England and Wales Court of Appeal (Civil Division) Decisions

IN THE COURT OF APPEAL (CIVIL DIVISION)
ON APPEAL FROM THE HIGH COURT OF JUSTICE
QUEEN'S BENCH DIVISION
ADMINISTRATIVE COURT
MR JUSTICE CHAMBERLAIN
[2021] EWHC 102 (Admin)

		Royal Courts of Justice Strand, London, WC2A 2LL
		13/10/2021

Lady Justice Andrews:

Introduction

1. Universal Credit ("UC") was introduced by the Welfare Reform Act 2012 ("the 2012 Act"), replacing a range of existing benefits with one holistic benefit. The 2012 Act, and the detailed regulations which implemented it, the Universal Credit Regulations 2013 (SI 2013/376, as amended by SI 2014/2887) ("the UC Regulations") revolutionised the benefit system, giving effect to a number of statutory aims and policy imperatives. This was a radical reform, intended to simplify the benefits system and make it fairer, and to remove the financial and administrative barriers to work inherent in the legacy welfare system which it replaced.
2. Although one of the fundamental aims of UC is to make work pay to the fullest possible extent, and thereby to encourage more people to see work as the best route out of poverty, any suggestion that the purpose of the reform could be distilled into the single aim of making work pay is apt to oversimplify. Other important aims include the more efficient targeting of financial support to those who most need it, the encouragement of claimants to take personal responsibility for their finances, the achievement of a fairer overall financial burden on the taxpayer, the reduction of fraud and error, and the simplification of the administrative process.
3. The amount of UC received by a claimant will depend on a number of factors which will vary from case to case, depending on individual circumstances. UC is a single payment which comprises a standard allowance (s.9 of the 2012 Act), plus (where applicable) various other elements, such as an amount in respect of responsibility for children (s.10); an amount in respect of housing costs (s.11); and amounts for "other particular needs or circumstances" (s.12), which include a need for payment of childcare costs whilst a single parent or both parents are at work. UC is not a full indemnity. If the claimant is working, it operates as a supplement to their earnings. If the claimant's earnings increase, the amount of UC will be reduced, and vice versa. It is designed to operate flexibly to meet the wide variety of needs and circumstances of those who claim it, including claimants whose earnings fluctuate, and can be adjusted to meet changes in those circumstances as and when they arise.
4. Payment of UC is made in arrears on a monthly basis, in the same way as a monthly salary would usually be paid. It will normally be made within seven days of the last day of the relevant monthly assessment period, which will be fixed by reference to the date on which the individual first makes a claim. That approach applies irrespective of whether the claimant is in or out of work, or moves between the two, and whether his or her earnings are from employment, self-employment, or both. The calculation of entitlement and payment each month creates certainty about what will be received and when. This encourages the recipient to budget on a monthly basis, and is seen as a means of fostering their independence.
5. The system is designed around the core principle of using monthly assessment periods and making a single monthly payment in arrears, after taking into account the claimant's actual earnings received in that period. This means that different elements of UC that feed into the overall calculation of what is due to a recipient cannot be separated out or paid before the end of the assessment period without upsetting the system and reintroducing the complexity and scope for error that it was designed to overcome.
6. This is an appeal by the Secretary of State for Work and Pensions ("the Secretary of State") against the decision of Chamberlain J that the requirement imposed by the UC Regulations on a person claiming the childcare costs element ("CCE") of UC to have already paid for the relevant childcare (i) subjected the claimant, Ms Salvato, and single mothers like her, to unlawful indirect discrimination on grounds of sex, contrary to Article 14 read with Article 8 and/or Article 1 of Protocol 1 to the ECHR, ("A1P1") and (ii) was irrational. The requirement was referred to in the judgment below as "the Proof of Payment Rule" and I will adopt the same terminology, although it is more accurate to describe it as a fact of payment rule.
7. The complaint in this case is not about a rule concerning qualification for entitlement to a particular welfare benefit, or limiting the recoverable amount, but rather, about the mechanism adopted for payment of an element of the benefit, as part of their UC award, to someone who is entitled in principle to claim it. As the Judge explained in paragraphs [2]-[4] of his judgment:

"2. … The effect of the UC Regulations is that a claimant is entitled to be paid the CCE as part of her UC award only if she has already paid the charges, rather than merely incurred them. Claimants therefore have to find ways of paying the charges from their own funds. They will only be reimbursed several weeks afterwards…

3. There is no such rule in relation to another element of UC – the housing costs element ("HCE"). There, provision is made for payment of "an amount in respect of any liability of a claimant to make payments in respect of the accommodation they occupy as their home (s.11 of the 2012 Act and reg 25ff of the UC Regulations). Guidance makes it clear that housing costs can be paid in various different ways, depending on the needs of the claimant, including by direct payment to the landlord.

4. The claimant is a single mother who wishes to work but would be unable to do so without help to cover childcare charges. She is in principle eligible to receive the CCE. She wants to work full-time, but says that, because of the Proof of Payment Rule, she cannot afford to pay the fluctuating costs of childcare and as a result has become indebted and ultimately had to reduce the number of hours she works."

8. When the claimant reduced her working hours to minimise the amount of childcare costs, thereby reducing her monthly income, her UC award became higher than it would have been if she had been able to continue working full-time. Her complaint was that the Proof of Payment Rule created a "Catch-22" situation in which she could not claim the element of the benefit that would enable her to pay for necessary childcare that she was otherwise unable to afford, without first making the payment that she could not afford to make without receiving the benefit. That argument, of course, assumes the absence of alternative financial resources. The Rule was said to be irrational because it undermined the key objective of getting people back to work and decreasing their dependency on benefits.
9. The Judge himself granted permission to appeal on four grounds, namely:

i) The Court erred in finding that the proof of payment rule was indirectly discriminatory;

ii) The Court erred in its approach to objective justification;

iii) The Court ought not to have made a finding that the application to the claimant of Regulation 33(1)(za) of the Universal Credit Regulations 2013 was incompatible with Article 14 (read with Art 8 and/or A1P1 ECHR) and

iv) The Court erred in determining that the choice of a proof of payment rule was irrational.

Subsequently Lewis LJ granted the Secretary of State permission to appeal on a fifth ground, namely, that the Court erred in considering that the Rule fell within the ambit of Article 8 and/or A1P1 so that Article 14 was engaged. He also directed expedition of the appeal.

10. Although there are five grounds of appeal (the first three of which are different facets of the same complaint), the issue at the heart of this case is whether the Proof of Payment Rule is objectively justified in the sense that it is proportionate to the legitimate aims of reducing/eliminating the scope for fraud or error, and achieving a simple, easily workable scheme. The Judge decided that it was not, because although it met the first of the four criteria set out by Lord Reed JSC in Bank Mellat v HM Treasury (No 2) [2013] UKSC 39 [2014] AC 700 at [54], it did not meet the others, which he considered together. His reasoning was based on the premise that a proof of liability rule, akin to that used for the HCE, might meet those aims without requiring claimants to incur debt in order to pay for childcare charges. The Judge's finding that there was a lack of objective justification also fed into his conclusion that the Rule was irrational.
11. I found some of the criticisms that Ms Dobbin QC, on behalf of the Secretary of State, made of the Judge's approach to be misplaced. Much of the Judge's legal analysis is exemplary, and he plainly gave careful consideration to the arguments presented to him. Nevertheless, I have concluded that the Judge did fall into material error when he sought to apply the principles he identified to the evidence in this case, and that there are deficiencies in the reasoning which led him to conclude that the Rule was indirectly discriminatory and irrational.

How the system works

12. The Proof of Payment Rule is set out in Reg 33(1)(za) which provides that:

"The childcare costs condition is met in respect of an assessment period if:

(za) the claimant has paid charges for relevant childcare that are attributable to that assessment period (see regulation 34A)." (Emphasis supplied.)

13. The Rule is further embodied in Reg 34A, which explains what charges are attributable to a particular assessment period. It provides that:

"Charges paid for relevant childcare are attributable to an assessment period where:

(a) Those charges are paid in that assessment period for relevant childcare in respect of that assessment period; or

(b) Those charges are paid in that assessment period for relevant childcare in respect of a previous assessment period; or

(c) Those charges were paid in either of the two previous assessment periods for relevant childcare in respect of that assessment period."

Thus the Regulations only permit claims for relevant childcare (as defined in Reg 35) that is attributable to the assessment period in question (as defined in Reg 34A) and which has already been paid for.

