R (on the application of SG and others (previously JS and others)) v Secretary of State for Work and Pensions

These appeals raise the question whether it was lawful for the Secretary of State to make subordinate legislation imposing a cap on the amount of welfare benefits which can be received by claimants in non-working households, equivalent to the net median earnings of working households. The legislation is challenged under the Human Rights Act 1998 primarily on the basis that it discriminates unjustifiably between men and women, contrary to article 14 of the European Convention on Human Rights (“the ECHR”) read with article 1 of Protocol No 1 to the ECHR (“A1P1”).

The discrimination arises indirectly. The cap affects all non-working households which would otherwise receive benefits in excess of the cap. Those are predominantly households with several children, living in high cost areas of housing. The heads of such households are entitled, in the absence of the cap, to relatively high amounts of child benefit, which is payable in direct proportion to the number of children. They are also entitled, in the absence of the cap, to relatively high amounts of housing benefit, which reflects the rental cost of the accommodation in which the household lives, and tends therefore to reflect to some extent the size of the household and, more particularly, the level of rental values in the area. In practice, this means that non-working households with several children, living in London, are most likely to be affected. The majority of non-working households with children are single parent households, and the vast majority of single parents are women (92% in 2011). A statistically higher number of women than men are therefore affected by the cap. The great majority of single parent non-working households are however unaffected by the cap.

It is argued that the cap also affects victims of domestic violence, because they may be temporarily housed in accommodation which is relatively expensive (the rent for such accommodation having tended to reflect the amount of housing benefit payable), and in that event are entitled, in the absence of the cap, to relatively high amounts of housing benefit. That will also be the position if they are entitled to housing benefit in respect of both the temporary accommodation and also other accommodation to which they hope to return. Victims of domestic violence are also predominantly women.

The justification put forward for the cap is one of economic and social policy, namely that it is necessary (1) to set a reasonable limit to the extent to which the state will support non-working households from public funds, (2) to provide the members of such households of working age with a greater incentive to work, and (3) to achieve savings in public expenditure at a time when such savings are necessary in the interests of the economic well-being of the country.

Article 14 provides:

As is apparent from its terms, article 14 can only be considered in conjunction with one or more of the substantive rights or freedoms set forth in the Convention. In the present case, the relevant right is that set forth in A1P1:

The general approach followed by the European Court of Human Rights in the application of article 14 was explained by the Grand Chamber in Carson v United Kingdom (2010) 51 EHRR 369, para 61:

A violation of article 14 therefore arises where there is:

In practice, the analysis carried out by the European court usually elides the second element – the comparability of the situations – and focuses on the question whether differential treatment is justified. This reflects the fact that an assessment of whether situations are “relevantly” similar is generally linked to the aims of the measure in question (see, for example, Rasmussen v Denmark (1984) 7 EHRR 371, para 37).

In relation to the third element, the court has referred to the criteria laid down in the second paragraphs of articles 8 to 11 of the Convention as legitimate aims, where article 14 has been read in conjunction with those articles. In Sidabras v Lithuania (2004) 42 EHRR 104, for example, the court stated at para 55 that the difference in treatment “pursued the legitimate aims of the protection of national security, public order, the economic well-being of the country and the rights and freedoms of others”. The court has also treated aims which are legitimate in the public interest in the context of A1P1, such as securing social justice and protecting the state’s economic well-being, as legitimate aims when article 14 has been read in conjunction with that article, as for example in Hoogendijk v The Netherlands (2005) 40 EHRR SE 189 and Andrejeva v Latvia (2009) 51 EHRR 650.

National authorities enjoy a margin of appreciation in assessing whether and to what extent differences in treatment are justified. The European court has emphasised the width of the margin of appreciation in relation to general measures of economic or social strategy, stating in its Carson judgment at para 61:

Article 14 is not confined to the differential treatment of similar cases: “discrimination may also arise where states without an objective and reasonable justification fail to treat differently persons whose situations are significantly different” (Pretty v United Kingdom (2002) 35 EHRR 1, para 87). An example is the case of Thlimmenos v Greece (2001) 31 EHRR 411, where this type of discrimination was first recognised.

The European court has also accepted that a difference in treatment may be inferred from the effects of a measure which is neutral on its face. In DH v Czech Republic (2007) 47 EHRR 59, the court stated at para 175:

The present case is essentially of a similar kind: the cap, in the form in which it has been established, affects a higher number of women than men because of differences in the extent to which the sexes take responsibility for the care of children following the break-up of relationships. Whether that differential effect has an objective and reasonable justification depends on whether the legislation governing the cap, which brings about that differential effect, has a legitimate aim and is a proportionate means of realising that aim.

When applying article 14 in the context of welfare benefits, the European court recognises the need for national rules to be framed in broad terms, which may result in hardship in particular cases. In its Carson judgment, for example, the Grand Chamber stated at para 62:

In considering the issues arising under article 14 in the present case, I shall begin by examining the process which led to the legislation with which we are concerned, in order to identify the aims pursued by the legislation and information relevant to the issue which the court has to determine. Consideration of the Parliamentary debates for that purpose is not inconsistent with anything said in the case of Wilson v First County Trust Ltd (No 2) [2003] UKHL 40; [2004] 1 AC 816: the purpose of the exercise is not to assess the quality of the reasons advanced in support of the legislation by Ministers or other members of Parliament, nor to treat anything other than the legislation itself as the expression of the will of Parliament.

On 22 June 2010 the Chancellor of the Exchequer laid before Parliament his Emergency Budget: Budget 2010 (HC 61). It set out a five year plan to rebuild the British economy by reducing the structural fiscal deficit. The plan involved reductions in Government spending of £32 billion per annum by 2014/15. These reductions would include £11 billion in savings achieved through reforms of welfare. The reforms were intended to make the welfare system fairer and more affordable, to reduce dependency, and to promote employment.

The following month, the Department for Work and Pensions (“the Department”) published a consultation document, 21^st century Welfare (Cm 7913), seeking views on options to reform the system of working age benefits. In response to a question about the steps which the Government should consider to reduce welfare dependency and poverty, many respondents answered that the most effective way would be to ensure that people were significantly better off working than on benefit, and suggested the introduction of a benefit cap to restrict the amount of welfare payments which people could receive while out of work: Consultation Responses to 21^st century Welfare (2010) (Cm 7971). A common view was that the cap should be set by reference to the national minimum wage. This idea was then discussed at the Department’s Policy and Strategy Forum, at which the Department engages with groups representing benefit recipients.

On 11 October 2010 the Secretary of State announced the Government’s intention to set a cap on benefits for non-working households. Further details were provided in the Spending Review 2010 (Cm 7942), which announced the intention to cap non-working household benefits at around £500 per week for couple and single parent households, and around £350 per week for single adult households, so that no non-working household would receive more in welfare than the median after tax earnings of working households. A “household” would comprise one or two adults living together as a couple, plus any dependent children living with them. The cap would be implemented by local authorities, which would assess the benefit income of housing benefit claimants, and reduce the payments of housing benefit where necessary to ensure that they did not receive more than the cap.

It is relevant to note, in relation to submissions concerning the impact of the cap upon children, that the Spending Review made clear the Government’s belief that the proposed reforms would promote the interests of children:

Contemporaneously with the Spending Review, HM Treasury published its Overview of the Impact of Spending Review 2010 on Equalities. This document considered the impact of the Spending Review on groups protected by equalities legislation, including women. It noted that decisions had been taken within the Spending Review which protected most of the services which women used more than men, in particular health, social care, early years and childcare. In order to protect those areas of spending, savings had to be made in other areas, including welfare. In relation to benefits, it was noted that any changes affecting single parent households would affect more women than men.

In November 2010 the White Paper, Universal Credit: Welfare That Works (Cm 7957) was published. It included the benefit cap as part of the design of universal credit. The Parliamentary Select Committee on Work and Pensions considered the White Paper, and received evidence from, amongst others, the two interveners in the present proceedings, the Child Poverty Action Group and Shelter, as to the likely impact of the cap: House of Commons Work and Pensions Committee, White Paper on Universal Credit, Oral and Written Evidence (2011) (HC 743). The impact on larger families, and those living in high cost areas, was highlighted. That reflected the fact, recognised from the outset, that the cap would primarily affect households receiving large amounts of child-related benefits and large amounts of housing benefit.

On 16 February 2011 the Welfare Reform Bill received its First Reading in the House of Commons. Clauses 93 and 94 set out the proposed provisions in respect of a benefit cap. As is customary in the area of social security, the clauses were drafted on the footing that the primary legislation would establish a framework for secondary legislation in which the rules would be set out in detail.

At the same time, the Department laid before Parliament an Impact Assessment for the Household Benefit Cap. That document explained the three policy aims: to deliver fiscal savings, to make the system fairer as between non-working households and working households, and to incentivise the non-working to work. It explained the policy options which had been considered, and the reasons for adopting the preferred option. In particular, it explained that consideration had been given to applying the cap to working households which also received benefits, but that it had been decided that they should be exempted, as “including recipients of working tax credit among those affected by the cap would seriously reduce incentives to work” (p 5). It had also been decided to exempt those in receipt of disability living allowance and constant attendance allowance, as disabled people with additional care or mobility costs had less ability to alter their spending patterns or reduce their housing costs in response to a cap on benefit. War widows and widowers would also be exempted, in order to recognise their sacrifice. Consideration had also been given to setting the cap at a different level, but it was decided that to base it on net median household earnings would best represent the average take home pay of working households.

The document explained that about 50,000 households would have their benefits reduced (representing around 1% of the out-of-work benefit case-load), and that affected households would lose an average of £93 per week. Those affected by the cap would need to choose between taking up work (in which event they would no longer be affected), obtaining other income (such as child maintenance payments from absent parents: other reforms were designed to make it more difficult for absent parents to evade their obligation to provide financial support to single parents), reducing their non-rent expenditure, negotiating a lower rent, or moving to cheaper accommodation.

In March 2011 the Department laid before Parliament its Household Benefit Cap Equality Impact Assessment. The document stated that the cap was intended to reverse “the disincentive effects and detrimental impacts of benefit dependency on families and children” (para 5). The likely impact was analysed according to disability, race, gender, age, gender reassignment, sexual orientation, religion or belief, and pregnancy or maternity. In relation to gender, it was estimated that around 60% of claimants who had their benefits cut would be single females, whereas 3% would be single men. That was because around 60% of households affected would comprise single parents living with children, and single parents living with children were predominantly women. The impact of the cap on single parents would be mitigated by the provision of support to help them to move into work. Single parents would also be exempt from the cap if they worked for only 16 hours per week, whereas other single claimants would have to work for at least 30 hours per week before they were exempt.

The policy was subjected to detailed and vigorous scrutiny by both Houses of Parliament, over a period of more than 12 months, during the passage of the Bill through Parliament. That scrutiny was assisted by a number of House of Commons Research Papers, and by briefings prepared by organisations opposed to the policy. During the Committee stage which followed the Second Reading debate in the House of Commons, the Public Bill Committee also received evidence from many organisations with an interest, including the interveners. Consideration was also given to reports on the Bill produced by the Office of the Children’s Commissioner, which focused upon the impact on children, and by the Equality and Human Rights Commission. The former report expressed concern about the potential impact on children if households affected by the cap moved home in order to reduce their housing costs. It also expressed concern about the potential impact if households were unable to reduce their housing costs.

The discussion in Committee, and in the earlier Second Reading debate, concerned a number of issues, including the impact of the cap on single parents, its impact on children, its impact on those living in temporary accommodation, and the appropriateness of fixing the cap according to the net median earnings of working households, when working households receiving net median earnings might also receive certain benefits.

In relation to the impact on single parents, it was argued that if such households included children under five years of age, there would be less likelihood of the parent being able to take up work, because of child-care responsibilities and the potential cost of child care. Amendments to the Bill were tabled in Committee that would have exempted households from the cap where a single parent had children under five years of age, or where work was not financially more advantageous due to child care costs.

In relation to the impact on children, it was argued that if households whose benefits were capped moved to areas where housing was less expensive, there could be consequent disruption in the supervision of children who were at risk of abuse, and also disruption of children’s schooling. If such households did not move to cheaper areas, they would have to economise in other ways. Amendments were moved in both Houses that child-related benefits should be excluded from the scope of the cap, and that the cap should be related to household size.

The potential impact on households living in temporary accommodation, at a relatively high cost, was also emphasised. Amendments were moved in both Houses that would have exempted households which were owed a duty by the local authority to be supported in temporary accommodation.

In relation to the use of net median household earnings as the benchmark, it was argued that the cap would leave the households affected worse off than working households with equivalent earnings, since some benefits were payable to households receiving average earnings. An amendment was tabled in Committee to require the cap to reflect net average earnings “plus in work benefits which an average earner might expect to receive”: Hansard (HC Debates), 17 May 2011, col 970. An amendment to similar effect was also proposed in the House of Lords.

In responding to these arguments during the discussion in Committee on 17 May 2011, the Minister emphasised the need to create a welfare system which was fair in the eyes of the general public and commanded public confidence, and the need to address a culture of welfare dependency. In relation to the former point, he stated that it did no service to welfare claimants if they were seen to be receiving amounts of money from the state that exceeded the average earnings of people who were working. That encouraged the view that there was something wrong, and it had the effect of stigmatising those claimants. It was important to help people into work, and it was also important to have a welfare system in which the public had confidence. At present, it was clearly demonstrable that that was not the case (col 950). In that regard, the Minister referred to the stigmatisation of non-working families who received high levels of benefit, and to the level of public support for the introduction of a cap on benefits. He went on to say that it was not reasonable or fair for out-of-work households to have a greater income from benefits than the net average weekly wage of working households (col 952). The proposed cap for couples and families was equivalent to an earned salary of £35,000 per annum, which was considered fair (col 984).

In response to the argument that average earnings were not a proper basis for comparison, since households on average earnings might also be in receipt of benefits, the Minister responded that it was necessary, for public confidence in the benefit system, to have a cap related to average earnings. He acknowledged that the proposed level of the cap was lower than the total income of a working household on average earnings which was receiving in-work benefits, but said that it was necessary to ensure that people were better off in work (cols 952 and 975). The Minister also observed that the policy would only succeed in its objectives of influencing behaviour and increasing public confidence in the benefits system if there was a simple rule which people could understand (col 954).

In relation to arguments based on the different needs of different types of household, such as those with several children, the Minister observed that there was a divide in philosophical view between those who thought that the cap should vary according to household size and other characteristics, and those who believed that there should be some limit to the overall benefits that the state should provide. Working people on low incomes had to cope with difficult circumstances, and they had to live within their means (cols 952, 973). Their earnings were not determined by the size of their families, and the Government believed that the same principle should apply to the level of the cap (col 975). Households whose benefits were capped might need to move to cheaper accommodation, but like other families they had to live in accommodation that they could afford.

In relation to those living in temporary accommodation, the Minister observed that local authorities had a legal duty to provide accommodation which was suitable for homeless applicants, and suitability included affordability. That observation was consistent with the decision in R (Best) v Oxford City Council [2009] EWHC 608 (Admin), approved by the Divisional Court in the present proceedings: [2013] EWHC 3350 (QB), [2014] PTSR 23, para 53. The Minister explained that, whatever the cost of the accommodation might be, the local authority could pass on only a charge that the applicant could afford. The issue of housing costs for those in temporary accommodation was being considered.

In relation to this matter, it is relevant to note the evidence given in these proceedings by Mr Robert Holmes, the Department’s lead official on the benefit cap policy. He explains that the Government used to reimburse local authorities, via the housing benefit system, the rent which they charged claimants for the provision of temporary accommodation, up to a maximum for each property of £500 per week in London and £375 per week elsewhere. It became clear that some local authorities were using this system to generate surplus revenues, by charging claimants at or about the maximum level regardless of the rental value of the accommodation in question. Claimants in temporary accommodation were then reluctant to seek employment, as they were concerned that they might lose their housing benefit and be unable to pay these artificially inflated rents. The Government was unwilling to exempt temporary accommodation from the cap, as it considered that to do so would continue to subsidise inflated rents and would discourage claimants from obtaining work. It decided instead to provide additional support for those in temporary accommodation through the discretionary housing payments scheme, to which it will be necessary to return.

The Bill was also considered in detail by the House of Lords, which was provided with an updated version of the Housing Benefit Cap Equality Impact Assessment (2011). The discussion in the House of Lords focused particularly upon the impact of the cap on households with children, and upon the use of median earnings, rather than income inclusive of benefits, as the benchmark. In the course of the discussion, the Minister gave an assurance that he had considered the requirements of the Human Rights Act 1998 and the ECHR in respect of the policy, and was satisfied that the way in which the Government would implement the clauses in question would meet those requirements (Hansard (HL Debates), 21 November 2011, col GC415).

In relation to the use of median earnings as the basis of the cap, the Minister explained that it necessarily followed, by definition, that half the working households in the UK would have earnings below the level of the cap (col GC425).

In relation to the impact of the cap on households with children, an amendment seeking to exempt single parents with children under five was opposed by the Government. In response to the argument that, since such parents were not obliged to seek work in order to be eligible to receive benefits, they ought also to be exempted from the cap on the amount of any benefits which they might receive, the Minister stated that the cap was intended to act as an incentive to work. Although single parents with children under five were not required to seek work as a condition of receiving benefits, that did not mean that the Government did not want to encourage them to find employment. The amendment would undermine the fundamental principles underpinning the cap: that ultimately there had to be a limit to the amount of benefit that a household could receive, and that work should always pay (col GC421).

A proposed amendment to exclude child benefit from the scope of the cap was opposed by the Government on the basis that its policy was that there should be a reasonable limit to the overall amount of support that non-working households could receive in welfare payments, that child benefit was as much part of that support as other welfare payments, and that it should therefore be taken into account in deciding whether the limit had been reached. It was estimated that excluding child benefit from the scope of the cap would reduce the savings from the cap by 40 to 50%, and that also excluding child tax credit would reduce the savings by 80 to 90% (Hansard (HC Debates), 28 November 2011, col 763W). There would be a similar impact upon the number of households affected (Hansard (HC Debates), 23 May 2011, col 496W).

The Bill was also scrutinised by the House of Lords and House of Commons Joint Committee on Human Rights, which considered the human rights effects of the Bill and published its report in December 2011 (HL Paper 233; HC 1704). In written evidence to the Committee, the Secretary of State stated that it was the Government’s view that, if A1P1 was engaged, the measures in the Bill were proportionate to the legitimate aim of securing the economic well-being of the country. He observed that the greater employment of single parents would have a positive effect on child poverty, and that there was a wide range of support available to single parents seeking employment, to take account of their role as the main carer for their child. He added that the Government believed that the effect of the cap was proportionate, taking into account (1) the amount of the cap and the fact that it would be based on average household earnings, (2) the fact that claimants would be notified of the cap and given time to adjust their spending to accommodate their new levels of benefit, and (3) the fact that the cap would affect relatively few households and that those affected would continue to receive a substantial income from benefits.

At the Report stage in the House of Lords, the Bill was amended so as to exclude child benefit from the scope of the cap. When the Bill returned to the House of Commons, the House considered and voted against that amendment. When the Bill subsequently returned to the House of Lords, the House agreed, on a vote, not to insist on the amendment.

