MM & Anor v Secretary of State for Work and Pensions (DLA)

These appeals raise the important question whether the application to refugees and their family members of the past presence test (“PPT”) in regulation 2(1)(iii) of the Social Security (Disability Living Allowance) Regulations 1991 amounts to unlawful indirect discrimination contrary to the provisions of Article 28 of EU Directive 2004/83/EC (the Qualification Directive), or Article 14 of the European Convention on Human Rights (ECHR).

I have decided that the PPT is unlawfully discriminatory on both bases and should be disapplied. This means that, if they meet the substantive conditions of entitlement to DLA, the Appellants should at the time of the decisions have been awarded DLA.

For present purposes, the relevant facts can be stated very briefly.

Both appellants were children at the time of the decisions which are the subject of these appeals. They both have substantial disabilities. MM is a Ugandan national. His mother was granted refugee status in the UK in March 2012. MM joined his mother in April 2013 with entry clearance on the basis of family re-union. A claim was made on his behalf for DLA in August 2013. SI is a Somali national. She arrived in the UK with her mother and sister in August 2013 and they were given indefinite leave to remain on arrival. A claim was made on her behalf for DLA in August 2013. The Secretary of State refused both appeals because neither Appellant had been present in the UK for 104 weeks.

They appealed, through their mothers who are their appointees, to the First-tier Tribunal on the ground that the PPT discriminated against them contrary to Article 28 of the Qualification Directive.

On 31 October 2014 the First-tier Tribunal doubted that Article 28 of the Qualification Directive was directly effective but it did not need to determine that point because it dismissed the appeals for other reasons. It decided that DLA was social assistance for the purpose of the Qualification Directive and that the PPT discriminated against refugees and those whose status is dependent on their family relationship to refugees, as compared with UK nationals. The measure was indirectly discriminatory but the tribunal found that the discrimination was justified. It decided that the aim of the PPT was to direct benefits to those who are permanent long term residents of the UK, that the imposition of a requirement for a period of residence was a rational means of achieving that aim and that the objective and the PPT were sufficiently connected for it not to be considered manifestly without reasonable foundation as a means of achieving that aim. The tribunal distinguished the case of Stewart v SSWP C-503/09, [2012] PTSR 1 on the ground that, here, the legitimate aim was not to establish a genuine link with the UK (as in Stewart) but was to direct benefits to permanent long-term residents of the UK.

On the basis of permission given by the First-tier Tribunal and supplemented by the Upper Tribunal, the issues in this appeal were identified as follows:

At the hearing I raised with the parties whether there should be a reference to the CJEU on the questions of direct effect or whether DLA was social assistance within the Directive. Mr Spencer submitted that none was needed, but Mr Henderson requested that the Secretary of State be allowed time to consider the matter. I gave permission to the representatives to make further written submissions on this and other specific points that had arisen in the course of the hearing.

In his subsequent written submissions, Mr Henderson said that the Secretary of State’s position was that no reference to the CJEU was required because the Secretary of State conceded that “Article 28 was in principle capable of having direct effect in respect of a claim by the Appellants for social assistance within the meaning of Directive 2004/83/EC… in the event that the UT finds that DLA does constitute social assistance, the Respondent concedes that Article 28 could be relied upon directly by the Appellants.” I do not understand what is meant by the use of the words “in principle” but, read as a whole, this is a concession that Article 28 is directly effective for these purposes. For reasons which I will briefly explain under the heading “Discussion”, that concession is correctly made and I so find.

Section 71 Social Security Contributions and Benefits Act 1992 makes provision for disability living allowance (DLA). By section 71(6) a clamant must satisfy prescribed conditions as to residence and presence in Great Britain. Those conditions are to be found in regulation 2 of the Social Security (Disability Living Allowance) Regulations 1991 (“the DLA Regulations”):

There is provision for exemptions from regulation 2(1)(a)(ii) in the case of absence from Great Britain of a serving member of the forces and those in other specified occupations, and absences for medical treatment. There is another exemption provided by regulation 2A, which was introduced following the decision of the Court of Justice in Stewart, as follows:

The Qualification Directive is headed as being “on minimum standards for the qualification and status of third country nationals or stateless persons as refugees or as persons who otherwise need international legal protection and the content of the protection granted”.

There are a number of relevant recitals to which I return later. Of particular importance is Recital 33 provides:

This is given effect by Article 28:

The effect of Article 28 is extended to the family members of those with refugee or subsidiary protection status by Article 23(2) of the Directive which provides that:

Article 23(3) excludes the above where the family member is or would be excluded from the relevant status pursuant to the provisions of the Directive.

Article 28 of the Qualification Directive

Direct effect

In the light of the Respondent’s concession, I can deal with this relatively briefly.

