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
The Hon Mrs Justice Whipple
CO24442020

		Royal Courts of Justice Strand, London, WC2A 2LL
		24 November 2021

Lord Justice Underhill and Lord Justice Baker:

1. 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.

BACKGROUND

2. 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.
3. On 31 January 2020, Public Health England announced the first cases of people in England to test positive for Covid-19.
4. 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.
5. In her statement in these proceedings, Ms Kantor, looking back to that point, stated:

"It was clear in light of social distancing measures announced by the Prime Minister that large parts of the economy would be very substantially affected, and that large emergency measures would be required to deal with those effects. There was a risk that many sound businesses would permanently cease trading as a reaction to short-term pressure on cash flow from fixed costs and disappearing revenues. Preventing failures of otherwise sound businesses and large-scale job losses could only be achieved by quickly moving to alternative sources of cash flow."

It was in that context that officials in the Treasury and Her Majesty's Revenue and Customs ("HMRC"), respectively the Defendant and the Interested Party in these proceedings ("the respondents") started work on policy initiatives to support employed and self-employed workers, which led to the establishment of two schemes – the Coronavirus Job Retention Scheme ("CJRS") for employed workers (commonly referred to as the "furlough scheme") and the SEISS for the self-employed – alongside other initiatives (including business loan arrangements and deferrals of income tax and VAT) devised in conjunction with other government departments. The development of the schemes was described through a series of Ministerial Briefing Notes from 22 March to 12 June 2020, considered in more detail below.

6. 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.
7. 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:

"Her Majesty's Revenue and Customs are to have such functions as the Treasury may direct in relation to coronavirus or coronavirus disease."

Under section 71, such directions may be signed by a single Treasury Commissioner or another Treasury minister.

8. 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:

"Providing such unprecedented support for self-employed people has been difficult to do in practice. And the self-employed are a diverse population, with some earning significant profits. So I've taken steps to make this scheme deliverable, and fair:

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.

95% of people who are majority self-employed will benefit from the scheme. HMRC are working on this urgently and expect people to be able to access the scheme no later than the beginning of June. If you're eligible, HMRC will contact you directly, ask you to fill out a simple online form, then pay the grant straight into your bank account. And to make sure no one who needs it misses out on support, we have decided to allow anyone who missed the filing deadline in January, four weeks from today to submit their tax return …"

9. 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.
10. 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.
11. 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.

THE LEGISLATION

THE FIRST DIRECTION – RATIONALE AND DETAILED PROVISIONS

12. 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.
13. 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.

"The tax liability of self-employed individuals … is based on their annual profits. [They] may choose their own accounting period which determines the tax year in which their profits must be reported. Tax returns must be filed by 31 January following the end of the tax year …. The tax return must contain information on each individual's income, expenses, and profit …. That means that, for the vast majority of self-employed individuals, the government obtains only an annual snapshot of that individual's financial position. The data so obtained is a minimum of 10 months old by the time self-assessment returns must be made and can be considerably older …."

In addition, the income of self-employed persons may vary substantially from year to year for a variety of reasons about which HMRC does not receive detailed information. Such reasons might include the incurring of high expenses in one year (through choice or necessity), ill-health, caring responsibilities, or pregnancy and maternity. Figures submitted in self-assessment tax returns are, however, the only reliable information available to HMRC for the purpose of assessing self-employed individuals' trading profits. Other sources of information were not a feasible basis for calculating support because

"the use of information which has not been verified by HMRC would have carried with it significant risks for the public purse and opened up the scheme to fraudulent activity."

14. 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.

"The overriding consideration when designing and implementing SEISS was to help as many eligible people as possible in as short a time as possible, without creating an unacceptable risk of fraud or error."

Although the key structure of the scheme was designed in about one week, very considerable work was involved in designing the details and putting the design into practice. Therefore,

"in order to achieve the design and implementation of SEISS in the limited time available, it was imperative that the scheme be simple and effective, and work on the basis of existing concepts and 'bright line' distinctions, and existing verifiable data, wherever practicable. Any complexity would substantially increase the practical work required to introduce SEISS, and also increase the complexity of its delivery in practice. Such complexity would risk jeopardising one of the main aims of SEISS, namely to provide support to those eligible as soon as practicable."

15. 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.
16. 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.
17. 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".
18. 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.
19. 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."

"[The] move to homeworking involved a substantial disruption to normal working practices …. The team working on SEISS overcame the significant hurdle of collaborating in developing SEISS not from a high-tech IT Delivery Centre, but from dining room tables, bedrooms and living rooms."

20. Mr Hacon continued (at paragraph 19):

"In the light of all of the above, certain fundamental principles underlay the design of SEISS. SEISS had to be put in place at great speed, for the reasons that I have set out. It had to be a reliable system, which would get money in people's bank accounts as quickly as possible. It had to be accurate, calculating grants on the basis of certain and verifiable underlying data, which was already in HMRC's possession. It had to have measures built in that would protect (to the extent possible) from errors and fraudulent claims. It had to be automated so far as possible, given the need for speed, reliability, and accuracy, and the constraints on HMRC's resource: ultimately computers are far more efficient at analysing and cross-referencing information than human beings. It also had to be simple for those seeking to make a claim: both in order to encourage people actually to use the system, and in order to ensure that they could do so with a minimum of 'live' administrative support from HMRC staff. Simplicity is not simply a matter of our administrative convenience: it was and is vitally important that all those eligible to claim SEISS, from whatever different backgrounds they come, and whatever their linguistic abilities and familiarity with administrative processes, are able in practice to do so."

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

"The purpose of SEISS is to provide for payments to be made to persons carrying on a trade the business of which has been adversely affected by the health, social and economic emergency in the United Kingdom resulting from coronavirus and coronavirus disease."

Paragraph 4, headed "Qualifying person", reads (so far as relevant to this appeal):

"4.1 A person is a qualifying person if the following conditions are met.

