The Motherhood Plan (R on the application of) v HM Treasury

This appeal concerns a challenge to the lawfulness of the Self-Employment Income Support Scheme (“SEISS”) introduced by the government in April 2019 during the first lockdown in the Covid-19 pandemic. The appellants contend that, contrary to Article 14 of the European Convention on Human Rights (“the Convention”), read with Article 1 of the First Protocol, the SEISS unlawfully discriminated against self-employed women who took a period of leave relating to maternity or pregnancy in any of the three relevant tax years since the level of support granted to them under the scheme was not representative of their usual profits. Their application for judicial review was dismissed by the judge. They now appeal against her decision with the leave of the single Lord Justice.

Although the early history of the pandemic in this country, and of the government’s response, is well known, it is important to recite it in outline to illustrate the speed with which events unfolded in March 2020. The summary set out below is substantially based on the witness statements of Suzanne Kantor, the Co-Director for Personal Tax, Welfare and Pensions at HM Treasury filed in these proceedings and dated 27 October 2020, and Max Hacon, the Project Director responsible for the SEISS.

On 31 January 2020, Public Health England announced the first cases of people in England to test positive for Covid-19.

On 3 March 2020, the government published its Coronavirus Action Plan. On 11 March, the Chancellor of the Exchequer announced a number of welfare support changes to provide assistance for those affected by the pandemic. On 12 March, the Prime Minister announced that from the following day, anyone with symptoms of the virus would need to stay at home for seven days, adding that it was likely that further social distancing measures would be required in the next few weeks. Four days later, on 16 March, he made a further announcement that those with symptoms should stay at home for 14 days, that all non-essential contact and travel should cease, and that people should work from home wherever possible. On 17 March, the Chancellor announced that government-backed and guaranteed loans would be made available to businesses, that business rates would be suspended for the forthcoming financial year and other rate relief made available, and that mortgage lenders would offer mortgage holidays of at least three months.

In her statement in these proceedings, Ms Kantor, looking back to that point, stated:

On 18 March, the Prime Minister announced the closure of schools taking effect two days later. On 19 March, the Coronavirus Bill was introduced into Parliament with the aim of granting the government wide-ranging emergency powers to manage the pandemic. On 20 March, the Chancellor announced that the government was setting up the CJRS under which the government would pay employers 80% of the costs of employment for furloughed workers up to a maximum of £2,500 per employee. Details of the scheme were announced on 26 March and it came into force on 20 April. At or around the same time as announcing the CJRS, the government set out its plans for tax deferrals, support for renters, and a £20 increase in universal credit. Meanwhile, pubs and restaurants were ordered to close from 21 March. On 23 March, the Prime Minister announced a further tightening of restrictions, introduced by statutory instrument and enforced by the police, preventing people from leaving their homes save for very limited purposes and prohibiting all social events. Thus in less than three weeks from the announcement of initial measures, the country had entered what subsequently became known as the first lockdown.

On 25 March, six days after being introduced into Parliament, the Coronavirus Act 2020 received Royal Assent. The powers granted to the government under the Act were to expire after two years and be subject to six-monthly reviews in the interim. Section 76 of the Act provides:

On 26 March, the Chancellor announced the details of the SEISS. Under the scheme, the government would pay self-employed people a taxable grant worth 80% of their average monthly profits over the last three years, up to £2,500 per month, for three months or longer if necessary. In his announcement, the Chancellor said:

To make sure that the scheme provides targeted support for those most in need, it will be open to anyone with income up to £50,000

To make sure only the genuinely self-employed benefit, it will be available to people who make the majority of their income from self-employment

And to minimise fraud, only those who are already in self-employment, who have a tax return for 2019, will be able to apply.

In the course of March and April, the details of the scheme were worked out, the process being described in the series of Ministerial Briefing Notes. Meanwhile, on 3 April and 27 April, the government announced details of two business loan schemes for larger and smaller enterprises. The latter scheme, called the Bounce Back Loan Scheme, included self-employed individuals and provided government-backed loans of between £2,000 and up to 25% of their turnover, up to a maximum of £50,000.

On 30 April, the Chancellor signed a direction under the Coronavirus Act giving effect to the SEISS – the Coronavirus Act 2020 Functions of Her Majesty’s Revenue and Customs (Self-Employed Income Support Scheme) Direction. It is this direction, hereafter referred to as “the First Direction”, which is the subject of the challenge in these proceedings. Following the First Direction, the SEISS opened for applications on 13 May and closed on 13 July 2020. The structure of the scheme is described in greater detail below.

On 1 July, a further direction (“the Second Direction”, not subject to challenge in these proceedings) was signed modifying and extending the SEISS. Under this direction, a second round of SEISS grants, worth 70% of average monthly trading profits, were made available by HMRC, with applications opening on 17 August and closing on 19 October 2020. Subsequently, further grants were introduced by later directions extending the support into 2021. The CJRS scheme has also been extended by parallel directions. Other schemes have been extended or introduced to mitigate the economic consequences of the pandemic which it is unnecessary to recite here.

In her statement, Ms Kantor identified a number of challenges facing the Treasury and HMRC when designing “a fair, workable and principled scheme for self-employment income support at great speed” which she put forward as justifying its structure.

First, delivery of support through HMRC was “the logical, and only realistic, choice for rapid delivery at scale … given HMRC’s possession of the relevant data and its general capabilities.” The information held by HMRC about the income of the self-employed, however, is “significantly less detailed” than about the income of those in employment. Both cohorts submit income data to HMRC but, in contrast to the employed population, in respect of whom the Revenue receive a complete set of information about their income and tax and national insurance payments from employers under the PAYE scheme contemporaneously and “in real time”, data on the self-employed is collected through their annual self-assessment tax returns.

Secondly, “for obvious reasons” there was “a pressing public interest” in setting up the SEISS as quickly as possible, given the significant public health restrictions introduced to combat the pandemic.

Thirdly, those concerns had to be balanced against the risk of error, fraud and perverse incentives. Given the scale of the public funds expected to be paid out, it was “essential to minimise fraud”. On the other hand, given the speed with which it was set up, “there was only a limited amount which could be done to protect against fraud risks”. One consequence was that it was decided that the newly self-employed, who were due to file their first tax returns covering earnings in the 2019/20 tax year, should be excluded from the scheme altogether.

Fourth, it was considered important to design a scheme that allowed the self-employed to continue to work during the pandemic where possible. Whereas the aim of CJRS was to provide a subsidy to businesses to prevent them making employed workers redundant and, as a result, the grant was only paid in respect of workers who were not working, it would have been “unnecessary and counter-productive” to require the self-employed to abstain from working since many of them were able to continue working and thereby sustain economic activity. For that reason, the SEISS was designed in a way that allowed people to receive the grant whilst maintaining a certain level of income, subject to meeting the eligibility criteria.

Finally, Ms Kantor stressed that SEISS needed to fit into the wider structure of government support, which reflects “an existing balance of different social and economic policy objectives”.

Describing the scheme and its rationale, Ms Kantor set out the qualifying conditions and a number of purposes, including focusing on those who most needed government assistance, minimising the risk of fraud, and ensuring its practicability. Basing the scheme on self-assessment tax returns filed before 23 April 2020 provided a safe, secure and speedy way to assess the validity of individual’s trading profits and therefore the amount of grant they were able to claim. Given the tendency of self-employed profits to vary markedly for many reasons, the use of average trading profits (“ATP”) as the basis of calculation was intended to “even out” the profits over the years of trading so as to provide the most accurate reflection possible of the profits generated by the business. No costings were done for the alternative models proposed by the appellants at the time the SEISS was implemented, but it was plain that any system that required individual qualitative assessment of evidence submitted would involve additional work and cost.