In Netherlands v Federatie Nederlandse Vakbeweging C-71/85, [1987] 3 CMLR 767 the European Court of Justice said at paragraph 13:

The requirement to be unconditional does not preclude the imposition of conditions of entitlement in implementing legislation. It means that:

The requirement of precision was explained by the Court of Justice in Van Duyn at [14]:

So legislation can be precise even if ambiguous, as explained by the Advocate-General Warner in Regina v Secretary of State for Home Affairs, ex p Santillo C-131/79 [1980] 2 C.M.L.R. 308:

The parties were able to find only one domestic authority in which the question of the direct effect of Article 28 has been addressed. This is Blakesley v SSWP [2015] 1 WLR 3150, in which at [60] Jackson LJ doubted that the provision had direct effect. However, with the greatest respect to him, his observation was obiter and does not take matters any further.

Article 28 satisfies the requirements for direct effect. First, it is unconditional in the above sense. There is no discretion in the UK Government as to whether to comply. The words of Article 28 are unconditional and Article 38(1) required the UK to bring into force the laws, regulations and administrative procedures necessary to transpose the Qualification Directive into domestic law before 10 October 2006. I note that even where a Directive contains a reservation as to scope rationae materiae, it may be directly effective. See Marshall v Southampton and South-West Hampshire Area Health Authority C-152/84, [1986] QB 401 at [54], and Netherlands [18] – [21].

Second, Article 28 is sufficiently precise. It imposes a clear requirement that those to whom it applies receive “necessary social assistance as provided to nationals”. As the Advocate General explained in Santillo, it can have the necessary quality of precision even though there is a need for a judicial determination as to what constitutes “necessary social assistance”.

This is also the position in relation to family members of refugees, despite the possibility under Article 23(2) of the imposition of conditions applicable to their benefits. In particular:

Accordingly I conclude that Article 28 of the Qualification Directive has direct effect in its application to refugees and their family members. In consequence, it is the responsibility of the national courts and tribunals to protect the rights conferred by the Directive.

Necessary social assistance

The First-tier Tribunal noted that the meaning of the term “social assistance” has not been considered by the courts in the context of the Qualification Directive but has been considered in other contexts, including the Coordination Regulation (Regulation 883/2004, replacing Regulation 1408/71). The tribunal said that, in that context, the CJEU has found that the care component of DLA is “sickness benefit” and not social assistance, and that the mobility component was a “special non-contributory benefit” and not social assistance. The tribunal noted that the interpretation of social assistance may depend on the context of the particular European instrument in which it arises.

The First-tier Tribunal did not find the question whether DLA was social assistance an easy one but concluded, on balance of probabilities, that it was. It thought that the distinction between DLA and means-tested benefits, which the Respondent had sought to make, was not clear cut and that European case law did not rule out non-means tested benefits as being social assistance. It decided that DLA was “part of the state’s provision to meet the basic subsistence needs of disabled people”.

The parties agree that the tribunal was wrong to decide this issue on “balance of probabilities”, which is the applicable approach to a finding of fact. Whether DLA is social assistance is a matter of law and it should have been determined as such.

As to the substantive issue of law, the Appellants agree with the tribunal’s conclusion but do not fully support the reasoning. They say that the tribunal’s approach was overly technical and that this Tribunal should adopt a broad purposive construction of the Directive and consider the nature of DLA in that context. The Respondent does not support either approach, and contends that DLA is not social assistance within Article 28. The Respondent submits that, while a purposive construction is appropriate, the tribunal’s task is to interpret the words of the Directive itself and that it is not necessary to consider the detail of DLA.

In the Upper Tribunal neither party has submitted that decisions of the CJEU as to the meaning of “social assistance” in other contexts are determinative in this context, although the Respondent did rely on some of those decisions in the First-tier Tribunal.

In Commission of the European Communities v European Parliament C-299/05 [2007] ECR 1-8695 the CJEU held that the care component of DLA was a “sickness benefit” and not “social assistance” for the purposes of the Coordination Regulation (Council Regulation 1408/71 on the application of social security schemes to employed and self-employed persona and their families moving within the Community). But the Coordination Regulation contained detailed provisions as to the benefits to which it applied and the judgment was informed by the technical classifications there. This contrasts with the Qualification Directive, which makes no attempt to classify different benefits save to distinguish between “social assistance” and “core benefits” (to which I return later).

In Pensionsversicherungsanstalt v Brey C-140/12, [2014] 1 WLR 1080 the Court of Justice considered the meaning of “social assistance” in Directive 2004/38, which is concerned with the free movement of EU citizens and their family members. The specific article in question grants the right of residence in another member state for EU citizens who do not have worker status provided they have sufficient resources not to become a burden on the social assistance system of the host state. In his opinion the Advocate General noted that the term “social assistance” does not necessarily have to be construed in the same way as in the context of other EU legislation, and that the term was not intended to have a precise meaning (paragraphs 40 and 41). In its judgment the Court said at [49]:

The Court said that the objective of the provision in question was to allow member states to impose legitimate restrictions in connection with the grant of social security benefits to non-economically active citizens of other member states so that those citizens do not become an unreasonable burden on the social assistance system of that member state ([57]), and then said:

The Court held that the fact that a benefit was not “social assistance” for the purpose of the Coordination Regulation did not mean that it could not be such for the purpose of the Directive.

In the light of the decision in Brey, the parties are agreed that the context of the Directive is key to the interpretation and so decisions as to the meaning of “social assistance” in other contexts are of limited if any assistance.