During the Bill’s passage, Ministers indicated that some of the concerns expressed in Parliament, many of them reflected in other proposed amendments, would be considered as the policy was developed. So it proved. One example was the introduction of a period of grace for benefit claimants who had previously been employed, so that their benefits would not be capped for a period of 39 weeks after they had last been in employment. That development reflected concerns which had been expressed about the application of the cap to households in which someone had been in work but had been made redundant or had left work in order to care for a child. It was also understood that child care responsibilities might make it difficult for some single parents to seek work and, by that means, to secure exemption from the cap. Measures were taken to address those difficulties by exempting benefits used to pay for child care (meeting 70% of the cost) from the cap, by providing single parents with job-focused interviews to assist them in finding work, and by setting the number of hours required to be worked by a single parent, in order to obtain exemption from the cap, at a lower level, of 16 hours per week, than for other claimants. Another development was the introduction of an exception to prevent payments covering the cost of accommodation in refuges, for women who had been victims of domestic violence, from being taken into account. It will be necessary to return to that matter. Measures were also taken to ensure that the supervision of children at risk of ill-treatment was not jeopardised in the event that their families moved to less expensive areas to live.

A decision was also taken to provide additional funding of £65m in 2013/2014 and £35m in 2014/15 for discretionary housing payments under the Discretionary Financial Assistance Regulations 2001 (SI 2001/1167) (“the DHP Regulations”). These are payments made by local authorities to claimants who require further financial assistance, in addition to any welfare benefits, in order to meet housing costs. They do not count towards the cap. As is stated in the guidance for local authorities published by the Government, the additional funding is intended to provide assistance to a number of groups who are likely to be particularly affected by the cap, including those in temporary accommodation, victims of domestic violence, families with children at school, and households moving to, or having difficulty finding, more appropriate accommodation. Households in those categories may be unable to avoid high costs in the short term: they may, for example, have to delay a move until suitable arrangements can be made for the education of children, or may require financial assistance to pay the deposit on a new home and the initial instalment of rent. The additional funding was intended to help them to meet those costs. The Government also undertook to review the operation of the cap, as had been recommended by the Joint Committee on Human Rights, and to lay before Parliament a report on its impact after a year of operation.

The Welfare Reform Act 2012 (“the 2012 Act”) received Royal Assent in March 2012. The provisions relevant to the cap are sections 96 and 97.

Section 96 enables regulations to provide for a benefit cap to be applied to the welfare benefits to which a single person or couple is entitled. For the purposes of the section, applying a benefit cap means securing that:

“Welfare benefits” are any benefit, allowance, payment or credit prescribed in regulations: section 96(10). The regulations cannot however prescribe as welfare benefits either state pension credit or retirement pensions: section 96(11). The “relevant amount” is an amount specified in regulations, which must be determined by reference to the average weekly earnings of a working household after deductions in respect of tax and national insurance: sections 96(5), (6) and (7). More detailed provision in respect of the benefit cap arrangements, including the welfare benefits or benefits from which a reduction is to be made, and any exceptions to the application of the benefit cap, are to be set out in the regulations: section 96(4). The regulations are to be made by the Secretary of State, and the first such regulations must be approved by Parliament under the affirmative resolution procedure: sections 96(10) and 97(3). Subsequent regulations must be approved under the negative resolution procedure.

Before laying draft regulations before Parliament, the Department consulted interested bodies, including the statutory Social Security Advisory Committee, Citizens Advice, Crisis and Shelter. That consultation influenced some of the policy changes which I mentioned in paras 44-45.

On 16 July 2012 the Benefit Cap (Housing Benefit) Regulations 2012 (SI 2012/2994) (“the Regulations”) were laid in draft before both Houses of Parliament. At the same time, the Department published updated impact assessments in respect of the cap. It was then estimated that 56,000 households would be affected (1% of the out-of-work benefit caseload), losing on average around £93 per week. 39% of households affected were expected to be couples with children, and 50% were expected to be single parents with children. Because single parents were predominantly women, 60% of affected claimants were expected to be single women, compared with 10% who were expected to be single men. Almost all the local authorities most affected were expected to be in London, reflecting the higher rents payable there.

Parliament received submissions on the draft regulations from a number of bodies, including Shelter. The draft regulations were considered by the House of Lords Secondary Legislation Scrutiny Committee, and were debated by the House of Lords Grand Chamber on 6 November 2012. They were also considered by the House of Commons Delegated Legislation Committee on the same date. The issues then considered included temporary accommodation, including women’s refuges and other accommodation for victims of domestic violence, the impact upon children of households moving to areas where housing was less expensive, and the greater difficulty which people who moved out of London might experience in obtaining work. The draft regulations were approved by both Houses of Parliament, and the Regulations were then made.

As had been announced, the Regulations fix the cap at £350 per week for single persons and £500 for families and couples, equivalent to gross salaries of £26,000 and £35,000 per annum respectively. These figures are slightly above the median earnings of single persons and couples respectively. They are well above the national minimum wage, which in 2012 was about £12,500 per annum for a 40 hour week. The Regulations list the benefits which are to be treated as welfare benefits. As anticipated, they include the main out-of-work benefits, together with child benefit, child tax credit and housing benefit. Again as anticipated, exceptions from the application of the cap are made in respect of households where a person receives specified benefits based on disability or service in the armed forces, and in respect of households where a single parent works for 16 hours per week or a couple work for 24 hours (provided one of them works for 16 hours). Provision is made for the 39 week period of grace.

In response to concerns expressed about the potential impact of the cap on households living in exempt accommodation (ie accommodation provided by housing associations, charities, other voluntary bodies or county councils to persons receiving care, support or supervision provided by or on behalf of the landlord), including in particular those living in refuges for victims of domestic violence, the Regulations were amended with effect from 15 April 2013 (when, as I shall explain, the cap first came into partial effect) by the Benefit Cap (Housing Benefit) (Amendment) Regulations 2013 (SI 2013/546). The effect of the amendment was that housing benefit provided in respect of such accommodation was to be disregarded for the purposes of the cap. In response to contentions that some women’s refuges fell outside the definition of exempt accommodation, the Minister announced in April 2013 that the issue was being addressed and that proposals would be brought forward at the earliest opportunity. The Housing Benefit and Universal Credit (Supported Accommodation) (Amendment) Regulations 2014 (SI 2014/771) (“the 2014 Regulations”) were subsequently made, after the present proceedings were under way. They replace the concept of “exempt accommodation” with a broader concept of “specified accommodation”, which encompasses a wider range of accommodation provided for vulnerable people, including the women’s refuges previously excluded.

The Regulations were made in November 2012, more than two years after the intention to introduce the cap had been announced. From April 2012 jobcentres and local authorities implemented arrangements to provide support to households that would be affected by the cap and assist them in deciding how to respond. In May 2012 jobcentres wrote to all claimants potentially affected by the cap, notifying them that they might be affected and explaining the support available. That support included assistance from dedicated staff in moving into the labour market, obtaining access to child care provision and negotiating rent reductions with private landlords, together with advice on housing options and household budgets. A help line was also set up to provide information about the changes and the support available. Employment events were organised with local employers and training bodies. Further letters were sent to claimants in October 2012, February 2013 and March 2013. Claimants were also contacted by telephone and, where that proved ineffective, were visited. The cap was then introduced in phases, during which its impact was monitored by the Department. On 15 April 2013 the cap was applied in four local authority areas in London. Between 15 July 2013 and the end of September 2013 the cap was applied in other local authority areas.

Since the introduction of the cap, its impact has been discussed at meetings of the Benefit Cap Project, a forum for meetings between the Department and interested bodies, including voluntary organisations working with children and the homeless.

From August 2013 the Department published a number of reports on the impact of the cap. The most recent report, at the time when these appeals were heard, was that published in March 2014, which contained data for the period to January 2014. It reported that 38,665 households had had their housing benefit capped. 28% of the households which had at one time been capped were no longer capped. 39% of those had become exempt because a member of the household had entered work. 27% were no longer claiming housing benefit or had reduced their rent so as to come below the cap. Of the 20 local authorities with the highest number of capped households, 19 were in London. 95% of capped households included children. 59% of capped households, and 62% of capped households with children, comprised a single parent with children.

In response to a request from this court, counsel also provided the Department’s analysis of the data for the period up to March 2014 in respect of single parent households including a child under five years of age. 29% of such households which had at one time been capped were no longer capped. 38% of those had become exempt because a member of the household had entered work. These figures are in line with those for all households.

According to the Department’s most recent estimate as at the date of the hearing, the cap is expected to save £110m in 2013/2014 and £185m in 2014/2015. This level of savings is expected to continue over the longer term. These figures do not take into account the implementation costs or the additional funding made available for discretionary housing payments. Nor, on the other hand, do they take account of any reduction in benefit payments, or any receipts from income tax or national insurance, resulting from claimants moving into work.

There is no challenge in these proceedings to the 2012 Act: it is not argued that section 96 is incompatible with the ECHR. It follows that there is no challenge to the principle of a cap, the impact of which is inevitably greatest for those who would otherwise be entitled to the highest amount of relevant benefits. Nor is there any challenge to the fixing of one “relevant amount” (ie the cap) for single claimants and another for all other households, rather than the relevant amount being tailored to individual circumstances. Nor is there any challenge to the fixing of the “relevant amount” by reference to estimated average net household earnings, rather than by reference to estimated average net household income inclusive of benefits. The challenge is primarily to the compatibility of the Regulations with article 14 of the ECHR read in conjunction with A1P1.

In considering the compatibility of the Regulations with article 14 in conjunction with A1P1, the first question is whether there is an interference with possessions. That is not a straightforward question: as the European court explained in Valkov v Bulgaria (Applications Nos 2033/04, 19125/04, 19475/04, 19490/04, 19495/04, 19497/04, 24729/04, 171/05 and 2041/05) (unreported) given 25 October 2011 at para 85, a cap may be regarded either as a provision limiting the amount of benefit after it has been calculated under the general rules, and thus an interference with a “possession” of the appellants, or as part of the overall set of statutory rules governing the manner in which the amount of benefit should be calculated, and thus as amounting to a rule preventing the appellants from having any “possession” in relation to the surplus. It is however unnecessary to resolve that question in the present appeal, since the applicability of A1P1 has not been contested on behalf of the Secretary of State.

The next question is whether the Regulations result in differential treatment of men and women. This is conceded on behalf of the Secretary of State. Given the statistics as to the proportion of those affected who are single women as compared with the proportion who are single men, that concession is understandable. It is indeed almost inevitable that a measure capping the benefits received by non-working households will mainly affect households with children, since they comprise the great majority of households receiving the highest levels of benefits. It follows inexorably that such a measure will have a greater impact on women than men, since the majority of non-working households with children are single parent households, and the great majority of single parents are women. That consequence could be avoided only by defining “welfare benefits” so as to exclude benefits which are directly or indirectly linked to responsibility for children, a possibility to which it will be necessary to return.

On the other hand, the argument that the Regulations also result in differential treatment of women because of their effect upon the victims of domestic violence has not in my opinion been established. In so far as the argument is based upon the failure of the Regulations, as originally made, to exempt housing benefit received in connection with all women’s refuges, the amendments effected by the 2014 Regulations were designed to address that problem, and it is not argued in these appeals that they have failed to do so. In so far as the argument was that women fleeing domestic violence may live in temporary accommodation rather than refuges, and may then be entitled to housing benefit in respect of both their original home and the temporary accommodation, that problem, which is inherently of a temporary nature, is capable of being addressed under the DHP Regulations by the use of discretionary housing payments; and the funding made available by Government for such payments has been increased for that very purpose. As I have explained, guidance has been issued by the Government to local authorities advising them that the funding is specifically aimed at groups including individuals or families fleeing domestic violence, and that payments can be awarded for two homes when someone is temporarily absent from their main home because of domestic violence. It cannot therefore be said that the Regulations have a disparate impact upon victims of domestic violence. Whether problems are avoided in practice will depend upon how the discretionary payments scheme is operated by local authorities in individual cases. It is not suggested that any problems have arisen in the cases with which these appeals are concerned.

The next question is whether the Regulations pursue a legitimate aim. In my view that cannot be doubted. They pursue, in the first place, the aim of securing the economic well-being of the country, as the Secretary of State explained to the Parliamentary Joint Committee on Human Rights, and as is evident from the legislative history since the policy of reducing expenditure on benefits was first announced in June 2010. A judgment was made, following the election of a new Government in May 2010, that the current level of expenditure on benefits was unaffordable. The imposition of a cap on benefits was one of many measures designed to reduce that expenditure, or at least to constrain its further growth. It was argued on behalf of the appellants that savings in public expenditure could never constitute a legitimate aim of measures which had a discriminatory effect, but that submission is inconsistent with the approach adopted by the European court in the cases mentioned in para 10. It is also inconsistent with the acceptance of the economic well-being of the country as a legitimate aim of interferences with Convention rights under the second paragraphs of articles 8 to 11, and under A1P1. An interpretation of the Convention which permitted the economic well-being of the country to constitute a legitimate aim in relation to interferences with the substantive Convention rights, but not as a legitimate aim in relation to the ancillary obligation to secure the enjoyment of those rights without discrimination, would lack coherence.

In relation to the case of Ministry of Justice (formerly Department for Constitutional Affairs) v O’Brien (Council of Immigration Judges intervening) [2013] UKSC 6; [2013] 1 WLR 522, para 69, on which the appellants relied, I would observe that acceptance that savings in public expenditure can constitute a legitimate aim for the purposes of article 14 does not entail that that aim will in itself constitute a justification for discriminatory treatment. As I have explained, the question whether a discriminatory measure is justifiable depends not only upon its having a legitimate aim but also upon there being a reasonable relationship of proportionality between the means employed and the aim sought to be realised.

The second aim, of incentivising work, is equally legitimate. It is, in the first place, an aspect of securing the economic well-being of the country. It has however a broader social objective which Ministers made clear to Parliament. That objective is based on the view that long-term unemployment is socially undesirable, because of its impact upon those affected by it (including the children brought up in non-working households), and that it is therefore important to make efforts to assist those capable of working to find work: efforts which can include the removal of financial disincentives.

The third aim, of imposing a reasonable limit upon the total amount which a household can receive in welfare benefits, is in my opinion equally legitimate. It is again an aspect of securing the economic well-being of the country: it is one of the means of achieving that objective. It also however has a broader aspect, namely to reflect a political view as to the nature of a fair and healthy society. As Ministers explained to Parliament, this objective responds in particular to a public perception that the benefits system has been excessively generous to some recipients: a perception which is related to the stigmatisation in the media of non-working households receiving high levels of benefit. The maintenance of public confidence in the welfare system, so that recipients are not stigmatised or resented, is undeniably a legitimate aim. In the language used by the European court in Hoogendijk and other cases, the benefit system is the means by which society expresses solidarity with its most vulnerable members. That being so, it is in principle legitimate to reform the system when necessary to respond to a threat to that solidarity.

The remaining question is whether the Regulations maintain a reasonable relationship of proportionality between the means employed and the aims sought to be realised.

It was argued by counsel for the appellants and interveners that the aim of setting a reasonable limit to the amount of benefits which a household can receive could have been achieved by using as a benchmark not the average earnings of working households but their average income inclusive of benefits. This would have been “fair”, adopting the adjective used by Ministers at some points during the Parliamentary debates, since it would have achieved parity between the maximum income received by non-working households and the average income of working households.

There are three problems with this argument. The first is that section 96 of the 2012 Act, whose compatibility with Convention rights is not challenged, requires the cap to be set by reference to “earnings”. The Regulations cannot be unlawful in so far as they follow that approach (Human Rights Act, section 6(2)(a)), and would be ultra vires if they failed to do so. Secondly, the assessment of the level at which a cap would represent a fair balance between the interests of working and non-working households is a matter of political judgment. Furthermore, the assumption that fairness requires an equivalence between the incomes of working and non-working households ignores the costs incurred by working households in earning that income: both financial costs in respect of such matters as travel and clothing, and non-financial costs in respect of the time spent commuting and working. As the Thlimmenos principle illustrates, non-discrimination does not require that different situations should be treated in the same way. Thirdly, and in any event, the Government has made a judgment, endorsed by Parliament, that a cap set at the level of the average income of working households would be less effective in achieving its aims. That is not an unreasonable judgment: plainly, the fiscal savings would be less, and the financial incentive to find work would be reduced. Indeed, if the cap were set at a level which achieved parity between the income of a person on benefits and the average income of a person in work, it would act as a disincentive to work for below average earnings. Whether the aim of securing a benefit system which was perceived by the public as fair and reasonable would also have been less effectively achieved is again a political judgment, which cannot be said to be manifestly unreasonable.

It was also argued that the short-term fiscal savings appear to be relatively marginal at best. It is true that the savings made are a small proportion of the total welfare budget, the bulk of which is spent on pensions. They nevertheless contribute towards the achievement of the objective of reducing the fiscal deficit. It is also necessary to bear in mind that the Regulations are designed to result in savings over the longer term, as the intended change in the welfare culture takes effect.

Other criticisms of the Regulations focused upon the impact of the cap upon the income of the households most severely affected, such as those of the appellants. Emphasis was placed in particular upon the inclusion of child benefit and child tax credit among the “welfare benefits”, and the difficulties which single parents faced with a loss of income might encounter in finding work, because of their child care responsibilities, or in moving to cheaper accommodation, because of the impact upon their children.

In relation to the reduction in income, it has to be borne in mind that the cap for a household with children has been set by Parliament at the median earnings of working households, equivalent to a salary of £35,000 per annum. By definition, half of all working households earn less than that amount. The exclusion of child benefit and child tax credit from the “welfare benefits” counting towards the cap would enable non-working households with children to receive an income from public funds in excess of that amount. Whether that level of benefits ought to be paid by the state is inherently a political question on which opinions within a democratic society may reasonably differ widely. It is not the function of the courts to determine how much public expenditure should be devoted to welfare benefits. It is also important to recognise that the households affected were given advance notice of the reduction in their income, and that assistance was made available to them to enable them to address the implications, as I have explained.

In relation to the related criticism that children in households affected by the cap are deprived of the basic necessities of life, that argument was rejected by the courts below, and I see no basis for reaching a different conclusion. As I have explained, the cap for a household with children is equivalent to a gross salary of £35,000 per annum, higher than the earnings of half the working population in the UK, almost three times the national minimum wage, and not far below the point at which higher rate tax becomes payable (in 2013/14, a salary of £41,450). Although the compatibility of the Regulations with article 14 does not depend on the individual circumstances of the appellants, as I have explained, the Court of Appeal considered in detail submissions to the effect that the cap would reduce them to a state of destitution, and concluded that their circumstances did not approach that level. The Divisional Court noted that even in cases where the cap had particularly adverse consequences, in the last resort the local authority was under a duty to secure suitable and affordable accommodation for the family.

