ASDA Stores Ltd v Brierley & Ors (Equal Pay Act) [2017] UKEAT 0011_17_3108

EAT

ASDA Stores Ltd v Brierley & Ors (Equal Pay Act)

[2017] UKEAT 0011_17_3108 · 2017-08-31

1. Although the point is not acte clair , the better view is that article 157 of the Treaty on the Functioning of the European Union is directly effective in a claim founded on equal pay for work of equal value. The Appeal Tribunal declines to make a reference to the Court of Justice of the European Union seeking a determination of this (or any) point.

2. Where there is a “single source” of pay and conditions for both claimant and comparator, a comparison between them is permitted independently of whether unequal treatment arises from legislation or collective agreements and whether or not the employment is in the same establishment or service.

3. Where no comparator works at the establishment where the claimant is employed, comparison is permitted applying the North hypothetical test. The better view is that the North hypothetical test remains good law and has survived the replacement of section 1(6) of the Equal Pay Act 1970 by section 79 of the Equality Act 2010 .

4. The Employment Judge did not err in law in deciding that the law is as stated above. He did not misapply the law. Nor were any of his findings of fact perverse. He reached conclusions that were open to him on the facts. There is no basis for interfering with his decision that the Claimants can compare themselves with their chosen comparators.

THE HONOURABLE MR JUSTICE KERR

Introduction: the appeal

1. This appeal arises from about 7,000 equal pay claims brought by supermarket employees of the Appellant (Asda), nearly all women. The appeal is against a Reserved Decision of Employment Judge Tom Ryan sitting in Manchester, dated 13 October 2016. The claims are founded on equal value. The issue in the appeal is whether the Claimants can compare their work with that of distribution workers based at depots, who are nearly all men and are paid more. The Judge held that they can, a decision Asda says is wrong and should be reversed.

2. Most of the claims were brought in 2014 and 2015 and relate back to a period before the Equality Act 2010 (“EqA”) came into force on 1 October 2010; in respect of which the old law in the Equal Pay Act 1970 (“EqPA”) applies. To decide whether comparability with the distribution workers is permissible, the Judge ruled in the Claimants’ favour on issues formulated variously in the grounds of appeal. Very broadly, they require resolution of the following issues:

(1) whether article 157 of the Treaty on the Functioning of the European Union (“TFEU”) is directly effective in equal value claims such as these;

(2) whether the Judge correctly applied EU law on the existence of a “single source” of pay and conditions of employment for both claimants and comparators; and

(3) whether the Judge properly applied domestic law (old and new) on comparison of terms between claimants and comparators, or the classes of employee to which they belong.

3. The Judge heard evidence and submissions over six days in June 2016. He made detailed findings of fact about the history and practice of Asda in setting terms for, respectively, the Claimants (retail staff) and the comparators (distribution staff). He then set out the relevant law at length before considering the parties’ submissions and ruling on them, in each case in the Claimants’ favour, and giving his reasons for preferring the Claimants’ arguments.

4. In Asda’s stores, the Judge found, the predominantly female retail workforce is not heavily unionised and terms are not set through collective bargaining, though there are differential rates based on cost of living differentials in three separate geographical areas. Pay and terms are imposed from the top down. In the separately located distribution depots, the predominantly male distribution workforce at each site now works on terms collectively bargained through the GMB, following national recognition agreements in 2012 and 2014.

5. The appeal concerns the proper limits of comparison for the purposes of equal pay claims. In ascertaining where those limits lie, the parties placed emphasis on different EU law principles of general importance to their respective cases. Some of the issues raised could well be appropriately resolved by a reference to the Court of Justice of the European Union. In my judgment, at least one of the grounds of appeal raises an EU law issue that is not acte clair in favour of either party. Neither party strongly submitted that I should make such a reference; nor did either positively object to one.

6. I have a power but not a duty to make a reference. After reflection, I have decided not to do so. To borrow from Stanley Burnton LJ ( R (Risk Management Partners Ltd) v Brent LBC

[2008] EWHC 692 (Admin)

, paragraph 47), my function is probably no more than juge rapporteur for the Court of Appeal. That court will be better placed to decide whether a reference is appropriate. The parties are prepared to tolerate the additional delay that will cause, and the uncertainty about the future jurisdiction of the Court of Justice.

7. The debate before me was wide-ranging, sophisticated, subtle and presented with the utmost skill both orally and in writing. Credit and my gratitude are due to juniors and solicitors as well as the silks who presented the oral argument. To clarify the issues in my own mind, I have found it necessary to summarise the parties’ submissions, sometimes in my own words. I have not attempted to cover all the nuances of every argument advanced. I hope I will do at least rough justice to the parties’ arguments. There is a transcript of the oral hearing, and extensive written skeletons, should I fail to do so.

8. For Asda, Mr Jeans QC emphasised the centrality of the individual employment relationship and, in particular, of the business unit to which the employee is assigned. Comparison is not of like with like, he said, where claimant and comparator work in different industries under different employment regimes. He described as “a basic tenet of EU law” the proposition that “[a]n employment relationship is essentially characterised by the link existing between the employee and the part of the business to which he is assigned to carry out his duties” ( Rockfon A/S v Specialarbejderforbundet i Danmark C-449/93,

[1996] IRLR 168

, paragraph 31; USDAW v WW Realisation 1 Ltd

Case C-80/14

(heard with other cases),

[2015] IRLR 577

, paragraph 44).

9. For the Claimants, Mr Short QC emphasised the policy of EU law of eliminating sex discrimination in the workplace. Without the ability to compare groups of workers at different workplaces, the policy would be undermined by employers able to perpetuate “occupational segregation” and pay less for undervalued jobs predominantly done by women such as “caring, cashiering, catering, cleaning and clerical”: see paragraph 15 in the February 2006 report of the Women & Work Commission chaired by Baroness Prosser and entitled Shaping a Fairer Future ; and Cox J in Ministry of Defence v Armstrong

[2004] IRLR 672

at paragraph 34:

“… pay discrimination is frequently systemic in character, arising as a result of gender job segregation or from discrimination in pay structures and grading systems, rather than from the terms of individuals’ contracts of employment. …”

Is article 157 of the TFEU directly effective in an equal value claim?

10. The first ground of appeal is that the Judge was wrong to hold that article 157 of the TFEU (formerly, article 141 of the Treaty Establishing the European Community (“TEC”) and, before that, article 119 of the Treaty of Rome which did not explicitly refer to equal value) is directly effective and may be relied upon directly by the Claimants in the present equal value claims. Asda contends, and contended below, that article 157 is not directly effective where a claim is founded on work of equal value rather than like work.

11. Asda’s unsuccessful argument below was that, to borrow the language of the Court of Justice (“ECJ”) in Defrenne v SABENA

[1976] ICR 547

, paragraph 21 and Worringham v Lloyds Bank Ltd

[1981] ICR 558

, paragraph 23, in an equal value claim the discrimination cannot be “judicially identified” on the basis of “a purely legal analysis of the situation” and “solely with the aid of the criteria based on equal work and equal pay referred to by the article”. A complex evaluation of the jobs using “expert methodology” is required.

12. The Judge heard submissions from the parties which, to Judge from his decision, were similar to those made to me. He preferred the argument of the Claimants that Worringham v Lloyds Bank had determined the issue in their favour. It did not establish only that article 119 (now 157) is directly effective in an equal value claim where equal value has already been established by internal job evaluation. Direct effect was not excluded by the need for expert evaluation in order to determine whether the jobs were of equal value, he decided.

13. In challenging the proposition that article 157 is directly effective here, Mr Jeans relied on the following main points:

(1) The ECJ in Defrenne v SABENA (paragraphs 18-22) had distinguished direct and overt discrimination (as in the facts in that case) from “indirect and disguised” discrimination only discernible “by reference to more explicit implementing measures”, community or national.

(2) The latter type of discrimination included equal value claims; while the former, identifiable “solely by reference to the criteria laid down in article 119”, included those founded on legislation or collective agreements and “which may be detected on the basis of a purely legal analysis …” or where unequal pay is paid to workers in the “same establishment or service”.

(3) This approach reflected the concept of direct effect articulated in earlier ECJ jurisprudence requiring the measure in question to be unconditional and sufficiently precise, starting with Van Gend en Loos

[1963] CMLR 105

(4) In Worringham v Lloyds Bank Ltd

[1981] ICR 558

(paragraph 23), the ECJ explained that the cases where discrimination may be “judicially identified” without community or national implementing measures were cases where “the court is in a position to establish all the facts enabling it to decide whether a woman receives less pay than a man engaged in the same work or work of equal value”.

(5) For the court to “establish all the facts”, it must not be required to engage in a complex fact-finding exercise or in evaluation of expert opinion evidence; it must be established, without such an exercise being necessary, that the work is of equal value; for example, by internal job evaluation, as in Worringham , or by a concession from the employer.

(6) That analysis was supported by obiter dicta from three distinguished Judges in domestic courts, all opining against direct effect in an equal value claim: Lord Oliver in Pickstone v Freemans

[1989] AC 66

HL, at 124B-F; Balcombe LJ in Leverton v Clwyd County Council

[1989] AC 706

, at 723H-724D; and Lord Eassie in City of Edinburgh v Wilkinson

[2012] IRLR 202

, paragraphs 41-42.

(7) Those views were based on the need for the operation of a detailed procedural regime set out in the tribunal’s rules, in order to determine whether the jobs being compared are of equal value or not. The same is true in this case, where equal value is not established; and whether or not the fact-finding exercise is done in the tribunal, using its procedures, or in the High Court, using the Civil Procedure Rules.

(8) Other cases relied on by the Claimants ( Commission v UK

[1982] ICR 578

ECJ; Murphy v Bord Telecom

[1988] ICR 445

ECJ; Royal Copenhagen ( Specialarbejderfor-bundet i Danmark v Dansk Industri )

Case C-400/93

[1996] IRLR 51

ECJ; and Walton Centre for Neurosurgery NHS Trust v Bewley

[2008] ICR 1047

EAT) are not authority that article 157 is directly effective here. In none of those cases did the issue arise for decision. Indeed, those cases are against direct effect.