In his statement, Mr Hacon described in detail the design and execution of the scheme, emphasising the scale of the challenges involved. Having identified the existing data systems held by HMRC, he continued (at paragraphs 11 to 18):

“Using this data was the only practical approach to ‘historic’ data that had been submitted and validated, and so would be impossible for applicants for the purposes of claiming SEISS. It would not have been possible to capture new sets of self-assessment information in the time available…”.

“The functionality required for the SEISS IT system … was very extensive: the ability to receive potentially millions of applications; to process those applications against HMRC’s existing data; to incorporate eligibility and fraud checks; and to pay out grant sums quickly, accurately and efficiently. It was, however, constrained by the practical requirement to use data which HMRC held in the course of HMRC’s usual statutory functions.”

“At March 2020, HMRC simply did not have a system which had that functionality as it was a unique ask, outside the remit of HMRC’s core statutory functions. It was obvious from the start that such a system would have to be developed. A large number of teams, from across government, led by HMT and HMRC, worked at great speed to do so….”

“An enormous amount of information and coding was brought into one system design, in order to create a streamlined, functional, efficient and reliable system that could operate efficiently at speed. The scale of this was beyond anything previously attempted by HMRC to such a short timescale.”

Mr Hacon continued (at paragraph 19):

Turning to the details of the scheme as set out in the First Direction, its purpose was summarised as follows in paragraph 2:

A separate mechanism was provided for persons subject to “the loan charge”, an anti-avoidance measure introduced in earlier tax legislation to address tax losses incurred by so-called “disguised remuneration” schemes. It is unnecessary to set out the details of the loan charge any further although for reasons set out below it is necessary to describe how persons subject to the loan charge were treated under the First Direction.

The “profits condition” referred to in paragraph 4.2(g) was defined in paragraph 5 of the First Direction:

In her statement, Ms Kantor explained the scheme in these terms:

The mechanism for calculating the quantum of grant was set out in paragraph 6 of the First Direction:

The choice of an average of trading profits in the three preceding years, as opposed to, for example, trading profits in 2018-19 alone, was explained by Ms Kantor as follows:

Although not subject to challenge in these proceedings, the Second Direction signed on 1 July 2020 is relevant to this appeal because of its modifications to the SEISS. Part 3 of the Second Direction extended the SEISS to two categories of persons not included under the First Direction – military reservists who had not qualified because they undertook military service in the tax year 2018-19 and persons who had not qualified because of the effect of infant care, pregnancy or maternity on that person’s trading profits or total income for the tax year 2018-19 (hereafter referred to as “2018-19 parents”). Under the Second Direction, these two categories could make a claim under the SEISS in respect of both the first and second grant, provided they met revised profit conditions which, in the case of a 2018-19 parent, were the same conditions as applicable to a person subject to the loan charge. In those circumstances, under paragraph 8(a) and (b) of the Second Direction, the amount of grant payable to a 2018-19 parent was calculated in the same way as for a person subject to the loan charge in paragraph 6.2(b) of the First Direction.

It was Ms Kantor’s evidence that the change to the SEISS to include 2018-19 parents was possible because it did not involve substantial amendment to the scheme but simply the application of the “loan charge” mechanism to the cohort of 2018-19 parents.

The development of the scheme, and the extent of consideration given to those whose claim was affected by circumstances relating to maternity, can be traced through the series of Ministerial Briefing Notes presented to the Chancellor and disclosed (subject to some redactions) in these proceedings. The Notes illustrate the careful and detailed work undertaken by officials at speed when developing the scheme.

The first Note, dated 22 March 2020, outlined the options for the scheme, including whether to use 2018-19 tax returns as the sole reference point or an average of returns over three years. It did not include an equalities assessment and although a number of “hard cases” were identified these did not include persons whose profits in the reference years had been reduced as a result of maternity or other caring responsibilities.

The second Note, drafted two days later on 22 March, did include an equalities assessment. By that stage, as the Note made clear, the Chancellor had indicated that ideally the award should be based on a three-year average of profits. In a section of the Note headed “Equalities Assessment and Family Test”, the authors noted that the proposed draft would mean that young people were less likely to benefit, although this was mitigated by the availability of other support, and that “Pakistani and Bangladeshi individuals were more likely to be self-employed than the general population”, although it was not thought that this gave rise to an equalities issue. The authors observed that they were not aware of any other impacts on protected characteristics, adding that, while 66% of self-employed persons were women, “we do not think female individuals are any more or less likely to meet the eligibility conditions than males, so this does not give rise to any equalities issues.” The Chancellor was asked to confirm that he had noted the equality impacts of the policy.

The third Note, prepared a day later on 25 March, which focused on the basis of averaging eligibility criteria and the impact on persons affected by the loan charge, made no reference to equalities issues. The fourth Note, dated 2 April, however, addresses the issue of maternity in the following section headed “Self-employment for part of the year”:

The fifth Note, dated 22 April, eight days before the First Direction was signed, was headed “SEISS: Final Steers for Treasury Direction”. It asked the Chancellor to make final decisions about the scheme by 27 April. Under the heading “Maternity leave”, he was informed:

(ii) is a rule which applies to anyone who did not file a tax return for 18-19, and does not outwardly discriminate against those on maternity leave. However, it is clear that there are hard cases in this group, and you may wish to address this.

(iii) is a consequence of the decision to average across three years to determine the level of award, which protects those who had unusually low income in 18-19. In previous advice, you ruled out pro-rating partial years of trading, which would have had a small mitigating effect on this issue, but led to oversubsidisation of many others.

The sixth Note, dated 5 May 2020, was headed “SEISS: Options for groups with periods of disrupted trade and special tax treatment”, and provided further information and options for accommodating reservists on longer term call ups and those with periods of maternity leave within the scheme. Within the introduction, the Chancellor was advised:

Under the heading “Policy options”, the Note continued:

The Note continued:

The seventh Note, dated 12 June 2020, the final document in this form disclosed in these proceedings, provided follow-up advice on options for groups with periods of disrupted trade. The authors stated:

The claim for judicial review was filed on 13 July 2020. The first claimant, The Motherhood Plan, also known as “Pregnant Then Screwed”, is a registered charity. Its aims were described in the Background and Grounds of Judicial Review as being to end discrimination faced by women and mothers by campaigning to change legislation, raising awareness in the media and working with employers to change business practice and culture. The second claimant, Ms Kerry Chamberlain, worked as a self-employed energy analyst. In the tax year 2017-18, she took a 39-week period of maternity leave after the birth of her second child and, in the following tax year, she took a further 39-week period of leave after the birth of her third child. As a result of her periods away from work, her trading profits were reduced. The defendant was the Chancellor of the Exchequer. By the claim, the claimants sought a declaration that the SEISS unlawfully discriminated against women who had taken a period of absence from work relating to pregnancy or maternity, contrary to Article 14 of the Convention read with Article 1 of the First Protocol, asserting that the discrimination had taken the form of what the judge called “conventional” indirect discrimination and also “Thlimmenos discrimination”. They also sought a declaration that the defendant had breached his public sector equality duty under section 149 of the Equality Act, and a mandatory order requiring him to reconsider his duties regarding the scheme. HMRC were joined to the proceedings as an interested party. On 22 September 2020, permission to apply for judicial review was granted, and the application was heard by Whipple J in January 2021.

