K Element & Ors. v Tesco Stores Limited

The appellants in this appeal, Ms Element and her colleagues, are claimants before the employment tribunal in equal pay claims againsttheir employer (or in some cases their former employer), Tesco Stores Limited, the respondent before both the tribunal below and this appeal tribunal. There are two groups of claimants: those represented by Leigh Day solicitors and those represented by Harcus Sinclair solicitors. I shall continue to refer to the two different groups by the names of their respective solicitorsor as the claimants and to the respondent to the appeal as the respondent or by its name. The appeal concerns two narrow points on the meaning and definition of a Job Evaluation Study (“JES”), a term set out and defined in s.80(5) Equality Act 2010 (“EqA 2010”) which is relevant to one of the three routes available for establishing equal work, and whether, and if so how, the statutory burden of proof provisions in s.136 EqA 2010 apply to the determination of the question of whether a study or scheme is a JESat a preliminary hearing.

The appeal is a small aspect of a large group of equal pay claims brought by approximately 10,000 mainly female stores based hourly paid Tesco employees, seeking pay parity with their better paid, predominantly male, colleagues who are based in Tesco’s distribution centres (the comparators) through the equality clause that is implied by statute into all contracts of employment. (Footnote: ¹) The gender disparity and pay differential as between the two groups of employees is not in dispute and Tesco accepts that the stores based staff receive a lower hourly rate of pay than the warehouse staff. In addition, the distribution centre staff receive payments in respect of weekend and night working enhanced overtime rates and bonus, which are either not paid to the claimants or are more favourable for the comparators.

But in order to compare themselves with their comparators, the claimants must first establish that their work is equal to the work of their chosen comparators pursuant to Chapter 3 EqA 2010. Although it risks over-simplification, it may be helpful to set out a very brief thumbnail outline of the equal work provisions to understand the context of the dispute before diving into the detail of the employment tribunal (“ET”) decision under challenge and the parties’ competing arguments.

The work of a female claimant (A) can be established as being equal to that of an identified male comparator or comparators (B) employed by the same employer (subject to various limitations not relevant to this appeal set out in s.79) in one of only three ways: (1) where A’s work is like B’s work, referred to as a “like work” claim where the jobs and roles are similar, typically where men and women perform the same job or do broadly similar work, but receive different pay for it; (2) where a JES has been conducted that has assessed or rated different jobs or work as being equivalent, in terms of the demands made on a person by reference to factors such as effort, skill and decision-making, often referred to as a “rated as equivalent” (RAE) claim; or, (3) if the work of the claimants and their comparators is of equal value, where the determination of whether the work is of equal value is undertaken in the litigation process, unlike the second route where the JES is freestanding to the litigation.

A claim under the second route, the RAE claim, will therefore usually be a short cut from the more lengthy equal value route to establishing equal work because the time-consuming comparative job analysis exercise has already been done. But a JES can cut both ways. Just as, in some circumstances, claimants may rely on a JES that has rated their work as equivalent to their male comparators to establish equal work (so-called sword cases since they are used to advance a claim), an employer can also, in certain circumstances, rely on a JES that has rated jobs or work as unequal, or not equivalent to the work of their comparators, to defeat a claim for equal value, thereby avoiding the lengthy litigation involved in equal value cases (so-called shield cases as they are used to resist a claim). The detailed provisions in shield and sword cases are not however symmetrical and are considered in detail below since both sides relied on them in support of their arguments.

Of the three routes to establishing equal work, equal value cases are the slowest and most tortuous route. Described by Lord Bridge in one of the early landmark cases of Leverton v Clwyd County Council [1989] IRLR 28 as “lengthy, elaborate and…expensive [33]”, in spite of valiant legislative and procedural amendments over the years intended to reduce the much-criticised complexity and delays in equal value claims, it remains slow and technically complex. There are now fairly prescriptive procedural rules in an equal value claim involving three stages, with time limits ostensibly attached.

If a claimant cannot establish equal work via one of the three routes, the claim will fail as the comparison between her work and that of the better paid male worker will not be an appropriate comparison for the simple reason that there is no entitlement to equal pay when the work is not equal. But if equal work is established via one of the three routes, and subject to an employer establishing a material factor defence that does not involve direct or indirect sex discrimination, a sex equality clause is implied into every contract of employment which has the effect of modifying any less favourable contractual term as between men and women so as not to be less favourable. The claimant’s contractual terms will thus be levelled up to the contractual terms of her comparator engaged on equal work.

The claimants in this case instituted claims in the employment tribunal in February 2018, for work RAE and for work of equal value to that of their comparators. A small number also have asserted like work claims which are not relevant for the purposes of this appeal.Job titles of the comparator roles have been identified, but individual comparators had not, at the date of the ET preliminary hearing been identified but have since been identified.

The claimants’ RAE claim is based on what Tesco described as an informal exercise that was carried out in 2014 which had scored the demands of certain roles or activities within Tesco, given them an overall score and ranked them in score order (“the 2014 Exercise”). The issue between the parties is whether or not the 2014 Exercise is a JES under s.80(5) on which the claimants are entitled to rely to establish equal work, thus obviating the need to establish equal work via the equal value route.

In all cases brought under EqA 2010, s.136 provides for the burden of proof to shift from a claimant to a respondent in certain circumstances. An issue between the parties at the hearing was whether s.136 had any application when the tribunal was considering the JES question and, if so, whether the burden had shifted to the respondent to disprove that the 2014 Exercise was a JES, instead of the claimants being required to prove that it was. It did not affect the outcome of the case before the tribunal, but it had been raised in a proposed cross-appeal by Tesco. At a preliminary hearing before this tribunal (which had been directed by HHJ Auerbach on a consideration of the appeal under the rule 3(7) sift procedure), HHJ Tayler gave the parties permissionfor the issue to be considered at the substantive appeal.

The case was decided by Employment Judge Manley, sitting alone, in the Watford employment tribunal at an open preliminary hearing over 6 days with 2 days of deliberation in October 2020. Judgment was promulgated and sent to the parties on 29 October 2020. The tribunal concluded that:

At the tribunal, it was common ground that in order for the 2014 Exercise to be a JES, it needed to: (1) intend to evaluate the demands of jobs done by some of Tesco’s employees; (2) assess the work by reference to particular factors, rather than the overall job content in order to be analytical; and (3) take into account factors only connected with the demands of the job. But the consensus ended there. The Leigh Day claimants and Tesco accepted the authority of Eaton v Nuttall [1977] ICR 272 EAT that in order to be a JES, often referred to as a “valid” JES (Footnote: ²), it must satisfy “the test of being thorough in analysis and capable of impartial application.” The Harcus Sinclair claimants accepted that the ET, as a first instance tribunal, was bound by the EAT authority of Eaton, but reserved the right to raise it on appeal, which they have duly done.

The respondent’s case was that yet more was required in order for a study to be a JES. The tribunal must test the study rigorously by reference to the following 9 factors, that it be: (1) thorough in analysis; (2) capable of impartial application; (3) objective; (4) transparent; (5) accurate; (6) internally sound; (7) internally consistent; (8) sufficiently detailed; and, (9) fair. There was then a further question which was whether it was complete, which would require the claimants to prove to the tribunal three aspects: that it was (1) established with the respondent’s authority; (2) a finished product and (3) accepted or adopted by the employer and employees as a valid study regulating their relationship.

On the burden of proof issue, it was the claimants’ contention that the burden had shifted to the respondent to prove that the 2014 Exercise was not valid, because it was undisputed that the 2014 Exercise had rated jobs, including some of those of the claimants. The respondent’s case was that the burden was on the claimants to show equal work and lower pay (as well as no material factor defence) before the burden shifted to the respondent under the statute. A dispute between the parties also arose as to whether what constitutes a JES is different depending on whether the study in question is being used as a shield or a sword.