4.2 The person must –

(a) carry on a trade the business of which has been adversely affected by reason of circumstances arising as a result of coronavirus and coronavirus disease,

(b) have delivered a tax return for a relevant year on or before 23 April 2020,

(c) have carried on a trade in the tax years 2018-19 and 2019-20,

(d) intend to carry on a trade in the tax year 2020-21,

(e) [condition relating to non-UK residents or persons not domiciled in the UK – not relevant to this appeal]

(f) be an individual, and

(g) meet the profits condition.

4.3 In paragraph 4.2, 'relevant tax year' means all or any of the tax years 2016-17, 2017-18 and 2018-19, as the case may be, for which a person's trading profit and relevant income must be determined for the purposes of SEISS."

22. 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.
23. The "profits condition" referred to in paragraph 4.2(g) was defined in paragraph 5 of the First Direction:

"5.1 The profits condition is met if

(a) where the person is not subject to the loan charge, the person meets condition A, B or C, or

(b) where the person is subject to the loan charge, the person meets condition D or E.

5.2 Condition A is met if

(a) the person's trading profits of the tax year 2018-19 were £50,000 or less but were more than nil, and

(b) those profits are equal to or more than the person's relevant income in that tax year.

5.3 Condition B is met if

(a) the person carried on a trade in the tax years 2016-17, 2017-18 and 2018-19,

(b) the average amount of the person's trading profits of those tax years was £50,000 or less but was more than nil, and

(c) the sum of those profits is equal to or more than the sum of the person's relevant income for those tax years.

5.4 Condition C is met if

(a) the person carried on a trade in the tax years 2017-18 and 2018-19 but did not carry on a trade in the tax year 2016-17,

(b) the average amount of the person's trading profits of the tax years 2017-18 and 2018-19 was £50,000 or less but was more than nil, and

(c) the sum of those profits is equal to or more than the sum of the person's relevant income for those tax years.

5.5 Condition D is met if

(a) the person carried on a trade in the tax years 2016-17 and 2017-18,

(b) the average amount of the person's trading profits of the tax years was £50,000 or less but was more than nil, and

(c) the sum of those profits is equal to or more than the sum of the person's relevant income for those tax years.

5.6 Condition E is met if

(a) the person did not carry on a trade in the tax year 2016-17,

(b) the person's trading profits of the tax year 2017-18 were £50,000 or less but were more than nil, and

(c) those profits are equal to or more than the person's relevant income in that tax year."

"Relevant income" was defined in paragraphs 7 and 8 of the First Direction to mean total income minus its trading income component. The effect was to exclude from the scheme any person who made less than 50% of their income from self-employment.

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

"These requirements had a number of purposes. These included focusing on those who most needed government assistance, minimising the risk of fraud, and ensuring the administrative practicability of the scheme. By targeting the scheme at only those who met the profit condition (no more than £50,000 and at least equal to the individual's non-trading income) and were continuing to trade (i.e. who had not stopped their business prior to the pandemic), the government ensured that the scheme was targeted at actively trading businesses who needed the support the most. By basing the SEISS scheme upon filed self-assessment tax returns, and permitting only those who had filed returns before 23 April 2020 to claim, HMRC could put in place a process whereby individual's applications could be automatically assessed against information from their filed self-assessment tax returns, and the amount of the grant to be automatically calculated by HMRC. This provided a safe, secure and speedy way to assess the validity of individuals' trading profits from self-employment, including how they compared to other income and therefore the amount of grant they were able to claim."

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

"6.1   The amount of the SEISS payment is the lower of

                   (a)     £7,500, and

                   (b)     3 x ( x 80%).

6.2    In paragraph 6.1, TP is

(a)     except where the person is subject to the loan charge, determined by the first to apply of the following principles:

         (i)      if the person carried on a trade in the tax years 2016-17, 2017-18 and 2018-19, the average trading profits of those tax years;

                   (ii)     if the person did not carry on a trade in the tax year 2016-17, the average trading profits of the tax years 2017-18 and 2018-19, and

         (c)     if the person did not carry on a trade in the tax year 2017-18, the trading profits of the tax year 2018-19, or

(b)     where the person is subject to the loan charge, the average trading profits of the tax years 2016-17 and 2017-18 or, if the person did not carry on a trade in the tax year 2016-17, the trading profits of the tax year 2017-18."

26. 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:

"Self-employed individuals' profits can vary markedly, for all sorts of reasons. At its simplest, a self-employed individual may have a particularly good, or particularly bad, year. He or she may also incur (by choice, or necessity) particularly high expenses in one year; for example if purchasing essential tools. He or she may choose to take a sabbatical; or be affected by unfortunate ill-health; or take on responsibility for caring for a relative (e.g. if elderly, or disabled), or an older child. Female self-employed individuals may also (as the claimants emphasise) take a period of absence from self-employed work because of pregnancy or maternity. All of these things are likely to affect that individual's self-employed profits, and so (if incurred in one of the relevant years) to affect the quantum of the SEISS grant. The use of ATP [average trading profits] as the basis for calculation was and is intended to 'even out' the individual's profits over the years in which they have traded, and so to provide the most accurate reflection possible of the general profits incurred by that business."

THE SECOND DIRECTION

27. 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.
28. 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 MINISTERIAL BRIEFING NOTES

29. 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.
30. 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.
31. 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.
32. 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":

"12. The current scheme design encounters issues for those who have only partial year earnings, for example because they started their trade mid-year, or if they paused trading care for a new baby. In particular, the issue around pausing trade to care for a new baby has been flagged by UK Music and Labour MPs including Jess Phillips. One way to address for some individuals this [sic] could be to annualise profits to calculate what an individual could have expected to earn if they had been trading for a full year. Eligibility and the level of grant could then be determined using the annualised figure.

13. This would be of most benefit to individuals who started trading mid-year and would allow for their SEISS payment to more accurately reflect their (assumed) normal income. It would provide unfair benefit to those whose partial-year profits are already an accurate reflection of what their total year profits would be if, for example, their trade is seasonal.

