Skip to main content
Skip to main content

Split Court of Appeal rules detention of asylum seekers unlawful

Written by
Rajkiran Barhey, UK Human Rights Blog
Date of Publication:
18 October 2018

R (on the application of Hemmati and Others) v The Secretary of State for the Home Department [2018] EWCA Civ 2122: The Court of Appeal has concluded, by a 2-1 majority, that the detention of five asylum seekers pending their removal to another country where they should first have claimed asylum had been unlawful, and that they were entitled to damages. This article (the first of two) will unpick the reasons behind this legally complex appeal.

Background to the case

The case concerned five individuals who had entered the EU via countries in Eastern Europe and eventually made their way to the UK illegally. All five claimed asylum in the UK.

The Dublin III Regulation – an EU-wide Regulation – creates a system which determines where a person should claim asylum. The key principle is that a person must claim asylum in the first Dublin country they reach, where they should stay.

However, as one example in this case, Mr Hemmati entered Bulgaria and claimed asylum, but then left and entered the UK illegally, where he claimed asylum. The UK therefore asked Bulgaria to take back Mr Hemmati. Bulgaria accepted and Mr Hemmati was detained before removal.

Article 28(1) of Dublin III says that a person cannot be detained if the only reason they are being held is because they are subject to 'the procedure established by this Regulation.' The 'procedure' is the process whereby a country determines the 'right' country where a person's asylum claim should be determined, in this case when the UK asked Bulgaria to take Mr Hemmati back, and Bulgaria agreed. In plain English, Dublin III prohibited the UK from holding Mr Hemmati in detention if their only justification was that they were submitting a take back request to Bulgaria and waiting for Bulgaria to reply.

However, Article 28(2) says that if there is a significant risk that the person will abscond, then a Member State may detain the person.

Moreover, Article 2(n) defines the phrase "risk of absconding" in Article 28 as meaning "the existence of reasons in an individual case, which are based on objective criteria defined by law, to believe that an applicant or a third country national or a stateless person who is subject to a transfer procedure may abscond." This creates an important safeguard regarding the exercise of the power under Article 28(2).

The key point is that in an Article 28 case, i.e. a person who is going through the Dublin III procedure, the only ground for detention is if there is a 'significant risk of absconding', which is defined as reasons based on objective criteria defined by law.

The Al-Chodor decision

The European Court of Justice decided the case of Case C-528/15 Al Chodor just before the appeal in Hemmati. The CJEU considered what the phrase "objective criteria defined by law" set out in Article 2(n) required. The specific issue was whether 'defined by law' required the criteria to be set out in legislation, or whether settled case law confirming a consistent administrative practice was sufficient.

The CJEU found that

the individual discretion enjoyed by the authorities concerned pursuant to Article 28(2) of the Dublin III Regulation, read in conjunction with Article 2(n) thereof, in relation to the existence of a risk of absconding, should be exercised within a framework of certain predetermined limits. Accordingly, it is essential that the criteria which define the existence of such a risk, which constitute the basis for detention, are defined clearly by an act which is binding and foreseeable in its application."[emphasis added].

But whilst legislation was not necessary, the CJEU further stated that the following requirements were needed:

… only a provision of general application could meet the requirements of clarity, predictability, accessibility and, in particular, protection against arbitrariness.

The Court's judgment

Sir Terence Etherton MR and Jackson LJ gave the majority judgment with Sales LJ dissenting.

At [5] Sales LJ summarised the key issues. This post will deal with the first issue, namely:

Whether the Hardial Singh principles and/or the Secretary of State's published policy in Chapter 55 of his Enforcement Instructions and Guidance ("EIG") satisfied the requirements of Article 28 and Article 2(n) of Dublin III in relation to the periods of detention of the appellants.

This can be be split two further sub-issues:

  1. Whether the policy published in Chapter 55 of the Secretary of State's Enforcement Instructions and Guidance ("EIG") satisfied Articles 2(n) and 28 and Al Chodor.
  2. Whether the Hardial Singh principles satisfied Articles 2(n) and 28 and Al Chodor.

It is worth clarifying at the outset that it was agreed that the Court could proceed on the assumption that the sole reason for the detention of the individuals was that the Secretary of State wished to remove them under the Dublin III procedure.

The EIG policy

I would recommend looking at the policy as, otherwise, the judgment is somewhat hard to understand. The operative part is 55.3.

The question of whether the EIG policy met the requirements outlined above can be separated into three smaller issues:

  1. The 'quality of law'
  2. The 'form of law'
  3. The 'author of the law'

The 'quality of law'

The 'quality of law' question was whether the EIG satisfied the requirements of clarity, predictability, accessibility and protection against arbitrariness within a framework of certain predetermined limits, as required by Al Chodor and Articles 28 and 2(n).

The majority found that the EIG did not satisfy these requirements. Crucially, the policy did not contain a clear, exhaustive list of factors that were relevant to whether a person was at risk of absconding. Furthermore, the policy did not refer to Dublin III, Article 28 or Article 2(n) and the respective requirements of those provisions and did not contain a list of factors which were stated to be relevant to the risk of absconding.

Whilst it did contain lists of factors which were relevant to the broader question of whether the power of detention should be exercised, a general list was not good enough for the majority. Within that general list, some of the factors were clearly irrelevant to the risk of absconding. The court found that a person would not be able to read the EIG and understand all of the criteria which would be applied to meet the requirements of Articles 28 and 2(n).