The tribunal conducted a helpful review of the authorities relied on by both parties, the statutory code of practice produced by the Equality and Human Rights Commission (EHRC) in 2011 and non-statutory guidance notes, ACAS guidance and textbook authority on statutory interpretation and various employment law textbooks such as Harvey on Industrial Relations and Employment Law and the IDS handbook.

The tribunal found that between 1991-2001 Tesco had used a JES (“the 1991 Scheme”) which had been agreed with the recognised trade union, USDAW. Tesco and USDAW agreed to abandon the 1991 Scheme in 2001. USDAW had asked for a replacement scheme to be introduced on several occasions since then, but it had not been progressed with them.

The tribunal found that at the end of 2013, a Tesco reward manager with responsibility for stores salaried managerial staff, Ms Caroline Yik, and Mr Paul Hunt (a reward manager with responsibility for stores hourly paid staff) embarked on an exploration of “what a new job sizing framework might look like”. It was thought to be a good idea since it could affect more than 38,000 staff at that time. The purpose was to explore how a simpler scheme than the 1991 Scheme might work. Work on the project was agreed by their joint line manager on the basis that it was exploratory. After Ms Yik was promoted to lead reward manager in May 2014, no further work or scoring was undertaken although the PowerPoint slides disclosed to the claimants in the course of this litigation were prepared. USDAW was not aware of the 2014 Exercise or anything to do with the process.

Ms Yik’s evidence, which the tribunal accepted as accurate, was “that it was never envisaged that there be a “fully fledged” job evaluation scheme but that issues with “job sizing” were to be explored.” The Tribunal made detailed findings of fact about the nature of the 2014 Exercise – that it was informal and exploratory and about “job sizing” rather than “job evaluation”. The tribunal found that:

Some work was done on proposed factors, factor definitions and job weighting to fashion a job evaluation tool. Seven initial roles were initially scored, but without any job descriptions or other written work, information, or involvement from any line managers for any of the seven roles, or the trade union, USDAW. Some further scoring was undertaken, some of which, for example customer assistant and fishmonger, resulted in very different and inconsistent outcomes when the role was scored more than once. Internal emails described the 2014 Exercise as “still a work in progress” in May 2014. In February 2015 Ms Yik described it as not “fully developed” and “needs a lot more work and ultimately would need to be verified.”

Four proposed factors were identified – (1) knowledge and skills; (2) interpersonal skills (3) judgment and decision making; and (4) responsibility for resources. It was decided not to include what were acknowledged to be the commonly used factors of “physical” because it was “not necessary/relevant for establishing role size; risk of discrimination”, nor “working conditions” because “[it] can recognise in pay e.g., night premiums but not relevant for job sizing” or “effort” for which the rationale was “include within interpersonal skills and responsibility for resources.”

In preparation for the preliminary hearing both sides had instructed job evaluation experts who both gave evidence to the tribunal and had produced reports. The claimants’ expert had been instructed to prepare her report and give evidence at the tribunal to assist the tribunal in determining whether the 2014 exercise “was a JES in compliance with s.80(5) EqA2010”. The respondent’s expert, Mr Michael Bourke, had been instructed “to prepare a report on the merits of” the 2014 exercise and “specifically, whether the 2014 exercise could be considered a valid JES”. On the face of it, although differently worded, their instructions were materially identical, but both experts were working to their respective legal team’s different interpretation of a JES. The respective experts did not disagree to any significant degree about the primary facts of the 2014 Exercise but applied those facts to their different respective understanding of what constituted a JES and unsurprisingly reached different conclusions. Mr Bourke considered that the 2014 Exercise was flawed and its outcomes unsafe to be relied on. The ET judgment sets out his executive summary which identifies 3 concerns about whether the 2014 Exercise was analytical. He concluded that the factors selected did not cover all the main job demands of the jobs covered by the scheme. He considered that physical effort, physical skills, dexterity and working conditions were significant features of some hourly paid jobs. Secondly that the factor and factor level definitions were inexact and, in many instances, did not describe the meaning and scope of each factor. Thirdly “there was no rationale established for the scoring and weighting system. Jobs were evaluated using a factor plan with four separate weightings which produced four separate evaluation outcomes. These were added together, and an average taken. No final weighting and scoring model were ever agreed.” In summarising his oral evidence, the judge noted that “he agreed that the 2014 Exercise was analytical, the factors chosen did not evaluate the jobs fairly, objectively and in a non-discriminatory manner.” There was also a lack of quality assurance or moderation of scoring or checking for disparate outcomes.

Ms Edgar concluded that although the 2014 Exercise was carried out in a relatively short timescale and good practice guidance not consistently followed, it was “a basic, analytical JES”. She accepted that physical effort was missing from the 2014 Exercise but considered that this could in part be included in “technical skills” and in any event did not know if it was a relevant demand. She accepted that working conditions such as risk of slips, trips, and falls, should be measured if they were significant risks, but was not sure that they would be a demand of the job.She accepted that the job holder and line manager would be best placed to have input into the demands of the job. She also accepted that a number of good practice aspects were missing and there was no written rationale for the scoring of the jobs considered, but overall considered it was a “very basic analytical exercise.”

From her recitation of the experts’ evidence the judge concluded that the facts were, for the most part, agreed [70]. None of the parties sought to argue that there was a lack of clarity or insufficient factual basis for the tribunal’s conclusions.

The tribunal first set out its conclusions on the dispute on the law. It accepted the claimants’ arguments on the burden of proof and concluded that the burden had shifted to the respondent to prove that the 2014 Exercise was not a JES.

She next considered the statutory interpretation question. She started with the words of the statute and the Harcus Sinclair interpretation of the meaning of a JES:

But since she considered herself bound by the Eaton v Nuttall test, she next considered if the 2014 Exercise wasthorough and capable of impartial application and found that it was not:

However, in spite of what the claimants submit, I do not stop there. The respondent and the Leigh Day claimants accept that, at the very least, the Eaton v Nuttall test, that the exercise must be thorough and capable of impartial application, must be applied. I accept that this is the case. Not only Eaton v Nuttall itself but several cases since have approved that test. When I look at various aspects of the 2014 exercise, I cannot accept either that it was thorough or capable of impartial application.”

In applying the Eaton v Nuttall test, she considered the thoroughness of the 2014 Exercise and found that it was not thorough for three reasons. Firstly, because the factors chosen did not cover the demands made on the job holders as there was no factor for physical effort or skills which, in particular, the judge considered to be a “serious omission”. Secondly because those designing the job evaluation tool and scoring had no job information, only very limited job descriptions for a small minority of the jobs and had not spoken to either job holders or line managers. Thirdly because there was no testing or checking as the scores were recorded and nor had they been subsequently moderated.

She then assessed whether the 2014 Exercise was capable of impartial application. She found that it was not because of the lack of training for either the scheme designers or the scorers, and also because there was no record of how the scores had been arrived at. The problem was compounded by the lack of either job holder or trade union involvement that created a real risk of the exercise not being applied impartially which were “fundamental errors” (a reference to Greene v BroxtoweDistrict Council [1977] ICR 241 discussed further below).

Although her conclusion decided the preliminary issue, out of respect to the parties’ detailed and lengthy arguments, the judge went on to deal briefly with the other matters raised. As she put it “I am not sure that I need to go further than that but appreciate that does not deal with all of the arguments before me.” [93]

Firstly, she found that there was no distinction in the definition between a JES raised as a sword and a JES raised as a shield to a claim [94]. Secondly, that if it was necessary for the elements of authority, completion, and acceptance to be present for an exercise to come within the definition of a JES (as per the authority of Arnold v Beecham Group [1982] ICR 744), the 2014 Exercise was not a valid JES. Although the 2014 Exercise had been conducted with the respondent’s authority to the extent that Ms Yik’s line manager and other managers knew the exercise was taking place “it was limited to an exploratory exercise only” and it was incomplete. The judge did not explore in detail the meaning of “acceptance” and the case of Arnold, beyond repeating that the 2014 Exercise was incomplete. She sensibly avoided being drawn into the debate about whether Arnold is correct, merely noting that she was bound by it [97].