14. "Relevant childcare" comprises care provided during the hours when a single parent is, or both parents are, engaged in paid employment, by a person who is suitably qualified to look after children and formally regulated. It covers care provided by a school proprietor to a child outside school hours, or to a child who has not reached compulsory school age, and care provided to a child by someone who is registered in England under Part 3 of the Childcare Act 2006 or under the Health and Social Care Act 2008 (or registered under the equivalent regulatory regimes in Wales and Scotland). It does not extend to care provided by a grandparent or other family member, even if that person is paid.
15. It is immaterial whether the payment was made in advance of the provision of the childcare (as many providers require) or afterwards. There was a dispute between the parties as to whether the effect of the Regulations (particularly Reg 35) is that the CCE will only be received by a claimant after the childcare has actually been provided, or whether it could also be claimed in circumstances where, for example, the provider requires payment in advance, which is non-refundable, but the child then falls ill and cannot attend. For the purposes of this appeal it is unnecessary to resolve that dispute; on any view, the claimant must have paid for the childcare before he or she can claim the CCE.
16. Reg 34(1) caps the amount of the CCE at "85% of the charges paid for relevant childcare that are attributable to that assessment period" up to prescribed maximum amounts – currently £646.35 per month for a single child, or £1,108.04 for two or more children. This percentage is higher than the 70% that was previously recoverable under the legacy system. If the childcare provided is split over two assessment periods, the amount paid in respect of any assessment period is calculated by reference to a formula set out in Reg 34(2).
17. In order to work out how much the claimant will receive by way of UC for a particular assessment period, the standard allowance plus the various elements to which they are entitled for that assessment period (including the CCE where applicable) will be added together to produce a total entitlement figure before deductions. Deductions will then be made from that figure in respect of unearned or earned income – in the latter case, the claimant's take-home pay in that period above a certain amount which is disregarded. Every £1 earned over the disregarded amount will reduce the amount of UC by 63 pence. The recoverable amount of UC will therefore depend on the claimant's earnings received in the relevant assessment period; the greater the earnings, the lower the sum recoverable.
18. A claimant's entitlement in any assessment period is subject to the so-called "benefit cap", which is set at an amount varying according to the characteristics of that individual. In order to encourage claimants to work, Reg 82 provides for the disapplication of the benefit cap in cases where their earned income in an assessment period is equal to or exceeds a minimum amount. The benefit cap has no relevance to the issues that arise in the present case.
19. The system for the payment of UC is automated and calculated on the basis of information provided by the applicant. This is done in an online "journal" which will prompt the applicant to upload supporting evidence of payment (e.g. bank statements or receipts from the provider). The system also enables the applicant to have an online dialogue with someone in a service centre who will provide assistance where required, and who may ask them for further information (e.g. to enable the provider of the childcare to be verified as a regulated provider).
20. In practical terms, a claimant must find the funds to make the initial payment for childcare (in the first assessment period). Thereafter, the payment of the UC award, which includes the CCE for the previous period, should generate sufficient funds to enable them to afford to pay for childcare in the next assessment period, despite the fluctuations in the amount of childcare that may be required from month to month or week to week. Leaving aside any difficulty of affording to make the initial payment, an inability to pay upfront is most likely to be experienced in circumstances where there is a significant increase in the amount of childcare required in the current assessment period – for example at the start of the school holidays in the summer.
21. The childcare covered by the CCE is in addition to the maximum 30 hours of free childcare available to all working parents, outside the UC system. Tax-free childcare is available to those who do not qualify for UC, under a separate scheme. Unlike the position under the legacy system, there is no minimum hours worked threshold. The evidence of Ms Niamh Parker, a UC policy team leader in the Department for Work and Pensions ("DWP"), is that the CCE was intended to provide a work incentive and give confidence to parents, particularly single parents (most of whom are women) to take up employment where their personal circumstances may not allow them to work a minimum of 16 hours a week.
22. Chamberlain J's exposition of the policy underlying the legislation (at [28]-[36]), and his synopsis of the factual evidence relied on by Ms Salvato (at [37]-[64]) and by the Secretary of State (at [65]-[87]) cannot be improved upon. It is unnecessary to repeat those passages of the judgment below, which I adopt in their entirety; I will simply refer to salient aspects when I come to address the issues to which they relate.

The relevant legal principles

23. Article 14 ECHR provides, so far as material that:

"the enjoyment of the rights and freedoms set forth in this Convention shall be secured without discrimination on any ground such as sex, race, colour, language, religion… or other status."

In order to show a breach of Article 14, it is sufficient that the facts fall within the ambit of another substantive Convention right; the claimant need not establish a violation of that other right. In the present case, Ms Salvato principally relies upon the right to respect for family life under Art 8, although she also contends that A1P1 is engaged.

24. In order to determine whether a measure is incompatible with Article 14, four questions need to be answered:

(1) does the alleged discrimination concern the enjoyment of a right set out in the ECHR – in other words, do the facts fall within the ambit of a Convention right?

(2) has the claimant been treated less favourably than a class of persons whose situation is relevantly similar?

(3) is the difference in treatment on the ground of a "status" recognized under Article 14?

(4) is there an objective and reasonable justification for the difference in treatment?

The answers to the third and fourth questions determine whether there is "discrimination" within the meaning of Art 14.

25. Art 14 extends to measures which, on their face, apply equally to all, but which have a disproportionately prejudicial effect upon members of a particular group. The most recent authoritative statement of the nature of indirect discrimination in the context of Art 14 is to be found in the judgment of Lord Reed (with which the other members of the Supreme Court agreed) in R (SC) v Secretary of State for Work and Pensions [2021] UKSC 26, [2021] 3 WLR 428, ("SC").
26. Like the present case, SC was a case of alleged indirect discrimination on grounds of sex, in the field of welfare benefits. It concerned a challenge to an aspect of one of the legacy benefits replaced by UC, child tax credit, the maximum amount of which was calculated without reference to any third or subsequent child born to a claimant after a specified date. The challenge was dismissed on the basis that the measure was objectively justified. The decision of the Supreme Court in SC post-dated Chamberlain J's judgment, but it was consistent with the approach taken by the Court of Appeal, (reported as R (C) v Secretary of State for Work and Pensions [2019] EWCA Civ 615, [2019]1 WLR 5687) in which the leading judgment was given by Leggatt LJ.
27. At [37] Lord Reed set out the general approach to Art 14 adopted by the European Court of Human Rights ("ECtHR"), which he distilled into the following four propositions:

i) only differences in treatment based on an identifiable characteristic, or 'status', are capable of amounting to discrimination within the meaning of Art 14;

ii) in order for an issue to arise under Art 14 there must be a difference in the treatment of persons in analogous, or relevantly similar, situations;

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

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

28. At [49] Lord Reed quoted a passage from the judgment of the ECtHR in Guberina v Croatia (2018) 66 EHRR 11 as follows:

"The court has also accepted that a general policy or measure that has disproportionately prejudicial effects on a particular group may be considered discriminatory notwithstanding that it is not specifically aimed at that group, and that discrimination potentially contrary to the Convention may result from a de facto situation. This is only the case, however, if such policy or measure has no 'objective and reasonable' justification, that is, if it does not pursue a 'legitimate aim' or if there is not a 'reasonable relationship of proportionality' between the means employed and the aim sought to be realised."

As Lord Reed observed, that statement is itself derived from the judgment of the Grand Chamber in DH v Czech Republic (2008) 47 EHRR 3. He continued:

"This is what is described in the Convention case law as 'indirect discrimination'. It can arise in a situation where a general measure or policy has disproportionately prejudicial effects on a particular group. It is described as 'indirect' discrimination because the measure or policy is based on an apparently neutral ground, which in practice causes a disproportionately prejudicial effect on a group characterised by a salient attribute or status."

29. After illustrating how the concept of indirect discrimination was developed in the ECtHR by reference to the decisions of the Court in Hoogendijk v The Netherlands (2005) 40 EHRR SE22, DH (above), and S.A.S. v France (2014) 60 EHRR 11, Lord Reed summarised the approach to be taken at [53]:

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

30. Lord Reed went on to consider a number of preliminary issues. These included whether, in the context of assessing the proportionality of a particular measure to a legitimate aim, the approach adopted by the Supreme Court in a number of cases, of respecting the policy choice of the executive or legislature in relation to general measures of economic or social strategy in the context of welfare benefits unless it was "manifestly without reasonable foundation," accurately reflected the approach of the ECtHR and should continue to be followed. He described this as "a difficult and important question which required careful consideration."
31. After a masterly analysis of the European case law, he summarised the approach of the ECtHR at [142]:

"In summary, the European court has generally adopted a nuanced approach, which can be understood as applying certain general principles, but which enables account to be taken of a range of factors which may be relevant in particular circumstances, so that a balanced overall assessment can be reached. As I have explained, there is not a mechanical rule that the judgment of the domestic authorities will be respected unless it is "manifestly without reasonable foundation". The general principle that the national authorities enjoy a wide margin of appreciation in the field of welfare benefits and pensions forms an important element of the court's approach, but its application to particular facts can be greatly affected by other principles which may also be relevant, and of course by the facts of the particular case.…

In the context of article 14, the fact that a difference in treatment is based on a "suspect" ground is particularly significant. The recent cases of JD, Jurčič and Yocheva and Ganeva, like many earlier cases, indicate the general need for strict scrutiny focused on the requirement for very weighty reasons, where the difference in treatment is based on a suspect ground such as sex or birth outside marriage, unless the issue concerns the timing of reform designed to address historical inequalities, where a wider margin is likely to be appropriate."