Turning then to context, Mr Spencer submits that Recital 33 of the Qualification Directive (set out at paragraph 13 above) shows that the purpose of Article 28 is to “avoid social hardship” by providing refugees and their families with “adequate social welfare and means of assistance.” Moreover, he says, Recital 2 shows that the purpose of the Directive as a whole is to secure the full and inclusive application of the Geneva Convention relating to the Status of Refugees (the Geneva Convention) across member states, and so the terms of the Convention and the view of the United Nations High Commissioner for Refugees (UNHCR) may be relied on as an aid to construction of Article 28. When viewed in this context he submits that it is clear that the term “social assistance” is a broad one and not confined to specific categories of welfare benefits based on formal criteria.

Mr Henderson submits that the words of Recital 33 demonstrate that “social assistance” is about the means of subsistence or covering the costs of everyday living. Income support is such a benefit but DLA is not. Rather, he submits, DLA is a specialist payment to cover the extra costs associated with long-term disability but not daily subsistence costs. Subsistence costs are met by benefits such as income support. In the First-tier Tribunal the Respondent said that subsistence costs of children are met by benefits such as child benefit or child tax credit, but not by DLA.

I have summarised the respective positions of the parties very briefly. Further detail appears in my analysis below.

I conclude that DLA is “social assistance” within Article 28. The term is intended to embrace a wide range of welfare benefits so as to ensure that refugees and their families are treated equally within the benefits systems of member states. But even if it is directed to a narrower group of subsistence benefits, DLA for these purposes is included. I now explain why I have reached these conclusions.

Despite Mr Henderson’s reliance on them, I derive little assistance from the precise words used in Article 28 itself. The Article does not define “social assistance” and it is not clear from the words alone what “necessary” adds. It appears from Recital 33 that Article 28 is concerned with “adequate social welfare and means of subsistence”, but neither of those terms is defined. Recital 34 provides some assistance in that it distinguishes between “social assistance” and “core benefits”. The latter are described as comprising “at least minimum income support, assistance in case of illness, pregnancy and parental assistance”, and “social assistance” is plainly wider than that.

Recitals 2 – 7 show that the purpose of the Qualification Directive is to assist in achieving the European Council’s commitment to the full and inclusive application of the Geneva Convention relating to the Status of Refugees. The Geneva Convention is therefore key to understanding the terms of the Directive. This is established by the case law of the CJEU, as to which a helpful summary and explanation is found in the opinion of the Advocate General in Warendorf v Alo and Osso v Hannover Joined Cases C-443/14 and C-444/14 delivered on 6 October 2015. Those cases were concerned with Directive 2011/95, which substitutes the Qualification Directive for those states which are bound by it (and the UK is not), and is in materially identical terms to the Qualification Directive. The Advocate General said as follows:

In HB v SSWP (IS) [2013] UKUT 433 (AAC) at [29], the Upper Tribunal accepted that the Geneva Convention can be used as an aid to interpret the Qualification Directive and its relevance was implicitly accepted by the Court of Appeal, on appeal from the decision in HB, in Blakesley v SSWP [2015] 1 WLR 3150. Moreover, as Jackson LJ said at [44] of his judgment in that case, the views of the UNHCR should always be taken into account when the courts are considering the interpretation of the Geneva Convention.

Article 23 of the Geneva Convention provides:

The term “public relief and assistance” is a broad one, not defined by reference to particular categories of benefits or technical rules of entitlement. The width of the term is confirmed by the UNHCR Annotated Comments on the EC Council Directive 2004/83/EC of 29 April 2004 (January 2005) which state:

This explains the wide terminology used in the Qualification Directive. To define “social assistance” in Article 28 by reference to formal criteria relating to specific benefit entitlements would create the risk that refugees may be treated differently depending on how member states organise their benefits systems (Brey at [59]), and undermine the aims of the Geneva Convention as explained by the UNHCR.

The inclusion of the adjective “necessary” in Article 28 does not alter this conclusion. I note that the word does not appear in Article 23 of the Refugee Convention. It would not be consistent with Article 23, nor with the aim of applying a common standard across member states, to interpret “necessary” as providing some discretion in member states to determine the scope of social assistance to which beneficiaries of the Directive are entitled. I conclude that “necessary” implies a standard for social assistance that is afforded to meet the needs of claimants, and reflects the terms of recital 6 of ensuring a “minimum level of benefits”.

Mr Henderson’s submissions pose too sharp a distinction between subsistence benefits and a disability benefit such as DLA. Even in the line of European case law which has addressed technical classifications of benefits in the Coordination Regulation, it is hard to find a clear line between means-tested (subsistence) benefits and social assistance.

The fact that DLA is awarded on the basis of an individual assessment of need does not prevent it from being a subsistence benefit. The Court of Appeal observed in Burnip v Birmingham City Council [2012] EWCA Civ 629 at [45] that DLA is intended to meet the ordinary living expenses of the disabled person, recognising that “Disability can be expensive” (Burnip at [1]). The essential or subsistence needs of people with disabilities can be greater than those of others. In terms of “social hardship”, the benefit is designed to avoid the hardship associated with disability.

