now applies the identical test to the release of various kinds of determinate sentence prisoners. I accept the submission of Mr Grodzinski that, by adopting this test, Parliament must have intended that this approach would apply equally to the release of recalled determinate sentence prisoners.

It follows that the December 2013 Guidance correctly stated that section 255B(3) and section 255C(3) does not require or permit the risk of offending to be balanced against the benefits of release to the prisoner and the community.

For all these reasons, I conclude that the Divisional Court reached the right conclusion for the right reasons. I have seen the judgment of Lord Justice Sales in draft and agree with it. I would dismiss this appeal.

I agree that the appeal should be dismissed for the reasons given by the Master of the Rolls.

In relation to the first issue I think the analysis of section 255B and section 255C which the Master of the Rolls sets out at para. [23] above is of particular importance. When a prisoner is to be released again after recall and before the end of his sentence the release is effected by exercise of the Secretary of State’s power of release in section 255B(2) and section 255C(2), respectively. The Board can only effect early release of a prisoner under section 255B and section 255C by requiring the Secretary of State to exercise her power under subsection (2) in each case, and in each section subsection (3) states that the Secretary of State cannot exercise that power unless the public protection test is satisfied. The Board is given no authority to trump the public protection test which Parliament has stipulated must govern the exercise of the relevant power of early release in subsection (2), so of necessity the Board must satisfy itself that that test will be satisfied before it directs the Secretary of State regarding the exercise of that power.

Mr Southey QC sought to escape the force of this logic by submitting that subsection (5) in each of section 255B and section 255C contained its own freestanding authorisation for the Secretary of State to release a recalled prisoner early which was not conditioned by the restriction in subsection (3). I do not accept this.

I can see that if subsection (5) were a provision which stood by itself one could readily imply into it authorisation for the Secretary of State to act in accordance with a direction given by the Board: a legal requirement to act implies a power to act; ‘ought’ implies ‘can’. But subsection (5) is not a stand-alone provision: in the case of each of section 255B and section 255C it appears in the context of a scheme which spells out expressly in subsection (2) the power of the Secretary of State to release a prisoner early. There is no necessity and no warrant for the implication of a distinct power of release in subsection (5) itself. Mr Southey’s submission depends upon tearing subsection (5) from the specific statutory context in which it appears in each case. In each section the only power of release is the express power in subsection (2), and that power is conditional upon satisfaction of the public protection test in subsection (3).

Sections 255A, 255B and 255C were first introduced into the 2003 Act by amendments contained in the 2008 Act. They were amended by LASPO to reflect the language in the new statutory scheme introduced by that Act. Reference to the wording of subsection (5) of section 255B and section 255C as originally enacted reinforced the true analysis under those sections. In each case, the Board was given power to “recommend” how the Secretary of State should proceed, rather than to issue a direction to her. In my opinion, this language indicates that the Board’s function under the scheme of each section was to make a recommendation as to how the Secretary of State should exercise her power under subsection (2), conditioned as it was by the limitation in subsection (3). As is common ground, there has been no material change of substance as a result of the shift with the enactment of LASPO from the language of “recommendation” to the language of “direction”.

Finally, I would like to register that apart from the question of interpretation of the public protection test addressed in the Master of the Rolls’ discussion of the second issue on this appeal, the precise content of the statutory public protection test was not the subject of debate before us. It is not obvious to me why the Board employs the “life and limb” approach when applying the statutory test. On the face of it, the public might require protection if, for example, an incorrigible fraudster were released early in circumstances where there was a significant risk he would again prey upon the public, even though he represented no threat to life and limb. I express no view about this aspect of the Board’s guidance because it was not in issue before us.

I agree with both judgments.