32. Lord Reed then went on to consider the domestic context, explaining that domestic courts have generally endeavoured to apply an analogous approach to that of the ECtHR, and why that is appropriate. At [145] he said that in domestic law, as at the Strasbourg level, one would expect closer scrutiny where the case concerns discrimination on a "suspect" ground such as sex or race, rather than a difference in treatment on less sensitive grounds, especially if it is simply a by-product of a legitimate policy. Distinctions drawn on "suspect" grounds are inherently appropriate for close judicial scrutiny, notwithstanding the respect due to the judgment of the executive or the legislature. He pointed out at [146] that the administrative law test of unreasonableness is generally applied in contexts such as economic policy and social policy with considerable care and caution; and the same is true of the Convention test of proportionality. Both tests have to be applied in a way which reconciles the rule of law with the separation of powers.
33. Lord Reed went on to explain that the balanced approach which he had described was adopted by the Supreme Court in the early case of Ghaidan v Godin-Mendoza [2004] UKHL 30, [2004] 2 AC 557, but that the reasoning on justification adopted in the case of Humphreys v HMRC [2012] UKSC 18, [2012] 1 WLR 1545 and followed by the Supreme Court in subsequent cases, (perhaps articulated most vehemently by Lord Wilson JSC in R (DA) v Secretary of State for Work and Pensions [2019] UKSC 21, [2019] 1 WLR 3289), did not reflect the Strasbourg jurisprudence entirely correctly. The "manifestly without reasonable foundation" formulation, as used in the Strasbourg judgments, does not express a test in the sense of a requirement whose satisfaction or non-satisfaction will in itself necessarily be determinative of the outcome. Rather, that phrase indicates the width of the margin of appreciation, and hence the intensity of the review, which is in principle appropriate in the field of welfare benefits, other things being equal. It does not replace or supersede the requirement for "very weighty reasons" where suspect grounds are in issue. Instead, the degree of deference usually appropriate in relation to social or economic policy choices may have to be taken into account in assessing whether very weighty reasons have been shown.
34. Lord Reed concluded that the "manifestly without reasonable foundation" formulation still had a part to play, but that the approach which the Court had followed since Humphreys should be modified in order to reflect the nuanced nature of the judgment which is required. He stressed the importance of avoiding a mechanical approach based on the categorisation of the ground of the difference in treatment. A more flexible approach will give appropriate respect to the assessment of democratically accountable institutions, but will also take appropriate account of such other factors as may be relevant. The Courts should generally be very slow to intervene in areas of social and economic policy such as housing and social security, but as a general rule, differential treatment on grounds such as sex or race nevertheless requires cogent justification.
35. The measure under challenge in the present case is a statutory instrument; the relevant provisions of the UC Regulations were subject to the affirmative resolution procedure (in their original version) and the made negative procedure (in their amended version): see the judgment at [171]. In Bank Mellat (above) Lord Sumption said this at [44]:

"When a statutory instrument has been reviewed by Parliament, respect for Parliament's constitutional function calls for considerable caution before the courts will hold it to be unlawful on some ground (such as irrationality) which is within the ambit of Parliament's review. This applies with special force to legislative instruments founded on considerations of general policy."

Those observations were specifically endorsed by Lord Reed in R (SG and others) v Secretary of State for Work and Pensions [2015] UKSC 16, [2015] 1 WLR 1449 at [94]. SG was another case in which regulations relating to welfare benefits were challenged on Article 14 grounds.

36. With that guidance in mind, I turn to consider the grounds of appeal in the present case.

Does the alleged discrimination fall within the ambit of a Convention right?

37. Although this issue was the basis of the fifth ground of appeal and was addressed by counsel in that order, logically it must be answered first, because if the Judge was wrong on this point, Article 14 would not be engaged at all, and there would be no basis for an indirect discrimination challenge.
38. Ms Dobbin submitted that the Judge was wrong to find that the Proof of Payment Rule fell within the ambit of Art 8 and A1P1. That was always an ambitious submission, at least so far as Art 8 was concerned, in the light of the analysis of Leggatt LJ in the Court of Appeal in SC (above) particularly at [57], which the Judge had carefully followed. The argument became impossible to sustain in the light of Lord Reed's observations on that issue.
39. At [41] of his judgment in SC, Lord Reed referred to the fact that a number of judgments of the ECtHR confirmed that welfare benefits which were designed to facilitate or contribute to family life, by supporting families with children, are likely to fall within the ambit of Art 8 for the purposes of complaints under that article taken together with Art 14. He said that since child tax credit was payable only to adults who are responsible for children, and was intended to provide financial support to families with children, he did not see any convincing basis for distinguishing it from the benefits with which those cases were concerned, and therefore agreed with the Court of Appeal that the complaint of the adult appellants in that case fell within the ambit of Art 8 taken together with Art 14.
40. In the present case, Chamberlain J approached the issue in the manner suggested by Leggatt LJ, by asking whether the CCE was "a way in which the state shows respect for children and the life of the family of which they are a part". Relying on the Executive Summary of the White Paper which preceded the 2012 Act, he said this at [164]:

"The purpose of the benefit is to break the cycle of worklessness, which blights the life chances of children, as well as their parents. Part of the philosophy underlying the benefit is that having a parent in work changes the way children view work. The benefit enables parents and children to have a different family life – one in which the parent contributes to society by working and the children have a role model who lives a productive adult life that they will come to regard as normal and aspire to for themselves. In those circumstances, applying the reasoning in C's case, the benefit is, in my judgment, clearly a measure by which the state shows respect for family life."

41. The Judge was entitled to take that view for the reasons that he gave, although, if he had had the benefit of Lord Reed's judgment in SC he might have reached the same conclusion by the more direct route of finding that the CCE was a benefit payable only to adults with children, and intended to provide financial support to families with children.
42. The Judge found that the complaint also fell within the ambit of Art 14 taken together with A1P1, on the basis that this conclusion followed from the reasoning of the Grand Chamber in its admissibility decision in Stec v United Kingdom (2005) 41 EHRR para 54:

"… The relevant test is whether, but for the condition of entitlement about which the applicant complains, he or she would have a right, enforceable under domestic law, to receive the benefit in question."

Ms Dobbin sought to distinguish the present case on the basis that the Proof of Payment Rule was not concerned with a condition of entitlement to a particular welfare benefit or an element of it, but with the mechanism and timing of payment of an element of the benefit to which the claimant was entitled in principle. The Judge was not persuaded that this was a valid distinction; he held that the Stec test was satisfied, finding that but for the Proof of Payment Rule, domestic law would have given the claimant an enforceable right to receive the CCE. He pointed out that the CCE does not depend on the exercise of a discretion.

43. The CCE is a contribution, made in arrears, towards the cost of childcare provided during working hours to enable the claimant to work. Whilst I appreciate that the Proof of Payment Rule could be characterised as a condition of entitlement, in the sense that only someone who has already paid for the childcare is able to receive the CCE, it is not a condition of entitlement in the sense in which that expression was used in Stec. There is no doubt that Ms Salvato qualifies to claim the childcare element in relation to her employment. The Rule does not delineate who qualifies in principle to claim the CCE, and it is artificial to characterise it as doing so. Rather, it accurately establishes what childcare was provided, by whom and when, and how much it cost. Without the Proof of Payment Rule, there would have to be some other mechanism by which a claimant proved that the childcare had been provided and how much it cost. Without that alternative mechanism, they would not be able to claim the CCE.
44. Therefore, although I appreciate the force of the argument on A1P1, I am more doubtful than the Judge was that the alleged discrimination fell within its ambit. However, it is unnecessary to reach a firm conclusion about this matter, because he was undoubtedly right about Art 8, and it was only necessary for the claimant to show that the alleged discrimination fell within the ambit of one Convention right to engage Art 14.

Does the Proof of Payment Rule have a disproportionate prejudicial effect on women?