This is given added force by two particular characteristics of DLA. First, athough DLA is not means-tested, it is (as the First-tier Tribunal acknowledged) ignored as income in the assessment of means for other benefits and so is implicitly treated as money paid to meet free-standing needs arising from disability. Second, DLA operates as a “passport” to a range of other benefits which are intended to meet subsistence and housing needs of disabled people or their carers. and which are, on the Respondent’s view (see the transposition note, discussed at paragraphs 53 – 55 below) “social assistance” within Article 28. For example, a person who is regularly and substantially engaged in caring for a person who is receiving the middle or highest rate of the care component of DLA is entitled to income support and is not required to take part in work-related activity; where a child is receiving DLA, a housing benefit claimant is entitled to a disabled child premium and is exempt from the benefit cap; and receipt of DLA care at the highest or middle rate is one of the conditions of entitlement to an additional room for a disabled adult or child for housing benefit purposes. Mr Spencer provided a number of other examples of DLA operating in this manner.

Mr Henderson says that these passported benefits are all social assistance because they are means-tested. This is not entirely accurate (carers allowance is not means-tested). But in any event that is a circular and self-serving argument which does not explain why social assistance should be limited to means-tested benefits.

Of course, the role of DLA as a passport to those benefits does not of itself mean that DLA has the same character as those benefits. But it shows that the benefits system recognises disability as giving rise to additional needs and associated costs.

Mr Henderson relies on the transposition note which was prepared by the Home Office to explain to the European Commission how the Qualification Directive was being implemented in the UK, and which was annexed to the Explanatory Memorandum to the Refugee or Person in Need of International Protection (Qualification) Regulations 2006. The note explained that, in implementation of Article 28, refugees were entitled to “the same income-related benefits as UK nationals, provided that they fulfil the normal conditions of entitlement” and then listed the benefits included as income support, state pension credit, housing benefit, council tax benefit and child benefit. Mr Henderson says that the note is an authoritative statement of the UK Government’s understanding of the intention behind the Directive, that the European Commission did not contradict it, and that it shows that DLA is not intended to fall within the definition of “social assistance”.

I do not find the transposition note to be of assistance. It reflects the Government’s understanding of the Directive but is not an aid to construction of the provisions of the Directive. Nothing can be read into the Commission not having objected to the note, not least because the note also stated that, once a person is granted refugee status or humanitarian protection, then they have access to public funds under section 115 of the Asylum and Immigration Act 1999. These funds include DLA. So, although the note went on to list a number of income-related benefits which were included within the entitlements, the proposals at that time did not suggest that there would be restricted access to DLA for persons to whom the Directive applied.

Finally, on this point, the inclusion in the list of child benefit (which is not means-tested, although since the date of the transposition note is subject to a tax charge for high earners) undermines the Respondent’s case that “social assistance” is restricted to means-tested benefits or those intended for “basic subsistence needs”.

Discrimination

The First-tier Tribunal accepted that refugees are intrinsically less likely to satisfy the PPT than UK nationals as a whole. But the Respondent contended that the correct comparator for refugees is not “UK nationals as a whole” but “UK nationals and other migrants arriving from abroad” and, on that basis, he submitted that both groups would be likely to have similar chances of satisfying the test and so there was no indirect discrimination. The First-tier Tribunal rejected this. Article 28 entitles refugees to the necessary social assistance “as provided to nationals of that Member State”. That therefore was the correct comparator. The tribunal found that the PPT constituted indirect discrimination.

In the Upper Tribunal Mr Henderson maintains the position that the correct comparator is “a migrant to the UK from a third country, or a UK national returning after a period living abroad”. A comparator for the purposes of a discrimination claim should be a person (or group), real or hypothetical, who is to the greatest extent possible identical to the claimant save for the purported ground of discrimination. He says that ‘UK nationals as a whole’ is too broad a category because it is not nearly close enough to the Appellants’ position to provide a sensible comparison. A third country migrant or UK national returning from some time abroad is in an identical situation to the Appellants, save for their refugee status (or that of their mothers), and so allows a proper comparison between similar groups. The question then arises whether the fact that the Appellants are family members of a refugee means they are less likely to fulfil the PPT than any number of other factors. He submits that, applying the correct comparator, the PPT is not indirectly discriminatory because a migrant or UK national returning home is just as likely to fail to satisfy the PPT as a refugee.

Mr Spencer submits that, if a strict comparator is required, the First-tier Tribunal was correct to adopt UK nationals as the correct comparator. However, he says that strict comparators are not necessary and it is sufficient that the PPT places refugees and their families at a particular disadvantage when compared with nationals, which requires justification.

I start with the words of Article 28 itself. It provides for equal treatment between refugees and nationals of the member state. It does not restrict the right to equal treatment by reference to any subset of member state nationals.

The Respondent’s approach seeks to impose such a restriction by defining the comparator by reference to the discrimination claimed but this defeats the purpose of the non-discrimination provision of the Directive. The approach does not find support in domestic or European case law.