In relation to the difficulties of finding work, data from the Office for National Statistics (ONS) indicate that 63.4% of single parents with dependent children were in work during the second quarter of 2014. An ONS analysis based on data for 2012 indicated that the employment rate for single parents with a dependent child under the age of 2 was 32%; for the age range 2-4 it increased to 42%; for the age range 5-11 it was 63%. Plainly, many single parents, including those on low incomes, make arrangements for the care of children in order to work. Their children over five years of age are required to attend school. Their younger children may attend nurseries or may be looked after by family members or child minders. The amount of work which a single parent has to perform, in order to be exempted from the cap, is only 16 hours per week. Even those hours need not necessarily be worked throughout the year: if a person works in a place of employment which has a recognisable cycle of employment, such as a school, the holiday periods during which she does not work are disregarded. As I have explained, assistance with meeting the cost of child care is available and is excluded from the cap. The statistics set out at paras 56 and 57 above do not support the contention that single parents with children under five have experienced greater difficulty in obtaining work than other claimants affected by the cap. Some people take the view that it is better for the single parent of a young child to remain at home full-time with the child, but there is no basis for requiring that view to be adopted by Government as a matter of law.

In relation to the argument that households with children cannot reasonably be expected to move house, because of the impact on the children, it is not merely a forensic point that one of the two adult appellants came with her family to the UK from Belgium, and that the other adult appellant came with her family to the UK from Algeria. Millions of parents in this country have moved house with their children, for a variety of reasons, including economic reasons. It is, in particular, not uncommon for working households to move out of London in order to find more affordable property elsewhere. It is also necessary to recognise that transitional financial assistance is available for households affected by the cap who cannot move until suitable arrangements have been made in relation to the children, as I have explained. Although assistance of that nature may not constitute a complete or satisfactory answer to a structural problem of a permanent nature arising from discriminatory legislation, such as the inadequacy of housing benefit to meet the cost of accommodation suitable for the needs of severely disabled claimants (as was held in Burnip v Birmingham City Council [2012] EWCA Civ 629; [2013] PTSR 117), it is relevant to an assessment of the proportionality of a measure which is liable to give rise to transitional difficulties in individual cases.

As I have explained, the court is concerned in a case of this kind with the question whether the legislation as such unlawfully discriminates between men and women, rather than with the hardship which might result from the cap in the cases of those most severely affected. In that regard, it is highly significant that no credible means was suggested in argument by which the legitimate aims of the Regulations might have been achieved without affecting a greater number of women than men. Put shortly, since women head most of the households at which those aims are directed, it appears that a disparity between the numbers of men and women affected was inevitable if the legitimate aims were to be achieved.

The greater number of women affected results from the inclusion of child-related benefits within the scope of the cap. If those benefits had been excluded from the cap, the legitimate aims of the cap would not have been achieved, as Ministers made plain to Parliament. The question is raised by Lady Hale whether taking child-related benefits out of the cap as it applies to single parents only would have an emasculating effect. I do not recall this point being raised with counsel for the Secretary of State, but the information available enables it to be considered. Parliament was informed that the exclusion of child-related benefits would reduce the savings, and the number of households affected by the cap, by 80 to 90% (para 41). According to the most recent statistics available at the time of the hearing, single parent households form 62% of the affected households receiving child-related benefits (para 56). It is therefore plain that the exclusion of child-related benefits, even if confined to single parent households, would have compromised the achievement of the legitimate aims of the Regulations.

An argument of a different character was put forward on the basis of article 3(1) of the United Nations Convention on the Rights of the Child (“UNCRC”), which provides that “in all actions concerning children … the best interests of the child shall be a primary consideration”. The argument developed during and after the hearing of the appeal. Initially, it was contended that the Secretary of State was obliged by section 6 of the Human Rights Act to treat the best interests of children as a primary consideration when making the Regulations, in accordance with article 3(1) of the UNCRC, since the cap had an impact upon the private and family lives of children forming part of the households affected. Article 8(1) of the ECHR was therefore applicable. Since the European court would have regard to the UNCRC when applying article 8 in relation to children, it followed that the Secretary of State was also obliged to comply with article 3(1) of the UNCRC, but failed to do so.

This argument raises a number of questions. In the first place, there is the question whether general legislation which limits welfare benefits, resulting in some cases in a reduction in household income, constitutes, by reason of the impact of that reduction in income upon the lives and circumstances of those affected, an interference with their right to respect for their private and family life. If it does, the ambit of article 8 is enlarged beyond current understanding so as to embrace legislation imposing increases in taxation or reductions in social security benefits. Secondly, on the assumption that such legislation falls within the ambit of article 8(1), article 8(2) permits an interference with the right to respect for family life to be justified as being necessary in a democratic society in the interests of the economic well-being of the country. The argument that justification on that ground is impossible unless the best interests of the children affected by the measure in question have been treated as a primary consideration - not only in the sense that they have been taken into account but, as counsel emphasised, in the sense that the legislation is in reality in the best interests of the children affected by it – has major implications for the effect of the ECHR in relation to legislation in the field of taxation and social security.

These issues were not addressed in the course of the argument. Most of the European authorities cited in support were concerned with the different question of the eviction of individuals from their homes, which is not an issue arising on the facts of the present cases. The cases indicate that a reduction in income may have consequences which are such as to engage article 8, as for example where non-payment of rent leads to the threat of eviction from one’s home, but they do not indicate that the reduction in income is itself within the ambit of article 8. The only other European authority cited was the case of Neulinger v Switzerland (2010) 54 EHRR 1087, which was concerned with the return of a child under a child abduction convention. It is unnecessary to say more than that the argument has not been made out.

A more closely reasoned argument has been developed in submissions lodged after the hearing, which treats article 3(1) of the UNCRC as forming part of the proportionality assessment under article 14 of the ECHR read with A1P1. In consequence, a test of compliance with article 3(1) is effectively substituted for the “manifestly without reasonable foundation” test which all parties agree to be applicable in the present context. On that basis, article 3(1) is argued to be decisive of the appeals. It is therefore necessary to consider carefully how, if at all, article 3(1) bears on the issues in these appeals.

As an unincorporated international treaty, the UNCRC is not part of the law of the United Kingdom (nor, it is scarcely necessary to add, are the comments upon it of the UN Committee on the Rights of the Child). “The spirit, if not the precise language”, of article 3(1) has been translated into our law in particular contexts through section 11(2) of the Children Act 2004 and section 55 of the Citizenship, Borders and Immigration Act 2009: ZH (Tanzania) v Secretary of State for the Home Department [2011] UKSC 4; [2011] 2 AC 166, para 23. The present case is not however concerned with such a context.

The UNCRC has also been taken into account by the European Court of Human Rights in the interpretation of the ECHR, in accordance with article 31 of the Vienna Convention on the Law of Treaties. As the Grand Chamber stated in Demir v Turkey (2008) 48 EHRR 1272, para 69, “the precise obligations that the substantive obligations of the Convention impose on contracting states may be interpreted, first, in the light of relevant international treaties that are applicable in the particular sphere”. It is not in dispute that the Convention rights protected in our domestic law by the Human Rights Act can also be interpreted in the light of international treaties, such as the UNCRC, that are applicable in the particular sphere.

The approach adopted is illustrated by V v United Kingdom (1999) 30 EHRR 121, where the European court had regard to articles 37 and 40 of the UNCRC when considering how the prohibition of inhuman and degrading treatment in article 3 of the ECHR applied to the trial and sentencing of child offenders, and, in a domestic context, by R (T) v Chief Constable of Greater Manchester Police [2014] UKSC 35; [2015] AC 49, where this court referred to article 40 of the UNCRC when considering whether legislation regulating the disclosure of offences committed by children was compatible with article 8 of the ECHR.

The case of X v Austria (2013) 57 EHRR 405, on which the appellants and the interveners principally rely, concerned the proposed adoption of a child by the female partner of the child’s biological mother. The effect of adoption under Austrian law was to sever the legal relationship between the child and the biological parent of the same sex as the adoptive parent. In consequence, therefore, Austrian law could not recognise a legal relationship between a child, an adoptive parent, and a biological parent of the same sex as the adoptive parent. An application to the European court was brought by the child, the mother, and her partner, all of whom lived together as a family, on the basis that they had been denied legal recognition of their family life by reason of the sexual orientation of the two adults, in violation of article 14 of the ECHR read together with article 8. The court considered their complaint on the basis that all three applicants enjoyed family life together, and all three were therefore entitled to complain of a violation of their rights. The effect of the Austrian law was to prevent second-parent adoption by same-sex couples. The justifications advanced were the protection of the family in the traditional sense, and the protection of the interests of children, both of which were legitimate aims. The question was whether the principle of proportionality was adhered to. In considering that question, the court identified a number of considerations which weighed in favour of allowing the courts to carry out an examination of each individual case, rather than imposing an absolute rule. The court added that this would also appear to be more in keeping with the best interests of the child, which was a key notion in the relevant international instruments. In that regard, the court had earlier referred to a number of provisions of the UNCRC, including article 3(1).

It is clear, therefore, that the UNCRC can be relevant to questions concerning the rights of children under the ECHR. There are also cases in which, although the court has not referred to the UNCRC, it has taken the best interests of children into account when considering whether an interference with their father’s or mother’s right to respect for their family life with the children was justified. An example is the case of Űner v Netherlands (2006) 45 EHRR 421, which concerned the deportation of an adult, resulting in his separation from his children. In circumstances of that kind, the proportionality of the interference with family life could not be assessed without consideration of the best interests of the children, a matter which was relevant to respect for his family life with them, as it was also to their right to respect for their family life with him. Indeed, they might themselves have been applicants, on the basis that their own article 8 rights were engaged.

The present context, on the other hand, is one of alleged discrimination between men and women in the enjoyment of the property rights guaranteed by A1P1. That is not a context in which the rights of the adults are inseparable from the best interests of their children. It is of course true that legislation limiting the total income which persons can receive from benefits, like any legislation affecting their income, may affect the resources available to them to provide for any children in their care, depending upon how they respond to the cap: something which will vary from one case to another. They may increase their income from other sources, for example by obtaining employment or by obtaining financial support for the upkeep of a child from an absent parent; or they may respond by reducing their expenditure, for example by moving to cheaper accommodation. Depending on how parents respond, the consequences of the cap for their children may vary greatly, and may be regarded as positive in some cases and as negative in others.

The questions (1) whether legislation of this nature should be regarded as “action concerning children”, within the meaning of article 3(1) of the UNCRC, (2) whether that provision requires such legislation to be in the best interests of all the children affected by it, and (3) whether the Regulations fulfil that requirement, appear to me to be questions which, for reasons I shall explain, it is unnecessary for this court to decide. Even on the assumption, however, (1) that article 3(1) of the UNCRC applies to general legislation of this character, (2) that article 3(1) requires such legislation to be in the best interests of all the children indirectly affected by it, and (3) that the legislation in question is not in reality in the best interests of all the children indirectly affected by it, that does not appear to me to provide an answer to the question whether the legislation unjustifiably discriminates between men and women in relation to their enjoyment of the property rights guaranteed by A1P1.

It is true that the benefits which are taken into account when deciding whether the cap has been exceeded include benefits payable to parents by reason of their responsibility for the care of children. It is also true that the differential impact of the measure upon men and women arises from the fact that more women than men take on responsibility for the care of their children when they separate. It is argued that it is therefore unrealistic to distinguish between the rights of women under article 14 read with A1P1, and those of their children under the UNCRC. There is nevertheless a clear distinction. In cases where the cap results in a reduction in the resources available to parents to provide for children in their care, the impact of that reduction upon a child living with a single father is the same as the impact on a child living with a single mother in similar circumstances, or for that matter a child living with both parents. The fact that children are statistically more likely to be living with a single mother than with a single father is unrelated to the question whether the children’s rights under article 3(1) of the UNCRC have been violated. There is no factual or legal relationship between the fact that the cap affects more women than men, on the one hand, and the (assumed) failure of the legislation to give primacy to the best interests of children, on the other. The conclusion that the cap is incompatible with the UNCRC rights of the children affected therefore tells one nothing about whether the fact that it affects more women than men is unjustifiable under article 14 of the ECHR read with A1P1. The contrary view focuses on the question whether the impact of the legislation on children can be justified under article 3(1) of the UNCRC, rather than on the question whether the differential impact of the legislation on men and women can be justified under article 14 read with A1P1, and having concluded that the legislation violates article 3(1) of the UNCRC, mistakenly infers that the difference in the impact on men and women cannot therefore be justified.

Nor is the argument made stronger by being recast in terms of domestic administrative law, on the basis that the decision to make the Regulations was vitiated by an error of law as to the interpretation of article 3(1) of the UNCRC. It is firmly established that UK courts have no jurisdiction to interpret or apply unincorporated international treaties: see, for example, J H Rayner (Mincing Lane) Ltd v Department of Trade and Industry [1990] 2 AC 418, 499; R v Lyons (Isidore) [2002] UKHL 44; [2003] 1 AC 976, para 27. As was made clear in R (Corner House Research) v Director of the Serious Fraud Office (JUSTICE Intervening) [2008] UKHL 60; [2009] AC 756, it is therefore inappropriate for the courts to purport to decide whether or not the Executive has correctly understood an unincorporated treaty obligation. As Lord Bingham of Cornhill said at para 44:

The case of R v Secretary of State for the Home Department, Ex p Launder [1997] 1 WLR 839, on which reliance is placed, is distinguishable from the present case on the same basis as it was distinguished in the Corner House Research case. In the first place, as Lord Bingham pointed out (para 44), there was in Launder no issue between the parties about the interpretation of the relevant articles of the Convention, whereas in Corner House, as in the present case, the court was being asked to determine, in the absence of any international judicial authority, the meaning of a provision of an unincorporated international treaty. Secondly, as Lord Brown noted (para 66), Launder was a case in which it was plain that the decision-maker would have taken a different decision had his understanding of the treaty been different: his clear intention was to act consistently with the United Kingdom’s international obligations, whatever decision that would have involved him in taking. In Corner House, on the other hand, the primary intention behind the decision was to save this country from a threat which it faced, and all that the Ministers were really saying was that they believed the decision to be consistent with the international obligation in question.

Finally, it has been explained many times that the Human Rights Act entails some adjustment of the respective constitutional roles of the courts, the executive and the legislature, but does not eliminate the differences between them: differences, for example, in relation to their composition, their expertise, their accountability and their legitimacy. It therefore does not alter the fact that certain matters are by their nature more suitable for determination by Government or Parliament than by the courts. In so far as matters of that nature have to be considered by the courts when deciding whether executive action or legislation is compatible with Convention rights, that is something which the courts can and do properly take into account, by giving weight to the determination of those matters by the primary decision-maker.

That consideration is relevant to these appeals, since the question of proportionality involves controversial issues of social and economic policy, with major implications for public expenditure. The determination of those issues is pre-eminently the function of democratically elected institutions. It is therefore necessary for the court to give due weight to the considered assessment made by those institutions. Unless manifestly without reasonable foundation, their assessment should be respected.

As I have explained, the Regulations were considered and approved by affirmative resolution of both Houses of Parliament. As Lord Sumption observed in Bank Mellat v H M Treasury (No 2) [2013] UKSC 39; [2014] AC 700, para 44:

Many of the issues discussed in this appeal were considered by Parliament prior to its approving the Regulations. That is a matter to which this court can properly have regard, as has been recognised in such cases as R (Williamson) v Secretary of State for Education and Employment [2005] UKHL 15; [2005] 2 AC 246, R (Hooper) v Secretary of State for Work and Pensions [2005] UKHL 29; [2005] 1 WLR 1681, R (Countryside Alliance) v Attorney General [2007] UKHL 52; [2008] AC 719, and R (Animal Defenders International) v Secretary of State for Culture, Media and Sport [2008] UKHL 15; [2008] AC 1312. Furthermore, that consideration followed detailed consideration of clause 93 of the Bill, which became section 96 of the 2012 Act. It is true that the details of the cap scheme were not contained in the Bill which Parliament was debating, but the Government’s proposals had been made clear, they were challenged by means of proposed amendments to the Bill, and they were the subject of full and intense democratic debate. That is an important consideration. As Lord Bingham of Cornhill observed in R (Countryside Alliance) v Attorney General [2007] UKHL 52; [2008] AC 719, para 45:

Giving due weight to the assessment of the Government and Parliament, I am not persuaded that the Regulations are incompatible with article 14. The fact that they affect a greater number of women than men has been shown to have an objective and reasonable justification. No-one has been able to suggest an alternative which would have avoided that differential impact without compromising the achievement of the Government’s legitimate aims. Put shortly, it was inevitable that measures aimed at limiting public expenditure on welfare benefits, addressing the perception that some of the out-of-work were receiving benefits which were excessive when compared with the earnings of those in work, and incentivising the out-of-work to find employment, would have a differential impact on women as compared with men. That followed from the fact that women formed the majority of those who were out of work and receiving high levels of benefit. The Government’s considered view, endorsed by Parliament, that the achievement of those aims was sufficiently important to justify the making of the Regulations, notwithstanding their differential impact on men and women, was not manifestly without reasonable foundation. I would accordingly dismiss the appeals.

Others have explained the factual and legal background of these appeals. The following issues were agreed between the parties for consideration by the Supreme Court:

Was the Court of Appeal wrong to have declined to decide whether the benefit cap, as formulated in the 2012 Regulations, had an unlawfully disproportionate impact on victims of domestic violence?

Was the Court of Appeal wrong not to have found that the disproportionate effect of the 2012 Regulations on victims of domestic violence was contrary to article 14 ECHR (read with article 8 and/or article 1 of Protocol 1) and unlawful?

Was the Court of Appeal wrong to have found that the discriminatory effects of the 2012 Regulations on lone parents were justified and lawful, and not contrary to article 14 (read with article 8 and/or article 1 of Protocol 1)?

Was the Court of Appeal wrong to have found that the Respondent has complied with his obligation to treat the best interests of children as a primary consideration when implementing the benefit cap scheme?

The boundaries between these heads of claim have not been very clearly delineated in the arguments before us. However, in agreement with both Baroness Hale and Lord Reed, I find it most helpful to concentrate on issues (iii) and (iv), with specific regard to article 1 of protocol 1 (“A1P1”). Like them I do not think that a case has been made, at least on the evidence before us, for separate treatment of the position of victims of domestic violence, the subject of issues (i) and (ii). Under issue (iii) it is common ground that the scheme falls within the ambit of A1P1, and that in the context of article 14 it is indirectly discriminatory against women, particularly lone parents. The only issue therefore is justification.

Article 8 was also mentioned under issue (iii), and was relied on by Mr Wise in his printed case. However, as I understood it, this was not by way of challenge to the Court of Appeal’s rejection of the “free-standing” claim under article 8, which is consequently not one of the agreed issues for this court. Rather he relied on article 8 either as an alternative route into article 14, or as supporting his “best interest” claim under issue (iv). I note that article 8 was not relied on by Mr Drabble QC for the Child Poverty Action Group. I have not been persuaded that either of Mr Wise’s formulations adds anything of substance to the claim based on A1P1.

