R v Secretary of State for Work and Pensions and others

This appeal concerns the fact that entitlement to one of the components of one of the welfare benefits available in the United Kingdom, namely the individual element of child tax credit, is limited to a maximum amount, calculated as the amount payable in respect of two children. That limitation is challenged in these proceedings as being incompatible with the European Convention on Human Rights (“the Convention”, or “the ECHR”), as given effect by the Human Rights Act 1998. The appeal raises a number of important questions in relation to the relevant articles of the Convention, and in relation to the constitutional law of the United Kingdom.

In view of the length of this judgment, it may be helpful at the outset to explain how it is laid out, and the conclusions reached.

The United Kingdom has for many years operated a system of welfare benefits in order to support individuals and families. Most families with new claims for benefit are supported by universal credit, a holistic benefit which was established by the Welfare Reform Act 2012. But a majority of families receiving benefits at the present time are supported by a variety of longer-established benefits. This appeal is concerned primarily with one such benefit, known as child tax credit, but the aspect of it which is challenged is also a feature of universal credit. It is sufficient for the court to consider the position in relation to child tax credit, since it is common ground that the relevant considerations are the same in each case.

Child tax credit is a non-contributory benefit intended to provide financial support to families with children. It was introduced by the Tax Credits Act 2002 (“the 2002 Act”). It can be claimed by persons who are in work as well as by those who are not. Persons in work who are earning up to £16,385 per annum (or more, depending on their entitlement to working tax credit) can continue to receive child tax credit in full. Above that level of earned income, the amount of child tax credit received is reduced in proportion to the claimant’s income until it eventually tapers out altogether.

Child tax credit can be claimed either jointly by a couple or by a single person who is not entitled to make a joint claim: section 3(3) of the 2002 Act. In order to be entitled to child tax credit, the claimant, or either or both claimants in a couple, must be responsible for one or more children or “qualifying young persons”: section 8(1). Children are defined as persons aged under 16, and qualifying young persons are defined as young people aged 16 to 19 who are in “advanced education” or “approved training”. For the sake of simplicity, this judgment will refer to “children” as encompassing both categories. Thus, the person entitled to child tax credit is not the child, but the person responsible for him or her. For this purpose, a person is treated as responsible for a child who is normally living with him or her: the Child Tax Credit Regulations 2002 (SI 2002/2007) (“the 2002 Regulations”), regulation 3.

Child tax credit consists of three elements: (1) a “family element” of £545 per annum (all figures are as at 2020/21), (2) an “individual element” of £2,830 per annum in respect of each child, subject to the limitation described in the next two paragraphs, and (3) a “disability element” of £3,415 per annum in respect of each child who is disabled, and of £4,800 per annum in respect of each child who is severely disabled: section 9 of the 2002 Act.

A limitation on the maximum amount of the individual element is imposed, and certain exceptions to that limitation are allowed, under section 9(3A) and (3B) of the 2002 Act, as amended by section 13(4) of the Welfare Reform and Work Act 2016 (“the 2016 Act”). They provide:

The effect of those provisions is that, in calculating the maximum amount of a person’s entitlement to the individual element, no account is taken of third or subsequent children born on or after 6 April 2017, unless one of the prescribed exceptions apply. Those exceptions are set out in the 2017 Regulations. They allow a person or couple to claim an additional individual element of child tax credit for a third or subsequent child born on or after 6 April 2017 for whom they are responsible in the cases of multiple births, adoption, non-parental caring arrangements and non-consensual conception.

Child tax credit is separate from, and additional to, numerous other benefits which are payable to families with children. For example, child benefit is payable to the person responsible for a child, at a rate of £1,076.40 per annum for the first child and £712.40 per annum for each subsequent child. There is no limit referable to the number of children in respect of whom child benefit is paid. A person’s entitlement to housing benefit, which is payable in respect of the cost of housing, will also increase if he or she requires a larger property, and therefore has higher housing costs, as a result of children being added to the family, without any limit referable to the number of children. Other support available to families with children includes 70% assistance with childcare costs as part of working tax credit, discretionary assistance with childcare costs for those working less than 16 hours per week, free childcare for younger children, free school meals, and a variety of other benefits. None of these benefits is subject to any limitation relating to the number of children.

In addition to the limitation on a person’s entitlement to the individual element of child tax credit, there is also an overall limitation or “cap” on the total amount of welfare benefits which a person may receive (subject to various exceptions). That cap was challenged unsuccessfully in earlier proceedings in this court, where it was argued to be incompatible with the Convention: see R (SG) v Secretary of State for Work and Pensions (Child Poverty Action Group intervening) [2015] UKSC 16; [2015] 1 WLR 1449 (“SG”), which concerned the cap as originally introduced by the Welfare Reform Act 2012, when it was set at a figure equivalent to the net median earnings of working households, and R (DA) v Secretary of State for Work and Pensions (Shelter Children’s Legal Services intervening) [2019] UKSC 21; [2019] 1 WLR 3289 (“DA”), which concerned a revised version of the cap which was introduced by the 2016 Act, when it was set at the figure of £20,000 per annum (or £23,000 per annum for claimants living in London).

Although this appeal is a test case brought with the support of a campaigning organisation, the Child Poverty Action Group, acting as the appellants’ solicitors, in order to challenge legislation which that organisation unsuccessfully opposed during its passage through Parliament, there are a number of individual appellants who are said to be the victims of violations of their Convention rights. The first group of appellants comprises SC and three children for whom she is responsible. The youngest of those children was born after 6 April 2017, and SC therefore receives no additional child tax credit by reason of the birth of that child. According to the witness statement which she provided when these proceedings were before the Administrative Court, she was receiving at that time welfare benefits in the form of income support, child tax credit and child benefit totalling £12,081.68 per annum, plus housing benefit which paid in full her rent of £5,720 per annum. She also received some financial support from the father of her youngest child. She stated that she was making ends meet, and that she managed but it was not easy. She was and remains subject to the overall cap on welfare benefits of £20,000 per annum, inclusive of housing benefit, and at the time of her statement was receiving £17,801.68.

The second group of appellants are CB and five children for whom she is responsible, the youngest of whom was born after 6 April 2017. According to the witness statement which she provided to the Administrative Court, she was at that time receiving welfare benefits in the form of maternity allowance, working tax credit, child tax credit, child benefit and housing benefit totalling £544.59 per week (as a recipient of working tax credit, CB was not subject to the benefit cap). She did not receive any financial support from the fathers of her children. She stated that she was budgeting to the best of her ability, but that her children were unable, for example, to emulate friends who held their birthday parties at commercial venues.

The proposal to introduce the limitation on entitlement to child tax credit was announced by the Government on 8 July 2015 as one of a number of measures intended to fulfil a commitment made in the Conservative Party manifesto for the 2015 General Election to reduce spending on welfare benefits by £12 billion. The Bill was introduced in the House of Commons the following day. In accordance with section 19(1)(a) of the Human Rights Act, the Minister in charge of the Bill made a statement to Parliament that in his view its provisions were compatible with Convention rights.

The Bill received its Second Reading in the House of Commons on 20 July 2015. Parliament was provided with an impact assessment produced by the Treasury and the Department for Work and Pensions. It identified the problem which the Bill was intended to address:

The impact assessment noted that tax credit expenditure had “more than trebled in real terms between 1999/00 and 2010/11, with total expenditure in 2014/15 estimated to be around £30 billion”. The proposed limitation on entitlement was “part of a package which will deliver a more sustainable welfare system and return expenditure on tax credits to 2007/08 levels in real terms”. The option of doing nothing was rejected on the grounds that it was unfair to families not eligible for state support and to the taxpayer, and would not return welfare spending to a sustainable level. The impact assessment also stated that delivering welfare savings was “a vital part of the government’s deficit reduction plan. Had the budget not announced such significant welfare savings, steep reductions in public service spending would have been required - or higher borrowing and debt or higher taxes”. It was estimated that the measure would result in annual savings of £1.365 billion by 2020/21 and that those savings would continue to rise thereafter. The number of households which would be affected by the limitation was accurately predicted. It was also noted that “women … are more likely to be affected, in the absence of behavioural change”, since “[a]round 90% of lone parents are women, and a higher proportion of this group are in receipt of CTC [child tax credit]”.

The Bill was the subject of considerable scrutiny and debate during its passage through Parliament. The proposed limitation on child tax credit, in particular, was politically contentious. Numerous documents were placed before Parliament to assist members in their consideration of the Bill, including various impact assessments, written evidence received from over 80 interested organisations and individuals, and briefing papers and notes provided by the libraries of both Houses. Parliamentary scrutiny included consideration of the Bill by the Joint Committee on Human Rights. A Government memorandum to the committee anticipated arguments made in these proceedings that the limitation would affect large families, and was more likely to affect women (because they were more likely to claim child tax credit), and was therefore incompatible with the ECHR and with the United Nations Convention on the Rights of the Child (“the UNCRC”). It rejected arguments based on article 14 of the ECHR on the basis that the measure pursued a legitimate aim and was necessary and proportionate. The stated justification was that:

Before the House of Commons Bill Committee (“the committee”), it was explained that the current level of spending on child tax credit, amounting to £30 billion per annum, was unsustainable. The limitation on entitlement to the individual element of child tax credit to the amount payable in respect of two children was proposed in light of the fact that the average number of dependent children in families in the UK was 1.7. In order to give families time to prepare, the changes would not be brought into effect until April 2017, and would only apply in respect of children born after that date.

The committee received submissions from the Equality and Human Rights Commission which stated that the proposed changes to child tax credit might affect the living standards of poor families with more than two children. The Commission expressed concern that, although the Government asserted that the proposals in the Bill were in the best interests of children, since the savings achieved by reducing spending on welfare would allow it to protect expenditure on education, childcare and health, and would improve the country’s economic situation, its impact assessment did not address the consequences of the proposals for the children directly affected.

The committee also received a large volume of other evidence about the potential impact of the Bill, including evidence concerned with the impact of the limitation upon children in families which might be affected by it. It included evidence from numerous organisations concerned with social policy, and more particularly with children and poverty, including the Child Poverty Action Group. The committee was provided with a 283 page document containing written submissions from that organisation and others, which was also made available to other members of both Houses of Parliament. The committee devoted the whole of one of its sessions to the proposed limitation. Amendments to the Bill moved by the Opposition in committee to retain entitlement to the individual element of child tax credit (and the child element of universal credit) without such a limitation, or to set the limitation at a greater number of children than two, were either defeated on a vote or withdrawn after debate. An Opposition amendment at the report stage in the House of Commons, to leave the current arrangements for child tax credit in place, was also defeated.

The Bill was then considered in the House of Lords, where the debate echoed that in the House of Commons. Evidence about the potential impact of the Bill on child poverty was circulated to all members. The proposed limitation, and the proposed exceptions to it, were extensively debated in committee and were the subject of public consultation. The ECHR and the UNCRC were prayed in aid by those opposing the Bill. The draft regulations implementing the exceptions (para 8 above) were referred to the independent Social Security Advisory Committee for its comments. The final debate in the House of Commons was on 23 February 2016, following which the Bill received Royal Assent on 16 March 2016. Steps were taken to ensure that all recipients of child tax credit were informed of the impending changes.

In these proceedings, it is argued that the limitation on entitlement to the individual element of child tax credit to the amount payable in respect of two children, where subsequent children have been born on or after 6 April 2017 and fall outside the scope of the prescribed exceptions, is incompatible with the adult appellants’ rights (1) under article 8 of the Convention, to respect for their private and family life, (2) under article 12, to the right to found a family, (3) under article 14 taken together with article 8, to the enjoyment of the right to respect for their private and family life without discrimination, and (4) under article 14 taken together with A1P1, to the enjoyment of the right to the peaceful enjoyment of their possessions without discrimination. It is also argued that the limitation is incompatible with the child appellants’ rights under article 14 taken together with article 8. It will be necessary to consider each of these arguments in turn. The appellants’ case in relation to article 14 is supported by the Equality and Human Rights Commission, which appears as an intervener.

The court has been requested to grant leave to allow the adult appellants to appeal on two additional grounds. The first is that the limitation is incompatible with their rights under article 14 taken together with article 9 (the right to freedom of thought, conscience and religion). The second is that the limitation is incompatible with their rights under article 14 (taken together, presumably, with article 8) in relation to their control of their bodies. Such leave is however refused. The proposed grounds of appeal do not have any basis in the pleadings or the evidence. They do not appear to the court to arise on the facts of the appellants’ cases, for the reasons explained in para 28 below, or to be arguable in any event.

The judge, Ouseley J, dismissed the claims for reasons given in a judgment dated 20 April 2018: see [2018] EWHC 864 (Admin); [2018] 1 WLR 5425. An appeal against that decision was dismissed by the Court of Appeal (Patten, Leggatt and Nicola Davies LJJ) on 16 April 2019, for reasons explained in a judgment given by Leggatt LJ: see [2019] EWCA Civ 615; [2019] 1 WLR 5687. This court has been greatly assisted by both the judgments below.

Article 8 provides:

Article 8 has never been held to impose an obligation on the state to have in place a programme of financial support for private or family life: see, for example, Petrovic v Austria (1998) 33 EHRR 14, para 26 (“Petrovic”). Accordingly, counsel did not seek to argue that article 8 directly imposed an obligation on the state to provide an unlimited entitlement to the individual element of child tax credit. Instead, counsel argued that article 8 indirectly imposed such an obligation.

First, focusing on the position of the adults receiving child tax credit, counsel argued that the limitation was known and intended to affect their reproductive choices, by discouraging them from having more than two children. They had to decide whether to engage in sexual abstinence, contraception or abortion, or to have another child for whom they would receive no additional support in the form of child tax credit. That was incompatible with respect for their dignity. Even if a woman decided to continue with a pregnancy, she would have to consider whether to have an abortion: something which, absent the limitation, she might not have needed to do. That affected her psychological integrity, which was an important aspect of private life. Reliance was placed on the case of Botta v Italy (1998) 26 EHRR 241, where the European Court of Human Rights observed at para 33 that the positive obligations imposed by article 8 “may involve the adoption of measures designed to secure respect for private life even in the sphere of the relations of individuals between themselves”.

Secondly, focusing on the position of children affected by the limitation (ie children living in households which include a third or subsequent child, born on or after 6 April 2017, not falling within one of the prescribed exceptions), counsel argued that the failure to make additional payments of child tax credit in respect of those children had a damaging effect on their integration into their families. Reliance was placed on the case of Marckx v Belgium (1979) 2 EHRR 330, where the European court stated at para 31 that “respect for family life implies in particular, in the court’s view, the existence in domestic law of legal safeguards that render possible, as from the moment of birth, the child’s integration in its family”.

I am unable to accept those submissions. Considering first the argument relating to adults, the first point to be made is that the argument does not arise on the facts of this appeal. SC states that, when she became pregnant with her youngest child, “the pregnancy was unplanned so even if I had heard of the two child rule, it would not have resulted in me not getting pregnant and, because of my views on abortion, it would not have made a difference to me continuing with the pregnancy”. CB states that “I was not aware of the two child rule at the time that I became pregnant but, as the pregnancy was completely unplanned and I was on the pill, if I had known about it, it would not have made a difference to my behaviour nor, because of my ethical principles, would it have changed my decision to keep the baby”. So the measure had no effect on either woman’s decision-making in relation to the birth of children.