In Patmalniece v SSWP [2011] 1 WLR 783 the issue was whether domestic legislation unlawfully discriminated against nationals of other EU member states as compared to UK nationals, contrary to article 3 of Council Regulation (EEC) 1408/71 which provided that employed persons who were insured by the social security system of a European Union member state (amongst others) should “enjoy the same benefits under the legislation of any Member State as the nationals of the State.” It was a condition of entitlement under domestic legislation that a claimant had a right to reside in the Common Travel area (which includes the UK). The Supreme Court held that, although the test applied to all claimants, a UK national was more likely to satisfy it than a national of another member state and so the provision was indirectly discriminatory (Lord Hope at [28] – [35], Lady Hale at [93]). The Supreme Court did not seek to restrict the pool of UK comparators so as to limit it to those who are in a situation akin to that of foreign nationals, for instance by limiting it to UK nationals who are coming into the UK following a period working abroad.

A similar approach can be seen in the judgment of the European Court of Justice in O’Flynn v Adjudication Officer C-237/94, [1996] ECR 1-2617 at [17] - [23], where the Court decided that regulations restricting funeral payments to cases where the funeral takes place in the UK were indirectly discriminatory against migrant workers. The relevant EU Regulation entitled a worker from one Member State to enjoy in the territory of another Member State the same social and tax advantages as national workers. The Court said:

This was the case even where the national provision was “applicable irrespective of nationality”: [18]. Moreover, it was not necessary to have statistics as to the impact. The question was simply whether the test was liable to have the discriminatory effect: [21].

The Court did not compare migrant workers with UK nationals who wished to hold a funeral abroad, or with UK nationals who had family connections abroad. The provision was indirectly discriminatory because migrant workers were more likely than nationals to have to arrange for burial in another member state in view of the links which the members of such a family generally maintained with their State of origin: [22].

Mr Henderson relies on the judgment of the CJEU in Dano v Jobcentre Leipzig C-333/13, [2015] 1 WLR 2519 in which it was held that it was lawful to impose a residence requirement on non-economically active nationals of other member states in order to qualify for certain benefits. Mr Henderson points to the similarity between the wording of the EU regulation in issue in that case, which entitled persons to whom it applied to enjoy the same benefits under national legislation “as the nationals thereof”, and that of Article 28 of the Qualification Directive. He says that the decision in Dano means that an equal treatment provision such as in Article 28 does not mean that it is impermissible to impose a residence condition.

This submission involves a selective reading of the judgment and is incorrect. The directive in issue in that case, which gave more specific effect to the general principle of non-discrimination in the EU regulation, permitted derogation including on the basis of residence.

There can be no doubt that the PPT places refugees and their families at a particular disadvantage compared to UK nationals. Refugees and their families are intrinsically less likely to satisfy the test than UK nationals. Indeed, that is not disputed by the Respondent.

The First-tier Tribunal decided that the PPT was justified on the following basis:

The Appellants’ case is that the tribunal fell into error because it failed to appreciate the true aim of the PPT, as stated by the Respondent, which was to ensure that limited public funds are directed towards those with the closest attachment to the UK and to protect the integrity of the UK benefit system. Directing benefits to those who are permanent long-term residents of the UK was the means employed to achieve those aims, and the PPT is the legislative means of achieving that. In wrongly confusing the aims of the measure with the means of achieving it, the tribunal allowed the PPT to be self-justifying as an end in itself.

Mr Henderson accepts that the First-tier Tribunal confused the aim (which is that in the preceding paragraph) and the means of achieving it, but he says that, despite that mistake, the measure is justified and so the tribunal reached the correct decision and the tribunal’s error was immaterial.

At the heart of this lies a disagreement between the parties as to the proper approach to justification in a case such as this.

It is common ground between the parties that, under both EU law and Article 14, the burden is on the Respondent to justify the discriminatory effect of the PPT and that the test for justification is whether the discriminatory measure can be objectively justified as a proportionate means of achieving a legitimate aim: Hockenjos v Secretary of State for Work and Pensions [2004] EWCA Civ 1749 at [156]. It is also common ground that the State enjoys a broad margin of appreciation in the area of social entitlement, but they disagree as to the scope of the margin in a case such as the present. I return to that issue below.

The Appellants say that, once the aim of the PPT is correctly identified, this case is indistinguishable from Stewart. The issue in that case included whether a condition of past presence in the UK for not less than 26 weeks in the preceding 52 weeks, as a condition of entitlement for short-term incapacity benefit in youth, was incompatible with Council Regulation (EEC) 1408/71 which entitled EU citizens to retain the right to such benefits when they resided in another member state. Having found that the condition amounted to a restriction on the right to free movement, the Court concluded that the presence condition was not justified. The aim of the measure was to ensure that there was a genuine link between a claimant and the competent Member State and to guarantee the financial balance of the national social security system, and this was legitimate: [89], [92]. The existence of such a link could effectively be established, in particular, by a finding that the person in question had been, for a reasonable period, actually present in the Member State: [93]. However the past presence condition in that case was too exclusive in nature because it:

The Court went on to say, at [96], that the existence of a connection between a claimant and the Member State might be established by other means. These included: the relationship between a claimant and the social security system of the competent Member State which was established in that case by the fact that the appellant was already entitled to DLA - [97]; that the appellant was credited with UK national insurance contributions - [98]; that she was dependent on her parents who received UK retirement pensions and had worked in the UK or received UK benefits before retiring - [100]; that the appellant had lived in the UK for a considerable part of her life - [101]. The Court said that those considerations also applied with regard to the objective of guaranteeing the financial balance in the social security system.