It is important also to understand how the interests of children affected by the scheme may be relevant to the legal analysis, either under the Convention itself, or indirectly by reference to article 3(1) of the UNCRC (best interests of children as “a primary consideration”). As to the Convention, the children have no relevant possessions under A1P1 in their own right; nor are they a protected class under article 14. However, as Lady Hale has said (para 218), the disproportionate impact on women arises because they are responsible for the care of dependent children. Elias LJ said in the Divisional Court (para 62):

The possible relevance of UNCRC article 3(1) requires a little more explanation. Before the Divisional Court (para 45) Mr Eadie QC was recorded as having submitted on behalf of the Secretary of State that, as “an international instrument with no binding effect in English law”, the Convention had no bearing on the case. This argument was rejected by Elias LJ and has not been renewed. The Court of Appeal said:

It is unnecessary to repeat the accounts given in other judgments of the nature of the discrimination, of the three-fold justification put forward by the Secretary of State, and of the criticisms made of it by the appellants, supported by the interveners. In short, it is said, the two objectives of fairness and increasing incentives to work are largely irrelevant or misconceived in their application to the group which is the object of discrimination; and that the third, saving money, cannot on its own justify discriminatory treatment in the enjoyment of a convention right. The essential objection was put shortly by Mr Drabble for the Child Poverty Action Group:

The Court of Appeal, in agreement with the Divisional Court, rejected these criticisms, holding in particular that there had been compliance with article 3(1) (para 72ff). Applying the approach of members of this court in H (H) v Italian Prosecutor [2013] 1AC 338, they held that it was not necessary for the decision-maker to adopt a “tightly structured” approach to consideration of the issues raised by article 3(1). It was enough for him to “give appropriate weight to the interests of children as a primary consideration in the overall balancing exercise”.

They found “ample evidence” that the Secretary of State had satisfied this test, citing five matters (para 74):

The 2010 Treasury Spending Review made clear that a principal objective was “to raise children out of long term poverty”;

The February 2011 Impact Assessment showed that the government was “keenly aware” of the likely impact on children;

The March 2011 Equality Impact Assessment stressed the objective of reversing the detrimental impact on families and children of benefits dependency, and indicated that the government was looking at ways to ease the transition for large families;

The Parliamentary debates “focussed time and again” on the interests of children; and

The July 2012 Impact Assessment revised the assessment of the number of children likely to be affected and addressed the issue of short term relief.

The comments in this court in H (H) predated, and therefore did not take account of, the most authoritative guidance now available on the effect of article 3(1). This is in “General Comment No 14”, adopted by the UN Committee on the Rights of the Child early in 2013. Although this guidance was not available at the time of the decisions under challenge, it is as I understand it intended as a restatement of established practice, rather than a new departure.

Paragraph 6 explains that “best interests” in this context is a “three-fold concept”: (a) a substantive right, (b) a fundamental, interpretative legal principle, and (c) a rule of procedure. The first and third are explained as follows:

Later paragraphs explain that the phrase “actions concerning children” is to be read in a “very broad sense” covering actions including children and other population groups, such as those relating to housing (para 19); that where a decision will have a major impact on children “a greater level of protection and detailed procedures to consider their best interests (are) appropriate (para 20)”; and that the child’s interests “have high priority and (are) not just one of several considerations … larger weight must be attached to what serves the child best” (para 39).

In relying on this guidance, Mr Wise accepted that it was not necessary for the decision-maker to address the issues in a “particular structured order”, as the Court of Appeal may have understood his argument. What matters is the substance of what is done rather than the form. However those passages do show in my view that the evaluation needs to consider, where relevant, the interests both of children in general and of those directly affected by the action. It also needs to indicate the criteria by which the “high priority” given to children’s interests has been weighed against other considerations. In so far as that evaluation shows conflict with the best interests of the children affected, it needs either to demonstrate how that conflict will be addressed, or alternatively what other considerations of equal or greater priority justify overriding those interests.

Accordingly, as the submissions and evidence stood at the end of the hearing, my view was that, judged by those criteria, the matters relied on by the Court of Appeal fell well short of establishing compliance. The Treasury’s long-term objective of taking children out of poverty, laudable in itself, was no substitute for an evaluation of the particular impact on the children immediately and directly concerned, and their parents. The February 2011 Impact Assessment and the March 2011 Equality Impact Assessment may have shown that the government was “keenly aware” of the likely impact on children, and was “looking at ways to ease the transition”, but they did not provide the answers. In any event, those assessments were related to the statute rather than the regulations which are now under challenge.

Those assessments also predated the report by the Children’s Commissioner in January 2012, which set out a number of “likely outcomes” of concern to the Commissioner. They included increase in child poverty (including diversion to housing costs of money which would otherwise have been spent on “necessities for children’s health and wellbeing”), children losing their homes, incentivising family breakdown, and disproportionate impact on children from some BME groups. The Commissioner expressed the view that “the universal imposition of the cap without regard to the individual circumstances of children” would conflict with the best interests principle under UNCRC article 3(1). This view had special significance, as that of the authority responsible under the Children Act 2004 for advising the Secretary of State on the interests of children.

The subsequent Equality Impact Assessment of July 2012, prepared by the Department in support of the regulations, did indeed make some revisions to the earlier figures, and mentioned the short-term relief to be provided by discretionary housing payments. But it did not in terms respond to the more fundamental points of concern raised by the Commissioner’s report. In his evidence for the Secretary of State, Mr Holmes observed simply that the government did not agree with the Commissioner’s assessment, but without further detail. The July assessment also indicated that there would in due course be a “full evaluation” of the operation of the benefit cap, to be published in autumn 2014. (We have not been given any information relating to this exercise, nor has it been suggested that it is relevant to our consideration of the legal issues relating to the decisions under challenge.)

For these reasons, my provisional view at the end of the hearing was that, in their application to lone parents and their dependent children, the regulations were not compatible with Convention rights, and that the court should so declare.

In post-hearing submissions permitted by the court, the point was taken on behalf of the Secretary of State that A1P1 (with or without article 14) was not the context in which article 3(1) UNCRC had hitherto been relied on by the appellants. I observe that this limitation is not apparent from the agreed wording of question (iv). Nor it seems was the discussion in the courts below so limited. Lord Dyson’s reference to this argument (paras 67-75), and to its treatment by the Divisional Court, came immediately after his discussion of “article 14 (read with A1P1)”; he observed that the argument had “featured prominently” in Mr Wise’s submissions on justification “in relation to article 14 (as well as in relation to article 8 which we deal with below”. It is fair to say however that at the hearing Mr Wise’s submissions in that connection were directed mainly to article 8. For this reason, and because of the importance of the issue for this case and others, counsel for the Secretary of State were given the opportunity to make further written submissions.

They summarised their submissions in the following six points:

Article 3(1) of the UNCRC is a provision of an unincorporated treaty which may only be relied on to the extent that it has been transposed into domestic law;

The ECtHR uses international law when determining the meaning of provisions of the ECHR, in accordance with the Vienna Convention on the Interpretation of Treaties;

Article 3(1) of the UNCRC is, as a matter of principle and in accordance with Strasbourg authority, not relevant to the question of justification of discrimination under article 14 read with A1P1. It has no role to play in determining the meaning of article 14 (read with A1P1 or otherwise), and does not inform or illuminate the question whether the differential impact on women of the benefit cap is proportionate;

Article 3(1) of the UNCRC does not supplant, dilute or compromise the Stec test which all parties have agreed, at every stage of these proceedings, applies both when considering whether the aims are legitimate and when determining whether the 2012 Regulations, having regard to their differential impact on women, are proportionate;

Even if the Court were to consider it foreseeable that the ECtHR may develop its case-law to have the effect that a breach of article 3(1) of the UNCRC renders legislation disproportionate, there are strong constitutional reasons why the Court should refrain from going beyond the current Strasbourg jurisprudence; and

In any event, the 2012 Regulations do not breach article 3(1) of the UNCRC. The Secretary of State fully took into account the best interests of children, as a primary consideration, and these were extensively debated in Parliament.

I have little difficulty with points (i), (ii), (iv) and (v). There has been no dispute as to the application of the Stec test to the issue of proportionality (iv), and no one has argued that we should go beyond existing ECHR jurisprudence (v). As to (i) it is of course trite law that, in this country at least, an international treaty has no direct effect unless and until incorporated by statute, but that it may be taken into account as an aid to interpretation in cases of ambiguity. To that extent the present case is to be contrasted with cases such as ZH (Tanzania) v Secretary of State for the Home Department [2011] 2 AC 166, in which as Lady Hale explained (para 23), UNCRC article 3(1) was reflected in the relevant statutory provisions. Ministerial statements of the government’s “commitment” to giving “due consideration” to the UNCRC articles (see Lady Hale para 214), may have political consequences but are no substitute for statutory incorporation.

It is equally clear (ii) that, under the ECHR and in accordance with the Vienna Convention, regard may be had to principles of international law, including international conventions, for the purpose of interpreting the “terms and notions in the text of the Convention”: see Demir v Turkey (2008) 48 EHRR 1272, paras 65, 67, 85. Demir itself is a good illustration of that proposition. For the purpose of determining whether article 11 (right to join a trade union) extended to civil servants, reference was made to article 22 of the International Covenant on Civil and Political Rights. It was noted by the court (para 99) that the wording of that article was similar to that of article 11 of the Convention, but that it was expressed to be subject to the right of the state to exclude the armed forces and the police, without referring to members of the administration of the state. Similarly, in Neulinger v Switzerland, to which Elias LJ referred, the court had regard to the Hague Convention on the Civil Aspects of International Child Abduction in determining whether forced return of a child to Israel would involve a breach of his rights under article 8 of the Convention.

Point (iii) questions the application of this approach in the context of article 14 taken with A1P1, and more specifically to the issue of justification. There seems to be no reason in principle why the Demir approach should not apply to article 14. Mr Drabble relies on X v Austria (2013) 57 EHRR 405, as the “clearest” example, in that case relating to article 14 taken with article 8. The court held that a law preventing second parent adoption in the case of same-sex marriages involved discrimination under 14, and, although the law served a legitimate aim, it had not been shown that an absolute prohibition was necessary for the protection of the families or children. Early in its judgment (para 49) the court had quoted UNCRC article 3, and also article 21 which requires that systems of adoption shall “ensure that the best interests of the child shall be the paramount consideration”. In considering the question of justification, the court listed the factors which seemed rather to “weigh in favour of allowing the courts to carry out an examination of each individual case” adding (with a reference to the earlier quotations):

Another Strasbourg case in which reliance was placed on the UNCRC as an aid to interpretation of the Convention, in this case in favour of the state, was Ponomaryov v Bulgaria (2011) 59 EHRR 799. The complaint was of a violation of article 14 taken with A2/P1 (right to education), by direct discrimination on the grounds of nationality with respect to the provision of secondary education. In dismissing the application, the court relied on UNCRC article 28 as supporting the view that the state enjoyed a greater margin of appreciation in relation to secondary as compared to primary education (para 57).

There are examples also in domestic jurisprudence. Lady Hale has referred to the judgment of Maurice Kay LJ in Burnip v Birmingham City Council [2013] PTSR 117, concerning discrimination in the application of housing benefit for a disabled person. Although the court was able to arrive at its decision on other grounds, Maurice Kay LJ would have relied if necessary on the UN Convention on the Rights of Persons with Disabilities (“CRPD”) to resolve any uncertainty over “the meaning of article 14 discrimination” in the circumstances of the case (para 22). Of this case Mr Sheldon comments:

I see no inconsistency between such reference to international treaties where relevant and the Stec test. In Burnip. Henderson J, giving the lead judgment, cited the passage in Stec which established the “manifestly without reasonable foundation” test as appropriate for review of “general measures of economic or social strategy”, and declined to adopt an “enhanced” test requiring “very weighty reasons” for the discrimination. It was in this context that Maurice Kay LJ, who agreed with Henderson J on the issue of justification (para 23), drew assistance from the CRPD.

Before considering the application of that approach to the present case, it is convenient to consider point (vi), that is whether the latest submissions throw any further light on the issue whether the regulations were in compliance with article 3(1).

It is not in dispute that, as asserted, issues in relation to the interests of children “were extensively debated in Parliament” or that the views so expressed were taken into account by Ministers. But article 3(1) is more than a restatement of the ordinary administrative law duty to have regard to material circumstances. The principles were summarised by Lord Hodge in Zoumbas v Secretary of State for the Home Department (AF (A Child) intervening) [2013] 1 WLR 3690 (paras 10-13) in seven points. I would emphasise the first and last: (1) “The best interests of a child are an integral part of the proportionality assessment under article 8 ECHR” …; (7) “A child must not be blamed for matters for which he or she is not responsible, such as the conduct of a parent”. On the other hand, as he added (by reference to H (H)) there may be circumstances in which “the weight of another primary consideration can tip the balance and make the interference proportionate even where it has very severe consequences for children” (para 13).

In considering how the government approached that task, rather than trawling through the parliamentary debates, we are entitled to rely on the evidence given in these proceedings on behalf of the Secretary of State. The Court of Appeal quoted (paras 32-33) the evidence of Mr Holmes that:

The difficulty with that response, in the context of a duty to treat the best interests of the child as a “primary consideration” is that it ignores the distinctive statutory purpose of the child related benefits. Lord Reed (para 35) refers to a ministerial response in the course of the Parliamentary debate, to the effect that working people on low incomes had to “cope with difficult circumstances” and “live within their means”; that their earnings were “not determined by the size of their families”, and that “the government believed that the same principle should apply to the level of the cap”.

As applied to child related benefits, in my view, this was a false comparison. No doubt for that reason it was not a point made by Mr Holmes. The benefits are paid regardless of whether their parents are in work or not. In this respect therefore workers and non-workers alike were (before the cap) able to rely on this extra assistance in “coping with difficult circumstances” in the interests of their children. Although paid to the parents, these benefits are designed to meet the needs of children considered as individuals. As Lady Hale said in Humphreys v Revenue and Customs Comrs (summarising the case for the Revenue):

As Mr Drabble QC submitted, the cap was a complete innovation in the combined benefits/tax system, which had always contained a mechanism to adjust for family size. The cap has the effect that for the first time some children will lose these benefits, for reasons which have nothing to do with their own needs, but are related solely to the circumstances of their parents. It is difficult to see how this result can be said to be consistent with the best interests of the children concerned, or in particular with the first and seventh principles in Zoumbas.

Lord Reed has referred to statements made to Parliament in November 2011 that excluding both child benefit and child tax credit would reduce the savings from the scheme by 80-90%, and so emasculate the scheme. It is not clear whether these are up-to-date estimates, or how they relate to the regulations as opposed to the Bill. If correct, they raise the questions why the viability of a scheme, whose avowed purpose is directed at the parents not their children, is so disproportionately dependent on child related benefits. There is nothing in Mr Holmes’ evidence which addresses or answers these questions.

Accordingly I remain of the view that the Secretary of State has failed to show how the regulations are compatible with his obligation to treat the best interests of children as a primary consideration.

The more difficult question, now that it has been put in issue, is how that finding in relation to the interests of children under UNCRC article 3(1) affects the resolution of issue (iii): that is the alleged justification for the admittedly discriminatory effects on women as lone parents. As Mr Sheldon submits, even if article 3(1) had a role to play in illuminating article 14, this could only be where the alleged indirect discrimination, or differential treatment, was in respect of children. In the present case, by contrast, the allegation is of discrimination, not against children, but against their mothers. The children, it is said, will be treated the same whether their lone parents are male or female. With considerable reluctance, on this issue agreeing with Lord Reed, I feel driven to the conclusion that he is right.

In all the article 14 cases to which we have been referred to in this context there was a direct link between the international treaty relied on and the particular discrimination alleged:-

In X v Austria (2013) 57 EHRR 405, where the complaint concerned discrimination by restrictions on adoption by single sex couples, the court referred not only to UNCRC article 3(1), but also to article 21 which applied the best interests principle specifically to adoption.

In Ponomaryov v Bulgaria (2011) 59 EHRR 799, where the complaint was of discrimination in respect of education, reference was made to UNCRC article 28 relating also to education.

In Burnip v Birmingham City Council [2013] PTSR 117, where the alleged discrimination related to the treatment of the disabled, reference was made to the CRPD, covering the same subject matter.

There is no such connection in the present case. The discrimination with which we are concerned under article 14 is in relation to women and their “possessions”. Those concepts require no relevant “illumination” by way of interpretation. It is true that the discrimination in this case is related to their responsibilities as lone parents, and to that extent, as Elias LJ accepted, the children are not “strangers to the article 14/A1P1 arguments”. But that is a comment on the facts, not on the interpretation of the convention rights. Indeed, as has been seen, it is the distinct interest of the children in the benefits as individuals that has reinforced my view of the breach under article 3(1). As Lord Reed says (para 89) the fact that children are statistically more likely to be living with a single mother than with a single father is unrelated to the question whether the childrens’ interests have been treated as a primary consideration as required by UNCRC article 3(1).

We have been shown no precedent in the Strasbourg jurisprudence for the use of an international treaty in this indirect way. Mr Sheldon argues that there are “strong constitutional reasons” why the court should not go beyond Strasbourg on an issue of this kind. Whether or not that is so, we have heard no argument that we should do so. The appellants and their supporters have relied simply on the principles to be extracted from the existing case-law.

In conclusion I would dismiss the appeal, albeit on grounds much narrower than those accepted by the courts below. I would hope that in the course of their review of the scheme, the government will address the implications of these findings in relation to article 3(1) itself. However, it is in the political, rather than the legal arena, that the consequences of that must be played out.

I agree with the judgment and conclusions of Lord Reed and would like him dismiss this appeal. I add only some additional observations in view of the difference of opinion which is disclosed by the judgments of Lady Hale and Lord Kerr.

There is much common ground.

The suggested discriminatory effect upon the victims of domestic violence adds nothing to the accepted discriminatory effect upon women. Moreover neither of the adult appellants is suffering any of the adverse effects of the cap relied upon as affecting such victims, so that the Court of Appeal was fully justified in declining to decide the issue of such victims. Further, the principal adverse effects peculiar to such victims which were relied upon (the treatment of refuges and the possible need for two sets of rent to be within housing benefit) have both been addressed by amendments to the original form of the Regulations.

It is agreed on all sides that the scheme has legitimate aims. At the very least, the principal aim of discouraging benefit dependence and encouraging work is agreed to be legitimate. For my part I agree that at a time of national economic crisis it was also legitimate to seek to reduce the overall expense on benefits, and that establishing a different balance between those who worked and paid taxes and those who did not was a further legitimate aim.

A1P1 to the First Protocol is agreed to be engaged to the extent that Stec v United Kingdom (2006) 43 EHRR 1017 establishes that, although it does not give an entitlement to benefits, the ECHR does require that if they are provided they must be administered in a manner which is not discriminatory contrary to article 14. Here a discriminatory effect of the regulations upon women is conceded, because they represent much the largest proportion of lone parents forming a household with children. Accordingly the scheme as a whole, including its discriminatory effect, must be justified. The test, in a case involving high level social/economic policy, is agreed by all parties to be that laid down in Stec, namely that it fails to be justified if it is manifestly without reasonable foundation.

The difference of opinion reduces itself to the place of article 3 of the UN Convention on the Rights of the Child (“UNCRC”). That in turn involves two questions:

does article 3 have legal effect in English law and if so by what route? and

if it does, has there been a breach of it such as to render the Regulations unlawful?

Article 3 UNCRC is contained in an international treaty ratified by the UK. It is binding on this country in international law. It is not, however, part of English law. Such a treaty may be relevant in English law in at least three ways. First, if the construction (ie meaning) of UK legislation is in doubt, the court may conclude that it should be construed, if otherwise possible, on the footing that this country meant to honour its international obligations. Second, international treaty obligations may guide the development of the common law. For these two propositions see for example R v Lyons (Isidore) [2002] UKHL 44; [2003] 1 AC 976, para 13. Neither has any application to this case. This case is concerned with legislation, not with the common law, and it is not suggested that there is any room for doubt about the meaning of the regulations. Thirdly, however, the UNCRC may be relevant in English law to the extent that it falls to the court to apply the European Convention on Human Rights (“ECHR”) via the Human Rights Act 1998. The European Court of Human Rights has sometimes accepted that the Convention should be interpreted, in appropriate cases, in the light of generally accepted international law in the same field, including multi-lateral treaties such as the UNCRC. An example is Demir v Turkey (2008) 48 EHRR 1272 which concerned the scope of article 11 (right of freedom of association), and which is cited by Lord Reed at para 83 above.