Her reasons were as follows:

The applicable current statutory provisions are contained in chapter 3 EqA 2010. The three routes for establishing equal work are set out in s.65. In a like work claim, a claimant must establish that her work is the same or broadly similar to that of her comparator and that such differences as there are between their work are not of practical importance in relation to the terms of their work (s.65(2) and (3)). The case law over the years has encouraged tribunals not to take too pedantic an approach, or undertake minute examination,or assess the matter “too narrowly” (see for example Capper Pass v Lawton [1977] 2 All ER 11 and Eaton Ltd v Nuttall [1977] ICR 272 @276 at E).

The definition of a JES is set out in the interpretation section at s.80(5)

In light of the parties’ submissions, it is necessary to set out a brief history to the current provisions. When the right to equal work was first introduced by the Equal Pay Act 1970 (EqPA 1970), there were only two routes: like work and RAE claims. The RAE route was defined as follows:

At that time a JES could only be used as a sword to advance a claim for equal pay since if the jobs being compared were not like work and there was either no JES, or the JES had rated the jobs as being of different value (subject to the JES not being gender specific), no claim could be brought. In 1984 the EqPA 1970 was amended by the Equal Pay (Amendment) Regulations 1983 (SI 1983/1794) enabling equal value claims to be brought as a third route to equal work. It also introduced a shield mechanism in the introduction of a new s.2A to EqPA 1970 by deeming that, where the work of a claimant (A) and her comparator (B) had been given different values in a JES, and there were no reasonable grounds for determining that the evaluation in the JES was made on a system that discriminated on grounds of sex, A’s equal value claim against B would have no reasonable prospects and would be struck out.

The logic of the amendment was obvious. If a JES had been undertaken and found that the two roles were not of equal value it would follow that a claim that the two jobs were of equal value under the new third equal value route, would not have reasonable prospects of success. The shield mechanism would act as an incentive for employers to undertake a JES. Where the JES showed the women were being underpaid, it was hoped it would enable the employer and employees and their trade union to resolve the inequality through negotiation and collective bargaining without the need for litigation. Where the work was found not to be equal by the JES, the employer could use the JES as a shield to avoid lengthy equal value litigation. However, it came with a safeguard for employees. If there were reasonable grounds for determining that the JES was tainted by sex discrimination, it could not be relied on by the employer as a shield to an equal value claim by those whose work had been rated as unequal.

The EqPA 1970 was further amended by the Equal Pay Act 1970 (Amendment) Regulations 2004 (SI 2004/2352) which marginally strengthened the use of a JES as a shield by introducing a presumption that a JES was not made on a system which discriminates on the grounds of sex. It also introduced an explicit provision enabling a claimant to continue with an equal value claim where a JES had found her job to be of a lower value than her comparator B’s job, if she could prove that it was “otherwise unsuitable” (s2.A(2A) EqPA 1970.) The various iterations of the legislation are set out in the annex.

The current provisions governing when a JES can block an equal value claim are set out in s.131 Equality Act 2010:

“s.131 Assessment of whether work is of equal value

The first case before the appeal tribunal to consider the meaning of a JES was Greene v Broxtowe District Council [1977] ICR 241. In that case it was accepted by the tribunal that there had been a JES which found that there should be no variation in grade between the full-time male rent collectors and part-time female rent collectors. But neither the employer nor the unions or apparently the claimants had accepted the JES’s conclusions in full, which led the tribunal to reject the RAE claim and decided it would only consider the claim as a like work claim. On appeal, the EAT (Kilner Brown J) found that the tribunal had misdirected itself and that:

The study could not be set aside because neither side liked its conclusions. There is an interesting discussion in the judgment about whether the case should be remitted back to the tribunal to decide or the EAT exercise its power to substitute the tribunal decision. The EAT came to the conclusion that “this tribunal is not the proper place to investigate the validity of the evaluation study” and remitted the matter back for the tribunal:

Next up was Eaton Ltd v Nuttall [1977] ICR 272, an appeal by an employer against a finding of like work by the tribunal, in which it transpired during the course of the appeal hearing that there may well have been a JES, although the case had not been brought as an RAE claim. Per curiam, to provide assistance in future cases, since Phillips J and the tribunal members had reached a conclusion on the matter, the appeal tribunal stated the following:

Thus, the Eaton test of “being thorough in analysis and capable of impartial application” was born. In the next case to reach the EAT, Kilner Brown J was the presiding judge in Hebbes v Rank Precision Ltd [1978] ICR 489 and noted that the proposition he had stated earlier in Greene v Broxtowe DC had been slightly modified (which can only be a reference to Eaton) but that:

The Eaton test, as it has come to be known, has been repeated with approval in many, if not all, the subsequent cases, with the possible exception of Bromley and ors v H & J Quick Ltd [1988] ICR 623 when the Court of Appeal first came to consider these provisions discussed below.

In England v Bromley Council [1978] ICR 1 the claimant’s challenge was to his receiving less pay than his female comparators on the basis that his work and that of his comparators had been rated as equivalent under a scheme that had been recommended to London local authorities by the Greater London Whitley Council known as the “London Scheme”. However, his employer, Bromley Council, had varied the London Scheme and introduced special factor points in the scheme. When they applied their customised scheme to their workforce, it resulted in Mr England’s comparators being awarded an additional 5 special factor points, resulting in the jobs not being considered of equal value under the schemeand the women were awardedhigher pay. Mr England’s appeal to the EAT was rejected:

The EAT further held:

In O’Brien v Sim-Chem Ltd [1980] ICR 573, the House of Lords considered s.1(5) EqPA 1970 and the RAE provisions for the first time. A job evaluation study had been devised and agreed by management and unions at Sim-Chem Ltd, the grading exercise had been undertaken and the employees informed of their new job grade and salary range. They were told it would take effect and amend their contracts on a particular date. There would also be a merit assessment scheme to decide the exact pay of each employee within the salary range for their pay grade. But the new graded pay structure was not implemented as Sim-Chem Ltd feared that its introduction would infringe a government pay and salary restraint policy then in force for which they would be sanctioned. Nor was the individual merit assessment undertaken. The issue before the court was whether the women claimants could rely on the new scheme to claim equal pay to their male comparators who had been given the same grade under the new scheme, but who were paid more than them under the pre-existing arrangement. The House of Lords noted that it was not compulsory to carry out a JES, but rather required the agreement of the relevant parties, including the employer, that there should be one (@578 H). The lead speech was given by Lord Russell:

The argument advanced by the employer that since they were under no statutory obligation to participate in a job evaluation exercise, they were under no obligation to implement it, was rejected by the court:

In Arnold v Beecham Group [1982] ICR 744 EAT a study was conducted, again with the agreement of both the union and the employer, but the results were objected to by a group of staff affected and there were also aspects that the employer found unsatisfactory. Following discussions between management and unions the study was not implemented and did not form the basis of the next wage settlement. The claimants were women who would have stood to benefit from the job evaluation exercise who were seeking to rely on it to claim equal pay. The EAT (Browne-Wilkinson J (as he then was)) framed the question as being “whether a job evaluation study can be treated as “complete” until the parties to it (namely the employers and the employees) have accepted it as a valid study”. He considered that the point had not been dealt with in Sim-Chem v O’Brien where both the employer and the employees had accepted the study. Relying on the authorities of England v Bromley LBC, Greene v Broxtowe DC and Hebbes v Rank Precision Industries discussed above, the EAT concluded that:

The EAT also considered that it accorded with good industrial relations and common sense:

Continuing chronologically, the next relevant authority was Bromley and ors v H & J Quick Ltd [1988] ICR 623 CA, after the equal value route had been introduced. The issue for determination was whether the employer could rely on a job evaluation study to defeat the claimants’ equal value claims pursuant to s. 2A EqPA 1970, since the study had given the female claimants’ jobs a different, lower, value to the jobs of their male comparators. Dillon LJ eloquently summarised what have come to be referred to as the sword and shield uses to which a JES can be put as follows:

Noting that the study was being relied on by the employer he identified that the onus was on them to show that there was a JES that satisfied the requirements of s.1(5) and also that there were no reasonable grounds for determining that the evaluation contained in the study was tainted by sex discrimination. On the latter point there was discussion about the need for objectivity in the evaluation of the jobs in order to achieve the non-discrimination principle. He cited the Eaton test, but with a word of caution given the inevitability of some subjective, value judgments being applied in a job evaluation exercise.