The Court concluded:

Mr Henderson does not agree that this reasoning applies to the present appeals because he submits that in the present case the legislature is to be afforded a greater margin of discretion than in Stewart. He relies on the decision of the Supreme Court in R (Lumsdon) v Legal Services Board [2015] 3 WLR 121 and says that this shows that the intensity of review of the proportionality of a measure which interferes with any of the EU’s fundamental “four freedoms”, is greater than that to be applied to a measure involving the exercise of a discretion under EU law, that Stewart was in the former category (involving a measure which interfered with the free movement of persons) and that this case is in the latter category. He relies on the following passages from the judgment in Lumsdon:

He submits that the PPT is a measure implementing the Qualification Directive in the exercise of a discretion “involving political, economic or social choices” and “a complex assessment”, particularly so in the current economic and fiscal climate where there is enormous pressure on the welfare budget and in the current political climate where there is considerable impetus to ensure that benefits are paid only to those with a clear and long-term connection to the UK. It follows, he submits, that the test of proportionality is whether the measure is “manifestly disproportionate” and this approach qualifies the approach in Hockenjos, (of which Stewart is an example, although not itself citing Hockenjos) which asks whether there were alternative means, which would not have been indirectly discriminatory, to achieve the desired aim.

I turn then to consider Lumsdon. The judgment of the Supreme Court was given by Lords Reed and Toulson, with whom the other members of the Court agreed. They gave detailed guidance to clarify the principle of proportionality in EU law, derived from the judgments of the Court of Justice as the “only authoritative interpreter” of the principle of proportionality: [23]. The starting point is that proportionality involves consideration of two questions: whether the measure in question is suitable or appropriate to achieve the objective pursued; and whether the measure is necessary to achieve that objective, or whether it could be attained by a less onerous method: [33]. The intensity with which the principle of proportionality is applied will vary according to a wide range of factors: [34].

That much is not in dispute in these appeals. The disputed issue is as to the intensity of review which should be applied to justification of the PPT. Paragraphs [73] and [74], relied upon by Mr Henderson, must be understood in the context of the analysis as a whole in which Lords Reed and Toulson considered three types of case: the review of EU measures, the review of national measures relying upon derogations from general EU rights, and the review of national measures implementing EU law. In relation to the second and third types, the Supreme Court’s broad summary was as follows:

The Court went on to say, at [50] that, in relation to the second type of case, compliance with the principle of proportionality is required not only in relation to permissible limitations on the four “fundamental freedoms” (free movement of workers, freedom of establishment, freedom to provide services, free movement of capital) but also other national measures falling within the scope of EU law including those which derogate from other rights protected by the Treaties such as the right to equal treatment or non-discrimination. In such cases, the Supreme Court said at [52] that the approach is as explained in Gebhard v Consiglio dell’Ordine degli Avvocati e Procuratori di Milano, C-55/94, [1995] ECR I-4165 at [37]:

As to the third and fourth of those conditions, at [55] the Court adopted the position as summarised by Advocate-General Sharpston in Commission of the European Communities v Kingdom of Spain C-400/08, [2011] ECR I-1915:

The Court went on to note that the justification for a restriction tends to be examined in detail, that a political justification may not be capable of being established by evidence but that evidence may be expected for an economic justification: [56]. Close examination is also given to whether other measures could have been equally effective but less restrictive, but the state is not required to exclude hypothetical alternatives: [61] and [63].

This is the context in which paragraphs [73] and [74] of the judgment are to be understood. They are the Court’s summary of the approach in the third type of case, ie to national measures implementing EU measures. The key distinction is between a directive which requires the national authority to exercise a discretion, in relation to which the test is that of “manifestly disproportionate”, and those where the state relies on a reservation or derogation in a directive in order to restrict one of the fundamental freedoms, in relation to which the proportionality of the measure will be subject to the same scrutiny as applies to the second type of case.

Turning then to the present appeals, I reject Mr Henderson’s submission that the applicable test is whether the PPT is “manifestly disproportionate”. My reasons are as follows.

Surprising though this may seem at first, the Qualification Directive is associated with promotion of the integration of the internal market and related social values. The preamble to the Directive has regard to the Treaty establishing the European Community and in particular Article 63.1(c), 2(a) and 3(a). Article 63 appeared under Title IV to the Treaty, headed “Visas, Asylum, Immigration and other policies relate to free movement of persons.” Article 61 described the purpose of adopting measures in accordance with Article 63 as:

Thus measures made under Article 63 are directly associated with the aim of ensuring free movement of persons.

In any event if, contrary to my conclusion above, the PPT is not directly concerned with the four “fundamental freedoms”, it is a measure falling with the scope of EU law which derogates from the right to equal treatment or non-discrimination which is protected by the Qualification Directive. The PPT is not an implementing measure: the Directive does not require member states to limit the application of Article 28 nor does it involve exercising a discretion in its implementation as considered in Lumsdon at [73].