(9) The infraction proceedings in Commission v UK would have been pointless if national measures to deal with equal value claims were rendered unnecessary by the direct effect of article 119, even though (see EC v Belgium

Case C-102/79

[1981] CMLR 282

, paragraph 12) direct effect is not a defence in such proceedings.

(10) In Commission v UK , A-G van Themaat had indicated (at 582H-584C) that direct effect of article 119 does not cover situations where “discrimination on grounds of sex cannot be directly determined by the court”; that this includes cases where job evaluation is required to determine whether jobs are of equal value; and that member states had discretion as to what implementing measures to enact, as long as the result is achieved.

(11) The evaluation of jobs to determine whether jobs are of equal value is not an ordinary fact-finding exercise for the tribunal. It is a system of classifying components of the different jobs, involving a methodology i.e. the system of methods and principles used in the particular discipline (per HHJ McMullen QC in Middlesbrough BC v Surtees (No 2)

[2008] ICR 349

, paragraph 8).

14. In support of the Judge’s decision on direct effect, the following were Mr Short’s principal counter- arguments:

(1) The evolution of European case law shows that the concept of direct effect in the context of workplace sex discrimination has broadened. While Defrenne was a case of direct and overt discrimination, the subsequent cases showed that direct effect is not confined to such cases. The later cases showed that the paradigm of overt direct discrimination embodied in Defrenne - unequal pay for equal work carried out in the same establishment or service - was only one example in a broader range of situations in which direct effect occurs.

(2) Thus, Macarthys Ltd v Smith

[1980] ICR 672

had established that comparison with a predecessor colleague was permissible; Jenkins v Kingsgate (Clothing Productions) Ltd

[1981] ICR 592

, that the direct effect of article 119 could cover cases of indirect discrimination (paying less to predominantly female part timers); and Worringham , at paragraph 23, that cases where discrimination may be “judicially identified” include cases of equal value among those where (emphasis added):

“the court is in a position to establish all the facts enabling it to decide whether a woman receives less pay than a man engaged in the same work or work of equal value .”

(3) A case of the type mentioned in Defrenne at paragraph 21, where the judicial analysis is purely legal because the facts are clear and uncontroverted, is but an example of a case where direct effect occurs; it is not the only such situation. The language used by the ECJ at paragraph 10 in Macarthys Ltd demonstrates that the ECJ did not in Defrenne set limits to the scope of direct effect in article 119 cases.

(4) The clarification provided by the ECJ in Worringham at paragraph 23, stating that direct effect can extend to equal value cases, followed an invitation from A-G Warner in Worringham (see at 566C-E) and in Jenkins (see 601G-602E) to clarify the scope of the direct effect of article 119 as expounded in Defrenne , some dicta in which the A-G found puzzling. It is thus clear that equal value cases where no concession of equal value is made and no job evaluation has been carried out, are not as such within the class of cases where national implementing measures are necessary to prohibit sex discrimination.

(5) Further support for this view can be found in Murphy v Bord Telecom , where the ECJ, rejecting an argument that article 119 could not be directly relied on in a case where the applicant’s job was, as the employer admitted and asserted, of higher value than the comparator’s, referred to “a consistent line of decisions of the court, in the case of work of equal value, and not in the case of work of unequal value” (paragraph 9).

(6) The distinguished ECJ Judge who had sat on five of the major cases in the field, Judge Pierre Pescatore, writing in the European Law Review ( Doctrine of Direct Effect; an Infant Disease of Community Law [1982] 8 ELRev 155, 158), referred to the line of cases in which article 119 had been held directly effective, as cases belonging to an “inner circle” at the “core” rather than the “fringe” of the provision, where “evaluation of the factual elements did not go beyond the normal powers of investigation and application of any court”.

(7) The conclusion that Worringham decided the point at issue in the claimant’s favour, is supported by the analysis of Elias J (P) in Walton Centre for Neurology and Neurosurgery NHS Trust v Bewley

[2008] ICR 1047

. While the decision was to reject, as outside article 141, a comparison with a successor employee, the comparison therefore necessarily being hypothetical, Elias J’s exposition of the law (at paragraph 15ff) treated the ECJ decisions as including equal value claims within the scope of direct effect.

(8) The cases in which article 119 and its successors are not directly effective must be those in which the comparators are hypothetical not actual, since hypothetical comparators are excluded from the scope of article 119 (as explained by Elias J in Walton at paragraphs 24-29) and their use must therefore depend on national implementing legislation. This is because the ECJ treated hypothetical comparisons as implying “comparative studies of entire branches of industry” ( Macarthys Ltd , paragraph 15, cited in Walton , paragraph 24).

(9) The ECJ had rejected, in Commission v UK

[1982] ICR 578

, the submission of the UK that “the criterion of work of equal value is too abstract to be applied by the courts” (paragraph 12) and had emphasised (at paragraph 42) that in the context of a defence of objective justification unrelated to discrimination, the national court must find the facts.

(10) There was nothing special about a court or tribunal making factual determinations about whether jobs are of equal value, whether using the specialised employment tribunal rules providing for expert evaluation, or in the High Court where an equal value claim may also be tried ( Abdulla v Birmingham City Council

[2012] ICR 1419

SC). The task of the court was not conceptually different from that of determining, for example, a factual defence of objective justification in an indirect discrimination claim, in which article 157 would beyond doubt be directly effective.

15. I have carefully considered these rival arguments. The Judge (at paragraph 167 of his Decision) was unimpressed by the argument that measuring the value of jobs that are being compared was beyond the scope of ordinary fact finding. He pointed out that an employment tribunal does not have to admit expert evidence measuring the value of jobs, though it would normally do so. He preferred the arguments of the Claimants.

16. In my judgment, the arguments on both sides are strong enough to rule out the view that the point is acte clair . In the earlier European jurisprudence on direct effect, in particular the Van Gend en Loos case, the court asked the question whether the measure in question was unconditional and sufficiently precise. Article 157 confers the right to equal pay for equal work, including equal pay for work of equal value. The right, in an equal value claim as in a like work claim, is unconditional.

17. The question is whether it is sufficiently precise. There is judicial support for the view that it is, and also for the view that it is not. I accept that in the 1982 infraction proceedings, Commission v UK , A-G van Themaat (at 582H-584C) was against direct effect of article 119 where “discrimination on grounds of sex cannot be directly determined by the court” and that, for his part, he treated that as including cases where job evaluation is required. He based his view on the variety of methods available for comparing jobs, reasoning that member states should have discretion on which method or methods to adopt.

18. I accept also that the domestic dicta relied on by Asda are against direct effect in a case such as the present where equal value is undetermined on the facts. The history was that in Pickstone v Freemans , the Court of Appeal Judges’ judgments support the claimant’s position (see per Nicholls LJ at 81C-E, Purchas LJ at 95H-96C, and Sir Roualeyn Cumming-Bruce at 96H-97C). But while an appeal to the House of Lords was pending in Pickstone v Freemans , in Leverton v Clwyd County Council , a different Court of Appeal took a different view.

19. Balcombe LJ (at 724B-D), supported by Stocker LJ, agreeing at 729A, expressed the contrary view for five reasons, the first four of which have largely been overtaken by the “single source” jurisprudence further considered below. The fifth was that in an equal value claim “detailed procedures are usually requisite to enable the industrial tribunal to determine the question before it”.

20. Lord Oliver in the House of Lords in Pickstone , at 124B-F confessed to “some doubt” whether article 119 “is directly enforceable in the circumstances of the instant case”. The House decided the case by statutory construction of the then wording of section 1 of the EqPA , so that it was unnecessary (per Lord Keith at 112E) to decide whether the Court of Appeal had been correct in its view that article 119 was directly effective in the case before it, in which the applicant relied on equal value. Lord Oliver’s doubt was such that he would have been minded to refer the issue to the ECJ, had it been necessary to do so.

21. His obiter reasoning was that, while the limits of direct effect were not clear from the European case law, the cases indicated that article 119 may not be directly effective in an equal value claim, “at any rate where there is no machinery in the domestic law by which the criterion of what is work of equal value can be readily ascertained”. He was concerned that the procedural machinery for determining equal value (then in 1985 Regulations ) could not be operated in a manner that went outside the statutory instrument creating it.

22. More than 20 years later, Balcombe LJ’s doubts were shared, on an obiter basis and without expressing a concluded view, by Lord Eassie, in the Inner House of the Court of Session, in City of Edinburgh v Wilkinson

[2012] IRLR 202

, paragraphs 41-42, on the ground that where a claim is founded not on like work or work rated as equivalent it is “difficult …. to say that discrimination on grounds of sex in the field of equal pay can be detected on a purely legal analysis” (paragraph 42).

23. On the other side of the argument, set against those dicta is the reference in paragraph 23 in Worringham to “work of equal value” which comes from the ECJ itself, in the context of an actual discussion about the width of direct effect of article 119; which followed an invitation from another distinguished Advocate-General, A-G Warner, to clarify the extent of what had been decided in Defrenne and Maca rthys Ltd v Smith .

24. In my judgment, that pronouncement carries more weight than the dicta from domestic Judges relied on by Asda and the view of A-G Themaat, which did not receive the express endorsement of the ECJ in Commission v UK and has not been endorsed since by that court or its renamed successor, the Court of Justice of the European Communities (“CJEU”).

25. It is established by Commission v UK that the concept of equal value is not too abstract to permit of national implementing measures. The question then becomes whether measuring the value of one job against another would be left by the CJEU to those implementing measures, or whether the measuring of jobs against each other is part of the exercise of judicially identifying sex discrimination.

26. That seems to me to depend, to borrow the words of Judge Pescatore in his article in the European Law Review, on whether “evaluation of the factual elements” is “beyond the normal powers of investigation and application of any court”. I think the better view is that CJEU would say it is not. The Luxembourg court has historically been prepared to take an expansive view of direct effect when it is necessary to ensure the efficacy of a Treaty right.

27. I do not accept the validity of the supposed distinction between, on the one hand, making a finding of fact on an issue where expert evidence is deployed to assist the court; and, on the other hand, making a finding on whether particular jobs done by particular claimants and their comparators are or are not of equal value. The enquiry may permit of more than one methodology and may be quite complex, though not necessarily so in every case.