Accordingly, elements of subjectivity will not necessarily mean that the evaluation contained in the study was tainted by sex discrimination and it will need to be considered carefully by the tribunal when considering s.2A(2)(b).

He continued:

Dillon LJ then rejected criticism of the use of the word “analytical” in the Eaton test as being a gloss on the statute:

For Mrs Bromley and her colleagues, whose jobs had not been valued in accordance with s.1(5), the employer could not rely on the job evaluation scheme which valued her job at less than that of her male comparators to show that there were no reasonable prospects in her equal value claim. Her grade had been allocated by “slotting-in” on a “whole job” basis and no comparison had been made by reference to factors between the demands made on her and her male comparator. There was no valid JES. Since the employer had not shown that the scheme was a study within s1(5), the appeal was disposed of. There was no need for the Court of Appeal to decide how far the tribunal was required to investigate matters under s.2A(2)(b) to decide that there were no reasonable grounds for determining that the evaluation contained in the study was made on a system which discriminated on grounds of sex. However, referring to s.2A(3), the concurring judgment of Woolf LJ (as he then was) noted that “a defective study could still, at least in theory, assist in establishing that there are no reasonable grounds for determining that the work is of equal value.” [639 @D-E]

The issue next came to the EAT in 2004 in two separate cases, both presided over by Mr Justice Burton, then President of this tribunal: Department for Environment Food and Rural Affairs (DEFRA) v Robertson and others [2004] ICR 1289 followed by Diageo plc v Thomson (EATS/0064/03) (unreported) in the Scottish EAT. In DEFRA, a sword case, the claimants sought to use a civil service wide job evaluation scheme (JEGS) and the main issue in the case concerned whether the claimants could rely on a JES to show that their jobs were RAE to that of female comparators in different government departments and whether they were “in the same employment” as defined in the legislation, which is not relevant for the purposes of this appeal. In dealing with a second, fact sensitive, somewhat minor issue Burton J endorsed both Eaton v Nuttall and Bromley v H & J Quick and described both authorities as helpful. The apparent qualification in Bromley v Quick of Eaton v Nuttall was not explored in the judgment.

In Diageo, an extempore unreported decision, the issue of s.2A EqPA 1970 was however at the centre of the appeal. The tribunal had found that the respondent’s job evaluation exercise that had rated jobs as being of different values was not a JES and could not be used as a shield to an equal pay claim. The employer appealed. Once again Burton J endorsed the Eaton test describing it as “extremely clear and concise. It had stood the test of time, is one which is to be welcomed, and could be operated by an employment tribunal.” [16] He noted that it is the study which must be capable of impartial application and a partial application of a study capable of impartial application does not make the study itself invalid [13]. In distinguishing between the robustness of the study itself as opposed to the workings and correctness or otherwise of the mathematical computations in applying the study, he gave the hypothetical example of a study that

Burton J noted that without following the approach set out in Eaton there was a very real risk of a tribunal falling into error.

The defects in the Diageo scheme that entitled the tribunal to conclude that it was invalid were an absence of a Central Evaluation Panel; the presence of only one trained evaluator with the consequent absence of checks and balances of a second evaluator; the absence, to an extent, of audit reports or continued monitoring; and the lack of rationale sheets.

We can move swiftly on from Home Office v Bailey and ors [2005] IRLR 757 CA, which although concerned RAE claims are not on point in this appeal (Footnote: ³).

The next case on JES and RAE to come before the appellate courts is Armstrong and ors v Glasgow City Council; McDonald and ors v Glasgow City Council [2017] IRLR 993 CSIH. It followed the move to single status pay bargaining in local government for different groups of employees – manual, administrative, technical, professional, and clerical - from the previous separate table bargaining and pay structures (identified by reference to the colour of the handbook, such as white book, purple book, green book and so on). Glasgow City Council carried out a bespoke job evaluation exercise to implement single status so that the council employees could be moved onto the single status pay scale. The claimants, who had been rated under the council’s scheme at lower than where they expected to be in the new pay scale, brought equal value claims. The employer relied on its scheme as a shield under s.2A EqPA 1970 to defeat the claimants’ equal value claims. The tribunal found that the Glasgow scheme could be relied on by the employer, it constituted a JES under s.1(5) and, applying s.2A(2A) there were no reasonable grounds for suspecting that the evaluation contained in the study was either made on a system which discriminated on grounds of sex, or was otherwise unsuitable to be relied on, resulting in the dismissal of the claimants’ equal value claims. In the Court of Session (Inner House), Lord Menzies giving judgment, found that under s.2A(2) the burden of proof lies with the respondent to show that its scheme was compliant with s.1(5). Before both the tribunal, and the CSIH on appeal, the parties were agreed that the respondent would have to provethat the exercise relied on “was (a) thorough in analysis, (b) objective, (c) transparent, (d) accurate, (e) internally sound and consistent, (f) sufficiently detailed and (g) fair” to be a JES within the meaning of s.1(5) EqPA 1970. The CSIH accepted the parties’ agreed position without analysis.

The Court then considered what was required of the respondent to discharge its burden of proof.

On the facts of the case, the respondent had not discharged its burden. The scheme relied on by the employer was bespoke, novel and untested with various unusual features and the appeal court found that without expert evidence before the tribunal the respondent had not led sufficient or sufficiently persuasive evidence that the scheme had met the relevant tests to comply with s.1(5). The absence of expert evidence meant that the respondent had failed to discharge the burden that the JES was compliant with s.1(5) and the decision was quashed: the case was remitted to the tribunal to consider the question of equal value. The shield had not protected the respondent from an equal value claim.

In addition to the statute and the case law, pursuant to s.14 Equality Act 2006, the Equality and Human Rights Commission (EHRC) issued a Statutory Code of Practice on Equal Pay (2011). Tribunals and courts considering an equal pay claim are obliged to take into account any part of the code that appears relevant to the proceedings. It provides the following guidance on RAE claims:

Job evaluation is a way of systematically assessing the relative value of different jobs. Work is rated as equivalent if the jobs have been assessed as scoring the same number of points and/or as falling within the same job evaluation grade. A small difference may or may not reflect a material difference in the value of the jobs, depending on the nature of the job evaluation exercise.

A job evaluation study will rate the demands made by jobs under headings such as skill, effort and decision-making. Because the focus is on the demands of the job rather than the nature of the job overall, jobs which may seem to be of a very different type may be rated as equivalent.

To be valid, a job evaluation study must:

If a job evaluation study has assessed the woman’s job as being of lower value than her male comparator’s job, then an equal value claim will fail unless the Employment Tribunal has reasonable grounds for suspecting that the evaluation was tainted by discrimination or was in some other way unreliable.