The judge began by describing the scheme and tracing its development through the Ministerial Briefing Notes summarised above. Having set out the parties’ respective submissions, she started her analysis of the issues by identifying some key points about the scheme, in particular that it had been devised under “enormous pressure of time, with the aim of getting payments out to the self-employed as soon as possible”; that “the quantum of payments under the scheme had to be based on some proxy” because it was not a flat-rate scheme but rather aimed at reflecting the established profitability of the business; that the proxy chosen was an average of the last three years’ trading profits shown in tax returns filed for those tax years; and that the application process could be “light touch” because it depended on accessing data already held.

The judge then considered the claimants’ challenge to the scheme on the grounds of unlawful discrimination contrary to Article 14. She recorded that it was common ground that the approach to indirect discrimination claims in the context of Article 14 was as described by Lady Black in by R (Stott) v Secretary of State for Justice [2018] UKSC 59,[2020] AC 51, at para. 8:

The judge observed that there was no dispute as to the first two elements. The circumstances plainly fell within the ambit of Article 1 of the First Protocol. Secondly, women who had been on maternity leave in the last three years shared a “protected characteristic” which qualified as “other status” for Article 14 purposes. As to whether the second claimant was in an analogous situation with others who had been treated differently, the judge (at para. 54) expressed some uncertainty as to how the asserted “uniqueness” described by the claimants to pregnancy and childbirth fitted with Lady Black’s third step but concluded:

The judge then considered the claimants’ alternative argument based on Thlimmenos, that:

The judge observed that the Treasury’s answer to the challenge depended in large part on a series of cases, namely Barry v Midland Bank PLC, Uppingham Trustees v Shillcock and, in particular, the recent decision of the Divisional Court in R (Adiatu) v HM Treasury, all of which we discuss more fully below. Whilst acknowledging the differences between those authorities (all considered below) and the present case, Whipple J concluded that the comments made by the Divisional Court in Adiatu were relevant, and this led her to conclude that she was “not persuaded that Article 14 is breached, whether on the basis of indirect discrimination of the conventional type or on a Thlimmenos analysis”.

The judge proceeded, however, to consider the question of justification in the event that there had been a breach. She recorded that on this point there was also common ground as to the legal approach, namely that the question was whether the measure was “manifestly without reasonable foundation” and, in determining whether a measure met that test, the court should proactively examine whether the foundation is reasonable. The Treasury relied on five separate justifications – the purpose of the scheme, policy delivery, the risk of fraud, the risk of perverse effects, and value for money – all of which were accepted by the judge. She concluded that, whether the various justifications are taken separately or in combination, the defendant's decisions were reasonable especially when judged in context and that the design of the scheme, specifically in the way the payments were calculated by reference to ATP, was not manifestly without reasonable foundation.

The judge then considered the second ground of challenge, based on the public sector equality duty, and concluded that there had been no breach of the duty. There is no appeal against that decision.

The appellants put forward three grounds of appeal against the judge’s dismissal of the application for judicial review.

First, they contend that she erred in concluding that there was no indirect discrimination against women who had not worked for reasons relating to pregnancy or maternity. They assert that she misapplied the test for indirect discrimination under Article 14, namely whether there had been a disproportionate effect on this class of applicant, not whether there were “hidden barriers” to entry to the scheme.

Secondly, the appellants submit that the judge erred in failing to consider whether there was a failure to treat differently persons whose situations are significantly different (so-called “Thlimmenos” discrimination). The question whether the circumstances of a group with a particular status call for that to be treated differently must be asked by reference to how the eligibility and quantum criteria applied to that group. Where, as here, those criteria are themselves backward looking to what has been earned in the past, then it is the impact on the group in the period during which the eligibility and quantum criteria apply which falls to be considered. There is no authority or reason in principle precluding requirements or conditions for eligibility to a particular benefit which is based on historical factors from potentially requiring justification in so far as they fail to treat some groups differently from others.

Thirdly, the appellants submit that the judge erred in her approach to justification by applying an excessively broad approach and failing to recognise that the degree of deference is reduced where (i) the measure involves adverse discriminatory effects, (ii) those effects relate to the core attribute of sex and (iii) the justifications advanced are ex post facto in nature. When the correct approach is applied, none of the five justifications put forward by the respondents stands up to scrutiny.

The appellants rely in the alternative on two kinds of discrimination – indirect discrimination and “Thlimmenos”discrimination. In the case of both, a prima facie discriminatory situation is capable of justification. In this section we consider only whether such a situation has been established. We take them in turn.

Article 14 of the Convention provides:

As is apparent from its terms, Article 14 can only be considered in conjunction with one or more of the substantive rights or freedoms conferred by the Convention or its protocols. In the present case, it is common ground that entitlement to benefit under SEISS falls within the ambit of Article 1 of the First Protocol to the Convention, which we need not set out.

The most recent authoritative statement of the nature of indirect discrimination in the context of Article 14 of the Convention is in paras. 49-53 of 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. At para. 49 Lord Reed quotes a passage from the judgment of the European Court of Human Rights (“the ECtHR”) in Guberina v Croatia(2018) 66 EHRR 11 as follows:

An important difference between indirect discrimination in the context of Article 14 and indirect discrimination in EU law (which continues to underpin domestic law) is that EU law proscribes only discrimination against persons or groups with specified protected characteristics, whereas under the Convention discrimination may be on the basis not only of the various specified characteristics but “other status”. But even apart from that, the formulations quoted above are rather more generally expressed than the definition of indirect discrimination adopted in the EU legislation and jurisprudence. The history of that definition is helpfully set out in the judgment of Lady Hale in Essop v Home Office[2017] UKSC 27, [2017] ICR 640, at paras. 18-22, and we need not repeat it here. The current version (see para. 21 of Lady Hale’s judgment) is:

The EU definition of indirect discrimination is in substance reproduced in English law by section 19 (2) of the Equality Act 2010, which reads:

Notwithstanding those differences in formulation, it is clear that the same principles underlie the concept of indirect discrimination in the Convention, EU and domestic contexts. Broadly speaking, the concept of a “measure” does the same work as a “PCP” in EU law, and the requirement that it has “disproportionately prejudicial effects” (or “affects a disproportionate number of members”) on the relevant group essentially corresponds to the requirement that it puts members of that group at “a particular disadvantage”. As Lord Reed notes, the Strasbourg case-law is not yet fully developed, and some differences in approach or application may emerge at the margins, but the essential similarities are such that it is in our view legitimate in an Article 14 case at least to have regard to the EU and domestic jurisprudence.

That being so, we think that it is useful to quote from a passage in the judgment of Lady Hale in Essop in which she identifies a number of “salient features” of indirect discrimination, as defined in the EU and domestic legislation. Three of those features are particularly material to the present issue:

A potential ambiguity about terminology should be noted. The status which is characteristic of the disadvantaged group is described in Article 14 as the “ground” of discrimination, without distinction between direct and indirect discrimination. But the terminology of “grounds” is sometimes used in a sense specific to direct discrimination, to connote the reason for the treatment complained of, whether in the sense of the criterion employed or the motivation of the relevant decision-taker. In the context of indirect discrimination, the reason why, in that sense, the putative discriminator acted as they did is of course irrelevant.

