England and Wales High Court (Administrative Court) Decisions

IN THE HIGH COURT OF JUSTICE
QUEEN'S BENCH DIVISION
ADMINISTRATIVE COURT

		Royal Courts of Justice Strand, London, WC2A 2LL
		19th December 2014

MR JUSTICE JAY:

Introduction

1. The Claimant is a national of Algeria. He came to the United Kingdom on the 15^th February 1998, aged only 15. He was granted exceptional leave to remain, on account of his youth, but that expired on 10^th July 2004. The Claimant has accumulated numerous criminal convictions, and spent periods of time in custody. On 10^th September 2007 the Defendant deemed that the Claimant's presence in the UK was not conducive to the public good, on account of his criminal record, and signed a deportation order against him. Subsequently, the Claimant received further criminal convictions leading to imprisonment, interspersed with periods of immigration detention.
2. On 20^th November 2012 the Claimant was convicted at Central Magistrates' Court of the offence of theft, and he was sentenced to 3 months 20 days' imprisonment. The custodial part of that sentence was completed on 14^th January 2013, and the Claimant was automatically on licence for the remaining period, having served half his sentence. However, he remained in detention at HMP Wandsworth under paragraph 2 (3) of Schedule 3 to the Immigration Act 1971, being the subject of a deportation order. Save for short periods when he was detained at Immigration Removal Centres ("IRCs"), the Claimant was held at HMP Wandsworth under immigration powers until 7^th November 2013, when he was transferred to HMP Wormwood Scrubs. During this period he made a number of bail applications, all of which were refused.
3. Eventually, the Claimant was transferred to Harmondsworth IRC on 21^st March 2014. He was released on immigration bail on 31^st July 2014. The Claimant has not been removed to Algeria, as he should have been a long time ago, owing to difficulties with his documentation. The Claimant asserts that he wishes to return to Algeria, and the authorities there are therefore respectfully encouraged to facilitate his wishes.
4. These judicial review proceedings, as originally constituted, sought to assail the lawfulness of the Claimant's detention after 14^th January 2013. There were two limbs to the challenge. By the first limb, the Claimant contended that he should not have been subject to administrative detention at all, because there was no prospect of his removal from the United Kingdom within a reasonable period. By his second limb, the Claimant contends that between 14^th January 2013 and 21^st March 2014 he should have been held in an IRC rather than a prison.
5. When these proceedings were last before the Court on 13^th May 2014, Holman J dismissed the claim on the first limb. He adjourned the second limb for future determination. At that point, it was understood that this issue would be addressed by the Administrative Court in another case listed for hearing in July 2014. However, it is clear from the transcript of the decision of HHJ Clive Heaton QC in Lemtelsi v SSHD [2014] EWHC 2750 (Admin) that the issue was not in fact determined. It therefore falls to me to rule on this important issue: I understand that at least one other claim is awaiting the outcome of these proceedings.
6. Although the Claimant is no longer in detention, the point he raises in these proceedings is far from academic. He seeks damages under section 8 of the HRA 1998 read in conjunction with Article 5(1) of the ECHR to reflect his unlawful detention in prison, rather than in a detention centre.
7. I mentioned that the Claimant's argument in relation to his second limb covers the period 14^th January 2013 to 21^st March 2014. However, it became clear during oral argument that the policy which the Claimant seeks to challenge was only applied to preclude his transfer to an IRC on and after 3^rd July 2013. I understand it to be common ground between the parties that the application of the impugned policy was the reason why the Claimant remained in prison from that date until 21^st March 2014.

Defendant's Policy Regarding the Place of Detention of Immigration Detainees

8. The Claimant is someone whom the Defendant chooses to call a "Time-Served Foreign National Offender" (a "TSFNO"). The Defendant signed a deportation order in this case owing to his criminal record. From the expiry of his last prison sentence, he was "time-served" and remained in detention under administrative powers, rather than in pursuance of any sentence of the court. In this sense, the Claimant is distinguishable from the majority of immigration detainees, even those subject to deportation orders, because they have not served sentences of youth detention or imprisonment.
9. Before January 2012, it was the Defendant's policy that TSFNOs in the Claimant's position should normally be held in IRCs, rather than in one of Her Majesty's prisons, unless there was some specific reason justifying incarceration.
10. Taking these policies in chronological order:-

(1) Rule 3 of the Detention Centre Rules 2001 provides:-

"(1) The purpose of Detention Centres shall be to provide for the secure but humane accommodation of detained persons in a relaxed regime with as much freedom of movement and association as possible, consistent with maintaining a safe and secure environment, and to encourage and assist detained persons to make the most productive use of their time, whilst respecting in particular their right to individual expression.

(2) Due recognition will be given at Detention Centres to the need for awareness of the particular anxieties to which detained persons may be subject and the sensitivities that this will require, especially when handling issues of cultural diversity."

(2) Section 10.1 of Chapter 55.10 of the Defendant's Enforcement Instructions and Guidance ("EIG"), promulgated on 26^th October 2010, provided (I use the past tense because it has been revised):-

"Immigration detainees should only be held in prison establishments when they present risk factors that indicate they pose a serious risk to the stability of Immigration Removal Centres or to the safety of others being held there.

Detainees moving from the Prison Estate into the IRC Estate will undergo an individual risk assessment. The existence of any of the following risk factors indicates that the detainee should be held in prison accommodation rather than an IRC but the list is not exhaustive … national security … criminality (serious offences) … behaviour during custody … security …control … health grounds."

(3) Clause 12.4 of the Service Level Agreement ("SLA") between the National Offender Management Service ("NOMS") (part of the Ministry of Justice) and what at that time was known as the United Kingdom Border Agency ("UKBA") (part of the Home Office), 2011-15, promulgated on 15^th June 2011 provides:-

"Immigration detainees should generally only be held in a prison when they present specific risk factors that indicate they pose a serious risk of harm to the public or for the good order of an Immigration Removal Centre, including the safety of staff and other detainees, which cannot be managed within the regime applied in IRCs. This regime derives from Detention Centre Rules and provides greater freedom of movement and less supervision than prisons, as well as access to the Internet and mobile telephones."

(4) Paragraph 2.68 of Prison Service Instruction ("PSI") 52/201, issued on 4^th November 2011, was in like effect to (3).

11. These policies recognised differences between the prison regime on the one hand and immigration detention on the other. The former, speaking generally, is concerned with punishment, deterrence, rehabilitation and reform (see the general purposes of sentencing set out in section 142 of the Criminal Justice Act 2003), the latter with the administrative detention of those liable to be removed from the United Kingdom who are assessed to be a real risk of absconding (the Defendant has detailed policies governing the use of immigration detention). Ordinarily an illegal entrant, or a person refused leave to enter, or an overstayer subject to a deportation order will not be detained under immigration powers at all (save possibly as an immediate prelude to removal). However, owing to the risk of absconding, a limited number must be, but – in the ordinary course – IRCs are designed to cater for them. The Defendant's previous policies recognised that, on specific grounds, it may be appropriate for such persons to be incarcerated rather than detained. I do not understand Mr Graham Denholm for the Claimant to be contending that the old regime was unlawful.
12. Immigration detainees held in prison are categorised by the Defendant as "unconvicted prisoners". Their status is more fully explained by Prison Service Order ("PSO") 4600 issued on 10^th February 2003. By paragraph 1.1:-

"Unconvicted prisoners have not been tried and are presumed to be innocent; the Prison Service's sole function is to hold them in readiness for their next appearance at Court. Their imprisonment should not deprive them of any of their normal rights and freedoms as citizens, except where this is an inevitable consequence of imprisonment, or the Court's reason for ordering their detention and to ensure the good order of the prison. Instructions or practices that limit their activities must provide only for the minimum restriction necessary in the interests of security, efficient administration, good order and discipline and for the welfare and safety of all prisoners." [emphasis as in the original]

Looking at the definition of "unconvicted prisoners" in Annex A of PSO/4600, it is clear that those held under immigration powers form part of a much wider category of persons who are not subject to formal terms of imprisonment. This category includes individuals held on remand awaiting trial. Remand prisoners are not held in specially designated institutions, but to my knowledge no one has sought to argue that Article 5 issues arise in their context.