45. The evidence before the court established that, as at August 2019, the vast majority of people whose UC payment included the CCE were single parents (41,928 of 50,269 households) and 81% of that cohort were single mothers (40,690 of the 41,928). This is hardly surprising, as CCE was specifically targeted at assisting that group into work, and eliminating the disadvantages they faced compared with couples. The fact that the Proof of Payment Rule (and therefore any inherent disadvantage it produces) affects far more women than men is simply a function of the fact that more women than men claim the CCE. It does not establish that the Rule has a disproportionate prejudicial effect on women in the sense that they are less able to comply with it because they are women.
46. The Judge was alive to this point. At [155] he acknowledged that the fact that a benefit is disproportionately claimed by a particular group is not enough to make every adverse rule which applies to that benefit indirectly discriminatory against the group in question: see R (Delve) v Secretary of State for Work and Pensions [2020] EWCA Civ 1199, [2021] ICR 236. Thus the fact that most claimants of CCE are women and the fact that a reimbursement model can make it difficult, or even impossible, for some of those who are entitled to the CCE to claim it, because they cannot afford to make the initial payment, would not be enough to establish a disproportionate prejudicial impact on women.
47. Nevertheless, the Judge held at [159] that the Proof of Payment Rule had a disproportionate detrimental effect on women in two ways; first, because the decision to make eligibility for the CCE - but not the HCE - dependent on proof of payment adversely affects all those UC claimants who are in principle entitled to the CCE, of whom more than 80% are women; and secondly because within the group who are in principle eligible for the CCE, the Rule is bound to have a greater adverse effect on women than on men, because women as a group earn substantially less than men as a group and are therefore substantially more likely to be unable to afford to pay the childcare charges out of their own funds before being reimbursed.
48. The first basis upon which the Judge found the Rule discriminatory was the result of a combination of using the wrong comparator, and reliance on the same flawed reasoning as he had already rejected at [155]. The issue is whether a single measure which is neutral, in the sense that it applies to all claimants, has a disproportionately prejudicial effect on women – in other words, whether women are less able to satisfy the Proof of Payment Rule than men are. The prejudicial effect complained of related to the ability of women to claim the CCE, and thereby get into work, not their ability to claim UC. Ms Salvato's ability to claim UC was unimpaired; indeed, as stated earlier, she received a higher award when she reduced her working hours because of her inability to pay for the childcare needed to enable her to work full-time.
49. At [156] the Judge compared the position of recipients of the HCE to recipients of the CCE, stating that "what is really under challenge here is the decision to make an exception for one group of claimants (those claiming HCE) but not another (those claiming CCE)." Instead of concentrating on the rule or measure under challenge, the Judge was comparing two completely different measures. It is no answer to say that they were both components of a single integrated benefit, because each component is aimed at different aspects of a claimant's needs. At this stage of the analysis it was irrelevant whether other elements of UC, targeted at different needs, were subject to the same or a different mechanism of claim and payment. The correct comparator group was men who were eligible to receive the CCE, and not all claimants of UC, let alone claimants of the HCE.
50. Mr Buttler QC sought to rely on the approach to comparators taken by the ECtHR, which is broader than the approach taken under EU Law or under the Equality Act 2010, to justify the first approach taken by the Judge. He relied in particular on the case of Di Trizio v Switzerland (Application No 7186/09) which concerned a benefit that was available to any person who could establish a minimum degree of disability. The complaint was that a particular method of assessment of the degree of disability was applied to part-time workers, but not to full-time workers or to persons who did not work at all. That method made it more difficult for those who chose to work part-time because of their household or parental responsibilities (most of whom were women) to qualify for the benefit than it was for anyone else, including those who chose to work part-time for other reasons, e.g. in order to pursue leisure activities. Although the method of calculation of the degree of disability had the same impact on male part-time workers as it did on female part-time workers, the latter were disproportionately disadvantaged in qualifying to claim the disability benefit.
51. Di Trizio was a classic case of indirect discrimination in which a particular class of persons eligible to claim the benefit, most of whom were women, were put at a disadvantage over all other claimants in demonstrating that they qualified for that benefit. Nothing in that case supports a comparison being made between a group of persons claiming one element of a welfare benefit with a group of persons claiming a different element of that benefit relating to different needs. The Judge's example of an employer with two offices, one of which has predominantly more women employees, subjecting workers in that office to less favourable terms of employment, seems to me to be a paradigm example of direct discrimination, but in any event it is not an apt analogy.
52. Even if one takes the more generous approach to comparators that has been adopted in Strasbourg, people claiming the HCE are not in a "relevantly similar" situation to people claiming the CCE. This is not just because the terms which apply to those disparate elements of UC are different. Housing costs are not directly comparable to childcare costs. The CCE is inherently "highly volatile and subject to regular fluctuations" both as regards school-age and pre-school-age children. That was the evidence of Ms Lawton of Save the Children UK, based on research conducted by that organisation, which the Judge accepted at [151(c)]. Housing costs reflect a need which is basic, continuous and broadly the same. Childcare provision is not monolithic or a fixed cost payable in the same way. Whilst some housing costs may fluctuate from month to month, the amount of rent (or equivalent) that a person has to pay is generally predictable and constant, whereas the amount of paid-for childcare that a parent may require during working hours in a given month (or even a given week) depends on many different factors, such as the number of hours the parent works and the timing of their shifts, the availability of a friend or relative to look after the child, and the child's state of health. The evidence demonstrated that the element of fluctuation in childcare costs is inevitably far greater, with more scope for error if payment is made when a liability is incurred rather than after it has been discharged.
53. At [159(a)] the Judge said that the decision to make eligibility for the CCE dependent on proof of payment was discriminatory because it adversely affects all those entitled to the CCE of whom the majority are women – the very line of reasoning he had already correctly rejected at [155]. He did not cure the underlying flaw in that reasoning by introducing in parenthesis a reference to a failure to apply a proof of payment rule to those claiming the HCE.
54. However, in his alternative analysis at [159(b)] the Judge did consider the correct comparator group, namely, men who are entitled to claim the CCE. This ground of appeal therefore turns on whether the Judge was entitled to reach the conclusion that the proportion of women who are able to comply with the Proof of Payment Rule was bound to be substantially lower than the proportion of men who can.
55. Ms Dobbin submitted that there was insufficient robust evidence to support that conclusion. As she pointed out, the Judge's comparison between the Proof of Payment Rule and a height requirement at [158] is inapt. Whereas the ability to meet a height requirement is fixed, in the sense that a person will inevitably be either tall enough or too short to meet it, and men are generally taller than women, the question whether someone can afford to pay for the first month's childcare, or for additional childcare at the start of the school holidays, may be contingent on a wide range of factors which are not static and which may be wholly unrelated to their sex – for example, how much they have saved, the extent of their individual need for childcare, and their access to childcare from relatives and friends. Ms Dobbin submitted that a broad comparison of earnings between male and female single parents does not demonstrate who within that pool entitled to CCE faces greater barriers to work because of the requirement to pay the provider first in full and await partial reimbursement after the end of the assessment period.
56. Ms Dobbin also pointed to the evidence of Ms Parker which indicated that since the last quarter of 2010, the employment rate of single parents had increased by 12%, which she said suggests that the CCE is having the desired effect of encouraging single parents into work. However, even if one could infer a causative link (the Judge was sceptical on that point) that figure tells one nothing about how many other single parents might have started working (or increased their working hours) if they did not have to find the resources to make the initial payment for the necessary childcare.
57. As a starting point, it is well-established that a prima facie case of indirect discrimination can be established without adducing direct statistical evidence of the number of people with a protected status who are adversely affected by the impugned rule: see the approach taken by the Grand Chamber of the ECtHR in DH ...above) at [188]–[193]. In that case, the statistical evidence relied upon by the claimants was undisputed. The respondent (the Czech government) submitted that it was "insufficiently conclusive," and the court accepted that "it may not have been entirely reliable". On the other hand, the respondent had produced no statistical evidence to contradict it. Irrespective of its flaws, the claimant's evidence was found to be sufficiently reliable for the purpose of making out a prima facie case of indirect discrimination; it illustrated a trend from which it was possible to infer that the number of Roma children in special schools was disproportionately high, even though the exact percentage was difficult to establish.
58. In the present case, the Judge accepted that there was a lack of robust quantitative evidence as to the numbers materially affected by the Proof of Payment Rule and as to the extent of its effects. Nevertheless 81% of those receiving the CCE were single mothers, and the evidence indicated that Gingerbread, Save the Children, and organisations representing childcare providers had all separately identified the Proof of Payment Rule as contributing to problems with paying for childcare, leading to UC customers getting into debt. He found this was consistent with the evidence given by the Child Poverty Action Group to the Work and Pensions Select Committee in September 2018, which identified the Proof of Payment Rule as the most common issue raised by its advisors in connection with childcare costs in UC.
59. Whilst that evidence did not enable the Court to make a finding as to the number or percentage of cases in which such problems occur, the Judge said it was unlikely that the issue would have been identified independently by these separate organisations if the number were not significant in real and percentage terms. That is fair comment, but (as the Judge acknowledged) it only indicates that in a significant but unquantified number of cases, parents claiming UC had difficulties in paying childcare providers out of their own resources. It did not explain why they were having such difficulties. Nor did it suggest that female claimants were less able to afford the payment than male claimants.
60. At that point in his analysis the Judge found that the evidence, taken in the round, demonstrated that the Proof of Payment Rule had materially adverse, and thus prejudicial effects on a significant number of those entitled to the CCE, most of whom are women; but he correctly identified that that was a different question from whether those prejudicial effects on women were disproportionate.
61. In that regard, the Judge relied on unchallenged evidence from Gingerbread, based on an analysis they had carried out of data from the Labour Force Survey conducted by the Office for National Statistics, that the median hourly pay of single mothers (both in the UK as a whole, and in London) is very substantially less than that of mothers in couples, single fathers, and fathers in couples. The statistics referred to in the Gingerbread report were used by them to compare the median value in the data set for each group with the living wage. Those statistics were gleaned from the whole working population, not just those who qualify for UC. The Judge acknowledged this, but he considered that nevertheless, the disparity between the figure for median hourly earnings for single mothers (just over £8 per hour) and single fathers (just over £13 per hour) was so significant that even among UC claimants in principle eligible for CCE, the median woman was bound to be earning significantly less per hour than the median man.
62. Three problems can be identified with that analysis, namely:

i) these were median figures;

ii) the same upper limit on earnings (approximately £62,000 per annum for a single parent with two children) applies to all claimants of UC;

iii) UC operates to supplement the earnings of lower-paid claimants in order to raise their monthly income to a level that would ameliorate, and possibly even eliminate, the disparity between them and higher-paid claimants (whose UC award will be lower).