Job evaluation studies must be non-discriminatory and not influenced by gender stereotyping or assumptions about women’s and men’s work. There has historically been a tendency to undervalue or overlook qualities inherent in work traditionally undertaken by women (for example, caring). A scheme which results in different points being allocated to jobs because it values certain demands of work traditionally undertaken by women differently from demands of work traditionally undertaken by men would be discriminatory. Such a scheme will not prevent a woman claiming that her work would be rated as equivalent to that of a male comparator if the sex-specific values were removed.”

Absent a provision to the contrary, as a general rule in the civil arena it is for a claimant to prove their claim by establishing facts from the evidence before the court or tribunal on the balance of probabilities. Facts can be proved by direct evidence or the drawing of inferences. In direct discrimination cases, there is rarely direct evidence of less favourable treatment because of, or in the language of the early days, on grounds of, a protected characteristic. In King v Great Britain China Centre [1991] 10 WLUK 157; [1992] ICR 516 CA, it was held that there had been no deviation from the fundamental principle that the person bringing the claim has to prove it, by the practice sometimes followed in the employment tribunal which was that where a claimant succeeded in establishing a difference in treatment and a difference in protected characteristic, such as race or sex, a court or tribunal was entitled to look to a respondent for an explanation. If no explanation for the less favourable treatment was forthcoming, or if it was unsatisfactory, in the sense of not being a credible explanation, a tribunal was perfectly entitled to draw the inference that the reason for the treatment was the protected characteristic. This was not a reversal of the burden of proof but a proper balancing of the factors which could be placed in the scales for and against a finding of unlawful discrimination (Neill LJ @ 529).

In Handels-og Kontorfunktionaerernes Forbund i Danmark v Dansk Arbejsgiverforening(Danfoss) (Case 109/88) [1991] ICR 74 the Court of Justice of the European Union (CJEU) considered where the burden of proof lay in the context of the principle of equal pay in Council Directive 75/117/EEC which “means, for the same work or for work to which equal value is attributed, the elimination of all discrimination on grounds of sex with regard to all aspects and conditions of remuneration…” (art.1). The male and female employees were doing the same work or work of equal value and received the same basic pay, but different individual pay supplements that were not transparent in that the criteria used in the application of individual pay supplements could not be identified. A random survey revealed that male employees were paid on average 6% more than their female colleagues. CJEU held that

In 1997 in EU Council Directive 97/80/EC on the burden of proof in cases of discrimination based on sex (“the Burden of Proof Directive”), a shifting burden of proof was introduced in sex discrimination and equal pay claims which required member states to take measures

The provision remained unaltered in EU Directive 2006/54/EC on the implementation of the principle of equal opportunities and equal treatment of men and women in matters of employment and occupation (recast) dated 5 July 2006 (“the Re-cast Directive”) at Article 19(1).

The Burden of Proof Directive was enshrined in s.63A Sex Discrimination Act 1975 (inserted by the Sex Discrimination (Indirect Discrimination and Burden of Proof Regulations 2001)) which provided that:

The Equality Act 2010 was with very few notable exceptions, a consolidating act to bring all the disparate sources of equality legislation into one document and to simplify and modernise some of the language. The burden of proof provisions is contained in s.136:

Danfossheld that where there is a lack of transparency and women are paid less than men, the burden is on the employer to show that its pay practice is not discriminatory and followed the Advocate General’s opinion that “the employer has the burden of proving facts which are exclusively within his sphere of influence” [41]. But it is important to bear in mind that the question before the tribunal in this case was not whether the employer had breached the claimants’ equality clause, but whether there was a JES that applied to the claimants’ and their comparators’ jobs. Information about the 2014 Exercise was known only to the respondent – it had not been shared with the claimants or the recognised trade union who had played no part in it – it was the employer’s exercise. It was a preliminary issue at a preliminary hearing at an early stage of proceedings, considered in isolation from all of the issues which need to be addressed in order to decide if there has been a contravention of the Act.

Mr Jones KC for the Leigh Day claimants submitted that in light of the tribunal’s finding that the 2014 Exercise was a “study undertaken with a view to evaluating…the jobs to be done” [88] and that the study was “analytical” in the sense that it had evaluated the jobs in terms of the demands made on a person by reference to such factors as effort, skill and decision-making, it followed that it must be a JES as defined in s.80(5). In concluding that the 2014 Exercise was not a JES, the tribunal had applied an unwarranted gloss to the statutory test.

The only requirements of s.80(5) were that there be a study, implicitly requiring an element of thought and analysis, and that the purpose of the study be to evaluate jobs, to be done with a methodology of measuring or evaluating the demands made on the employees by reference to demand factors.The section is not prescriptive in terms of the choice of demand factors, beyond providing examples of possible factors of effort, skill, and decision-making. Since the section is silent on matters such as the intended use of the study, consultation or engagement with the employees or trade unions or agreement, completion, or adoption of the study, it follows that none of those matters are preconditions of a JES. A JES may even be sex-specific, and by implication discriminatory, yet still be a JES, although certain consequences follow in the case of such schemes (see s.65(4) and s.131 further discussed below).

Mr Jones submitted that the test was accurately described in Bromley v H & J Quick Ltd [1988] ICR 623 by Dillon LJ at 633B-D:

The grammatical meaning of s.80(5) supported the claimants’ interpretation, but, if necessary, Mr Jones also relied on the purpose and parliamentary intention of the Act which is to secure that employers give equal treatment as regards terms and conditions of employment to men and women performing equal work. Tesco’s interpretation of a JES would create a perverse incentive for employers to ensure that their studies were not rigorous and make it more difficult for claimants to establish equal work. Tesco’s interpretation was at odds with the expansive and less restrictive approach to the definition of “like work”.

The tribunal should have gone no further than their conclusion in paragraph 88 which was a complete answer to the question. It should have stopped there and if it had, the 2014 Exercise would have been found to be a JES.

In analysing the case law, Mr Jones relied on Nuttall and Bromley to submit that a JES must evaluate, analytically, the jobs of the claimants and comparators by reference to factors, rather than on a whole job basis, and therefore be thorough; and that the study must be impartial and objective (even if including some unavoidably subjective elements). However, the interpretation sought to be drawn from the authoritiesrelied on by Tesco – principally Diageo v Thomson EATS/0064/03 and Armstrong (HJB Claimants) v Glasgow City Council CSIH [2017] IRLR 993 - went beyond the statutory language and had been misinterpreted. It resulted in too high a threshold being applied. This tribunal was asked by Mr Jones to either clarify what is meant by the guidance in Eaton v Nuttall or confirm that tribunals should focus on the statutory language.

He also submitted that reliance on early case law on authority, completion and acceptance such as Arnold v Beecham Group [1982] ICR 744 EAT and also O’Brien v Sim-Chem Ltd [1980] ICR 573 and England v Bromley LBC [1978] ICR 1, were also misunderstood, misapplied or misplaced.

In short, the tribunal had misunderstood the statutory test and in its reasoning in paragraph 89 had placed an impermissible gloss on the statute and set the bar of what constituted a JES too high by applying unwarranted and additional criteria.

Mr Jones also submitted that the test in defensive, shield cases is more demanding than the test when a JES is used by a claimant as a sword, because of the different policies underpinning sword and shield cases and their asymmetrical treatment. The logic of his submission was that the authorities in shield cases were of less assistance in sword cases such as here.