At para. 40 Whipple J summarised the way that the appellants put their case of indirect discrimination as follows:

First, the “general policy or measure” (or, in Lord Reed’s formulation, “neutrally formulated measure”) which is said to have a disproportionately prejudicial effect is the adoption of average trading profits over the past three years as the basis of calculation for the SEISS. We will refer to that (with apologies for using the word “measure” in two different senses) as “the ATP measure” and to the tax years in question as “the measurement period”.

Second, the group on which the ATP measure is said to have that effect is “women who have not worked in the preceding three tax years for maternity reasons” (i.e. the demands of pregnancy, childbirth and looking after a young baby). That is the “salient attribute or status”, in the language of DH – or, in Lord Reed’s formulation, the “characteristic” which is the ground of discrimination. Before us the shorthand “recent mothers” was used. (Another shorthand sometimes employed was recent mothers who had taken “maternity leave”; that is acceptable, provided it is understood that the phrase has no formal meaning in the context of a woman who is self-employed.) There was some discussion at the hearing of what was comprised in the reference to recent mothers having “not worked” in the measurement period. In principle the effect of which the appellants complain would be produced by any reduction for maternity reasons in the working hours of recent mothers, whether that took the form of periods when they did not work at all or periods when they worked fewer hours (or even the same hours but less effectively). But that refinement does not affect any of the issues of principle.

Third, the nature of the prejudicial effect of the ATP measure on recent mothers is that a period of non-work (or reduced or less effective work) during the measurement period will generally mean that their profits in that period will not be representative of the earnings that they would have been expected to earn in the current year but for the pandemic (her “hypothetical no-Covid earnings”) and thus will result in SEISS payments that do not truly represent the loss of earnings that it has caused. We will call that “unrepresentativeness” for short. The appellants offer no specific evidence of that effect, still less any evidence by which it could be quantified, beyond a very few individual cases. However, they say that its existence can be confidently inferred as a matter of common sense and common experience: recent mothers almost always take significant time off work, and in the typical case that is bound to impact the profits of recent mothers who are self-employed.

Fourth, although of course the profits of other SEISS claimants in the measurement period might also be unrepresentative, for any number of reasons, the appellants’ case is that in the case of recent mothers the risk of unrepresentativeness is peculiar and specific, over and above the level of risk from the ordinary vicissitudes of self-employment, and will eventuate in the great majority of cases: to put it in the language of the EU jurisprudence, the use of the ATP measure puts recent mothers at a particular disadvantage.

It is worth noting that although the appellants chose to include in their definition of the group the fact that the recent mothers in question have “not worked” in the measurement period, it would have been possible to omit that element and define the group simply as all women eligible for the SEISS who have been pregnant or given birth in the period in question, irrespective of whether they stopped work – or indeed as all women eligible for the SEISS, irrespective of maternity. If the appellants’ case is otherwise good, the ATP measure would still in principle have a disproportionate impact on the groups so characterised, albeit that it would be diluted by the presence in it of women who had not stopped work (or who were not recent mothers). It would also have been possible to define it as “recent parents”, in recognition of the fact that sometimes, albeit more rarely, self-employed fathers will find their ability to earn adversely impacted by the demands of looking after a young baby.

Mr Milford argued that the Appellants’ way of putting the case meant that there was no comparator group, which is essential in any claim of indirect discrimination: in as much as the disadvantage that Ms Chamberlain was relying on was that her benefits would be less than they should have been, her comparator was in effect herself. It will be apparent from the foregoing that we do not believe that that is a valid criticism. The important point to appreciate is that the comparative disadvantage complained of is not the unrepresentative payments as such but the enhanced risk of unrepresentative payments: see para. 14 above.

Mr Milford also argued that there might be other groups at special risk of unrepresentativeness. Obvious examples are self-employed people who suffered serious illness for part of the measurement period but had recovered by the current year or who had had to take time off to care for an elderly or ill representative but where those responsibilities had ended. Mr Bunting acknowledged that in principle such groups might also be entitled to claim prima facie indirect discrimination, provided they were recognised as having a “status”, but he submitted that that was not a reason for denying such a claim based on recent mothers as a group. We agree.

Whipple J held that in the circumstances of the present case there was no prima facie indirect discrimination. Her reasoning was based in part on three decisions relied on by the Respondent – Barry v Midland Bank plc [1999] UKHL 38,[1999] 1 WLR 1465, Trustees of Uppingham School Retirement Benefits Scheme v Shillcock [2002] EWHC 641 (Ch), [2002] IRLR 702, and R (Adiatu) v HM Treasury [2020] EWHC 1554 (Admin), [2020] PTSR 2198. We shall have to consider those decisions more fully later. It is enough to say at this stage that they are all cases where, in different circumstances, a claimant was held not to be entitled to rely on problems caused by his or her past level of earnings as the basis of a claim of indirect discrimination.

At paras. 57-61 of her judgment, Whipple J summarised each of those decisions, though her principal focus was on Adiatu. She acknowledged that there were differences between their circumstances and those of the present case, and that they were concerned with claims under domestic and/or EU law rather than under the Convention. But she says that they nevertheless provide “valuable pointers … to the analysis of cases like this”. Paras. 62-64 read:

The core reasoning is in para. 64, though it must be read in the light of what the Judge takes from Adiatu and Barry in paras. 62-63. It needs to be unpacked with a little care. The argument is best followed through sequentially.

The first sentence simply states the conclusion, for which the reasons follow in the rest of the paragraph. We should say that the reference to “approaching the matter on a conventional analysis” is simply in contradistinction to “Thlimmenos discrimination”, which we consider separately.

The second and third sentences make the point that there are no “hidden barriers”, either as regards eligibility for the SEISS or as regards the quantum of any entitlement. That picks up a similar statement in Adiatu (see para. 149 of the judgment, quoted at para. 90 below), which itself takes the phrase from para. 25 of Lady Hale’s judgment in Essop (see para. 57 above); but we are not sure what it means in this context. Even if it is in some sense true, it is not an answer to the claim that the adoption of the ATP measure has a disproportionate impact on recent mothers. We say a little more about this at para. 94 (2) below.

The fourth sentence begins with the observation that “the same rule applies to all”. That cannot of course be an answer to the appellants’ case: the essence of prima facie indirect discrimination is that the same rule applies to everyone but that it has a disproportionate impact on the group in question. We are sure the Judge was well aware of that, and we assume she was merely emphasising that this was not a case of directly different treatment; but that in itself does not advance the argument.

The statement in the second part of the fourth sentence that “it is no harder for a woman who has been on maternity leave to qualify or calculate their payment, than someone who has not” is true, but it does not address the essence of the appellants’ case. The complaint is not that recent mothers are disqualified from receiving a SEISS payment, or that calculation of their entitlement is more difficult, but that the adoption of ATP as the measure of their entitlement puts them at a particular disadvantage: see para. 63 above.

That brings us to the final sentence of para. 64. We should say by way of preliminary that the reference to some claimants receiving “lower payments than others” is not quite right. We are not in fact sure which “others” are being referred to; but in any event, as discussed above, the appellants’ point is not as such that recent mothers will receive lower SEISS payments than other self-employed people but rather that, compared with such people, they are at disproportionate risk of receiving payments which are unrepresentative of their hypothetical no-Covid earnings. However, that need not undermine the judge’s essential point, the essence of which is that the prejudice complained of is immaterial because it “reflects the fact of lower earnings in past years”.