13. The Defendant's previous policy reflected international standards. For example, in Adela Bero v Regierungspräsidium Kassel (C-473/13), CJEU 30^th April 2014, Advocate General Bot said this:-

"43[…] prisons, which are either penal establishments answering specific purposes related to the very notion of penalty, or remand centres, are not to be confused with the specialised detention centres provided for by the Directive. A person is held in prison in two cases only, either before being tried or in order for them to serve a criminal penalty, each of those cases being part of a procedure attaching to a serious criminal offence."

14. According to Article 17 (1) of the International Convention on the Protection of all Migrant Workers and Members of their Families (1990):-

"(1) Migrant workers and members of their families who are deprived of their liberty shall be treated with humanity and with respect for the inherent dignity of the human person and for their cultural identity.

(3) Any migrant worker or member of their family who is detained … shall be held, insofar as practicable, separately from convicted persons or persons detained pending trial."

15. By Council of Europe (PACE), Resolution 1707 (2010), 15 European Rules Governing Minimum Standards of Conditions of Detention for Migrants and Asylum Seekers:-

"Detainees shall be accommodated in centres specially designed for the purpose of immigration detention and not in prisons (paragraph 9.2.2).

The material conditions shall be appropriate to the individual's legal and factual situation (paragraph 9.2.5)

The detention regime must be appropriate to the individual's legal and factual situation (paragraph 9.2.6)"

16. According to the European Committee for the Prevention of Torture and Inhumane or Degrading Treatment or Punishment ("CPT"), CPT Standards 2011, page 65:-

"Where it is deemed necessary to deprive persons of their liberty for an extended period under aliens' legislation, they should be accommodated in centres specifically designed for that purpose, offering material conditions and a regime appropriate to their legal situation and staffed by suitable qualified personnel. Obviously, such centres should provide accommodation which is adequately-furnished, clean and in a good state of repair, and which offers sufficient living space for the numbers involved. Further, care should be taken in the design and layout of the premises to avoid as far as possible the impression of a carceral environment. As regards regime activities, they should include outdoor exercise, access to a day room and to radio/television and newspapers/magazines, as well as other appropriate means of recreation (e.g. board games, table tennis). The longer the period for which persons are detained, the more developed should be the activities which are offered to them." [emphasis as in the original]

17. Mr Denholm referred me to other authoritative statements to similar effect.
18. These materials do not provide definitive answers to the key questions arising in this judicial review application, but they are illuminating, particularly since the Defendant's previous policy was aligned to those standards. The key questions arising here are, in essence, whether the Defendant was entitled in public law terms to implement a different policy and/or whether that policy conflicts with Article 5 of the ECHR. But the answers to those questions are not to be found solely in these international materials. I have to consider all of the available evidence, and the legal submissions upon them.
19. The Defendant's policy changed on 24^th January 2012, in the form of version 13 of Chapter 55 of the EIG. There have been further versions since then but the relevant wording has remained the same.
20. The relevant parts of Chapter 55.10.1 ("Criteria for Detention in Prison") provide:-

"NOMS and the Home Office have a Service Level Agreement governing the provision of bed spaces within prisons. Under that agreement, NOMS make a number of bed spaces available for use by the Home Office to hold immigration detainees. It is for the Home Office to determine how those bed spaces are used and the type of detainees who are held in them.

"The normal expectation is that the prison beds made available by NOMS will be used to hold TSFNOs before any consideration is given to transferring such individuals to the IRC estate. This position will apply if there are free spaces among the beds provided by NOMS and even if the criteria or risk factors outlined below are not presented by the FNOs concerned. More generally, decisions to allocate specific detainees, whether TSFNOs or otherwise, to prison accommodation will be based on the presence of one or more of the risk factors or criteria below. "

[the risk factors are the same as before: see paragraph 10(2) above]

…

"(note: the existence of any of the above risk factors indicates that a detainee should be held in prison accommodation rather than an IRC but the list is not exhaustive and DEPMU staff should also satisfy themselves that no other risks exist which would make it inappropriate for the detainees to be held in an IRC, rather than a prison.)

"The normal expectation is that any remaining prison bed spaces made available under the agreement with NOMS after allocation of prison beds to individuals, presenting one or more of the criteria or risk factors above, will be filled by TSFNOs not falling into the above categories. Subject to risk assessment, such individuals will be placed on a waiting list, operated by DEPMU, for transfer to an IRC but will remain in prison accommodation pending that transfer.

"The transfer of such individuals to IRCs will take place only where the prison beds they are occupying are required either by individuals (FNO or otherwise) falling into one or more of the categories above or by more recently detained TSFNOs (that is, FNOs detained under immigration powers on completion of or released from custodial sentence). In the absence of the criteria or risk factors set out above, the length of time that an FNO has been held in a prison bed solely as an immigration detainee will be the main factor in deciding when to transfer to an IRC. In other words, priority for transfer to an IRC will be given to those FNOs who have been held in prison beds the longest.

"Separately from the use of the prison beds made available to the Home Office under the agreement with NOMS, and in the interests of maintaining security and control in the Home Office Detention Estate as a whole, a cap is placed on the total number of TSFNOs who may be held in the Detention Estate at any one time. The cap may also be used as part of the day to day management of the Home Office Detention Estate in order to meet changing operational priorities for the use of IRC beds, which will have a consequence for the number of beds that will be available for allocation to TSFNOs at any one time … where the current level of the cap is reached, TSFNOs will continue to be held in prison accommodation, even in the event that the prison bed spaces made specifically available to the Home Office by NOMS are full: the expectation in such circumstances is that additional bed spaces would be sought from NOMS.

"If transfer to an IRC is agreed, it should be effected as soon as reasonably practicable. Reasons for deciding not to transfer an individual must be recorded, as must the reasons for any delay in affecting agreed transfers.