Furthermore, the factual premise of the argument is that the limitation on entitlement to child tax credit was known and intended to have the effect of discouraging adult recipients from having more than two children. That contention was rejected by both courts below in the light of the relevant evidence.

Counsel relied before the judge on statements in the impact assessment (paras 14-15 above) that “encouraging parents to reflect carefully on their readiness to support an additional child could have a positive effect on overall family stability”, and that “in practice people may respond to the incentives that this policy provides and may have fewer children”. Those statements were argued to demonstrate that the measure was intended to influence intimate behaviour by creating an incentive for people receiving child tax credit to have smaller families. The judge accepted that it was anticipated that an effect of the measure might be that some people would decide not to have a child when they might otherwise have done so. But he rejected the contention that discouraging larger families could properly be described as an aim of the legislation. He also found that there was no evidence from which it could be inferred that the legislative change was actually having an effect on decisions made about family size. He noted that studies in the United States of the impact of analogous legislation had found little or no effect on the number of children born per family.

The Court of Appeal saw no basis for challenging those findings. Leggatt LJ concluded that there was no reason to attribute to the Government or to Parliament any aims in introducing the limitation other than those which were repeatedly stated during the legislative process. Those aims included encouraging people in receipt of tax credits to consider, before having additional children, whether they could afford to support them, and incentivising people to support themselves and their families through work. But the aims of the measure did not include any goal of reducing the size of families. The purpose was not to discourage people on lower incomes from having larger families, but to reduce public expenditure by limiting welfare benefits and to leave choices about family size to the individuals concerned in the knowledge of what state support would be available.

As explained at paras 166 and 174-176 below, the relevant intention, when one is considering the intention of primary legislation, is that of Parliament, not that of the Government. Parliament’s intention is ascertained primarily from the language which it has used. It is also legitimate to look at other materials in order to identify the problem or “mischief” which Parliament was seeking to remedy. In the present case, there is nothing in the legislation itself which indicates an intention to interfere with the reproductive choices of recipients of child tax credit. Nor is there the slightest indication in the other material before the court, summarised at paras 13-20 above, that their reproduction rate was regarded as a problem which needed to be addressed. The most that can be said is that one of the effects of the legislation, which Parliament can be taken to have intended, is that recipients of child tax credit have to take decisions about whether or not to have more than two children in the knowledge that their income, to the extent that it is derived from child tax credit, will not increase as a consequence of the birth of a third or subsequent child, unless one of the exceptions applies. But it is an ordinary fact of life that couples take decisions about the size of their families in the knowledge that their income will not automatically increase as the number of their children increases. Article 8 does not oblige the state to alter that situation by ensuring that parents are provided with additional income for every additional child that they may choose to have.

Considering next the argument relating to children affected by the limitation, it rests on an assertion that the failure of the state to pay additional child tax credit on the birth of a third or later child has a damaging effect on their integration into their families. There is no evidence to support that assertion, and there is no reason to believe that it is true. Counsel’s submissions suggested that such a child would be treated differently from the other children in the family because of the absence of an additional amount of child tax credit, but there is no reason to suppose that families behave in that way. The amount of the child tax credit which a parent receives is not hypothecated to particular children in their household. Parents are unlikely to respond to the demands placed on a limited budget by treating their third child less generously than their first and second children. Unsurprisingly, the statements of SC and CB contain no such suggestion. On the contrary, SC states that “it is not so much the baby [the third child] who loses out but rather the older children”. As an example, she states that her eldest child cannot attend school clubs over the summer, as they cost £20 per day. CB states that the fifth child’s arrival has resulted in “the same amount having to stretch to five children rather than four”. SC’s and CB’s statements contain nothing to suggest that their youngest children have not been fully integrated into the life of their families.

Article 12 provides:

Counsel for the appellants made no separate submissions to support the contention that the limitation violated article 12, but relied on his submissions in relation to article 8. For the reasons already explained, those submissions cannot be accepted. They are no more persuasive in the context of article 12. According to the case law of the European court, that article only protects the right to found a family within marriage: Goodwin v United Kingdom (2002) 35 EHRR 18, para 98. Neither claimant has any intention of marrying, or founding a family with, the father of her youngest child (or, so far as appears from the evidence, anyone else). Even if they had such an intention, article 12 has been held not to impose a positive obligation on the state to provide the material means which would enable them to found a family: Cannatella v Switzerland (Application No 25928/94) (unreported), European Commission on Human Rights, 11 April 1996. In short, article 12 has no application.

Article 14 provides:

The general approach adopted to article 14 by the European court has been stated in similar terms on many occasions, and was summarised by the Grand Chamber in the case of Carson v United Kingdom (2010) 51 EHRR 13, para 61 (“Carson”). For the sake of clarity, it is worth breaking down that paragraph into four propositions:

“The court has established in its case law that only differences in treatment based on an identifiable characteristic, or ‘status’, are capable of amounting to discrimination within the meaning of article 14.”

“Moreover, in order for an issue to arise under article 14 there must be a difference in the treatment of persons in analogous, or relevantly similar, situations.”

“Such a difference of treatment is discriminatory if it has no objective and reasonable justification; in other words, if it does not pursue a legitimate aim or if there is not a reasonable relationship of proportionality between the means employed and the aim sought to be realised.”

“The contracting state enjoys a margin of appreciation in assessing whether and to what extent differences in otherwise similar situations justify a different treatment. The scope of this margin will vary according to the circumstances, the subject matter and the background.”

I shall address those issues first in relation to the complaints made on behalf of the adult appellants, and then in relation to the complaints made on behalf of their children.

According to the case law of the European court, the alleged discrimination must relate to a matter which falls within the “ambit” of one of the substantive articles. This is a wider concept than that of interference with the rights guaranteed by those articles, as Judge Bratza explained in his concurring judgment in Adami v Malta (2006) 44 EHRR 3, para 17.

For example, in Petrovic the refusal to grant a father a parental leave allowance which was paid to mothers was held not to constitute an interference with the right guaranteed by article 8, “since article 8 does not impose any positive obligation on states to provide the financial assistance in question” (para 26). Nevertheless, since “[b]y granting parental leave allowance states are able to demonstrate their respect for family life within the meaning of article 8”, it followed that “the allowance therefore comes within the scope of that provision” (para 29), with the consequence that article 14 taken together with article 8 was applicable.

A number of other judgments of the European court confirm that welfare benefits which are designed to facilitate or contribute to family life, by supporting families with children, are likely to fall within the ambit of article 8, for the purpose of complaints under that article taken together with article 14. Examples include Okpisz v Germany (2005) 42 EHRR 32 (child benefits), Niedzwiecki v Germany (2005) 42 EHRR 33 (child benefits), Weller v Hungary (Application No 44399/05) (unreported) given 31 March 2009 (maternity benefits) and Fawsie v Greece (Application No 40080/07) (unreported) given 28 October 2010 (an allowance for large families). Since child tax credit is payable only to adults who are responsible for children, and is intended to provide financial support to families with children, I do not see any convincing basis for distinguishing it from the benefits with which those cases were concerned. I therefore conclude, in agreement with the Court of Appeal but contrary to the view of the judge, that the complaint of the adult appellants in the present case falls within the ambit of article 8 taken together with article 14.

The Court of Appeal, differing in this respect from the judge, held that the complaint also fell within the ambit of article 14 taken together with A1P1, on the basis that persons claiming child tax credit are denied the individual element in respect of a third and subsequent child on an allegedly discriminatory ground. The court considered that that conclusion followed from the reasoning of the Grand Chamber in its admissibility decision in Stec v United Kingdom (2005) 41 EHRR SE18, para 54 (omitting citations):

It was, however, accepted by counsel for the appellants before the Court of Appeal that only the persons entitled to claim child tax credit - that is to say, the adult appellants - could bring a complaint on this basis. There was no departure from that position before this court, where the focus of counsel’s submissions was on article 14 in conjunction with article 8.

It is argued that the measure in question treats women differently from men, and that the adult appellants, SC and CB, have therefore been the victims of discrimination on the ground of sex. Sex is undoubtedly a relevant characteristic or status: discrimination on the ground of sex is expressly prohibited by article 14.

No difference in the treatment of men and women is apparent on the face of the measure: the legislation applies “in the case of a person or persons entitled to child tax credit where the person is, or either or both of them is or are, responsible for a child or qualifying young person born on or after 6 April 2017 … unless - … he is (or they are) claiming the individual element of child tax credit for no more than one other child or qualifying young person”. No distinction is drawn according to whether “the person” is a man or a woman.

Counsel argues, however, that although the legislation is couched in neutral terms, statistically it affects more women than men. That is accepted on behalf of the Secretary of State. On that basis, it is argued that there is indirect discrimination, as explained by the European court in the case of DH v Czech Republic (2007) 47 EHRR 3. Consideration of this argument requires an examination of the concept of indirect discrimination in the Convention case law.

“Generally, in order for an issue to arise under article 14 there must be a difference in the treatment of persons in analogous, or relevantly similar, situations”: Guberina v Croatia (2016) 66 EHRR 11, para 69 (“Guberina”). That is the situation in an ordinary case of direct discrimination: there is an actual difference in treatment between comparable cases, directly based on a prohibited ground of discrimination.

In addition, “the right not to be discriminated against … is also violated when states without an objective and reasonable justification fail to treat differently persons whose situations are significantly different”: Guberina, para 70. In other words, article 14 may impose a positive duty to treat individuals differently in certain situations. One of the judgments cited by the court was Thlimmenos v Greece (2000) 31 EHRR 15, which illustrates the nature of the discrimination in such cases. The applicant had received a criminal conviction as a result of his refusal, for religious reasons, to wear a military uniform. He was refused admission to the profession of chartered accountant because he had been convicted of a serious crime. Since his conviction did not imply any dishonesty or moral turpitude which might render a person unsuitable to enter the profession, the court held that “there existed no objective and reasonable justification for not treating the applicant differently from other persons convicted of a felony” (para 47). The discrimination lay in not introducing an exception to a general rule.

Thirdly, “[t]he court has also accepted that a general policy or measure that has disproportionately prejudicial effects on a particular group may be considered discriminatory notwithstanding that it is not specifically aimed at that group, and that discrimination potentially contrary to the Convention may result from a de facto situation. This is only the case, however, if such policy or measure has no ‘objective and reasonable’ justification, that is, if it does not pursue a ‘legitimate aim’ or if there is not a ‘reasonable relationship of proportionality’ between the means employed and the aim sought to be realised”: Guberina, para 71. The judgments cited in support of that proposition included DH v Czech Republic. This is what is described in the Convention case law as “indirect discrimination”. It can arise in a situation where a general measure or policy has disproportionately prejudicial effects on a particular group. It is described as “indirect” discrimination because the measure or policy is based on an apparently neutral ground, which in practice causes a disproportionately prejudicial effect on a group characterised by a salient attribute or status.

The concept of indirect discrimination has only gradually come to be recognised by the European court. An early example is Hoogendijk v The Netherlands (2005) 40 EHRR SE22, where a requirement to qualify for a social security benefit affected more women than men. The court held that “where an applicant is able to show, on the basis of undisputed official statistics, the existence of a prima facie indication that a specific rule - although formulated in a neutral manner - in fact affects a clearly higher percentage of women than men, it is for the respondent Government to show that this is the result of objective factors unrelated to any discrimination on grounds of sex” (p 207). The government having failed to do so on the facts of the case, the court held that “the question therefore arises whether there is a reasonable and objective justification for the introduction of [the measure in issue]”. On the facts, it was held that there was.

The Grand Chamber adopted a broadly similar approach in DH v Czech Republic, which concerned indirect discrimination on the ground of ethnic origin. That aspect of the case was highly significant, since a difference in treatment based exclusively or to a decisive extent on ethnic origin is incapable of being justified (as the court noted at para 176). As in Hoogendijk, the starting point was for the applicants to submit evidence (again based on official statistics) giving rise to a prima facie case, or “presumption”, of discrimination on the ground of ethnic origin (paras 180, 189 and 195). The onus then shifted to the respondent government to “show that the difference in the impact of the legislation was the result of objective factors unrelated to ethnic origin” (para 195). In that regard, the government argued that the relevant difference - the disproportionate number of Roma children attending schools for children with special needs - was the result of their lower intellectual capacity, as assessed by neutral testing, and their consequent placement in appropriate schools. The court then had to consider whether that constituted an objective and reasonable justification: whether the government was pursuing a legitimate aim, and whether there was a reasonable relationship of proportionality between the means employed and the aim sought to be realised (para 196). Although the aim was accepted to be legitimate, the court concluded, in view of inadequacies in the testing regime, that the results of the tests were not capable of constituting an objective and reasonable justification for the difference in treatment.

A different type of situation arose in SAS v France (2014) 60 EHRR 11, which concerned a measure which made it unlawful for anyone to conceal their face in public places. The measure had a disproportionate impact on Muslim women, and was argued to constitute indirect discrimination on the ground of religion. The relevant question was therefore whether the measure pursued a legitimate aim and was proportionate (para 161). It was held that the measure met those requirements.

Following the approach laid down in these and other cases, it has to be shown by the claimant that a neutrally formulated measure affects a disproportionate number of members of a group of persons sharing a characteristic which is alleged to be the ground of discrimination, so as to give rise to a presumption of indirect discrimination. Once a prima facie case of indirect discrimination has been established, the burden shifts to the state to show that the indirect difference in treatment is not discriminatory. The state can discharge that burden by establishing that the difference in the impact of the measure in question is the result of objective factors unrelated to any discrimination on the ground alleged. This requires the state to demonstrate that the measure in question has an objective and reasonable justification: in other words, that it pursues a legitimate aim by proportionate means (see, in addition to the authorities already cited, the judgment of the Grand Chamber in Biao v Denmark (2016) 64 EHRR 1, paras 91 and 114).

The question whether that test is satisfied in the present case, in relation to the fact that the limitation on the individual element of child tax credit affects more women than men, is discussed below at paras 188-199.

The child appellants claim to be the victims of discrimination contrary to article 14 read together with article 8, first as children, as compared with adults, and secondly as children with two or more siblings, as compared with children who have fewer than two siblings. It is necessary to consider these complaints separately.

It is argued that the child appellants have been treated differently from adults because they are children, and that they are therefore the victims of direct discrimination. It is also argued that the child appellants have been the victims of indirect discrimination against children as compared with adults. There is no doubt that discrimination on the ground of age falls within the scope of article 14.

The measure in question is said to be directly discriminatory because it excludes third and subsequent children in a household from the scope of benefits intended to provide children with financial support, whereas there is no corresponding exclusion of adults from the scope of benefits designed to provide adults with equivalent support, such as income support, jobseeker’s allowance, employment and support allowance, working tax credit and pension credit.