We should note that Whipple J appears to have accepted that the use of the ATP measure did in fact lead to disproportionately unrepresentative payments to recent mothers, despite the limited nature of the evidence (see para. 62 above): her point was that that was immaterial. That seems to us realistic, and in any event the point is not the subject of a Respondent’s Notice.

At first sight, the final sentence of para. 64 of Whipple J’s judgment is puzzling. Lord Reed’s formulation of the circumstances giving rise to prima facie indirect discrimination in the Article 14 context requires only that the measure in question has “disproportionately prejudicial effects on a particular group”: see para. 49 of his judgment in SC. As noted above, Whipple J appears to have accepted that the ATP measure has a disproportionately prejudicial effect on recent mothers. That being so, it is unclear why it should matter that it “reflects [i.e. results from] the fact of lower earnings in past years”. As a matter of principle, the fact of the disproportionate effect is enough. As Lady Hale says at para. 24 of her judgment in Essop:

However, it is fact clear that Whipple J believed that her conclusion followed from the decisions in Barry, Shillcock and Adiatu. It is therefore necessary to consider those cases with some care.

We start with Barry, which is the only decision that is binding on us. The claimant was a female employee who was dismissed for redundancy. She had originally worked full-time but had recently changed to part-time working (described as “key-time”) following the birth of her child. The employer operated a severance scheme under which redundant employees were entitled to a lump sum based on their final salary at the date of dismissal multiplied by their years of service. That method of calculation was less advantageous to the claimant as a part-timer than one where the multiplicand was based on her average earnings over her career. She claimed that it was indirectly discriminatory, contrary to the Equal Pay Act 1970 and EU law, because women were more likely, on account of maternity, to have a pattern of working under which (like her) they switched from full-time to part-time working. The House of Lords by a majority of 4:1 held that prima facie indirect discrimination had not been established.

Identifying the ratio of the majority in Barry is not entirely straightforward: all four members of the majority gave speeches, but none of them explicitly adopted the reasoning of any of the others. The best starting point is to see how the employer put its case. The submissions of its counsel, Mr Patrick Elias QC, are conveniently summarised in the speech of Lord Nicholls (who was in fact the dissentient on this issue, though he found for the employer on justification), as follows:

Lord Slynn gave the first and longest speech. His essential reasoning as regards prima facie indirect discrimination is at pp. 1468-9 and reads as follows:

Lord Steyn’s speech is very short. Having observed that he had found the issue difficult, he says simply, at pp. 1478-9:

Lord Hoffmann approaches the issue in rather a different way from Lord Slynn. Most of his speech is addressed to establishing that the employer’s scheme was not discriminatory in itself and that in truth, as Lord Steyn had said, the claimant’s complaint was that she would have done better from a different kind of scheme which gave a severance payment based on total hours of service. However, he concludes, at p. 1481 A-C:

In his short speech, Lord Clyde acknowledged that it might appear that the scheme was disadvantageous to women compared to men. But, he continued (p. 1481 E-F):

On the basis of those passages, we believe that the ratio of the majority can be summarised as follows. It is essential to a claim of indirect discrimination that the group to which the claimant belongs should be differently treated from persons not in that group. In deciding whether there has been such a difference of treatment it is necessary to identify the true substance of the measure which gives rise to the claim, and that may involve a consideration of its purpose. The purpose of the scheme in Barry was to mitigate the impact on employees of the loss of the income that they would have received but for their dismissal. That being so, it cannot be regarded as a difference in treatment between part-time and full-time employees to leave out of account, in calculating the benefits payable, a factor (i.e. levels of earnings over the employee’s career as a whole) which was of its nature irrelevant to the loss which it was the purpose of the scheme to mitigate: what the part-time employees had lost was their part-time earnings. That reasoning is clearly expressed in the first paragraph of the passage which we have quoted from Lord Slynn’s speech, and in Mr Elias’s submissions as summarised by Lord Nicholls, and it is sufficiently clear that it was shared by the other members of the majority. In terms of the overall structure of indirect discrimination as expounded in paras. 53-58 above, the effect of the reasoning is that in such a case the claimant’s group is not prejudiced or “put at a particular disadvantage”: they have, in the material respects, been treated in the same way as everyone else.

It is instructive, particularly in view of Lord Nicholls’ eminence as a discrimination lawyer, to consider the reasons for his dissent. He acknowledged that Mr Elias’s argument “has some attraction” but he declined to accept it for what he described as “two practical reasons” which are in fact related: he did not regard it as “plain and obvious” that Mrs Barry’s career earnings were immaterial to the purpose of the scheme, and he thought accordingly that that issue should be considered in the context of the justification argument, where the burden of proof was on the employer (see pp. 1473H-1474A). It appears, therefore, that he accepted the validity in principle of Mr Elias’s approach but believed that it should be reserved for plain cases. There is accordingly no gulf of principle between his reasoning and that of the majority.

A not dissimilar situation to that in Barry was considered by the CJEU in Stadt Lengerich vHelmig C-399/92, [1996] ICR 35. The effect of that decision is conveniently summarised by Lord Nicholls in Barry at p. 1474 C-E as follows:

Once the ratio of Barry is understood, it can be seen that, contrary to para. 63 of Whipple J’s judgment, itprovides no analogy to the present case. The purpose of SEISS is to compensate self-employed persons for their loss of profits in the current year as a result of the pandemic. The ATP measure works by using past profits to represent, in however rough-and-ready a manner, their likely hypothetical no-Covid profits. If its use in the case of new mothers produces results which are disproportionately unrepresentative of those profits, as compared with others, that necessarily puts them at a particular disadvantage. By contrast, Mrs Barry’s previous (whole-career) earnings were irrelevant to the earnings that she would have received but for her dismissal.

We need not consider Shillcock in detail. It concerned a term in an employer’s pension scheme under which membership was only open to employees with a minimum level of earnings. That level was not attainable by part-time employees, who were disproportionately female. The claimant contended that for that reason it was indirectly discriminatory, contrary to section 62 of the Pensions Act 1995 and EU law. Neuberger J held, purporting to follow Barry, that there was no prima facie indirect discrimination. He avowedly found the point difficult and felt uncomfortable with the result, and he went on to hold that any prima facie indirect discrimination would in any event have been justifiable. We do not believe that it would be useful to decide whether his conclusion on the prima facie indirect discrimination issue was correct. His analysis of Barry broadly corresponds, as we understand it, to our own; and the only question is whether it was correctly applied to the facts of the case before him, which are very different from those of the present case.

We turn, finally, to Adiatu. This is a decision of the Divisional Court (Bean LJ and Cavanagh J) which considered a number of challenges to the provisions designed to provide support to workers affected by the pandemic. For present purposes we are concerned only with one of those challenges, which concerned statutory sick pay (“SSP”). The claimant’s case was that the rate at which SSP was set indirectly discriminated against women and members of ethnic minorities, contrary to EU law, because

The Court rejected that challenge. At para. 141 it says:

We respectfully agree with the conclusion of the Divisional Court and with the substance of its reasoning in para. 149. The essential point being made is the same as that in Barry, although Mr Adiatu’s claim was in truth a good deal more ambitious than Mrs Barry’s. SSP is not a benefit the purpose of which is to mitigate the effects of low income, still less the effect of the pandemic on those with low income; and the fact that a claimant has low earnings is simply immaterial. That being so, the fact that a higher rate of SSP would be beneficial to those on low incomes cannot found an argument that they are materially disadvantaged by the actual rate. That is what the Divisional Court is referring to when it says that “the absence of other financial resources … is not caused or related to the rate of SSP in any way”.