Any individual may request a transfer from prison accommodation to an IRC. Prompt and evidence based consideration must be given to such a request and, if rejected by DEPMU, the individual concerned will be given written reasons for this decision…"

21. The new policy alters the emphasis altogether. Put at its lowest from the Claimant's perspective, the general rule is that TSFNOs will be held in prison unless the beds they are occupying are required by incoming detainees, in which circumstances they may be transferred to an IRC on a "first in first out" basis. Furthermore, the Defendant expressly contemplates that a situation may arise - even "where the current level of the cap is reached" - where TSFNOs would continue to be held in prison accommodation: in such circumstances additional bed spaces would be sought from NOMS. As will soon be seen, this is exactly what happened in practice. Putting the matter at its highest from the Claimant's perspective, this was a blanket policy that TSFNOs in this Claimant's position should remain in prison until they came to the very end of this notional pipeline, regardless of their personal circumstances.
22. The new policy is quite fluid in its terms, and without knowing at all material times the availability of beds in the prison estate and in the IRC estate, and the number of existing and incoming immigration detainees, including TSNFOs, it would be difficult to predict exactly how the policy might impact on any given case. Once the number of bed spaces in the prison estate was raised in the autumn of 2012 (as to which, see further below), these predictive difficulties subsided.
23. The reasons for the change in policy were investigated during the course of these proceedings. The SLA for 2011-15, comprising non-legally binding arrangements between NOMS and UKBA, came into being on 15^th June 2011, but it underwent a period of gestation. It is clear from intra- and inter-departmental communications that the intention was to align the EIG with the SLA. The relevant provisions of the SLA, apart from Clause 12.4 (set out under paragraph 10(3) above), are:

"13.1 Section 12 above sets out the circumstances in which immigration detainees will be held in prisons and this SLA recognises that the number of detainees actually held there is subject to variability. The structure for payments will reflect this by setting a payment per place for an agreed minimum number of places to be provided across the year and a higher payment for places beyond this.

"For the financial year 2011/12, NOMS will make available 600 places in prisons for holding immigration detainees at £x per place per year.

"Payment for places beyond this will be at an annual rate of £x supplied monthly and based on the average number of immigration detainees in prison that month (as recorded by national operations group from establishments' annex 1 returns).

…

"13.11 Any agreement to extend these arrangements to future years (and, in doing so, any agreement to vary the number of places being made available or applicable payments) must be recorded in writing."

24. Mr Denholm relied heavily on the email communications which the Defendant has fairly disclosed pursuant to its duty of candour in judicial review proceedings. I shall focus on the highlights: -

"UKBA email dated 9^th May 2011

… This is why chapter 55 will need to be changed so that it reads the same as the SLA ... we don't want to put specific numbers in because these can and do frequently change … it's not a case of making the SLA match Chapter 55 it's the other way round … we have been holding TSFNOs in prison for purely immigration reasons for years while they are either protocol cases or awaiting a place in the IRCs …"

"Email 9^th May 2011, probably UKBA

[The SLA] doesn't make it clear that the 600 beds would not be used for "ordinary" immigration detainees (i.e. not FNOs or those whose behaviour is such that they can't be safely accommodated in IRCs). I understand that Ministers gave a commitment in 2002 to end the routine use of prison detention for immigration purposes. Any change to this position would require us to go back to Ministers for approval (notwithstanding this was a commitment given under the previous administration), so we ought to make clear in the SLA that we are not backtracking on that position …"

"UKBA email dated 11^th May 2011

… we will be trying to keep the numbers of TSFNOs in the prison estate as close to the 600 as we can … the process will be that as the FNO finishes their sentence they will be risk assessed by DEPMU for suitability to come into our estate and those that are risk assessed as suitable will be put on a list and brought across into our estate in as close to list order as we can, given their geographical location and transport limitations etc.

At this moment in time we have approximately 900 males (our full capacity) in our estate and approximately 550 in the prison estate and so the additional length of time that detainees will have to wait (given we bring around 110 detainees out of prisons each week) will not be more than a couple of days."

"UKBA email dated 24^th May 2011

In essence, what we are seeking to do is to change completely our approach to bed space allocation.

Instead of us saying that all FNOs are eligible for a transfer to an IRC bed unless their personal risk profile dictates that they are not suitable or we have reached our numerical cap, we are saying that all FNOs are liable to remain in a prison bed irrespective of their personal risk profile unless we have filled the quota of beds allocated to us by NOMS (as you say currently set at 600) and their risk profile dictates that they can come across.

It is foreseeable that we will have empty beds in our estate which under current arrangements FNOs might fill, but in the future they will not. These beds will be provided to the rest of the business to use for the removal of failed asylum seekers, overstayers, etc.

I don't think we should be afraid to say what we are doing or why we are doing it – indeed we would be criticised severely if we were to be seen to operate some form or clandestine policy or allocation criteria.

As I say, lawyers should look over this carefully before we publish."

"UKBA email dated 14^th September 2011

The revised criteria in 55.10.1 for allocation to prison beds now reflect those agreed with NOMS in the SLA. In addition, I've tried to set out the basis on which the 600 beds are expected to be used. The expectation is that, as we are paying for the beds whether used or not, they will be kept as full as possible at all times. This will be achieved in the large part, simply through operation of the allocation criteria for particular individuals, whether FNOs or not. Any remaining beds (on current figures, likely to be in the region of 150-200) will be filled by managing the flow of FNOs from prisons to IRCs, with individuals effectively being held back in prison until the pressure of more recently TSFNOs pushes them across to our estate (subject, of course, to the overall cap on FNOs in our estate)."

25. Mr Denholm criticised the Defendant's officials for failing to seek Ministerial approval for the new regime. I do not know whether such approval was sought or not (on one interpretation of the document mentioned under paragraph 37 below, it was), but it is quite clear to me that it ought to have been. Nonetheless, Mr Denholm's point does not constitute a judicial review ground.
26. In my judgment, two substantive points emerge from a proper consideration of these materials. First, the real reason for the change in policy was nothing to do with assessment of risk or the interests of immigration detainees, but everything to do with administrative convenience. The overall pressures on the system, both physical and financial, conspired to create a state of affairs whereby UKBA needed to purchase a number of bed spaces from NOMS and, having done so, those bed spaces needed to be kept as full as possible. Secondly, the SLA and the EIG are now misaligned. Clause 12.4 of the SLA reflects the previous policy, and also reflects the range of other materials which I have previously mentioned. The EIG may have been intended to "track" the SLA, rather than the other way round, but it conspicuously failed to do so. Version 13 of the EIG promulgated on 24 January 2012 matched the SLA in numerous respects, but clashed with clause 12.4. Ultimately, though, nothing specifically turns on this, because it is clear from all the available evidence that the policy which was applied to the Claimant, and others in like position, was that contained in the new EIG, not in clause 12.4 of the SLA. To my mind, it does not matter that there is a mismatch; the only real judicial review point concerns the lawfulness of the new policy properly understood.
27. By the autumn of 2012 it became clear that 600 beds were insufficient, and so consideration was given to increasing the availability of beds in the prison estate. The evidence relating to that consideration is sparse. The best evidence bearing on the raising of available places is contained in UKBA's email dated 1^st November 2012, the relevant part of which provides: -

"FOR ACTION: Use of Detention Beds

Colleagues (including those in NOMS)

In order to temporarily meet the increasing demand for detention space I am today making a number of decisions.

1. DEMAND: To increase our use of NOMS beds within the terms of the SLA for the remainder of this financial year up to a maximum of 1,000 …

2. RESOURCES & STRATEGY: … ensure that the UKBA Board can make a decision on the possible transfer of NOMS estate to UKBA. This is reasonable pressing [sic] as I know NOMS wants an answer and operational [sic] we could be at demand risk.

3. EFFECTIVENESS: we need to use our existing detention space more effectively, I will ask … to write to our enforcement teams to ensure better review of medium and long-term detainee [sic] to really check the likelihood of removal."

28. On 11^th February 2014 NOMS emailed all deputy directors of custody as follows:

"As part of the SLA for detention services provided by NOMS for the Home Office, NOMS are committed to making available 600 places in prisons for holding immigration detainees. Towards the end of 2012 it was agreed, at the request of the Home Office, that NOMS would make available a further 400 places, bringing the total to 1000 places in prisons for holding immigration detainees.