28. But that is often the case where factual issues are resolved with the assistance of expert evidence. It may be a simpler exercise for a court or tribunal than adjudicating on a defence of objective justification in an indirect discrimination case, where article 157 is, admittedly, directly effective. I accept Mr Short’s submission that there is nothing inherently special about job evaluation. It may be an art as well as a science, and it is not an exact science, but it is not alchemy. It is opinion evidence on the basis of which a tribunal makes a finding of fact.

29. The Claimants are right to say that Lord Templeman’s procedural concerns and those of Balcombe LJ, have been shown not to be well-founded. An equal value claim may be brought in the High Court ( Abdulla v Birmingham City Council , cited above) where the court may itself determine the issue of equal value as one of fact, with or without the assistance of expert evidence. In a case where the claimant’s job was obviously of equal (or superior) value to the comparator’s, expert evidence might be unnecessary.

30. Usually it would be needed but the High Court is well versed in the art of determining factual issues with assistance from experts. The employment tribunal is also suitably equipped for the task, with experienced lay members and procedural rules for receiving the evidence. The wording of the domestic legislation does not inhibit this but, in any case, the scope of direct effect should not vary from EU member state to member state, depending on the content of its domestic legislation.

31. On balance, I consider that a bsent national implementing measures, the CJEU now would be inclined to hold that direct effect can supply the omission where specific jobs are being compared with specific other jobs being done by actual not hypothetical workers, but not where the comparison is between a claimant and a hypothetical comparator.

32. As Elias J (P) noted in the Walton Centre for Neurology case (see paragraphs 24 and 25), the ECJ in Maca rthys Ltd v Smith identified the case of a hypothetical male comparator as one of covert or disguised discrimination which, as had been explained in Defrenne , “implies comparative studies of entire branches of industry and therefore requires, as a prerequisite, the elaboration by the Community and national legislative bodies of criteria of assessment” ( Maca rthys Ltd , at paragraph 15).

33. A-G Capotorti in Maca rthys Ltd (see at 677B-D) referred to what the ECJ in Defrenne had called “discrimination which can only be identified by reference to more explicit implementing provisions of a community or national character”; and described discrimination of that type in the following terms:

“… what led the court to make that distinction and to hold that article 119 had no direct effect in regard to the second type of discrimination was, above all, the existence of cases in which it is impossible to compare the work of men and women, as occurs, for example, when discrimination takes place not simply within the same establishment but in relation to whole sectors of industry which are characterized by exclusively or predominantly female employment, not to speak of the disguised discrimination which arises by virtue of restrictions on access by women to certain posts or levels of employment.”

34. And he drew a contrast in the next sentence between that type of discrimination and the other type that was amenable to direct effect:

“Direct comparability between the functions performed by employees of both sexes for the same undertaking is not, on the other hand, prevented by the fact that the services are rendered at different times; and detailed legislation is surely not necessary to enable a court to establish whether work is “equal” within the meaning of the first paragraph of article 119.”

35. As Elias J commented, while it is not apparent why the court took the view that hypothetical comparisons would necessarily involve comparisons across branches of industry, the rejection of hypothetical comparisons is “clear and unambiguous”. As explained by A-G Capotorti in the passages just quoted, I think it is such hypothetical comparisons that are regarded as beyond the normal powers of judicial investigation and application, and thus beyond the scope of direct effect, rather than concrete comparisons between actual jobs done by real people.

36. For those reasons I respectfully agree, on balance, with the Employment Judge. The better view is that article 157 of the TFEU is directly effective in this case, and the Claimants can rely on it. The analysis is not perfect, because it could be said that since hypothetical comparisons are outside article 157 altogether, the question of direct effect in such a case does not arise. I would willingly have referred the point to the CJEU if the parties had enthusiastically invited me to do so.

37. As they do not and the decision falls to me, I base it on the judgment of the ECJ in Worringham ; my preference for the approach of A-G Capororti to that of A-G van Themaat; and the proposition that determining whether jobs are of equal value on the facts is well within the ordinary capability of an English court or employment tribunal. I see no good reason for treating equal value as incapable of judicial identification so as to allow the applicability of article 157 to be determined by a legal analysis. The first ground of appeal therefore fails.

The “single source” grounds of appeal

38. I come next to the “single source” issues in the appeal. The second ground of appeal is that the presence of a single source of pay and conditions of work as between claimant and comparator is not sufficient to permit comparability. The third ground is that the Judge misdirected himself in various ways when considering the single source issue. The fourth ground is that his finding of fact that the pay and conditions of the retail staff Claimants and the distribution staff comparators emanated from a single source, was perverse.

39. In support of those grounds, Asda’s main submissions were as follows:

(1) When the concept of a “single source” of pay and conditions was introduced by the ECJ in 2002 (see Lawrence v Regent Office Care Ltd

[2003] ICR 1092

), it did not alter the restrictions on permissible comparability established in Defrenne ; rather, it added the need for a single source as a further restriction, over and above those derived from Defrenne .

(2) The law now is, therefore, that article 157 applies where there is sex discrimination arising from legislation, collective agreement or employment in the same establishment or service; and there is, in addition, not as an alternative, a single source of pay and conditions of employment.

(3) Domestic case law to the contrary effect, supporting the proposition that a single source is a freestanding gateway to comparability, is wrong and should not be followed, for it is based on a misreading of Lawrence and of Allonby v Accrington & Rossdale College

[2004] ICR 1328

, in both of which comparability was not permitted. The Judge below was wrong to reject Asda’s submission to that effect.

(4) Furthermore, the additional requirement to point to a single source of pay and conditions as between claimant and comparator comprises two separate tests, each of which must be satisfied: there must be a single body, not necessarily a legal person, responsible for the inequality in pay; and that body must be one that could restore equality of treatment.

(5) It is insufficient to show that claimant and comparator work for the same employer; or that a body exercises budgetary control over a different body which actually sets the pay in question. It is also insufficient to show that a body which, legally speaking, could restore equality, is responsible for the inequality in pay unless the body actually exercises its power to determine pay and conditions ( DEFRA v Robertson

[2005] ICR 750

, per Mummery LJ at paragraphs 13 and 41).

(6) The present case is not distinguishable from DEFRA in which, while the Crown was the employer of both claimants and comparators, it did not set the pay of both in practice. Similarly here, neither Asda’s executive board nor that of its parent company, Wal-Mart, exercised a power to set pay and conditions for both groups of employees; their pay and conditions were determined by separate bodies and mechanisms.

(7) Accordingly, the Judge below was bound to find that this was a case falling on the same side of the line as DEFRA , and also Armstrong v Newcastle-upon-Tyne NHS Hospital Trust

[2006] IRLR 124

. The case was factually different from North Cumbria NHS Trust v Potter

[2009] IRLR 176

, because in that case, although the claimants and comparators worked at different NHS establishments, the EAT upheld the finding of the Tribunal that there had been a real exercise of the centralised power to set pay at the different establishments.

(8) The reasoning in two non-binding Scottish decisions, that of the EAT sitting in Scotland under the presidency of Langstaff J (P) in Fox Cross Claimants v Glasgow City Council

[2013] ICR 954

and, on appeal in the same case, of the Court of Session Inner House in Glasgow City Council v UNISON Claimants

[2014] IRLR 532

, was unsound in that a single source for pay and conditions was found where a local authority owned and had a power to control two limited liability partnerships, which it did not in practice exercise.

(9) The Judge had been wrong to treat as exceptional a situation in which a single employer did not constitute a single source and to place an “evidential burden” on the employer to displace the proposition that there is a single source where there is a common employer. There was no basis for that approach to the factual appraisal.

(10) The Judge had been wrong to treat the oversight exercised by Asda’s executive board, and that of Wal-Mart, as different in quality from the oversight exercised by the Treasury over the pay of civil servants in the DEFRA case. He should have found that here, as in the DEFRA case, the power of the common employer to set pay and conditions was legal only and unexercised in practice. Furthermore, the Judge had found that there were two single sources, Asda and Wal-Mart, which is a contradiction in terms.

(11) On the facts, the only conclusion open to the Judge was that there was no single source of the pay and conditions of, respectively, the retail staff Claimants and the distribution staff comparators and the contrary conclusion the Judge reached was perverse.

40. The main points made in answer by Mr Short, for the Claimants, were as follows:

(1) Where there is a single source of pay and conditions, comparability is without more permissible. Lawrence and Allonby did not restrict the scope of comparability; that would undermine the policy of article 157 and is inconsistent with subsequent domestic law authority up to and including the Supreme Court’s decision in Dumfries and Galloway Council v North

[2013] ICR 993

(2) There, Lady Hale JSC’s analysis at paragraphs 37-40 led to the conclusion that “for the principle of equal pay to have direct effect, the difference in treatment must be attributable to a single source which is capable of putting it right”. On the facts in the Dumfries and Galloway case, the Supreme Court found a single source even though the claimants and their comparators worked at different workplaces and even though their pay and that of their comparators were fixed under different national agreements.

(3) The Judge was right to distinguish the present case from the factual position in Lawrence , Allonby , DEFRA and Armstrong . He correctly applied the test identified from the European jurisprudence by Mummery LJ in DEFRA , at paragraph 13:

“… the focus of this rather imprecise approach [of the ECJ in Lawrence] is on the location of the body responsible for making decisions on levels of pay in the relevant employment or establishment rather than on the identification of the relevant legal source of that decision-making power.”

(4) In DEFRA in the EAT (

[2004] ICR 1289

), the then President, Burton J, at paragraph 18 had advanced the proposition that while common employment between claimant and comparator is insufficient, in asking whether the two-pronged single source test articulated by the ECJ is satisfied, “[t]his is almost invariably the case where there is a common employer: the best example of such a single source of such a body is a common employer”.

(5) In the same judgment at paragraph 21 (in a passage cited with approval by Mummery LJ in the Court of Appeal, at paragraph 28) Burton J stated that Lawrence is authority “that what underlines the applicability of article 141 is that which is ordinarily exemplified by common employment, namely the existence of a common source, the existence of a central responsibility for terms and conditions”.