Whipple J believed that Adiatu supported her conclusion that new mothers’ earnings in the measurement period could not be relied on to found a case of particular disadvantage: see para. 62 of her judgment. But again, once its ratio is understood, we do not believe that it does so, for the same reason as we give at para. 87 above. The purpose of SEISS is to compensate self-employed people for the loss of the earnings that they would have received in the current year but for the pandemic and to use past earnings as the measure of those lost hypothetical earnings. In those circumstances, the past earnings in question are not immaterial: on the contrary, they are crucial.

It follows that in our view Adiatu does not afford Whipple J’s conclusion the support which she takes from it at para. 62 of her judgment.

Before we leave Adiatu, we should note two other points about Whipple J’s reasoning in relation to it:

For those reasons we respectfully believe that Whipple J was wrong to find that the use of the ATP measure did not constitute prima facie indirect discrimination.

At para. 48 of his judgment in SC, following his consideration of direct discrimination, Lord Reed says:

The nature of the appellants’ case under this head, and Whipple J’s reasons for rejecting it, appear from paras. 65-67 of her judgment, which read:

As appears from that passage, Whipple J’s reasons for rejecting this formulation of the appellants’ case largely overlap with her reasons for dismissing the claim of indirect discrimination; and we are inclined to believe that they are flawed for the same reasons as we have already given. But we do not believe that it is necessary to reach a definitive view on that point. Our conclusion as regards indirect discrimination means that the adoption of the ATP in the case of new mothers must be justified in any event; and that issue will not be any different whatever form the discrimination took.

The proper approach to justification in cases of indirect discrimination starts with the principle expressed by Lord Reed in R (SG) v Secretary of State for Work and Pensions)[2015] UKSC 16, [2015] 1 WLR 1449 at para. 13:

In Bank Mellat v HM Treasury (No.2)[2013] UKSC 39, [2015] 1 WLR 1449, (at para. 74), Lord Reed structured the assessment of proportionality in four distinct elements:

In his oral submissions Mr Milford emphasised Lady Hale’s statement in SG that it is the measure rather than its disproportionate impact that has to be justified. But, at the risk of stating the obvious, that does not mean that the disproportionate impact is irrelevant: justification involves weighing the importance of the measure against its discriminatory impact.

At the hearing before Whipple J, there was common ground as to the correct legal approach to the issue of justification. Unsurprisingly, therefore, she felt able to summarise that approach in a few brief paragraphs. She identified the issue to be determined as whether the measure met the test of “manifestly without reasonable foundation”. That was the appropriate test because, even though the alleged discrimination went to a “core characteristic (sex)”, the case concerned state benefits: Humphreys v Revenue and Customs Commissioners[2012] UKSC 18, [2012] I WLR 1545. In addressing that issue, the court should proactively examine whether the foundation advanced by the defendant was reasonable. The margin of appreciation was broad because the measure was a general one, aimed at meeting economic and social needs.

The judge’s succinct summary of the law was entirely appropriate at the time of her judgment in the light of the then state of the Supreme Court authorities. Having declared in Humphreys that “manifestly without reasonable foundation” was the correct test, and reiterated that approach in subsequent cases in 2015 and 2016, the Supreme Court had confirmed it again in emphatic terms in R (DA) v Secretary of State for Work and Pensions[2019] UKSC 21, [2019] 1 WLR 3289. At para. 65 of his judgment in DA, Lord Wilson JSC said:

As it turned out, however, there was a doubt. Just over two years later, a few days before the hearing of the appeal in the present case, the Supreme Court revisited this issue again in its judgment in SC. The case concerned the entitlement to the individual element of child tax credit which under the relevant regulation was limited to a maximum amount, calculated as the amount payable in respect of two children. In his judgment (with which the other six Justices agreed) Lord Reed, identified at the outset (in para. 2) a number of important questions arising in the case, including:

Lord Reed proceeded to conduct an extensive review of the Strasbourg case law. He summarised the approach of the European court in these terms:

Lord Reed noted that the European Court had originally developed a requirement for “very weighty reasons” needed to justify a difference in treatment based on sex or gender (para. 101). He proceeded to trace the development of this requirement through the Strasbourg cases relating to the other so-called “suspect” grounds of discrimination (birth status, nationality, sexual orientation, race of ethnic origin, religious belief , disability) and noted that the counterpart of that strict approach was less strict scrutiny of differences of treatment on other grounds (including age, immigration status, prisoner status and marital status) (paras. 104 to 114). He then observed that the European court’s approach in cases under Article 14 where it has used the phrase “manifestly without reasonable foundation” had been consistent with the points summarised above. He traced the use of the phrase starting with the decision in Stec v United Kingdom app no 65900/01, (2006) 43 EHRR 74, a case concerning differences in entitlement to a state benefit on the grounds of sex, and illustrated how the points he had identified relating to justification generally were mirrored in the jurisprudence concerning the phrase. He then considered whether the approach had been changed by a more recent line of authorities starting with JD and A v United Kingdom app no 32949/17, [2020] HLR 5, noting the submission made by the appellants in SC that the reasoning in that case established a simple rule that complaints of discrimination on “suspect” grounds fell outside the scope of the wide margin and “manifestly without reasonable foundation” approach.

At para. 142, Lord Reed summarised his analysis of the Strasbourg jurisprudence in these terms:

Lord Reed then considered the approach of our domestic courts to justification. Although the concept of the margin of appreciation was specific to the European court, domestic courts had generally endeavoured to adopt an analogous approach, partly to keep pace with Strasbourg jurisprudence as it developed over time but also because of the importance of respecting the separation of powers between the judiciary and the elected branches of government. Nevertheless:

He then considered the line of Supreme Court authorities starting with Humphreys. He cited Lady Hale’s observation in that case (para. 19) that:

Having considered the cases following Humphreys, up to and including DA and Lord Wilson’s emphatic endorsement of the “manifestly without reasonable foundation” approach, he set out his final conclusions on that approach at paras. 157-161:

Lord Reed added this warning (at para. 162):

The effect of the decision in SC was summarised by Andrews LJ at para. 34 of her judgment in R (Salvato) v Secretary of State for Work and Pensions[2021] EWCA Civ 1482 (handed down following the hearing of the present appeal) in which this Court allowed an appeal from the decision of Chamberlain J (which was considered by the judge in the present case):

As explained above, none of this learning was available to the judge in the present case who proceeded to apply the “manifestly without reasonable foundation” test to the measure under consideration. On the five justifications advanced by the Treasury, she reached the following conclusions.

Purpose: The stated purpose of the scheme – to provide support for self-employed people whose businesses were adversely affected by the pandemic – was “reasonable” and it was “reasonable to seek to advance that purpose by calculating payments by reference to average trading profits”. The claimants’ proposal that the purpose would have been better achieved by designing in an adjustment for women who had been on maternity leave went far beyond the stated purpose and “would, in effect, be to give it a new purpose, namely correcting perceived inequalities in the past, unrelated to the scheme” (paras. 72-73).

Policy delivery: “A move away from a method of calculation based on actual profits, reflected in the data collected and held by HMRC, would have involved expense and led to delay, both of which were antithetical to the required quick delivery. The defendant had good reason for adopting an approach that was simple and quick, which used one rule, one approach, applicable to all. IT overheads and manual intervention [were] kept to a minimum so that the Scheme could be implemented quickly. This was not unreasonable” (paras. 77).