This argument is unfounded in fact, and depends on a false comparison for such plausibility as it may appear to have. The welfare benefits payable to adults, which counsel treated as analogous to child tax credit, are paid to adults in order to support their individual needs. They are therefore payable to individual adults whether they live alone or with others, and are calculated on an individual basis. Children have no entitlement to receive welfare benefits: benefits are paid instead to the adults who are responsible for them. Child tax credit, in particular, is paid to the responsible adult as a lump sum in respect of the children living with him or her. As was explained in paras 6-9 above, the limitation on the individual element of child tax credit sets a cap on the maximum amount of one part of one benefit which an adult responsible for children can receive. As was explained in para 33 above, the element of child tax credit which is subject to the limitation is not hypothecated to the care of particular children to the exclusion of other children in the household. The limitation does not, therefore, exclude any children from the scope of the support provided by child tax credit, let alone from the scope of the support provided by the benefits system as a whole. As was explained at para 9 above, and as is illustrated by the facts of these appeals (paras 11-12 above), adults who are responsible for children, and meet the relevant qualifying conditions, are eligible to receive a wide variety of benefits designed to support families, such as child tax credit, child benefit, housing benefit, assistance with childcare costs, free childcare and free school meals. The idea that third and subsequent children are excluded from the scope of benefits is therefore mistaken, and the argument that there is, on that basis, direct discrimination against children as compared with adults is not made out.

It is also necessary to bear in mind that not all differences in treatment are relevant for the purposes of article 14. The difference is only relevant, for the purpose of assessing whether there has been discrimination, if the claimant is comparing himself with others who are in a relevantly similar situation. An assessment of whether situations are “relevantly” similar generally depends on whether there is a material difference between them as regards the aims of the measure in question.

Considering the comparability of children and adults in the light of the aims of the limitation, those aims can be summarised as (1) maintaining public expenditure on welfare benefits at a sustainable level, and (2) requiring adults receiving welfare benefits, like other adults, to make choices as to whether to have a larger than average number of children without a guarantee from the state that their income will rise substantially with the birth of every additional child. In relation to the first of those aims, it was noted at para 15 above that expenditure on child tax credit had more than trebled in real terms between 1999/00 and 2010/11. There is no evidence that there has been a comparable increase in expenditure on the other benefits with which counsel sought to compare child tax credit. In relation to the second aim, it was the automatic ratcheting of child tax credit in line with the number of children for whom a person was responsible, without any limit, under the law as it stood prior to the 2016 Act which prompted the legislation that is challenged in these proceedings. Since the adult benefits mentioned by counsel are not linked to the number of children for whom a person is responsible, they do not give rise to any comparable issue, and are therefore not relevantly analogous.

The limitation is said to be indirectly discriminatory because the households affected by it contain a larger number of children than adults. That is, of course, an inevitable consequence of the measure: households affected by it necessarily contain at least three children, but generally contain only one or two adults, depending on whether the household is headed by a single parent or a couple. In 2015/16, statistics indicated that there were 1.4m adults and at least 3m children living in households claiming child tax credit where there were more than two children.

I should observe at the outset that there may be an issue as to the scope of the concept of indirect discrimination. As explained earlier, the concept is concerned with measures which, although neutral in appearance, have a disproportionately prejudicial impact upon a group sharing a common characteristic. Every case to date in which the European court has treated the concept as relevant has concerned a group sharing a common characteristic corresponding to a “suspect” ground of differential treatment such as sex, sexual orientation, ethnic origin, nationality, religion or disability. “Suspect” grounds are discussed at paras 100-113 below. They do not include age: see para 114. The view has been expressed that indirect discrimination is confined to “suspect” grounds: Theory and Practice of the European Convention on Human Rights, 5th ed (2018), ed Van Dijk and others, p 1006. Without necessarily going so far, one might in any event question whether children, considered as a whole, constitute a group of the relevant kind. I do not, however, need to examine that point, which was not raised in argument.

There is in any event a fundamental problem with the argument. Child tax credit does not affect children and adults in comparable ways, as has been explained. It is not paid to adults for their own benefit, but in order to assist them in meeting the needs of the children for whom they are responsible. A rule which limits the amount of child tax credit affects the children in the household, since it limits the amount of money which the responsible adults can spend on their care. It does not have any comparable effect upon the adults themselves.

I should add, for the sake of completeness, that if one were to accept that children and adults were affected by the limitation in comparable ways, the issues arising in considering whether the measure had an objective and reasonable justification would be much the same as those considered at paras 200-209 below, and would lead to the same conclusion.

For these reasons, I reject the contention that the measure in question results in a difference in treatment between children as a class and adults as a class, and need not consider that aspect of the case further.

Counsel’s final argument is that the measure in question treats children with two or more siblings - or, as I shall re-formulate the category, children living in households containing more than two children - differently from other children. That is the effect of the terms of the legislation: as explained in para 8 above, it limits an adult claimant’s entitlement to the individual element of child tax credit to the amount payable in respect of two children, unless one of the prescribed exceptions applies. I have to confess to some doubt as to whether counsel for the child appellants has demonstrated sufficiently the extent to which the limitation has affected their family life for the purposes of article 8: their complaint has been presented as being primarily of a financial nature. However, I do not dismiss the complaint on that basis.

The judge considered with care the question whether being a child with two or more siblings constituted a relevant characteristic or status for the purposes of article 14. He considered that a relevant characteristic or status had to have an existence separately from the difference in treatment: otherwise, the requirement of a relevant status would cease to be distinct from the existence of a difference in treatment, and article 14 might as well stop after the words:

Considering the facts of the case, the judge pointed out that the application of the limitation does not depend on the number of siblings that a child has, but on the number of children for whom the claimant is responsible. There is no requirement that those children should be siblings. Nor are siblings taken into account if someone else is responsible for them. SC, for example, has three other children who do not live with her and are therefore irrelevant to her entitlement to child tax credit. Nor does referring to siblings capture the age requirements of the legislation: siblings over the age of 19 are never relevant, and siblings between the ages of 16 and 19 are relevant only if they are in “advanced education” or “approved training”, so as to count as qualifying young persons. The judge considered that the child appellants either had to argue for a status which was too broadly defined to correspond to the difference in treatment complained of, or to define the status in accordance with the terms of the legislation, in which event there was no distinction between the relevant status and the difference in treatment. He concluded that being a child with multiple siblings was not a relevant status for the purpose of article 14.

The Court of Appeal was able to consider this issue in the light of the discussion of article 14 in R (Stott) v Secretary of State for Justice [2018] UKSC 59; [2020] AC 51. Leggatt LJ agreed with the judge that, in article 14, the words from “on any ground such as” to “or other status” (para 36 above) were intended to add something to the requirement of discrimination. It followed that status could not be defined solely by the difference in treatment complained of: it must be possible to identify a ground for the difference in treatment in terms of a characteristic which was not merely a description of the difference in treatment itself. On the other hand, he also observed that there seemed to be no reason to impose a requirement that the status should exist independently, in the sense of having social or legal importance for other purposes or in other contexts than the difference in treatment complained of. In that regard, Leggatt LJ referred to some illustrations in the European and domestic case law, such as the judgment of the European court in Paulik v Slovakia (2006) 46 EHRR 10, where “there was no suggestion that the distinction relied upon had any relevance outside the applicant’s complaint but this did not prevent the court from finding a violation of article 14” (Clift v United Kingdom (Application No 7205/07) (unreported) given 13 July 2010, para 60).

Applying that approach to the facts of the case, Leggatt LJ agreed with the judge that the term “sibling” was not strictly apt, as what mattered under the legislation was the number of children for whom the claimant was responsible, rather than the relationship between those children. But the basic distinction which the legislation sought to draw was a simple one, between households containing one or two children, and households containing more than two children. Being a child member of a household containing more than two children could be regarded as an individual characteristic or status for the purposes of article 14. That was so even if that status was given more precise definition by the legislation.

I respectfully agree with that reasoning, and with that conclusion. I would add that the issue of “status” is one which rarely troubles the European court. In the context of article 14, “status” merely refers to the ground of the difference in treatment between one person and another. Since the court adopts a stricter approach to some grounds of differential treatment than others when considering the issue of justification, as explained below, it refers specifically in its judgments to certain grounds, such as sex, nationality and ethnic origin, which lead to its applying a strict standard of review. But in cases which are not concerned with so-called “suspect” grounds, it often makes no reference to status, but proceeds directly to a consideration of whether the persons in question are in relevantly similar situations, and whether the difference in treatment is justified. As it stated in Clift v United Kingdom, para 60, “the general purpose of article 14 is to ensure that where a state provides for rights falling within the ambit of the Convention which go beyond the minimum guarantees set out therein, those supplementary rights are applied fairly and consistently to all those within its jurisdiction unless a difference of treatment is objectively justified”. Consistently with that purpose, it added at para 61 that “while … there may be circumstances in which it is not appropriate to categorise an impugned difference of treatment as one made between groups of people, any exception to the protection offered by article 14 of the Convention should be narrowly construed.” Accordingly, cases where the court has found the “status” requirement not to be satisfied are few and far between.

A relevant difference in treatment has therefore been established in relation to the treatment of children living in households containing more than two children, as compared with children living in households containing one or two children. The remaining question is whether the difference in treatment is justifiable.

In relation to that question, the submissions of counsel for the appellants and of the Equality and Human Rights Commission, appearing as an intervener, raise three preliminary issues of general importance, which it will be necessary to discuss at some length. The first concerns the question whether it is appropriate for our domestic courts to determine whether the United Kingdom has violated its obligations under unincorporated international law when considering whether a difference in treatment is justified under the Human Rights Act. The second is whether the approach to proportionality under article 14 set out by this court in Humphreys v Revenue and Customs Comrs [2012] UKSC 18; [2012] 1 WLR 1545 (“Humphreys”), and followed in several later cases, to the effect that the court will respect the policy choice of the executive or the legislature in relation to general measures of economic or social strategy unless it is “manifestly without reasonable foundation”, accurately reflects the approach of the European Court of Human Rights and should continue to be followed. The third concerns the use which can be made of Parliamentary debates and other Parliamentary material when considering whether primary legislation is compatible with Convention rights, having regard to Parliamentary privilege.

According to the statement of facts and issues agreed between the parties, the issue which has to be determined, in relation to the question of justification, is “whether the UK’s obligations under the UNCRC have been breached in the present case, and if so whether in the circumstances the two child limit is compatible with Convention rights”. The primary question for the court to decide is therefore supposed to be whether, by introducing the limitation on entitlement to child tax credit, the United Kingdom has breached its obligations under the UNCRC. The primacy of this question is argued, on behalf of the appellants, to follow from the decision of this court in DA [2019] UKSC 21; [2019] 1 WLR 3289, which is said to be “authority for the proposition that a court must, where applicable, assess whether the UNCRC has been breached”. As counsel noted, the Court of Appeal’s approach to the present case is irreconcilable with that proposition.

This approach is mistaken. The matter can best be explained if it is approached in stages, considering, first, the rights and obligations created by unincorporated treaties such as the UNCRC on the international plane; secondly, the question whether unincorporated treaties create rights and obligations in domestic law; and thirdly, the question whether the Human Rights Act has given domestic legal effect to unincorporated treaties.

In relation to the first point, although treaties are agreements intended to be binding upon the parties to them, they are not contracts which domestic courts can enforce. As Lord Oliver of Aylmerton explained in JH Rayner (Mincing Lane) Ltd v Department of Trade and Industry [1990] 2 AC 418, p 499 (“International Tin Council”):

In relation to the second point, it is a fundamental principle of our constitutional law that an unincorporated treaty does not form part of the law of the United Kingdom. Lord Oliver explained this in the International Tin Council case at p 500:

That dictum was cited with approval, and the principle which it lays down reasserted by 11 justices of this court, in R (Miller) v Secretary of State for Exiting the European Union (Birnie intervening) [2017] UKSC 5; [2018] AC 61, paras 56, 167 and 244. As was there explained, the dualist system, based on the proposition that international law and domestic law operate in independent spheres, is a necessary corollary of Parliamentary sovereignty. It is only because “treaties are not part of UK law and give rise to no legal rights or obligations in domestic law” (para 55) that the prerogative power to make and unmake treaties is consistent with the rule that ministers cannot alter the law of the land.

The remaining question is whether the Human Rights Act has given domestic legal effect to unincorporated treaties. Clearly, it has not. As Baroness Hale of Richmond succinctly stated in R (Animal Defenders International) v Secretary of State for Culture, Media and Sport [2008] UKHL 15; [2008] 1 AC 1312, para 53:

A misunderstanding appears to have arisen in this jurisdiction from the fact that the European court frequently has regard to international law, and to its interpretation by competent institutions, when interpreting the Convention, for reasons which were correctly identified in this case by the Court of Appeal. In the first place, article 31(3)(c) of the Vienna Convention on the Law of Treaties (1969) requires that, in the interpretation of treaties, “[t]here shall be taken into account, together with the context … any relevant rules of international law applicable in the relations between the parties”. So, for example, the court has regard to international law in seeking to avoid a conflict between the Convention and other treaties, so far as it can. It also sometimes takes account of international law where it is incorporated into the law of a contracting state, and is for that reason relevant to the issue before the court. International law also has a broader significance, along with the contents of the domestic law of the contracting states, Council of Europe texts and other relevant materials, as evidence of a European consensus, or at least of relevant developments or evolving principles, which can inform the interpretation of the Convention, the width of the national margin of appreciation, and the court’s assessment of proportionality. This eclectic approach is facilitated institutionally by the presence on the court of judges from each of the contracting states, and by the work of the court’s registry in researching the relevant materials and establishing judicial information-sharing networks.

This approach to the interpretation of the Convention is illustrated by the two European authorities on which counsel for the appellants relied. The first was the judgment of the Grand Chamber in Demir v Turkey (2008) 48 EHRR 54, where the question arose whether civil servants were “members … of the administration of the state” within the meaning of article 11(2) of the Convention, which permits a state to impose restrictions on the exercise by such persons of rights protected by that article, such as the right to form and join trade unions. The court took account of international instruments which recognised the right of civil servants to form and join trade unions in concluding that civil servants were not “members … of the administration of the state” for this purpose. The court explained its approach at para 85:

The European court’s approach is also illustrated by the dictum, on which counsel relied, in Neulinger v Switzerland (2010) 54 EHRR 31, para 131, that “[t]he Convention cannot be interpreted in a vacuum but must be interpreted in harmony with the general principles of international law … as indicated in article 31(3)(c) of the Vienna Convention on the Law of Treaties”. In that case, the Grand Chamber took account of the 1980 Hague Convention on the Civil Aspects of International Child Abduction, and the UNCRC, when determining the obligations imposed on states by article 8 of the ECHR in the context of international child abduction. It also observed at para 134 that the child’s best interests must be the primary consideration in that area, as was apparent from the Hague Convention.

On the other hand, the European court has not treated provisions of international treaties as if they were directly incorporated into the Convention itself, so as to impose specific obligations on the contracting states via the Convention. Nor does it refer to international materials for the purpose of determining whether contracting states have complied with their obligations under unincorporated international treaties, recognising that it possesses no jurisdiction to make such a determination. Even where rules of international law are incorporated into domestic law, the court has said that its role “is confined to ascertaining whether those rules are applicable and whether their interpretation is compatible with the Convention”: Neulinger, para 100.

There is, accordingly, no basis in the case law of the European court, as taken into account under the Human Rights Act, for any departure from the rule that our domestic courts cannot determine whether this country has violated its obligations under unincorporated international treaties.