Sales LJ dissented on the basis that neither Article 2(n) nor Al Chodor required that the criteria were set out separately or identified clearly as relevant to the risk of absconding. In his view, a person could look at the EIG and understand which criteria would be applied to them. Some of the criteria were so obviously not relevant to a risk of absconding that, even though they appeared on the same list as criteria that were relevant, an individual was unlikely to find it confusing.

The 'form of law'

The issue was whether a policy document could constitute 'law' such that the requirement of Article 2(n) for 'objective criteria defined by law' could be satisfied, i.e. 'an act which is binding and foreseeable in its application.' (Al Chodor).

Sales LJ found that, as a result of the doctrine of legitimate expectations and of consistency in the application of policy, policies were capable of binding decision makers.

Normally, a decision maker can depart from a policy if they have a good reason. This would not meet the requirement of legal certainty. Sales LJ got around this by saying that there would never be a good reason for departing from the policy, because departing from the policy would be contrary to EU law, and no reason that requires a decision maker to depart from EU law can be a good reason. Therefore, in his view, the EIG was capable of falling within the concept of 'law' in Article 2(n).

The majority expressed no firm view, but made some critical comments of Sales LJ's position. They were sceptical of Sales LJ's reliance on case-law decided under the ECHR. They accepted that whilst the case law showed that 'in accordance with the law' in ECHR jurisprudence has been held to be wide enough to accommodate policies, it did not necessarily follow that the same analysis could be applied to Dublin III.

They also pointed out an ambiguity in the Al Chodor judgment. The Advocate General's Opinion had expressed the view that the only sort of law which would satisfy Article 2(n) was legislation. The CJEU itself did not explicitly express this view, stating the view that 'an act which is binding and foreseeable in its application' is required. However, it endorsed the above paragraphs in the AG's Opinion. This created some ambiguity as to what the judgment actually meant.

The 'author of the law'

The author of the law issue concerned the question of whether it mattered that the EIG was written by the same branch of the executive who were charged with its application, or whether a more independent body, e.g. the legislature, has to set the criteria.

In Al Chodor, the Advocate General endorsed legislation as the appropriate form of law partly on the basis that legislation is created by a 'institutionally separate authorities' and therefore there is a greater guarantee of control over the powers of the executive.

The CJEU was more measured, stating that

criteria established by a binding provision are best placed for the external direction of the discretion of those authorities for the purposes of protecting applicants against arbitrary deprivations of liberty. [emphasis added]

In the present case the policy was written and applied by the same body – the executive.

Sales LJ did not find this fatal, however, because the individual officials applying the policy would rarely, in reality, be the same people who had created the policy. He further noted that the Secretary of State was still bound by the policy as a result of the doctrine of legitimate expectations, which acted as a safeguard against arbitrariness.

The majority was careful to not express a view on this point, on the basis that Al Chodor was ambiguous as to the state of EU law and the point was of great importance. This issue is therefore not settled.

The Hardial Singh principles

The second sub-issue for the court was whether the Hardial Singh principles satisfied Article 2(n) and Article 28.

First, a recap of the Hardial Singh principles. In R (on the application of Hardial Singh) v Governor of Durham Prison [1983] EWHC 1 (QB), Woolf J (as he then was) found that although there were no express limitations upon the length of time a person could be held in immigration detention, the power was subject to limitation. The four principles reaffirmed at [169] of Hemmati are as follows:

  • The Secretary of State must intend to deport the person and can only use the power to detain for that purpose.
  • The deportee may only be detained for a period that is reasonable in all the circumstances.
  • If, before the expiry of the reasonable period, it becomes apparent that the Secretary of State will not be able to effect deportation within that reasonable period, he should not seek to exercise the power of detention.
  • The Secretary of State should act with reasonable diligence and expedition to effect removal.

The majority clarified that they did not understand it to be disputed between the parties that the Hardial Singh principles did not satisfy the requirements of Article 2(n).

They summarised the principles in Hardial Singh as essentially imposing a reasonableness test on the executive. In their view, this was a long way from providing criteria for deprivation of liberty which had the clarify, predictability, accessibility and protection against arbitrariness within a framework of certain pre-determined limits as required by Al Chodor.

Sales LJ stated that there was no requirement that the Hardial Singh principles ought to be informed by the requirements of Dublin III. In his view the principles were

well settled principles of domestic law arising as a matter of statutory interpretation of the powers of detention in the 1971 Act.

In other words, Hardial Singh governs the exercise of a power which emerges from domestic law and is confined to the domestic sphere. There was no reason in his view why Dublin III principles ought to inform its interpretation.


As readers who have made it this far will no doubt have gathered, Hemmati is a complex judgment raising a number of issues, ranging from the technical to the jurisprudential.

In my view, one of the most interesting discussions raised by the case is the proper definition of 'law.' What are the implications of holding a Home Office policy to be law? What is the boundary between policy and law? Can we really extrapolate anything of significance from this judgment or is the definition of law adopted relevant only to that particular Dublin III Regulation?

An interesting postscript to the Hemmati litigation is that the Secretary of State promulgated a set of Regulations to fix the problem. The Regulations — which have the catchy title of The Transfer for Determination of an Application for International Protection (Detention) (Significant Risk of Absconding Criteria) Regulations 2017 — set out the criteria to be used to determine if an individual was at risk of absconding under Article 28 and 2(n).

A similar challenge to the 2017 Regulations was mounted, namely that they were inconsistent with Al Chodor, but the challenge was recently dismissed by Lambert J in R (on the application of Omar) v SSHD [2018] EWHC 689 (Admin).

Therefore, what will happen next remains to be seen.

Watch this space for Part 2 — damages in this case.