14. However, annualising would not capture the majority of individuals who took time off to care for a new baby, as through self-assessment data we cannot identify breaks in trade of any sort which last for less than one year.

15. This option could also create hard cases. For example, someone who started trading in January 2019 with £15,000 profits from three months of trading would become ineligible as pro-rating would bring them above the upper income cap.

16. Annualising amounts would also carry a risk that we were seen to be departing from the principle of using the amounts returned, and annualising part year figures for existing self-employed could increase pressure to provide additional support for 19/20 starters.

17. Initial estimates suggest that the magnitude of cost for allowing pro-rata of partial year earnings would be c.£1bn for the three months. However, this is an initial estimate and so highly uncertain at this stage.

18. An alternative, where trading commenced mid-year in 2016-17 or 2017-18 would be to ignore all amounts reported in the return for the commencement year in the calculation of eligibility and amount of grant. However, this option could not apply to 2018/9 commencements and, like the other options would also not address temporary pauses in trading activity covering less than a full tax year. This would be seen as unfair by those who started trading in 2018/19.

DECISION: Do you agree not to pro-rate profits for individuals with partial year self-employment?

DECISION: For trades which commenced in 2016/17 or 2017/18, would you like to ignore the partial year results in calculating eligibility and level of grant?"

33. 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:

"17. You have a statutory obligation through the Public Sector Equalities Duty to consider the equality impacts of the SEISS, and consider mitigating action if a protected characteristic is discriminated against, whether directly or indirectly.

18. Since we initially advised you on the equality impacts of the SEISS, several stakeholders have argued that the SEISS disproportionately discriminates against pregnancy and maternity, a protected characteristic, as (i) a person currently on maternity leave may be unable to show that their trade has been adversely affected by coronavirus, and (ii) a trader who was on maternity leave for the whole of 18-19, and did not file a tax return in that year, would not be eligible for the scheme, and (iii) a trader who took maternity leave during any of the relevant years may receive a grant which does not reflect their usual level of trading profit, due to the averaging of trading profits.

19. [Words redacted.] HMRC still consider a trader to be trading even if she is on maternity leave, and therefore remains fully eligible for the SEISS. Therefore situation (i) does cause no detriment on the basis of pregnancy or maternity. We recommend making this position clear in guidance.

20. [Words redacted.] situations (ii) and (iii). Our view is that:

(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.

21. [Words redacted.] We are exploring, [words redacted], whether we could invite affected people to make themselves known to us through the disputes process, and apply some special rules to disregard the maternity period when calculating their claim. This would be outside the fully automated process, so we need to explore the numbers and impacts involved before determining the feasibility of this. If you decide to take mitigating action, this will need to be achieved through a second Treasury Direction.

DECISION: Please note that you have considered the further equality impacts of the SEISS."

34. 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:

"We are recommending that you amend SEISS eligibility conditions for reservists and all forms of parental care. [Words redacted.] We are recommending you do not amend the grant calculation for these groups as the three year averaging process mitigates for volatile incomes, and is a strength of the scheme.

The scheme will go live without reservists or maternity leave solutions in place and therefore those who are currently outside the scope of the SEISS will be ineligible until any solution is in place. Any bespoke treatment for these cohorts will require a second direction to HMRC.

[Words redacted.]

We believe that it is possible for HMRC to administer a scheme to cover reservists and those who took any form of parental leave, recognising that the capability of HMRC to assure eligibility on the basis of these criteria is limited and there [sic] this is therefore a risk of abuse. This bespoke solution would also require a more resource intensive approach, requiring additional IT resource. Payments would be available to those in scope by the middle of June; later than the main scheme. To note, this additional resource need could impact HMRC's capacity to deliver other priorities in June, such as a SEISS extension."

35. Under the heading "Policy options", the Note continued:

"For reservists and for those with periods of maternity leave, the challenge is twofold;

(i) Individuals are ineligible because they didn't have trading profits or did not submit a Self-Assessment return in 2018-19, or have failed the profits condition ….

(ii) The level of the grant awarded may be reduced if an individual has paused trading at any point between 2016-17 and 2018-19 …."

The Note proceeded to outline options for both cohorts. Under the heading "Adapt the scheme for reservists only", it identified eligibility options and (in paragraph 11) "three options for addressing the averaging challenge" namely "do not change the grant calculation", "use only whole trading years" and "ignore the tax year in which majority of call up occurred". The authors recommended the first option. Under the heading "Adapt SEISS for reservists and for maternity leave", they advised:

"13. We estimate at least 65,000 self-employed mothers have had a child between 2016-17 and 2018-19. The first challenge is identifying maternity leave as no definition of maternity leave exists for the self-employed. Therefore, this solution needs to target those who have had children in the reference years, and who as a result of the provision of care have seen a pause to their trading activity or a reduction in their profits.

Eligibility options

14. Individuals not currently eligible for SEISS may be unknown to HMRC. In order to proactively identify and contact this population, HMRC can utilise datasets on recipients of the maternity allowance (held by DWP) and new child benefit claimants (held by HMRC). This information can be cross-referenced against self-assessment returns to provide an eligibility solution in line with that recommended for reservists above. However, these maternity datasets do not cover the entire population as individuals who took maternity leave would not necessarily have drawn on these benefits, either voluntarily or through ineligibility due to high earnings.

Averaging options

15. As above, once eligibility has been determined, HMRC could adjust formula for calculating the grant in the same ways listed in paragraph 11, with the same issues. Our recommendation for this group is also the same, that you take no action and rely on the averaging calculation inherent within the SEISS to account for the volatility profits.

Consequences of action

16. [Words redacted.] It may also exacerbate presentational issues regarding other groups, protected or otherwise, who have experienced periods of paused trading, such as the long-term sick or the carers of elderly relatives [words redacted], if you wanted to take action for those with periods of maternity leave, the solution would also have to apply to at least all other forms of parental leave. Introducing those groups into the SEISS will result in a consideration of at least 70,000 cases, with an estimated cost of at least £20m a month."