In that regard, however, a misunderstanding has appeared in some recent judgments of this court. The starting point appears to have been the court’s consideration of the judgment of the Grand Chamber in X v Austria (2013) 57 EHRR 14, which concerned a complaint under article 14 together with article 8 relating to a law which allowed adoption by unmarried different-sex couples but not by unmarried same-sex couples. In considering whether the differential treatment was justified, the European court listed at para 146 a number of factors weighing in favour of a case-by-case approach to adoption by same-sex couples rather than an absolute prohibition, and added that “[t]his would also appear to be more in keeping with the best interests of the child, which is a key notion in the relevant international instruments”. In that regard, the court referred to an earlier part of the judgment, setting out relevant provisions of international law, in which it had cited articles 3 and 21 of the UNCRC. Article 3.1 provides that in all actions concerning children, “the best interests of the child shall be a primary consideration”. Article 21 provides, in relation to adoption, that “the best interests of the child shall be the paramount consideration”. The Grand Chamber did not inquire into whether Austria had complied with its obligations under the UNCRC: its focus was on the question whether the law in question was justified under article 14, not on whether it had been adopted in compliance with Austria’s obligations under the UNCRC. Nor did the Grand Chamber adopt the provisions of the UNCRC as part and parcel of the ECHR: it did not say, for example, that the best interests of the child were the paramount consideration. The case is an example of the court’s treating international instruments as relevant to its application of the Convention, rather than of its directly applying or “passporting” the UNCRC.

The principle of law which can be derived from that judgment and others (eg EB v France (2008) 47 EHRR 21), and applied by domestic courts, is that in assessing whether differential treatment is justifiable under article 14 of the Convention together with article 8, in a matter concerning a child, the best interests of the child are a relevant consideration. The judgment does not suggest that domestic courts should approach the question of justification by applying the provisions of the UNCRC, or by deciding whether, in adopting the measure in question, the national authorities complied with their obligations under the UNCRC. That approach has, however, been adopted, obiter or in dissenting judgments, in a number of domestic cases.

The first such case is SG [2015] UKSC 16; [2015] 1 WLR 1449, which concerned the compatibility of the benefit cap in its original form with article 14 read together with A1P1. Although the measure was agreed to affect a higher number of women than men and therefore to be indirectly discriminatory, it was held by the majority of the court to be justified because, in the words of the headnote:

It was, however, accepted by Lord Carnwath at paras 117-119, under reference to X v Austria and Ponomaryov v Bulgaria (2011) 59 EHRR 20 (“Ponomaryov”) (where international law was taken into account in assessing the width of the national margin of appreciation, as explained in para 80 above), that the question whether the Government had complied with the UNCRC could be relevant to the consideration of article 14 by domestic courts. He concluded at para 128 that the Secretary of State had failed to establish that the delegated legislation in question complied with the United Kingdom’s obligation under article 3.1 to treat the best interests of children as a primary consideration. Those observations were obiter.

The issue arose again in Mathieson v Secretary of State for Work and Pensions [2015] UKSC 47; [2015] 1 WLR 3250, which concerned delegated legislation under which an allowance paid to the parents of a disabled child ceased to be payable if he was a long-term patient in hospital. The question was whether the legislation resulted in unjustifiable discrimination contrary to article 14 taken together with A1P1. Lord Wilson, in a judgment with which a majority of the court agreed, began his discussion of the question of justification by contrasting the basis on which the legislation had been made - namely, that the disability needs of children in hospital were met by the NHS rather than by their parents - with the evidence before the court. That evidence demonstrated that the personal and financial demands made on parents who helped to care for their disabled children in hospital were, at least, no less than when they cared for them at home.

Lord Wilson next considered whether the Secretary of State, in making the delegated legislation in question, had complied with the United Kingdom’s obligations under article 3.1 of the UNCRC and article 7.2 of the United Nations Convention on the Rights of Persons with Disabilities, both of which required the best interests of children to be treated as a primary consideration. Lord Wilson concluded that the Secretary of State had violated these provisions of international law. In considering the relevance of that finding, Lord Wilson cited X v Austria, the judgment of Lord Carnwath and the dissenting judgments of Lady Hale and Lord Kerr in SG, and a passage in the judgment of Lord Hughes in SG in which he had referred to the European court’s practice of citing international materials as an aid to its interpretation of the Convention. Lord Wilson concluded that “[a] conclusion, reached without reference to international Conventions, that the Secretary of State has failed to establish justification for the difference in his treatment of those severely disabled children who are required to remain in hospital for a lengthy period would harmonise with a conclusion that his different treatment of them violates their rights under two international Conventions” (para 44). Lord Wilson did not, however, address the precise relevance of the international conventions, or what should happen if a conclusion reached without reference to them did not “harmonise” with a conclusion that they had been violated.

Since Lord Wilson’s conclusion that the differential treatment was unjustified was reached without reference to international law, but “harmonised” with his finding that the Secretary of State had violated two unincorporated international treaties, his remarks about international law should not be regarded as forming part of the ratio of the decision. With great respect, they must also be regarded as having been made per incuriam. As I have explained, for a United Kingdom court to determine whether this country is in breach of its obligations under an unincorporated international treaty, and to treat that determination as affecting the existence of rights and obligations under our domestic law, contradicts a fundamental principle of our constitutional law. The court was not referred to relevant authorities, such as the International Tin Council case.

One might add that what Lord Wilson took from the unincorporated international treaties was that the Secretary of State had been under a duty to treat the best interests of children as a primary consideration before making the legislation. There could have been no objection if he had instead treated the best interests of children as a relevant factor in the court’s assessment of whether the differential treatment resulting from the legislation was justified under article 14 of the Convention: an approach which could have been taken directly from article 14 taken together with article 8 as interpreted in X v Austria and other cases.

Reliance was also placed by counsel for the appellants on In re McLaughlin [2018] UKSC 48; [2018] 1 WLR 4250. The case concerned a provision under which the eligibility of a surviving parent to receive widowed parent’s allowance depended on whether she had been married to the parent who died. The provision was held to be incompatible with article 14 read with article 8. Lady Hale, giving a judgment with which the majority of the court agreed, reached that conclusion at para 39 for reasons which were independent of any consideration of international law, but added at para 40 that her view was “reinforced” by the international obligations to which the United Kingdom was party, on the basis that the provision in question was incompatible with the UNCRC and with the International Covenant on Economic Social and Cultural Rights. That part of the judgment was therefore obiter.

The most significant case which needs to be considered in relation to this point is DA [2019] UKSC 21; [2019] 1 WLR 3289. The issue in that case was whether the benefit cap, in its revised form (see para 10 above), unlawfully discriminated against single parents with young children. The court held, in the words of the headnote:

In the course of his judgment, Lord Wilson, with whom Lord Hodge and Lord Hughes agreed, identified the critical question, in relation to justification under article 14 read together with article 8, as being whether the reasons for the differential treatment were “manifestly without reasonable foundation” (para 65). He then considered whether the Government had complied with the requirement in article 3.1 of the UNCRC that the best interests of children be a primary consideration, on the basis that “a foundation for the decision [not to exempt the relevant category of children from the cap] not made in substantial compliance with article 3.1 might well be manifestly unreasonable” (para 78). In approaching the matter in that way, he again relied on the obiter dicta of Lord Carnwath, and the dissenting judgments of Lady Hale and Lord Kerr, in SG. In the event, he concluded at para 87 that there had been no breach of the UNCRC, finding that the Government had considered the best interests of children before introducing the legislation in question. In the next section of his judgment, Lord Wilson stated (para 88) that he was “also” driven to conclude that the Government’s decision was not “manifestly without reasonable foundation”: the Government’s belief as to the beneficial impact of the measure on children generally was not challenged, and discretionary payments were available in cases of particular hardship. Lord Carnwath agreed that “the ‘best interests’ principle under article 3.1 [of the UNCRC] is potentially relevant” (para 102), in a judgment with which Lord Hughes and I expressed agreement.

Given the cautious language used by Lord Wilson when discussing the significance of non-compliance with the UNCRC (“might well be manifestly unreasonable”), and the explicit basing of the decision on the Government’s belief about the beneficial effects of its policy, and on the availability of discretionary payments, it does not appear to me that Lord Wilson’s remarks about the UNCRC formed part of the essential grounds of the decision. In the words of Sir Frederick Pollock, cited by Lord Denning in Close v Steel Co of Wales Ltd [1962] AC 367, 388-389:

The second preliminary matter to be considered is whether the approach to proportionality under article 14 set out by this court in Humphreys [2012] UKSC 18; [2012] 1 WLR 1545, and followed in several later cases, to the effect that the court will respect the policy choice of the executive or the legislature in relation to general measures of economic or social strategy in the context of welfare benefits unless it is “manifestly without reasonable foundation”, accurately reflects the approach of the European court and should continue to be followed. The question has been raised particularly in the light of the recent decision of the First Section of the European court in JD and A v United Kingdom [2020] HLR 5 (“JD”). This is a difficult and important question which requires careful consideration.

According to the settled case law of the European court, the question whether there is an “objective and reasonable” justification for a difference in treatment is to be judged by whether it pursues a “legitimate aim” and there is a “reasonable relationship of proportionality” between the aim and the means employed to achieve it: see Carson (2010) 51 EHRR 13, para 61, cited at para 37 above. It is also well settled that states have “a margin of appreciation in assessing whether and to what extent differences in otherwise similar situations justify a different treatment”: ibid. Crucially, in relation to the present issue, “[t]he scope of this margin will vary according to the circumstances, the subject matter and the background.”

The court has not itself provided, in its judgments, a systematic analysis of relevant factors or an explanation of how they interact. Its accounts of the general principles it applies are stated at a high level of generality. Nevertheless, patterns emerge, and inferences can be drawn, from a survey of its case law, as I shall explain. It is doubtful whether the nuanced nature of the approach which it follows can be comprehensively described by any general rule. It is more useful to think of there being a range of factors which tend to heighten, or lower, the intensity of review. In any given case, a number of these factors may be present, possibly pulling in different directions, and the court has to take them all into account in order to make an overall assessment. The case law indicates, however, that some factors have greater weight than others.

One particularly important factor is the ground of the difference in treatment. In principle, and all other things being equal, the court usually applies a strict review to the reasons advanced in justification of a difference in treatment based on what it has sometimes called “suspect” grounds of discrimination. However, these grounds form a somewhat inexact category, which has developed in the case law over time, and is capable of further development by the European court. Furthermore, a much less intense review may be applied even in relation to some so-called suspect grounds where other factors are present which render a strict approach inappropriate, as some of the cases to be discussed will demonstrate.

The court originally developed a requirement that “very weighty reasons” must be shown to justify a difference in treatment based on sex or gender, on the basis that the advancement of gender equality is a major goal in the contracting states: see Abdulaziz v United Kingdom (1985) 7 EHRR 471, para 78. This approach was soon applied in cases concerning entitlement to welfare benefits: see, for example, Van Raalte v Netherlands (1997) 24 EHRR 503, para 39.

On the other hand, while continuing to state that “very weighty reasons” are required for a difference in treatment on grounds of sex to be regarded as compatible with the Convention, the court has found there to be no violation where a state has been acting to correct a historical inequality between the sexes and the pace of change has been reasonable in the context of the contracting states as a whole. For example, in Petrovic (1998) 33 EHRR 14, which concerned the lack of provision of parental leave allowances for fathers, the court noted that there was no common standard across the contracting states at the material time, and that the idea of providing financial assistance to fathers was a relatively recent development, as society moved towards a more equal sharing of responsibilities for the bringing up of children. It concluded that Austria’s introduction of legislation on this issue in a gradual manner, reflecting the evolution of society in that sphere, fell within its margin of appreciation. Fourteen years later, in the Grand Chamber judgment in Markin v Russia (2012) 56 EHRR 8 (“Markin”), the court reiterated the “very weighty reasons” approach (para 127), and found, having regard to the evolution of society in the contracting states, that the different treatment of fathers in Russia with respect to parental leave unjustifiably perpetuated gender stereotypes.

These cases illustrate three points of wider significance. The first is that the court’s statements that “very weighty reasons” are required to justify a difference in treatment on a particular ground do not necessarily exclude the possibility that a relatively wide margin of appreciation, and a correspondingly less intense standard of review, may nevertheless be appropriate in particular circumstances, as for example where historical inequalities are being addressed in pace with changes in social attitudes. The second is that the court’s case law evolves in the light of the development of common standards among the contracting states. The third is that the court has moved over time towards explaining the need for weighty reasons to justify certain grounds of differences in treatment in terms of the link between those grounds and problems of stereotyping, stigma and social exclusion, which prevent participation in society on an equal footing to others. It is to be noted that this approach differs from that suggested by Lord Walker of Gestingthorpe in R (RJM) v Secretary of State for Work and Pensions (Equality and Human Rights Commission intervening) [2008] UKHL 63; [2009] 1 AC 311, para 5, which focused instead on the relationship between the characteristic in question and the development of an individual’s personality.

The court next required “very weighty reasons” in cases concerned with differences in treatment based on birth status, referring to a consensus in the legislation of the contracting states: Inze v Austria (1987) 10 EHRR 394, para 41. This approach has been applied both to discriminatory treatment based on a person’s birth outside marriage, and also to such treatment based on a person’s being adopted: Pla v Andorra (2004) 42 EHRR 25, para 61.

The “very weighty reasons” requirement was next extended to differences in treatment based exclusively on nationality: Gaygusuz v Austria (1996) 23 EHRR 364, para 42. That case, like many later cases concerned with differential treatment on the ground of nationality, was concerned with entitlement to welfare benefits, but predated the formulation of the “manifestly without reasonable foundation” approach in the latter context. It will be necessary to consider later some more recent cases of the same kind.

There are, nevertheless, circumstances in which a wider margin is applicable to a difference in treatment based exclusively on the ground of nationality. An example is provided by the case of MS v Germany (Application No 44770/98) (unreported) given 20 January 2000, which concerned Germany’s willingness to extradite non-nationals living in Germany, but not its own citizens. Given that extradition is a matter of international law, and that the responsibility of a state for its own nationals and for others living on its territory may vary in that field, the decision of the German authorities was considered to fall within their margin of appreciation. The court has also accepted that there are circumstances in which a state may justifiably differentiate between different categories of aliens residing in its territory, for example so as to comply with EU law: Ponomaryov (2011) 59 EHRR 20, para 54.

The court has also stated that differences in treatment based on sexual orientation require “particularly serious reasons” by way of justification: see, for example, EB v France (2008) 47 EHRR 21, para 58. Nevertheless, a relatively wide margin was allowed in Schalk v Austria (2010) 53 EHRR 20 (“Schalk”), which concerned the unavailability of legal recognition for same-sex relationships. The court stated at para 97 that, on the one hand, it had repeatedly held that differences based on sexual orientation require “particularly serious reasons” by way of justification, but on the other hand, “a wide margin is usually allowed to the state under the Convention when it comes to general measures of economic or social strategy”, citing its judgment in Stec v United Kingdom (2006) 43 EHRR 47, para 52 (“Stec”). It went on at para 98 to reiterate that the scope of the margin of appreciation will vary according to the circumstances, the subject matter and its background, and said that, in that respect, one of the relevant factors might be the existence or non-existence of common ground between the laws of the contracting states. In that regard, it noted at para 105 that there was an emerging European consensus towards legal recognition of same sex couples, which had developed rapidly over the previous decade but was still confined to a minority of contracting states. The area was therefore one of evolving rights with no established consensus, where states must enjoy a margin of appreciation in the timing of the introduction of legislative changes. By the time of the judgment, Austria had legislated to recognise such relationships, and the legislator could not be reproached for not having introduced the legislation any earlier. Like Petrovic, this judgment illustrates how, even when a case concerns a ground of differential treatment which has been held to require “particularly serious reasons” (or “very weighty reasons”) by way of justification, there may be other factors present which require greater latitude to be allowed to the national authorities, as for example where the issue concerns the timing of legislative changes designed to address existing inequalities in response to changes in social attitudes.