(6) Burton J was therefore correct to identify at paragraph 22 “an evidential burden on an employer to displace the ordinary conclusion” and there was nothing wrong with the Judge here doing the same. Another past President, Underhill J (as he then was) in Beddoes v Birmingham City Council

[2011] 3 CMLR 42

, at paragraph 52 had also described as “certain highly unusual cases” those where there is a common employer but not a single source.

(7) The Claimants reserved the right to argue in a higher court that the domestic decisions in DEFRA and Armstrong were wrong. The correctness of the former was left open by Lady Hale in Dumfries and Galloway Council v North (at paragraph 41). For present purposes, they argue that the Judge was right to distinguish them on the facts, essentially for the reasons he gave.

(8) The Judge had been right to avoid the mistake made by the Employment Tribunal in Fox Cross Claimants , in which the then EAT President, Langstaff J, having reversed the decision on other grounds, correctly concluded (in obiter observations at paragraph 84) that the notion of “responsibility” to correct unequal treatment referred to ongoing rather than causative responsibility; and (at paragraph 87) that strategic management and powers in the context of a large organisation is sufficient without day to day intervention.

(9) The Inner House of the Court of Session on appeal in the same case, whose opinion was delivered by Lord Brodie, was likewise correct to observe in its obiter remarks at paragraph 49, that the focus should be on the ability of the single body to restore equal treatment and that in the DEFRA case, the authority to negotiate and set most aspects of the pay of civil servants employed by DEFRA was “specifically delegated by the Minister for the Civil Service by statutory instrument”, and the instrument would have to be revoked for the delegated power to be removed.

(10) As to the allegation of perversity in the fourth ground of appeal, the single source issue was for the Judge of fact and must stand unless perverse (per Nelson J in North Cumbria Acute Hospitals NHS Trust , paragraph 108); and the Judge had ample factual material before him, which he assessed without error, enabling him to distinguish cases such as DEFRA and Armstrong . His decision that this case fell on the other side of the line was eminently defensible and disclosed no error of law or impermissible finding of fact.

41. I come to my reasoning and conclusions on this issue. I note, first, that the Judge below (at paragraphs 180-182 of his Decision) rejected Asda’s argument that the presence of a single source of pay and conditions for both claimant and comparator is insufficient and that there must in addition be a single establishment, collective agreement or statutory framework.

42. He was right to do so. Mr Jeans accepts that recent case law assumes that the single source concept is a freestanding gateway to comparability, while contending that the assumption is wrong and should be corrected. I do not see any good reason for supposing the assumption is wrong. I think the reason Mr Jeans’ argument has been only implicitly rather than explicitly rejected is probably that until now, either no one thought of the point or no one thought it worth arguing.

43. That may be because in most cases the issue is unlikely to be decisive. But there is no good reason why article 157 (then 141) should be construed more narrowly after Lawrence than before it was decided. The ECJ in Lawrence did not say it was restricting the scope of article 141. The concept of a single source can be traced, as far as I am aware, back to the opinion of A-G Geelhoed in Lawrence at paragraphs 49-53. The ECJ adopted his single source concept.

44. The Advocate-General explained that the three situations in which comparison beyond the actual establishment where the claimant works - statutory rules for more than one undertaking, a collective agreement or regulations covering more than one undertaking and cases where employment terms are laid down centrally for more than one organisation within a conglomerate - all share the feature that employment terms are traceable back to one source.

45. He went on to say that since article 141 is aimed at the body responsible for unequal treatment, the relevant body - the legislature, parties to a collective works agreement or management of a corporate group, as the case may be - can be “held accountable in that regard”. The reason for identifying the single source of unequal treatment is because that source is the body with the power to cure the unequal treatment and therefore the body which should bear responsibility for it.

46. A-G Geelhoed’s analysis identified three separate categories of case where comparability is permitted outside a single establishment and identified what would ordinarily be their common feature. His account places responsibility where it properly lies and may obviate the need to enquire into issues of comparability across establishments that may otherwise arise. But I can see no basis in the case law and no policy reason why it should be treated as narrowing the scope of comparability, if the traditional Defrenne categories of comparability are still needed.

47. I do not find any misdirection of law in the Judge’s treatment of the single source jurisprudence. I conclude that the second ground of appeal fails. I consider next whether the findings of fact he made, and the conclusions to be drawn from them, were open to him. The findings of fact range from paragraphs 18-127 of the Decision. He traced the history of Asda and its acquisition in 1999 by Wal-Mart. He described the absence of co-location of stores and distribution centres or depots.

48. He went on to describe terms and conditions in stores and how they are set, in particular regarding pay, which is done without negotiation and without the involvement of any recognised union. Put simply, pay rates are signed off by a subcommittee of Asda’s executive board, including Ms Hayley Tatum who is responsible for pay in stores and consults with members of Wal-Mart’s international compensation team when formulating proposals. The “Total Remuneration Strategy” (“TRS”) is the name given to this process.

49. The Judge then did the same exercise for distribution staff terms and conditions, noting the growth of collective bargaining through the increasing influence of the GMB union which started to achieve recognition at some depots and eventually, after a period of difficult relations and difficult pay negotiations, entered into a national recognition agreement with Asda in May 2012.

50. The Judge described the establishment from about January 2010 of Asda’s “Industrial Relations Board” (“IRB”), its membership, including members of the executive board and the chief executive officer (“CEO”), the head of the employment law team, the general counsel and others including Mr Ian Stansfield, who regularly updated the IRB during the negotiations he was conducting with the GMB in 2010-2012. The Judge noted that pay rates were specific to individual depots.

51. The Judge then undertook a comparison of the terms of retail staff and distribution staff, to which I will return. He then considered “[f]acts material to the North hypothetical”, to which I will also return. Both Mr Stansfield and Ms Tatum were asked, and answered, questions about what pay rates and what terms would be if, hypothetically, a distribution worker should do his job at a store location, or a retail staff worker should do hers at a depot. The Judge rejected any suggestion that Asda had deliberately engineered the separation of retail staff terms from distribution staff terms in order to protect itself against equal pay claims.

52. The Judge set out his factual reasoning and conclusions on the single source issue at paragraphs 183-190. These paragraphs were subjected to intense criticism by Mr Jeans. The Judge started by saying that since there was a single employer an “evidential burden” rested on Asda to displace the proposition that there was a single source of pay. Despite Mr Jeans’ objection, I see no error there; the Judge’s observation was in line with that of Burton J in the DEFRA case.

53. The Judge reasoned that a single employer had “allocated responsibility to different internal structures for the purpose of setting terms and conditions”. The persons operating those internal structures do so under the authority of, and are accountable to, the common employer, personified by the Asda executive board. He reasoned that the position was not the same as in the DEFRA case. There, the delegation of power was not by a commercial undertaking but an arm of government. The delegation was embodied in statutory and constitutional arrangements between the Crown and the public service.

54. The Judge found at paragraph 185 that those charged with the decisions to set terms and conditions were “subject to a degree of oversight so that there was control in practice as well as in theory of those matters by the respondent’s [Asda’s] Executive Board”. That was in turn subject to the power of Asda’s sole shareholder, Wal-Mart, to “direct” Asda’s executive board “in the way that it chose” (paragraph 186).

55. There was “an absence of true autonomy” in the setting of both retail and distribution terms (paragraphs 187-188), said the Judge, referring back to his findings of fact on the way in which terms were set. He noted at paragraph 189 that the Respondent did not challenge the proposition that it could restore equal treatment, if there was unlawful equality of treatment. He concluded, at paragraph 190:

“… The Executive Board of Asda and the members of subcomittees of that Board had and exercised budgetary control and oversight over both Distribution and Retail at all material times. The Executive Board was responsible for the differences in pay and could, or could subject to the overarching control of Wal-Mart, have introduced equality.”

56. For all the intensity of Asda’s attack on the Judge’s reasoning and conclusions, far from considering them perverse I find them unassailable. This was an ordinary case of a large organisation delegating the setting of pay to separate internal organs. The Judge avoided the mistake made by the Tribunal in the Glasgow City Council case (see, in particular, paragraph 69 in Lord Brodie’s opinion) of treating the DEFRA decision as creating a category of dual source cases where power to set pay is delegated and then not interfered with on a regular basis.

57. In my judgment, the DEFRA case, if as I assume it is correctly decided, was wholly exceptional and turned on the unique position within the civil service, where the setting of pay and most other terms of civil servants on a departmental not national basis, was enshrined in statutory and other instruments that would have to be expressly revoked if the power was to be reclaimed by the Treasury and exercised centrally.

58. Here, by contrast, Asda or Wal-Mart could interfere at the stroke of a pen or, more likely, the click of a mouse. The Industrial Relations Board (“IRB”) included senior employees including the Asda’s CEO, with an interest in pay across both retail and distribution staff. Moreover, the IRB received regular updates on distribution staff pay negotiations, while retail pay was signed off by the executive board.

59. It would be unjust and little short of absurd if the presence of a single source turned on whether Asda’s or Wal-Mart’s board happened to have overruled Mr Stansfield, or the CEO or the board happened to have overruled Ms Tatum. It was quite enough to justify the Judge’s conclusion that the evidence showed that they regularly scrutinised the work done to set pay in, respectively, retail and distribution, and could overrule their decisions at any time.

60. In reply, Mr Jeans submitted that the Judge should have found the position analogous to that in the DEFRA case because there was a parallel here between the oversight and budgetary control exercised by Asda’s board, or Wal-Mart’s, and that exercised by the Treasury in the DEFRA case, as demonstrated by the finding highlighted in the EAT’s decision (at paragraphs 23 and 24 of Burton J’s judgment).

61. Burton J there related that each department had to submit an “annual remit”, and a “remit” before each negotiating round, to the Treasury showing how proposed changes to terms and conditions were to be funded. The Treasury then exercises a monitoring role by vetting the remits against “certain general principles” and in one case going beyond this, though it is not clear how. There is also a “management code” containing statements such as that bonuses should be linked to achievement, and the like. And pensions are reserved for central control.

62. I was shown some of the documentary evidence that was before the Judge. It included the following. There was an email exchange between Mr Stansfield and Asda’s CEO in 2008 answering a request for an update on recent distribution pay negotiations relating to particular depots, answered by a “steer” from the CEO, including a reference to 3 per cent in distribution being “dangerous vs [versus] retail”.