Mr Jones submitted that the tribunal had correctly analysed the burden of proof provisions in s.136 EqA 2010 to conclude that it was for the respondent to prove to the civil standard that the 2014 Exercise was not a JES. He relied on Handels-og Kontorfunktionaerernes Forbund i Danmark v Dansk Arbejsgiverforening(Danfoss) (Case 109/88) [1991] ICR 74, in which the CJEU had held that where there is a lack of transparency and women are paid less than men, the burden is on the employer to show that its pay practice is not discriminatory. In his opinion, the Advocate-General had stated that the approach meant that “the employer has the burden of proving facts which are exclusively within his sphere of influence.” The reasoning in Danfoss applies to the JES question as a classic example of matters exclusively within the employer’s sphere of influence and knowledge. He argued that since s.69 expressly provides how the burden of proof applied to material factor defences, it follows that s.136(4) must be interpreted as applying the burden of proof to the other elements of an equal pay claim, such as equal work.

He went on to submit that although the tribunal had correctly found that the burden of proof had shifted, it had wrongly concluded that the respondent had discharged it, principally because the tribunal had misunderstood the statutory test and had erroneously given the respondent the benefit of the doubt where there was a lack of job information, or other gaps in the evidence.

Mr Bryant KC for the Harcus Sinclair claimants described his submissions as complementing the Leigh Day submissions which he adopted. There was nothing in s.80(5) and the definition of a JES relating to how thoroughly the evaluation was conducted; whether it was capable of impartial application; whether it was a complete or accepted study; whether there was authority for it to be fully fledged; or whether it contained fundamental errors. The tribunal had relied on Greene v Broxtowe [1977] ICR 241 in support of its conclusion that partly because the 2014 Exercise contained fundamental errors, it was not a JES. The conclusion was criticised in two respects – firstly the failure to identify the so-called fundamental errors and secondly because the reference to fundamental errors in the judgment of Kilner-Browne J sitting in the EAT in Greene v Broxtowe needed to be read in the context of the case as a whole in which he had found that:

Once the tribunal had found that the 2014 Exercise was analytical, the unavoidable and necessary conclusion was that the 2014 Exercise was a JES. He accepted that this tribunal was bound by Bromley.

Unlike Mr Jones, Mr Bryant KC considered that the same definition of a JES applied across the board, whether it was being used as a sword or a shield in any particular case. He also considered that the meaning of a JES should be gleaned only from the statutory wording, and to that extent Eaton v Nuttall had been wrongly decided. Requiring a study to be objective and analytical adds qualitative terms not found in the legislation. The fact that it had been repeated and adopted in many cases since meant nothing more than that the error had become set in stone. Frequent repetition does not make something correct.

Mr Bryant also adopted Mr Jones’ arguments on the burden of proof. Enough had been done to establish a prima facie case since it was accepted that at least some of the claimants’ roles were paid less than at least some of the comparator roles and there was a gender disparity in those performing the roles. There was an exercise of some sort that had assessed the roles on the basis of the demands on the workers and the respondent’s own staff thought it looked as though the study had rated the roles as equivalent or of equal value. Those factors were sufficient for the reverse burden provisions to “bite”.

Mr Bryant too concluded that even without the benefit of the burden of proof having shifted the claimants had done enough to prove that the 2014 Exercise was a JES.

For the respondent, Mr Epstein KC’s central submission was that save for the burden of proof point, the tribunal had made the right decision for the right reasons and there were no errors of law in its decision and its judgment which had correctly understood the facts and applied the correct law.

He argued that the claimants’ analysis of the decision did not bear close scrutiny. In paragraphs 88-89 the tribunal correctly identified that the 2014 Exercise was not sufficiently analytical to satisfy the statutory definition of a JES and that the tribunal had done no more that apply the test in Eaton v Nuttall which he relied on as iterating the natural meaning of s.80(5), not applying an impermissible gloss on the statute.

Whilst he did not formally concede that Armstrong was wrongly decided, he said that he did not need to defend it as the tribunal had not relied on it, nor need he address it.

The claimants’ submissions would, if adopted, produce absurd results enabling an acknowledged unreliable study to be deemed valid which would be contrary to the purpose of the RAE route to establishing equal work. If the claimants’ arguments were correct it would prevent an employer from undertaking any exploratory work and embarking on a process of pay and grading analysis, as to do so would risk a rated as equivalent claim on the basis of no more than an early-stage initial exercise. It would wrongly elevate such tentative first steps to the status of a JES.

Mr Epstein contended however that the tribunal had erred in its application of the burden of proof. As a matter of statutory interpretation s.136 can only apply so as to reverse the burden of proof if the tribunal could hold that there was a “contravention” of the Act, in the context of this case a breach by Tesco of the sex equality clause in the contract of employment. In order for a claimant to succeed in an RAE claim four matters must be established: (1) that she is engaged on work that is RAE, (2) that she can make a comparison with a named comparator, (3) that there is a less favourable difference in contractual entitlement to pay, and (4) that there is no material factor defence. The question of whether the 2014 Exercise was a JES could not have led the tribunal to find that there had been a contravention of the EqA 2010 since it was only the first hurdle to be cleared. The application of the burden of proof provisions would occur at a later stage of the proceedings. The case law prior to the introduction of s.136, applied first principles that claimants, as parties seeking to prove the allegation, have the burden of proving it, (see for example England v Bromley [4D] and Bromley v Quick [636F]) which remained good law notwithstanding the introduction of the burden of proof statutory provisions.

Analysis and conclusions

It is worth remembering first principles and the history of the development of the common law and prior to the introduction of statutory provisions set out above.King v Great Britain ChinaCentre was just one of a number of important cases at that time which also included May LJ in North West Thames Regional Health Authority v Noone [1988] ICR 813through to the later case of Anya v University of Oxford [2001] EWCA Civ 405. Neill LJ’s discussion in King of the proper balancing of the factors which could be placed in the scales for and against a finding of unlawful discrimination is drawn from the observation of May LJ in North West Thames Regional Health Authority v Noone [1988] ICR 813. He held that it was almost unnecessary and unhelpful to introduce the concept of a shifting evidential burden of proof when tribunals were adopting an almost common-sense approach. At the conclusion of all the evidence the tribunal should make findings as to the primary facts and draw such inferences as they consider proper from those facts. They should then reach a conclusion on the balance of probabilities, bearing in mind both the difficulties which face a person who complains of unlawful discrimination and the fact that it is for the complainant to prove his or her case.

The effect of the introduction of the statutory provisions merely mandated the practice that was previously being routinely adopted in tribunals.

However, the statutory regime (which is presaged in the antecedent case law) is clear that a shift in the burden of proof only occurs “if there are facts from which the court could decide, in the absence of any other explanation, that a person (A) contravened the provision concerned” (s.136(2) EqA 2010). In other words, if a prima facie case has been established. The respondent’s argument that the burden of proof cannot have yet shifted when this was a preliminary hearing dealing with a single issue in isolation of the several other matters claimants must prove to succeed in an equal pay claim, which were still in dispute, is a powerful one.

Mr Jones’ reliance on Danfoss does not assist since the claimants in that case had established a prima facie case: the work of the men and women was equal; the gender disparity had been established and no explanation for it had been forthcoming. Mr Jones’ emphasis on the observation in the Advocate General’s opinion in Danfoss that “the employer has the burden of proving facts which are exclusively within his sphere of influence” [41] is not to be read in isolation of all the evidential aspects of a case. It is important to bear in mind that the question before the tribunal in this case was not whether the employer had breached the claimants’ equality clause, but whether there was a JES that had rated as equivalent the jobs of the claimants and their comparators. The 2014 Exercise was certainly known only to the respondent – it had not been shared with the claimants or the recognised trade union who had played no part in it – it was the employer’s exercise. But this was a preliminary issue at a preliminary hearing at an early stage of proceedings, considered in isolation from all of the issues which need to be addressed in order to decide if there has been a contravention of the Act.