The court has also adopted a strict approach to differential treatment on the ground of race or ethnic origin, stating that no difference in treatment based exclusively or to a decisive extent on a person’s ethnic origins is capable of being justified in a contemporary democratic society: Biao v Denmark (2016) 64 EHRR 1, para 94 (where the discrimination was indirect). The Grand Chamber’s observation (ibid) that discrimination on the basis of ethnic origin is a form of racial discrimination indicates that discrimination on the ground of race would be treated in the same way. The case concerned immigration law, and the Grand Chamber had stated at para 93 that a “wide margin is usually allowed to the state when it comes to general measures of economic or social strategy”, citing its judgment in Carson, para 61. The case is therefore another reminder that different principles, pulling in different directions, may be relevant to the same set of facts.

The court has generally adopted a strict approach also to differences in treatment on the ground of religious belief, in the light of the importance of the right enshrined in article 9 in guaranteeing the individual’s self-fulfilment. It has repeatedly said that “a distinction based essentially on a difference in religion alone is not acceptable”: see, for example, Vojnity v Hungary (Application No 29617/07) [2013] 2 FCR 495, para 31. The principle does not, however, appear to be as absolute as that language might suggest. In Vojnity, the court went on to say at para 36 that such treatment “will only be compatible with the Convention if very weighty reasons exist”. The court described its approach as being similar to that applied in the context of differences in treatment on the basis of sex, birth status, sexual orientation and nationality.

The court has, however, taken a less strict approach in some cases concerned with discrimination on the ground of religious belief, where other factors were relevant. The point is illustrated by the case of Eweida v United Kingdom (2013) 57 EHRR 8, which concerned a number of complaints of religious discrimination. Three were considered under article 14, taken together with article 9. The first concerned rules at a hospital which restricted the wearing of necklaces by staff handling patients, in the interests of clinical safety, and in consequence prevented the applicant, a nurse, from wearing a cross around her neck. The court accepted that this was an interference with her freedom to manifest her religion, but also noted the importance of the reason for it. It observed that this was a field where the domestic authorities must be allowed a wide margin of appreciation, and concluded that the measure was a proportionate interference with her rights under article 9. It added that the factors to be weighed in the balance when assessing the proportionality of the measure under article 14 would be similar, and that there was no basis on which it could find a violation of that article.

The second complaint concerned the introduction of a requirement that the applicant, a registrar of births, deaths and marriages, must also register civil partnerships between same-sex couples. As she had a religious objection to same-sex unions, she refused to register them, and consequently lost her job. The court accepted that the requirement that registrars must register civil partnerships had a particularly detrimental impact on her because of her religious beliefs. But the requirement had a legitimate aim, namely to provide a service without discrimination on the ground of sexual orientation. The court observed that it “generally allows the national authorities a wide margin of appreciation when it comes to striking a balance between competing Convention rights”, and held that there had been no violation of article 14 (para 106). The court applied similar reasoning in relation to the third complaint, which concerned a sex therapist who had lost his job because he refused to provide his services to same-sex couples.

A relatively strict approach has also been adopted in cases concerned with persons with disabilities, in order to “foster their full participation and integration in society”: Glor v Switzerland (Application No 13444/04) (unreported) given 30 April 2009, para 84 (“Glor”). In the more recent case of Guberina (2016) 66 EHRR 11, which concerned a refusal to grant a tax exemption for persons with special accommodation needs to the father of a disabled child, the court noted that, on the one hand, a wide margin is usually allowed to the state under the Convention when it comes to general measures of economic or social strategy, including measures in the area of taxation (para 73). On the other hand, it continued:

The court has also applied the same approach to a difference in treatment based on a person’s being HIV positive: Kiyutin v Russia (2011) 53 EHRR 26, paras 63-64. The decision illustrates how the court’s jurisprudence in this area can evolve over time, generally in response to a consensus emerging from international and European law, and the law of the contracting states. It is also a further illustration of the link between the need for “very weighty reasons” and the stigmatising of particular groups in society. It was indicated in Kiyutin that the same approach would also apply to a difference in treatment based on a person’s mental faculties.

The counterpart of the strict approach taken, other things being equal, to differences in treatment on the grounds discussed above, is less strict scrutiny, other things being equal, of differences in treatment on other grounds, such as age (see, for example, Schwizgebel v Switzerland, Reports of Judgments and Decisions, 2010-V, 29, para 92, and British Gurkha Welfare Society v United Kingdom (2016) 64 EHRR 11, para 88), immigration status (see, for example, Bah v United Kingdom (2011) 54 EHRR 21, para 47), prisoner status (see, for example, Stummer v Austria (2011) 54 EHRR 11, para 101), and marital status (see, for example, Yiğit v Turkey (2010) 53 EHRR 25, para 72). However, as indicated in para 99 above, the case law does not support a straightforwardly binary approach, as a range of factors may be relevant in particular circumstances. For example, although age has not been treated as a “suspect” ground, the best interests of children have been treated by the European court as an important factor in assessing proportionality under article 14 (see, for example, para 86 above), reflecting the fact that individuals in that age group have particular needs and vulnerabilities.

In summary, therefore, the court’s approach to justification generally is a matter of some complexity, as a number of factors affecting the width of the margin of appreciation can arise from “the circumstances, the subject matter and its background”. Notwithstanding that complexity, some general points can be identified.

As the cases demonstrate, more than one of those points may be relevant in the circumstances of a particular case, and, unless one factor is of overriding significance, it is then necessary for the court to make a balanced overall assessment.

The court’s approach in cases under article 14 where it has used the phrase “manifestly without reasonable foundation” is consistent with the foregoing points. The phrase seems first to have appeared, in the context of a complaint under article 14, in the Grand Chamber’s judgment on the merits in Stec (2006) 43 EHRR 74. The case concerned a difference in treatment on the ground of sex in relation to entitlement to a state benefit. As the benefit was linked to the statutory pension scheme, women initially qualified for it at 60, whereas men qualified at 65, those being the ages at which they qualified for a state pension. The scheme was designed to achieve equality in the qualifying age by 2020.

After explaining that the contracting states enjoy a margin of appreciation in assessing whether and to what extent differences in otherwise similar situations justify a difference in treatment, the court continued at para 52:

On the facts of Stec, the third and fourth of the points identified in para 115 were also relevant. In relation to the third point, it was considered significant that many of the other contracting states continued to maintain a difference in the ages at which men and women became eligible for the state retirement pension. In relation to the fourth point, Parliament could not be condemned for deciding to introduce the reform slowly and in stages, given the extremely far-reaching and serious implications for women and for the economy in general (para 65).

Like Petrovic (para 102 above) and Schalk (para 107 above), Stec illustrates how, where a ground of differential treatment is one which could generally be justified only by “very weighty reasons”, other factors may nevertheless make it appropriate to allow the state a wide margin of appreciation. It also demonstrates that the fact that a case falls within the general scope of the “manifestly without reasonable foundation” approach is not the whole story, where other relevant factors are also present.

Later cases where reference has been made to the “manifestly without reasonable foundation” approach support that view. Considering first cases in which suspect grounds were involved, the cases of Zeman v Austria (Application No 23960/02) (unreported) given 29 June 2006 and Runkee v United Kingdom [2007] 2 FCR 178 both concerned differential treatment on the ground of sex in relation to entitlement to widows’ benefits. In each case, the court repeated what it had said in para 52 of Stec (para 118 above). In Zeman, where further differentiation between men and women had recently been introduced, the court observed that “very strong reasons” had to be put forward to justify it (para 40), and concluded that no convincing reason had been shown. In Runkee, on the other hand, the court noted that the measure in question formed part of a process of reform designed to achieve equality between the sexes over time, and concluded that it was not unreasonable of the legislature to decide to introduce the reform slowly. The same approach was followed, on similar facts, in Andrle v Czech Republic (2011) 60 EHRR 14.

The same approach has been followed in cases concerned with differences in treatment under social security schemes on the ground of nationality: another of the so-called suspect grounds, as explained in para 105 above. In the case of Luczak v Poland (Application No 77782/01) (unreported) given 27 November 2007, the court repeated what had been said in para 52 of Stec, mutatis mutandis. The court concluded that “the Government have not adduced any reasonable and objective justification for the distinction [between nationals and non-nationals] such as to meet the requirements of article 14, even having regard to their margin of appreciation in the area of social security” (para 59). The same conclusion was reached by the Grand Chamber in the case of Andrejeva v Latvia (2009) 51 EHRR 28. Having referred to the wide margin of appreciation and to its approach of generally respecting the legislature’s policy choice unless it is “manifestly without reasonable foundation” (para 83), the court went on to state that “very weighty reasons” would have to be put forward before it could regard a difference in treatment based exclusively on nationality as compatible with the Convention (para 87). No such reasons were found to exist. Accordingly, “while being mindful of the broad margin of appreciation enjoyed by the state in the field of social security” (para 89), the court concluded that there had been a violation of article 14.

Those cases might be contrasted with British Gurkha Welfare Society v United Kingdom (2016) 64 EHRR 11, which also concerned differential treatment on the ground of nationality, in the context of a pension scheme. The differential treatment arose in the context of transitional arrangements designed to correct an historical inequality over time. The court repeated its earlier statements that “very weighty reasons” would have to be put forward to justify a difference in treatment based exclusively on the ground of nationality. It also repeated its customary statements about the wide margin usually allowed when it comes to general measures of economic or social strategy, and about its willingness generally to respect the legislature’s policy choice unless it is “manifestly without reasonable foundation”. When it came to consider the facts, the court stated that “in considering whether such ‘very weighty reasons’ exist, the court must be mindful of the wide margin usually allowed to the state under the Convention when it comes to general measures of economic or social strategy” (para 81). It added that that was “particularly so where an alleged difference in treatment resulted from a transitional measure forming part of a scheme … which was carried out in good faith in order to correct an inequality”.

The decisions in Luczak and Andrejeva might also be contrasted with Tomás v Spain (2016) 65 EHRR 24, which concerned differential treatment in the field of pensions based on ability to marry, and consequently on sexual orientation, where the court followed the approach which it had taken in Schalk (para 107 above) and allowed a wide margin.

The cases of Stec, Zeman, Runkee, Andrle, Luczak, Andrejeva, British Gurkha and Tomás, all concerned with “suspect” grounds of differential treatment, might be contrasted with others concerned with non-suspect grounds. The case of Carson (2010) 51 EHRR 13 concerned the differential treatment of recipients of the state pension depending on whether they were resident in the United Kingdom or overseas. Unlike sex and nationality, residence is not one of the so-called suspect grounds. The Grand Chamber’s approach reflected that difference. It repeated at para 61 what had been said in Stec (para 118 above) about a wide margin, and respecting the legislature’s policy choice unless it was “manifestly without reasonable foundation”. It made no reference to “very weighty reasons”. It also underlined that, in the context of welfare benefits and pensions, it will look at the compatibility of the system overall, without giving undue weight to the circumstances of the individual, since welfare systems, to be workable, have to deal in broad categorisations which will inevitably affect some people more prejudicially than others.

A similar approach was adopted by the Grand Chamber in Stummer (2011) 54 EHRR 11. The case concerned the different treatment of prisoners engaged in prison work, as compared with other workers, in relation to the Austrian old-age pension system. Since the ground of differential treatment was not “suspect”, the court made no reference to a need for “very weighty reasons”. It noted that the issue of working prisoners’ affiliation to the old-age pension system raised choices of social strategy, where the court would only intervene if it considered the legislature’s policy choice to be “manifestly without reasonable foundation”. It also attached weight to the fact that there was no common ground between the contracting states at the material time. Although there was an evolving trend, and “the respondent state is required to keep the issue … under review” (para 110), there had been no violation. Other cases where a similar approach was adopted include Tarkoev v Estonia (Application No 14480/08) (unreported) given 4 November 2010, Bah v United Kingdom (2011) 54 EHRR 21 and B v United Kingdom (Application No 36571/06) (unreported) given 14 February 2012.

Two more recent cases should also be noted. First, the judgment in Vrountou v Cyprus (2015) 65 EHRR 31 was expressed in terms which might be thought to suggest a stricter approach, in the context of welfare benefits, to differences of treatment on the ground of sex. The case concerned the provision of housing assistance to the children of men, but not women, displaced by the Turkish invasion of Cyprus. The respondent government argued that the case fell within the scope of the “manifestly without reasonable foundation” approach. However, although the court accepted that the case concerned welfare benefits (paras 63 and 66), it did not refer to the wide margin of appreciation usually allowed in relation to general measures of economic or social policy, or repeat the “manifestly without reasonable foundation” formulation. Instead, it said that “[i]n cases where the difference in treatment is on grounds of sex, the general principles which apply in determining this question of justification, were restated by the Grand Chamber in Markin” (cited at para 102 above: it was not a welfare benefit case). Those principles included the statement that the scope of the margin of appreciation will vary according to the circumstances, the subject matter and its background, and that “very weighty reasons” would have to be put forward to justify a difference in treatment on the ground of sex (para 75).

Secondly, the Grand Chamber judgment in Fábián v Hungary (2017) 66 EHRR 26 modified the standard statement taken from para 52 of Stec (para 118 above), repeated time and again in the case law. The case concerned a difference in treatment between civil servants and private sector employees under the state pension scheme. That was not, of course, a “suspect” ground, and the court made no reference to the need for “very weighty reasons”. However, after repeating what had been said in para 52 of Stec from “a wide margin” to “manifestly without reasonable foundation”, the Grand Chamber added (para 115, omitting most of the citations):

Up to this point, the cases from Stec onwards concerned with welfare benefits and pensions can be seen to have generally followed a consistent approach. Although that approach is far from mechanical, the points noted in para 115 above can readily be discerned.

The remaining point which can be drawn from the case law up to this point is how the court has balanced different factors when they pull in different directions: notably, where the need for “very weighty reasons” co-exists with the generally wide margin of appreciation described by the “manifestly without reasonable foundation” formulation. In Luczak, the court concluded that the government had not adduced a reasonable and objective justification, “even having regard to their margin of appreciation in the area of social security” (para 59). In Andrejeva, the court concluded, “while being mindful of the broad margin of appreciation enjoyed by the state in the field of social security”, that there had been a violation (para 89). In British Gurkha, the court stated that “in considering whether such ‘very weighty reasons’ exist, the court must be mindful of the wide margin usually allowed to the state under the Convention when it comes to general measures of economic or social strategy” (para 81). It appears, therefore, that in cases involving “suspect” grounds in the field of welfare benefits and pensions, the determinative factor has generally been whether “very weighty reasons” have been shown, but that the court has taken account of the wide margin generally applicable in that field when making that assessment. Whether a measure has formed part of a scheme intended to address historical inequalities, and the presence or absence of common standards among the contracting states, have also been important factors.