63. A median is the value separating the higher half from the lower half of a particular data sample or range. The median is ascertained by ordering the values in the relevant data set from lowest to highest, and picking the one in the middle. Sometimes the median will be the same number as the mean, the traditional mathematical average of all the numbers in the data set, but often it will not. Thus in a data set which comprises 2, 3, 7, 8 and 25 the median will be 7, but the mean, the value produced by adding all the numbers together and dividing them by the number of values in the set, (in this example, 45 divided by 5) will be 9. A weighted average (eliminating the "outlier" figures at the top and bottom of the range and dividing the product of the remaining 3 values by 3) would be 6. The median is not the value exactly halfway between the lowest and highest values in the data set, which in the example above would be 13½, unless they are the only two figures.
64. That example illustrates that the median will depend on the identity and number of different figures within the data set and the breadth of that data set. In broad terms, if there are more high numbers than low numbers in the data set, the median figure will be higher than it would be in the reverse situation. If the figures in each data set are different, the median is likely to be different. Therefore the extrapolation of information based on the median figure in one data set, to a different data set, particularly one with narrower parameters (as in this case) is unlikely to be a productive exercise. The median figure in a larger group is unlikely to tell one anything about the median figure in data relating to a smaller group encapsulated within the larger one.
65. The discrepancy in the median hourly rates referred to by Gingerbread is readily explicable if the highest-paid workers in the country are men, because this would mean that the figures in the relevant data set for men would have higher values than the figures in the data set for women. However, by contrast with the range of earnings applicable to the public at large, a single upper limit of earnings applies to all claimants for UC. Once the highest-paid groups of workers in the country are eliminated from the two data sets that are being compared, it does not necessarily follow that the median rates of hourly earnings for men and women will be different from each other, let alone that they will be significantly different. In order for the median figures to be different, the figures in the data set for single parent male claimants of UC would still need to show higher values than the figures in the data set for single parent female claimants of UC.
66. Therefore, even if median figures could be used to provide a meaningful comparison between the hourly earnings of single mothers and single fathers eligible for the CCE, in order to carry out that comparison one would really need to look at the data for the earnings of single mothers on UC and the data for the earnings of single fathers on UC and find the midway point of the range of values of earnings in each data set, bearing in mind that the upper limit of earnings (the UC earnings limit) and the lower limit (zero earnings, because UC can be claimed by a person who does not work) would be the same for both groups. That exercise has not been done.
67. Due to the way in which UC is claimed, the DWP has access to accurate data (provided directly by HMRC) concerning the monthly earnings of all those within the cohort of UC claimants, including those eligible to claim the CCE. Those data could have been used to demonstrate (if it were the case) that the £5 per hour disparity between the median figures in the data considered by Gingerbread was caused by the people who were earning over £62,000 per annum; or that for some other reason there is no significant disparity in the hourly earnings between men and women within the relevant cohort, but neither exercise was attempted.
68. The Judge was aware that the figures which demonstrated a "marked difference" between the median hourly earnings for men and women applied to the population as a whole (he specifically acknowledged this at [157]). He therefore did not fall into the trap of assuming that the same discrepancy would be shown by the data for single parent UC claimants, if that data had been made available. However, he took the view that because the disparity in the wider population was so large, narrowing the group to single parent claimants of UC would not have so significant an impact as to make it unreasonable to infer that single mothers claiming UC would still have significantly less income than single fathers claiming UC, making it more difficult for them to make the initial payment for childcare costs out of their own resources.
69. I have grave doubts about whether there was a sound basis for drawing any inferences about a disparity in earnings between "median" members of the relevant cohort from the statistics relating to the median hourly rates of pay of single parents within the wider population. However, Ms Dobbin did not contend that the Judge was not entitled to draw the broad-brush inference that within the group of single parents entitled to claim the CCE, the women would still earn significantly less than the men. The Secretary of State's case focused on the lack of robust evidence of a causal connection between the inferred disparity in income and an individual's ability to afford to pay for childcare. Ms Dobbin submitted that there was nothing within the data relied on which demonstrated how many within the group claiming CCE could not afford to make the initial payment for childcare out of their own resources. However, as I have already pointed out, the claimant did not need to descend to that level of detail.
70. In making the assessment that it would be more difficult for women within the cohort to pay for the initial childcare upfront from their own resources, the Judge does not appear to have factored in the further salient point that UC is designed to supplement a claimant's earnings with the aim of eliminating wage disparities, so that even if there were a substantial difference in earnings between single-parent male and female claimants of UC at the point of entry, the receipt of a higher award of UC by the lower-earning female would go at least partway towards closing the gap.
71. However, that too was not a point on which any significant reliance was placed by the Secretary of State. Ms Dobbin did not submit that the payment of UC to the single female claimants of CCE would narrow any disparity in earnings to such an extent that no reliable inference could be drawn that the female single parent would be less likely to be able to afford to pay for childcare. That is perhaps understandable in the light of the failure by the Secretary of State to adduce any relevant evidence to support such a submission.
72. The figures on which the Judge relied related only to hourly rates, telling one nothing about the number of hours worked, which would also have an impact on an individual's overall take-home pay. However, given that one of the purposes of the CCE was to eliminate the disadvantages experienced by single mothers who could not meet the previous 16-hour per week minimum work requirement, it seems reasonable to assume that single-parent females eligible to claim the CCE are unlikely to work longer hours than their male counterparts.
73. Mr Buttler submitted that the Judge was entitled to regard it as self-evident that women within the UC cohort earned substantially less than men, and to draw the conclusion from this that a disparate number of women would be prejudiced by the Proof of Payment Rule. He relied upon the approach taken by the Divisional Court in R (Adiatu) v HM Treasury [2020] EWHC 1554 (Admin), [2020] PTSR 2198 especially at [158]-159], where the inference was drawn from data relating to the percentages of women doing certain types of job, including part-time work, that it was likely that a greater proportion of women than men earn below the lower earnings limit for statutory sick pay of £120 per week. Indeed, in that case the respondent had conceded the point. Mr Buttler pointed out, as the Judge appears to have accepted, that the quality of the evidence in the present case was better than that adduced by claimants in many other cases of indirect discrimination.
74. Mr Buttler also placed reliance on the fact that, as in DH, there was an absence of contradictory data. He submitted that the inference that the Judge drew was a fair and reasonable inference, which had not been rebutted, though it would have been within the power of the Secretary of State to demonstrate that it was wrong.
75. The question for the Court of Appeal is not whether we would have drawn the same inference as the Judge, but whether the inference that he drew was one that was open to him. I have not found this point an easy one, partly because of the difficulties in extrapolating any meaningful information from the data relating to the population at large on which the Judge relied, but also because of the range of imponderables, wholly unrelated to gender, which also have an impact on whether an individual can or cannot afford to pay for something out of their own resources. A person's take-home pay is only one factor.
76. That said, it is probably the most substantial one, and the one that could make all the difference in terms of whether a payment is affordable or not, even though the amount of childcare needed (and thus the amount of the initial outlay required) will vary from person to person. Those random variations in circumstances will apply throughout the population. The level of a person's earnings will obviously have some impact on their ability to save. People working on zero hours contracts, or part-time for less than 16 hours per week, are inherently disadvantaged in building up substantial savings, and even if they were able to, it may take them longer than someone earning more or working longer hours. It is not unreasonable to infer that the more a person earns, the more likely they are to be able to afford to make a payment from their own resources, or the sooner they will be able to afford to do so. Other factors having a bearing on affordability might be expected, statistically, to be reflected more or less equally among men and women.
77. I have concluded, not without some hesitation, that looking at all the evidence in the round, including the evidence that various different organisations had identified the Proof of Payment Rule as creating a real practical difficulty, the Judge was entitled to reach the conclusion that he did. Whilst a broad comparison of earnings between male and female single parents eligible to claim the CCE did not provide all the information necessary to enable him to conclude how many females within that group would be unable to afford to make the initial childcare payment, when all the evidence was taken in the round, there was just enough information to support the inference that significantly fewer women than men are likely to be able to afford to do so, and that was enough to raise a prima facie case of indirect discrimination.
78. It was therefore necessary to consider whether the Rule was objectively justified, which in turn involved consideration of whether it had a legitimate aim and whether it was a proportionate means of achieving that aim.

Was the Rule objectively justified?

79. Chamberlain J addressed the proper approach to objective justification at [167]. Having correctly identified that in a case of indirect discrimination it has to be shown that the measure itself has a legitimate aim and is a proportionate means of achieving that aim, he said that assessing proportionality involves answering the four questions posed by Lord Reed in Bank Mellat at [74], namely:

i) Whether the objective of the measure is sufficiently important to justify the limitation of a protected right;

ii) Whether the measure is rationally connected to the objective;

iii) Whether less intrusive measures could have been used without unacceptably compromising the achievement of the objective; and

iv) Whether, balancing the severity of the measure's effects on the rights of the persons to whom it applies against the importance of the objective, to the extent that the measure will contribute to its achievement, the former outweighs the latter.

He then held that, as this was a case involving a welfare benefit, the Court had to proactively examine whether the measure was "manifestly without reasonable foundation".

80. Next, the Judge identified that the intensity with which the Court will review the asserted justification for the measure depends on the context ([167(e)]. At [167(f)] he identified three factors relevant to determining the appropriate intensity of review in a case where the question is whether the measure is manifestly without reasonable foundation, namely, (i) the nature of the ground of differential treatment, (ii) whether and to what extent the matter involved a real socio-economic policy choice present to the mind of the decision-maker, and (iii) whether and to what extent the measure had been approved by Parliament.
81. The Judge held at [168] that the Proof of Payment Rule had a legitimate objective, namely "making UC simpler than the various benefits it replaced, with a view to reducing fraud and error, which had previously been significant problems" and that that objective was sufficiently important to justify the limitation of a protected right. Having then said that the three remaining limbs of the Bank Mellat test could conveniently be considered together, he first considered each of the three factors relevant to the proper intensity of review, and how they applied in this case, before concluding at [172]:

"Drawing the threads together, the decision not to deliver the CCE by direct payment to the childcare provider was made by Ministers. By contrast, there is no evidence that the decision to make payment of the CCE dependent upon proof of payment (rather than proof that the charges have been incurred) was ever directly considered by Ministers. In considering whether that decision was manifestly without reasonable foundation, it is therefore not appropriate to apply a particularly wide discretionary area of judgment. However the fact that the decision was given effect in an instrument approved by Parliament, and the amended version was laid before Parliament, must be recognised when applying the test." (Emphasis supplied).

82. He next considered the evidence concerning the decision by Ministers not to deliver the CCE by direct payment to the childcare providers and found, at [173(d)], that:

"The decision to favour a simpler system, with a single monthly payment, so as to minimise deviations from the architecture of the scheme, is the kind of socio-economic choice on which democratic decision-makers are entitled to a relatively broad discretionary area of judgment. It is not the kind of decision this Court can properly stigmatise as lacking a reasonable foundation".