We have since been informed by the Home Office that the number of prison places required for immigration detainees will revert back to 600."

29. Mr Denholm made a series of submissions directed to the raising of the cap, which policy the Defendant chose to label "Operation 1000". In essence, he submitted that the Defendant was operating a separate, indeed secret, policy outside the terms of the SLA. Mr Denholm devoted considerable attention, both in writing and orally, to this submission. I indicated during the course of his oral argument that I was not remotely attracted by these submissions. The SLA clearly contemplated that the 600 figure was not writ in stone, and might be exceeded by agreement between NOMS and UKBA. The EIG was in similar terms. It is true that the SLA contemplated that any variation to it would be in writing, but these are not commercial agreements and both "parties" to the SLA (I use inverted commas because the Crown is indivisible) proceeded on the basis that there should be 400 additional beds between November 2012 and February 2014. These beds were paid for. Accordingly, all that has happened here is that the figure specified in the SLA as being the starting point was raised by agreement from 600 to 1,000 beds. That in itself raises no arguable judicial review issue.
30. However, I do take Mr Denholm's further point that raising the threshold or cap had a contingent or knock on effect on those in the Claimant's position. The increase in the numbers, in the events which happened, meant that it took much longer for people in the Claimant's position to work their way through this "pipeline" and arrive at the top of the queue. Although the Claimant had been risk assessed for a transfer in July 2013, I infer that he did not reach the top of the queue until March 2014, and that was the reason for his transfer to an IRC (there is some suggestion that the Claimant was suffering from mental health problems, but whether those were the reason for his transfer is unclear). This point is made clearer when consideration is given to the Minister of State's letter referred to under paragraph 35 below.
31. The precise number of immigration detainees held in the prison estate over the relevant period is not altogether clear, but the evidence seems to me to be as follows. Between October 2012 and November 2013 the numbers fluctuated between 552 and 959. In December 2013 there appear to have been 1,214 immigration detainees held in prison notwithstanding that the cap had been exceeded. A notional health warning must be placed against that figure since there is also evidence that on 31^st December 2013 there were only 850 immigration detainees in this position. In 2014 the numbers varied between 562 and 863.
32. I have already observed that put at its lowest, version 13 of chapter 55 of the EIG, promulgated on 24^th January 2012, established a general rule that those in the Claimant's position should remain in prison. I need to address the evidence bearing on Mr Denholm's higher submission that version 13 did more than that; it created a blanket policy. Mr Mathew Gullick for the Defendant strongly disputes this inference, pointing to the wording of the EIG – "normal expectation" – and the reference to consideration of individual cases, with the need to give reasons in writing. Mr Denholm invites me to examine the matter more closely, and consider how the new policy was applied in practice. I propose to take up his invitation.
33. The Claimant's pre-action protocol letter dated 22^nd November 2013 stated: "the current policy is unlawful as it presumes that immigration detainees will be held post sentence in prisons and is applied absent individual consideration." An executive officer within the Defendant replied to Bhatt Murphy's letter on 11^th December 2013, and addressed the foregoing point in these terms: -

"At the present time, the Home Office is limiting transfers for only those detainees who have removal directions in place as we have not used all the bed spaces available to us in the NOMS estate. It is for this reason that your client will currently remain in the NOMS estate at this time unless there is a significant change to his medical wellbeing."

34. Thus, the Defendant was making clear that a blanket policy was operating against the Claimant save to the extent that he might be able to show a clear medical reason justifying a transfer to an IRC.
35. Previously, on 20^th May 2013 the Minister for Immigration, Mr Mark Harper MP, had written to the director of the Association of Visitors to Immigration Detainees in these terms:

"In the autumn of 2012, a decision was reached to halt temporarily the transfer of TSFNOs into immigration detention. This agreement was reached to assist with capacity issues within the IRCs. The decision to temporarily halt transfers came at a time when there were around 600 TSFNOs in the prison estate, with around 20 prisoners becoming time expired each week. It was agreed with the NOMS that the transfers would be halted temporarily, until the numbers of TSFNO reached 1,000. TSFNOs continue to be transferred to the detention estate for removals and court appearances, but these are exceptions to the current agreement".

36. The terms of this letter are clear. I do not interpret the Minister as saying that there could be absolutely no circumstances in which a TSFNO could be transferred to an IRC, but it is plain that the policy was being applied in a fairly mechanistic way. Furthermore, the Minister was recognising that the effect of increasing the number of beds within the prison estate would, and did, have the practical consequences he has specified. Thus, it was not the new EIG per se which was the problem; it was the new EIG applied in conjunction with "Operation 1000".
37. The Claimant has filed evidence from solicitors and other practitioners with experience in this area (see paragraph 49 of Mr Denholm's skeleton argument) to the effect that the policy was applied in practice without appropriate flexibility. Mr Denholm's skeleton argument also drew my attention to an email from NOMS to the Head of Detention Operations, UKBA, dated 18^th March 2014. This email was not discussed during oral argument, and it is open to at least one interpretation. My preferred interpretation is that the author of the email is referring to a submission – it is unclear to whom, but possibly a Minister – which was probably sent in the autumn of 2012. Paragraph 5 of the submission provided:-

"The additional detainees will be generated by suspending transfers out of prisons into IRCs. UKBA have today instructed their managers to take this action, although transfers for imminent removal from the UK will be maintained as normal. This means that the population in prisons will grow gradually as foreign national prisoners become time-served and are detained under immigration powers. We anticipate that the number of detainees in prisons (552 on Monday 29^th October [2012]) will increase by between 30-50 per week until the total number is around 1,000".

This email provides strong support for Mr Denholm's submission that there was a blanket policy operating in practice.

38. I asked Mr Gullick if there was any evidence in the voluminous bundles before me (running to thousands of pages – excessive for an application of this nature) demonstrating that the Defendant had given individual consideration to cases on an exceptional basis. I was told that there was none. It is clear from a consideration of the Claimant's own case that he was moved to IRCs for the purposes of bail hearings and to be interviewed by the Algerian Embassy. Thus, it would be incorrect to say that the policy permitted of absolutely no exceptions. Nonetheless, it would be fair to say that operational considerations meant that the Claimant had to be moved on a purely temporary basis; this is different from the sort of exceptional case I had in mind when I posed my question to Mr Gullick. I was contemplating permanent transfers, subject always to a subsequent grant of bail, removal from the United Kingdom, or transfer back into the prison estate, having regard to a detainee's individual circumstances.
39. On 4^th September 2013 the Claimant requested a transfer to an IRC, and for full reasons in the event that the Defendant was unwilling to transfer him. I am not aware of any document evidencing the furnishing of such reasons. Instead, there are internal emails expressing concerns for the Claimant's safety in prison, and he remained assessed as suitable for transfer to an IRC. As I have said, there is also some evidence that the Claimant was experiencing mental health difficulties. Although these matters were appreciated and understood by the Defendant, they appear to have had no impact on the place and conditions of his detention. The Claimant remained in prison because that is what the policy provided.
40. In the light of the Director's email to BID dated 30^th September 2013, Mr Gullick's failure to point to the Defendant giving individual consideration to cases on an exceptional basis was unsurprising. BID asked the Director to provide clarification of the type of circumstances under which transfers to IRCs might take place. According to this email:

"There have been some individuals who have been moved to an IRC for a short period, for example to ensure that they have been able to attend appeal hearings where the prison timings would prevent this. These individuals will then be moved back to a prison bed following their court hearing."

