England and Wales Court of Appeal (Civil Division) Decisions

IN THE COURT OF APPEAL (CIVIL DIVISION)
ON APPEAL FROM The Employment Appeal Tribunal
Mr Recorder Luba QC and 2 lay members

		Royal Courts of Justice Strand, London, WC2A 2LL
		09/12/2015

Lord Justice Underhill:

1. This appeal concerns alleged discrimination in the remuneration of chaplains employed in HM Prison Service^[1]. By section 7 of the Prison Act 1952 there is a statutory obligation on the Secretary of State to appoint "a clergyman of the Church of England" as a chaplain at every prison in England. By section 10 he is empowered also, where "the number of prisoners who belong to a religious denomination other than the Church of England is such as in the opinion of the Secretary of State to require the appointment of a minister of that denomination", to appoint such a minister to that prison. Some chaplains are employed (full-time or part-time); others are engaged on a sessional basis. As at November 2010 there were 353 employed chaplains in the Prison Service, of whom 109 were Church of England, 112 Christian of other denominations, and 89 Muslim.
2. The Claimant in these proceedings, the Appellant before us, is an imam. He is of Pakistani origin. In June 2001 he was appointed a chaplain at HMP Bullingdon. He initially worked on a sessional basis but with effect from 25 October 2004 he started employment as a full-time chaplain. He and the Church of England chaplain have at all material times been the only employed chaplains at Bullingdon, though a number of others have worked on a sessional basis.^[2]
3. As is usual in the public sector, employees in the Prison Service are assigned to "pay bands" on an incremental scale, which determine their basic pay. Employees can expect to progress up the scale, and thus receive increased earnings, year-by-year. There has for many years been a performance element in the rate of progression. The details are irrelevant and have changed over the period with which we are concerned: all that matters for our purposes is that basic pay will go up each year provided the employee receives at least an "achieved" marking on their annual appraisal. Increases as a result of progress up the scale are distinct from any increase on account of increases in the cost of living.
4. When he was first employed in 2004 the Claimant was assigned to "Chaplaincy Payband 1", to which I will for convenience refer simply as "the chaplaincy payband"^[3]. At that time it comprised 101 "spinal points". The Claimant started at the bottom of the band. On the basis of an "achieved" assessment, progression was typically at a rate of six points per year, so that it would take him no fewer than seventeen years to progress from the bottom of the scale to the top. However, the Prison Service has for some time recognised that many or most of its pay-bands are too long, and it has been its target gradually to reach a point at which movement from the bottom of a band to the top would take no more than six years. In accordance with that policy, accelerated progression applied to the chaplaincy payband in the four years 2007/8 to 2011/12, and the lower points on the scale were deleted; but the process was then stalled as a result of public sector pay restraint. The result is that at the date that these proceedings were commenced, on 21 April 2011, it would take nine years (again, assuming an "achieved" assessment each year) to get from the bottom of the band to the top. The Claimant was less than half-way up, earning a basic salary of £31,606 p.a. Chaplains at the top of the band were earning £36,703. (I should say that total gross remuneration was substantially higher – in the Claimant's case totalling £39,735 – as a result of his entitlement to a "Chaplain Required Hours Allowance" and a "Local Pay Allowance". But this claim is concerned with basic pay.)
5. The starting-point for the Claimant's complaint is the fact that the average basic pay of Muslim chaplains on the chaplaincy payband is lower than the average basic pay of Christian chaplains. This was established from figures supplied by the Prison Service in response to a questionnaire and was not in issue before the ET. As at 1 April 2011 the average basic pay for Muslim chaplains was £31,847 whereas for Christian^[4] chaplains it was £33,811.
6. It is common ground that the higher average pay of Christian as compared with Muslim chaplains is the result of the fact that the Prison Service only began to employ Muslim chaplains in 2002. Before that date they were engaged, as the Claimant was between 2001 and 2004, only on a sessional basis. The result is that whereas there were in 2011 a substantial number of Christian chaplains who started employment before 2002, all employed Muslim chaplains will have started employment on or after that date. It necessarily follows that the average length of service of Christian chaplains as a group is greater than the average of length of service of Muslim chaplains as a group – the figures are 5.76 and 9.43 years respectively; and the operation of the incremental pay system means that that produces a differential in pay.
7. I should at this stage set out one important finding made by the ET which goes beyond the purely factual. In her submissions the Claimant's then counsel, Ms Catrin Lewis, had described the fact that no Muslim chaplains had been employed prior to 2002 as "a religion bar". The tribunal did not accept that characterisation. It said, at para. 20.4 of its Reasons:

"We think that that phrase was an unfortunate one, which we cannot endorse in fairness to the Respondent. It implies an arbitrary and potentially directly discriminatory decision to exclude Muslims from employment. That was not the case advanced by the Claimant, and we find as a fact that it was not what happened. As the data referred to above indicate, there were no Muslim Chaplains in employment at a time when there appeared objectively to be insufficient numbers of Muslim prisoners to warrant the employment of Muslim Chaplains, as opposed to their engagement on a sessional basis."

In short, the ET found (a) that until 2002 the number of Muslim prisoners was not such that there was judged to be a need for employed Muslim chaplains and (b) that that judgment was "objective". On that basis it held that the non-employment of Muslim chaplains prior to 2002 was not discriminatory: if that is not quite explicit in the passage quoted, it is clear from para. 20.7 (quoted in full at para. 31 below), where the tribunal referred to having found "that the reason for Muslims' late joining of service was not of itself discriminatory or for an unlawful reason".

8. Those findings reflected the evidence of Ms Michelle Crerar, the Policy and Support Manager within Chaplaincy HQ at NOMS, who made a witness statement and also gave oral evidence. She said in her witness statement that "the change to enable Muslim Chaplains to be employed by the Service was part of a programme of work to meet the needs of the growing Muslim prisoner population": the number of Muslim prisoners had grown from 2,745 in 1995 to 5,046 in 2001 and 7,216 in 2005 (and indeed – though this may be strictly immaterial – to 11,261 in 2011). She also referred to "the combating extremism agenda".
9. Mr Sean Jones QC, who appeared for the Claimant before us, made it clear that he did not seek to challenge the ET's findings as to the reason why no Muslim chaplains were employed before 2002 or its decision that that reason was not discriminatory. He would in any event have had difficulty doing so given that, as the ET records, no case that the pre-2002 practice was discriminatory had been advanced.^[5]
10. Since Christian chaplains are predominantly (though not exclusively) white and Muslim chaplains predominantly (though not exclusively) non-white, there is also a disparity between the average basic pay of white and non-white chaplains on the chaplaincy payband; but, for reasons which will appear, I need not set out the figures.
11. Against that background, the Claimant's complaint in these proceedings was summarised by the Employment Judge, with the agreement of both parties, as follows:

"The Respondent has a long pay ladder with slow progression up the ladder. Most entrants start at the bottom rung. Muslim Chaplains have been unable to join the ladder before 2002. Anglican Chaplains have been able to do so for many years. Until the Claimant reaches the top of the ladder, it is likely that he will be paid less than his Christian comparators, because they were able to start on the ladder long before him, and pay relates predominantly to time spent on the ladder."

The Claimant says that that inequality entitles him to claim for indirect religious, alternatively racial, discrimination. There is no claim of direct discrimination.

12. The relevant provisions of the Equality Act 2010 are as follows:

(1) Section 4 defines the protected characteristics, which include race and religion or belief. Those characteristics are the subject of more detailed provisions in sections 9 and 10 respectively, but there is nothing in those provisions material for present purposes.

(2) Section 13 defines direct discrimination. I need only set out sub-section (1), which reads:

"A person (A) discriminates against another (B) if, because of a protected characteristic, A treats B less favourably than A treats or would treat others."

(3) Section 14 deals with "combined discrimination". I mention it only for completeness, because it is referred to in section 23 (1), to which I refer below. It is not yet in force.

(4) Section 19 defines indirect discrimination. It reads, so far as material:

"(1) A person (A) discriminates against another (B) if A applies to B a provision, criterion or practice which is discriminatory in relation to a relevant protected characteristic of B's.

(2) For the purposes of subsection (1), a provision, criterion or practice is discriminatory in relation to a relevant protected characteristic of B's if —

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

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

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

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

(3) The relevant protected characteristics are —

…

race

religion or belief

… "

I will for convenience use the shorthand "justification" for the requirement covered by head (d) under sub-section (2), and I will refer to heads (a)-(c) as constituting "prima facie indirect discrimination".

