HM Land Registry v Benson & Ors (Age Discrimination : no sub-topic) [2011] UKEAT 0197_11_1002

EAT
HM Land Registry v Benson & Ors (Age Discrimination : no sub-topic)
[2011] UKEAT 0197_11_1002 · 2012-02-10
[1]The Land Registry, which is the Appellant, is the government body responsible for maintaining a register of land in England and Wales. It is a self-financing trading fund under the aegis of the Ministry of Justice which charges fees for its services which are used to pay staff and meet costs. It is required to break even year-on-year: in years where there is a surplus this is held by way of reserve and cannot be spent without Treasury approval. The Appellant’s employees are civil servants. The majority are Registration Executives (“REs”), Registration Assistants (“RAs”) or Registration Officers (“ROs”). REs fall into two grades, RE1 and RE2. RE2s are sub-divided into “upper” and “lower” (“RE2U” and “RE2L”). The split between RE2Ls and RE2Us is roughly two-to-one.[2]The Appellant has believed for some time that it is seriously over-staffed. An “Organisational Blueprint” issued in late 2006 proposed a gradual reduction from 8,000 employees to 6,000 in 2014. However, those proposals were overtaken by the property slump in 2008, which had a very serious effect on the Appellant’s income: it made an operational loss of almost £80m in 2008/9. It needed to cut costs, and in particular staffing costs, urgently. It obtained Treasury approval to spend £50m from its reserves on schemes to encourage staff to leave by voluntary redundancy or early retirement. It originally proposed two such schemes, an “RO/RA Scheme”, for the most junior grades, and an “Approved Early Retirement Scheme”. However, when it became clear that the cost of these would not amount to the full £50m available, it also decided to accelerate proposals in the original blueprint to merge three pairs of offices which were geographically close to each other. The offices in question were in Birkenhead, South Wales and Lytham St. Anne’s: the two Lytham offices were known respectively as the “Lytham” and “Lancashire” offices, with the merged office to be known as “Fylde”. This was referred to as the “Merging Offices Scheme”.[3]This appeal is concerned with the Merging Offices Scheme, to which we will refer simply as “the Scheme”, and in particular with the merger of the Lytham and Lancashire offices. The Scheme was aimed at REs, since ROs and RAs could take advantage of their own scheme. Under the Scheme REs at the merging offices who were over 50 would be offered “compulsory” early retirement, and those who were under 50 would be offered the opportunity to volunteer for redundancy (“compulsory early severance”) under the terms of the Civil Service Compensation Scheme (“the CSCS”). The benefits available under the CSCS can be very substantial and are correspondingly attractive to many employees. In the case, specifically, of early retirement, employees would be entitled to immediate payment of a lump sum and pension, based on their years of service at the date of retirement but, in the case of the pension, unreduced for early receipt.[4]The Scheme was announced on 15 December 2008. Staff were invited to apply. It was made clear that the available budget was finite and that all applications would not automatically be accepted: depending on the level of interest in the Scheme, a selection process might be required. That indeed turned out to be the case. One of the criteria for the selection process which was applied – we give more detail below about how it operated – was how much it would cost to release any given applicant: that is, subject to certain qualifications, the Appellant selected those whose entitlements under the CSCS would be the lowest. We refer to this as “the cheapness criterion”.[5]The selection process took place in early 2009, with the principal decision-taking meeting occurring on 25 March. Letters to applicants informing them of the outcome went out shortly thereafter. That was not the end of the process since formal offers, quantifying the benefits payable in each case, had to be made, and it was possible that applicants might change their minds at that point – though that was not very likely and there were in fact only fourteen such cases. There was a further meeting on 17 April which resulted in offers being made in seven further cases, in two of which the applicants were from Lytham/Lancashire. The retirements/dismissals took effect at the end of May 2009.[6]We should mention at this stage, though it is only indirectly relevant to the issues on this appeal, that the Appellant subsequently, in the autumn of 2009, put forward a radical “Accelerated Transformation Programme” for its business (“the ATP”), involving further very substantial staff reductions. The first draft of its business case, prepared for the purpose of obtaining Treasury approval, was produced in October 2009: it showed an overall cost, to be funded from the Appellant’s cash reserves, of £128m. A version of the ATP was in due course implemented.