36. The Note continued:

"We recommend you enable reservists and those with periods of all forms of parental leave in 2018-19 to use 2017-18 as the reference year for submitting a self-assessment return if required and that the grant calculation for both groups is not amended."

37. 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:

"We continue to recommend that as a minimum you act [words redacted] by amending the SEISS eligibility conditions to ensure a grant is available to those whose pregnancy/childbirth affected their 2018-19 trading profits trading profits."

They identified two options: (1) a "minimal solution narrowly targeted at those who took time out of their trade due to pregnancy or childbirth" and (2) "a solution more widely targeted at those who took time out of their trade to care for a child within the first 12 months of the child's birth or adoption." The decision between these two options was described as "finely balanced". The authors proceeded to recommend the adoption of the mechanism applicable to those persons subject to the loan charge that was ultimately adopted in the Second Direction. They also included an equalities impacts assessment on the options provided.

THE CLAIM

38. 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 JUDGMENT

39. 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.
40. 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:

"In order to establish that different treatment amounts to a violation of article 14, it is necessary to establish four elements. First, the circumstances must fall within the ambit of a Convention right. Secondly, the difference in treatment must have been on the ground of one of the characteristics listed in article 14 or "other status". Thirdly, the claimant and the person who has been treated differently must be in analogous situations. Fourthly, objective justification for the different treatment will be lacking. It is not always easy to keep the third and the fourth elements entirely separate, and it is not uncommon to see judgments concentrate upon the question of justification, rather than upon whether the people in question are in analogous situations."

41. 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:

"In the end I understood the argument to be that the index group were disadvantaged by comparison with everyone else who was eligible for a payment under the Scheme, because these women's trading profits had been lower during the relevant tax years for reasons connected with maternity and childbirth, a unique state which did not affect others who were eligible for payments under the Scheme."

42. The judge then considered the claimants' alternative argument based on Thlimmenos, that:

"their unique situation required a unique solution, and that the calculation for these women should have been different, so as to remove the disadvantage which affected them if they were treated in the same way as everyone else who was claiming payment under the Scheme."

She decided it was unnecessary to resolve an issue which had arisen in the course of argument about the nature of Thlimmenos discrimination, concluding that Article 14 plainly encompassed both kinds of discrimination and citing the analysis of Chamberlain J in R (Salvato) v Secretary of State for Work and Pensions [2021] EWHC 102 (Admin).

43. 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".
44. 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.
45. 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 ISSUES

46. The appellants put forward three grounds of appeal against the judge's dismissal of the application for judicial review.
47. 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.
48. 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.
49. 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.

DISCRIMINATION

50. 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.

INDIRECT DISCRIMINATION

(1) Background Law

51. Article 14 of the Convention provides:

"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, political or other opinion, national or social origin, association with a national minority, property, birth or other status."

52. 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.
53. 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:

"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 he observes, that statement is itself derived from the judgment of the Grand Chamber in DH v Czech Republic (2008) 47 EHRR 3 (see para. 175). He continues:

"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."

At para. 50 he observes that "[t]he concept of indirect discrimination has only gradually come to be recognised by the European court". He goes on to illustrate the development of the concept 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. He concludes, at para. 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)."

Before Whipple J the parties treated the decision in DH as the authoritative statement of the correct approach, but it was common ground before us that Lord Reed's summary is not materially different.

54. 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:

"[I]ndirect discrimination shall be taken to occur where an apparently neutral provision, criterion or practice would put persons [with a protected characteristic] at a particular disadvantage compared with other persons, unless that provision, criterion or practice is objectively justified by a legitimate aim and the means of achieving that aim are appropriate and necessary."

The key concepts as regards prima facie indirect discrimination – that is, a situation which requires justification – are thus that there should be a "provision, criterion or practice" ("PCP") which puts persons with a protected characteristic at "a particular disadvantage" compared with others". (We should note for completeness that in Enderby v Frenchay Health Authority C-127/92, [1994] ICR 112, the CJEU identified what appears on its face to be a distinct kind of indirect discrimination; but that is an unnecessary complication for present purposes.)

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

"A provision, criterion or practice is discriminatory in relation to a relevant protected characteristic of B's if -

(a) A applies, or would apply, it to persons with whom B does not share the characteristic,

(b) it puts, or would put, persons with whom B shares the characteristic at a particular disadvantage when compared with persons with whom B does not share it,

(c) it puts, or would put, B at that disadvantage, and

(d)     A cannot show it to be a proportionate means of achieving a legitimate aim."

We should also note section 23 (1) of the Act, which provides that "[o]n a comparison of cases for the purposes of section … 19 there must be no material difference between the circumstances relating to each case".

56. 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.
57. 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:

"24. The first salient feature is that, in none of the various definitions of indirect discrimination, is there any express requirement for an explanation of the reasons why a particular PCP puts one group at a disadvantage when compared with others. Thus there was no requirement in the 1975 Act that the claimant had to show why the proportion of women who could comply with the requirement was smaller than the proportion of men. It was enough that it was. There is no requirement in the Equality Act 2010 that the claimant show why the PCP puts one group sharing a particular protected characteristic at a particular disadvantage when compared with others. It is enough that it does. Sometimes, perhaps usually, the reason will be obvious: women are on average shorter than men, so a tall minimum height requirement will disadvantage women whereas a short maximum will disadvantage men. But sometimes it will not be obvious: there is no generally accepted explanation for why women have on average achieved lower grades as chess players than men, but a requirement to hold a high chess grade will put them at a disadvantage.