Risk of fraud: “The desire for claims to be verifiable by reference to data already held by HMRC is a powerful justification for the design of the Scheme. It meant that the claims could be automated, which achieved speed and cost savings, and the risk of fraud was reduced. The concern about fraud is canvassed in the ministerial submissions right from the first, and the Defendant's evidence confirms that it was at the heart of the Scheme's design from the outset. I accept that evidence. The Claimants' suggested alternative, based on a woman self-certifying that she had been on maternity leave for certain periods, would in the Defendant's view have been "wide open to fraud", because such certifications could not realistically have been checked and for that reason would not have been acceptable. That is an understandable position to take …. Overall, it was reasonable for the Government to accept some areas of fraud risk while seeking to minimise that risk in other areas. The design of the Scheme involved a balance of various interests and factors (paras. 79-80).

Perverse effects: It was “not possible to arrive at a solution to the claimants’ problems which does not in itself create a range of hard cases and anomalies, which fall on the wrong side of the line … For every tweak to the simple formula, a new cohort of hard cases would have been created which fell on the wrong side of the tweaked line. The bright line solution was preferred. This, again, was a political decision for Government to make. The making of some changes (tweaks) for some groups (e.g. reservists and 18/19 parents) does not require wider inroads. This is a political decision, for the architects of the Scheme. It is not a matter for lawyers” (paras. 81-82).

Value for money: The claimant’s proposals would have cost money. Simplicity was the key to the Scheme and kept implementation costs down, enabling swift payments to be made (paras. 83-84).

The judge expressed her conclusion on the issue of justification in these terms (at para. 85):

Having agreed before the judge that “manifestly without reasonable foundation” was the appropriate test of whether indirect discrimination could be justified, the appellants now contend, in the light of the Supreme Court’s judgment in SC, that it was an over-simplification of a multi-faceted test.

The appellants argue that the judge’s observation that the margin of appreciation was “broad” was wrong in law for three reasons. First, the court erred in its starting point. They submit that it is the adverse discriminatory effects of the measure that determine the degree of deference to be granted to a decision-maker, not the measure itself. Given the fact that the judge approached the issue of justification on the basis that there was no discrimination to justify, her errors of law on Article 14 undermined her whole analysis of the issue. Secondly, the fact that the discrimination in issue related to a core status under Article 14, namely sex, ought to have narrowed the broad margin of appreciation applied by the judge. Thirdly, she failed to have regard to the fact that the justification advanced was ex post facto in nature. It is well established that an ex post facto justification requires particular scrutiny: Belfast City Council v Miss Behavin’ Ltd[2007] UKHL 19, [2007] 1 WLR 1420, In re Brewster[2017] UKSC 8, [2017] 1 WLR 519.

The appellants submit that, when the correct approach is applied, none of the five justifications put forward by the respondents stands up to scrutiny. With regard to the purpose of the scheme, the judge was wrong to hold that the appellants’ case was inconsistent with that purpose. A non-discriminatory scheme was entirely consistent with the scheme’s purpose of providing a subsidy reflecting the general profitability of the individual’s business. The failure to make a payment that properly reflected a woman’s trading profits was contrary to the purpose of the scheme. Secondly, the judge applied the wrong approach to policy delivery. Although it is correct that the scheme was introduced quickly, it continued to be considered in detail over the following weeks during which the discrimination in issue was squarely raised. The second Briefing Note recognised the need for an equality impact assessment, but failed to recognise the impact on women in general or women on maternity leave in particular. By the time the potential impact on maternity was recognised, the decision to use ATP over three defined years as the basis for calculation had been made. In the event, some alterations were made to the scheme after it was brought into force to ensure that it provided support to otherwise ineligible 2018-19 parents. The need to introduce the scheme quickly therefore provided no justification for the ongoing discrimination. If it was unfair for those people in the loan charge category to be required to rely on three years’ ATP, it was equally unfair to impose the requirement on those who had taken maternity leave during those years. Thirdly, the judge wrongly concluded that the approach to fraud provided sufficient justification. The entire design of the scheme was based on a recognition and acceptance of the risk of large-scale fraud associated with permitting all self-employed workers to self-certify, without independent verification, that their businesses had been adversely affected by the pandemic. The risk of fraud associated with the comparatively small cohort of women who had taken periods of absence for reasons relating to maternity was minuscule by comparison. Fourth, the judge erred in finding that the discrimination was justified because any solution would itself create a range of hard cases and was therefore justified to avoid perverse effects. In doing so, she failed to recognise the special status of women on maternity leave. Finally, the judge erred in her analysis of the issue of value for money. There was no contemporaneous evidence about the extra cost which would have been incurred had the scheme been extended in accordance with the appellants’ case.

In reply, the respondents contend, first, that it is not the function of the appeal court to redetermine proportionality issues. When considering a proportionality assessment conducted by a lower court, an appellate court should only intervene if the lower court has made a significant error of principle or there is a flaw in the judge’s reasoning: In re B (A Child) [2013] UKSC 13, [2013] 1WLR 1911; R (R) v Chief Constable of Greater Manchester Police[2018] UKSC 47, [2018] 1 WLR 4079; R (Z) v Hackney London Borough Council[2020] UKSC 40, [2020] 1 WLR 4327. In the Hackney case, the Supreme Court endorsed the approach advocated by Lewison LJ in the Court of Appeal ([2019] EWCA Civ 1099) at para. 66:

Secondly, they submit that the adjustment to the test made by the Supreme Court in SC has not materially altered the approach to be adopted in the present case. It remains the case that the bar to a challenge to an instrument of social and economic policy is a high one: R (RJM) v Secretary of State for Work and Pensions, [2008] UKHL 63, [2009] 1 AC 311 per Lord Neuberger at paras. 54-57. Determining the scope of instruments of welfare provision is primarily the responsibility of the executive and legislature. The Direction was secondary legislation – an economic and social measure worked up in a few days in response to a pandemic, with enormous social and economic impact and costing several billion pounds of public money. Accordingly, the judge was right to apply a broad margin.

Mr Milford submitted that it would be absurd to say that very weighty reasons were required to justify the indirect discrimination in this case. In support of this submission, he referred us to the decision of this Court in AM (Somalia) v Entry Clearance Officer [2009] EWCA Civ 634. In that case the claimant sought to challenge the provision of the Immigration Rules that required a person seeking entry clearance as a spouse to show that the couple would be able to maintain themselves without recourse to public funds on the basis that it was discriminatory, contrary to Article 14, against people with a disability because they were less likely to be able to earn. She argued, drawing on the Strasbourg case-law as it then stood, that such discrimination could only be justified on “weighty grounds”. That submission was rejected by both Maurice Kay LJ and Elias LJ, on the basis that a less stringent approach was required to justification where the discrimination in question is indirect. We need only quote from the judgment of Elias LJ, who said, at para. 61:

Furthermore, the respondents reject the argument that it was the adverse discriminatory effects of the measure that were determinative. On the contrary, it was the measure itself that had to be justified: see SG, per Lady Hale at para. 189. They add that the appellants’ contention that the judge’s analysis of the issue of justification was tainted by her finding that there was no discrimination is wrong because she expressly approached the issue on the alternative basis that the claimants had established the discrimination for which they contended.