[7]The Claimants in these proceedings are six RE2s at the Lytham office whose applications under the Scheme in early 2009 were not accepted. The cases of five of them – the First to Fifth Respondents – are substantially identical. The sixth is different and was only heard with the other cases as a matter of convenience. Specifically:(1) Mrs Benson, Mr Melling, Mr Wakefield, Mrs Hunt and Mrs Jardine were all aged between 50 and 54 and applied for early retirement. They claim that the application of the cheapness criterion involved unlawful discrimination against applicants in their age group, contrary to the Employment Equality (Age) Regulations 2006 , essentially because, for reasons explained below, the benefits payable were higher for them than for employees in other age groups and they were accordingly more expensive to release. We refer to these as “the age discrimination Claimants”.(2) Mrs McGlue was aged 38 at the time of the offers under the Scheme. At that time she was absent on a career break, which she had elected to take in March 2008 following the end of her maternity leave. The decision was made at the selection meeting on 25 March 2009 not to make offers under the Scheme to employees on a career break unless they were due to return before 1 April 2010 (i.e., near enough, within the next year). Mrs McGlue was excluded on the basis that she was not due to return before that date. She claims that the application of that policy constituted indirect sex discrimination within the terms of section 1 (2) (b) of the Sex Discrimination Act 1975 .[8]Those claims were heard by an Employment Tribunal sitting at Manchester, chaired by Employment Judge Brain, over eight days in November 2010. The age discrimination Claimants were not legally represented, but Mr Melling presented the case on behalf of all of them. Mrs McGlue was represented by Mr Charles Apthorpe of counsel. The Appellant was represented by Mr Mark Whitcombe of counsel. By a Judgment and Reasons sent to the parties on 3 February 2011 the claims of all six were upheld. We should say at this stage that the Tribunal’s Reasons are commendably thorough and careful.[9]The Appellant has appealed. The age discrimination Claimants are now represented by counsel, Mr Stuart Brittenden. Mr Apthorpe continues to represent Mrs McGlue. Ms Marina Wheeler now appears for the Appellant.[10]We will deal separately with the claims of the age discrimination Claimants and of Mrs McGlue. A. THE AGE DISCRIMINATION CLAIMANTS THE REGULATIONS AND THE ISSUE 11. Reg. 7 (2) of the 2006 Regulations provides (so far as relevant) that:
“(2) It is unlawful for an employer, in relation to a person whom he employs at an establishment in Great Britain, to discriminate against that person — (a)-(c) … (d) by dismissing him, or subjecting him to any other detriment.”
Discrimination is defined at reg. 3 as follows:
“(1) For the purposes of these Regulations, a person ('A') discriminates against another person ('B') if- (a) on grounds of B's age, A treats B less favourably than he treats or would treat other persons, or (b) A applies to B a provision, criterion or practice which he applies or would apply equally to persons not of the same age group as B, but- (i) which puts or would put persons of the same age group as B at a particular disadvantage when compared with other persons, and (ii) which puts B at that disadvantage, and A cannot show the treatment or, as the case may be, provision, criterion or practice to be a proportionate means of achieving a legitimate aim. (2) A comparison of B's case with that of another person under paragraph (1) must be such that the relevant circumstances in the one case are the same, or not materially different, in the other …”
We adopt the usual shorthands of referring to discrimination within the meaning of limbs (a) and (b) of reg. 3 (1) respectively as “direct” and “indirect” discrimination, and to the requirement of the final part of the definition as “justification”. We will also use the abbreviation “PCP” for the phrase “provision criterion or practice” in limb (b).