25. A second salient feature is the contrast between the definitions of direct and indirect discrimination. Direct discrimination expressly requires a causal link between the less favourable treatment and the protected characteristic. Indirect discrimination does not. Instead it requires a causal link between the PCP and the particular disadvantage suffered by the group and the individual. The reason for this is that the prohibition of direct discrimination aims to achieve equality of treatment. Indirect discrimination assumes equality of treatment - the PCP is applied indiscriminately to all - but aims to achieve a level playing field, where people sharing a particular protected characteristic are not subjected to requirements which many of them cannot meet but which cannot be shown to be justified. The prohibition of indirect discrimination thus aims to achieve equality of results in the absence of such justification. It is dealing with hidden barriers which are not easy to anticipate or to spot.

26. A third salient feature is that the reasons why one group may find it harder to comply with the PCP than others are many and various ... They could be genetic, such as strength or height. They could be social, such as the expectation that women will bear the greater responsibility for caring for the home and family than will men. They could be traditional employment practices, such as the division between 'women's jobs' and 'men's jobs' or the practice of starting at the bottom of an incremental pay scale. They could be another PCP, working in combination with the one at issue, as in Homer v Chief Constable of West Yorkshire [2012] UKSC 15, [2012] ICR 704, where the requirement of a law degree operated in combination with normal retirement age to produce the disadvantage suffered by Mr Homer and others in his age group. These various examples show that the reason for the disadvantage need not be unlawful in itself or be under the control of the employer or provider (although sometimes it will be). They also show that both the PCP and the reason for the disadvantage are 'but for' causes of the disadvantage: removing one or the other would solve the problem."

58. 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.

(2) How the Appellants Put their Case

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

"[T]hey argue that the Scheme has a disproportionately prejudicial effect on women who have not worked in the preceding three tax years for maternity reasons; the prejudice is that those women receive smaller payments than they would otherwise be entitled to, and thus the Scheme indirectly discriminates against such women. They cite [DH], paragraph 175 to define indirect discrimination:

'[the application of] a general policy or measure that has disproportionately prejudicial effects on a particular group [which] may be considered discriminatory notwithstanding that it is not specifically aimed at that group.'"

It is necessary to identify the elements in that case with some care.

60. 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".
61. 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.
62. 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.
63. 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.
64. 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.
65. 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.
66. 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.

(3) The Judge's Reasoning

67. 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.
68. 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:

"62. I consider the Divisional Court's comments in Adiatu to have relevance to this case, especially the passage at paragraph 149. In this case, too, the disadvantage is not caused by the Scheme itself; rather it is a disadvantage which flows from an absence of or reduction in a person's income in the past; for the group of women represented by the Claimants, it is the consequence of a self-employed woman being unable to earn while on maternity leave. I accept the point made in the Defendant's evidence and by submission on behalf of the Defendant, that there may be many reasons why a self-employed person is unable to work. This is not to draw comparisons between the different reasons; it is simply to recognise the fact that for self-employed people, absence from work is likely to translate into lower earnings.

63. Barry is also of assistance in providing a factual analogy with this case: just as lower final earnings as a part-time employee could be used to calculate the termination payment, so here the lower ATP of a recent mother could be used to calculate the Scheme payments, without in either case being discriminatory.

64. I am not therefore persuaded that there is any indirect discrimination, approaching the matter on a conventional analysis. The measure imposes no hidden barriers to eligibility. So far as quantum of payment is concerned there is no hidden barrier either: quantum is based on past (average) trading profits, which are a matter of past fact. The same rule applies to all and it is no harder for a woman who has been on maternity leave to qualify or calculate their payment, than someone who has not. The fact that some claimants will receive lower payments than others reflects the fact of lower earnings in past years; I agree with the Defendant that the reasons for lower earnings in past years, in the context of this Scheme with its stated purpose, are not relevant."

69. 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.
70. 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.
71. 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.
72. 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.
73. 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.
74. 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".
75. 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.

(4) Discussion and Conclusion

76. 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:

"There is no requirement in the Equality Act 2010 that the claimant show why the PCP puts one group sharing a particular protected characteristic at a particular disadvantage when compared with others. It is enough that it does."

The reasoning is all the more surprising where it is clear that the reason for the "lower earnings in past years" – that is, lower than would be expected in the hypothetical no-Covid current year – is gender-related.

77. 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.
78. 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.
79. 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:

"The first stage raises the question whether the principles of discrimination are applicable at all. Article 119 [of the Treaty of Rome] is concerned with discrimination: equal pay for men and women for equal work or work of equal value. Discrimination means treating like cases differently or, as is claimed in the present case, treating unlike cases the same … Mrs. Barry worked full-time for eleven years and key-time for 17.5 hours per week for two years, but she is treated the same as a key-time worker who had worked for 17.5 hours per week for thirteen years. Cases are 'unlike', and ought to be treated differently, when the difference between them is material for the purpose in hand. The purpose in hand here is the calculation of severance payment. Whether the difference relied upon by Mrs. Barry as a material difference should be so regarded depends upon whether the bank could properly adopt a severance payment scheme with the objects of this scheme. If the bank could do so, the difference relied upon is immaterial. Her hours of work history is irrelevant. In that event, discrimination not having been established, Mrs. Barry's claim would fail at the first hurdle."

80. 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:

"The first question which arises is whether there is a difference in treatment at all between full-time and part-time workers for the purposes of the Act and the Treaty. In that regard it is not sufficient merely to ask whether one gets more or less money than the other. It is necessary to consider whether, taking account of the purpose of the payment, there is a difference in treatment. The purpose of the payment here is to provide support for lost income during the period immediately following redundancy. As the Industrial Tribunal put it, it is 'to cushion employees against unemployment and job loss'. It is not to remunerate for past service (when it would be necessary to have regard to actual service at different periods) even if the payment takes into account years of service to reflect loyalty to the employer. …

The weekly amount lost during the redundancy period is thus the amount of salary being paid at the end of the employment; it is not therefore a relevant difference in treatment to base all employees' severance payments on their final salary. In principle the position is the same here as under the statutory scheme for redundancy pay and the statutory scheme for payment in lieu of notice when payment is related to years of service but is based on final salary. The payment reflects the actual salary the employee would have received during the notice period and not some notional amount calculated on types and hours of service over the whole period of employment. In the present case there is no relevant difference in treatment because all employees, men and women, full-time and part-time, of all ages, receive a payment based on final salary."