83. Having reached that conclusion, the Judge then went on to identify the key focus of the argument in these proceedings as being "the difference between proof of payment (on the one hand) and proof of liability to pay (on the other)." He identified that there was no evidence that Ministers, as distinct from officials, had ever directed their minds to the distinction between those two systems, and concluded that it was not obvious why a proof of payment rule should be more simple, easy to understand or better at avoiding errors in the system than a proof of liability rule, which might be capable of addressing the identified difficulties. For a series of reasons that he set out at [173(e) to (j)], he concluded at [174] that the Rule lacked a reasonable foundation and was therefore not objectively justified, which made it incompatible with Art 14.
84. Ms Dobbin made three criticisms of the Judge's approach, namely: (1) that the level of scrutiny was unduly strict; (2) that it was impermissible for him to examine whether the decision-maker had confronted the problem and considered solutions; and (3) the Judge was not entitled to consider whether an alternative rule existed which would have solved or ameliorated the problem. Mr Buttler submitted that there was no substance in any of these criticisms. The Judge was entitled to reach the conclusion that he did. There was no error of principle nor any identifiable flaw in the Judge's reasoning, and the Court was being invited to substitute its own evaluation of proportionality for a decision the Judge had reached after weighing all the relevant factors and applying the correct legal test.
85. The Judge's analysis of the relevant legal principles at [167] was prescient. It epitomises the "balanced approach" to justification subsequently articulated by Lord Reed in SC. He rightly recognised that this was a case where the requirement for "convincing and weighty reasons" by way of justification remained, notwithstanding the degree of deference to be afforded to the decision-maker, because the measure under challenge was an aspect of a welfare benefit. At [169] he said that although this was a case of indirect discrimination, the group which is subject to disproportionately prejudicial effects is defined by reference to gender:

"This means that more convincing and weighty reasons will be needed than if the discrimination were on another, non-suspect ground."

I can see no reason to criticise that formulation.

86. In the light of the judgment of the Supreme Court in SC, the Secretary of State's contentions that the manifestly without reasonable foundation test has superseded the Bank Mellat test, and that in the context of welfare benefits it also replaces the "weighty reasons" test, must be rejected. In SC, Lord Reed endorsed the approach of the Court of Appeal in R (Joint Council for the Welfare of Immigrants) v Secretary of State for the Home Department [2020] EWCA Civ 542, [2021] 1 WLR 1151 ("JCWI") and by the Divisional Court in Delve. In JCWI (having referred with approval to Leggatt LJ's approach in SC in the Court of Appeal), Hickinbottom LJ put the matter in this way at [140]-[141]:

"In my view, the [manifestly without reasonable foundation] criterion simply recognises that, where there is a substantial degree of economic and/or social policy involved in a measure, the degree of deference to the assessment of the democratically-elected or -accountable body that enacts the measure must be accorded great weight because of the wide margin of judgment they have in such matters. The greater the element of economic and/or social policy involved, the greater the margin of judgment and the greater the deference that should be afforded. That is, for obvious reasons, particularly so when that body is Parliament. However, if the measure involves adverse discriminatory effects, that will reduce the margin of judgment and the degree of deference. That will be particularly so where the ground of discrimination concerns a core attribute such as sex or race…

If that analysis is right, whether seen in terms of the application of the manifestly without reasonable foundation criterion or simply in terms of the usual balancing exercise inherent in the assessment of proportionality, the result should be the same."

87. Lord Reed agreed with that analysis. He said this in SC at [161]:

".. the ordinary approach to proportionality will accord the same margin to the decision-maker as the "manifestly without reasonable foundation" formulation in circumstances where a particularly wide margin is appropriate."

Therefore, in a case where the "manifestly without reasonable foundation" criterion is in play, the application of the Bank Mellat test should produce the same result.

88. In the present case, the measure to be justified was the Proof of Payment Rule, a provision which requires a person claiming CCE to have paid for relevant childcare before claiming the element of UC relating to that childcare. As the Judge recognised, it was a measure of socio-economic policy relating to a welfare benefit and the "manifestly without reasonable foundation" criterion applied. The analytic framework set out by the Judge at [167], including the three factors he identified as relevant to the intensity of review in a case where that criterion applies, is an accurate reflection of the case law as it then stood. It also accords with the approach adopted by Lord Reed in SC.
89. It was only when the Judge began to examine the second of those factors - whether there was a deliberate policy choice present to the mind of the decision-maker - that he took a wrong turn which, in my judgment, fatally undermined his application of the principles to the facts. All that the evidence needed to show in order to establish that factor was that Ministers expressly considered and decided to adopt the Proof of Payment Rule as part of the structure of the CCE.
90. The evidence established that the policy was formed by an iterative process that ultimately involved Ministers. They were concerned to arrive at a scheme for delivering the CCE that met the policy imperatives, an important aspect of which was reducing fraud and error. Fraud had not been identified as a major issue, but error was: - it had been established that around 25% of monies paid for childcare under the legacy system (based on estimated needs) were wrongly paid. The system design also took account of what stakeholders who were consulted said they wanted (e.g. not vouchers).
91. A report made by DWP officials to Ministers of one meeting with stakeholders on 31 October 2011 said that:

"Stakeholders appeared to like the simplicity of the reporting paid costs and removing the need for estimates and complex calculations….

As expected they did raise some concerns about childcare payments always being in arrears as part of the Universal Credit award and the time lag for some claimants being as long as 4 weeks. However both provider and family representative organisations suggested that work could be done with providers to change payment/billing behaviour and offered to work with the Department to do this.

… Although stakeholders were pleased with our intention to allow parents to claim upfront childcare costs, they did raise a concern about the work incentive implications if parents have to pay that initial amount out of their own money.

Action: to consider interaction with other support available to help parents pay upfront costs – i.e. budgeting loans within Universal Credit and support via the JCP Flexible Fund."

Ministers were therefore aware of issues that had been raised about the proposal that payment of the CCE would be made in arrears as part of the UC award, and that claimants would have to find the money for initial payments from their own resources. They adopted the model for delivery of the CCE, including the Proof of Payment Rule, notwithstanding those concerns.