It is necessary to consider the judgment in JD [2020] HLR 5 against that background. It is worth mentioning at the outset that the decision has been selected by the Bureau of the court (comprising the President, Vice-Presidents and Section Presidents) as a “key case” following a proposal by the Jurisconsult, whose role is to monitor the case law and promote its consistency.

The case concerned two complaints of discrimination, one on the ground of disability, and the other on the ground of gender. Those are both “suspect” grounds, as explained earlier. The first complaint had been rejected unanimously, and the second by a majority of this court, in R (MA) v Secretary of State for Work and Pensions (Equality and Human Rights Commission intervening) [2016] UKSC 58; [2016] 1 WLR 4550 (“MA”). The European court dismissed the first complaint, but upheld the second, by a majority. The judgment is an example of the application of the Thlimmenos approach to discrimination (para 48 above), in that the breach of article 14 comprised a failure to treat differently persons whose situations were significantly different.

For present purposes, however, it is another aspect of the judgment which is relevant. The material passage in the judgment of the majority appears at paras 87-89, under the heading “the general principles”, and needs to be set out in full:

Given the reliance placed on this passage by counsel for the appellants in the present proceedings, the court must scrutinise the reasoning with care, notwithstanding the great respect which it has for the European court. Considering, first, para 87 of the judgment, the passages cited from Fábián, para 115, Hämäläinen v Finland, para 109, and Andrejeva, para 83, were not concerned with “the context of article 1 of Protocol 1 alone”. Each of those passages, in judgments of the Grand Chamber, was concerned with article 14. The passage in Fábián appeared in the discussion of article 14 under the heading, “Relevant principles”. The passage in Hämäläinen appeared in a similar discussion of article 14 under the heading “General principles”. The passage in Andrejeva appeared under the heading, “Compliance with article 14 of the Convention”. In reality, the principles set out in para 87 of JD are familiar aspects of the court’s jurisprudence on article 14. They are not even confined to cases concerned with article 14 taken together with A1P1: the same principles were applied, for example, in Bah, which concerned article 14 read together with article 8, and in Muñoz Díaz, which concerned article 14 taken together with articles 8 and 12.

Considering next para 88 of the judgment, the first two sentences reflect what was said in Fábián, para 115. I have to confess to some difficulty in understanding the third sentence. The court’s case law does not support the statement that, in the sphere of economic or social policy, “the court has limited its acceptance to respect the legislature’s policy choice as not ‘manifestly without reasonable foundation’ to circumstances where an alleged difference in treatment resulted from a transitional measure forming part of a scheme carried out in order to correct an inequality”. Certainly, Stec, Runkee, Andrle and British Gurkha involved circumstances of that kind; but the court has also applied the “manifestly without reasonable foundation” formulation, and found no violation, in many cases not concerned with transitional measures, as has been explained. Examples include Tomás, Carson, Tarkoev, Stummer, Bah and B v United Kingdom. Two of those (Carson and Stummer) are Grand Chamber judgments. But the court in JD will have known that.

Considering next para 89 of the judgment, disability and gender are both “suspect” grounds for a difference in treatment, and therefore require to be justified by “very strong reasons”. That appears clearly from Glor, Guberina and Markin (paras 102 and 112 above). But para 89 departs from the court’s previous judgments concerning article 14 in the field of social security (a field with which Glor, Guberina and Markin were not concerned), even those concerned with differences in treatment on a suspect ground (apart from some isolated examples such as Vrountou), in omitting any mention of the wide margin, signified by the “manifestly without reasonable foundation” formulation, which is usually enjoyed by the national authorities in that field. As I have explained, even in the context of cases concerned with “suspect” grounds, including gender, the court has in the past stated that there is in principle a wide margin applicable in the field of welfare benefits, reflected in the “manifestly without reasonable foundation” approach. That is true not only of cases concerned with transitional measures, such as Stec, Runkee and British Gurkha, but also of cases which were not concerned with such measures, such as Zeman, Luczak, Andrejeva, Andrle and Tomás. It has balanced the factors pointing towards a strict standard of review, and those pointing towards a wider margin, as I explained in para 130 above under reference to Luczak, Andrejeva and British Gurkha.

It is argued in the present proceedings that the reasoning in JD establishes a simple rule: complaints of discrimination on “suspect” grounds fall outside the scope of the wide margin and “manifestly without reasonable foundation” approach usually accorded in the field of welfare benefits, unless the case concerns “transitional measures”. Instead, cases concerned with suspect grounds are governed by the principles laid down in cases from outside that field, such as Markin, in relation to discrimination on the ground of gender, and Guberina, in relation to discrimination on the ground of disability. That approach would represent a significant modification of a substantial body of case law, as I have explained.

More recently, the same section of the court decided (unanimously) the case of Jurčič v Croatia [2021] IRLR 511, which also concerned a complaint of discrimination on the ground of sex in relation to welfare benefits. The court reiterated the “manifestly without reasonable foundation” standard (para 64), following the formulation set out in Fábián, para 115 (para 128 above), and added (para 65, omitting some of the citations):

This reasoning is broadly in line with the court’s case law prior to JD, and treats that case as establishing merely that cases concerning transitional measures designed to correct historical inequalities form an exception to a general principle that, notwithstanding that a wide margin, reflected by the “manifestly without reasonable foundation” formulation, is generally appropriate in the field of welfare benefits, a stricter approach, calling for “very weighty reasons”, is appropriate where a difference in treatment is based on sex.

More recently still, the Fourth Section of the court decided the case of Yocheva and Ganeva v Bulgaria (Application Nos 18592/15 and 43863/15) (unreported) given 11 May 2021, which concerned a difference in treatment under social security legislation on the suspect ground of birth outside marriage. At para 101, the court reiterated the “manifestly without reasonable foundation” standard, following the formulation set out in Fábián, para 115, then repeated the second and third sentences of para 88 of the judgment in JD. The decision required very weighty reasons to justify the difference in treatment.

In the meantime, the court has continued to allow a wide margin, and to repeat the “manifestly without reasonable foundation” formula without qualification, in cases concerned with differences in treatment on non-suspect grounds: see, for example, Popovic v Serbia (2020) 71 EHRR 29.

In summary, the European court has generally adopted a nuanced approach, which can be understood as applying certain general principles, but which enables account to be taken of a range of factors which may be relevant in particular circumstances, so that a balanced overall assessment can be reached. As I have explained, there is not a mechanical rule that the judgment of the domestic authorities will be respected unless it is “manifestly without reasonable foundation”. The general principle that the national authorities enjoy a wide margin of appreciation in the field of welfare benefits and pensions forms an important element of the court’s approach, but its application to particular facts can be greatly affected by other principles which may also be relevant, and of course by the facts of the particular case. Indeed, this approach is not confined to cases concerned with article 14, but can be seen in other contexts where the state generally enjoys the wide margin of appreciation signified by the “manifestly without reasonable foundation” formula, but where other factors may indicate a narrower margin of appreciation, and the court accordingly balances the relevant factors: see, for a recent example, LB v Hungary (2021) 72 EHRR 28, paras 48-50. In the context of article 14, the fact that a difference in treatment is based on a “suspect” ground is particularly significant. The recent cases of JD, Jurčič and Yocheva and Ganeva, like many earlier cases, indicate the general need for strict scrutiny, focused on the requirement for very weighty reasons, where the difference in treatment is based on a suspect ground such as sex or birth outside marriage, unless the issue concerns the timing of reform designed to address historical inequalities, where a wider margin is likely to be appropriate.

The concept of the margin of appreciation is specific to the European court. Nevertheless, domestic courts have generally endeavoured to apply an analogous approach to that of the European court. They have done so for two reasons. The first was explained by Lady Hale in R (Countryside Alliance) v Attorney General [2007] UKHL 52; [2008] AC 719, para 126:

The second reason is that domestic courts have to respect the separation of powers between the judiciary and the elected branches of government. They therefore have to accord appropriate respect to the choices made in the field of social and economic policy by the Government and Parliament, while at the same time providing a safeguard against unjustifiable discrimination. As Lord Neuberger of Abbotsbury observed in R (RJM) v Secretary of State for Work and Pensions, para 57, “there will come a point where the justification for a policy is so weak, or the line has been drawn in such an arbitrary position, that, even with the broad margin of appreciation accorded to the state, the court will conclude that the policy is unjustifiable”.

In domestic law, as at the Strasbourg level, one would expect closer scrutiny where the case concerns discrimination on a ground such as sex or race, rather than a difference in treatment on less sensitive grounds, especially if it is simply a by-product of a legitimate policy. Distinctions drawn on “suspect” grounds are inherently appropriate for close judicial scrutiny, notwithstanding the respect due to the judgment of the executive or the legislature.

There is nothing alien or new about an approach which, in general, accords a high level of respect to the judgment of public authorities in the field of economic or social policy, but balances that with the need for close scrutiny where differences of treatment are based on “suspect” grounds. On the one hand, unjustifiable discrimination by public authorities is likely to be irrational and therefore unlawful at common law: as Lord Hoffmann stated in Matadeen v Pointu [1999] 1 AC 98, 109, “treating like cases alike and unlike cases differently is a general axiom of rational behaviour”. Indeed, the classic example given of unreasonable behaviour in the Wednesbury case was of discrimination (Associated Provincial Picture Houses Ltd v Wednesbury Corpn [1948] 1 KB 223, 229). At the same time, the general need for judicial restraint in the area of economic policy was made clear by Lord Bridge of Harwich’s statement in R v Secretary of State for the Environment, Ex p Hammersmith and Fulham London Borough Council [1991] 1 AC 521, 597, that where a “statute has conferred a power on the Secretary of State which involves the formulation and the implementation of national economic policy and which can only take effect with the approval of the House of Commons, it is not open to challenge on the grounds of irrationality short of the extremes of bad faith, improper motive or manifest absurdity”. In other words, the administrative law test of unreasonableness is generally applied in contexts such as economic policy and social policy with considerable care and caution; and the same is true of the Convention test of proportionality. Both tests have to be applied in a way which reconciles the rule of law with the separation of powers.

This balanced approach was adopted in the early case of Ghaidan v Godin-Mendoza [2004] UKHL 30; [2004] 2 AC 557, concerned with discrimination in housing legislation on the basis of sexual orientation. Lord Nicholls of Birkenhead acknowledged at para 19 that “Parliament is charged with the primary responsibility for deciding the best way of dealing with social problems”. National housing policy was a field where the court would be less ready to intervene, since Parliament had to take into account “broad issues of social and economic policy”. Nevertheless, he said, “even in such a field, where the alleged violation comprises differential treatment based on grounds such as race or sex or sexual orientation the court will scrutinise with intensity any reasons said to constitute justification”. The reasons had to be “cogent” if such differential treatment was to be justified. That approach closely reflects the approach adopted by the European court.

The fact that some grounds of differences in treatment can only be justified by “very weighty reasons”, whereas others can more easily be justified, was also discussed in some of the early cases in the House of Lords. For example, in R (RJM) v Secretary of State for Work and Pensions, para 56, Lord Neuberger of Abbotsbury, having noted that the ground of the differential treatment in relation to a social security benefit was homelessness, observed that the context was “an area where the court should be very slow to substitute its view for that of the executive, especially as the discrimination is not on one of the express, or primary, grounds”.

The case which led to some divergence from the reasoning of the European court, as it now appears, was Humphreys [2012] UKSC 18; [2012] 1 WLR 1545, which concerned indirect discrimination on the “suspect” ground of sex. The argument was that the payment of child tax credit to the parent with primary responsibility for the child, in cases where the child lived part of the time with one parent and part of the time with the other, adversely affected more men than women, as most children in that situation lived mainly with their mothers. It was common ground that the relevant question was whether the legislative policy was “manifestly without reasonable foundation”. The only European cases cited were Stec and Runkee, which concerned a difference of treatment on a “suspect” ground but in the special context of transitional arrangements, and Carson, which did not concern a “suspect” ground. It is unfortunate that cases concerned, like Humphreys, with non-transitional measures and a “suspect” ground, such as Zeman, Luczak and Andrejeva, were not cited.

Lady Hale (with whom the rest of court agreed) interpreted Stec, in accordance with the appellant’s submission, as applying what she described as “the ‘manifestly without reasonable foundation’ test” (para 17) to discrimination on the ground of sex. Referring to the fact that in R (Carson) v Secretary of State for Work and Pensions [2005] UKHL 37; [2006] 1 AC 173 a distinction had been drawn between discrimination on so-called “suspect” grounds, such as race and sex, and discrimination on less sensitive bases, she stated at para 19:

I do not doubt the correctness of that decision, but it now appears to me that the reasoning on justification, following the submissions made, did not reflect the Strasbourg jurisprudence entirely correctly. First, it seems to me that the “manifestly without reasonable foundation” formulation, as used in the Strasbourg judgments, does not express a test, in the sense of a requirement whose satisfaction or non-satisfaction will in itself necessarily be determinative of the outcome. The phrase indicates the width of the margin of appreciation, and hence the intensity of review, which is in principle appropriate in the field of welfare benefits, other things being equal. As I have explained, however, a number of other factors may also be relevant in the circumstances of particular cases, some of which may call for a stricter standard of review. One might then ask, for example in a case concerned with “suspect” grounds, whether “very weighty reasons” have been shown, having regard to the wide margin generally available in that field (as the European court has done: para 130 above). However it is put, the question is more complex than a “test” of whether the policy choice is “manifestly without reasonable foundation” might appear to be if that were regarded as the entirety of the inquiry.

Secondly, the reasoning departed from the Strasbourg approach in its interpretation of Stec and Runkee, and its consequent rejection of a distinction between “suspect” and other grounds of differences in treatment in the field of welfare benefits (matters which could only be appreciated in the light of other Strasbourg authorities which were not before the court). As explained at paras 100-113, 117-124 and 136-142 above, differential treatment on a suspect ground, if it is capable of justification at all, generally (but not always) requires to be justified by “very weighty reasons”. That is so even in the context of measures of social and economic policy which would usually benefit from the “manifestly without reasonable foundation” approach. The “manifestly without reasonable foundation” approach does not, therefore, replace or supersede the requirement for “very weighty reasons” where “suspect” grounds are in issue. Instead, the degree of deference usually appropriate in relation to social or economic policy choices may have to be taken into account in assessing whether “very weighty reasons” have been shown.

Nevertheless, Lady Hale’s insistence on the need for “careful scrutiny” was capable of ensuring an appropriate balance between the competing considerations. In Humphreys itself, Lady Hale carried out a scrupulously careful and balanced analysis. Although her judgment was not expressed in the same language as the European court would have employed, I do not doubt that it was Convention-compliant.

In later domestic cases, the approach adopted in Humphreys has continued to be followed. In SG [2015] UKSC 16; [2015] 1 WLR 1449, the issue concerned a complaint of indirect discrimination on the ground of sex. The majority approached the issue by considering whether the legislation in question had a legitimate aim and was a proportionate means of achieving that aim, applying a “manifestly without reasonable foundation” test, as in Humphreys.