I agree with Mr Henderson that a national measure can be treated as implementing EU legislation even though it is not characterised as such in domestic law. But the judgement of the CJEU in Aklagaren v Akerberg Fransson C-617/10, [2013] 2 CMLR 46, especially paragraphs [27] and [28] upon which Mr Henderson relies, does not assist him. The Directive in question in that case required member states to take measures to ensure collection of VAT and prevent evasion. The Court found that national measures imposing tax penalties and criminal proceedings for tax evasion had not been adopted to transpose the Directive, but the measures were designed to penalise an infringement of the Directive and so was intended to implement an EU obligation. The position there was that the national legislation in effect reflected the terms of the Directive. The same cannot be said of regulation 2 of the DLA Regulations.

I conclude therefore that the PPT falls within the second type of case addressed by the Supreme Court in Lumsdon and attracts the close scrutiny to be afforded to proportionality in such cases.

Mr Henderson submits that, even if “manifestly inappropriate” is not the applicable test, the PPT is nonetheless justified. He submits that, given the particular circumstances of refugees, there was no other way of establishing their link with the UK, that the Government considered and made appropriate exemptions from the test but there was no reason to exempt refugees, and that the period of required past presence was adopted so as to achieve consistency with other contributory benefits.

For reasons which I will now explain I conclude that, although the aim of establishing a sufficiently close link with the UK is a legitimate one, the Government has not established that the application of the PPT to refugees is a proportionate means of achieving that aim.

Proportionality involves consideration of whether the measure is appropriate to achieve the objective pursued and whether it is necessary to do so (or, in other words, whether the measure could be attained by a less onerous method): Lumsdon at [33] and [55]. The first limb does not require the measure to be the most appropriate of all possible measures, but it must not be inappropriate for the purposes of achieving the aim. As for the second limb, a measure is not necessarily disproportionate just because the least intrusive measure was not selected. Rather, the question is whether a less intrusive measure could have been used without unacceptably compromising the objective: Lumsdon at [105]).

In relation to the PPT, the rationale for a requirement of a substantial period of presence in order to establish a genuine or substantial link with the UK does not apply to refugees in the way that it does to other migrants. Refugees generally have no or little choice as to where they live and are unable to return to their country of origin. It cannot reasonably be said that most refugees have come to the UK on a temporary basis or for the purpose of accessing the benefits system, as other migrants may do. On the contrary, most refugees will have firmly established their genuine link with the UK by applying for and obtaining refugee status and severing their links with their home countries.

The Respondent has not disputed these propositions as to the different circumstances of refugees. In its Report on the Income-related Benefits Schemes (Miscellaneous Amendments)(No.3) Regulations 1994 (Cm 2609), which introduced the habitual residence test for income support, housing benefit and council tax benefit, the Social Security Advisory Committee (whose recommendations were accepted by the Government) made a similar point at paragraph 42:

In addition, as a result of the particular vulnerability of refugees (see the same Social Security Advisory Committee report), exemptions have been made from presence or residence tests in relation to some other benefits including a number of means-tested benefits: income support, jobseekers allowance, employment and support allowance, state pension credit and universal credit, and from entitlement rules in relation to other benefits requiring a claimant to have lived in the UK or Great Britain for a period of three months.

Mr Henderson says that these rules are not true exemptions. His argument is as follows. A person who is designated “a person from abroad” may not claim certain benefits. A “person from abroad” is defined as being a person who is not “habitually resident” in the Common Travel Area. Refugees are automatically considered “not a person from abroad” and so they do not need to demonstrate habitual residence. However refugees still have to meet all the substantive eligibility criteria for the benefits in question and this is consistent with them having to meet the PPT, which is a substantive eligibility criterion.

This analysis fails to address the concrete reality that in order to obtain some benefits refugees are not required to satisfy rules as to residence or presence which others must satisfy. These are, in substance, exemptions. The Social Security Advisory Committee the Government in its response accepted that these were exemptions. In any event, the Respondent’s technical approach breaks down in relation to other benefits such as child benefit and child tax credit, in respect of which a period of past residence is clearly a condition of entitlement from which refugees are exempt.

Mr Henderson also says that the reason why refugees are defined as “not a person from abroad” is because, by the time their asylum claims have been processed, they would have been resident in the UK for some time and “it would be obvious that they intend to reside permanently in the UK”. He points to the Government’s response to the Social Security Advisory Committee which made this point at paragraph 9(2). The same is true for the three month “living in” requirement. In making this submission Mr Henderson seeks to distinguish between the justification for not requiring refugees to satisfy the habitual residence test but requiring them to satisfy the PPT. In my view it undermines the rationale for applying the PPT to refugees because if it is clear that refugees intend to reside permanently in the UK, then it is difficult to see why this is not sufficient to establish their link with the UK.