63. I was shown the slides for a presentation by the President and CEO of Wal-Mart to an “IR [Industrial Relations] Strategy Discussion” in September 2009, which included his “IR Vision”, embracing both retail and distribution staff. In January 2010 Mr Kuchinad, Asda’s executive people director, received the emailed thoughts of Mr Gooddie, an industrial relations expert member of the IRB, on industrial relations issues in both retail and distribution, in advance of a discussion to be held the following Friday.

64. In March 2010, Mr Gooddie was again emailing Mr Kuchinad, with a “back of an envelope” strategy for developing distribution pay and indicating that they should also discuss “the various links with retail”. In November 2010, something called the “Pay Strategy Working Group” received a tabulated overview of “Retail Pay” and “Distribution Pay”. It included a section containing some bullet points on “Culture and Communication across retail & depots”, and a comparative “Performance” table comparing retail staff against distribution staff.

65. I was also shown notes of a meeting held on 14 December 2010 of the executive committee of Asda’s executive board (referred to at paragraph 66 of the Judge’s Decision). It records that Mr Kuchinad updated the board on progress with the “3 year reward strategy” and “how this fitted into the overall people plan”. Mr Gooddie and Mr Stansfield then “gave an update on the ongoing work being done with the distribution pay strategy - this had been covered at the IR board”. The meeting decided that the “Distribution strategy should go back to the IR Board for discussion and approval”.

66. That evidence shows how far removed the facts were in this case from those in the DEFRA case, on which Asda relied so heavily. They also demonstrate that there was ample material before the Judge to justify the conclusion he reached. I conclude without difficulty that the third and fourth grounds of the appeal are not well founded.

67. I would add for completeness that there is nothing in the point that a “dual” source of pay and terms of employment, namely Asda and Wal-Mart, rules out the presence of a single source. If that submission were correct, equal pay law could be avoided at will. Mr Short is right that a single source means a common source and the “vertical” presence of a parent company sitting above its subsidiary does not make the source of pay and conditions other than common to both claimant and comparator.

The scope of comparability in domestic law

68. The fifth ground of appeal challenges the Judge’s approach to the interpretation of domestic law. The point is put in various ways but boils down to the complaint that the Judge was too much influenced by the policy of EU law rather than the meaning of domestic statute law. The Claimants say the Judge’s approach to interpretation of domestic law was impeccable and rightly influenced by the content of EU law; and they pointed to the obligation of an English tribunal to disapply domestic law where it conflicts with directly effective EU law.

69. I was referred to authority on how EU law, as it is now called, should or should not influence domestic law and its interpretation: on the Claimants’ side, Halsey v Fair Employment Agency

[1989] IRLR 106

NICA, per Lord Lowry CJ at paragraph 16; Scullard v Knowles

[1996] ICR 399

, per Mummery J at 403G-405C; and the Dumfries and Galloway case, per Lady Hale JSC at paragraph 45; and on Asda’s side, the traditional approach embodied in the speech of Lord Templeman in Duke v GEC Reliance

[1988] IRLR 118

, citing from that of Lord Diplock in Garland v British Rail Engineering Ltd

[1982] IRLR 257

, at 259.

70. This ground of appeal adds nothing of substance, for it can only succeed if the appeal is otherwise well founded. In its skeleton argument, Asda says the Judge’s conclusions on the application of domestic law are “unsafe”, as if they were a criminal conviction. If they are unsafe, they may be unsafe yet correct or unsafe and also wrong. If it is the latter, one or more other grounds of appeal will succeed and the fifth ground will not be needed. But if, though unsafe, they are nonetheless correct, the fifth ground cannot save the appeal from failure.

71. I will therefore pass on to more substantial matters. The sixth and seventh grounds, taken together, assert misdirection (the sixth ground) and perversity (the seventh) in relation to the Judge’s treatment of the issue of common terms and conditions of work observed, or applying, generally as between claimants and comparators, for the purposes of section 1(6) of the EqPA and, from 1 October 2010, section 79 of the EqA .

72. Until 30 September 2010, section 1(6) of the EqPA provided that men shall be treated as in the same employment with a woman (where they do not work at the same establishment) if both work at establishments in Great Britain “and common terms and conditions of employment are observed either generally or for employees of the relevant classes”.

73. From 1 October 2010 onwards, section 79 of the EqA defines who may be a comparator. Where the comparator works at a different “establishment”, the case must be one where “common terms apply at the establishments (either generally or as between A [the claimant] and B [the comparator]” (section 79(4)(c)). The reference in section 1(6) to “employees of the relevant classes” was not repeated (a point relevant to the tenth and last ground of appeal, addressed later).

74. In support of the sixth and seventh grounds, Mr Jeans for Asda made the following main points (the quotes are from his skeleton argument, save where indicated otherwise):

(1) The terms of the distribution staff and the retail staff are set by reference to the physical location of the staff concerned:

“the set of terms which applies depends on where you work”

. Therefore, there are no common terms across stores and depots. Further, each individual depot has its own pay rates.

(2) Comparing the content of the respective sets of terms is the wrong approach. The test is whether the claimants and comparators work under “essentially different employment regimes” (per Lord Bridge in Leverton v Clwyd County Council

[1989] 1 AC 706

HL, at 746C-E, whether for geographical or historical reasons, or a combination of both.

(3) Even if it were legitimate to compare the content of the two sets of terms, they constitute different packages of terms with different terms relating to matters such as working hours, sickness absence, productivity targets, breaks, probationary periods, increased pay after an initial period of employment, different forms of pay (basic, overtime, night pay and bank holiday pay) and Asda’s right to vary terms.

(4) The Judge should have found that the facts here are “remarkably similar” to Lord Bridge’s second example in Leverton (at 746E-G) of a case falling outside section 1(6): that of an employer who takes over a company which has a pre-existing collective agreement with a union covering one set of employees, which the employer continues to operate; such that at a different factory, a different collective agreement has produced different structures governing pay and other terms of work.

(5) The Judge wrongly failed to take account of the employment history and genesis of the two distinct sets of terms and conditions. He also erred by taking account of non-contractual terms and conditions; under the wording of section 1 of the EqPA , and by section 80(2) of the EqA , these are irrelevant to the “common terms” issue. Any discrimination outside contractual terms of employment is remediable under the (former) Sex Discrimination Act 1975 , not under the equal pay legislation.

(6) By asking himself whether it was “easy” or “straightforward” to undertake a “broad comparison” between the content of the two sets of terms, the Judge applied the wrong test. He should have asked himself whether the two sets of terms were common and should have regarded differences between them, e.g. in hourly rates of pay, night working, sick leave and holidays, as indicating non-commonality, not (as he did) commonality.

(7) The Judge erred by treating as material to the “common terms” issue the point that both sets of terms emanated from the same employer. By definition, in an equal pay claim such as this the claimant and comparator must be employed by, at least, “associated” employers, if not the same employer. The fact that both sets of terms are set by the same employer is therefore a neutral factor, not a pointer in the direction of the terms being common.

(8) As to the seventh ground (perversity), Mr Jeans explained that Asda did not submit that the Judge’s findings of fact were perverse and should be overturned but, rather, that the conclusion the Judge drew from them was perverse. Although this was Asda’s general position, it does complain that the Judge wrongly treated the distribution workers’ productivity targets as non-contractual and in the nature of a mere aspiration.

(9) On the facts found, and treating the productivity term as contractual, the only conclusion open to him was that the terms were not common. The two sets of terms were geographically, historically, and qualitatively distinct. The retail staff terms apply at stores only and are set by Asda’s board or a subcommittee thereof without reference to productivity. The distribution staff terms are set mainly by collective bargaining.

75. In defence of the Judge’s reasoning and conclusions, Mr Short countered with the following main arguments:

(1) The “common terms and conditions” requirement is a low threshold to cross because it operates at the preliminary stage, where the employer still has the opportunity to defend against claims either on the basis that jobs are not of equal value or by asserting objective justification for unequal treatment. The threshold requirement must not allow the employer to invoke “differences which may themselves be discriminatory in creating a barrier to an appropriate comparison” (Decision below, paragraph 210).

(2) Accordingly, a broad comparison of terms is appropriate and broad similarity between the two sets of terms suffices (per Lord Bridge in Leverton at 745F-746C; per Lord Slynn in British Coal Corporation v Smith

[1996] ICR 515

, at 526H-528B, 530D-E). And it is clear from Leverton , British Coal v Smith and Dumfries and Galloway Council v North (per Lady Hale JSC at paragraphs 12-13, and 33-41) that a difference in workplace location as between claimant and comparator does not preclude a finding of common terms.

(3) Asda’s erroneous approach is to try to fit the facts here into the straitjacket of Lord Bridge’s second example in Leverton (at 746E-G) and to demand the same factual result here, when each case is different on its facts and it is for the tribunal to find the primary facts and draw its conclusion from them, applying the statutory test, as the Judge did here. It is not the law that, because terms are negotiated separately over different periods for different groups, they cannot be common.

(4) The Judge did not misdirect himself by finding that the differences in the history and genesis of the two sets of terms were not determinative. Nor is there any bar to a comparison which looks at the content of the two sets of terms, as distinct from, or in addition to, the source of the respective sets of terms. The issue is one for the ordinary factual judgment of the tribunal ( South Tyneside MBC v Anderson

[2007] ICR 1581

CA, per Sedley LJ at paragraph 21).

(5) The Tribunal did not err by including in the comparison terms that were not contractual. The Judge agreed that the relevant terms were contractual ones and correctly looked at those terms in the proper context of the overall employment relationship. The word “observed” in section 1(6) of the EqPA , and the word “apply” in section 79(4) of the EqA , indicate that the Tribunal should focus on what happens in practice; cf. Holyland v Asda Stores Ltd

[2006] IRLR 468

Court of Session, Inner House, deciding that discretionary bonuses are outside the scope of the (former) Sex Discrimination Act 1975 .

(6) The Judge did not err by taking into account that the terms of both claimants and comparators were set by the same employer; they could have been set by associated employers but were not in this case; nor, as in Lord Bridge’s second example in Leverton , by a prior employer whose business and workforce had transferred wholesale to the current employer. The Tribunal did not treat the point as determinative but as part of the explanation for the similarities in the two sets of terms.