41. Given the way in which the Claimant's case was advanced from the outset, I would have expected the Defendant to place evidence before the Court that it had given consideration to transfers to IRCs on an exceptional basis – if such evidence existed. In a case of this nature, having regard to the weight of all the evidence ranged against the Defendant, it is insufficient merely to draw attention to the possibility of transfer under the terms of the revised EIG. I infer that there is no such evidence. Given the existence of medical facilities in prisons, the reference to "medical wellbeing" in the executive officer's email is somewhat theoretical, and is not supported elsewhere. That inference is fortified by all the material to which Mr Denholm refers, the most compelling parts of which I have summarised. In practice, therefore, I conclude that the Defendant did operate a blanket policy which permitted of no exceptions.

The Claimant's First Ground: Public Law Error

42. I have renumbered, and distilled, the Claimant's grounds to reflect the ebb and flow of the oral argument.
43. On my understanding of his case as it became refined during the course of the hearing, Mr Denholm adhered to two principal submissions under the rubric of his public law ground. First, he submitted that it was unlawful for the Defendant to implement a policy which eschewed any individualised risk assessment, and mandated incarceration within the prison estate until certain thresholds were exceeded and the head of a queue was reached. Secondly, Mr Denholm contended that it was unlawful for the Defendant to operate a blanket policy which ignored the circumstances of any particular case.
44. Given my finding that the Defendant did operate a blanket policy, it follows that the second submission is made out (see R(Lumba) v SSHD [2012] 1 AC 245, paragraphs 40-55). Where that leaves the Claimant in the context of this case is another matter. Not merely has the Claimant been released from prison, he is no longer detained under immigration powers at all. Moreover, the Defendant made material changes to her policy in February 2014 which - without more evidence as to the current position – suggest that my finding may well be academic.
45. Mr Denholm's first submission is more difficult. I asked Mr Gullick to explain the reason for the amendments to chapter 55 of the EIG made in January 2012. He said, in line with paragraph 10 of his skeleton argument:-

"The Claimant's place of detention is linked to the grounds for his detention; he was a TSFNO who was serving a sentence of imprisonment and had immediately upon expiry of the custodial part of that sentence been detained in prison pending his deportation, which was taking place as a consequence of his criminal offending".

46. As I have already observed, Mr Gullick also referred to the Claimant's status as an "unconvicted prisoner", and in addition the Defendant's letter dated 11^th December 2013 makes the point that under the Immigration (Places of Detention) Direction 2008, prisons are designated places of detention for immigration detainees. Even so, in my view the Defendant cannot circumvent the substantive issues which arise by deploying labels and nomenclature which suit.
47. The key question which arises is whether paragraph 10 of Mr Gullick's skeleton argument sets out a rational basis for holding the Claimant in prison conditions. In my judgment, it does not. It is quite correct that the Claimant's deportation was taking place in consequence of his criminal offending, and the Defendant was entitled to detain the Claimant in consequence of her deportation order pending the Claimant's removal. It is also both correct, and trite, to observe that criminal offenders subject to sentences of imprisonment will be held in Her Majesty's prisons. But it does not follow from all or any of the above that those whose custodial terms have expired should be held in prison. That would require a finding not merely that such persons ought to be detained, but that the risks are such that detention in a designated facility is inappropriate. In the absence of such a finding, there would appear to be no link between the Claimant's place of detention in prison, qua TSFNO, and the grounds for his detention.
48. More specifically, the grounds for the Claimant's detention are that the conditions expressly or impliedly contained in Schedule 3 to the Immigration Act 1971 are satisfied. It may well be that possession of a criminal record bears on the absconding risk, but that – without more - has no logical connection with the place, as opposed to the fact, of detention. The only justification put forward by the Defendant is, in reality, that the Claimant has been someone subject to several terms of imprisonment. That does not ipso facto make him an appropriate candidate for incarceration, and the Defendant accepts that the Claimant did not require imprisonment after 3^rd July 2013. Thus, it is not being said that the exercise of the Schedule 3 power required incarceration: all that it required was detention. In my judgment, the Defendant's reasoning is based on a frank non-sequitur and must, therefore, be rejected.
49. I say nothing about whether the Defendant could devise a lawful policy which entails holding immigration detainees in prisons. It is clear that there are circumstances in which such persons could, and should, be held in a more secure environment. Exactly how a policy might lawfully be formulated which achieves those legitimate objectives is for those advising the Defendant and not for this Court. My decision is limited to the finding that the reason the Defendant has advanced for holding the Claimant in prison conditions after 3^rd July 2013 cannot be upheld. Furthermore, and for the avoidance of doubt, my finding is closely tethered to the increase in the threshold from 600 beds to 1,000 beds in the autumn of 2012, with the practical consequences I have explained.
50. Given that the Claimant is no longer subject to the Defendant's policy, the Claimant's public law ground brings him marginal or no practical benefit. But, the conclusions I have reached are capable of being relevant to the Article 5 issue, which for clear practical reasons seems to me to be the most important issue in these judicial review proceedings. Mr Denholm urged me to reflect my finding of public law error in the form of declaratory relief, but in the circumstances of this case that, in my judgment, would be judicial overkill. There is no need for a declaration; the matter is historical – certainly as regards this Claimant, and possibly for all purposes, given the change of policy in February 2014. The justice of the matter is met by the Claimant being able to read the terms of the narrative judgment that I have given.

Ground 2: Article 5

51. This ground sub-divides into a series of issues, which I now set out in what I believe to be their correct logical order:-

(i) whether my finding that the Defendant has perpetrated public law error is sufficient to justify a breach of Article 5.

(ii) if not, whether the Claimant's detention in prison, as opposed to an IRC, was "arbitrary" for Article 5 purposes because there is no link between the ground or reason for the Claimant's detention, and its location and conditions.

(iii) if so, whether binding Court of Appeal authority precludes me from finding a breach of Article 5 on that basis (or on the public law basis referred to under (i) above).

(iv) if so, whether in any event the Claimant's incarceration was "unduly harsh" so as to found a breach of Article 5.

Issue 1 – Public Law Error and Article 5

52. Mr Denholm accepts that a public law error, without more, cannot give rise to a claim at common law. As Lord Dyson JSC pointed out, at paragraph 68 of his judgment, in Lumba:-

"I do not consider that these arguments undermine what I have referred to as the correct and principled approach. As regards Mr Beloff's first point, the error must be one which is material in public law terms. It is not every breach of public law that is sufficient to give rise to a cause of action in false imprisonment. In the present context, the breach of public law must bear on and be relevant to the decision to detain. Thus, for example, a decision to detain made by an official of a different grade from that specified in a detention policy would not found a claim in false imprisonment. Nor too would a decision to detain a person under conditions different from those described in the policy. Errors of this kind do not bear on the decision to detain. They are not capable of affecting a decision to detain or not to detain".

53. The Claimant relies heavily upon the decision of Bean J in R(Luis Rozo-Hermida) v SSHD [2011] EWHC 695 (Admin) in support of the proposition that it is sufficient for Article 5 purposes to demonstrate (as I have found) a public law error in his case.
54. In Rozo-Hermida Bean J held that the decision to hold the Claimant in prison on a fact-specific basis under the old policy was Wednesbury unreasonable. At paragraph 24 of his judgment he said this:-

"Whether the matter is put on the basis of Article 5 or of domestic public law, in my judgment the consideration of the Claimant's case both on 12^th August 2010 and thereafter was flawed. Much of the reasoning of the decision letter of 12^th August 2010 is unassailable, but in two respects it is plainly wrong".