As we understand it, that reasoning in substance amounted to an acceptance of Mr Elias's submissions.

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

"At the end of the hearing I was inclined to the view that in its actual operation the scheme has a discriminatory effect. On further consideration I have been persuaded that properly analysed the appellant's claim is that she has been treated less favourably than she would have been treated if a different scheme, with different objectives, giving greater credit to a feature to her advantage, had been adopted. I would now on the special facts of this case accept the argument of Mr Elias QC that the rules of discrimination are not engaged. In my view therefore the scheme does not offend against the principle of equal pay for equal work and is therefore not unlawful."

82. 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:

"I quite accept that in examining the method of payment adopted by the employer to see whether it has a discriminatory effect, the court is concerned with substance rather than form. The question is how the scheme actually operates. But there is nothing artificial about the criteria adopted in the present scheme. The emphasis is upon final salary because the main purpose of a redundancy payment is to cushion the employer against the hardship of a sudden cessation of his salary, whatever it may have been."

Thus Lord Hoffmann too regards the purpose of the scheme as crucial.

83. 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):

"… the primary objective of the scheme was to cushion employees against unemployment. In that context the calculation by reference to the level of final salary seems to me entirely appropriate and it is applied to men and women without distinction."

Read in isolation, that might seem rather opaque, but in context it is clear that his essential reasoning was the same as Lord Slynn's.

84. 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.
85. 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.
86. A not dissimilar situation to that in Barry was considered by the CJEU in Stadt Lengerich v Helmig 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:

"Part-time workers complained that pay at overtime rates was only available once an employee had worked for longer than the ordinary working week of a full-time worker. They contended they were treated differently from full-time workers because, unlike full-time staff, they were not paid overtime for each hour in excess of their contracted hours. The court held there was no difference in treatment because part-time and full-time workers received the same amount of money for the same number of hours worked. In other words, the purpose of pay in the form of salary was to remunerate for hours actually worked. The court decided that, in the light of this conclusion at stage (1), it was not necessary to proceed to consider objective justification."

Lord Slynn regarded Stadt Lengerich as supporting his analysis – see p. 1469 B-D; and, consistently with what we have said above, Lord Nicholls also accepted it but distinguished it on the basis that on the facts it was, unlike Barry, "a plain case" (p. 1474C).

87. Once the ratio of Barry is understood, it can be seen that, contrary to para. 63 of Whipple J's judgment, it provides 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.
88. 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.
89. 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

"given that female and BAME employees are disproportionately represented in the lowest earning groups, they are disproportionately likely to be unable to have the resources to manage with such a low income, and are accordingly disadvantaged by the rate of SSP (either losing income or going to work when they ought not to do so)"

(see para. 140 of the judgment of the Court).

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

"In our judgment, this argument is misconceived. The rate of SSP is not a PCP which places certain categories of employees at a particular disadvantage. The classic PCP which does so is a requirement that must be satisfied in order for persons to qualify for a particular opportunity or benefit, such as a height requirement in order to be permitted to join a police force, or the requirement to be a full-time worker in order to qualify for a pension. These examples place women at a particular disadvantage because women are less likely than men to be tall, and are more likely to be part-time workers (because of child-care responsibilities). The rate of SSP is not a barrier or gateway in this sense. It is a sum that is paid, in exactly the same way, to everyone who receives SSP, regardless of their protected characteristics. It does not place women or BAME employees at a particular disadvantage: everyone is treated the same."

It goes on to consider Barry and Shillcock, together with the final sentences of para. 25 of the judgment of Lady Hale in Essop (quoted above), referring to "hidden barriers". It concludes at para. 149:

"In relation to the rate of SSP, there is no 'hidden barrier'. Essop is not authority for the proposition that something places those with protected characteristics at a particular disadvantage because their circumstances, unconnected with the PCP, are less favourable than those of others. In our judgment, the Defendant is right to submit the Claimants do not rely upon any disadvantage that is caused by the rate of SSP itself. Rather, they rely upon an alleged disadvantage, the absence of other financial resources, which is not caused or related to the rate of SSP in any way. This does not turn the rate of SSP into a PCP which places women or BAME employees at a particular disadvantage. In our view the EU law challenge to the rate of SSP is wholly unsustainable."

91. 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".
92. 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.
93. 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.
94. Before we leave Adiatu, we should note two other points about Whipple J's reasoning in relation to it:

(1) In para. 62 she says that the disadvantage to new mothers "is not caused by the Scheme itself" but by their reduced level of earnings while on maternity leave. But that mis-identifies the disadvantage being alleged. That disadvantage is the fact that in the case of recent mothers their earnings in the measurement period will be disproportionately unrepresentative of their hypothetical no-Covid earnings, resulting in lower SEISS payments for recent mothers as a group than if they had been representative: that disadvantage is caused by "the scheme itself" – or, more particularly, by the use of ATP as the relevant measure.

(2) As already noted, her statement in para. 64 of her judgment that there are in the present case no "hidden barriers" echoes para. 149 of the judgment in Adiatu. Its use there derives from para. 25 of Lady Hale's judgment in Essop. But Lady Hale was plainly not saying in that passage that the existence of a "hidden barrier" is the essence, or a touchstone, of indirect discrimination: she was making a much more general point in relation to the distinction between direct and indirect discrimination. In fact, many or most PCPs are far from "hidden"; and even the reasons why they have a disproportionate impact on the group in question are in many, though not all, cases obvious (as indeed Lady Hale observes in para. 24 of her judgment).

95. 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.

"THLIMMENOS DISCRIMINATION"

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

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

Discrimination of this kind is often referred to as "Thlimmenos discrimination".