In MA [2016] UKSC 58; [2016] 1 WLR 4550, a wholesale attack was mounted upon the “manifestly without reasonable foundation test”, but was rejected. Lord Toulson, giving the judgment of the majority of the court, accepted that examples could be found of state benefit cases where the European court had spoken of a need for “very weighty reasons” to justify discrimination, and cited Andrejeva, Zeman, Luczak and Vrountou, but saw no inconsistency between those judgments and Humphreys. He concluded that Humphreys laid down the correct test for cases of alleged discrimination on the grounds of sex and disability. That decision was followed by complaints by the unsuccessful appellants to the European court, and the judgment in JD.

More recently, in DA [2019] UKSC 21; [2019] 1 WLR 3289, the court expressed its adherence to “the ‘manifestly without reasonable foundation’ test” in particularly strict terms. Lord Wilson stated at para 65, in reliance on Humphreys and MA:

In the light of the foregoing discussion, I am not persuaded by the argument, based on JD, that the “manifestly without reasonable foundation” formulation can never have any part to play, even in relation to differences of treatment on “suspect” grounds, outside the context of transitional measures. I am not convinced that JD should be understood as going as far as that, in the light of Jurčič v Croatia, Yocheva and Ganeva v Bulgaria and the earlier case law. There is not in any event “a clear and constant line of decisions” to that effect (Manchester City Council v Pinnock (Secretary of State for Communities and Local Government intervening) [2010] UKSC 45; [2011] 2 AC 104, para 48).

Nevertheless, it is appropriate that the approach which this court has adopted since Humphreys should be modified in order to reflect the nuanced nature of the judgment which is required, following the jurisprudence of the European court. In the light of that jurisprudence as it currently stands, it remains the position that a low intensity of review is generally appropriate, other things being equal, in cases concerned with judgments of social and economic policy in the field of welfare benefits and pensions, so that the judgment of the executive or legislature will generally be respected unless it is manifestly without reasonable foundation. Nevertheless, the intensity of the court’s scrutiny can be influenced by a wide range of factors, depending on the circumstances of the particular case, as indeed it would be if the court were applying the domestic test of reasonableness rather than the Convention test of proportionality. In particular, very weighty reasons will usually have to be shown, and the intensity of review will usually be correspondingly high, if a difference in treatment on a “suspect” ground is to be justified. Those grounds, as currently recognised, are discussed in paras 101-113 above; but, as I have explained, they may develop over time as the approach of the European court evolves. But other factors can sometimes lower the intensity of review even where a suspect ground is in issue, as cases such as Schalk, Eweida and Tomás illustrate, besides the cases concerned with “transitional measures”, such as Stec, Runkee and British Gurkha. Equally, even where there is no “suspect” ground, there may be factors which call for a stricter standard of review than might otherwise be necessary, such as the impact of a measure on the best interests of children.

It is therefore important to avoid a mechanical approach to these matters, based simply on the categorisation of the ground of the difference in treatment. A more flexible approach will give appropriate respect to the assessment of democratically accountable institutions, but will also take appropriate account of such other factors as may be relevant. As was recognised in Ghaidan v Godin-Mendoza and R (RJM) v Secretary of State for Work and Pensions, the courts should generally be very slow to intervene in areas of social and economic policy such as housing and social security; but, as a general rule, differential treatment on grounds such as sex or race nevertheless requires cogent justification.

It may also be helpful to observe that the phrase “manifestly without reasonable foundation”, as used by the European court, is merely a way of describing a wide margin of appreciation. A wide margin has also been recognised by the European court in numerous other areas where that phrase has not been used, such as national security, penal policy and matters raising sensitive moral or ethical issues.

It follows that in domestic cases, rather than trying to arrive at a precise definition of the ambit of the “manifestly without reasonable foundation” formulation, it is more fruitful to focus on the question whether a wide margin of judgment is appropriate in the light of the circumstances of the case. The ordinary approach to proportionality gives appropriate weight to the judgment of the primary decision-maker: a degree of weight which will normally be substantial in fields such as economic and social policy, national security, penal policy, and matters raising sensitive moral or ethical issues. It follows, as the Court of Appeal noted in R (Joint Council for the Welfare of Immigrants) v Secretary of State for the Home Department (National Residential Landlords Association intervening) [2020] EWCA Civ 542; [2021] 1 WLR 1151 and R (Delve) v Secretary of State for Work and Pensions [2020] EWCA Civ 1199; [2021] ICR 236, that the ordinary approach to proportionality will accord the same margin to the decision-maker as the “manifestly without reasonable foundation” formulation in circumstances where a particularly wide margin is appropriate.

It is also important to bear in mind that almost any legislation is capable of challenge under article 14. Judges Pejchal and Wojtyczek observed in their partly dissenting opinion in JD, para 11:

The third preliminary matter to be considered concerns the use which can be made of Parliamentary debates and other Parliamentary material when considering whether primary legislation is compatible with Convention rights, having regard to Parliamentary privilege. The issue arises because counsel for both parties placed reliance upon such materials in order to demonstrate the adequacy, or inadequacy, of the consideration given to matters which were argued to be relevant to the proportionality of the legislation, such as its impact upon the interests of the children affected. In broad summary, counsel for the appellant relied upon statements and documents emanating from the Government, including the impact statement and the statement of compatibility made by the Minister in charge of the Bill (paras 13-15 above), in order to show that the Government’s and Parliament’s consideration of relevant issues had been inadequate and in breach of the UNCRC, whereas counsel for the Secretary of State relied upon material placed before Parliament by other bodies as well as the Government, and the discussion of the Bill in committee and in debates (paras 16-20 above), in order to demonstrate that Parliament had had regard to all material factors.

Parliamentary privilege is given statutory expression in article 9 of the Bill of Rights 1689 (1 Will & Mary, sess 2, c 2):

As that statement makes clear, the law of Parliamentary privilege is not based solely on the need to avoid any risk of interference with freedom of speech in Parliament. It is underpinned by the principle of the separation of powers, which, so far as relating to the courts and Parliament, requires each of them to abstain from interference with the functions of the other, and to treat each other’s proceedings and decisions with respect. It follows that it is no part of the function of the courts under our constitution to exercise a supervisory jurisdiction over the internal procedures of Parliament. That principle was affirmed by this court in R (Buckinghamshire County Council) v Secretary of State for Transport [2014] UKSC 3; [2014] 1 WLR 324, in my own judgment at para 110 and in the judgment of Lord Neuberger and Lord Mance at paras 203-206, where they observed (at para 206) that “[s]crutiny of the workings of Parliament and whether they satisfy externally imposed criteria clearly involves questioning and potentially impeaching (ie condemning) Parliament’s internal proceedings, and would go a considerable step further than any United Kingdom court has ever gone”.

Another constitutional fundamental which needs to be borne in mind is that the Government is separate from Parliament, notwithstanding the many connections between the two institutions. As a matter of daily reality, ministers and party whips have to negotiate and compromise in order to secure the passage of the legislation which the Government has promoted, often in an amended form. In fact, as well as in theory, “the legislative function belongs to Parliament not to the executive”: Wilson v First County Trust Ltd (No 2) [2003] UKHL 40; [2004] 1 AC 816, para 111 (“Wilson”) (Lord Hope of Craighead). Accordingly, as Lord Hope observed (ibid), “it is the intention of Parliament that defines the policy and objects of its enactments, not the purpose or intention of the executive”. The reasons which the Government gives for promoting legislation cannot therefore be treated as necessarily explaining why Parliament chose to enact it.

Two other aspects of Parliamentary proceedings are important in this context. First, the will of Parliament finds expression solely in the legislation which it enacts. Parliament does not give reasons for enacting legislation: it simply votes on a motion to approve a proposed legislative text. There is no corporate statement of reasons, and the individual members of Parliament do not give their reasons for voting in a particular way. As Lord Hobhouse stated in Wilson, para 143, “[i]t is not part of the duty of any Member of Parliament to provide or state definitively in Parliament the justification for legislation which the legislature is content to pass”.

Secondly, the decisions which Parliament takes are not necessarily capable of being rationalised in any event. In the first place, Parliament does not operate only, or even primarily, as a debating chamber. It is also a forum for gathering evidence, and for extra-cameral discussion, negotiation and compromise. Furthermore, the way in which members of Parliament vote will usually, but by no means always, reflect party policy, and may be influenced by the discipline imposed by the party whips.

It follows that Parliamentary methods of resolving disputes are very different from judicial methods, aimed at the production of decisions arrived at by an independent and transparent process of reasoning. That is by no means a criticism of Parliament. Its methods reflect the nature of its task: the management of political disagreements within our society so as to arrive, through negotiation and compromise, and the use of the party political power obtained at democratic elections, at decisions whose legitimacy is accepted not because of the quality or transparency of the reasoning involved, but because of the democratic credentials of those by whom the decisions are taken.

A number of consequences follow from this. One is that a ministerial statement of compatibility, made in accordance with section 19 of the Human Rights Act, cannot be ascribed to Parliament. As Lord Hope explained in Anderson v Scottish Ministers [2001] UKPC D5; [2003] 2 AC 602, para 7, it is no more than a statement of opinion by the relevant minister.

A more far-reaching consequence is that the courts have to be careful not to undermine Parliament’s performance of its functions by requiring it, or encouraging it, to conform to a judicial model of rationality. That model is not suitable for resolving differences of political opinion. An insistence on transparent and rational analysis would be liable to make the process of resolving political differences through negotiation, compromise and the exercise of democratic power more difficult and less likely to succeed.

A further consequence is that the intention of Parliament, or (otherwise put) the object or aim of legislation, is an essentially legal construct, rather than something which can be discovered by an empirical investigation. The point is illustrated by Lord Bingham’s comment in R (Countryside Alliance) v Attorney General [2007] UKHL 52; [2008] AC 719, para 40, after identifying the rationale of the legislation there in question, “that this rationale was nowhere expressed in the Act, that this did not reflect the Government’s intention in introducing the Bill and that virtually no Parliamentary statement expressed the rationale in this way”.

It follows from the foregoing that considerable care has to be taken when considering the use of Parliamentary materials in connection with the Human Rights Act. The matter was considered in depth in Wilson, where the House had the assistance of submissions on behalf of the Speaker of the House of Commons and the Clerk of the Parliaments. Lord Nicholls, with whom the other members of the committee were substantially in agreement, stated at para 61 that the courts were now required to evaluate the effect of primary legislation in terms of Convention rights and, where appropriate, make a formal declaration of incompatibility. In carrying out that evaluation, the court had to compare the effect of the legislation with the Convention right. If the legislation impinged on a Convention right, the court must compare the policy objective of the legislation with the policy objective which, under the Convention, might justify a prima facie infringement of the Convention right. When making those two comparisons, the court would look primarily at the legislation, but not exclusively so:

The legislation must also satisfy a proportionality test. The court must decide whether the means used by the legislation to achieve its policy was appropriate and not disproportionate in its adverse effect. For those purposes, reference could be made to Parliamentary debates and other Parliamentary material. Lord Nicholls explained why this was so at paras 63-64:

“To that limited extent”, Lord Nicholls said at para 65, “there may be occasion for the courts, when conducting the statutory ‘compatibility’ exercise, to have regard to matters stated in Parliament”. That followed by necessary implication from the Human Rights Act. In the next paragraph, he said that he expected “that occasions when resort to Hansard is necessary as part of the statutory ‘compatibility’ exercise will seldom arise”, and added that, should such an occasion arise, “the courts must be careful not to treat the ministerial or other statement as indicative of the objective intention of Parliament … It should not be supposed that members necessarily agreed with the minister’s reasoning or his conclusions”.

Lord Nicholls explained the continuing importance of Parliamentary privilege at para 67, which may conveniently be broken up into three principles:

In more recent cases, the Speaker and the Clerk of the Parliaments have accepted Lord Nicholls’ explanation of the position: see R (Heathrow Hub Ltd) v Secretary of State for Transport (Speaker of the House of Commons intervening) [2020] EWCA Civ 213; [2020] 4 CMLR 17, para 158, and R (Project for the Registration of Children as British Citizens) v Secretary of State for the Home Department (Speaker of the House of Commons intervening) [2021] EWCA Civ 193; [2021] 1 WLR 3049, paras 87 and 90.

Authorities since Wilson have generally followed Lord Nicholls’ approach, relying on Parliamentary material relatively rarely (at least until recent times), and using it as an aid to ascertaining or confirming the purpose of the legislation: see, for example, R (Jackson) v Attorney General [2005] UKHL 56; [2006] 1 AC 262, paras 65-66. There are also a number of authorities in which the courts have taken account of material before Parliament, and Parliamentary debates, in assessing the proportionality of legislation. This has been taken to extreme lengths in some recent cases, where counsel have trawled through debates in an effort to establish whether or not the Government complied with the United Kingdom’s obligations under unincorporated international treaties: an illegitimate exercise, as I have explained. But even the older cases raise the question whether, when the court is considering whether a legislative provision is a proportionate means of achieving a legitimate aim, the fact that Parliament can be seen to have been aware of the various interests involved, and can therefore be taken to have considered how a balance should be struck between them, can legitimately be taken into account.

Guidance in relation to this question can be derived from Lord Bingham’s speech in R (Countryside Alliance) v Attorney General, where it was necessary to consider the proportionality of a legislative ban on fox-hunting. Lord Bingham referred in his speech to the fact that the legislation had been passed by a majority of the country’s democratically elected representatives after prolonged and intense debate in Parliament, in the course of which the different views on the subject had been fully expressed (paras 1, 8 and 45). He acknowledged that the existence of duly enacted legislation did not conclude the issue, illustrating the point by reference to Dudgeon v United Kingdom (1981) 4 EHRR 149 and Norris v Ireland (1988) 13 EHRR 186, where legislation criminalising homosexual relations had been found to violate article 8 (para 45). But he pointed out that the legislation there in issue had been enacted a century earlier and was not enforced, as it had ceased to reflect moral perceptions. Here, on the other hand, the House of Lords was dealing with a law which was very recent and must be taken to reflect the conscience of a majority of the nation. He continued (ibid):

As Lord Bingham explained, the degree of respect which the courts should show to primary legislation in this context will depend on the circumstances. Among the relevant factors may be the subject-matter of the legislation, and whether it is relatively recent or dates from an age with different values from the present time. Another factor which may be relevant is whether Parliament can be taken to have made its own judgment of the issues which are relevant to the court’s assessment. If so, the court will be more inclined to accept Parliament’s decision, out of respect for democratic decision-making on questions of political controversy.

In that regard, it is apparent from cases such as Animal Defenders International v United Kingdom (2013) 57 EHRR 21, para 108, and Hirst v United Kingdom (No 2) (2005) 42 EHRR 41, para 79, that the European court takes account of whether the legislature has considered the matters which are relevant to a measure’s compatibility with the Convention, although that is by no means determinative of its decision. Since the European court is likely to take that into account, the objective of the Human Rights Act suggests that domestic courts should do likewise, in order to enable Convention rights to be properly enforced domestically and not only by recourse to Strasbourg.