It was on this third basis that the UNCRC was advanced in argument before this court and, as I understand it, in the courts below. Until post-hearing submissions in this court, this argument was confined to praying in aid article 3 UNCRC upon the application or content of article 8 of the ECHR (respect for private and family life). In turn, the complaint of infringement of article 8 was based upon the rights of the children affected by the cap, not of their mothers except to the extent that they were, as carers, directly involved in the article 8 rights of their children. Article 3 UNCRC was not, until the post-hearing submissions, advanced as relevant to the justification of the admitted indirect discrimination against women in relation to their A1P1 rights.

For the reasons set out by the Court of Appeal, the article 8 rights of children are not arguably infringed by the benefit cap scheme. Elastic as that article has undoubtedly proved, it does not extend to requiring the State to provide benefits, still less benefits calculated simply according to need, nor does it require the state to provide a home. See Chapman v United Kingdom (2001) 33 EHRR 399, para 99; R (TG) v Lambeth London Borough Council (Shelter intervening) [2012] PTSR 364, paras 34 and 40; AM v Secretary of State for Work and Pensions [2014] EWCA Civ 286, para 22 and the cases there cited. Winterstein v France [2013] ECHR 984 depended upon the long toleration of itinerants on the land from which they were evicted and the absence of provision of alternative accommodation, and does not lead to a different conclusion. Moreover, the likely impact of this scheme upon some children who are members of larger families living in high-rent homes is at most to make it unavoidable for the family to move; the duty of Local Authorities to provide accommodation under the Housing Act 1996, Part 7, remains. None of the judgments suggests that article 8 is engaged. I agree that it is not. It follows that article 3 UNCRC cannot have effect in English law on the grounds that it is relevant to its interpretation.

The additional argument now formulated before this court and accepted by Lady Hale and Lord Kerr would give article 3 UNCRC the force of domestic English law on the grounds that it bears on the issue of whether the agreed discrimination against women in relation to their A1P1 rights was justified. Lord Kerr would additionally give article 3 direct effect on the grounds that the UK’s signature to the convention is sufficient to impose a domestic duty to comply with it. Like Lord Reed and Lord Carnwath, I am unable to accept these arguments.

It may not be difficult to see that in interpreting the content of the article 8 rights of children, it may be legitimate to take into account the international obligation contained in article 3 UNCRC. ZH (Tanzania) v Secretary of State for the Home Department [2011] UKSC 4; [2011] 2 AC 166 was an article 8 case where the relevance to that article of the interests of the children of a potential deportee was conceded. Similarly, Neulinger v Switzerland (2010) 54 EHRR 1087 depended upon article 8. It concerned an order directly about the upbringing of a child, namely an order for return to another state pursuant to the Hague Convention on the Civil Aspects of Child Abduction, and the very first words of that convention declare the interests of children to be of paramount importance in matters relating to their custody. If article 8 rights are engaged, the question will often become: is such impairment of respect for private and family life nevertheless permissible under article 8(2)? If the article 8 rights relied upon are those of children, as was asserted here, or of their parents in the form of their relationship with their children, as in ZH (Tanzania), there is scope for the argument that an internationally recognised duty to approach the children’s interests in a particular way bears on whether article 8(2) is satisfied – in the context of these regulations whether any impairment of children’s article 8 rights was permitted on the grounds that it is necessary in a democratic society in the interests of the economic well-being of the country or the protection of the rights and freedoms of others, such as those taxpayers who do not claim benefits.

The Demir approach is not of course limited to article 8, as that case itself shows. And it may extend to cases where discrimination is in issue. Opuz v Turkey (2009) 50 EHRR 695 was an article 2/article 14 case involving a complaint of failure to protect from domestic violence. The court relied in part on the Convention on the Elimination of All Forms of Discrimination against Women (“CEDAW”) in determining the scope of article 14: see paras 185-187. Ponomaryov v Bulgaria (2011) 59 EHRR 799 was a complaint of discrimination against foreign nationals by charging for education when Bulgarian nationals received free provision. Obiter, the court referred at paras 56-57 to international conventions which indicated that the state’s margin of appreciation increased as one moved from primary, through secondary, to tertiary education. Burnip v Birmingham City Council [2013] PTSR 117 was a benefits case involving A1P1 and a derivative article 14 claim. In the Court of Appeal Maurice Kay LJ would have been prepared to adopt a similar approach by gaining assistance on the scope of article 14 from the UN Convention on the Rights of Persons with Disabilities (“CRPD”) if the extent of article 14 had been in doubt. Obiter, he also offered the opinion that CRPD might illuminate the approach to justification, but the occasion to test this did not arise. But before the Demir approach to the interpretation of the ECHR can be relevant, there has to be the necessary connection between the international law invoked and the Convention right under consideration. This was clearly present in each of Opuz, Ponomaryov and Burnip. In each, the international instruments referred to were directly concerned with the particular form of discrimination in issue. Demir does not mean that the UNCRC (in this case) becomes relevant to every ECHR question which arises, simply because children are as a matter of fact affected by the decision or legal framework under consideration.

It is said that the Strasbourg court has invoked article 3(1) UNCRC in the context of a discrimination claim in X v Austria (2013) 57 EHRR 405. That was a case in which the same-sex partner of a child’s mother wished to adopt the child, who lived with the two ladies. The effect of Austrian law was that adoption substituted the adoptive parent for the natural parent of the same sex. Thus ‘second parent adoption’ (adoption by the partner of the natural parent) by a same-sex partner was legally ineffective, since if the adoption order were made the same-sex partner of the mother would achieve parental rights, but in place of the natural mother, leaving the legal relationship of the absent father to the child unaltered. Conversely, ‘second parent adoption’ by the different-sex partner of the natural parent was effective. The claimants in that case were scrupulous in limiting their complaint about Austrian law to the resultant difference of treatment between, on the one hand, a different-sex unmarried couple and, on the other, a same-sex unmarried couple such as themselves. They disclaimed any complaint about any different treatment as between married couples and unmarried couples, which the court had previously found to be within the margin of state appreciation: see Gas & Dubois v France [2014] 59 EHRR 22.

The court decided the case on the grounds advanced by the claimants. The discrimination between different-sex couples and same-sex couples was based upon sexual orientation alone. Where such discrimination is in question, the margin of appreciation is narrow and proportionality requires not merely that the measure in question pursues a legitimate aim but also that it is necessary: see paras 140-141. The relevant Convention rights to which the derivative article 14 claim to discrimination was attached were the article 8 rights of all three people, the mother, her partner and the child. In the absence of any evidence submitted to suggest that a child was generally better brought up by a different-sex couple than by a same-sex couple, there was no justification for the different treatment as between such couples. The court adopted its usual practice of setting out international instruments in the field, and thus included article 3(1) UNCRC. The decision in question (adoption) related directly to the upbringing of the child. It is unsurprising that the court referred (somewhat in passing) at para 146 to the fact that its conclusion was also more in keeping with the best interests of the child, which it noted to be a key notion in the relevant international instruments. It might have added that in the great majority of developed states there is consensus that questions of a child’s upbringing must be determined by his or her best interests or welfare as the dominant or paramount consideration: in England this principle is long-established law and now encapsulated in section 1(1) of the Children Act 1989.

At its highest, this decision is another in which the UNCRC is referred to as relevant to the content of article 8 rights, and thus to the issue of justification for discrimination in relation to such rights. That is a very long way from saying that article 3(1) is relevant to justification upon any kind of discrimination issue, whether or not the decision is about the child’s upbringing, and whether or not either the ECHR rights of the child or article 8 rights of his family are at stake. Such issues simply did not arise in X v Austria.

If the rights in question are the A1P1 property rights of women, and their associated derivative right not to be discriminated against in relation to those rights, it is an impermissible step further to say that there is any interpretation of those rights which article 3 UNCRC can inform. In the case of article 8, the children’s interests are part of the substantive right of the parent which is protected, namely respect for her family life. In the case of A1P1 coupled with article 14, the children’s interests may well be affected (as here), but they are not part of the woman’s substantive right which is protected, namely the right to be free from discrimination in relation to her property. There is no question of interpreting that article 14 right by reference to the children’s interests. The protected right to respect for family life under article 8 is entirely different from the protected right to property under A1P1. Nor can the article 8 rights of the child be said to be in need of interpretation when it is clear for the reasons given in all the judgments that they are not infringed. The necessary connection between the ECHR right under consideration and the international instrument is not present. That can be seen by considering the position of the appropriate comparator, namely a lone non-working father with the same children and household outgoings. The interests of the children would be exactly the same in his case, but he would have no article 14 claim to discrimination.

I also agree that to treat failure to comply with article 3(1) UNCRC as determinative of the present case would be tantamount to departing from the Stec test for justification which has been agreed on all sides throughout this litigation.

It is unnecessary to decide this question, but I ought to say that in my view it is clear that there was in any event no breach of article 3.

The language of article 3(1) does give rise to some difficulty. It is in these terms:

It might be thought that article 3 was intended to apply to decisions directly about a child, or perhaps to those and to others directly affecting him, such as for example decisions relating to the provision of education or child-support facilities, and that “a primary consideration” therefore imports some priority for the best interests of children even if short of making them determinative, as the paramountcy principle does. That might perhaps be suggested by article 3(3) which clearly is specific to the care and protection of children, while article 3(2), which requires states to take appropriate legislative and administrative measures to ensure that the child has such protection and care as is necessary for his well-being, is also perfectly consistent with this. This is not, however, the view taken in General Comment 14, adopted by the UN Committee on the Rights of the Child at its 2013 session, referred to by Lord Carnwath at para 105, and foreshadowed by earlier similar documents.

That Comment suggests (at para 19) that article 3 extends well beyond decisions directly about children to those which indirectly affect either individual children or children in general, “eg related to the environment, housing or transport”. If the meaning of article 3(1) is as broad as this, then all manner of court decisions may fall within it; a planning decision relating to housing development might be one, whilst the making of a possession order against a tenant who has children, or the enforcement of money judgments against the family motor car, or the sentencing of him for a serious criminal offence might be others.

Pace Lord Carnwath, I do not take it as read that the Committee’s views, although entitled to careful consideration coming from the source that they do, can be regarded as binding upon party States as to the meaning of the treaty to which they agreed. But it is neither necessary nor appropriate to attempt to resolve these issues in this case, especially since we heard no argument upon them. All that needs to be said is that it is clear that the wider the reach of the concept of “decisions concerning” either an individual child or children in general, the less possible it is to impose the best interests of such child or children as a determinative or even priority factor over the frequently complex legal or socio-economic considerations which govern such decisions. The committee’s general comment gives some acknowledgement to this problem in, for example, para 20, which recognises that although all State actions may affect children, a full and formal process of assessing their best interests is not called for in every case, and in para 32 where it is stated that the concept of the child’s best interests is flexible and adaptable.

The Committee’s General Comment also realistically recognises that the relevant best interests of children will, in relation to decisions which are not simply about identified individual children, include those of children generally. This is apparent throughout the document, including in those passages from para 102 cited by Lord Carnwath. I respectfully agree with Lady Hale that where article 3(1) applies it is not enough to consider only the interests of children generally, without also evaluating the interests of any likely to be particularly affected by the legislation in prospect, but the converse is also true. It is obvious that in the context of this kind of socio-economic legislation, there will be a tension between, on the one hand, the interests of children generally in promoting the legitimate aims of reducing a culture of benefit-dependency and encouraging work and, on the other, the special interests of those children most likely to suffer an adverse effect of the cap, such as the present appellants. This is realistically recognised by the UN Committee in, for example, para 32 of the Comment, which reads:

Whilst the appellants in the present case relied upon article 3(1) as “substantive and not merely procedural” they did not analyse the extent to which it was asserted that priority ought to be given to children’s best interests, still less the interests of which children. Their chief reliance was upon the suggested failure of the Secretary of State properly to have analysed and considered the best interests of children. Relying upon paragraph 6 of the UN Committee’s General Comment, the principal submission of Mr Wise QC was that the article 3 obligation required (a) careful consideration of how many children will be or are likely to be affected by the cap, (b) asking what the effect on those children particularly affected by it would be, (c) asking whether the cap could be implemented in a manner protecting such children from adverse effects, and (d) asking whether the general proposition that the cap will lift children out of welfare dependency outweighs the risk to those particularly affected.

Like both courts below I regard it as plain that the Secretary of State did not fail to undertake all these exercises. There was the fullest public debate about not only the concept of the cap but its proposed details. This country has four Children’s Commissioners, charged with the duty of monitoring children’s interests and advocating them publicly. All participated in the debate and made strongly the case now made by the appellants that the general benefits to families and children which would be brought by the cap were outweighed by the likely adverse consequences for particular children in situations exactly like those of the present appellants. The two Impact Assessments and the Equality Impact Assessment written by the Government recorded the likely adverse consequences for children such as these, in particular those in larger one-parent families living in high rent areas. The Parliamentary debate on the detailed proposals returned time and again to this topic. There was a specific proposal, supported by the House of Lords, to amend the Bill by excluding child benefit from the cap, which, as Lady Hale observes, would no doubt remove the adverse impact on the appellants here relied upon; this proposal was considered but rejected by the House of Commons and withdrawn in consequence by the House of Lords. The Secretary of State concluded, and still concludes, that to do this would drive a coach and horses through the whole policy. The evidence could not really be clearer that the Secretary of State did indeed ask the questions which Mr Wise contends are required by article 3 UNCRC. The appellants’ real complaint is that he reached what they say is the wrong value judgment when it came to balancing the interests of children (and society) in general against those of particular children likely to suffer adverse effects from the cap. Reasonable people may well either agree or disagree with this value judgment, but to say that one disagrees is not the same as saying that the decision is unlawful.

The “benefit cap” is one of a package of measures provided for in the Welfare Reform Act 2012. The total amount of benefit to which a couple or a single person is entitled is capped at a prescribed sum, irrespective of how much they would otherwise be entitled to. The bare bones of the scheme are provided for in the 2012 Act, but its detailed implementation is contained in the Benefit Cap (Housing Benefit) Regulations 2012.

The appellants do not challenge the compatibility of the Act with their rights under the European Convention on Human Rights, but they do challenge the compatibility of the way in which it has been implemented by the 2012 Regulations. They argue that it has a disproportionate impact upon lone parents and upon the victims of domestic violence; both groups are predominantly, although not exclusively, composed of women; hence the scheme is indirectly discriminatory on grounds of sex. As the scheme falls within the ambit of the protection of property rights in article 1 of the First Protocol to the Convention, this violates their right, under article 14 of the Convention, to enjoy such rights without discrimination unless it can be justified. The Secretary of State accepts that the scheme falls within the ambit of article 1 of the First Protocol and that it is indirectly discriminatory against lone parents and thus against women. The question, therefore, is whether it can be justified. A further question, which has only emerged after the hearing in April 2014, is the extent to which, if at all, the obligations of the United Kingdom under the United Nations Convention on the Rights of the Child is relevant to that issue.

Both the Divisional Court and the Court of Appeal held that it can be justified: [2013] EWHC 3350 (QB) and [2014] EWCA Civ 156. This raises several questions: whether the justification advanced relates to the scheme as a whole rather than to its discriminatory effect; what is the test to be applied in deciding whether the discrimination is justified; and what is the part played by the international obligations of the United Kingdom under the United Nations Convention on the Rights of the Child in assessing that.

The benefit cap is, of course, quintessentially a matter of social and economic policy. In such matters, as Lord Hope of Craighead observed in R v DPP, Ex p Kebilene [2000] 2 AC 326, at p 381, it will be easier for the courts to recognise a discretionary “area of judgment within which the judiciary will defer, on democratic grounds, to the considered opinion of the elected body or person whose act or decision is said to be incompatible with the Convention”. As Lord Reed explains, the introduction of the cap was indeed extensively debated in Parliament and various amendments were proposed and resisted which would have mitigated the adverse effects with which we are here concerned. But the details of the scheme, including those adverse effects, were deliberately left to be worked out in regulations. It is therefore the decisions of the Government in working out those details, rather than the decisions of Parliament in passing the legislation, with which we are concerned.

Furthermore, as Lord Hope went on to say in In re G (Adoption: Unmarried Couple) [2009] 1 AC 173, para 48, protection against discrimination, even in an area of social and economic policy, falls within the constitutional responsibility of the courts:

It is not necessary to go into the scheme in great detail, but it is necessary to understand the essentials. Section 96(1) of the Act provides that “Regulations may provide for a benefit cap to be applied to the welfare benefits to which a single person or couple is entitled”. Section 96(2) provides that where their total entitlement to “welfare benefits” exceeds the “relevant amount”, their entitlement is reduced by the amount of the excess. This is the cap. The “relevant amount” is to be specified in Regulations (section 96(5)), but “is to be determined by reference to estimated average earnings” (section 96(6)). By this is meant “the amount which, in the opinion of the Secretary of State, represents at any time the average weekly earnings of a working household in Great Britain after deductions in respect of tax and national insurance contributions” (section 96(7)). “Welfare benefits” means any benefit, allowance, payment or credit prescribed in regulations (section 96(10)); but retirement pensions and state pension credit may not be prescribed (section 96(11)). Regulations may also provide for exceptions to the application of the cap (section 96(4)(c)) and also for the benefit or benefits from which the reduction is to be made (section 96(4)(b)).

Thus it will be seen that all the details of the scheme are to be covered in the regulations. The only principle required by the Act, should the Government decide to introduce a cap at all, is that it is set by reference to average weekly earnings net of tax and national insurance contributions. This, as Mr Holmes, the lead official in the Department of Work and Pensions responsible for the benefit cap policy, points out, produces a much higher figure than would be produced by working 40 hours a week for the minimum wage or even the London living wage. But the Government was left a free hand in deciding what working age benefits would count towards the cap.

In fact, the cap operates by way of a deduction from housing benefit. Hence the 2012 Regulations amend the Housing Benefit Regulations 2006, principally by introducing a new Part 8A, entitled “Benefit cap”. The “relevant amount” is set at £350 for a single claimant (without dependent children) and £500 for all other claimants (that is, couples and lone parents with dependent children) (regulation 75G). This is the equivalent of a gross annual salary of £35,000 a year and £26,000 net. A long list of welfare benefits is prescribed, most importantly for our purposes including housing benefit, child benefit and child tax credit (regulation 75G). Once the cap is reached, therefore, no account is taken of the number of children in the family. On the other hand, the benefit cap does not apply at all where the claimant, the claimant’s partner or a child or young person for whom either is responsible is receiving any of a long list of benefits; these are mainly disability-related but include a war pension (regulation 75F).