Thirdly, the respondents challenge the appellants’ assertion that the justification for the discrimination set out in the evidence is ex post facto rationalisation. They describe it instead as an “elaboration” or “fleshing out” of considerations which were manifestly in the first respondent’s mind when devising the scheme, as is evident from the Ministerial Briefing Notes cited above. They submit that, in any event, it is well-established that retrospective assessments, particularly those within the decision-maker’s sphere of expertise, are fully admissible on issues of proportionality: see Miss Behavin, supra.

The respondents reject the argument that the five justifications identified by them and accepted by the judge do not stand up to scrutiny. Although it is not the appeal court’s function to reassess proportionality, a review of the judge’s analysis should lead to the conclusion that her conclusions were correct on all five factors. As to the purpose of the scheme, the ATP calculation reflected its purpose of providing support for self-employed individuals to mitigate the actual adverse effect of the pandemic on their business. Time away from work, whether for maternity or other reasons, translates into lower income. The scheme was designed to minimise the impact of variations in income so that the actual trading position was better reflected. In any event, the appellants had not come up with a tenable alternative method of calculation. The options proposed by the appellants would have caused other problems for claimants. With regard to policy delivery, the key aim was to put a scheme in place quickly. Adding complexity would have significantly added to the time taken to devise and deliver the scheme. The respondents detect in the appellants’ argument an implicit challenge to the Second Direction which is not the subject of this application for judicial review. With regard to fraud, it is clear that it was a major concern in the design and implementation of the SEISS. Given the limited time available to set up the scheme, there were limits to what fraud protections could be put in place. Any suggestion that the fact that there was an inevitable fraud risk made it impermissible to seek to limit the vulnerability of the scheme to additional risk is untenable. Tackling the additional fraud risk would have introduced further complexity and delay. The alternatives suggested by the appellants would have led to perverse effects in that individuals who were covered by the SEISS as drafted would have fallen outside it. Disregarding tax years would have led to calculations of grants being based on a smaller sample of filed returns, which would have led to greater distortions. Finally, with regard to value for money, any alternative of the sort proposed by the appellants would have caused substantially increased costs, as was demonstrated by the evidence put before the judge.

At the date of the judgment, the case law on justification in cases of indirect discrimination was clear and, to paraphrase Lord Wilson, not open to doubt. Whipple J was therefore entirely right to adopt a straightforward approach, reflected in the relatively brief terms in which she dealt with the law in her judgment. Since then the topic has been re-assessed in 65 paragraphs of the judgment in SC. As a result, a modified approach must be adopted to “reflect the nuanced nature of the judgment which is required”.

We are not persuaded that in the circumstances of the present case the reformulation of the law in SC is material to how Whipple J approached the issue of justification. Her assessment was appropriately nuanced, gave appropriate respect to the assessment of democratically accountable institutions, and recognised not only the need for caution before intervening in areas of social and economic policy but also that cogent justification was required for a measure having a differential impact on women. In our view she applied a level of scrutiny appropriate to the circumstances of the case. There is no merit in the appellants’ contention that the judge’s assessment of justification was tainted by her decision about discrimination. She was careful to say that it was conducted on the basis that she was wrong on the first issue and that the circumstances did amount to indirect discrimination. Furthermore, the appellants have not satisfied us that she failed to take into account any factor relevant to the assessment. Their attack on her assessment was directed at her attribution of weight to the factors identified. We were not persuaded that there was any error of principle or any flaw in her reasoning. In those circumstances, it would be contrary to established authority for this Court to re-determine this issue which is fundamentally a question of proportionality.

However, in the light of SC we think it appropriate to say that even if it fell to this Court to revisit the judge’s assessment we would come to the same conclusion as she. We can state our reasons fairly briefly.

We start by reminding ourselves of Lord Reed’s warning at para. 162 of SC:

The respondents have advanced reasons by way of justification in witness statements that are clear, detailed and in our judgment cogent. As regards the appellants’ contention that that evidence constitutes ex post facto rationalisation rather than contemporary reasoning, we note the observation of Lord Kerr in In re Brewster, at para. 52:

The measure complained of is, as noted at para. 60 above, the use of the ATP model as the means of assessing the profits lost by self-employed traders as a result of the pandemic. The choice of that measure plainly had legitimate aims – effectiveness and speed of policy delivery, ease of verifying the figures supplied in order to reduce the risk of fraud (and other errors), the need to avoid perverse effects, and cost. In truth, it was the obvious means for the purpose, for all the reasons given by the respondents in their evidence: accounts submitted to HMRC over past years were the only verified information about self-employed earnings that was readily available.

The appellants’ real complaint, however, is not the use of the ATP measure as such but the failure to introduce a special modification to cater for the case of recent mothers whose profits in the measurement period will have been unrepresentative of their likely hypothetical no-Covid earnings in 2020. As to that, we have deliberately set out the history of the development of the scheme in some detail to illustrate the extreme if not unique circumstances in which it was devised. It is evident that it was not until the middle of March 2020 that the government appreciated the seriousness of the threat to lives and to the economy posed by Covid-19. By that time, the die was cast and the need for urgent action obvious. Time was of the essence. The scheme therefore had to be devised and constructed with the utmost speed. To obtain additional information in the case of recent mothers would have involved a complex exercise which in turn would have significantly delayed the implementation of the scheme. Furthermore, any further information would have had to be based on some form of self-reporting by the individual taxpayer of impact of maternity on their earnings in the measurement period: whatever method were adopted, its reliability and accuracy of which would be hard to verify. Perhaps, given time, it would have been possible to devise a modification that allowed recent mothers to submit verifiable information that could form the basis for a special assessment of their likely lost earnings. But in our judgment it would plainly not have been possible to do so in the limited time available without compromising the essential requirements of speed, simplicity and verifiability.

The judge also concluded that it would not have been possible to fashion a scheme that resolved the difficulties identified by the claimants without creating another range of hard cases and anomalies. On any view, devising a scheme that addressed the difficulties of those who took maternity leave without creating other perverse effects would plainly have taken time and resources. We do not accept that that difficulty can be ignored simply by reference to the special status of women on maternity leave.

In short, given the cardinal features required of the scheme – and above all speed and simplicity – the first respondent was in our view justified in introducing the scheme in a form which did not contain special provision for the position of recent mothers.

We believe this approach is supported by consideration of the Ministerial Briefing Notes summarised at paras. 29-37 above. Although the authors of the second Note evidently failed to spot the potential difficulty for recent mothers, that is understandable given the great speed at which the hard-pressed civil servants were working on this scheme alongside other projects to meet the economic crisis caused by the pandemic. But the issue was raised and considered in the fourth to sixth Notes. The discussion in those Notes speaks for itself. It is clear that the pros and cons of various ways of addressing the issue of the possible unrepresentativeness of the earnings of recent mothers were evaluated, albeit in an abbreviated fashion having regard to the urgency of the task. The sixth Note did make a recommendation, which was implemented in the Second Direction, of a partial mitigation which enabled persons with periods of all forms of parental leave in 2018-19 to use 2017-18 as a reference year. It does not follow from the fact that the respondents were able to introduce this limited and specific mitigation in the Second Direction that it would have been practicable to have introduced some much broader modification addressing the problem of unrepresentative earnings for recent mothers, let alone to have done so in the First Direction.

In short, we do not consider that the impact of the use of the ATP measure, in unmodified form, on recent mothers was disproportionate to the benefit of the impugned measure. Given the exigencies under which the Treasury and HMRC were operating, as clearly described in the material put before the court, we conclude that the judge was right to conclude that if there was indirect discrimination (as we have found) it was justified.

We would therefore dismiss the appeal.

I agree.