55. On the basis of that conclusion, paragraph 4 of Bean J's Order reflected that there had been a finding that Article 5 had been breached, and the question of assessment of damages was remitted to a Master of the Queen's Bench Division for determination. It seems clear that Bean J proceeded on the basis that a flaw in the Defendant's decision-making process was sufficient to establish a breach of Article 5. It is unclear whether there was any contrary argument on the point.
56. Arguably, the facts of the instant case are more serious than Rozo-Hermida. I have found that the reasoning the Defendant has advanced for holding the Claimant in prison conditions after 3^rd July 2013 was fundamentally flawed. Moreover, that flaw was not case-specific, it was systemic. No doubt mindful of the test applicable to Article 5 in this context (as more fully explained below), the Defendant has sought to identify a link between the Claimant's immigration status and the place and conditions of his incarceration (believing, correctly, that this was incumbent on her), but in my judgment she has failed to do so. Judging the matter according to the Defendant's own lights, it would follow that there was no such link.
57. Plainly, Rozo-Hermida avails Mr Denholm's argument, but in my respectful view it was too broadly expressed. The consistent jurisprudence of the ECtHR is that a violation of the substantive and procedural rules of national law is insufficient to found a breach of Article 5. It does not appear that the relevant ECtHR cases were cited to Bean J. Furthermore, there is binding Court of Appeal authority (discussed further below) which post-dates and is inconsistent with Rozo-Hermida, and applies even where the Defendant accepts that there was no basis to incarcerate the immigration detainee according to the terms of her policy.
58. Accordingly, Mr Denholm's argument may be more fruitfully examined under the rubric of the second issue.

Issue 2 – No link

59. Mr Gullick urges me to find that the instant case is far removed from the extreme type of case where the ECtHR has ruled that there was no link between the ground of permitted deprivation of liberty relied on, and the place and conditions of detention.
60. In Aerts v Belgium [2000] 29 EHRR 50, the applicant had profound psychiatric problems and should have been held in a Social Protection Centre under provisions of Belgian law which reflected Article 5(1)(e) of the Convention. Instead, he was held in the psychiatric wing of a prison for seven months: it appears that there were no spaces available. At paragraph 46 of its judgment, the Court said this:-

"The Court reiterates that in order to comply with Article 5(1), the detention in issue must take place "in accordance with a procedure prescribed by law" and be "lawful". The Convention here refers essentially to national law and lays down the obligation to conform to the substantive and procedural rules of national law, but it requires in addition that any deprivation of liberty should be in keeping with the aim of Article 5, namely to protect the individual from arbitrariness" [emphasis supplied].

61. I have supplied the relevant emphasis to underscore the point that a public law error is insufficient to found a breach of Article 5. In Aerts, the breach of Article 5(1)(e) was constituted by the following:

"The reports … show sufficiently clearly that the psychiatric wing could not be regarded as an institution appropriate for persons of unsound mind, the latter not receiving either regular medical attention or a therapeutic environment … the proper relationship between the aim of the detention and the conditions it took place was therefore deficient."

62. What was critical in Aerts was the finding that there was an inadequate or "deficient" link between the applicant's incarceration in the psychiatric wing of a prison and the aim/ground of his detention for the purposes of Article 5(1)(e). The ECtHR did not hold that no therapy whatsoever was provided in the psychiatric wing; rather, the environment, having regard to the limited regime available, was not appropriate. This entailed a broad evaluative assessment. I doubt whether the ECtHR would have come to the conclusion it did unless it considered that the prison regime was clearly inappropriate. Furthermore, it is worthy of note that the applicant's Article 3 claim was not upheld.
63. In Mayeka v Belgium [2008] 46 EHRR 23, a decision of the Grand Chamber, the five year old applicant was detained in a closed centre intended for illegal immigrants in the same conditions as adults. She was in a position of extreme vulnerability. Although her detention came within Article 5(1)(f) of the Convention (as did this Claimant's):

"… that … does not necessarily mean that it was lawful within the meaning of this provision, as the Court's case law requires that there must be some relationship between the ground of permitted deprivation of liberty relied on and the place and conditions of detention (paragraph 102 of the judgment)".

64. The ECtHR held that there was no relationship between the ground of permitted deprivation of liberty relied on (Article 5(1)(f)) and the place and conditions of this applicant's detention. She was held in a closed centre in the same conditions as adults. Specifically:

"… these conditions were consequently not adapted to the position of extreme vulnerability in which she found herself as a result of her position as an unaccompanied foreign minor (paragraph 103)".

65. Whereas it is correct to point out that the ECtHR upheld an Article 3 complaint in that case, I do not read paragraph 103 of the judgment as depending on that finding. It was inherently inappropriate to hold a small child in an adult detention facility which could not properly address her needs. That said, Mayeka was clearly an extreme case.
66. In Saadi v UK [2008] 47 EHHR 17, the applicant asylum-seeker was held in an IRC, not a prison. The Grand Chamber held that it was neither necessary nor sufficient for the applicant's purposes to prove a violation of the procedural or substantive requirements of domestic public law; a breach of Article 5 could be made out if the detention was "arbitrary". However, in all the examples of arbitrariness the Grand Chamber gave, the Administrative Court would in fact find a breach of our domestic public law (the position might differ across the signatory States). More importantly, at paragraph 74 of its judgment the ECtHR expounded the general principle that the place and conditions of detention should be appropriate, bearing in mind that the coercive measure (under Schedule 2 to the 1971 Act in that case) was applicable (both on the facts of the case before it, and typically) to those who have not committed criminal offences. The Grand Chamber concluded that there had been no violation of Article 5(1) on the facts of Saadi: amongst other things, the detention was lawful under domestic law, and the IRC in question was specially adapted for asylum seekers.
67. In both Aerts, Mayeka and Saadi the "ground of permitted deprivation of liberty relied on" was, respectively, Article 5(1)(e) and (f) of the Convention. In the language of the Convention, this was the "legitimate aim". Equally, in the language of the Convention, the relationship or nexus was held not to be established because in the first two cases in this trilogy the methods Belgium adopted to pursue that legitimate aim were either inconsistent with it (Mayeka) or clearly inadequate to achieve it (Aerts). The reason why the United Kingdom successfully upheld its administrative detention of Mr Saadi is important: the Defendant could show that the place and conditions of detention were appropriate, for the reasons summarised above. Had, for example, Mr Saadi been held in prison, the outcome would surely have been different.
68. Equipped with these authorities, Mr Gullick submitted that the threshold is a very high one and that in the instant case there was at least some link and/or a legally sufficient relationship between the aim of detention, or the ground of permitted deprivation of liberty, and the locale and conditions in which it took place. But the difficulty with that submission is that the link he puts forward in his written and oral argument is not well founded, for the reasons I have already ventured to explain.
69. Whether there was in fact no link between the ground of permitted deprivation of liberty, and the place and conditions of the Claimant's detention, regardless of the reasons given by the Defendant, potentially raises a more difficult question. Foreign criminals are liable to deportation and to be detained by the Defendant pursuant to the exercise of administrative powers. Would the Defendant be in breach of Article 5(1)(f) if she closed all her IRCs and placed all immigration detainees in prisons? In my view, probably but not necessarily; it would depend on an analysis of the conditions of detention. If immigration detainees were treated in exactly the same way as remand prisoners, for example, I apprehend that there would be difficulties from the Defendant's perspective. The Claimant's case clearly does entail making comparative judgments between incarceration on the one hand and IRCs on the other. On one view, the problem arises from the Defendant's perspective because she has chosen to build IRCs which constitute a separate detention facility.
70. Furthermore, it is not in dispute that circumstances may arise where it is entirely appropriate to hold immigration detainees in the more secure environment of HMPs. As previously observed, that requires a risk assessment on a case by case basis.
71. But in a situation where the Defendant routinely holds immigration detainees in prison conditions, and abstains from conducting a risk assessment on any individualised basis, the link between the ground of permitted deprivation of liberty (viz. from the perspective of Article 5(1)(f), meeting the risk of absconding, in an immigration context) and the place and conditions of detention becomes extremely tenuous. This is all the more so where the Defendant herself has failed to put forward a link which is remotely compelling.
72. In my judgment, a correct approach to this issue in the context of Article 5(1)(f) of the Convention cannot ignore (a) the link the Defendant herself advances as satisfying the requirements of that sub-Article, and (b) the mass of international and other materials of high-standing to which I have previously referred. The Defendant has created a separate facility for immigration detainees precisely because she recognises that the paradigm case of immigration detention neither requires nor justifies incarceration. The Defendant has consistently recognised the discrete nature of immigration detention – until, that is, she changed her policy in the EIG in January 2012 (without, as it happens, amending it elsewhere), and then procured an increase in the NOMS facility so that those in the Claimant's position would remain incarcerated for lengthy periods. As I have already said, this change in policy had nothing to do with the express and implied requirements of Article 5(1)(f).
73. In the context of what I have numbered the fourth issue, Mr Denholm seeks to persuade me that the conditions of detention in prisons are "unduly harsh". What exactly that means in the context of the fourth issue will need to be examined, but for the purposes of this second issue I consider that all Mr Denholm requires is to demonstrate the existence of significant differences between incarceration and IRCs. In my judgment, he has clearly succeeded to that extent. I will examine the available evidence when I come to address the fourth issue, but the Defendant has never disputed that there are material differences between the prison and IRC regimes.
74. Mr Denholm also relies on paragraph 17 of the judgment of Bean J in Rozo-Hamida, where he refers to the views of CPT (see paragraph 16 above):-