[12]The formulation of the test of justification in reg. 3 (1) is an expression of the well-known proportionality test deriving from EU law. That test has been formulated in different ways, though without any intended difference of substance. In Bilka-Kaufhaus GmbH v Weber von Hartz [1987] ICR 110 the test was said to be whether “the measures [in question] correspond to a real need on the part of the undertaking, are appropriate with a view to achieving the objectives pursued and are necessary to that end” (see para. 35 of the judgment of the Court, at p. 126). It was expressed more succinctly by Balcombe LJ in Hampson v Department of Education and Science [1989] ICR 179 (see at p. 191), in a formulation approved on many subsequent occasions [1] , as requiring “an objective balance between the discriminatory effect of the condition and the reasonable needs of the party who applies the condition”.13. It is, superficially, a curiosity of the present case that the Claimants’ complaint is not that they have been dismissed but that they have not been dismissed – or, more accurately in the case of the age discrimination Claimants, that they have not been offered the opportunity to retire. But, as we have said, the enhanced benefits available under the CSCS mean that dismissal or early retirement may, depending on the employee’s circumstances, be a very attractive option, and it was common ground before us that exclusion from the opportunity to receive those benefits was capable of constituting a detriment within the meaning of reg. 7 (2).14. The Appellant accepted before the Tribunal that it applied to the Claimants a PCP or PCPs which put employees in the age group 50-54 at a particular disadvantage compared with other employees applying under the Scheme, and that accordingly prima facie indirect discrimination falling within the terms of reg. 3 (1) (b) was established. That concession was made on the basis of a statistical analysis carried out by the Appellant following the implementation of the Scheme which showed that proportionately fewer such employees were selected for release. Specifically, only 26% of applications from this age group (who represented about a quarter of all applicants) were successful, as against 69% for those aged 40-44, 48% for those aged 45-49 and 75% for those aged 55-59 (those groups between them representing the great majority of the applicants). The Appellant identified the PCP(s) which had this effect as being “all factors taken into account … at the decision-taking meetings” (see Reasons, paras. 206-207); but it is in fact clear that the only one of those factors which caused the disadvantage complained of is the cheapness criterion. The reason why the cheapness criterion disadvantaged applicants in the 50-54 age group is straightforward. A n employee taking retirement at age 50 instead of at the normal retirement age of 60 would receive under the Scheme ten additional years of pension, at 51 an additional nine years and so on. It is thus inherently more expensive, all other things being equal [2] , to release such employees than to release either their younger colleagues, who are entitled only to “compulsory early severance” benefits under the CSCS (which cannot exceed six years’ pay and are generally much less), or their colleagues in their later fifties, who will be entitled to fewer additional years’ pension. [3]15. It followed from that concession that the application of the PCP was unlawful unless it could be justified. Justification was the sole issue before the Tribunal. THE CHEAPNESS CRITERION IN MORE DETAIL16. There were in the event more applicants under the Scheme than the Appellant felt it could accept. 404 employees applied from the six offices, with 142 applying from the Lytham and Lancashire offices: all but a handful were RE2s. The cost of releasing that number of applicants on CSCS terms was far more than could be accommodated within the available budget. [4] Out of the £50m originally allocated by the Treasury, £38m was taken up by the other two schemes, so that only £12m was available for the Merging Offices Scheme. If all those who applied from the six offices had been granted early release the cost would have been £33.6m. That left a prima facie shortfall of £21.6m; but in fact £1.9m could be found from a separate pot, so that the actual shortfall was £19.7m.17. The selection process which was operated against that background was clearly explained in the Appellant’s evidence and is fully set out by the Tribunal at paras. 106-126 of the Reasons. It can be sufficiently summarised for present purposes as follows: (1) The decision-takers – being a Mr Simm, a Ms Llewellyn-Jones and a Mr Evans – understood that they had a budget of £12m (see above). Within that budget they needed as a minimum to lose a sufficient number of employees at each of the pairs of merged offices to get down to the number that represented (on a fairly tight allocation of space) the maximum capacity of the retained premises. But they also sought to release more employees than that minimum target. As Mr Simm put it:
“The idea was to spend the whole budget in order to reduce staff numbers as much as possible as it was felt that the more staff we could release the better it would be for the Land Registry and the most costs would be saved and the offices would be less cramped.” (2) It necessarily followed that the primary criterion was cheapness: by choosing those applicants whose entitlements would be the least they would not only ensure that they hit their minimum, space-related, target but would also be able to release the maximum additional number within the constraints of the £12m budget. However, that could not be the sole criterion. It was necessary to retain essential specialist expertise (though that was not in fact an issue as regards RE2s) and, more pertinently for present purposes, to maintain a proper balance between RE2Us and RE2Ls: since RE2Ls were less well-paid than RE2Us and on average had lesser length of service the cheapness criterion would lead to a disproportionate number of them being released. (3) Against that background, a spreadsheet was produced at the meeting of 25 March 2009 which showed the cost of releasing each applicant in all of the six merging offices, with the most expensive to release at one end and the cheapest at the other. The decision-takers first drew a line at the point necessary to achieve the minimum reduction necessary at each pair of offices. In the case of Fylde that was 91, so the 91 cheapest applicants from the two Lytham offices were selected. There was then an adjustment to restore the current proportion of RE2Us and RE2Ls: i.e. some “cheaper”
RE2Us who had most nearly missed (provisional) selection were swapped with some of the more “expensive” RE2Ls who had only just qualified for it. (4) That process did not bring the total cost to £12m, so the decision was taken to select a further 25 employees from each pair of offices. These were selected on the same basis, i.e. by “cheapness” subject to an adjustment to maintain the RE2U/RE2L balance. (5) The meeting on 17 April was not of real significance for present purposes. (6) The total released from the Lytham and Lancashire offices was 117 as against 142 applicants. The total for all six offices was 238, as against 404 applicants. 18. The Tribunal found that other selection criteria – in particular, some form of assessment of performance or of value to the business – had been considered prior to the key meeting, but they were not in the end employed (save that performance assessment may have been used in one or two cases by way of “tie-break”). The recognised trade union, PCSU, had been strongly opposed to any criterion which involved the making of judgments about performance or value and had on that basis advocated the cheapness criterion which was in the event employed. THE TRIBUNAL’S REASONS 19. The Tribunal’s conclusions on the age discrimination claim are to be found at paras. 231-260 of the Reasons. Its reasoning can be summarised as follows. 20. Paras. 231-235 are introductory. The only point that we should note for present purposes is that at para. 231 the Tribunal expressly accepts the Appellant’s definition of the relevant PCP(s): see para. 14 above. 21. At paras. 236-237 the Tribunal sets out, following the language of the justification test in reg. 3 (1), to identify the aim(s) which the relevant PCP(s) sought to achieve. At para. 236 it accepts Mr Whitcombe’s definition of the aims in question, which had been recorded at para. 208 as follows:
“… to reduce headcount to ensure that the remaining staff would fit as comfortably as possible into a single site; to reduce payroll costs, both immediately and also in the longer term; and to leave the right balance of grades in place at the merged offices. The respondent says that the reduction of headcount is the primary aim.” (We should note at this stage that Ms Wheeler in her skeleton argument submitted that that characterisation of the Appellant’s aims was incomplete. It should have added “within a budget of £12m” or “in accordance with the approved available budget”). At para. 237 the Tribunal holds that the aims identified by Mr Whitcombe were legitimate. It says: “The Tribunal has little difficulty in upholding the respondent’s case that the three aims of the scheme are legitimate. The respondent recognised some three years prior to the implementation of the accelerated Merging Offices scheme that there was a need to reduce its estate and its staffing numbers. There can be no question but the reduction of the excess costs being incurred by the respondent in running its organisation entailed by excess staff numbers and excess estates is a legitimate aim to be pursued by a public body which has a requirement to break even. Acceleration of the Merging Offices programme to assist with that costs reduction was clearly in pursuit of a legitimate aim.” 22. Having sought to identify the Appellant’s relevant aims, the Tribunal turns – again, following the language of reg. 3 (1) – to consider the means adopted to achieve those aims. Paras. 238-240 identify the basis of selection adopted, as we have summarised it above. At the end of para. 240 the Tribunal endorses a submission by Mr Whitcombe that “… it makes all the difference in the world that the scheme was over subscribed. It was not possible to release everyone within the available budget and therefore it was necessary to adopt a method of selection.” 23. At para. 241 the Tribunal finds that the basis of selection was not purely on grounds of cost, because of the adjustment made to ensure a proper balance between RE2Us and RE2Ls. It holds that this was accordingly a “cost plus” case. That is plainly a reference to the judgment of this Tribunal in Cross v British Airways plc [2005] IRLR 423 (see para. 63, at p. 435) and means that it rejects any argument that the justification argument must fail because it is based on “cost alone”