(7) On the seventh ground of appeal (perversity): the Judge was entitled to reach the conclusion that he did on the facts despite Asda’s two bases for criticising as perverse his finding that there were common terms as between claimants and comparators; namely first, that each set of terms had a “distinct genesis” and, secondly, that their “shape and content was fundamentally different”.

(8) As to the first of those (distinct genesis), the Judge was alive to the point but entitled to find that it was not determinative. As to the second (shape and content of the respective terms) the Judge was entitled to place little weight on differences between terms that were the very subject of the complaints (rates of pay, night and bank holiday premiums); and to disagree with Asda’s factual assessment by deciding that, analysing the differences and the similarities, the latter were sufficient. His finding to that effect is unassailable, not perverse.

76. The Judge’s account of the relevant law is at paragraphs 145-156 of his Decision. Asda does not criticise it. This is not surprising since it is mostly quotation from statute and case law. As to the facts, his findings concerning the source of the respective terms of retail and distribution staff have already been mentioned when considering the third and fourth grounds of appeal. He also compared the contents of the respective sets of terms, at paragraphs 88-110.

77. He noted that the processes for setting terms for retail and distribution staff were separate, because Asda had chosen to make them so. There was no prior TUPE transfer of sites en bloc creating inherited terms for either class of employee. He looked at Asda’s “handbooks” containing terms at different times. I was also shown some of these. He then examined tabulated comparisons of terms and analyses produced by the parties. These used different methodologies and naturally sought to influence the Judge towards opposing conclusions. I was also shown this material.

78. He then commented in detail on the similarities, and the differences, between the terms of retail staff and distribution staff, on a term by term basis. These comments are found at paragraphs 95-110 of his Decision. He noted that the Claimants’ table acknowledged that some terms are “not contractual or are derived from statute” (paragraph 99). In paragraph 100, he introduced his comments on specific terms, saying it appeared to him that:

“… subject to such weight as I attach to the respondent’s arguments to the contrary, there could be said to be a greater degree of variation in respect of the following: …”

79. He must, in my judgment, have meant to refer to a greater degree of similarity, or a lesser degree of variation, rather than to a greater degree of variation. Neither party contended otherwise. The narrative that follows is directed more to similarities than differences, under the following headings: pay cycle, paid holiday entitlement, holiday pay, breaks, paternity pay, mobility, working hours, variation clause, notice periods, probation, references and working for other employers.

80. That commentary is followed by comments on the “remaining areas of the analysis” which “show more substantial differences”, including principally that of pay, which he considered not relevant since the differences in pay were the very subject of complaint. The other areas were also at least in part pay related: shift pay, bank holidays, overtime and sick pay. He noted the areas of similarity and the differences under those headings.

81. He returned to the issue of common terms at paragraphs 192-217, in the part of his Decision setting out his reasoning and conclusions. He recorded the arguments of the parties, which were similar to those before me. He noted (paragraph 208) that Asda placed particular reliance on the productivity clause which was added to distribution staff terms when the national recognition agreement with the GMB was entered into.

82. At paragraphs 210-217, he gave his reasons for preferring the Claimants’ arguments. Essentially, he accepted the argument that the terms were broadly similar and rejected Asda’s arguments based on terms being tied to a specific location. In effect, he found that they created a danger of legitimised discrimination by job segregation. He found, on a factual comparison, “a significant correlation … in a broad way between the terms …” (paragraph 212).

83. The setting of terms by the same employer was not determinative but “is reflected in the strong similarities in the Handbooks and is for that reason a material factor” (paragraph 212). Such differences as he found were not “so extensive as to undermine the broad comparison which has to be made” (paragraph 213). He rejected the suggested technique of considering whether the facts fitted into Lord Bridge’s second example in Leverton of an impermissible comparison.

84. As to the productivity term for distribution staff relied on by Asda, at paragraph 216 the Judge referred back to its wording (at paragraph 78), set out in a “Memorandum of Understanding” and in a statement of distribution staff terms and conditions annexed to the national recognition agreement. It referred to the common goal of increased productivity. It said that “there may be a need to apply professionally measured work standards to the planning of work and methods” and that “colleagues will be required to perform in line with these set work standards and demonstrate the flexibility and collaboration required”.

85. As the Judge pointed out at paragraph 79, there was no specific link to pay, such as a provision that failure to meet productivity targets set by management would result in a loss of pay, bonus or other financial or other benefit. This was not a performance related pay term. Retail staff also had “productivity targets” but these were, on the evidence, not used to measure individual performance but, rather, for budgeting purposes.

86. At paragraph 216, the Judge said that he did not accept Asda’s argument that the productivity term represented “a significant difference between the group of Distribution employees as a whole and the Retail employees”. He said the term was “as much a statement of aspiration or intent than creating a contractual obligation”; so that if an employee were alleged to be in breach of the term, it would be an issue of capability or conduct rather than a breach of any specific obligation to meet a target.

87. The comparison of terms required under section 1(6) of the EqPA and, subsequently, under section 79 of the EqA , is an ordinary matter of factual judgment for the tribunal of first instance, subject to the normal constraints of law, as Sedley LJ pointed out in South Tyneside MBC v Anderson at paragraph 21. Aside from the Judge’s treatment of the productivity term, Mr Jeans did not criticise the Judge’s findings, only the conclusion he drew from them, which he said was perverse.

88. I reject Mr Jeans’ submission that the Judge was bound to find that the terms could not be common because they were tied to particular locations. That submission confuses the question with the answer to it. A comparison is being made of terms observed at different establishments. There are no comparators working at the Claimants’ establishment, and vice versa. So the starting point is that the terms that are being compared are observed at different establishments. You then look at whether they are “common” in the statutory sense, even though they are observed at different establishments. You cannot answer that question merely by repeating the point that they are observed at different establishments.

89. The submission that the terms cannot be common because their genesis and history are different is also wrong. Terms which have a different genesis and history could be identical in content. An example would be an employer who applies men’s terms to women at a different establishment, to avoid discrimination, recognising that the women’s work is of equal value.

90. Nor is there any bar to comparing the content of the respective sets of terms. To decide whether terms are common or not, it is useful to know what the terms are. The Judge was right to ascertain them and the parties were right to put them in evidence before him. I find his approach in line with that held to be correct by May LJ in his dissenting judgment in Leverton (approved in the House of Lords) and the approach of Lord Bridge in Leverton , of Lord Slynn in British Coal and of Lady Hale JSC in Dumfries and Galloway , at paragraph 12.

91. I do not accept that the Judge wrongly attached weight to the fact that the terms emanated from the same employer. That they did was “reflected” in their similarities, as he found; it was not a “determinative” point and it was not treated as itself establishing or contributing to the similarities he found.

92. The Judge did not err in attaching weight to non-contractual terms; he recognised them as such and as merely contextual. Nor did he err in attributing no weight to pay differences that were the very subject of complaint. Nor is it right to say that terms mandated by statute must be disregarded. Compulsory statutory terms can be and usually are also contract terms. There is no warrant for excluding them from the comparison exercise.

93. The sixth ground of appeal therefore fails. There was no misdirection. The seventh ground is an ordinary perversity challenge. It depends, first, on impugning the Judge’s treatment of the productivity term for distribution workers and, secondly, on his overall conclusion that the terms had sufficient broad commonality to satisfy the statutory tests. I find no lack of rationality or perversity in the Judge’s reasoning on either count and therefore no flaw in the conclusion he drew from it, strongly though Asda disagrees with it.

94. The productivity term may or may not be correctly described as contractual but, even if it was contractual, the Judge was entitled to find that it did not have “teeth” over and above the applicability of ordinary capability and conduct procedures. It was not suggested that it was a performance related pay term in the usual sense of that term or that it created a specific disciplinary offence of failing to meet a target.

95. Thus, an unproductive store worker could be dismissed for being unproductive, albeit without the benchmark of specific targets. An unproductive distribution worker’s lack of productivity is easier to measure if you have targets to measure against, but it was for the Judge to evaluate that difference and set it against other similarities, and alongside other differences, when applying the test.

96. The Judge’s overall conclusion that the terms were common in the statutory sense is not impeachable. He weighed the differences and the similarities, evaluated the arguments for and against commonality, and came down on the side of commonality on the facts and arguments presented to him. He was entitled to do so. The seventh ground of appeal also fails.

The “ North hypothetical” grounds of appeal

97. That brings me to the remaining three grounds of appeal. They all concern what is called the “ North hypothetical”, in the jargon of the equal pay legal community. This refers to a hypothetical comparator as discussed by Lord Slynn in British Coal Corporation v Smith and by Lady Hale JSC in Dumfries and Galloway v North . The issue it addresses is that raised by section 1(6) of the EqPA : whether common terms “are observed … for employees of the relevant classes”, i.e. the retail staff and the distribution staff.

98. As no retail staff work at distribution depots and no distribution staff work in stores, the Tribunal considered the hypothetical question (the “ North hypothetical”) whether, if distribution staff did their jobs at store locations, broadly similar terms would apply, such that it can be said that common terms are observed at both types of establishment. The Judge answered yes to that question.

99. The eighth ground of this appeal is that that the Judge misdirected himself when applying the North hypothetical test, and that he misapplied the test. The ninth is that the conclusion he reached having applied it - that common terms and conditions of employment were observed at the stores and depots for employees of the relevant classes, namely retail staff and distribution staff - was perverse and not open to him on the evidence.

100. The further issue arising under section 79 of the EqA (i.e. from 1 October 2010 onwards) is whether that issue is still open, or whether as Asda contends, comparison by reference to “classes” of employees is no longer available to claimants from that date. The tenth and final ground of appeal is that the Judge misinterpreted section 79 of the EqA by deciding that it continued to permit comparison of terms by reference to the North hypothetical test despite the absence of any reference to “classes” of employees.

101. In British Coal Corporation v Smith , Lord Slynn said at 526F-G:

“What … has to be shown is that the male comparators at other establishments and at her [the claimant’s] establishment share common terms and conditions. If there are no such men at the claimant’s place of work then it has to be shown that like terms and conditions would apply if men were employed there in the particular jobs concerned.”