(5) Section 23 (1) reads:

"(1) On a comparison of cases for the purposes of section 13, 14, or 19 there must be no material difference between the circumstances relating to each case.

(2)-(4) ..."

(6) Section 39 (1) outlaws discrimination by an employer against an employee as to (inter alia) his terms of employment.

Those provisions reflect the terms of EU Council Directives 2000/43/EC ("the Race Directive") and 2000/78/EC ("the Framework Directive"), which they are intended to implement. I need not set out the relevant articles here.

13. The claim was heard by an employment tribunal sitting at Reading, chaired by Employment Judge Lewis, in April 2012. The case was conducted on the basis, accepted by both parties, that the claim of racial discrimination could not succeed if the claim of religious discrimination did not, and the focus was accordingly entirely on the latter. By a reserved judgment promulgated on 12 July 2012, the tribunal held that the Claimant had established prima facie indirect discrimination against him as a Muslim but that the claim failed because the Prison Service had established justification.
14. The Claimant's appeal was heard by the Employment Appeal Tribunal, chaired by Mr Recorder Luba QC, in November 2013. By a judgment handed down on 15 January 2014 the appeal was dismissed. However, the EAT's reasoning differed from that of the ET. It held that there was, on the ET's findings, no prima facie indirect discrimination, although it went on to say that if there had been it would not have upheld the ET's finding of justification.
15. On the appeal before us the primary case of Mr Tom Linden QC, for the Respondent, was that the EAT was right to find that there was no prima facie indirect discrimination (although he advanced an alternative analysis leading to that result); but it was also his case that if such discrimination were established the ET had been right to find that it was justified.
16. We heard the submissions of both parties on the issue of prima facie indirect discrimination first, and at the conclusion of the argument we announced our decision that the appeal would be dismissed, with the result that it was unnecessary to consider the issue of justification. My reasons for reaching that conclusion are as follows.
17. The issue raised by this appeal is essentially one of legal analysis, and I do not need to set out the facts in any more detail than I have already done. Nor will I at this stage consider the reasoning of the ET and the EAT but will go straight to the analysis under section 19 and the parties' submissions.
18. The starting-point must be to identify the "provision criterion or practice" within the meaning of section 19 (2) – in the jargon, "the PCP" – which is complained of. The PCP pleaded by the Claimant, in response to a request from the Respondent, was:

"that within the pay system applicable to those staff in the Chaplain payband 1 grade, in order to be at the top of the relevant pay scale an individual needs to have been employed by the Respondent for a certain period of time."

At para. 20.2 of the Reasons the ET identified the PCP, in slightly different terms but to essentially the same effect, as "the … pay system, which incorporates pay progression over time, which in turn is related to length of service". Both formulations focus on length of service. Mr Linden said that that was over-simple. He submitted that the PCP should be defined as the totality of the system for establishing basic pay, thus including both the performance element and the various changes to the range designed to reduce its length (see para. 4 above). I do not agree. The relevant PCP is the feature of the system which is said to have a discriminatory effect, and that is plainly the relationship between basic pay and length of service. The features referred to by Mr Linden are not said to, and plainly do not, put Muslims at a particular disadvantage: there is, for example, no reason to suppose that their performance assessments differ in any way from those of Christian chaplains. What we are concerned with in this case is the effect of a length of service criterion.