I agree with the respondent’s submission that s.136(3) only bites and shifts the burden of proof so that a court must hold that a contravention has occurred unless the respondent shows that they did not contravene the provision under s.136, when a prima facie case on all aspects of a claim has been established (either on the evidence or because the basic facts are not in dispute). That was not the case here. Much is still in dispute and it was therefore premature to consider that the burden of proof had formally shifted under s.136 at this preliminary hearing. It follows that the burden of proof had not shifted to the respondent under the statute.

However, my reading of the employment judge’s conclusions in paragraphs 86 and 87 of her judgment is that although she has referred to s.136 and the shifting burden of proof, her approach was more akin to the application of basic principles to findings of fact and the evaluation of evidence. She conducted a proper balancing of the factors which could be placed in the scales for and against an assessment of whether the 2014 Exercise was a JES. The claimants had to prove that the 2014 Exercise was a JES. The evidence established that an exercise was carried out, that various jobs were given various scores and a process of sorts was undertaken. On the basis of that undisputed evidence the employment judge was perfectly entitled to look to the respondent for evidence of why it was not a JES. She had made primary findings of fact where matters were in dispute after hearing the evidence and had reached a conclusion on the balance of probabilities, bearing in mind that the evidence about the 2014 Exercise came only from the employer.

But although the tribunal had expressed itself in the language of s.136, it was largely applying a common-sense approach to the evidence and its task of analysing whether the 2014 Exercise was a JES. If the tribunal had instead said that because the 2014 Exercise contained a number of features consistent with a JES and because the facts about the 2014 Exercise were exclusively within the knowledge of the respondent, it would scrutinise the respondent’s evidence with particular care, to test if the factors were inconsistent with a JES, there could have been no sensible objection. She did not need s.136 to conduct the perfectly proper exercise that she did. It was an error however for the tribunal to state that the burden of proof had shifted under the statute.

In any event the tribunal’s conclusion on the burden of proof had no effect on the outcome of the case. As is often the way in these cases, whether or not it had shifted was a distinction which made no difference to the decision.

Although the wording of EqA 2010 s.80(5) is not exactly, as in word for word, the same as the predecessor legislation, Equal Pay Act 1970 (EqPA 1970) s.1(5), the sections are materially identical and, as noted above, the EqA 2010 was a consolidating, not an amending act, save in respect of a number of limited specific areas, such as private members clubs. It did not intend to, and nor did it, change the law in relation to equal pay in force prior to the commencement date of the EqA 2010. Indeed, the parties had not taken me to the history of equal pay legislation to demonstrate that the EqA had changed the law, but to give a full picture of the development of the law and changes introduced in 1984 and 2004 to provide a better understanding of the context and intention of parliament to understand the legal meaning of s.80(5).

Mr Jones and Mr Bryant stressed the purpose of the EqPA 1970, set out at the start of the Act in s.1(1): “The provisions of this section shall have effect with a view to securing that employers give equal treatment as regards terms and conditions of employment to men and to women [performing like work or work rated as equivalent within the meaning of section 1(5)]”, in which they emphasised the words “with a view to securing that employers give equal treatment”. So far, so good, but central to the notion of equality in this context is the equal treatment of men and women doing equal work: the comparison between the jobs done by women and jobs done by men is at the heart of the legislation. If the women are not doing equal work, they are not entitled to be paid the same as the better paid men and there is no breach of the equality clause.

All parties agreed that the Eaton test in considering the meaning of a JES of “thorough in analysis and capable of impartial application” formulated in this tribunal in 1977 has been repeated in the subsequent case law, adopted in the statutory code of practice, and coined the term a “valid JES” as shorthand for the Eaton test. The use of the term “analytical” in the Eaton test was approved by the Court of Appeal in Bromley v Quick as a convenient indication of the general nature of what is required of a JES. It was not a gloss on what is now s.80(5) EqA 2010.

But in what has become an overlooked passage in Bromley v Quick the same cannot be said of the second limb of the test: “and capable of impartial application.” The Court of Appeal firstly noted that there may necessarily be subjective elements in an objective process and “capable of impartial application” must be read in the context of the inevitable involvement of some value judgments in the process (@632 C-E). Dillon LJ goes on to find that the question of whether discrimination is not, inadvertently, let in, falls to be considered not at thestage of considering whether the study is “such as is mentioned in s.1(5)”, but when a JES is being raised as a defence or shield to a claim under s.2A(2)(b) (@F p.632), now s.131(6). In other words the definition of a JES is the same whatever purpose it is being used for: that must be the case as the wording of s.1(5) did not differentiate between the two different potential uses of a JES and s.2A.2(2)(a) and s2A.2(3) expressly referred to “a study such as is mentioned in s.1(5) above” and EqA 2010 s.131(9) states that “job evaluation study” has the meaning given to it in s.80(5). But if a respondent seeks to rely on a JES that complies with s.80(5) (formerly s.1(5)) to show that the jobs have not been given an equal value under the JES in a bid to strike out the equal value claim, it is for the claimant to show reasonable grounds for suspecting that the evaluation contained in the study was flawed – either because of a taint of sex discrimination or because it was otherwise unreliable.

The distinction set out in Bromley v Quick between what constitutes a JES and the separate question of what a claimant has to show to dislodge a presumption that a JES ascribing different values to men’s and women’s work is sufficient to strike out an equal value claim, appears to have been lost in the cases that followed. In subsequent cases, such as DEFRA v Robertson the fallacy that the Court of Appeal had wholeheartedly approved the Eaton test in Bromley v Quick became accepted wisdom.

It is probably time to go back to Bromley v Quick which stressed the importance of the words of the statute without further gloss or additional requirement:

It was a powerful Court of Appeal. Attention is rightly focussed on the lead judgment of Dillon LJ, but equally pertinent are the judgments of Neill LJ and Woolf LJ (as he was then). Neill LJ did not depart from the wording of the statute in holding:

He was even relaxed about whether a detailed written job description for each of the individuals would be necessary (@638 C-D). Woolf LJ also found that it was only because the two claimants’ jobs had never been evaluated under various headings as required by s.1(5) that the employer’s scheme was not a JES.

Bromley v Quick trumps Eaton and the EAT is bound by the Court of Appeal. In light of Bromley v Quick the last limb of the Eaton test when considering if there is a JES under s.65(4) and s.80(5) under the EqA 2010 has shaky foundations. Bromley v Quick explicitly states that the test of “capable of impartial application” only becomes relevant if s.131(6) applies and the respondent is seeking to rely on a JES to show that the work of the claimant and the comparator is unequal. At the s.65(4) and s.80(5) stage, all that is required is there to be a study undertaken with a view to evaluating jobs in terms of the demand made on a worker under various headings, for instance, effort, skill and decision-making, as per the words of the statute.Bromley v Quick agreed only that implicit in the words of the statute is that the study be thorough and analytical. What seems to have happened over time is that the test of a JES under s.1(5) EqPA 1970 and s.80(5) EqA 2010 has become conflated with the test under s.131(6) which only applies when a respondent is using a JES as a shield.

Mr Epstein accepted that relevant Court of Appeal authority cannot be ignored by this tribunal and it must be followed, but that this tribunal should respect Armstrong, as a judgment of the Court of Session Inner House in Scotland. It goes much further than Eaton introducing the additional requirements for a study to comply with s.1(5): that it must be objective; transparent; accurate; internally sound; internally consistent; sufficiently detailed; and fair. There is a clear conflict between Bromley v Quick and Armstrong. Mr Epstein submitted that it was to be expected that the EAT would follow a Court of Session judgment. However, since there is Court of Appeal authority on point, I cannot accept that proposition. In Marshall Clay Products Ltd v Caulfield [2004] ICR 1502 CA Laws LJ [32] reiterated that it is the decisions of the Court of Appeal that bind the EAT, not CSIH and I am therefore bound by Bromley v Quick not Armstrong.