It is of course true that the relevant question, when considering the compatibility of legislation with Convention rights, is not whether Parliament considered that issue before making the legislation in question, but whether the legislation actually results in a violation of Convention rights. In order to decide that question, however, the courts usually need to decide whether the legislation strikes a reasonable balance between competing interests, or, where the legislation is challenged as discriminatory, whether the difference in treatment has a reasonable justification. If it can be inferred that Parliament formed a judgment that the legislation was appropriate notwithstanding its potential impact upon interests protected by Convention rights, then that may be a relevant factor in the court’s assessment, because of the respect which the court will accord to the view of the legislature. If, on the other hand, there is no indication that the issue was considered by Parliament, then that factor will be absent. That absence will not count against upholding the compatibility of the measure: the courts will simply have to consider the issue without that factor being present, but nevertheless paying appropriate respect to the will of Parliament as expressed in the legislation.

However, it is important to add two caveats. First, the courts should go no further than ascertaining whether matters relevant to compatibility were raised during the legislative process, if they are to avoid assessing the adequacy or cogency of Parliament’s consideration of them, contrary to Lord Nicholls’ third principle (in my numbering: para 176 above). The distinction between determining whether, as a question of historical fact, an issue was before Parliament, on the one hand, and determining the cogency of Parliament’s evaluation of that issue, on the other hand, is real and must be respected. Undertaking a critical assessment of Parliamentary debates would be contrary to both authority and statute. Furthermore, as I have explained at paras 167-171 above, it would mistake the nature of Parliamentary processes, and create a risk that the courts might undermine Parliament’s effectiveness. Trawling through debates should not, therefore, be necessary, and is unlikely to be appropriate: a high level review of whether a topic was raised before Parliament, whether in debate or otherwise, should suffice.

Secondly, the courts must not treat the absence or poverty of debate in Parliament as a reason supporting a finding of incompatibility.

In practice, cases of the present kind generally raise issues which were at the forefront of debate, as they were in the present case. The dispute between the Government and the Opposition in Parliament focused on the impact of the measure on poorer households containing several children: paras 19-20 above. Typically, the organisations which bring cases of the present kind will themselves have campaigned against the legislation during its passage through Parliament, as Lord Bingham noted. They will have made sure that their concerns were drawn to the attention of Parliamentarians, as the Child Poverty Action Group did in the present case: para 19 above.

To recap, the contention that the legislation directly discriminates against children as compared with adults has been rejected on the basis that children and adults are not in relevantly similar situations (paras 56-60 above). The contention that the legislation indirectly discriminates against children as compared with adults has been rejected, on the assumption that the doctrine of indirect discrimination can apply in that context, on the basis that children and adults are not in relevantly similar situations (paras 61-64 above).

There remain (1) the contention that the legislation indirectly discriminates against women as compared with men, contrary to article 14 read together with article 8 and with A1P1, and (2) the contention that the legislation discriminates against children living in households containing more than two children, by comparison with children living in households containing one or two children, contrary to article 14 read together with article 8. That is, of course, the differentiation deliberately made by Parliament in enacting the legislation. The question whether that difference in treatment is justified depends upon whether it pursues a legitimate aim and, if so, whether there is a reasonable relationship of proportionality between the means employed and the aim sought to be realised: Carson (2010) 51 EHRR 13, para 61.

As previously explained, a presumption of discrimination on the ground of gender having been raised as a result of the fact that the limitation affects a greater number of women than men, it is necessary to consider whether the measure has an objective and reasonable justification: that is to say, whether it pursues a legitimate aim, and does so by proportionate means. In that regard, the European court has held that very weighty reasons have to be put forward before a difference in treatment on the ground of gender can be regarded as compatible with the Convention, whether the alleged discrimination is direct or indirect (Di Trizio v Switzerland (Application No 7186/09) (unreported) given 2 February 2016 (“Di Trizio”), paras 82 and 96).

That approach might be contrasted with the approach followed by the United States Supreme Court under the Constitution’s Equal Protection Clause. It has distinguished between direct and indirect discrimination, and held that the disproportionate impact of a measure on a particular group, where a suspect ground is in issue, does not trigger the rule applicable in cases of direct discrimination that the measure must be subjected to the strictest scrutiny and is justifiable only by the weightiest of considerations: Washington v Davis 426 US 229 (1976), p 242. The majority pointed out, at p 248, that a rule that a statute designed to serve neutral ends was nevertheless invalid, absent compelling justification, if in practice it benefited or burdened one group more than another would have far reaching consequences. There seems to me to be force in that observation, particularly in view of the wide scope of article 14. I shall, however, follow the approach adopted by the European court in Di Trizio.

It is apparent from the background material described in paras 13-20 above that there were two related “mischiefs” or problems which prompted the introduction of the legislation. The first was an excessively high level of public spending on welfare benefits, resulting in a large fiscal deficit. Addressing this was a major priority of the Government’s macro-economic policy at the time, and had been a manifesto commitment at the 2015 General Election. Expenditure on tax credits was a particular concern, as it had more than trebled over the previous ten years or so. It was understood that the introduction of the proposed limitation on entitlement to the individual element of child tax credit would result in significant savings: see paras 13-17 above.

The second problem was the fact that persons in receipt of child tax credits were guaranteed a rise in income for every additional child they might choose to have, without limit. That situation was regarded as unfair to persons supporting themselves solely through work, and as an unreasonable burden to impose on the taxpayers who pay for the scheme. Since that issue relates specifically to the design of the legislation so as to focus on households containing three or more children, it is considered at paras 201-209 below in the context of the allegation of discrimination against children belonging to such households. Nevertheless, I take that discussion into account in so far as it is also relevant to the allegation presently under consideration.

Focusing for present purposes on the objective of protecting the economic well-being of the country, that is undoubtedly a legitimate aim for the purposes of the Convention: see, for example, Andrejeva, para 86.

The remaining question, in relation to that objective, is whether the legislation is a reasonably proportionate means of realising Parliament’s aim. There is clearly a rational connection between the objectives pursued by the legislation and Parliament’s decision to limit entitlement to the individual element of child tax credit to the amount payable in respect of two children. It is not in dispute that the measure, by imposing that limitation, will achieve savings in public expenditure, and thus contribute to reducing the fiscal deficit.

A number of criticisms are made of the legislation. They are, for the most part, relevant primarily to the allegation that the limitation discriminates against children living in larger households, and it is convenient to consider them in that context, at paras 205-209 below. Although that discussion is relevant also in the present context, and has been taken into account, I shall focus at present on considerations which bear specifically on the allegation of indirect discrimination against women.

The most important point to be made in the present context is that it was inevitable, if the aims of the legislation were to be achieved, that there would be a greater numerical impact on women than on men. That is because, as counsel for the Secretary of State explained, women constitute 90% of single parents bringing up children, as well as 50% of parents jointly bringing up children. That statistic is accepted by counsel for the appellants. It was also brought to the attention of Parliament: see paras 15-16 above. Since women are disproportionately represented among parents bringing up children, it is inevitable that they will be disproportionately affected by legislation affecting parents bringing up children, including legislation making changes to child-related benefits paid to parents.

That explanation was accepted by the judge. He found as a fact, at para 108 of his judgment, that 90% of single parents are women and that single parent families make up about 33% of families in receipt of child tax credit. He concluded at para 147 that “[i]t is inevitable that, if child-related benefits, paid to a parent and used by the household, are reduced or not made available for a third or further child, that that will affect more women, because of the higher proportion of single-parent households which they make up”. That finding was upheld by the Court of Appeal (para 126), and is undisputed.

In short, more women than men are affected because more women than men are bringing up children. That is an objective fact. There is no suggestion that that is itself the result of discrimination on the ground of sex.

The differential impact on women is not, therefore, a special feature of this measure. It is inherent in any general measure which limits expenditure on child-related benefits. Indeed, even if Parliament had chosen to limit spending on benefits across the board rather than focusing on child tax credit, that approach would have had a greater differential impact on women, according to the Government’s uncontradicted evidence: see the judgment of the Court of Appeal, para 127. The appellants have not suggested any way in which the legitimate aims of the measure might have been achieved without affecting a greater number of women than men. The judge, and the Court of Appeal, discussed the possibility that single parents might have been excluded from the scope of the limitation (with, presumably, a correspondingly stricter limitation on child tax credit paid to couples). As they concluded, however, to have treated single parents more favourably than couples in identical financial circumstances would itself have encountered obvious objections under article 14. In addition, it would have contradicted the second aim of the legislation, namely to address the unfairness and unreasonableness of a situation in which recipients of child tax credits were guaranteed a rise in income for every additional child they might choose to have, without limit.

Once it is understood that the legitimate aims of the measure could not be achieved without a disproportionate impact on women, arising from the demographic fact that they form the majority of parents bringing up children, the only remaining question which can be asked, in relation to proportionality, is whether the inevitable impact on women outweighed the importance of achieving the aims pursued. Parliament decided that the importance of the objectives pursued by the measure justified its enactment, notwithstanding its greater impact on women. I see no basis on which this court could properly take a different view.

There remains the argument that the legislation discriminates against children living in households containing more than two children, by comparison with children living in households containing one or two children, contrary to article 14 read together with article 8. That is, of course, the differentiation deliberately made by Parliament in enacting the legislation.

Parliament’s aims in enacting the legislation, and consequently differentiating between the treatment of households with one or two children and households with three or more children, were explained at paras 190-191 above: first, to promote the economic well-being of the country by reducing excessive public expenditure on welfare benefits, with spending on child tax credit being a particular concern; and secondly, to address what was regarded as an unfair and unreasonable aspect of the child tax credit system, namely that recipients were guaranteed a rise in income for every additional child they might choose to have, without limit. In that regard, the decision to limit the individual element of child tax credit to the amount referable to two children ensured that the measure would not affect families of average or below-average size.

As explained at para 192 above, the objective of protecting the economic well-being of the country is undoubtedly a legitimate aim for the purposes of the Convention. In particular, a welfare benefits scheme such as child tax credit “has limited resources and must therefore be guided in part by the principle of control of expenditure”, as the European court observed in Di Trizio, para 96. In that regard, the objective of ensuring that a benefits system is fair and reasonable must also be legitimate. The benefits system is sometimes described as an expression of social solidarity: the duty of any community to help those of its members who are in need. The system must be fair and reasonable (not least in the case of non-contributory benefits), if that solidarity is not to be weakened.

The remaining question is whether the legislation is a reasonably proportionate means of realising Parliament’s aims. In answering that question, it is important to note that the basis of the differential treatment - namely, whether the number of children living in a household is two or less, or is greater than two, is not one of the grounds of differential treatment calling for “very weighty reasons”: see para 114 above. Since the legislation is a general measure of social and economic strategy, involving an assessment of priorities in the context of the allocation of limited state resources, it follows that Parliament’s assessment that the difference in treatment is justified should be treated by the courts with the greatest respect. At the same time, since the measure affects children, the courts also have to bear in mind the significance of their best interests to the assessment of proportionality.

Approaching the matter on that basis, there is clearly a rational connection between the objectives pursued by the legislation and Parliament’s decision to limit entitlement to the individual element of child tax credit to the amount payable in respect of two children. It is not in dispute that the measure, by imposing that limitation, will achieve savings in public expenditure, and thus contribute to reducing the fiscal deficit. It is true that that does not in itself explain why households should be treated differently, depending on the number of children they contain. In that regard, it is also necessary to take account of the second objective pursued: namely, to ensure that the scheme is fair and reasonable, by limiting the extent to which recipients of child tax credit are guaranteed a rise in income if they have additional children. Plainly, a difference in treatment based on the number of children living in a household is unavoidable if that aim is to be realised. Parliament’s choice has been to set the limitation on entitlement at a level which will not affect families of an average size.

That decision is criticised in these proceedings on a variety of grounds. It is pointed out, for example, that persons supporting themselves solely through work generally have higher incomes than persons receiving child tax credit, and are therefore better able to afford to have additional children. This criticism seems to me to miss the point. The concern expressed about the fairness and reasonableness of the child tax credit system as it stood prior to the introduction of the limitation was that persons not in receipt of child tax credit had to make decisions about the size of their families in the knowledge that they would have to fund the cost of supporting additional children from their own resources, whereas persons in receipt of child tax credit were provided out of public expenditure with a guaranteed increase in their income for every additional child that they chose to have, without limit. Counsel’s point about absolute income levels does not address that concern.

It is also pointed out that a couple may decide to have a third or subsequent child at a time when they reasonably believe that they will be able to support the child out of their own resources, only for some misfortune to render them dependent on welfare benefits. However, there are, as explained in para 9 above, a variety of benefits payable to families with children which provide protection against risks of that kind. How far the welfare system should go to protect families against the vicissitudes of life is a matter on which opinions in our society differ greatly, and of which Parliament is the best judge. It is also pointed out that some pregnancies are unplanned. However, an exception exists under the legislation for pregnancies which result from non-consensual sex: see para 8 above. Beyond that, to create an exception for unplanned pregnancies, resulting for example from casual relationships or from the failure of contraceptive measures, would appear to be completely impractical: how would such exceptions be applied in practice? This is an example of a situation in which it is legitimate for the legislature to adopt a general rule, even if it may have unfortunate consequences in some individual cases: as was observed in Carson, para 62, any welfare system, to be workable, may have to use broad categorisations.

It is also argued that the legislation is not in the best interests of children living with persons whose entitlement to child tax credit is affected by the limitation. The argument was advanced on the basis that the Government had breached the UK’s obligations under unincorporated international treaties. For the reasons I have explained, the court cannot entertain such an argument. But the best interests of the children affected remain relevant to the assessment of proportionality. Plainly, the amount of money provided under the scheme for the support of such children is less, per child, than is provided to persons whose number of children is below the limit. That is something which Parliament must have taken into account, as the debate over that issue formed an important part of the background to the legislation, and the effect on children in larger households was in any event an obvious consequence: see paras 18-20 and 185 above. It was, inevitably, something to be taken into account, rather than a conclusive argument. It might be argued that children’s best interests would always be better served by a more generous benefits system. But Parliament was told that reducing spending on welfare benefits would allow the Government to protect other expenditure of benefit to children: on education, childcare and health (para 18 above). Furthermore, the difficult question is not so much what would be in the best interests of children, but the extent to which it is fair, economically desirable and socially acceptable to impose the cost of supporting children, whose parents lack the means to do so themselves, on other members of society. Parliament must have considered that the impact of the limitation upon the interests of the children who would be affected by it was outweighed by the reasons for introducing it.

The assessment of proportionality, therefore, ultimately resolves itself into the question as to whether Parliament made the right judgment. That was at the time, and remains, a question of intense political controversy. It cannot be answered by any process of legal reasoning. There are no legal standards by which a court can decide where the balance should be struck between the interests of children and their parents in receiving support from the state, on the one hand, and the interests of the community as a whole in placing responsibility for the care of children upon their parents, on the other. The answer to such a question can only be determined, in a Parliamentary democracy, through a political process which can take account of the values and views of all sections of society. Democratically elected institutions are in a far better position than the courts to reflect a collective sense of what is fair and affordable, or of where the balance of fairness lies.

That is what happened in this case. The democratic credentials of the measure could not be stronger. It was introduced in Parliament following a General Election, in order to implement a manifesto commitment (para 13 above). It was approved by Parliament, subject to amendments, after a vigorous debate at which the issues raised in these proceedings were fully canvassed, and in which the body supporting the appellants was an active participant (para 185 above). There is no basis, consistent with the separation of powers under our constitution, on which the courts could properly overturn Parliament’s judgment that the measure was an appropriate means of achieving its aims.

For the foregoing reasons, I conclude that the appeal should be dismissed.