19. It is common ground that the PCP so defined is applied equally to Muslim and Christian chaplains, so that section 19 (2) (a) is satisfied. The question then is whether it puts Muslim chaplains "at a particular disadvantage" within the meaning of section 19 (2) (b).
20. As to that, it was Mr Linden's case, advanced by way of Respondent's Notice, that there is nothing in the length of service criterion that inherently disadvantages Muslim chaplains. That criterion allows them to progress up the incremental scale in exactly the same way as their Christian colleagues; and the Claimant expressly acknowledged, as the ET recorded, that a Christian chaplain who started employment on the same day as he did would (assuming equivalent performance) be paid the same as him. The only reason that Muslim chaplains have, on average, shorter service than their Christian colleagues is that none were employed earlier than 2002. That fact did not form part of the PCP complained of, and it did not in any event, as the ET held, constitute unlawful discrimination: see paras. 7-9 above.
21. Mr Jones's case as initially formulated in his skeleton argument placed considerable reliance on the decision of the EAT in Essop v Home Office [2014] UKEAT 480/13, [2014] ICR 871, which was decided after the decisions of both tribunals in this case. However, the wind was taken out of those particular sails by the recent reversal of that decision by this Court – see [2015] EWCA Civ 609, [2015] ICR 1063. I will have to say something more about Essop in due course, but I will focus at this stage on how Mr Jones put his case in his oral submissions. He submitted that it was essential to consider the effect of the PCP not simply in the abstract but in its effect on the actual population to which it was applied. He drew an analogy with the case of Ministry of Defence v DeBique [2010] IRLR 471, where the claimant was held by the EAT to have been unlawfully discriminated against when she suffered a disadvantage as a result of the combination of her position as a single mother and her non-British nationality. In the present case, a length of service criterion was being applied to a population in which by reason of the pre-2002 policy or practice Muslim chaplains had, on average, been employed for shorter periods; and that meant that as a class they were put at a particular disadvantage. The reason for the pre-2002 practice, and in particular the fact that it had been held not to be discriminatory, was immaterial at this stage: it was enough that it was particular to Muslims. In so far as it was relevant, it could only be to the issue of justification.
22. I can accept that submission in part. Specifically, I agree that it is necessary to consider the impact of the length of service criterion on the actual population to which it is applied. However, I do not agree that if it is shown, as it is here, that the use of the criterion leads to a disparity in the pay of Muslim and Christian chaplains, the enquiry under section 19 (2) (b) must end at that point. In my view it was and is open to the Respondent to go behind the bare fact that Muslim and Christian chaplains have different lengths of service and seek to establish the reason why that was so. What has been established in this case – indeed it was never in dispute – is that the reason for the difference is that there was no need for employed Muslim chaplains prior to 2002^[6]. That being so, I do not believe that it can properly be said that it is the use of the length of service criterion which puts Muslim chaplains at a disadvantage, within the meaning of section 19 (2) (b). The concept of "putting" persons at a disadvantage is causal, and, as in any legal analysis of causation, it is necessary to distinguish the legally relevant cause or causes from other factors in the situation. In my view the only material cause of the disparity in remuneration relied on by the Claimant is the (on average) more recent start-dates of the Muslim chaplains. But that does not reflect any characteristic peculiar to them as Muslims: rather, it reflects the fact that there was no need for their services (as employees) at any earlier date.
23. I do not need to express a definitive view about what the position would be if the non-recruitment of Muslim chaplains before 2002 were itself the product of discrimination (e.g. because the Prison Service assessed "need" according to some different criterion in the case of Muslims), since the ET has held that that was not so. Mr Linden submitted that even in such a case the Claimant would be obliged to complain about that discrimination in its own right – i.e. about the policy of not recruiting Muslim chaplains prior to 2002 – and to prove that his own recruitment had been delayed as a result. I am not so sure. I am inclined to think that in such a case it would be right to regard the shorter service of Muslim chaplains as a characteristic of the population to which the length of service criterion was applied since it was related to their religion in a different sense from the relationship in the present case. But the point is not quite straightforward and can be left to be decided in a case where it arises.
24. It may be instructive to compare the position of the Muslim chaplains in this case with that of the claimants in the litigation about the use of a length of service criterion by the Health and Safety Executive in determining the pay of its inspectors – see Cadman v Health and Safety Executive [2004] EWCA Civ 1317, [2005] ICR 1546, (CA) and [2006] ICR 1623 (CJEU) and Wilson v Health and Safety Executive [2009] EWCA Civ 1074, [2010] ICR 302. In that litigation female inspectors claimed under the Equal Pay Act 1970 on the basis that their pay was, on average, lower than that of their male colleagues because they had, on average, shorter service. To that extent the claims parallel the claim in these proceedings. But the difference is that it was acknowledged in those cases that a length of service criterion had an inherent tendency to put women at a disadvantage because women are liable to start their careers later than men and/or to take career breaks because of family and childcare responsibilities: see, e.g., para. 2 of the judgment of Arden LJ in Wilson (p. 305D). Accordingly the use of the criterion had to be justified. If the employer had been able to establish that the only reason for the disparity in average lengths of service had been that the proportion of women being recruited had increased in recent years – say, as a result of a changes in social attitudes, so that women were more willing to contemplate working in an industrial environment – the analysis would have been different.^[7] The essential point about the present case is that the shorter average length of service of Muslim chaplains can be shown to be attributable to a factor – that is, the change in the need for their services as the proportion of Muslim prisoners grew – which does not operate to the disadvantage of Muslims.
25. This seems to me to the right result in principle. It is not always straightforward to prove that a particular pay practice is justified, and indeed some practices may prove on examination to be to a greater or lesser extent unfair: very lengthy incremental pay scales may be a good example. But the correction of unfair pay practices is not the business of the 2010 Act and respondents ought not to have to defend them unless there is in truth an element of discrimination. That point was insisted on in two well-known decisions of the House of Lords – Strathclyde Regional Council v Wallace [1998] ICR 205 and Glasgow City Council v Marshall [2000] ICR 196 – concerned with the construction of the Equal Pay Act. Although the language of section 1 of that Act – and in particular sub-section (3), which gives the employer a defence when he can show that a pay differential "is genuinely due to a material factor which is not the difference of sex" – pre-dates the development of modern discrimination law, it is now well-established that it must be interpreted in the same sense as that of the familiar, and EU-derived, terms of the 2010 Act and its predecessors. In Wallace Lord Browne-Wilkinson said, at p. 213 B-D:

"… in considering section 1(3) of the Equal Pay Act 1970, the only circumstances in which questions of 'justification' can arise are those in which the employer is relying on a factor which is sexually discriminatory. There is no question of the employer having to 'justify' … all disparities of pay. Provided that there is no element of sexual discrimination, the employer establishes a subsection (3) defence by identifying the factors which he alleges have caused the disparity, proving that those factors are genuine and proving further that they were causally relevant to the disparity in pay complained of."

Likewise, in Marshall Lord Nicholls said, at p. 203 B-C, "if the employer proves the absence of sex discrimination he is not obliged to justify the pay disparity".

26. The analysis in the preceding paragraphs seems to me consistent with two lines of authority concerned with a similar, indeed probably the same, issue, albeit in a different context. I take them in turn.
27. First, I would refer to the line of authority which begins with the decision of this Court in Armstrong v Newcastle upon Tyne NHS Hospitals Trust [2005] EWCA Civ 1608, [2006] IRLR 124, though ultimately it derives from the decisions of the House of Lords in Wallace and Marshall to which I have already referred. The development of the cases can be summarised as follows (ignoring some twists and turns that are immaterial for our purposes):

(1) Armstrong was one of the many equal pay claims brought in recent years by employees in local government and the NHS in jobs done mainly by women who complained that they had not had access to bonus schemes made available for jobs done predominantly by men. There was an issue whether it was a defence for the employer to show that the differential was due to factors other than the difference of sex or whether they had also to show that it was objectively justified. This Court held that, as Buxton LJ put it at para. 110 of his judgment (p. 139), "if the employer proves the absence of sex discrimination he is not obliged to justify the pay disparity".

(2) The decision in Armstrong was at first controversial, and some uncertainty was expressed about its implications in a case explicitly based on indirect discrimination. The argument was complicated by the fact that the language of the 1970 Act is idiosyncratic, even though – as already noted – it is clear that it has to be read so as to conform with the language of the EU Directives and the other domestic discrimination legislation. There was also a question whether it was compatible with the well-known decision of the ECJ in Enderby v Frenchay Heath Authority [1994] ICR 112. But in Middlesbrough Borough Council v Surtees [2007] ICR 1644, which was another bonus case, the EAT analysed the issue in depth and concluded not only that what was said in Armstrong applied to cases of indirect discrimination but that it was right in principle and consistent with Enderby – see paras. 24-55 (pp. 1652-1660). As Elias P put it at para. 46 of his judgment (p. 1658 F-G):

"… it ought in principle to be open to an employer to show that even though there is disparate adverse impact, that it is not in any way related to any act of the employer which is sex tainted, and thereby avoid the need to establish justification".