The cap does not apply at all where the claimant is, or the claimant and her partner are jointly, entitled to working tax credit (regulation 75E(1), (2)). This effectively exempts most working households from the cap; the rules are complicated, but a lone parent responsible for a child would qualify for working tax credit if she worked at least 16 hours a week, while a couple responsible for a child would qualify if they worked a total of 24 hours a week, as long as one of them worked for at least 16 hours a week; the normal requirement is 30 hours’ work a week (Working Tax Credit (Entitlement and Maximum Rate)) Regulations 2002, regulation 4, as amended by the Tax Credits (Miscellaneous Amendments) Regulations 2012, regulation 2). Not only that, if the claimant or her partner have been employed or engaged in work for payment for 50 out of the preceding 52 weeks, the benefit cap will not apply for 39 weeks from their last day of work (regulation 75E(1), (3) – (5)). This gives a period of grace in which to find another job or to move house.

The final regulation which is relevant for our purposes is that which provides, in effect, that the housing benefit payable for what is now (following a recent amendment) to be termed “specified accommodation” is disregarded (regulation 75C(2)(a)). The amendment means that women’s refuges are now covered, whereas previously many of them were not. However, there is no comparable exemption for housing benefit paid in respect of temporary accommodation provided under the homelessness provisions of Part 7 of the Housing Act 1996.

The benefit cap was introduced in April 2013 in four London boroughs, “rolled out” in July 2013 to a further 335 local authorities and in August 2013 to the remaining 40 authorities in England, Wales and Scotland. It has not yet been implemented in Northern Ireland. Between April 2013 and January 2014, a total of 38,655 households were capped, 47% of these in London and the vast majority in England.

As Elias LJ, giving the judgment of the Divisional Court, observed at [2013] EWHC 3350 (Admin), para 11, “It was obvious from the outset that the introduction of the cap would have severe and immediate consequences for claimants who had been receiving substantially in excess of the relevant amount”. To mitigate this, the Government provided additional funds to local authorities to enable them to make discretionary housing payments (DHPs) to claimants affected by the cap (along with the other purposes for which such payments may be made). This was specifically intended as a short term solution where transitional help was necessary and not as a long term solution to the needs generated by the cap (see Holmes, Witness Statement No 1, para 130).

Elias LJ continued, at para 12:

The four appellants are the lone mother and her youngest child in two families (a third family has now withdrawn from the case as the cap no longer applies to them). The following evidence of their circumstances was before the court when the case was heard in April 2014.

Ms SG and her family live in Stamford Hill, North London. This is important because they are members of a particular orthodox Jewish sect. The school age children attend a local Jewish school, kosher food is readily available (but expensive) in the local shops, they can walk to the synagogue and there is a support network of family and friends there. Their lone mother has six children in all, but only three of them live with her: a son now aged four, a daughter now aged seven and another daughter now aged nine. The family used to live in Belgium, but SG left her husband and came to live near her relatives in Stamford Hill in order to escape from her husband’s abusive behaviour towards her and their eldest daughter, now aged 18. The daughter was made a ward of court to prevent her father removing her from this country. Because of her behavioural and psychological difficulties she was placed by the local authority in foster care within the same community. She has since married but still lives locally and relies heavily upon her mother for support. The oldest son studies in a yeshiva abroad and is unlikely to rejoin the family, but there are currently proceedings in Belgium about the residence of the second son, now aged 12, whom his mother earnestly hopes can return to live with the family in London.

The family live in a two bedroomed flat rented from a private landlord. This is already too small for them and would be quite unsuitable were the 12 year old boy to come and live with the family again. When these proceedings began, the rent was £300 per week, but the landlord was proposing to put it up. They were entitled to £289.20 housing benefit, £71.70 income support for SG, £167.30 Child Tax Credit (all means-tested benefits), and £47.10 child benefit. Hence their total benefit entitlement before the cap was £575.30 a week. The cap has therefore resulted in a reduction of £75.30 in their weekly income. The landlord has notified an increase in her rent to £420 from 31 January 2014, which would leave them with only £80 to live on.

The Secretary of State correctly points out that housing benefit would not in any event meet such a high rent in full (because it exceeds the local housing allowance limit for that part of London). He also argues that there are cheaper two bedroomed flats available in the area, but the appellants dispute this. We are not in a position to resolve such factual disputes. However, it is obvious that SG has very good reasons for wanting to continue to live in Stamford Hill, that accommodation there is in short supply because of demand from the local community, and that if she does stay there her weekly income will fall well below that which the State deems necessary for her and her three young children to live on.

For a time, she did have part time work for 16 hours a week and thus the benefit cap did not apply. But she was unable to sustain this, owing to the demands of the court proceedings relating to her children, both here and in Belgium, and the need to care for the younger children. The 39 week grace period expired in November 2013, since when her benefits have been capped. She has been receiving a discretionary housing payment to meet the shortfall between her rent and her housing benefit, but only until 30 June 2014, when it was due to be reviewed having regard to the steps she has taken to avoid the cap.

Mrs NS is also the lone mother of three children, daughters now aged 4, 11 and 12. There is a long history of sexual abuse and domestic violence within her marriage, much of it witnessed by the children. She had left her husband to stay in a women’s refuge with the children on two previous occasions before their final separation in December 2012. After a period in unsuitable accommodation, she obtained orders excluding her husband from the family home, and returned there with the children in April 2013. Her husband is prohibited from contacting the family there, but last summer they had to turn to him for help with transport when one child suffered an accident requiring surgery and the other two became ill. NS is concerned that the local children’s services authority will consider her children to be at risk of harm if they have contact with their father.

Their home is also a two bedroomed flat rented from a private landlord. It is also too small for them but is close to the children’s schools. The rent is £270 a week. She is entitled to £270 housing benefit, £71.70 income support for NS, child tax credit for the children of £166.94 (although she says that she gets only £162.44), and child benefit of £47.10. Her total entitlement therefore should amount to £555.74 (although she says that she gets only £550.44). Whichever it is, the cap reduces it to £500.

NS was awarded discretionary housing payments, but only after a delay during which arrears accrued to her rent account, and only until 31 March 2014. The local authority has yet to decide upon its DHP budget for this year and so she does not know whether or not she will get it. She is of course concerned that the landlord may seek to evict her if she falls into arrears.

NS did not work outside the home during her marriage, nor has she done so since it ended. She was allowed very little freedom by her husband and speaks very little English.

It is common ground that the scheme falls within the ambit of article 1 of the First Protocol, which protects the right to peaceful enjoyment of possessions. Possessions for this purpose includes entitlement to welfare benefits, not only those which have been “paid for” by national insurance contributions, but also those which the State provides on a non-contributory basis to supply its people with the basic necessities of life. As the Strasbourg court explained in Stec v United Kingdom (2006) 43 EHRR 1017, para 53:

It has not been argued that the benefit cap is itself a violation of article 1 of the First Protocol, on the basis that it deprives affected households of the benefits to which they would be entitled under the usual rules relating to needs-related welfare benefits. Instead, it is argued that it violates article 14, which provides that “the enjoyment of the rights and freedoms set forth in the Convention shall be secured without discrimination on any ground such as sex …”. It is not suggested that the scheme is directly discriminatory against women, as it affects all benefit claimants in the same way, irrespective of their sex. However, as the Divisional Court observed, “It is clear, and indeed conceded, that the benefit cap has a disproportionate adverse impact on women” (para 71). This brings it within the concept of indirect discrimination, which was recognised by the Grand Chamber of the European Court of Human Rights in DH v Czech Republic (2007) 47 EHRR 59, at para 175 (see also para 184):

The prejudicial effect of the cap is obvious and stark. It breaks the link between benefit and need. Claimants affected by the cap will, by definition, not receive the sums of money which the state deems necessary for them adequately to house, feed, clothe and warm themselves and their children. Furthermore, the greater the need, the greater the adverse effect. The more children there are in a family, the less each of them will have to live on. Ms SG, for example, will receive no more benefit if her 12 year old son rejoins the family, even though a court (either here or in Belgium) has decided that it is in his best interests to do so. This prejudicial effect has a disproportionate impact upon lone parents, the great majority of whom are women, and is also said to have such an impact upon victims of domestic violence, most of whom are also women.

The disproportionate impact upon lone parents is relatively straightforward to explain. The relevant comparison is between those housing benefit claimants who are, and those who are not, affected by the benefit cap. Lone parents constitute around 24% of all claimants for housing benefit, but have so far constituted between 59% and 74% of those affected by the cap. This is more than double their proportion in the housing benefit population as a whole. Overall some 92% of lone parents are women. Hence it is not surprising that the Government predicted, in its first Equality Impact Assessment of the Benefit Cap (March 2011, para 27), that single women, mostly lone parents, would constitute 60% of those affected.

The reasons for this are fairly obvious. It is much more difficult for lone parents to move into paid employment, even for the 16 hours which would take them out of the cap. It is more difficult for them to do so, the more children they have, because of the problems of delivering and collecting children from different schools or day care placements, the problems of making appropriate day care arrangements for very young children and for all children during the school holidays, the problems of responding to their children’s illnesses, accidents and to casual school closures. The more children they have, the harder it will be for them to move into work; and the more children they have, the harsher will be the effects of the cap. These problems arise irrespective of the ages of the children, but are obviously more acute when any or all of them are under school age.

The disproportionate effect which the cap is said to have upon victims of domestic violence, most of whom will also be parents, is a little more complicated. It stems from the limited options available to victims who wish to escape, with their children, from the violence and abuse which they are suffering at home. Some victims are fortunate enough to be able to stay in their own homes while the perpetrator either agrees or is ordered to leave and having done so can be relied upon to stay away. But many are not so fortunate. Their only way of escaping the violence, at least in the first instance, is to leave home. If they go to a refuge, the problem is that the costs may easily take them over the cap. Under the original scheme, some refuges counted as “exempt” accommodation, which effectively created an exception to the cap, but many did not. Very recently, the Government has addressed this, by amendments which will create an exception for all refuges.

But not all victims can go to a refuge. Their other alternative is to apply to the local authority for accommodation under the homelessness provisions of Part 7 of the Housing Act 1996. Unlike the cost of refuges, the cost of other types of temporary accommodation is not exempt. Temporary accommodation is often in the private sector and much more expensive than permanent accommodation in social or other forms of affordable housing. Furthermore, as the intervention from Shelter makes clear, a homeless person has very little choice about where she is housed. She has to accept any offer of “suitable” accommodation or risk becoming literally without a home (and even having her children taken away from her as a result). In areas of high housing need, families may stay for a very long time in so-called “temporary” accommodation before affordable permanent housing becomes available.

Some of these victims will want to keep open the possibility of returning to the family home, or securing a transfer, once the family court has decided who is to live there. Hence, very sensibly, the housing benefit scheme provides that in certain circumstances councils may continue to pay benefits in respect of two homes for a certain length of time (Housing Benefit Regulations 2006, regulation 7(6)(a)). But this, of course, means that the total amount of housing benefit, when taken together with other benefits, will take the claimant over the limit where the cap applies.

Thus, even with the recent change relating to refuges, the effect of the cap is to undermine the humane treatment given to victims of domestic violence both by the homelessness regime and by the housing benefit scheme. However, although both of the families whose cases are before us have suffered from domestic violence and abuse, they have not suffered these particular adverse effects (we do not know whether Mrs NS’s family was in receipt of dual housing payments between December 2012 and April 2013, but in any event that was before the cap came into force), nor do they claim to be at risk of suffering them in the future. For this reason, the Divisional Court and the Court of Appeal declined to decide whether the cap did have a disproportionate effect upon the victims of domestic violence. Mr Wise QC, for the appellants, complains that they should have done so. The appellants have both suffered domestic violence and abuse and Mrs NS might well have to flee to expensive temporary accommodation while wishing to retain the family home should her husband once again try to assert his control over her.

In my view, however, the problems suffered by the victims of domestic violence are principally suffered because they are parents who have every reason to separate from the other adult in the household, not only for their own sake but also for the sake of their children. Of course, there may be some victims of domestic violence who are not responsible for the care of children, but it has not been shown how likely it is that they will be affected by the cap or how difficult they would find it to escape its adverse impact. I would therefore treat the victims of domestic violence as a subset of lone parents, who may be more likely to be affected by the cap because of the high cost of temporary accommodation and the dual payments problem, and who will have the same problems in escaping its effects.

The applicable principles are set out in the Grand Chamber judgment in Stec v United Kingdom (2006) 43 EHRR 1017, para 51:

However, it is important to understand that what is needed to justify indirect discrimination is different from what is needed to justify direct discrimination. In direct discrimination, it is necessary to justify treating women differently from men. In indirect discrimination, by definition, women and men are treated in the same way. The measure in question is neutral on its face. It is not (necessarily) targeted at women or intended to treat them less favourably than men. Men also suffer from it. But women are disproportionally affected, either because there are many more of them affected by it than men, or because they will find it harder to comply with it. It is therefore the measure itself which has to be justified, rather than the fact that women are disproportionately affected by it. The classic example is a maximum age bar on recruitment to particular posts; it applies to all candidates, women and men; but it disadvantages women because they are more likely to have taken a career break to have or care for children than are men. The question therefore is whether the age bar can be independently justified. This long-standing position is reflected in the definition of indirect discrimination in section 19 of the Equality Act 2010. It was also the approach of the Strasbourg court in Hoogendijk v Netherlands (2005) 40 EHRR SE 189, a case of indirect discrimination in relation to welfare benefits.

Turning to the explanations offered for the cap, it is important to recognise that the Government has never claimed that its aim is to encourage claimants to limit the size of their families or to penalise those who already have large families (had they done so, they might perhaps have faced discrimination claims on other grounds). The evidence before the court is contained in two witness statements from Mr Holmes. He states that the Government had three specific aims in introducing the benefit cap:

to “introduce greater fairness in the welfare system between those receiving out-of-work benefits and tax payers in employment”;

to make financial savings (anticipated to be £110m in 2013/14 and £185m in 2014/15) and “more broadly, help make the system more affordable by incentivising behaviours that reduce long term dependency on benefits”; and

to increase incentives to work. This is later described as “the main aim of the policy” (Holmes, Witness Statement No 1, para 107).

It is accepted that achieving fairness between those in work and those out of work is a legitimate aim. As Elias LJ recognised, “the fairness concept has sometimes been justified by relying on the notion that those on benefit should face difficult decisions of the kind facing those in work” (para 94). But there are many different ways of defining such fairness. It could be that a family on benefits should never be better off than a working family of the same size living in the same accommodation. It could be that a family on benefits should always be worse off than the equivalent working family. Or it could be that a family on benefits should always be much worse off than the equivalent working family.

The criticism levelled at the Government’s concept of fairness, in particular in the intervention from the Child Poverty Action Group, is that the benefit cap scheme as implemented does not compare like with like. It compares the maximum level of benefit with average earnings, thus ignoring the benefits which are also available to people who are in work. CPAG have produced tables (not challenged in these proceedings) comparing the income available to each of the appellant families according to whether they are (a) not working but without the cap, (b) working 16 hours per week on the minimum wage, (c) working for average household earnings, and (d) working 35 hours a week for the minimum wage. These show that both Ms SG and her children and Mrs NS and her children would be (in round figures) £94 a week better off in scenario (b) than in scenario (a), £163 better off in scenario (c), and £122 better off in scenario (d). In other words, they would always be significantly better off in work than not in work. CPAG have also produced tables which show that this would also be the case wherever in the country these families were living. The effect of the cap is simply to increase the differential which is already there.

Thus, it is said, there was no need to introduce the benefit cap in order to ensure that families on benefit have to make the same difficult choices that working families have to make. They already do have to make those choices. If this is so, the focus shifts to the other two objectives.

The savings projected by the Treasury in the 2013 budget were £110m in 2013-2014 and £185m in 2014/15. These did not take into account the possible implementation costs or the additional funding made available for DHPs of £65m and £35m respectively. On the other hand, nor did they take into account any resulting behavioural changes. The aim was not merely to make savings in the short term but to “produce a positive cultural shift” (Holmes, Witness Statement No 2, para 36).

It has to be accepted that the savings made are a drop in the ocean compared with the total benefit bill, let alone the total housing benefit bill. The Government predicted that only 1% of housing benefit claimants would be affected by the cap. In May 2013, there were approximately five million housing benefit claimants, yet in January 2014 there were less than 28,000 households subject to the cap, not much over half a percent of all claimants. Lone parents subject to the cap were 1.37% of all claimants (further demonstrating that they are disproportionately affected).

However, the main argument made against this aim is that, standing alone, it is not sufficient to justify discriminatory treatment in the enjoyment of a convention right. The authority cited for this proposition is O’Brien v Ministry of Justice [2013] UKSC 6, [2013] 1 WLR 522. This was a case about discrimination between full time and part time workers, which is prohibited by the Framework Agreement on Part-time Work, annexed to Council Directive 97/81/EC.

However, in Andrejeva v Latvia (2009) 51 EHRR 650, the Strasbourg court accepted that the “protection of the country’s economic system” is a legitimate aim which is “broadly compatible with the general objectives of the Convention” (para 86). They therefore looked to see whether there was a reasonable relationship of proportionality between that legitimate aim and the means employed. As the discrimination in that case was based solely on nationality, for which “very weighty reasons” would be required for compatibility with the Convention, the court held that it was not justified (paras 87-88). The same would apply to sex discrimination. If the state introduces a benefit, for example for older people, but denies it to women on the basis that this will save money, this would be contrary to article 14 read with article 1 of the First Protocol, unless there were some other justification for the difference in treatment. The court found such a justification in Stec, because the difference complained of was the result of the difference between the retirement ages of men and women, itself a response to the disadvantage suffered by women in the workplace. This brings the focus back to the proportionality of any discrimination involved in a money-saving measure.

Mr Holmes also refers in his evidence to “a clear, simple message that there has to be a maximum level of financial support beyond which claimants cannot expect the state to provide” (Witness Statement No 1, para 98) and “one of the key drivers for introducing the cap, that ultimately there has to be a limit to the overall amount of financial support that households in receipt of out of work benefits can expect to receive in welfare payments” (para 104). However, it is difficult to see how the delivery of such a message can be an aim in itself if the message is the product of a measure which cannot be justified.

On analysis, it is therefore said, the Government’s aims come down to incentivising work and promoting long term behavioural change. Again, no-one doubts that these are legitimate aims, not only in order to save public money but also, as Mr Holmes put it, “to achieve long term positive behavioural effects by changing attitudes to welfare and work and encouraging responsible life choices, which will benefit adults and children alike” (Witness Statement No 1, para 121). Put another way, it is not good for children to grow up in a household which is wholly supported by the state, if thereby they absorb the message that there will be no need for them to support themselves when they grow up.

However, the Government has accepted that certain people should not be expected to seek work in order to escape the cap. Thus retirement pension and state pension credit are not taken into account because “the policy is primarily a work incentive aimed at people of working age” (Holmes, Witness Statement No 1, para 100). Thus also the cost of “supported accommodation” is not taken into account because “households in supported accommodation are likely to be in vulnerable situations … and they will not generally be in a position to make quickly the behavioural changes required to remove themselves from the cap” (para 105). Thus also the “disability-related exemptions mean that the cap will not apply to people who are least likely to be able to work and who perhaps have the least scope to adjust their circumstances to improve their employment prospects” (para 112). Lone parents of children under five are also not expected to seek work, but they are subject to the cap.