92. The Judge, however, approached the matter (from [170 (e)] onwards) on the basis that in order to establish that factor it was necessary to show that all alternatives for delivery of the CCE, including a proof of liability rule, had been considered and rejected by Ministers. The impact of that error is most clearly illustrated at [172]. The Judge's view that it was inappropriate to accord a particularly wide margin of appreciation to the decision-maker was expressly premised on his characterisation of the decision which required justification as a "decision to make payment of the CCE dependent upon proof of payment (rather than proof that the charges have been incurred)". He adopted that characterisation despite the fact that there never was a Ministerial decision to reject a proof of liability rule, because that choice was never presented to Ministers. The evidence showed that a proof of liability rule had never even been suggested to officials by the stakeholders with whom they had engaged, nor had it been raised independently by those officials for consideration by Ministers.
93. The mischaracterisation of the measure under challenge also affected the Judge's approach to justification. The Judge approached the matter at [173] as if the Secretary of State was attempting to provide justification after the event for a selection of a proof of payment rule over a proof of liability to pay rule. He was wrong to do so; not only was there never any such selection, but that was not what the Secretary of State was required to justify.
94. Moreover, that was not the case which the Secretary of State had to meet. The claim for judicial review was primarily for a declaration that Reg 33(1) was unlawful, because it infringed Art 14 read together with Art 8 and/or A1P1. As this was an indirect discrimination case, it was the measure itself, rather than its discriminatory effect, which had to be justified by the Secretary of State as a proportionate means of achieving a legitimate aim. The claim was not articulated as a challenge to a decision by the Secretary of State to reject a proof of liability rule in favour of a proof of payment rule. Insofar as it was ever articulated as a challenge to a decision, it was a challenge to an ongoing decision by the Secretary of State refusing to disapply Reg 33(1).
95. In paragraph 49 of the claimant's Statement of Facts and Grounds, it was contended that there were "obvious ways of promoting an operational aim of combating fraud without undermining the overarching objective of the scheme (increasing work)". Four examples were given, none of which was a system by which CCE was paid against proof of liability. They were, respectively, a childcare deposit scheme; direct payments to childcare providers; the creation of childcare accounts; and a discretion to allow upfront payments (i.e. payments in advance of the UC award). Each of these had been espoused, to varying degrees, by different stakeholders when the policy was in the process of formulation.
96. In response to that case, the Secretary of State did not accept that discounting alternatives was part of the legal test for justification. However, the Grounds of Defence and Ms Parker in her evidence addressed each of the four mechanisms suggested by the claimant, in order to demonstrate that they were at odds with fundamental aspects of the scheme of UC (as the Judge ultimately accepted they were). The absence of evidence about the potential alternative of a proof of liability rule was understandable, because the failure to adopt such a rule was not a ground of criticism in the pleaded case.
97. It was only when the claimant's skeleton argument was lodged that the argument began to focus on the difference between the system adopted for the HCE and the system adopted for the CCE. It appears that the dichotomy between the Proof of Payment Rule and an alternative regime of payment against invoice only emerged clearly in argument at the hearing itself. Claims for judicial review quite often become more narrowly focused as the case progresses, and no application was made by the Secretary of State for an adjournment or to put in further evidence. However, whilst no criticism is or could be made of the claimant about the way in which matters developed, Ms Dobbin submitted, and I accept, that this forensic context should have been reflected in the Judge's approach to the issue of justification. There was no reason to have expected the Secretary of State's evidence to have addressed a point which did not emerge until late in the day, let alone to have dealt with it in the same detail as the proposals which were carefully examined in the process of policy formulation and expressly referred to in the claimant's pleaded case.
98. The Judge correctly identified what had to be justified, at [167(a)], as the Proof of Payment Rule, or the decision to apply it to the CCE without exceptions. No doubt because of the way in which the argument then developed, it was only when he began to consider the justification given for the Rule, that he first approached the matter on the erroneous basis that it was incumbent on the Secretary of State to justify the rejection of (or failure to adopt) alternatives which might also have served the identified legitimate aim. That appears to have been what led him to discount the explanation given by Ms Parker for the adoption of a system based upon a reimbursement model, which fitted in with the core design of UC as a single lump-sum payment made monthly in arrears, which could operate flexibly to meet the different needs and circumstances of individual claimants. This last aspect was particularly important, given the highly volatile nature of childcare needs.
99. That error led the Judge to apply a greater degree of intensity of review than was actually warranted. In fact, this was a classic example of a case in which a substantial degree of deference should have been afforded to the judgment of the decision-maker, especially as the impugned measure had been subjected to Parliamentary scrutiny, notwithstanding that this deference would be tempered to some extent by the fact that the ground of discrimination complained of was on the basis of sex, which would require more cogent reasons to justify it than if the differential treatment were on the basis of some other non-suspect ground.
100. If the Judge had dealt with the second, third and fourth limbs of the Bank Mellat test sequentially, instead of attempting to deal with them compendiously, it is unlikely that he would have fallen into the error that he did. The answer to the second question posed by Lord Reed was plainly yes. The Proof of Payment Rule was rationally connected to the objective that the Judge had already held was sufficiently important to justify, in principle, some limitation of a protected right. A system which identifies with precision how much has been expended on childcare in the relevant assessment period plainly reduces the scope for fraud or error, and is simple to understand and administer.
101. The third question posed by Lord Reed is whether less intrusive measures could have been used without unacceptably compromising the achievement of the objective. This seems to me to be the key question in this case. It is an objective question, but the Judge did not address it, because he focused instead on whether there was justification for a hypothetical (and non-existent) subjective decision to choose one system over another. He made no finding that a proof of liability rule would be as simple, easy to understand or as good at avoiding fraud or errors of over- or under-payment, simply observing at [173(f) and (h)] that it was "not obvious" why a system of awards based on liability to pay would be any more likely to achieve those aims than a system based on proof of liability to pay. That placed an unwarranted burden on the Secretary of State to demonstrate why an alternative system was not chosen, rather than to justify the system that was.
102. There is a self-evident simplicity and certainty about a system of reimbursement for childcare that has already been paid for. The amounts paid for childcare and the dates to which the payments relate are easy to ascertain and prove, and can be cross-checked by the DWP against the income for the hours worked by the parent. On the face of it, therefore, the Rule has a reasonable foundation, even though some single parent mothers may find it harder than some single parent fathers to comply with it when they first seek to claim the CE element of UC.
103. Although a rule based on proof of liability to pay might also achieve simplicity without compromising a requirement for accuracy (in the sense that an invoice which tells you how much someone is liable to pay for childcare is more precise than an estimate of future costs), I am not persuaded that the evidence would support a conclusion that it would be as simple, or as good at avoiding errors, as a rule based upon proof of what was actually paid. As Ms Dobbin pointed out, in a system based upon an invoice, particularly where the liability to pay arises in advance of provision of the services, the certainty that the payment made in respect of childcare is used for childcare during working hours diminishes. A liability may not be discharged – for perfectly legitimate reasons, e.g. if the services are no longer required, or are required for a shorter period than initially expected. The child or the parent may fall ill, or a relative may offer to look after the child instead. That means the system would have to cater for overpayments and refunds.
104. In any event, switching to a proof of liability system would only make a practical difference in a case where the provider required payment upfront, and the claimant could not afford it, if the UC payment was made in time to enable the claimant to discharge that liability from the benefit payment before the childcare was provided (which would cut across the fundamental architecture of a system based on payment in arrears for something already provided to the claimant). Even the HCE is payable in respect of accommodation which has already been provided. As the Judge recognised at [173 (i)], in order to fit in with the general overall scheme of UC, even if payment of the CCE were made on proof of liability to pay the relevant childcare costs, instead of on proof of payment, the payment would still have to be made in arrears (after the childcare was provided).
105. The Judge said it was "possible to conceive of" a system based on proof of liability, where the charges related to childcare provided in a past complete assessment period. That is true, but as the Judge recognised in his analysis at [170], no such system had ever been put forward for consideration. Under such a system, the provider of the childcare might raise an invoice at the start of the assessment period, but the payment of UC (factoring in the CCE) would still not be made until after the end of that period. Even in that type of case, making payment of the CCE dependent upon the production of an invoice rather than a receipt would only make a difference to the ability of the impecunious single parent to work (or work longer hours) if the childcare provider were willing to wait for payment.
106. The Judge accepted this premise; but he then engaged in speculation that "[the provider] might be willing to do so if they knew the CCE would cover their invoices. If so, a system in which entitlement is based on proof of liability to pay would prevent claimants having to incur debt in order to pay childcare charges." [Emphasis supplied]. He therefore reached no conclusion on the question whether such a system would make a difference in terms of achieving the objective of getting people back to work. At most, he found that such an alternative measure might be capable of addressing the problem identified by the claimant, but only if the market reacted favourably to it.
107. There was no basis for concluding on the evidence that the postulated alternative measure would achieve a different practical outcome for single mothers, and the Judge did not reach that conclusion. Indeed, such evidence as there was about the attitude of the market providers suggested that they would not react favourably to the idea of waiting for payment (even though payment by the State reduces the credit risk). Due to their own cashflow concerns, the vast majority of childcare providers required payment in advance of providing the childcare – 90%, according to evidence from Save the Children. Often this was in the form of a non-refundable deposit. Unless there was a radical change in their attitudes, even under a proof of liability system, a claimant would still have to pay for all the childcare in the initial period out of their own resources, instead of just finding 15%, because there would still be a time lag between submitting the invoice as proof of liability and receiving reimbursement as part of the UC award, and the childcare would not be provided without payment. In other words, the complaint about the Catch-22 situation would arise even under a proof of liability system which related to childcare provided within the assessment period in question.
108. At [173 (j)] the Judge said that there was nothing in Ms Parker's statement which explained why a system based on liability to pay (such as that adopted in the HCE, which is also paid in arrears) would be any less able to deal with the fluctuating costs of childcare than the current system. I have already explained why there was no reason to have expected her evidence to address that specific point. In any event, I do not accept that the HCE is truly comparable, for the reasons already articulated at [52] above. The system for delivery of the CCE was designed to meet a need for flexibility that does not exist in the sphere of housing costs.
109. On the basis of the Judge's findings, the third Bank Mellat criterion was not met. Moreover, the broad-brush nature of the inference drawn by the Judge that substantially more female claimants than male claimants would be unable to afford to meet the Proof of Payment Rule made it that much harder to condemn that Rule as disproportionate. What Mr Buttler termed a "systemic problem" of not being able to afford to pay first from one's own resources only arises when a claimant first claims the CCE, or if their need for childcare suddenly increases; the problem does not apply throughout the entire period when that individual is claiming the CCE. It was impossible to gauge how many people entitled in principle to claim the CCE could not afford to pay the initial childcare costs from their own resources, let alone the extent to which, in practice, the Rule inhibited single female parents from working or working full-time.
110. In circumstances in which there was no evidence that a significant number of claimants were even affected, and there was no means of gauging how severe the impact was on them, it would have required compelling evidence to justify a conclusion that the third Bank Mellat criterion was satisfied. At most, there was a possibility that the postulated alternative measure might achieve the same objectives less intrusively, but only if there were a radical change in the attitude currently taken by childcare providers. On a proper analysis of the evidence there was only a slender and improbable prospect that single female claimants of CCE might be any better off under a proof of liability rule. A measure cannot be impugned as a disproportionate interference with a protected right on the basis of such a conditional and speculative finding.
111. In the light of those matters, there was no basis for concluding that, when all relevant matters are weighed in the balance and the appropriate degree of deference is afforded to the decision maker, the problems for however many female claimants are affected are so grave as to outweigh the adoption of a Proof of Payment Rule which has been shown to have had the desired effect of significantly reducing fraud and error as well as achieving simplicity.
112. An application of the correct approach, whether as reflected in limb 4 of Bank Mellat, or by affording the appropriate margin of appreciation to the decision maker as is required by the "manifestly without reasonable foundation" formulation, taking into account the three factors identified by the Judge at [167(f)], inevitably leads to a different answer from the one reached by the Judge at [178]. The difference in treatment is not manifestly disproportionate to the legitimate aim identified by the Judge at [168].
113. Ms Parker's evidence provided ample justification for the adoption of the Proof of Payment Rule. The reimbursement model of payment of costs actually expended on childcare fitted with the overall architecture of UC, it was simple to understand and operate and it significantly reduced the problems that had caused major difficulties with the legacy system. The Judge's critique of the justification put forward for the Rule failed to accord the appropriate degree of respect for the measure of judgment afforded to Ministers in this area of policy. It seems to me that the Judge's observations about the decision to reject direct payments to the providers at [173(d)], quoted in paragraph 82 above, are just as apposite and apply with equal force to the decision to adopt the Proof of Payment Rule.
114. For those reasons I would allow this appeal on the discrimination issue.