Finally Mr Henderson’s submits that there is no particular reason to exempt refugees from the PPT because all those who would (other than for the PPT) be entitled to DLA are vulnerable and it would be hard to justify a rule by which some vulnerable claimants receive DLA and others do not. As the Social Security Advisory Committee said, and the Government at the time agreed, as a class refugees are particularly vulnerable. The point was explained by Blake J in MM (Lebanon) v SSHD [2013] EWHC 1900 (Admin) at [99]:

Mr Henderson says that the Respondent “would have explored” alternatives to the PPT but that “no better alternative policy was available which would have achieved the aims pursued”. The Respondent has produced no evidence of having considered the position of refugees in the context of the PPT. In the light of the considerations which I have set out above, the aim of the PPT could have been achieved without requiring refugees to satisfy it. Moreover, the Government’s response to Stewart (while that case is not on all fours with the present appeals) illustrates the possibility of other measures which could have been applied to refugees. Following the decision in Stewart the Respondent amended the DLA Regulations to add regulation 2A. This replaced the PPT for EU citizens exercising free movement rights with the test of demonstrating a genuine and sufficient link to the UK social security system. The Respondent has not explained why, as an alternative to a simple exemption of refugees from the PPT, other conditions could not have been formulated for refugees so as to provide an opportunity for them to satisfy the aim by other means. For instance, a “genuine and sufficient link” condition would have allowed refugees to rely on factors such as their links with others in the UK, receipt of other UK benefits, and a range of other relevant personal circumstances.

For the above reasons, I conclude that the PPT, in its application to refugees, is an inappropriate means of achieving the Government’s aim and the Respondent has failed to show that there would not have been a less restrictive way of achieving its aim.

Finally, in the light of the Respondent having failed to justify the measure by reference to its aim of establishing a link with the UK, it cannot be justified as a means of protecting the integrity of the benefits system. As the CJEU said in Stewart at [103],

The Respondent has not advanced any basis for suggesting that making different provision for refugees would adversely affect the integrity of the benefits system. It would lead to some additional benefits expenditure, although the Respondent has produced no evidence as to this nor how any such savings are to be balanced against the impact of the application of the test on refugees and their families.

In the light of my conclusions on discrimination contrary to the Article 28 of the Qualification Directive, I can address Article 14 relatively briefly.

The Appellants rely on Article 14 by reference to Article 1 Protocol 1. DLA falls within the scope of A1P1: Mathieson v SSWP [2015] UKSC 47 at [18]. The Respondent does not disagree with that but does not accept that A1P1 is engaged because there has been no assessment as to whether the Appellants satisfy the substantive conditions of eligibility for DLA. For present purposes, I assume that they do. Plainly if, on assessment, they are found not to do so then there could be no question of a violation of Article 14.

The Respondent does not dispute that the Appellants’ status as a refugee or the family member of a refugee constitutes “other status” for the purposes of Art. 14. That is plainly correct: see Hode and Abdi v UK (app no 22341/09).

A measure which is ostensibly of general application may engage Article 14 either because it disproportionately affects a particular group or because it treats people who are in a materially different situation as if they are the same: Burnip v SSWP [2012] EWCA Civ 629. It follows from my conclusions as to discrimination above, that the PPT discriminates (subject to justification) against refugees and their family members because they are disadvantaged by the application of apparently neutral criteria and because the test fails to recognise that refugees are in a materially different situation from others to whom it applies.

In Humphreys v Revenue and Customs Commissioners [2012] 1 WLR 1545 at [16] Baroness Hale said as follows, referring to the decision of the Grant Chamber in Stec v United Kingdom (2006) 43 EHRR 1017:

She went to observe that, that does not mean the justification for the measure should escape careful scrutiny. See also Mathieson v SSWP [2015] UKSC 47 at [25]-[26].

For the same reasons as I have given in relation to discrimination contrary to EU law, I conclude that the application of the PPT to refugees is not justified in the context of Article 14.

The First-tier Tribunal erred in law in finding that the PPT for refugees was proportionate under EU law. The test unlawfully discriminates against refugees and their family members contrary to Article 28(1) and 23(2) of the Qualification Directive. It also discriminates contrary Article 14 ECHR.

The First-tier Tribunal’s decision cannot stand in the light of its error and I set it aside pursuant to section 12(2)(a) Tribunals Courts and Enforcement Act 2007.

I must either then remit the appeals to the First-tier Tribunal or remake the decision. The only issue before the First-tier Tribunal was that of discrimination and so there is no point in my remitting the appeals. It is appropriate to make the decision that the First-tier Tribunal should have made.

In the light of my finding of discrimination contrary to the Qualification Directive and applying section 2(1) European Communities Act 1972, it is appropriate to disapply the offending provision in the DLA Regulations. See Hockenjos, at [88]. In relation to discrimination contrary to Article 14 ECHR, the same result follows from the application of section 6(1) Human Rights Act 1998.

There is no dispute that the Appellants satisfied the other conditions of regulation 2(1) of the DLA Regulations. So the decision which I make is that at the relevant time the Appellants satisfied the residence and presence conditions for DLA in section 71(6) Social Security Contribution and Benefits Act 1992.

The Secretary of State must now determine the Appellant’s claims for DLA in accordance with this decision. If the Appellants are unhappy with the decisions that are made by the Secretary of State, they will have a further right of appeal to the First-tier Tribunal.