As well as moving into work, the other choices the Government wished to encourage as a way of avoiding the cap included persuading the landlord to take less rent, moving to cheaper accommodation, reducing expenditure on non-housing items, and in the case of lone parents seeking child maintenance from the other parent, which is wholly disregarded for the purpose of the cap (Holmes, Witness Statement No 1, para 124).

Against this, both the appellants and the interveners argue that these expectations are simply unrealistic in the case of the families of lone parents and victims of domestic violence, upon whom the policy has such an adverse effect. For the reasons already mentioned, lone parents, especially those with more than one child, find it particularly difficult to obtain even part time work which will fit in with their child care responsibilities. It is accepted, of course, that there are some lone parents, even of very young children, who do manage to do this. Adequate and subsidised day care is now more readily available. But it is unrealistic to assume that parents will always be able to find acceptable solutions without prejudice to their children’s welfare. The Government accepts that lone parents of children under five should not be expected to look for work, no doubt partly because of the difficulties of finding acceptable and affordable child care, but perhaps also because many parents and child care professionals consider it better for very young children to have the full time loving care of a committed parent rather than be separated from them and placed in institutional settings, however competent, for a large part of the day. Even if we accept that it is justifiable to deny this choice to those lone parents who are subject to the benefit cap, we should not accept that their children’s welfare should be put at risk by their having to make unsatisfactory child care arrangements or (as in the case of Mrs NS) to rely upon assistance from a violent partner which the local children’s services authority fears may put the children at risk.

Nor is it realistic to assume that they will eventually be able to move to cheaper accommodation. Many private landlords, particularly in the more expensive areas, are unwilling to take tenants who are dependent on housing benefit. In any event, they will require deposits and rent in advance, which the family will not be able to afford (unless they can persuade the local children’s services authority to help out under section 17 of the Children Act 1989). Social housing is in short supply, with long waiting lists which may well require a qualifying period of residence in the area before a person is even placed on the list. The allocation criteria under Part 6 of the Housing Act 1996 do give preference to those homeless families to whom the full housing duty is owed under Part 7 of that Act (1996 Act, section 167(2)(a)). But if the family try to move to another local authority area where housing is cheaper or more plentiful, they may be refused on the ground that they have no local connection with that area. It will be particularly difficult for them to move if they have rent arrears, but the benefit cap is very likely to lead to rent arrears unless there is a speedy grant of a discretionary housing payment to fill the gap, which certainly cannot be guaranteed.

The Court of Appeal has recognised that discretionary housing payments are not an answer. In Burnip v Birmingham City Council [2012] EWCA Civ 629, [2013] PTSR 117, the Court of Appeal held that it was unjustifiably discriminatory to limit a severely disabled man who needed an overnight carer to the housing benefit payable for a one-bedroomed flat. As Henderson J explained at para 46, where there is a gap between objectively verifiable need and the housing benefit payable,

It was predicted that there would be an increase in evictions and homelessness as a result of the benefit cap. If the family does become homeless because of the cap, the Government hopes that neither the local housing authority nor the courts will regard them as intentionally homeless. They will have a priority need and should therefore be owed the full housing duty under Part 7 of the 1996 Act (1996 Act, sections 189(1)(b), 193(2)). Nevertheless, it may take a very long time before permanent accommodation becomes available, during which time they will be placed in temporary accommodation, often in the private sector. This is known to be more expensive than permanent accommodation. In other words, if they become homeless as a result of the cap, they are equally likely to be capped in their temporary accommodation. They do not have a choice. Provided the accommodation is “suitable” they have to take what is offered. The Government points out that affordability is part of suitability, but there may well be nothing else available. Local housing authorities have difficulty finding enough accommodation, and it is simply unrealistic to expect a homeless family to turn down an offer of otherwise suitable accommodation on the basis that it is not affordable. The Government wishes to encourage local authorities to move people out of temporary accommodation as soon as possible (Holmes, Witness Statement No 1, para 114), but the question is whether depriving homeless families of the full cost of such accommodation is a proper way to put pressure on local authorities to do so.

In addition, there are many other reasons why it may be quite unreasonable to expect a lone parent to move to another area. Finding new schools for several children in an unfamiliar area is not straightforward, nor is it good for the education which will in the long term be the best way of lifting those children out of poverty. Thus the Divisional Court concluded (at para 27, echoed almost precisely at para 22 in the judgment of the Court of Appeal):

As CPAG point out, the Government accepted in its grounds of resistance to the claim that “the aim of incentivising claimants to work may be less pertinent for those who are not required to work” (such as parents with young children). Hence it has to fall back on “making fiscal savings and creating a system which is fairer as between those receiving out of work benefits and working households”.

The Strasbourg court will, of course, allow Contracting states a margin of appreciation in assessing whether the difference in treatment is justified. As is well known, the width of that margin differs according to the subject-matter. In Stec, the court went on to explain, in para 52:

The references cited for the “manifestly without reasonable foundation” test were James v United Kingdom (1986) 8 EHRR 123, para 46, and National & Provincial Building Society v United Kingdom (1997) 25 EHRR 127, para 80, both cases complaining of a violation of article 1 of the First Protocol. In AXA General Insurance Ltd v HM Advocate [2011] UKSC 46, [2012] 1 AC 868, both Lord Hope, at para 31, and Lord Reed, at para 124, treated this test as directed towards whether the measure is “in the public interest”, in other words to whether it has a legitimate aim. They dealt separately with whether the interference with property rights was proportionate. They relied upon cases such as Pressos Compania Naviera SA v Belgium (1995) 21 EHRR 301, at para 38, where the Strasbourg court appears to have regarded this as a separate question:

When it comes to justifying the discriminatory impact of an interference with property rights, a distinction might similarly be drawn between the aims of the interference and the proportionality of the discriminatory means employed. However, it has been accepted throughout this case that the “manifestly without reasonable foundation” test applies to both parts of the analysis; but that, as this court said in Humphreys v Revenue and Customs Commissioners [2012] UKSC 18, [2012] 1 WLR 1545, at para 22, “the fact that the test is less stringent than the ‘weighty reasons’ normally required to justify sex discrimination does not mean that the justifications put forward for the rule should escape careful scrutiny”.

In Burnip’s case, at para 21, Maurice Kay LJ pointed out that “In the recent past, the European court has shown an increased willingness to deploy other international instruments as aids to the construction of the Human Rights Convention”. He cited, among others, the important case of Opuz v Turkey (2009) 50 EHRR 695, at para 185:

Burnip was concerned with discrimination against disabled people by failing to make reasonable accommodation for their special needs. The United Nations Convention on the Rights of Persons with Disabilities (CRPD) was cited to the Court of Appeal, but not, it appears, the case of Glor v Switzerland, Application No 13444/04, 30 April 2009, where the Strasbourg court reiterated that the Convention must be interpreted in the light of present-day conditions, including the European and worldwide consensus on the need to protect people with disabilities from discriminatory treatment, citing in particular the CRPD. The Court of Appeal in Burnip felt able to determine the issue without resort to the CRPD, but had he not been able to do so, Maurice Kay LJ would have resorted to that Convention, which would have resolved any uncertainty in favour of the claimants. He continued “It seems to me that it has the potential to illuminate our approach to both discrimination and justification” (para 22).

Likewise, our approach to both discrimination and justification in this case may be illuminated by reference to other international instruments to which the United Kingdom is party, including not only the Convention on the Elimination of All Forms of Discrimination against Women (CEDAW), which was the relevant instrument in Opuz v Turkey, but also most notably the United Nations Convention on the Rights of the Child (UNCRC). In Neulinger v Switzerland (2010) 54 EHRR 1087, for example, the Grand Chamber observed, at para 131:

This may be putting matters a little too high. The relevant international instruments relied upon by the Grand Chamber were, principally, article 3(1) of UNCRC:

As this court recognised in ZH (Tanzania) v Secretary of State for the Home Department [2011] UKSC 4, [2011] 2 AC 166, para 25, “a primary consideration” is not the same as “the primary consideration” still less “the paramount consideration”. Nevertheless, the obligation to treat their best interests as a primary consideration in all actions concerning children is binding upon the Government of this country in international law. It has also become relevant in domestic law in at least two ways. First, section 11 of the Children Act 2004 places a duty on a wide range of bodies providing public services to carry out their functions “having regard to the need to safeguard and promote the welfare of children”. This duty has also been placed on the Secretary of State for the Home Department in the exercise of her functions in relation, among other things, to immigration, asylum or nationality, by section 55 of the Borders, Citizenship and Immigration Act 2009.

This duty has not yet, however, been extended to all Government departments, including the Department of Work and Pensions, with whose decisions we are concerned in this case. Nevertheless, in a Written Statement to Parliament on 6 December 2010, the Minister of State for Children and Families made

However, the international obligations which the United Kingdom has undertaken are also taken into account in our domestic law insofar as they inform the interpretation and application of the rights contained in the European Convention, which are now rights in UK domestic law. There is no reason at all why those obligations should not inform the interpretation of the Convention right to the enjoyment of the substantive Convention rights without discrimination just as much as they inform the interpretation of the substantive Convention rights. ZH (Tanzania) happened to be a case about article 8, as were H (H) v Deputy Prosecutor of the Italian Republic, Genoa [2012] UKSC 25, [2013] 1 AC 338, and Neulinger itself. The Strasbourg court has taken the UNCRC into account in construing other articles of the Convention, most notably article 6 in relation to the fair trial of juvenile offenders, in T v United Kingdom (1999) 30 EHRR 121.

For these reasons, echoing Maurine Kay LJ in Burnip, I agree that our international obligations under the UNCRC and CEDAW have the potential to illuminate our approach to both discrimination and justification. Whatever the width of the margin of appreciation in relation to the subject matter of a measure, the Strasbourg court would look with particular care at the justification put forward for any measure which places the United Kingdom in breach of its international obligations under another human rights treaty to which we are party.

Hence it is no surprise that the Divisional Court held that the court should have regard to the UNCRC as a matter of Convention jurisprudence and the Secretary of State did not challenge that view in the Court of Appeal (see para 69 of their judgment) or, initially, in this court. The Statement of Facts and Issues agreed between the parties for this court included:

The Secretary of State makes two main arguments against taking article 3(1) of UNCRC into account in deciding whether this discrimination can be justified. The first is that the UNCRC, like other international conventions, can inform the substantive content of the Convention rights, but not the approach to proportionality and discrimination. As to proportionality, this argument is clearly negated by the Grand Chamber decision in Neulinger v Switzerland (2010) 54 EHRR 1087, where the best interests of the child were taken into account in deciding whether the interference with the parties’ rights to respect for their family life, entailed in an order to return to the child’s home country of Israel, was proportionate. Reference was also made to the long line of cases dealing with the expulsion of aliens, “according to which, in order to assess the proportionality of an expulsion measure concerning a child who has settled in the host country, it is necessary to take account of the child’s best interests and well-being” (para 146). In those cases, the best interests of a child have been taken into account in assessing the proportionality of an interference with the Convention rights of others: see Uner v Switzerland (2006) 45 EHRR 421, at paras 57-58.

It is no doubt for that reason that the Secretary of State for the Home Department conceded, in ZH (Tanzania) [2011] 2 AC 166, that removing the mother would be a disproportionate interference with the children’s article 8 rights. This concession was rightly made, irrespective of section 55 of the Borders, Citizenship and Immigration Act 2009. The relevance of the duty in that section was to whether the decision was “in accordance with the law” (see para 24) rather than to its proportionality.

As to discrimination, the Secretary of State’s argument is clearly negated by the Grand Chamber decision in X v Austria (2013) 57 EHRR 405. This was a case of alleged discrimination on grounds of sexual orientation. A same sex partner could not adopt so as to become a joint parent with the birth parent partner, whereas an opposite sex partner could do so. When dealing with the relevant international law, at para 49, the court begins with the article 3(1) of the UNCRC, before turning to article 21 and other specific provisions on adoption. When discussing the suggested justifications for the discrimination, at para 146, the court concludes that “Unless any other particularly convincing and weighty reasons militate in favour of such an absolute prohibition, the considerations adduced so far would seem rather to weigh in favour of allowing the courts to carry out an examination of each individual case. This 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”. The footnote refers back to para 49. In common with Lord Carnwath, I read this case as clearly indicating that the best interests of the child are to be taken into account in determining whether discrimination is justified under article 14.

The second argument now advanced by the Secretary of State is that the discrimination in this case is not against the children involved but against their mothers. It is not suggested that the rights of the children themselves have been infringed. This case may be contrasted with Neulinger, and indeed ZH (Tanzania), in which the complaint was of interference with the children’s right to respect for their family lives, as well as their mothers’. However, the same cannot be said of X v Austria. The child was a complainant, but it was not suggested that there had been discrimination against her; rather it was that the discrimination against her mother and her mother’s same sex partner affected (but did not infringe) her right to respect for her family life. It is difficult indeed to see how the family life of the child in that case was any more affected by the legal status of the people looking after her than is the family life of the children involved in this case by the financial situation in which the benefit cap has placed their parents.

There is the further point, most clearly articulated by Lord Reed at para 89 of his judgment, that the children living with lone parent fathers suffer just as much as the children living with lone parent mothers. Their welfare cannot therefore be relevant to justifying the discrimination between them. However, for the reasons explained in para 189 earlier, this point does not arise when the discrimination complained of is indirect rather than direct. It is of the nature of indirect discrimination that the measure in question applies to both men and women. What has to be considered is whether the measure itself, which in this case I take to be the benefit cap as it applies to lone parents, can be justified independently of its discriminatory effects. In considering whether that measure can be justified, I have no doubt at all that it is right, and indeed necessary, to ask whether proper account was taken of the best interests of the children affected by it.

Both the Divisional Court and the Court of Appeal concluded that the Government had complied with its obligation to treat the best interests of the children concerned as a primary consideration (paras 75 and 49, respectively). They were, of course, correct to say that “the Government was keenly aware of the impact the benefits cap would be likely to have on children” (Court of Appeal, para 74(2)). But it does not follow from that that the “the rights of children were, throughout, at the forefront of the decision-maker’s mind” (para 75, emphasis supplied). Still less does it follow that their best interests were being treated as a primary consideration. In agreement with the powerful judgments of Lord Carnwath and Lord Kerr on this point, it is clear to me that they were not.

The Government’s contention was that “the long term shift in welfare culture”, or “reversing the impact of benefit dependency on families and children”, would be beneficial to children in the longer run. This may well be so, although it is interesting how little prominence was given to this aspect of the matter in the justifications put forward by the Government for their policy. But in any event, this is to misunderstand what article 3(1) of the UNCRC requires. It requires that first consideration be given to the best interests, not only of children in general, but also of the particular child or children directly affected by the decision in question. It cannot possibly be in the best interests of the children affected by the cap to deprive them of the means to provide them with adequate food, clothing, warmth and housing, the basic necessities of life. It is not enough that children in general, now or in the future, may benefit by a shift in welfare culture. Insofar as the Secretary of State relies upon this as an answer to article 3(1), he has misdirected himself.

It may be worth noting that the UNCRC contains some specific obligations which go beyond treating children’s interests as a primary consideration when making decisions concerning them. Article 27(1) provides that “States Parties recognise the right of every child to a standard of living adequate for the child’s physical, mental, spiritual, moral and social development”. Although parents have “the primary responsibility to secure, within their abilities and financial capacities, the conditions of living necessary for the child’s development” (article 27(2)), States Parties have to “take appropriate measures to assist parents and others responsible for the child to implement this right and shall in case of need provide material assistance and support programmes, particularly with regard to nutrition, clothing and housing” (article 27(3)). The usual approach of the Strasbourg court is that the Convention confers no right to be provided with any particular welfare benefit but that, if it is provided, it must be provided in a non-discriminatory manner. The United Kingdom performs its obligations towards children, among other ways, through the welfare benefits system, which provides specific benefits in order that children shall be free from want. The benefit cap deprives some children, principally those in larger families living in high cost accommodation, of provision for their basic needs in order to incentivise their parents to seek work, but discriminates against those parents who are acknowledged to be least likely to be able to do so. The children affected suffer from a situation which is none of their making and which they themselves can do nothing about.

This case is therefore very different from the case of Humphreys v Revenue and Customs Commissioners [2012] UKSC 18, [2012] 1 WLR 1545, in which the Government had to justify the discriminatory effect of paying child tax credit to the parent with the main responsibility for looking after the child, even though the care of the child was shared with another parent. This was indirectly discriminatory against fathers, but the object was to concentrate the help for the child where it was most needed and to maximise the amount of public money available to support children. As the Government put it, “the benefit attaches to the child rather than the parent” (para 25).

Viewed in the light of the primary consideration of the best interests of the children affected, therefore, the indirect discrimination against women inherent in the way in which the benefit cap has been implemented cannot be seen as a proportionate means of achieving a legitimate aim. Families in work are already better off than those on benefits and so the cap is not necessary in order to achieve fairness between them; saving money cannot be achieved by unjustified discrimination; but the major aim, of incentivising work and changing the benefits culture, has little force in the context of lone parents, whatever the age of their children. Depriving them of the basic means of subsistence cannot be a proportionate means of achieving it.

The claimants seek both declaratory relief and an order quashing Part 8A of the Housing Benefit Regulations. The latter would not be appropriate, given that it is not suggested in this case that the implementation of the cap in relation to single person and two parent households is incompatible with the Convention rights. It is the implementation in relation to lone parents, some of whom will be fleeing domestic violence, and their dependent children, which has been shown to be incompatible.

There are several different ways in which that incompatibility might be cured, most notably perhaps by taking the child tax credit and/or child benefit payable to lone parents out of the list of welfare benefits taken into account in calculating the cap. It is true, of course, that the Government resisted amendments to take housing benefit, child benefit and child tax credit out of the cap, on the ground that this would be to emasculate its policy objectives. It is easy to see how this might be so, if it were done for all claimants. But it has not been shown that taking the child-related benefits out of the cap as it applies to lone parents would do so. In any event, it is obvious that there is sufficient flexibility in the statutory scheme to enable appropriate solutions to be crafted. It is not for this court to suggest any particular way in which the problem might be solved.

In my view, therefore, the appropriate relief would be a declaration that Part 8A of the Housing Benefit Regulations is incompatible with the Convention rights in that its application to lone parents is indirectly discriminatory on grounds of sex, contrary to article 14 of the Convention read with article 1 of the First Protocol.

As Lord Hughes has observed, there is much common ground among the members of the panel about the issues that arise on this appeal. He has helpfully outlined the areas of agreement in para 135 of his judgment. I am also in broad agreement with virtually all of Lord Carnwath’s judgment (except as to outcome) and am in complete agreement with Lady Hale that the appeal should be allowed for the reasons that she has given. On one view, therefore, there is nothing to be gained from my contributing further to the debate. But I have changed the view that I originally held about the direct effect of article 3 of UNCRC and wish to explain why. If I am wrong in my revised view, there remain two particular issues which separate the majority from Lady Hale’s approach (which I would favour as an alternative to my principal conclusion) that I believe are of vital importance and which have implications well beyond this appeal. For that reason, I feel constrained to say something of them as well.

The two issues are these: (i) if article 3 does not have direct effect, what is the use to which it may be put in considering the proportionality of a measure which interferes with a Convention right; and (ii) whether there is a sufficient identity of interest between a child and her or his lone parent so as to render discrimination against the child discrimination against the parent. Before turning to those issues, however, I wish to begin by examining the role of unincorporated treaties.