(3) When Surtees was appealed to this court – [2000] EWCA Civ 885, [2009] ICR 133 (sub nom Redcar and Cleveland Borough Council v Bainbridge) – it was unnecessary for "the Armstrong point" to be decided, although Mummery LJ, delivering the judgment of the court, was inclined to think that Elias P's analysis was correct (see para. 60 (p. 157C)). But the point was put beyond doubt at this level in Gibson v Sheffield City Council, [2010] EWCA Civ 63, [2010] ICR 708, in which both Smith and Maurice Kay LJJ expressed the clear view that Armstrong was both binding and correctly decided (though Pill LJ disagreed) – see at paras. 59-72 (pp. 726-731) and para. 74 (p. 731) respectively. At para. 66 (p. 728 D-E) Smith LJ said:

"My conclusion is that whether the alleged indirect discrimination arises in the field of pay or non-pay, it is always open to a defendant to demonstrate that, notwithstanding the appearance that the practice puts women at a particular disadvantage, in fact the apparent disadvantage has arisen due to factors which are wholly unrelated to gender."

The reference to "pay or non-pay" is important because it shows that Smith LJ was equating the position under the 1970 Act with that under the Sex Discrimination Act 1975 and thus with the language used in the discrimination legislation generally. That is made quite explicit in the preceding reasoning, where she says, at para. 63 (p. 727 F-G) that where a claimant in a "non-pay claim" proved, in accordance with section 1 (2) (b) of the 1975 Act (which corresponds to section 19 (2) of the 2010 Act), that a PCP had put women at a particular disadvantage it remained open to the respondent "to demonstrate that what might appear to be a disadvantage to women arises from factors wholly unrelated to gender".

(4) I came to a similar conclusion, sitting in the EAT, in a further round of the Armstrong litigation, Newcastle upon Tyne NHS Hospitals Trust v Armstrong [2010] ICR 674 – see at paras. 66-73 (pp. 704-5).

28. The point established by the Armstrong line of cases – that is, that an employer can rebut a claim of indirect discrimination by showing that an apparent disparate impact is the result of non-discriminatory factors – is in line with my own reasoning at para. 22 above. As Smith LJ makes clear in Gibson, there can be no difference between the approach in an equal pay case and in a claim under the 2010 Act.
29. Secondly, I return to the decision of this Court in Essop. In that case a group of black and ethnic minority ("BME") candidates aged over 35 who had failed the "core skills assessment" necessary to achieve promotion in the Civil Service brought proceedings under the 2010 Act claiming that the requirement to pass the assessment was indirectly discriminatory. There was statistical evidence that BME candidates aged over 35 had a markedly lower pass-rate than younger white candidates. An ET held that those statistics alone did not establish prima facie indirect discrimination against any such candidates who failed the test, and that it was necessary for the claimant in each case to show what was the reason why he or she failed the test. The EAT held that that was not so. At para. 24 of his judgment (p. 879 D-E) Langstaff P pointed out that the wording of section 19 did not require members of a disadvantaged group to show why they had suffered the disadvantage alleged, but only the fact that they had done so. The leading judgment of this Court was given by Sir Colin Rimer, with whom the Chancellor and Lewison LJ agreed. He said, at paras. 57-59 (pp. 1081-2):

"57. ... I respectfully disagree with Langstaff J and [counsel for the claimant] that it is not necessary in indirect discrimination claims for the claimant to show why the PCP has disadvantaged the group and the individual claimant.

58. The legislation shows that in a direct discrimination claim the claimant must prove that the respondent's discriminatory treatment of him was because of, or on the grounds of, his protected characteristic. He must therefore, if he is to succeed, show that to be the answer to 'the reason why' question that arises in such claims. In indirect discrimination claims, there is also a necessary 'reason why' question but it is of a different nature. It does not go to the employer's motive or intention, whether conscious or unconscious. It is as to why the PCP disadvantages the group sharing the protected characteristic.