. This finding might be debatable, but it is not challenged in the Claimants’ Answer and we need not accordingly become embroiled in the question about the validity of the cost plus approach raised in Woodcock v Cumbria Primary Care Trust [2011] ICR 143 , at para. 32 (pp. 155-6). 24. At para. 242 the Tribunal makes the following finding:
“The cost of release of the applicants was plainly the most significant factor but we find that costs considerations were mitigated by considerations of merged office size and grade balance. The Tribunal agrees with the respondent that, subject to that mitigation, no other satisfactory method of selection was shown to be possible (whether at the material time or subsequently). There appears to have been no unequivocal agreement with the trade union as to how to go about the selection process. The union’s suggestion was to select simply upon the basis of cost. There are real difficulties with the methods of selection that have been suggested. The vast majority of staff were assessed in the second of three categories (see paragraph 105) and therefore some other way of distinguishing between them would have had to be found. The risk of being seen to reward poor performance or failure on the one hand and the risk of allowing the better employees to go on the other are profound difficulties which do not give of an easy answer.” (The “three categories” referred to relate to performance assessment, which, as noted above, had previously been considered as a selection criterion.) At para. 243 the Tribunal points out that discrimination could have been avoided by allocating extra funds and releasing all applicants, but the point is not developed (though it is returned to later – see below). It then observes that if the Appellant had simply released everyone in the Claimants’ age group that would probably have involved discrimination against applicants in other age groups. At para. 244 the Tribunal notes that the only other way of avoiding discrimination was not to proceed with the Scheme at all; but, importantly, in para. 245 it finds that there was “a real need to accelerate the Merging Offices Scheme” in view of the steep decline in revenue and the losses anticipated for 2008/9. (That finding is repeated at para. 253.) 25. To anticipate, Ms Wheeler submitted that the Tribunal’s findings in paras. 242-244 should have been treated by it as decisive of the justification issue: it had found that the means adopted to achieve the identified aims were the only ones possible. That was not, however, the Tribunal’s conclusion. The reasoning in the following paragraphs is elaborate and it is not, with respect, entirely easy to follow how its various parts fit together; but, as we understand it, there are essentially two stages to the reasoning, which we consider in turn. 26. The first stage starts with the conclusion, at the start of para. 246, that “ the Tribunal holds that there was not a real need to embark upon a selection exercise of this nature”
. What is meant by “of this nature” appears from the remainder of the paragraph and paras. 247-251, where the Tribunal makes a number of criticisms of the process followed by the Appellant. In particular, it finds that the process was carried out with unnecessary haste and without performing an equality impact assessment, which would have identified the disproportionate impact of the selection criteria – and specifically the cheapness criterion – on applicants in the Claimants’ age group. At para. 252 it says:
“The Tribunal holds, therefore, that the respondent has not shown there to be a real need to have adopted the means which they did in order to accelerate the Merging Offices programme. The reality is that the respondent did not have to adopt the chosen means entailing selection from amongst those who applied for early release.” 27. The way that that conclusion is worded suggests that the Tribunal regarded it as dispositive of the justification issue. However, that does not appear to be the case, since it proceeds to a second stage. At para. 253 it repeats its previous finding “that there was a reasonable need to accelerate the Merging Offices programme” and that “… other than not embarking on the Merging Offices scheme at all … the only non-discriminatory way of proceeding with such a scheme would have been for the respondent … to allow all of those who applied to be released.”