97. 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:

"65. Applying his alternative approach based on Thlimmenos, Mr Bunting suggests that women who have recently been on maternity leave, who are thus in a unique situation, must be afforded different treatment to reflect the fact that they have lost out on self-employed earnings in the relevant tax years; in other words, the calculation for them must be adjusted to take account of the period of lost earnings related to childbirth. This is an argument that such women have been treated similarly when they should have been treated differently. Many of the same points already outlined could be made in this context. But in my judgment, there are (at least) two problems specific to this argument.

66. First, accepting for present purposes that pregnancy and maternity are unique situations for which no comparator exists and in relation to which special protections are warranted, they are circumstances which for the Second Claimant and the group she represents exist in the past. The effect of the Claimants' argument would be to demand redress by means of the Scheme in relation to a unique situation in the past. None of the six factors noted by the Court of Appeal in Ali [the reference is to Ali v Capita Customer Management Ltd [2019] EWCA Civ 900, [2020] ICR 87, para. 68] are relevant at this distance of time. I was shown no authority to support the proposition that uniqueness, or difference, in the past is a basis on which to require different treatment in the present, such that failure to accord that different treatment in the present amounts to unlawful discrimination.

67. Secondly, and as stated above, the disadvantage identified by the Claimants follows from the fact - for that is what it is - that they earned less in past years. I fail to see how that state of affairs requires them to be compensated through the benefits system now, by receiving a higher level of benefit. This is the Adiatu point: the disadvantage is not caused by the measure but rather it exists independently of the measure. I do not accept that the Scheme's failure to take account of and rectify historic disadvantage amounts to discrimination."

98. 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.

JUSTIFICATION

(1) THE LAW

99. 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:

"Whether that difference in treatment has an objective and reasonable justification will depend on whether the rule which results in the difference in treatment has a legitimate aim and is a proportionate means of realising that aim."

In her judgment at para. 189, Lady Hale observed:

"It is important to understand that what is needed to justify indirect discrimination is different from what is needed to justify direct discrimination. In direct discrimination, it is necessary to justify treating women differently from men. In indirect discrimination, by definition, women and men are treated in the same way. The measure in question is neutral on its face. It is not (necessarily) targeted at women or intended to treat them less favourably than men. Men also suffer from it. But women are disproportionally affected, either because there are many more of them affected by it than men, or because they will find it harder to comply with it. It is therefore the measure itself which has to be justified, rather than the fact that women are disproportionately affected by it. The classic example is a maximum age bar on recruitment to particular posts; it applies to all candidates, women and men; but it disadvantages women because they are more likely to have taken a career break to have or care for children than are men. The question therefore is whether the age bar can be independently justified …."

100. 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:

"(1) whether the objective of the measure is sufficiently important to justify the limitation of a protected right, (2) whether the measure is rationally connected to the objective, (3) whether a less intrusive measure could have been used without unacceptably compromising the achievement of the objective, and (4) 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 …. In essence, the question at step four is whether the impact of the rights infringement is disproportionate to the likely benefit of the impugned measure."

101. 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.
102. 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.
103. 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:

"in relation to the Government's need to justify what would otherwise be a discriminatory effect of a rule governing entitlement to welfare benefits, the sole question is whether it is manifestly without reasonable foundation. Let there be no future doubt about it."

104. 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:

"whether the approach to proportionality under article 14 set out by this court in Humphreys … and followed in several later cases, to the effect that the court will respect the policy choice of the executive or the legislature in relation to general measures of economic or social strategy unless it is "manifestly without reasonable foundation", accurately reflects the approach of the European Court of Human Rights ("the European court") and should continue to be followed."

Lord Reed indicated at the outset that:

"The answer, put shortly, is that the case law of the European court supports a nuanced approach which is not fully captured by a "manifestly without reasonable foundation" standard of review, and which in some circumstances calls for much stricter scrutiny."

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

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

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

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

106. 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.
107. At para. 142, Lord Reed summarised his analysis of the Strasbourg jurisprudence in these terms:

"In summary, the European court has generally adopted a nuanced approach, which can be understood as applying certain general principles, but which enables account to be taken of a range of factors which may be relevant in particular circumstances, so that a balanced overall assessment can be reached. As I have explained, there is not a mechanical rule that the judgment of the domestic authorities will be respected unless it is "manifestly without reasonable foundation". The general principle that the national authorities enjoy a wide margin of appreciation in the field of welfare benefits and pensions forms an important element of the court's approach, but its application to particular facts can be greatly affected by other principles which may also be relevant, and of course by the facts of the particular case. Indeed, this approach is not confined to cases concerned with article 14, but can be seen in other contexts where the state generally enjoys the wide margin of appreciation signified by the "manifestly without reasonable foundation" formula, but where other factors may indicate a narrower margin of appreciation, and the court accordingly balances the relevant factors…. 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 v Croatia [2021] IRLR 511 and Yocheva and Ganeva v Bulgaria [2021] ECHR 385 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."

108. 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:

"In domestic law, as at the Strasbourg level, one would expect closer scrutiny where the case concerns discrimination on a ground such as sex or race, rather than a difference in treatment on less sensitive grounds, especially if it is simply a by-product of a legitimate policy. Distinctions drawn on "suspect" grounds are inherently appropriate for close judicial scrutiny, notwithstanding the respect due to the judgment of the executive or the legislature" (para. 145).

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

"It seems clear from Stec, however, that the normally strict test for justification of sex discrimination in the enjoyment of the Convention rights gives way to the 'manifestly without reasonable foundation' test in the context of state benefits."

In Lord Reed's opinion, however, this reasoning did not reflect the Strasbourg jurisprudence entirely correctly:

"It seems to me that the "manifestly without reasonable foundation" formulation, as used in the Strasbourg judgments, does not express a test, in the sense of a requirement whose satisfaction or non-satisfaction will in itself necessarily be determinative of the outcome. The phrase indicates the width of the margin of appreciation, and hence the intensity of review, which is in principle appropriate in the field of welfare benefits, other things being equal. As I have explained, however, a number of other factors may also be relevant in the circumstances of particular cases, some of which may call for a stricter standard of review" (para. 151).