102. In Dumfries and Galloway Council v North , Lady Hale JSC said at paragraph 12 that the “common terms and conditions” in section 1(6) are:

“… on the one hand, the terms and conditions under which the male comparators are employed at different establishments from the women and, on the other hand, the terms and conditions under which those male comparators are or would be employed if they were employed at the same establishment as the women.”

103. And at paragraph 33, she accepted the submission set out earlier, at paragraph 30 (with the italics in the original), that:

“… the tribunal should not speculate about the adjustments to the comparators’ present terms and conditions which might be made in the unlikely event that they were transferred to the claimants’ workplace. The hypothesis is that the comparators are transferred to do their present jobs in a different location. The question is whether in that event, however unlikely, they would remain employed on the same or broadly similar terms and conditions to those applicable in their current place of work. As Lord Slynn had recognised in British Coal Corpn v Smith … the object of the legislation was to allow comparisons to be made between workers who did not and never would work in the same workplace. An example might be a manufacturing company, where the (female) clerical workers worked in an office block, whereas the (male) manufacturers worked in a factory.”

104. I leave aside for the moment the tenth ground which raises a distinct point of statutory construction. Mr Jeans, for Asda, supported the eighth and ninth grounds of appeal with the following main arguments (the quotes are from his skeleton argument, with underlining in the original):

(1) In law, the North hypothetical test can only be satisfied if the “ scope ” or “ reach ” of distribution staff terms and retail staff terms was such that they are observed at, respectively, stores and depots. Whereas in the public sector context, terms are often national and not geographically grounded, here the anchoring of depot terms to the depot’s physical location precludes the North hypothetical test being satisfied.

(2) While accepting that the question is a hypothetical one, Mr Jeans nevertheless submitted that the Tribunal must not consider whether distribution staff terms “ might apply in stores in an alternative reality …”; instead, the Judge should have asked “do Distribution terms apply to those doing depot jobs in stores, and do Retail terms apply to those doing store jobs in depots?”

(3) The Judge was also wrong to suppose that a hypothetical distribution worker working in a store would not be in physical proximity to retail staff and customers. When applying the comparison, he should have rejected the supposition of two separate operations, one retail, one distribution, on a single site.

(4) The Judge should have asked the question whether a distribution worker doing his job at a store would work on terms broadly similar to “actual depot terms”; whereas the Judge asked the wrong question: whether “depot terms” would “apply to depot workers in a hypothetical new depot”.

(5) The Judge erred in law by finding that distribution workers would be unwilling to see their terms extended to store workers and would insist on retaining their terms despite working at a store location and not a depot location. There was no evidential foundation for the former proposition; as to the latter one, the likely negotiating position of a group of workers is not the focus of the test; rather, it is the geographical reach of the contractual terms concerned.

(6) As to the ninth ground (perversity): the Judge misunderstood the evidence of Asda’s two witnesses, Mr Stansfield who had responsibility for distribution staff pay and conditions, and Ms Tatum who had that responsibility in relation to retail staff. The challenge in cross-examination of those two witnesses was limited to the issue of pay. The Judge had effectively rejected their evidence and given insufficient reasons for doing so, amounting to perversity. They were not challenged on the point that as regards terms other than pay, neither group would have kept their terms on transfer from a depot to a store, or vice versa.

(7) Furthermore, the Judge took account of the irrelevant consideration that a distribution worker would retain his terms if temporarily seconded to a store, e.g. to deal with an emergency such as a flood at nearby competitor stores causing a surge in customer demand, as, the Judge found, had occurred in one particular instance in 2016. That was not challenged but was not relevant and should have been disregarded.

(8) On the evidence the Judge heard, the only conclusion open to him was that the North hypothetical test could not be satisfied. The terms were not geographically transferrable from a depot to a store and vice versa. The Judge was bound so to find and no other finding was available to him, such that his finding was perverse and the EAT should substitute the finding he was bound to make.

105. Mr Short defended the Judge’s approach to the law on the North hypothetical test and his findings of fact supporting his conclusion that the test was satisfied here. He made the following main points in answer to Asda’s submissions:

(1) It cannot be the law that both sets of terms under consideration must actually be observed at both depots and stores. That would mean the exercise is not hypothetical at all and the legislative purpose would be defeasible by the expedient of job segregation and gender segregation.

(2) The hypothetical question was one of “fact” - i.e. what would have happened - for the Tribunal. The Judge applied his mind to the question correctly and answered it in a manner that was open to him on the facts as found. There was no error of law.

(3) The Judge was right or, at least entitled, to assume that a depot job would not be done in physical proximity to retail staff and customers. Ms Tatum had herself observed that a depot worker would not drive his fork lift truck up and down the aisles of a store. To be employed at the “same establishment”, as the hypothesis requires, is not the same thing as the depot worker actually having to work in a store face to face with the public. That is an unwarranted gloss since an “establishment” can comprise a store and a depot at the same geographical location.

(4) When applying the hypothetical test, the Judge was entitled to make the finding that depot staff would be unwilling to give up their more favourable terms, on the basis of an inherent likelihood rather than direct evidence; it is in the nature of the exercise that there will rarely be such direct evidence. A tribunal is entitled to bring to bear industrial experience and common sense to the issue.

(5) It is thus a misplaced criticism to accuse the Tribunal of impermissible speculation about the likely negotiating stance of the depot workers. By the same reasoning, it is permissible and relevant for the Tribunal when applying the test to ask itself what would happen if the distribution workers did their jobs at the same establishment as a store; and the likely attitude of those workers, applying industrial common sense, is logically relevant to answering that question.

(6) The Judge did not err in taking account of the real fact (as distinct from hypotheses) that temporarily redeployed distribution workers had in the past worked at stores without losing their terms of employment. That fact is not in principle irrelevant and it was for the Judge to give it such weight, if any, as he thought fit. He gave it limited weight and that was not wrong.

(7) The Tribunal’s finding at paragraph 117 was that, according to both Mr Stansfield and Ms Tatum, distribution workers doing their job at a store would be “paid the rate for the job they were actually doing”; and that, as regards terms other than pay, “Retail terms would apply to Distribution employees deployed to work in stores and Distribution terms to Retail employees deployed to work in depots”. Those findings are unassailable and sufficient.

(8) The perversity challenge is a restatement of the same complaint but giving it a different label. It is merely a disagreement with the “factual”, i.e. hypothetical, finding made by the Judge. There is no flaw in the Judge’s treatment of the issue.

106. In evaluating these arguments, I start by reminding myself what the Judge said. He made findings relevant to this issue at paragraphs 111-127. He noted that there was evidence of temporary secondment of distribution workers to stores, without any change to their terms. He noted that concern was expressed within Asda about the potential implications for equal pay claims which this might have, but rejected any suggestion that the separation was artificially engineered to help provide a defence to any such claims.

107. He noted the difficulty arising from the unreality of supposing that retail jobs could be done alongside distribution jobs without some physical separation. He postulated some degree of separation within what could be called co-location within an establishment, such as “some small section of a Distribution centre being given over to a small retail outlet”.

108.

The Judge then recorded the evidence of Mr Stansfield and Ms Tatum on the issue, at paragraph 117:

“Both … were asked in evidence what would be the position in the event of Distribution employees, however unlikely that might be, performing Distribution work in stores. Both clearly answered that if the Distribution employees were carrying out Distribution work they would be paid the rate for the job they were actually doing. … Both witnesses also maintained their primary position that Retail terms would apply to Distribution employees deployed to work in stores and Distribution terms to Retail employees deployed to work in depots.”

109. I accept Mr Short’s submission that the last sentence must, to make sense of the paragraph, be taken to refer to terms other than pay. That interpretation is also consistent with the fact that his challenge in cross-examination of Mr Stansfield and Ms Tatum was limited to pay, and did not extend to other terms and conditions. Mr Jeans referred me to the written witness statements of Mr Stansfield and Ms Tatum dealing with this point. In footnote 178 in Asda’s skeleton, he submitted that the Judge may have misunderstood their oral evidence. I was not persuaded that the Judge misunderstood the evidence on that or any other point.

110. The Judge’s reasoning and conclusion are found at paragraphs 218-242 in his Decision. After summarising the parties’ submissions, he gave his reasons for preferring the Claimants’ arguments. He rejected Asda’s argument that weight should be attached to the proposition that a worker moving from one depot to another would take the new depot terms, each depot’s terms being separate and specific to it. He rejected the suggestion that distribution workers doing their jobs at a store location should be taken to be working face to face with the public.

111. He accepted that temporary redeployments from depots to stores, and the retention of depot terms during the period of secondment, is “not properly comparable”, though it “provides some slight support for the claimants’ case” (paragraph 240). He attached greater weight to what Mr Jeans called the speculation that distribution staff enjoy more favourable terms and (paragraph 241):

“… it is inherently unlikely that depot workers would be willing to see those extended to Retail employees if hypothetical relocation of Retail employees occurred in that direction and equally unlikely that depot workers would be willing to give up their terms if there were hypothetical relocation of them into stores.”

112. He also accepted (paragraph 242) that weight could be attached to the point that there is “significant similarity” between parts of the two sets of terms and that these would be “maintained in the hypothetical situation”, whether or not they were sufficient to amount to common terms generally.

113. I do not accept that the Judge should have adopted Asda’s proposition that the North hypothetical test cannot be satisfied unless terms are actually observed across establishments. That is to deprive the test of its hypothetical character, as Mr Short pointed out. Once again, it is no answer for Asda without more to show that, as things stand, depot terms are tied to specific depots. The Judge had to consider whether that would still be the case if the workers at a particular depot were to move away from their depot and work at a store.

114. I have no difficulty with the Judge’s treatment of the hypothetical co-located depot and store premises. Many establishments accommodate more than one type of activity. They do not have to be all mixed up together. In the local authority cases, you do not have to apply the test by imagining the comparator gardener watering plants in the classroom alongside the claimant teaching assistant. The hypothetical gardener can tend the plants outside while the children learn inside and the catering staff prepare the lunch in the kitchens, physically separated from each other and all within the same establishment.

115. Mr Jeans was unhappy with the Judge taking account both of direct evidence and indirect evidence on this issue. He criticised the Judge for taking account of certain direct evidence, namely of temporary redeployment. It was irrelevant, he argued. I reject that. The evidence of temporary redeployment was potentially probative but the Judge was careful to attach only very limited weight to it. It was not irrelevant in principle.