And it repeats, at the start of para. 254, its finding that “in practical terms selection could only be on cost”. It continues:
“… The question that arises is whether there a real need for the respondent to limit the amount of money spent upon the Merging Offices scheme to £12m which led to the indirectly discrimination selection exercise? Put another way, was there a reasonable need not to allocate some further funds to enable all of the applicants to go?”
It then proceeds, over paras. 255-260, to address that question. Its answer is, in short, that the Appellant could and should have avoided the problem by releasing everyone who applied under the Scheme – i.e. the point already floated at para. 243 (see above). The Tribunal acknowledges that to release all the applicants would have cost a further £19.7m, but it believed that that cost should have been incurred. In that connection it makes the following points: (1) Neither the total cost of £33.6m nor the “additional” £19.7m represents a sum that would have had to be paid out in 2009. Rather, it represents the total cost of benefits that would be paid out over a period of ten years, i.e. until the last 50-year-old early-retirer reached what would otherwise have been the normal retirement age of 60. The Tribunal acknowledges that the relevant accounting conventions required the cost to be attributed to the year in which the retirements took place, but that did not reflect the cash outlay: see Reasons para. 256. (2) The costs figures did not incorporate the saving to the Appellant of not having to pay the salaries of the released employees: see para. 257. (3) The Appellant at all material times had significant cash reserves: see para. 256. At the start of 2008/9 they stood at £293.7m, though they would clearly be reduced by the operating losses during the year. (4) It relied on some features of the ATP as undermining the justification offered for imposing a cost-related cap on the numbers who could be released under the Scheme: see in particular paras. 256 and 257. The Tribunal concludes this section as follows, at para. 258:
“The Tribunal concludes, there, that the respondent has not shown that it was unaffordable to mitigate the adverse discriminatory impact of the means chosen to allow all applicants to be released. In the Accelerated Transformation Programme, the respondent contemplated a significant draw upon capital reserves to fund significantly greater redundancy costs. The burden is upon the respondent to show that the chosen means correspond to a real need and are reasonably necessary to that end. The Tribunal was not satisfied with the cogency of the respondent’s evidence upon the crucial question of the costings in the scheme and of the affordability or lack of affordability of avoiding the disparate impact complained of by releasing everyone who applied under the Merging Offices scheme.”
We note that those paragraphs contain two rather different kinds of point. There are some specific points about whether the costings showed a true picture; but there is in addition the broader point that £19.7m was only a fraction of the Appellant’s reserves and that only a few months later, in the ATP, it was prepared to spend a far higher part of those reserves on a restructuring exercise. It seems to us that it is the latter which must be the real point. On the Tribunal’s approach, even if the costings were right they were affordable. It is only if the reserves had been, say, £20m that details about the precise figures or accounting approach would matter. 28. The relationship between the two stages of the reasoning analysed in the two preceding paragraphs is not explicit. It may be that they were intended as two alternative reasons for rejecting the justification defence; or the Tribunal may have regarded only the second stage as decisive, recognising as it did that the question of whether it was justifiable for the Appellant to select between the applicants would have arisen even if the process had been unimpeachable. In any event, in our view the focus should be on the reasoning in paras. 255-258. It is hard to see how the procedural failings which the Tribunal criticised at the first stage of its reasoning could have undermined the justification for the use of the cheapness criterion if it was in principle a proportionate means of achieving a legitimate aim. 29. Although para. 258 appears to conclude the Tribunal’s reasoning on the justification issue, there are in fact two further paragraphs in this section, as follows: (1) Para. 259 refers to the ATP. Precisely what point is being made about the Scheme with which we are concerned is unclear. (2) Para. 260 reads as follows:
“The Tribunal considers, upon the evidence, that it was affordable for the respondent to have avoided the disparate impact by allowing all applicants to be released. At any rate, the respondent has not proved, the burden being upon it, that it was not. The discriminatory impact upon the claimants is significant as recognised by the respondent. The Tribunal is not satisfied that the additional costs that would have been incurred by the respondent is disproportionate to the benefit to the claimants in terms of eliminating the discriminatory impact. Accordingly, the age discrimination claimants’ claims succeed.”