He added:

"Differential treatment on a suspect ground, if it is capable of justification at all, generally (but not always) requires to be justified by "very weighty reasons". That is so even in the context of measures of social and economic policy which would usually benefit from the "manifestly without reasonable foundation" approach. The "manifestly without reasonable foundation" approach does not, therefore, replace or supersede the requirement for "very weighty reasons" where "suspect" grounds are in issue. Instead, the degree of deference usually appropriate in relation to social or economic policy choices may have to be taken into account in assessing whether "very weighty reasons" have been shown" (para. 152).

110. 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:

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

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

159. It is therefore important to avoid a mechanical approach to these matters, based simply on the categorisation of the ground of the difference in treatment. A more flexible approach will give appropriate respect to the assessment of democratically accountable institutions, but will also take appropriate account of such other factors as may be relevant….

160. It may also be helpful to observe that the phrase "manifestly without reasonable foundation", as used by the European court, is merely a way of describing a wide margin of appreciation….

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

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

"In practice, challenges to legislation on the ground of discrimination have become increasingly common in the United Kingdom. They are usually brought by campaigning organisations which lobbied unsuccessfully against the measure when it was being considered in Parliament, and then act as solicitors for persons affected by the legislation, or otherwise support legal challenges brought in their names, as a means of continuing their campaign. The favoured ground of challenge is usually article 14, because it is so easy to establish differential treatment of some category of persons, especially if the concept of indirect discrimination is given a wide scope. Since the principle of proportionality confers on the courts a very broad discretionary power, such cases present a risk of undue interference by the courts in the sphere of political choices. That risk can only be avoided if the courts apply the principle in a manner which respects the boundaries between legality and the political process."

112. 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):

"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."

(2) THE JUDGE'S DECISION

113. 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.

(1) 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).

(2) 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).

(3) 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).

(4) 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).

(5) 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).

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

"Whether the various justifications are taken separately or in combination, the Defendant's decisions were reasonable ones, especially when judged in context. The Scheme was a macro-economic policy involving substantial public expenditure to mitigate the effects of a global pandemic. The Government had a wide margin of appreciation. The design of the Scheme, specifically in the way the payments were calculated by reference to ATP, was not manifestly without reasonable foundation."

(3) THE SUBMISSIONS TO THIS COURT

115. 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.
116. 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.
117. 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.
118. 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:

"It is not enough simply to demonstrate an error or flaw in reasoning. It must be such as to undermine the cogency of the conclusion. Accordingly, if there is no such error or flaw, the appeal court should not make its own assessment of proportionality."

119. 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.
120. 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:

"Like Maurice Kay LJ, I would accept that any rule which differentiated benefits or rights specifically by reason of disability would require weighty reasons; prima facie it is hard to see how it could be justified and there would need to be very good reason to explain why it was being adopted. But it would be absurd to apply the same requirement to cases of indirect discrimination, particularly in circumstances where there is equality of treatment and the contention is that there should not be. The range of characteristics linked to one of the identified forms of status is potentially very wide indeed, and it would severely inhibit a state's power to legislate if it had to provide weighty reasons for adopting policies which adversely impacted on groups not by reason of status alone, but for reasons connected to it. Furthermore, the need for weighty reasons is in any event less prominent where questions of social policy are in issue."

121. 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.
122. 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.
123. 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.

(3) DISCUSSION AND CONCLUSION

124. 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".
125. 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.
126. 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.
127. We start by reminding ourselves of Lord Reed's warning at para. 162 of SC:

"Since the principle of proportionality confers on the courts a very broad discretionary power, such cases present a risk of undue interference by the courts in the sphere of political choices. That risk can only be avoided if the courts apply the principle in a manner which respects the boundaries between legality and the political process."

The provisions implemented in the First Decision plainly came within the field of social and economic policy. The Decision was made by a democratically accountable government minister under powers vested in him by Parliament. We would add that although the level problem of unrepresentative earnings in the measurement period is likely to be disproportionately experienced by recent mothers it is not peculiar to them: this is a case of indirect, not direct discrimination. Accordingly a less intense level of review may be appropriate. Mr Bunting did not seek to argue that the observations of Elias LJ in AM (Somalia) had been overtaken by any subsequent case-law, and in our view the distinction which they recognise between direct and indirect discrimination is legitimate and consistent with Lord Reed's judgment in SC. The binary distinction between whether or not "weighty reasons" are needed may no longer be relevant in the light of the more nuanced approach which he endorses. But the underlying point that a less intense level of review may be appropriate where the discrimination complained of is indirect (even involving a "suspect ground") rather than direct seems to us to be in fact a good example of that nuanced approach. We also note that in Humphreys Lady Hale appears to have taken it for granted that the fact that the discrimination in that case was indirect operated to reduce the appropriate level of scrutiny (see para. 19 of her judgment): although of course SC marks a departure from some of the observations in Humphreys we do not believe that it affects that point.

128. 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:

"Obviously, if reasons are proffered in defence of a decision which were not present to the mind of the decision-maker at the time that it was made, this will call for greater scrutiny than would be appropriate if they could be shown to have influenced the decision-maker when the particular scheme was devised. Even retrospective judgments, however, if made within the sphere of expertise of the decision-maker, are worthy of respect, provided that they are made bona fides."

In this case we see no reason to challenge the bona fides of the reasons advanced in the evidence. Further, we accept the respondents' submission that those reasons are elaborations of arguments which were plainly under consideration when the scheme was being developed. And we believe that they have received an appropriate level of scrutiny both at first instance and in this Court.

129. 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.
130. 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.
131. 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.
132. 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.
133. 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.
134. 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.

DISPOSAL

135. We would therefore dismiss the appeal.

Nicola Davies LJ:

136. I agree.