116. The criticism that the Judge engaged in speculation when considering the probabilities of what would have happened, is in my judgment equally misplaced. The Judge had to decide the hypothetical issue on such evidence, direct and indirect, as he had. The indirect evidence comprised the entire edifice of Asda’s and Wal-Mart’s governance systems and the history of collective bargaining and the setting of terms. More narrowly, it included the existence of differentials and the GMB’s success in securing recognition. Indeed, there was evidence in one of the emails I and the Judge were shown (dated 20 January 2010), of the shop workers’ union (USDAW) being regarded by Asda as a lesser force to contend with than the GMB.

117. I do not think the Judge was required to abjure what he clearly regarded as a sensible assessment of the probabilities, applying industrial common sense. Speculation is merely a pejorative term to describe the task the Judge was required to carry out. Hypothesis would be a more generous term. The Judge’s treatment of the hypothetical comparator exercise was not in my judgment flawed by error of law or perversity. The eighth and ninth grounds therefore do not succeed.

118. I come to the tenth and final ground, which concerns the period from 1 October 2010 when section 79 of the EqA entered into force. Mr Jeans submitted that section 79 had repealed the North hypothetical test; there was no repetition of the reference in section 1(6) to “classes” of employees. The Judge disagreed with the submission that the law had changed. In support of this ground of appeal, Asda’s main points were:

(1) Under section 79(2) and (4)(c), “A” may only compare herself with “B”, where they work at different establishments, if “common terms apply at the establishments (either generally or as between A and B)”. The EqA is a reforming statute as well as a consolidating statute. Parliament must be taken to have deliberately omitted the reference to “classes” of employees which is the foundation for the North hypothetical jurisprudence expounded in the British Coal and Dumfries and Galloway cases.

(2) The comparison is now, therefore, restricted to a comparison between A’s terms and B’s terms. They must now therefore share the same broad package of terms, even if that package does not apply “generally”, i.e. to all or most other workers at those establishments. In the circumstances mentioned by Lord Slynn and Lady Hale JSC, the comparison is no longer permitted.

(3) That view is supported by the authors of the Incomes Data Services (“IDS”) Handbook entitled Equal Pay , published in August 2011, who there drew attention to the changed wording and commented that “section 79(4) has the effect that a situation highlighted by the British Coal analysis is no longer covered by the new law …” and that “the scope for showing ‘same employment’ where claimant and comparator are employed at different establishments has been narrowed”.

(4) The government’s explanatory notes accompanying the Equality Bill, do not shed light on the issue one way or the other, contrary to the position of the claimants. A rectifying construction along Inco Europe lines is not possible as the stringent conditions required for such a construction to be adopted are not met.

(5) The new wording and corresponding new law did cut down the rights of claimants but has the virtue of simplicity and avoidance of the need to address difficult hypothetical questions. It meets complaints that the wording of the former section 1(6) was opaque and difficult to apply.

119. Mr Short, for the Claimants, supported the Judge’s reasoning and conclusions, offering the following riposte on this ground of appeal:

(1) A purposive construction of section 79 is needed to avoid licensing sex discrimination by job segregation, just as its predecessor has repeatedly been given a purposive construction in cases from O’Brien v Sim-Chem Ltd

[1980] ICR 573

HL, per Lord Russell at 580F; to Pickstone v Freemans plc

[1989] 1 AC 66

, per Lord Keith at 111F-112E; to Redcar and Cleveland BC v Bainbridge

[2007] IRLR 984

, per Maurice Kay LJ at paragraph 22; to Dumfries and Galloway per Lady Hale JSC at paragraph 34.

(2) Article 141 of the TEC and subsequently article 157 of the TFEU both contained express reference to equal value as well as equal work, clarifying, as explained in case law interpreting article 119 of the Treaty of Rome , what was already inherent in article 119. The principle of equal pay “is one of the foundations of the Community and … article 119 creates rights for individuals which the national courts must safeguard” ( Coloroll Pension Trustees Ltd v Russell

[1995] ICR 179

ECJ, at paragraph 26).

(3) EU law is supreme in this country, as Lord Denning MR explained in Macarthys Ltd v Smith at 692F-H. In similar vein, Lord Lowry CJ in Halsey v Fair Employment Agency

[1989] IRLR 106

, at paragraph 16, emphasised the corresponding importance of construing, if possible, domestic statutory provisions consistently with article 119; and the principle that article 119 confers an independent right against the employer where national equal pay legislation falls short of what article 119 requires.

(4) If that is not possible, then the statutory provision must be disapplied: per Lady Hale JSC in Dumfries and Galloway , paragraph 41. While that should be done if necessary, here it is unnecessary because, as she observed earlier in her judgment, at paragraph 3, the provisions in the EqA are “intended to be of equivalent effect” to those in the EqPA which they replaced. It follows that the analysis in British Coal v Smith and her own analysis in Dumfries and Galloway still holds good.

(5) The authors of the IDS Handbook extract relied on by Asda did not opine that the change of wording necessarily reflected a change in substantive law; they merely posited that it could do as a matter of language and left open “whether it reflects an intention to restrict the circumstances in which comparison may be made, or if it is merely the result of infelicitous drafting”.

(6) No violence is done to the language of section 79 of the EqA if Asda’s retrograde and literalist construction is rejected by reading into section 79(4)(c) the proposition derived from section 1(6) of the EqPA that the comparator “B” may be a hypothetical worker rather than an actual person.

(7) Acknowledging in the usual way, in line with high authority, the caution that must be exercised when looking at explanatory notes accompanying a Bill or Act of Parliament, it is striking that in this case the notes commenting on section 79 (at paragraphs 275-281), while not explicitly mentioning preservation of the North hypothetical test, say nothing about changing the substantive law and assert that article 157, “which has direct effect, will ensure that existing case law on the breadth of possible comparisons is carried forward”.

120. I now consider these rival contentions. The Judge below reached the conclusion that the law had not changed, on two alternative bases. The first (paragraph 251) was that there was “no suggestion at the time of the enactment of the EA 2010 that any change in the law was contended for or was intended by Parliament”; and that the elimination of the reference to “classes” of employees was probably to achieve consistency of language throughout the statute where “A” and “B” are used to denote claimant and comparator respectively.

121. He also regarded a narrowing of claimants’ rights of comparison as unlikely in the light of the EU law jurisprudence, which recognised comparisons across establishments and without a common employer. That, he reasoned, also pointed in the direction of a purposive construction of section 79 in harmony with that of the former section 1(6) of the EqPA (paragraph 252).

122. In the alternative, he held that if Asda’s construction of section 79 were correct, it would be inconsistent with the directly effective right under article 157 of the TFEU and to the extent that it narrowed the scope of the former section 1(6) by excluding the North hypothetical, it would have to be disapplied (paragraphs 254-256).

123. Care must be exercised in approaching this issue of statutory construction. No reference was made to any Hansard materials, either before me or below. The government’s explanatory notes are not an authoritative source of interpretation. It is undeniable that the plain wording of the statute no longer covers the North hypothetical case and I accept Mr Jeans’ submission that the requirements for an Inco Europe rectifying construction are not met.

124. Nevertheless, on balance I have come to the conclusion that the Judge’s construction of section 79(4)(c) is correct. I think it is unrealistic to attribute to Parliament an intention to reduce the scope of the right to bring a claim for this particular type of sex discrimination. The history of discrimination law generally over the past half century is one of steady expansion into new fields such as age and sexual orientation.

125. I cannot think of any other legislative initiative to reduce rights in the field of discrimination law. None was cited to me. It would be surprising if this were the solitary exception. In reaching this view, I take comfort from being in the exalted albeit obiter company of the incoming President of the Supreme Court, who in Dumfries and Galloway at paragraph 3 commented that the EqA “has now been repealed and replaced by provisions in the Equality Act 2010 which are intended to be of equivalent effect”.

126. I do not think the interpretation contended for by the Claimants strains the language of the provision to the point where it would be characterised as a rectifying construction, which may be adopted only in very narrow circumstances where there has been a clear drafting error. I think it is a slightly strained and strongly purposive construction arising from the need to read section 79 in its proper historical context and in line with the codifying function of the EqA and the use of simple language in it.

127. For those reasons, I agree with the Judge. I think the function of section 79 was consolidation rather than reform and I read 79(4)(c) as bearing the meaning that comparison is permitted where “common terms apply or would apply at the establishments (either generally or as between A and B (or B’s hypothetical equivalent) )”.

128. For completeness, I add two further points. The first is that I do not find the explanatory notes of much assistance either way and I do not base my view on them. They do not answer the question of interpretation of section 79 clearly one way or the other. I do note, though, that they are not inconsistent with what I regard as the better view of its meaning.

129. Secondly and finally, I have doubts about the alternative basis on which the Judge decided the question of interpretation. I am, as already explained, unconvinced that article 157 and its predecessors cover cases of hypothetical comparison at all, at any rate without further community or national implementing measures. The decision of the ECJ in Macarthys Ltd v Smith at paragraphs 7, 14 and 15 is authority against the proposition that it does.

130. The Court of Appeal’s question 2(a) was whether the principle of equal pay for equal work in article 119 applies “where a worker can show that she receives less pay … than she would have received if she were a man doing equal work for the employer” (my emphasis). The ECJ answered the question by characterising the case of what it called a “hypothetical male worker” as “indirect and disguised discrimination” requiring “comparative studies of entire branches of industry”. Unless EU law has changed since then, hypothetical comparisons are thus outside, at least, the direct effect of article 119, now article 157.

131. The decision of Elias J (as he then was) in the Walton Centre for Neurology case is additional domestic authority against the proposition that what is now article 157 embraces hypothetical comparisons: see his explanation of Macarthys Ltd at paragraphs 20-25. I therefore do not base my interpretation of section 79 on the Judge’s alternative reasoning. For the reasons already given, and not that reason, the tenth ground of appeal does not succeed.

Conclusion

132. It follows that the appeal must be dismissed and the Decision of the Employment Judge upheld. The Claimants can compare themselves with their chosen comparators for the purposes of their equal pay claims.