59. Langstaff J was correct to note, at [24], that 'section 19 … does not in terms require members of a disadvantaged group to show why they have suffered the disadvantage, in addition to the fact that they have done so'. But that, perhaps somewhat literal, interpretation of the statutory language overlooks that it is conceptually impossible to prove a group disadvantage for the purpose of section 19(2)(b) without also showing why the claimed disadvantage is said to arise. Group disadvantage cannot be proved in the abstract. Its proof necessarily requires a demonstration of why the comparative exercise inherent in the section 19(2)(b) inquiry results in the claimed disadvantage. ..."

Sir Colin went on at para. 60 (p. 1082) to cite a number of authorities in support of that analysis. It will be noted that although the actual issue raised by the decision of the ET related to the reason why the individual claimant failed the test – that is, the section 19 (2) (c) question – his reasoning expressly extends to the question of group disadvantage under section 19 (2) (b).

30. That passage clearly establishes that in considering whether a claimant can rely on a disparity in the treatment of two groups for the purpose of section 19 (2) (b) it is permissible to consider the reason for the disparity complained of, in the sense of the factors which caused it to occur. That is contrary to Mr Jones's case – as is indeed highlighted by the fact which I have already mentioned that in his skeleton argument he relied heavily on the reasoning of Langstaff P which has now been disapproved. (I should note that the passage quoted might in isolation be taken as suggesting that the burden of proving the reason is in all cases on the claimant; but it is clear from the later reasoning in his judgment that that was not Sir Colin's view – see in particular paras. 63-65 and the reference to the provisions of section 136 of the 2010 Act (p. 1083).)
31. Having set out what I believe is the correct analysis, I should show how the ET departed from it. That is in fact straightforward. At para. 20.7 of its Reasons it said:

"If we ask, what is the reason why all Muslims cannot achieve the same length of service as some Christians, one answer is that Muslims could not join the service before 2002. We find that it is not relevant that the reason for Muslims' late joining of service was not of itself discriminatory or for an unlawful reason."

A little later, at para. 20.14, it accepted a submission by Ms Lewis that "the Claimant need not show why the PCP put him at a disadvantage, but whether it does". That is precisely the argument which I have rejected above. (I should perhaps say at this point that although I disagree with the ET's conclusion, the care and clarity with which it analysed the case is impressive.)

32. I turn to the judgment of the EAT. Its reasoning was based on section 23 of the 2010 Act (see para. 12 (5) above). It held that there was no prima facie indirect discrimination because there was "a material difference in the circumstances relating to" those chaplains employed prior to 2002, namely that at that date no Muslim chaplains were employed; and that accordingly those chaplains should be excluded from "the pool" when making the comparison inherent in the exercise under section 19 (2).
33. It is not surprising that the route via section 23 should lead to the same result as my own approach, which focuses on section 19 (2) itself, since section 23 does not represent a free-standing ingredient in the definition of discrimination. Its purpose is simply to emphasise, what would in my view be necessarily implicit in sections 13 and 19 in any event, namely that like must be compared with like (cf. the observation to the same effect by Dillon LJ about the predecessor provision in the 1975 Act in Bain v Bowles [1991] IRLR 356, at para. 16 (p. 358)). But I do not, with respect, think that section 23 affords the best route to the right result. It tends to bring in an elaborate jurisprudence about "pools" which attached to the predecessor provisions of the pre-2010 legislation in the context of arguments about statistical analysis. No such arguments arise in the present case (and indeed they arise less frequently generally, following the changes made to the definition of indirect discrimination to bring it into line with the language of the Directives). Also, the word doing all the work is "material", which – at least in a difficult case – begs the question of what differences are or are not material; and that can only be answered by going back to the analysis under section 13 or 19, as the case may be.
34. The EAT's focus on section 23 led to us being referred to some of the numerous authorities on the selection of the correct pool for comparison. Mr Jones relied in particular on Spicer v Government of Spain [2004] EWCA Civ 1046, [2005] ICR 213 and Pike v Somerset County Council [2009] EWCA Civ 808, [2010] ICR 46. Neither, however, seemed to me to shed light on the essential question in the present case, as I have sought to analyse it above.
35. Mr Jones also drew our attention to a passage in the opinion of Advocate General Kokott in the recent decision of the CJEU in CHEZ Razpredelenie Bulgaria AD v. Komisia za Zashtita ot Diskriminatsia Case C-83/14, [2015] IRLR 746, in which she discusses the nature of indirect discrimination – see paras. 89-93 (p. 754); and to the answers given by the Court to the sixth to ninth questions posed by the referring court – see paras. 92-109 (pp. 767-9). But again, as I think in the end Mr Jones accepted, the issues there considered seem to me to have no application to the circumstances of the present case.
36. In short, therefore, I believe that the ET's explicit conclusion that the average shorter length of service of Muslim chaplains was not the result of any discriminatory practice on the part of the Prison Service means that they were not put at a particular disadvantage within the meaning of section 19 (2) (b); and the claim of indirect discrimination must accordingly fail.