Two dominant principles have traditionally restricted the use of international treaties in British domestic law. The first is that domestic courts have no jurisdiction to construe or apply treaties which have not been incorporated into national law; that they are effectively non-justiciable. The second is that such treaties, unless incorporated into domestic law, are not part of that law and therefore cannot be given direct effect to create rights and obligations under national or municipal law. This is a matter of constitutional orthodoxy. It underpinned the series of decisions in which courts consistently refused to give effect to Convention rights before the coming into force of the Human Rights Act 1998. See, for instance, R v Secretary of State for the Home Department, Ex p Brind [1991] 1 AC 696, 747-748, Lord Bridge of Harwich and at 762B-D Lord Ackner; NALGO (1992) 5 (Admin) LR 785, 798, Re McKerr [2004] UKHL 12; [2004] 1 WLR 807: Lord Nicholls of Birkenhead at para 25, Lord Steyn at para 48, Lord Hoffmann at para 63, Lord Rodger of Earlsferry at para 80 and Lord Brown of Eaton-under-Heywood at para 90.

Perhaps the high water mark of the dualist conception of the restriction on the use of international law was reached in J H Rayner (Mincing Lane) Ltd v Department of Trade and Industry (“International Tin Council”) [1990] 2 AC 418. The House of Lords reaffirmed the two principles of non-justiciability and no direct effect. This was on the basis that domestic courts had no competence in respect of the legal relations between sovereign states, nor was the royal prerogative reviewable. At 499F/500C Lord Oliver of Aylmerton said:

Of course the prerogative can now be reviewed, in appropriate circumstances – see, for instance, Council of Civil Service Unions v Minister for the Civil Service [1985] AC 374; R (Abbasi) v Secretary of State for Foreign and Commonwealth Affairs [2002] EWCA Civ 1598, [2003] UKHRR 76. The conduct of foreign affairs, including the making of treaties is still considered to be beyond the reach of judicial review, however. In R (Campaign for Nuclear Disarmament) v Prime Minister of the United Kingdom [2002] EWHC 2759, 126 ILR 727 the High Court held that domestic courts will not determine the meaning of an international instrument (in this case a UN Security Council Resolution) operating purely on the plane of international law. It was said that the only cases in which the court would pronounce on an issue of international law are those where it is necessary to do so in order to determine rights and obligations under domestic law, so as to “draw the court into the field of international law” (at paras 36-40, 47(i)).

Despite the seemingly comprehensive ban on the use by the courts of unincorporated international treaties to recognise rights on the domestic law plane, there are three possible ways which have been considered by the courts in which such treaties may have an impact on national law – (i) as an aid to statutory interpretation; (ii) as an aid to development of the common law; and (iii) as a basis for legitimate expectation.

Where a legislative provision is ambiguous there is a presumption that Parliament intended to legislate in a manner which does not involve breach of international treaty obligations: Salomon v Commissioners of Customs and Excise [1967] 2 QB 116, 143E-G; Garland v British Rail Engineering Ltd [1983] 2 AC 751, 771 (Lord Diplock). See also Sir John Laws [1993] PL 58, 83. While New Zealand allows non-ambiguous legislation to be ‘read down’, or additional words to be read in for the purpose of consonance with international treaties (eg Sellers v Maritime Safety Inspector [1999] 2 NZLR 44, CA), this is not currently the case in the UK – see Collco Dealings Ltd v Inland Revenue Commissioners [1962] AC 1, 19; Quazi v Quazi [1980] AC 744, 808D-E; JH Rayner (Mincing Lane) Ltd v Department of Trade and Industry [1990] 2 AC 418, 481F-H, 500E (the International Tin Council case); R v Secretary of State for the Home Department, Ex p Brind [1991] 1 AC 696, 747-749, 760D-G; Brown [1994] 1 AC 212, 256E-F; J A Pye (Oxford) Ltd v Graham [2002] UKHL 30; [2003] 1 AC 419, 65; R v Lyons (Isidore) [2002] UKHL 44, [2003] 1 AC 976, 13; Boyce v The Queen [2004] UKPC 32, [2005] 1 AC 400, 25 and 81; Abdirahman v Secretary of State for Work and Pensions [2007] EWCA Civ 657, [2008] 1 WLR 254, at 35-40.

But the presumption of compatibility of domestic legislation with international law is well established. A recent example is to be found in Assange v Swedish Prosecution Authority [2012] 2 AC 471 where at para 122 Lord Dyson said: “there is no doubt that there is a ‘strong presumption’ in favour of interpreting an English statute in a way which does not place the United Kingdom in breach of its international obligations”.

It is clear that unincorporated treaties may have a bearing upon the development of the common law: Lyons [2002] UKHL 44, [2003] 1 AC 976, 13 per Lord Bingham. Developments of the common law should ordinarily be in harmony with the UK’s international obligations: A v Secretary of State for the Home Department (No 2) [2005] UKHL 71; [2006] 2 AC 221 Lord Bingham at para 27. Unincorporated treaties may also be used to resolve ambiguities in the common law: Derbyshire County Council v Times Newspapers Ltd [1993] AC 534. See also Director of Public Prosecutions v Jones (Margaret) [1999] 2 AC 240 Lord Slynn at 265D-F, Lord Hope at 277E-278F: reference to the ECHR for guidance was found to be inappropriate in context as there was no doubt about the content of the common law. By implication, at least, where such doubt is present, consideration of an international convention or treaty such as ECHR would be appropriate in order to determine what the common law position is or should be.

The proposition that the common law cannot be used to incorporate treaties “through the backdoor” has, however, been reasserted in, for instance, A v Secretary of State for the Home Department (No 2) [2004] EWCA Civ 1123; [2005] 1 WLR 414 Laws LJ at paras 266-267, Neuberger LJ at para 434.

In Chundawadra v Immigration Appeal Tribunal [1998] Imm AR 161 it was argued that every citizen had a legitimate expectation that, if the ECHR was relevant to a matter under consideration, the Minister would take it into account when deciding how to exercise his powers. The Court of Appeal refused to accept this argument, holding that it was not appropriate to introduce the Convention into domestic law by the back door in this way.

Arguments based on the Australian authority of Minister of State for Immigration and Ethnic Affairs v Teoh (1995) 183 CLR 273; 128 ALR 353 were considered by the Court of Appeal in Behluli v Secretary of State for the Home Department [1998] Imm AR 407, at para 415. The court expressly refused to follow Teoh; it held that mere ratification of a treaty could not generate a legitimate expectation that the treaty would be followed. Two months later, however, a different division of the Court of Appeal indicated a willingness to adopt and follow Teoh in relation to decisions taken under the Royal Prerogative. In R v Secretary of State for the Home Department, Ex p Ahmed and Patel [1998] INLR 570, the Court of Appeal held that the entering into a treaty by the Crown could give rise to a legitimate expectation because, subject to any indication to the contrary, ratification amounted to a representation that the Crown would act in accordance with the obligations imposed on it by the treaty in question. The High Court followed this approach in R v Uxbridge Magistrates' Court, Ex p Adimi [2001] QB 667, DC, (Simon Brown LJ at para 686, Newman J at paras 690-691), although apparently without having Chundawadra or Behluli cited to it.

In the High Court in R v Director of Public Prosecutions, Ex p Kebilene [2000] 2 AC 326 Lord Bingham CJ rejected an attempt to base a legitimate expectation on the ratification of the ECHR, observing that ratification took place nearly half a century ago at a time when it was generally assumed that ratification would have no practical effect on British law or practice. (This view was endorsed by the House of Lords). Laws LJ at pp 353-356, agreeing with Lord Bingham, referred to what had by then become the somewhat hackneyed theme that the legitimate expectation argument would effectively introduce the ECHR through the back door. He acknowledged, however, that the Convention had “plainly” informed the common law and he noted Teoh but suggested that any extension of this area would have to be at a higher level, to overcome the House of Lords authority of Brind.

The proposition that the doctrine of legitimate expectation can generate a right to rely on the provision of an unincorporated treaty in the interpretation and application of domestic law is, at least, controversial. But treaties concerning human rights are, for reasons that I will develop, in a different position.

A small opening for an exception in relation to human rights treaties can perhaps be seen in Lewis v AG of Jamaica [2001] 2 AC 50 PC, where Lord Slynn, although upholding the traditional principles of non-justiciability and no direct effect, acknowledged the argument that an exception might be read into these rules when the treaty in question was a human rights treaty: “even assuming that that [rule] applies to international treaties dealing with human rights ...”: p 84. In Foreign Relations and the Judiciary (2002) 51 ICLQ 485, 496 Lord Collins has commented on this passage:

In Re McKerr Lord Steyn cast doubt on the applicability of the fundamental principles set out in International Tin Council so far as they governed the position in relation to human rights treaties, arguing that “the rationale of the dualist theory, which underpins the International Tin Council case, is that any inroad on it would risk abuses by the executive to the detriment of citizens. It is, however, difficult to see what relevance this has to international human rights treaties which create fundamental rights for individuals against the state and its agencies. A critical re-examination of this branch of the law may become necessary in the future” at paras 49-50.

While acknowledging that the point had not been the subject of argument, Lord Steyn referred to some academic criticism of International Tin Council and highlighted what he termed “growing support for the view that human rights treaties enjoy a special status”, citing the views of Murray Hunt (Using Human Rights Law in English Courts (1998)), at pp 26-28) and the extra-judicial comments of Lord Collins quoted above.

In International Law in Domestic Courts: The Developing Framework (2008) 124 LQR 388 Philip Sales and Joanne Clement attack this argument, pointing out that the rationale for International Tin Council is that the Crown cannot change domestic law by the exercise of prerogative powers as this would infringe the sovereignty of Parliament. In adopting what might be regarded as a somewhat absolutist position, Sales and Clement argue at para 388:

The Sales and Clement article provides a comprehensive survey of international law in this area. They argue forcefully that unincorporated treaties should not be extended so as to have direct effect in national law. The dualist structure of our law, which treats international law as operating on a separate plane, has, they suggest, been repeatedly upheld as a central constitutional, legal and political principle. They conclude at 421:

In an article entitled Human Rights Treaties in the English Legal System published in [2011] PL, 554-576, Dr Bharat Malkani has challenged the central thesis of Sales and Clement. He argues that one needs to question why Parliament should be treated as the “proper locus of law-making … power”. Dr Malkani suggests that the enactment of the Human Rights Act and the incorporation of ECHR into domestic law brought about a change in the constitutional order and that parliamentary sovereignty is no longer the principal basis of the British constitution. This was, rather, the rule of law. On this basis he argues that the constitutional principle of parliamentary sovereignty does not require that international conventions on human rights be “transformed” into domestic law in order to create rights, citing Alan Brudner “The Domestic Enforcement of International Covenants on Human Rights: A Theoretical Framework” (1985) 35 University of Toronto Law Journal. Brudner propounds a theory which would be regarded as highly radical in UK law to the effect that a convention while in origin a treaty between independent states, is in content “the legislation of a universal community of rational beings”. On this account, he argues that since international conventions on human rights articulate principles rationally connected to the common good, they do not require to be transformed into national law.

In light of the authorities that I have earlier considered, it may safely be said that such a far - reaching approach is unlikely to find favour in the courts of this country. It is perhaps noteworthy, however, that other commentators have been critical of the courts’ adherence to the dualist theory of international law, especially in relation to human rights conventions – see, for instance, Brice Dickson, “Safe in Their Hands? Britain’s Law Lords and Human Rights” (2006) 26 Legal Studies 329, 335; D Beyleveld “The concept of a human right and incorporation of the European Convention on Human Rights” [1995] PL 577; M Hunt “Using Human Rights Law in English Courts” (Oxford: Hart, 1997).

I consider that the time has come for the exception to the dualist theory in human rights conventions foreshadowed by Lord Slynn in Lewis and rather more firmly expressed by Lord Steyn in Re McKerr to be openly recognised. This can properly be done in relation to such conventions without espousing the complete abandonment of the theory advocated by some of the commentators referred to above.

If Lord Steyn is right, as I believe he is, to characterise the rationale for the dualist theory as a form of protection of the citizen from abuses by the executive, the justification for refusing to recognise the rights enshrined in an international convention relating to human rights and to which the UK has subscribed as directly enforceable in domestic law is not easy to find. Why should a convention which expresses the UK’s commitment to the protection of a particular human right for its citizens not be given effect as an enforceable right in domestic law?

Standards expressed in international treaties or conventions dealing with human rights to which the UK has subscribed must be presumed to be the product of extensive and enlightened consideration. There is no logical reason to deny to UK citizens domestic law’s vindication of the rights that those conventions proclaim. If the government commits itself to a standard of human rights protection, it seems to me entirely logical that it should be held to account in the courts as to its actual compliance with that standard. This is particularly so in the case of UNCRC. On its website UNICEF has stated that: “The CRC is the basis of all of UNICEF’s work. It is the most complete statement of children’s rights ever produced and is the most widely-ratified international human rights treaty in history”.

I therefore consider that article 3(1) of UNCRC is directly enforceable in UK domestic law. A primacy of importance ought to have been given to the rights of children in devising the regulations which bring the benefits cap into force. For the reasons given by Lady Hale, I have concluded that this has not taken place.

In the Court of Appeal Lord Dyson MR said at para 69:

One starts therefore with the proposition that UNCRC is, as Lord Carnwath has put it, legally relevant. Its legal relevance stems from the fact that, as again Lord Carnwath has put it, under ECHR and in accordance with the Vienna Convention, regard may be had to principles of international law, including international conventions in order to interpret the “terms and notions of the Convention” – Demir v Turkey (2008) 48 EHRR 1272. Lord Carnwath has said that in the cases of X v Austria (2013) 57 EHRR 405, Ponomaryov v Bulgaria (2011) 59 EHRR 799 and Burnip v Birmingham City Council [2013] PTSR 117, the court used international materials to “fill out, or reinforce, the content of a Convention article” (para 130). I would prefer to put the matter slightly differently.

What the courts did in those cases, following the Demir approach was to recognise that the nature and content of Convention rights could be informed by international instruments which expressed standards that were internationally recognised. This does not involve directly applying the provisions of an international treaty which had not been incorporated into domestic law. It does not introduce those provisions “by the back door”. Rather, it reflects the courts’ obligation, charged as we are with the duty of obtaining a proper understanding of the nature of an avowed right, to have regard to standards which have found expression in those treaties. We should do this for the prosaic but extremely important reason, as I have said in para 256 above, that they have been the product of extensive and, hopefully, enlightened consideration.

If the rights enshrined in those treaties are not directly enforceable in domestic law it is, of course, open to domestic courts to refuse to allow such treaties to have any influence whatever on our conclusions as to the content of the right. Such an approach would be justified where, for instance, the right was too broadly expressed or too remote from the subject under consideration. Or we could conclude that the right was too ambivalently stated to allow any influence to be brought to bear on the content of a right asserted under domestic law. But where the claimed right is directly relevant to the domestic issue to be decided, then recourse to the standards that the international instrument exemplifies is not only legitimate, it is required. How, otherwise, are we to acquire a true understanding of the proper contours and content of the right under discussion? This is not applying an unincorporated international treaty directly to domestic law. It is merely allowing directly relevant standards to infuse our thinking about what the content of the domestic right should be.

Article 3(1) of UNCRC is unquestionably directly relevant to the question of whether a primacy of importance was given to the interests of children in formulating the regulations which give effect to the benefits cap. As I have already said, I agree with Lady Hale that it was not. I will say no more on that topic. The critical issue now is whether there is a sufficient connection between the interests of the children and those discriminated against, viz their lone mothers, to make discrimination against the children of those mothers discrimination against them also. Put another way, as Lord Carnwath does in para 115 of his judgment, is there a direct link between the international treaty relied on and the particular form of discrimination alleged?

In this case the government accepts that the benefits cap discriminates against women as lone parents. Its defence of this admitted discrimination rests exclusively on its claim that it is justified by the social objectives which it pursues. It claims, however, that justification of those objectives does not require it to give primary consideration to the impact of the benefit cap on the children of lone mothers. That, the government says, is because the interests of lone mothers can be disassociated from those of their children. Lord Carnwath has accepted this argument. He considers that the interests of children are distinct from their single parent mothers. I cannot agree.

The particular species of discriminatory impact here is on women who, by reason of their position as lone mothers, claim to suffer a disproportionate interference with their Convention rights. Justification of the interference must be related to the condition which provides the occasion for the discrimination viz the position of these women as lone parents. A mother’s personality, the essence of her parenthood, is defined not simply by her gender but by her role and responsibility as a carer of her children, particularly when she is a lone parent.

Justification of a discriminatory measure must directly address the impact that it will have on the children of lone mothers because that impact is inextricably bound up with the women’s capacity to fulfil their role as mothers. If you take money away from children which mothers would receive on their behalf, money which they use to realise their role as mothers, the discrimination that you perpetrate involves withholding resources necessary to fully discharge their maternal role. Because, therefore, one cannot segregate the interests of the deprived children from those of their mothers, the discrimination against mothers and their children is of the same stripe. No hermetically sealed compartmentalisation of their interests is possible.

A lone mother’s interests, when it comes to receiving state benefits, are indissociable from those of her children. The rate of her benefits is fixed by reference (among other things) to the number and needs of those children. Her capacity to care for her children is likewise directly connected to the amount of state benefits that she receives. The interests of single mothers are, therefore, inextricably bound up with the interests of their dependent children, for the trite and prosaic reason that they are parents. Any adverse impact on a lone parent’s financial position is inevitably transmitted to the child because of her or his dependence (financially as well as otherwise) on the parent. For these reasons, when one comes to consider the justification for interference with a lone parent’s Convention right, the interests of the children of that parent cannot be left out of account.

If the disproportionate effect on lone parents can only be justified by addressing their position as the providers for dependent children, attention to the interests of those children is an integral part of the process. How, otherwise, are their interests to be taken into account? As Lord Reed has said, regard has been had to the UNCRC by the European Court of Human Rights in the application of the ECHR, when considering how its substantive guarantees apply to children. When considering the rights of children as a component part of their mothers’ rights under A1P1 and article 14, there is no reason that UNCRC should not likewise infuse the determination of what the content of those rights should be. I therefore agree with Lady Hale that, in considering whether the particular species of interference in this case is justified, the interests of the children affected are, by reason of article 3(1) of UNCRC, to be treated as a primary consideration.

Once this position is reached, the question for the government is how to meet the challenge of showing that the measures which discriminate against the child (and ergo the mother) are no more intrusive than they need to be. In this context, I have no difficulty in accepting that the test set out in Stec v United Kingdom (2006) 43 EHRR 1017 continues to apply. So, as a yardstick of the proportionality of this general measure of economic or social strategy, the question is whether it was manifestly without reasonable foundation. But, if article 3(1) of UNCRC has to play its part in deciding whether the benefits cap was without reasonable foundation, it requires that first consideration be given to the best interests of the children directly affected by the decision.

For the reasons given by Lady Hale in para 220, it cannot be in the best interests of the children affected by the cap to deprive them of the means of having adequate food, clothing, warmth and housing. Depriving children of (and therefore their mothers of the capacity to ensure that they have) these basic necessities of life is simply antithetical to the notion that first consideration has been given to their best interests.

I would therefore allow the appeal and make the order that Lady Hale proposes.