"Although this opinion is not binding on me, the views of the CPT are entitled to great respect. Certainly it would be disturbing to most people's sense of fairness that an immigration detainee who has not been convicted of any criminal offence should be confined in a prison save in the most exceptional circumstances".

75. Approaching the issue at this stage without reference to binding Court of Appeal authority, my approach would be very similar. My sense of fairness is similarly disturbed. However, I am not sure that holding immigration detainees in prison requires "the most exceptional circumstances". It does require a sound and proper justification within the context of Article 5(1)(f), and the policy matrix which the Defendant has devised and implemented. A policy which either systematically or invariably (it matters not which for this purpose) has a consequence of holding those in the Claimant's position in prison, rather than in an IRC, cannot be properly justified. Moreover, the implementation of such a policy severs the requisite link which must exist in cases such as these to justify detention under Article 5. The severance of that link is conclusively demonstrated in the Claimant's case by the fact that he was assessed as being suitable for detention in an IRC on 3^rd July 2013. On balance, therefore, if the matter were free from authority, I would hold that the Defendant's incarceration of the Claimant between 3^rd July 2013 and 14^th March 2014 was in breach of his rights under Article 5(1)(f) of the ECHR; and a sufficiently serious breach to sound in damages.

Issue 3 – stare decisis

76. Mr Gullick relied very heavily on the reserved judgment of the Court of Appeal in Krasniqi v SSHD [2011] EWCA Civ 1549, decided on 19^th December 2011. Carnwath LJ gave the sole reasoned judgment, with which Moses and Sullivan LJJ agreed. In that case, Mr Krasniqi claimed damages for breach of Article 5 on the basis inter alia that he was held in prison rather than an IRC. The judge found, and it was not disputed, that if the Defendant's published policy had been correctly applied, the Appellant would, and should, have been transferred to an IRC on completion of his sentence of imprisonment.
77. It was submitted on behalf of the Appellant that his detention in prison was "arbitrary", and therefore in breach of Article 5 of the Convention. Exactly how the submission was advanced is unclear from an examination of paragraph 16 of the judgment. However, it is clear that Rozo-Hermida was not cited to the Court of Appeal. It was not an authority of co-ordinate jurisdiction, but it would have been helpful.
78. At paragraph 18 of his judgment, Carnwath LJ said this:-

"Mr Roe accepts that, in accordance with decisions of the Strasbourg Court, detention will not be lawful if it is "arbitrary", which might include detention in bad faith, or not genuinely for the purpose of the relevant exception, or where there is not "some relationship between the ground of permitted deprivation of liberty relied on and the place and conditions of detention" (Saadi v UK [2008] 47 EHRR 17, paragraphs 68-74). However, none of the cases relied on supports a claim based solely on an irregularity in the selection of the place of detention, at least in the absence of any evidence that the conditions of detention were unduly harsh."

79. I find this a difficult passage. Carnwath LJ recognised that, in the absence of a requisite link, or relationship, there could be a breach of Article 5. Aside from correctly holding that a public law error could not, without more, amount to a breach of Article 5, Carnwath LJ did not appear to identify the nature of the link in that case between the permitted deprivation of liberty relied on, and the selection of the place and conditions of detention. Instead, the Court of Appeal appeared to approach the issue on the basis of an overall qualitative assessment of the Appellant's incarceration, and concluded that it did not meet a high threshold ("unduly harsh"). In Saadi the Grand Chamber made clear that the place and conditions of detention must be appropriate, having regard to the fact that asylum seekers have not committed criminal offences. I do not read paragraph 18 of the judgment of Carnwath LJ as applying such a test. I have already made the point that the criminal record of the Claimant in the instant case is historical, in the sense that it could not, without more, justify incarceration once he was automatically on licence.
80. Plainly, the Court of Appeal did not receive the breadth and depth of submission which emanated from Mr Denholm in the present case, nor was Mr Krasniqi the victim of the systematic application of a flawed policy. Furthermore, it is unclear whether the Court of Appeal was treated to any evidence which pointed to the differences between prison and IRC regimes. Even so, in my view the Court of Appeal must have held that there was some link or relationship between the ground of detention and its place and conditions, or else it would have found a breach of Article 5. For the purposes of the doctrine of precedent, it seems to me that the Court of Appeal's exact route to that conclusion does not matter.
81. Mr Denholm sought to distinguish Krasniqi by relying on the clause after the final comma in paragraph 18, and cross-referencing that with paragraph 19 of the judgment, which provides:-

"The only specific example [Counsel for the Appellant] was able to offer of a finding of illegality under Article 5, depending solely on the place of detention, was Mayeka v Belgium [2007] 1 FLR 1726. There, a five year old Congolese girl was detained for two months, alone among strangers, in an adult detention centre. This, as the court found, caused such distress and potential psychological damage as to amount both to inhuman treatment contrary to Article 3, and to a violation, under Article 5, of the principle that the place and conditions of detention must be related to the permitted ground of deprivation of liberty. This was clearly an extreme case, and the contrast with the present case is striking. In my view, it underlines the force of Mr Roe's submission. The judge was correct to reject the claim under this head, in the absence of any specific evidence challenging the conditions of detention".