Was the Proof of Payment Rule irrational?

115. The Judge also upheld the rationality challenge, which depended on the submission that the Proof of Payment Rule undermines the fundamental purpose of UC of encouraging work and making work pay to the fullest possible extent. He held, for essentially the same reasons, that the justifications given by the Secretary of State for maintaining the Rule did not adequately explain the need for entitlement to be based on proof of payment rather than on proof of liability, and that for the reasons set out at [177] there were "striking parallels" with the decision that was successfully challenged in R (Johnson) v Secretary of State for Work and Pensions [2020] EWCA (Civ) 778 [2020] PTSR 1872 ("Johnson"). He concluded at [178] that:

"the maintenance of the Proof of Payment rule, insofar as it precludes a system where eligibility is based on liability to pay, is irrational" [Emphasis supplied.]

116. I do not accept the analogy with Johnson, which was a very different case. The case concerned an acknowledged arbitrariness in the way in which Reg 54 of the UC Regulations operated for certain claimants, because the system for identifying the income earned by a claimant in a particular assessment period did not accommodate the fact that people who were usually paid their salary on a particular date each month, would be paid on a different day if the designated date fell on a non-banking day. If, as a result of this shift, two monthly salary payments fell within the same assessment period for UC, the claimant's award would be artificially reduced. Conversely, if the non-banking day shift meant that in a different assessment period there was no monthly salary payment, the claimant would get a much higher award of UC for that period. The resulting fluctuations in income made it very difficult for these claimants to budget their monthly outgoings, and caused some of them serious hardship.
117. Importantly, the decision which was challenged as irrational was a decision by the Secretary of State not to provide a solution to this acknowledged problem. It was not a challenge to the lawfulness of any of the Regulations themselves.
118. Rose LJ, who delivered the leading judgment, described the way in which the Regulations applied to the claimants in that case as "odd in the extreme". She pointed out, at [47], that there was no policy reason why they should face those difficulties. In assessing rationality, the Court considered the disadvantages of adopting a solution to the problem (by amending the Regulations, altering the computer program, or both); whether a solution would be consistent or inconsistent with the nature of the UC regime; and whether it was possible to say that no reasonable Secretary of State would have struck the balance in the way in which the Secretary of State did in that case: see Rose LJ's judgment at [50].
119. The Court concluded that the high hurdle of establishing irrationality had been met. Although it might be difficult, it was not impossible to draft an exception to the Regulations to cover the problem; furthermore, it ought to be possible to modify the computer program to recognise the non-banking day problem and make the necessary adjustment so that if it arose, the two monthly salary payments would be treated as falling into different assessment periods. The Minister had been assured that the computer might be programmed to recognise unexpected and significant fluctuations so that they might be catered for (though in the event that was not done).
120. The Court also took into account the substantial size of the cohort affected (possibly as many as 85,000 claimants); the duration of the impact upon them, and the fact that it was not really possible for a claimant to cure the problem; the arbitrary nature of the occurrence; and the fact that the Secretary of State's refusal to deal with the problem failed to promote the purpose of the UC regime. Indeed, as Underhill LJ pointed out in his short concurring judgment at [114], the present operation of the scheme ran positively counter to its declared purpose.
121. The Court of Appeal went out of its way to confine the decision in Johnson to its own peculiar facts. At [107] Rose LJ described the case as:

"one of the rare instances where the SSWP's refusal to put in place a solution to this very specific problem is so irrational that I have concluded that the threshold is met."

Underhill LJ added, at [116]:

"…I regard this as a case which turns on its own very particular circumstances. It has no impact on the lawfulness of the universal credit system more generally."

122. The narrow confines of the decision have been highlighted more recently by the Court of Appeal in R (Pantellerisco) v Secretary of State for Work and Pensions [2021] EWCA Civ 1454, the facts of which were much closer to Johnson than the facts of the present case. Pantellerisco concerned claimants who were paid their salary on a 28-day cycle rather than by the calendar month. This provided certain disadvantages in respect of the disapplication of the benefit cap (referred to as "the pay cycle effect"). Unlike Johnson, but like the present case, the challenge was directly to the lawfulness of the Regulations which gave rise to the problem, and the failure to cater for exceptions. Initially the claim was brought both on grounds of irrationality and under Art 14, but Garnham J, who found for the claimant on the irrationality challenge, did not find it necessary to deal with the Art 14 challenge, and it was not pursued when the Secretary of State appealed.
123. On appeal to this Court in Pantellerisco, the Secretary of State submitted that it was inevitable in the case of legislative schemes of this kind that there would be cases where people are disadvantaged by falling just on the wrong side of a "bright line", as the claimant in that case and others in her position could be said to do; but that did not mean that the line is arbitrary in the sense of being illegitimate. The Court of Appeal agreed, and reversed Garnham J's decision.
124. At [59] Underhill LJ (who gave the leading judgment in that case) repeated what he had said in Johnson at [113]:

"I recognise, as does Rose LJ, the extraordinary complexity of designing a system such as universal credit, and that it necessarily involves a range of practical and political assessments of a kind which the Court is not equipped to judge. I also accept that in order to be workable any such system may have to incorporate bright-line rules and criteria which do not discriminate fully between the circumstances of different individuals. … I fully accept that a Court should avoid the temptation to find that some particular feature of such a system is 'irrational' merely because it produces hard, even very hard, results in some individual cases."

He continued:

"I would add that the very complexity and difficulty of the exercise is bound to mean that following the implementation of the scheme it may become clear with the benefit of experience that some choices could have been made better. But it does not follow that the Regulations were in the relevant respect irrational as made, or that it would be irrational not to correct the imperfections in question once identified: the Court cannot judge the lawfulness of such schemes by the standard of perfection. Whether any errors or imperfections in the scheme are of such a nature or degree as to impugn the lawfulness of the relevant regulations must depend on the circumstances of the particular case, having regard to the appropriate intensity of review."

125. Having considered the evidence put forward by the Secretary of State in that case, justifying the adoption of a system by which a person's entitlement to the disapplication of the benefit cap is based upon actual receipts during the monthly assessment period, and the difficulties of derogation from that rule, Underhill LJ accepted that there was a real risk that any departure from the "actual receipts principle" would seriously impair the workability and reliability of the assessment of entitlement for the relevant group of claimants. That being so, it was not open to the Court to hold that it was irrational for the Secretary of State not to have modified the effect of Reg 54 so as to eliminate or mitigate the pay cycle effect. He said this at [80]:

"…deciding whether or to what extent to derogate from a general principle of this kind in order to address the interests of a particular group is quintessentially a question for the Secretary of State (with the assistance, of course, of her civil servants) and not the Court. It requires a detailed understanding of a highly complex scheme, and the technicalities of its administration, which the Court does not have, so as to be able to assess the advantages and disadvantages of implementing any particular solution. It will also ultimately require the striking of a balance between those advantages and disadvantages, which is an exercise of judgment that is the province of the legislator. If it were established that there was a straightforward solution which it was irrational for the Secretary of State not to have pursued the Court could and should nevertheless intervene; but that is not the case."

126. The challenge in the present case is not to a decision refusing to "fine tune" the UC system to cater for an arbitrary harmful impact of the type identified in Johnson. Rather, as in Pantellerisco, it seeks to stigmatise as unlawful one of the elements of the system which was deliberately chosen, and which potentially gave rise to hardship for those who were unable to comply with its requirements or found it difficult to do so. The Proof of Payment Rule was adopted in order to reduce fraud and error and promote simplicity. It fits within the overall scheme of UC as a benefit to be paid in arrears, akin to a salary, supplementing income from work. It operates on the basis of actual outlay in the period of assessment, just as the rule under challenge in Pantellerisco operates on the basis of actual receipts in that period. It plainly has a rational justification; moreover, it cannot be said that no reasonable Secretary of State would have taken the decision to adopt such a Rule as part of the UC system or to have decided to apply it without making exceptions for individual hardship. What Underhill LJ said in Pantellerisco about derogation from such a rule being quintessentially a matter for the legislator, applies with equal force here.
127. Moreover, the Judge's analysis at [177] was based on the same fundamental mischaracterisation of the measure under challenge as a decision to make entitlement dependent on proof of payment rather than on proof of liability to pay. This was not a case where there was an obvious straightforward solution which it was irrational for the Secretary of State not to have pursued. Given that, for the reasons already set out, the introduction of a proof of liability rule would not have produced any substantially different outcome for Ms Salvato and other claimants in her position, unless providers of childcare had a radical change of attitude, it cannot be said that "precluding a system where eligibility is based on liability to pay" struck the balance in a way which no reasonable Secretary of State would have done. The threshold for irrationality is a high one, and in my judgment the challenge fell short of meeting it by some margin.
128. For those reasons, I would allow this appeal on the rationality ground also.

Warby LJ:

129. I agree with both judgments.

Underhill LJ:

130. I agree that this appeal should be allowed and the claim accordingly dismissed, for the reasons given by Andrews LJ.  Like her, I think there are real difficulties in the case that the Proof of Payment Rule had a disproportionate impact on women; but I am not prepared to say that on the material before him the Judge was not entitled to reach the conclusion that he did on the second of his two alternative bases.  However, the uncertainty about the extent of the disparate impact on women is a further reason why when it comes to justification the proportionality balance should be struck in the Secretary of State's favour.