So far as we can see, this essentially re-states the points made in the previous paragraphs, and in particular para. 258. THE APPEAL 30. The Notice of Appeal identifies five grounds of appeal, which are set out in some detail. But Ms Wheeler in her oral submissions focused on a single point which underlies most of the various pleaded formulations. She submitted that the Appellant’s aim, for the purpose of the issue of justification, should be characterised as being to achieve the maximum staff reduction consequent on the merger of the relevant offices that was possible within the available budget of £12m (cf. para. 21 above), subject to the retention of necessary skills and a proper balance between the grades; and that the Tribunal had accepted at para. 237 of the Reasons that that aim was legitimate. It inevitably followed, she submitted, that selection was necessary, because allowing all the applicants to go would have involved over-spending that budget by more than 150%; and the Tribunal had itself found at para. 242 that, if selection was necessary, cheapness was the only practicable criterion. Ergo the claim of justification should have succeeded. Where the Tribunal went wrong was in disregarding the fact that the budget of £12m was an essential part of the Appellant’s aim. Ms Wheeler referred to the decision of this Tribunal (Elias P presiding) in Chief Constable of the West Midlands v Blackburn [2008] ICR 505 (subsequently approved by the Court of Appeal – [2009] IRLR 135 ), in which the employment tribunal was criticised for falling into the error, first identified by Lord Nicholls in Barry v Midland Bank plc [1999] ICR 319 , of substituting a different aim for that in fact being pursued by the employer – see para. 30 (at p. 513A). 31. Mr Brittenden’s submission in response was that it was illegitimate to incorporate the budget of £12m as part of the definition of the Appellant’s aim. The relevant aims were those initially defined by Mr Whitcombe and accepted by the Tribunal. The number of employees to be released – and thus the question whether selection was required – was part of the means of achieving those aims and was accordingly subject to the proportionality test, which the Tribunal had properly applied. He drew attention to observations of this Tribunal (myself presiding) in Pulham v London Borough of Barking and Dagenham [2010] ICR 333 to the effect that employers “cannot automatically justify a failure to eliminate discrimination by allocating the costs of doing so to a particular budget and then declaring that budget to be exhausted” (see para. 41, at p. 355 F-G). 32. The first step in the analysis must be to identify the PCP or PCPs which had the discriminatory effect complained of. As noted above, the Appellant defined the relevant PCP(s) as “all factors taken into account … at the decision-taking meetings”; and the Tribunal accepted that definition. But, as we have said, that seems too general a description. For the purpose of a claim of indirect discrimination the PCP should be defined so as to focus specifically on the measures taken – that is, the thing or things done – by the employer which result in the disparate impact complained of (cf. Kraft Foods UK Ltd v Hastie [2010] ICR 1355 , at paras. 9-10 (pp. 1359-1360)). In the present case that would appear to mean that the relevant PCP is the cheapness criterion. No doubt other features of the selection process – in particular, the establishment of the maximum capacity of the retained offices and the adjustment to maintain the balance between RE2Us and RE2Ls – potentially affected the outcome; but the only feature which had a disparate impact as between applicants of different ages was the underlying selection on the basis of relative cost of the benefits payable under the CSCS (cf. para. 14 above). 33. The next step must be to identify the aim for the pursuit of which the cheapness criterion constituted the means. Plainly the criterion was a means of selecting between applicants, but it is necessary to identify what aim selection was intended to achieve. This is rather less straightforward. The immediate aim of selection was to bring the number of applicants down to a level the cost of which came within the £12m budgeted for the exercise. But it could be argued that it is necessary to include within the definition of the aim the carrying out of the redundancy/early retirement exercise itself, and perhaps also to ask what the aim of the exercise was. In that case the answer would be that the aim of the exercise was to reduce headcount, which in turn was a means of ensuring that the Appellant’s costs did not exceed its revenue. The truth is that the distinction between means and aim is not always easy to draw. 34. The next question is whether the relevant aim or aims were “legitimate”. The uncertainty about how to characterise them discussed in the preceding paragraph does not, fortunately, matter since in our view all the various potential elements are plainly legitimate. It is (to put it no higher) legitimate for a body such as the Appellant, like any business, to seek to break even year-on-year and to make redundancies in order to help it do so where necessary. It is likewise legitimate to offer voluntary redundancy/early retirement schemes of the kind with which we are here concerned: the Tribunal found in terms that the Appellant had a “real need” to implement the Merging Offices Scheme in 2008/9 (see paras. 