82. It is true that in Mayeka the applicant succeeded both under Article 5 and Article 3 of the Convention. On my reading of his judgment, Carnwath LJ was equating the two. In other words, the Court of Appeal was holding that the presence or absence of "some relationship" depends on a qualitative assessment of the conditions of detention, and the surpassing of some threshold. This explains the reference to "unduly harsh" in the final sentence of paragraph 18. However, this approach does not quite chime with my reading of the judgments of the ECtHR in Aerts, Mayeka and Saadi. In my view, those cases hinged on the absence or plain inappropriateness of a link, not on any direct qualitative judgment of the conditions of detention judged in isolation. Some form of qualitative assessment is required, but that is not the criterion. If, on the facts of Aerts, there had been no material difference between the conditions of detention in the psychiatric unit of a prison, and the conditions in the institution in which Mr Aerts should have been held, no breach of Article 5 would have been made out. The assessment goes that far, and no further.
83. As for the final sentence of paragraph 19 of Krasniqi, it seems clear that the Appellant adduced no evidence about the conditions of his detention, and relied merely on the fact that he was being held in prison. Plainly, he did not seek to contend that there was a breach of Article 3, which would have been close to an impossible submission. However, on any fair reading of his judgment, Carnwath LJ is holding that a claimant would need evidence to prove conditions close to inhuman treatment before an Article 5 breach could be made out.
84. Notwithstanding my respectful disagreement with this decision, it is necessary to identify what Krasniqi decides. In my judgment, Krasniqi is authority for the proposition that a public law error is insufficient. Mr Gullick is right in pointing out that the public law error in Krasniqi was as decisive in bringing about the place of detention as was the error in the present case. However, I do not agree that the systemic error in the present case is equally or less serious; the ECtHR has consistently regarded generic errors as being of greater concern. On the other hand, the Strasbourg jurisprudence shows that domestic law violations are insufficient to establish a breach of Article 5.
85. Further, Krasniqi must be regarded as authority for the proposition that, even in a case where detention was in the "wrong institution", namely one of Her Majesty's prisons, there was nonetheless some link between the ground of permitted deprivation of liberty relied on, and the place and conditions of detention, unless those conditions were "unduly harsh". In the light of the Appellant counsel's submissions, it was necessary for the Court of Appeal to arrive at that conclusion in order to dismiss the appeal. Finally, Krasniqi is also authority for the proposition that an Article 5 claim directed to the place and conditions of detention will only succeed if the evidence establishes something approximating a breach of Article 3. It is true that the Appellant adduced no evidence challenging the conditions of his detention, but the adducing of such evidence would not, without more, cure the difficulty. Carnwath LJ made clear that the evidence must be so compelling as to demonstrate that the conditions of detention were "unduly harsh" and tantamount to a violation of Article 3.
86. In my judgment, the effect of Krasniqi, properly understood, is that the Claimant must fail before this Court on the first and second issues. Given the conclusions that I would have reached had I not been constrained by binding authority, it seems to me that these issues are fit for further consideration by the Court of Appeal.

Fourth issue – "unduly harsh"

87. Taking up with vigour the invitation offered by Carnwath LJ in paragraph 19 of his judgment in Krasniqi, the Claimant adduced a mass of evidence bearing on conditions of detention for immigration detainees at Her Majesty's prisons as compared with IRCs. The majority of this evidence is of a generic nature, and I take Mr Gullick's point that the Claimant's own witness statement is far from ideal, in that it fails to link that generic evidence with the circumstances of his own case. However, I am prepared to take a commonsense approach and to consider the matter at the level of some generality, taking into account what I know about category B prisons.
88. Mr Denholm was very keen that I should consider all the generic, systemic evidence, including the evidence which was submitted very late (and in respect of which I am prepared to give permission). I can assure Mr Denholm that I have considered that evidence, but in the light of the conclusions I draw from it, it is unnecessary for this section of my judgment to be particularly lengthy.
89. In his recent evidence the Claimant draws my attention to a report published in September 2014 by Bail for Immigration Detainees ("BID"), Denial of Justice: the Hidden Use of UK Prisons for Immigration Detention. The executive summary appears at page 5 of the report. Although the authors have not refrained from a modicum of hyperbole, the summary is helpful:-

"Detainees held within the prison estate suffer from multiple, systemic, compounding barriers to accessing justice, with an often devastating effect on their ability to progress their immigration case, seeking independent scrutiny of their ongoing detention from the Courts and Tribunals, and seek release from detention, as well as on their physical and mental wellbeing".

90. This report describes these practical barriers, which include but are not limited to:-

"• No automatic access to on-site immigration advice like that provided for detainees in IRCs.

• The existence of financial disincentives to legal aid providers who wish to work with detainees in prisons under current legal aid agency contracts.

• Immigration detainees routinely held under serving prisoner regimes.

• Prison regimes and restrictions that preclude the holding of mobile phones, adequate access to wing telephones during working hours and a slow internal postal system in prisons, which delay and frustrate timely communication with legal advisors, the Courts, and the Home Office.

• Lack of internet access in prisons which hinders legal research for unrepresented detainees, and makes cooperation with the Home Office re documentation process very difficult.

• Home Office escorting failures resulting in failures to produce detainees at bail hearings.

• Time limited video link connections to prisons.

…

• Failure to fit electronic tags within the prescribed two working days resulting in extended detention in prison".

91. The Claimant's recently served evidence points to the practical difficulties in obtaining legal advice in Her Majesty's prisons, particularly in the current financial climate. There is no duty scheme available in prisons. According to the second statement of James Read, on 14^th May 2014 the Detention Advice Service, a charity which provided advice to immigration detainees and foreign nationals in prison, ceased trading. Further, a charity called Praxis no longer has funding to support legal advice in prisons. An email dated 4^th December 2014 corrects an error contained in the second witness statement of Karen Abdul-Hady of the same date. Elsewhere, the Defendant's rebuttal evidence served to temper some of the Claimant's strictures, and to demonstrate a degree of overstatement, but I am able to express my conclusions in the following way. I am entirely satisfied that there are significant differences between HMPs and IRCs. I have already pointed out that they are designed for different purposes. The existence of these significant differences no doubt avails Mr Denholm's Article 5 case in relation to the "some link" point, although he is precluded from running that case at this level. On the other hand, and this is decisive for the purposes of this fourth issue, the Claimant's evidence is quite insufficient to show anything like the level of "undue harshness" which Carnwath LJ had in mind in Krasniqi. This would be so whether "unduly harsh" is to be understood as meaning "tantamount to a breach of Article 3", or something slightly less serious. These are absolute, not comparative, judgments.
92. A finding, without more, that prisons are significantly harsher than IRCs does not avail the Claimant. For the reasons I have already given, binding authority precludes treating this as a freestanding basis for finding a breach of Article 5. Although the Claimant has picked up the gauntlet notionally thrown down by Carnwath LJ in the final sentence of paragraph 19 of his judgment in Krasniqi, he has failed by some margin to meet the threshold test which the Court of Appeal had in mind.

Conclusion

93. With considerable reluctance, I must dismiss this application for judicial review. I do, however, grant permission to appeal to the Court of Appeal. I must leave it to that Court to determine the proper bounds of Krasniqi.