Lord Justice Lewison:

37. I agree with the judgment of Underhill LJ. My reading of Essop (to which I was a party) is that the key to the decision is to be found in section 136 of the Equality Act 2010. That section provides:

"(1) This section applies to any proceedings relating to a contravention of this Act.

(2) 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, the court must hold that the contravention occurred.

(3) But subsection (2) does not apply if A shows that A did not contravene the provision."

38. In Essop the statistical disparity between members of the BME group who failed the relevant exam and other groups who passed was enough to raise a prima facie case of indirect discrimination: see [63] to [65]. Section 136 then shifts the burden to the employer to prove that the "reason why" was not the impugned PCP, either in relation to the group as a whole or in relation to a particular claimant. It is only in that sense that in Essop Sir Colin Rimer said that it was for the claimant to prove his case: see [66] and [67]. The mistake of the EAT in Essop was to jump straight from the establishment of a prima facie case to the question of justification, which only arises once the indirect discrimination has been proved (with or without the benefit of section 136). Thus on the approach of the EAT the employer was deprived of the chance to explain the "reason why". That was the approach originally advocated on Mr Naeem's behalf in our case. However, in our case the employer did explain the "reason why"; and the ET found that it was based on objective, non-discriminatory, factors.
39. For the reasons given by Underhill LJ, together with these reasons, I joined in the decision to dismiss the appeal.

Lord Dyson:

40. I agree that this appeal should be dismissed for the reasons given by Underhill LJ.  I see no inconsistency between para 30 of his judgment and para 38 of the judgment of Lewison LJ.

Note 1    The Prison Service is part of the National Offender Management Service (“NOMS”), but it was originally distinct and for convenience I will in this judgment continue to refer to it as such. The Respondent Secretary of State is responsible for NOMS.     [Back]

Note 2    The religions/denominations to which such sessional chaplains belonged were, in 2011, Roman Catholic, Free Church, Muslim, Sikh, Buddhist, Jewish and “pagan” (the last being engaged, I should say, only for one session per month).     [Back]

Note 3    There is a “Chaplaincy Payband 2”, affording a lower level of basic pay, to which some 97 out of the 353 employed chaplains in 2010 were assigned, but it is irrelevant for the purposes of this claim.    [Back]

Note 4    The figures were supplied, responding to how the Claimant formulated his questionnaire, by reference to “Christian” chaplains. The ET in its Reasons appears to proceed on the basis that all the “pre-2002-employed” Christian chaplains were Church of England – see, e.g., the Employment Judge’s summary quoted at para. 11 below. It is not entirely clear to me that that is the case, but the point does not affect the issues which we have to decide.     [Back]

Note 5    In fact discrimination on the grounds of religion and belief only became unlawful with effect from 2 December 2003, when the Employment Equality (Religion or Belief) Regulations 2003 came into force. But Mr Jones made it clear that it was not part of his case that the non-employment of Muslim Chaplains prior to 2002 would have been unlawful if the Regulations had been in force. (And in any event race discrimination, on which the Claimant relied in the alternative, has been unlawful since the 1970s.)     [Back]

Note 6    I accept that that is a slight over-simplification. The ET’s actual finding is that there was “objectively”, and non-discriminatorily, perceived to be no such need. But for present purposes that comes to the same thing. Indeed it is a useful thought-experiment to consider how the claim would look if, however unrealistically, there had been literally no Muslim prisoners before 2002.    [Back]

Note 7    Indeed if it were otherwise a employer who made positive efforts to increase the diversity of his workforce – say, by advertising vacancies in media with a greater appeal for women or members of ethnic minorities – would be making a rod for his own back, at least if length of service were a criterion in his pay system.     [Back]