In any event, the list of requirements said to be necessary for a JES under s.1(5) EqPA 1970 set out in Armstrong was based on a concession and was not the subject of any argument or discussion (see [40]) and as such do not have precedential value: the Court merely accepted the agreed position of both parties. As stated in R(Kadhim) v Brent Housing Board [2001] QB 955 CA:

The Court of Session in Armstrong fell into exactly the error identified in Bromley v Quick and conflated the test under s.1(5) with the test under s.2A(2A) EqPA1970 (Footnote: ⁵) when setting out the agreed list of 9 adjectives to be applied to considering whether a scheme is a JES under s.1(5). What the respondent had to show to satisfy s.2A(2A)(b) EqPA1970 is a different question to whether the study the respondent was relying on was a JES under s.1(5). The Court of Session elided the two issues. (Footnote: ⁶)

Armstrong was a case in which the JES was being used as a shield by the respondent employer. It is therefore not necessary for me to decide if it should be followed in the determination of the question of what an employer has to show to a tribunal when advancing a s.131(6) EqA2010 defence to an equal value claim as it is not relevant to this appeal. I note however that it would appear to add a layer of complexity and tautology not contained in the statutory wording and some of the words in the list have similar meanings and risk resembling a thesaurus.

Since the second limb of the Eaton test and the phrase “thorough and capable of impartial application” is not in the statute and was rejected by Bromley v Quick, this tribunal must follow Bromley v Quick. Where does that leave 45 years of case law that has found the Eaton test to be helpful and easy to apply and which has found its way into the statutory code of practice? Mr Epstein submitted that it would produce absurd results to jettison the last limb of Eaton and would enable a claimant to rely on a JES that was tainted by sex discrimination which would defeat the purpose of the legislation. Mr Epstein conceded that it would or might be possible for a respondent to raise the fact that a JES is unsatisfactory in some way as a material factor defence to defeat an RAE claim, since there was nothing in the language of statutory defence or the case law to suggest otherwise, but it is uncharted territory. He submitted that it would at best involve an additional, and ultimately pointless step in the already tortuous process and at worst be illogical.

When a tribunal is considering whether “a study evaluates jobs in terms of the demand made on a worker under various headings, for instance, effort, skill and decision-making”, it will focus on the need for the exercise to be a study, which denotes a degree of detail and rigour and evaluation, which is also self-explanatory. The tribunal’s consideration of the scheme being put forward as a JES will no doubt assess how well such a study is capable of being applied impartially: if the study is not capable of evaluating the jobs by reference to the demands on the workers, it will not be capable of being applied impartially. But it does not require extra words or an additional limb of the statutory test to do so. So, whilst Bromley v Quick must be followed and the second limb of the Eaton test should not have remained such common currency, it is unlikely to affect the way that tribunals undertake their fact-finding task when considering whether a study is a JES.So, whilst tribunals no doubt routinely consider studies relied on as a JES with care and scrutiny in their fact-finding exercise, it is unhelpful to add words to the statute. The absurd results feared by Mr Epstein are avoided by a common-sense approach and natural reading of the statute as by simply applying the words of the statute and assessing if a scheme is thorough in analysis, tribunals are well able to decide if a scheme relied on is a JES as defined by s.80(5), as this tribunal did. Only where a JES is relied on by a respondent to strike out an equal value claim under s.131(6) must the tribunal consider if the study is tainted by sex discrimination or is otherwise unsuitable to be relied on. If a study is tainted by sex discrimination it may well be that it also means that it has not given the job in question a value by reference to the demand made on a worker under various headings, but that is as a consequence of applying s.80(5).

I wonder if the problem stems from a tendency, over the years, to over complicate and micro-manage what is a basic fact-finding exercise culminating in Armstrong. It is informative to return to the earlier cases to see how the law has developed from the initial approach of Greene v Broxtowe District Council in 1977 and the confidence then expressed in the ability of the tribunal to decide for itself how far it can go in examining the validity of the evaluation study, to Armstrong. With now nearly 50 years of experience in examining schemes relied on as a JES by both claimants and respondents’ tribunals have shown themselves to be and are well able to understand and apply the words of the statute.

For understandable reasons the tribunal followed the Eaton orthodoxy which, as I have set out above, is not quite the straitjacket that it has been made out to be, even though, it makes little, if any practical difference in how tribunals approach their task. The employment judge did not have the luxury of being taken through the entire canon of the case law on s1(5) in the depth and detail to which I have been. It follows that the tribunal has over-stated its task in its direction at paragraph 89 of its judgment. The real question is whether, notwithstanding its direction, the tribunal merely applied the words of the statute, or imposed impermissible additional requirements and raised the bar too high for the claimants, which would call into question its conclusion that the 2014 Exercise was not a JES.

The answer is found in the tribunal’s findings of fact and conclusion at [90]. The tribunal found that although the study was analytical, the factors chosen, “did not cover the demands made on the job holders. The omission of, in particular, any factor for physical efforts or skills where the jobs certainly demand those features was a serious omission.”

That finding alone is fatal to the claimants’ appeal.The employment judge found that it was not a JES because it did not evaluate the demands made on a person, as required by the definition in s80(5) since physical efforts or skills were both omitted which the judge considered to be serious omissions. It is a fundamental error of the type identified in Green v Broxtowe DC that proves that the jobs were not evaluated in terms of the demands made as required by the plain words of s.80(5) EqA 2010.

The argument advanced by both the Leigh Day and Harcus Sinclair claimants was that the tribunal applied additional requirements of authority, completion and acceptance gleaned from the early case law which do not form part of the statutory language in s.80(5) EqA 2010. The finding was made in the alternative by the employment judge [95 – 97] and was not part of the ratio of the decision. I need only deal with one aspect.

As always, it is worth returning to the wording of the statute which requires there to be “a study undertaken” with a view etc.(s.80(5)). The use of the past tense implies a level of completion and finality – a work in progresswill be unfinished and will not have been “undertaken.” Ignoring the issues of authority and acceptance for a moment, it is no more than common sense that the study must have been complete, in the sense of an evaluation having been made, in order for it to have results. Whether something is sufficiently complete to amount to a study undertaken is a finding of fact for the tribunal. In some cases, there may be grey areas that require careful judgment and fact finding by the tribunal. But on the facts in this case the tribunal had little difficulty in concluding that the 2014 Exercise was still a work in progress and thereforenot a JES. It was “an exploratory exercise only” [95] and was “not complete in any meaningful sense of the word.” [96]. The point is well illustrated by the initial work done comparing the work of a customer assistant and fishmonger which produced different and inconsistent results which were unresolved because the study was incomplete, in the sense of not having been undertaken but still being a work in progress. It was thus too early and premature to say which of the contradictory results was the evaluation under the study. It would be like an early draft of a judgment being elevated to the status of something more fully worked through, which is an anxious making thought. In any event, it was a purely factual exercise by the tribunal and the challenge reaches nowhere near the high test of perversity. Furthermore, the decision was demonstrably right.

It would, as Mr Epstein said, produce absurd results if preparatory work or ideas were elevated to the status of a JES. For the purpose of deciding the issues in this appeal it is not necessary to examine the case law on authority and acceptance since the tribunal made an unchallengeable finding that the 2014 Exercise was not a JES as it was incomplete.

The judge’s finding that the 2014 Exercise was not sufficiently worked up to be complete in the sense of being a study undertaken would itself be a sufficient finding to support her conclusion that there was no JES under s.80(5). So even if I am wrong in my primary findings, the employment judge’s alternative finding at paragraphs 96 and 97 is fatal to the claimants’ appeal.

Definition of Equal Pay

History of the introduction of the use of a JES as a shield or defence to an equal value claim following the equal value amendment