245 and 253 of the Reasons – para. 24 above). And, most pertinently, it was in our view legitimate for the Appellant to impose a budget on the amount to be spent on such schemes in 2008/9, even if that might mean that selection had to be made between applicants. Like any business, it was entitled to make decisions about the allocation of its resources. The imposition of a maximum spend did not of course mean that selection was inevitable: that would depend on whether more than “£12m-worth” of employees applied. And, if selection were necessary, it would not inherently involve adopting a method which had a disparate impact on particular age-groups (or other groups with a protected characteristic): there could in principle have been selection by an assessment of value to the business (or indeed by lot) [5] . Any discriminatory impact is the result of the particular selection criterion chosen rather than of the aim being pursued. 35. The question then is whether the adoption of the cheapness criterion was a proportionate means of selection in order to meet the £12m limit (and, if this adds anything, the other aims which selection served). As we have said, it was not in principle the only means; and others were in fact considered. But the Tribunal found in terms that it was the only practicable criterion – see para. 242 of the Reasons (para. 24 above) – and that finding is not challenged. That being so, it is hard to escape the conclusion that its use was justifiable. The logic of the contrary position is that the only choices lawfully open to the Appellant were either to abandon the Scheme altogether, notwithstanding that it had on the Tribunal’s findings a real need to implement it, or to pay more than twice what it had decided to allocate to it (or, to put it another way, 40% more than it had, legitimately as we find, decided to allocate to redundancy/early retirement schemes in that year). That does not seem to us to be an acceptable conclusion: we return to this point below. 36. The Tribunal’s analysis of means and aims was different from ours. It treated the question of the £12m limit as an aspect of the means adopted by the Appellant to achieve more broadly-defined aims. We do not think that that is right, and it might have found its conclusions less comfortable if it had asked whether the imposition of the limit was “legitimate” rather than whether it was “proportionate”. But we do not think that the outcome of this appeal should turn on nuances of language or on the problems of drawing the distinction between aim and means. In Pulham (above) we said, at para. 15 (pp. 343-4):
“As Elias J observed in MacCulloch v Imperial Chemical Industries Ltd [2008] ICR 1334 (see para 10(2), at p 1338 a-b ), what the language of regulation 3 i s intended to express is the classic proportionality test, which has been applied to resolve issues of justification in discrimination cases at least since the decision of the European Court of Justice in Bilka-Kaufhaus GmbH v Weber von Hartz (Case 170/84) [1987] ICR 110 . The exercise required of the tribunal on such a test is: “to weigh the reasonable needs of the undertaking against the discriminatory effect of the employer's measure and to make its own assessment of whether the former outweigh the latter”: see para 10(4) in MacCulloch , at p 1338 c-d . That exercise of course necessarily involves identifying the “legitimate aim” which the employer is seeking to achieve by taking the measure complained of—that measure being the “means”
. But the dichotomy of “aim” and “means” is not always clearcut and the two elements can sometimes reasonably be formulated in more than one way. In Loxley v BAE Systems Land Systems (Munitions and Ordnance) Ltd [2008] ICR 1348 Elias J (President) observed in relation to the justifications relied on by the employer in that case that “whether [they] are better described as aims or as proper means of achieving the aims is perhaps a matter of semantics” (see para 37, at p 1356 c ). Tribunals need not cudgel their brains with metaphysical inquiries about what count as aims and what count as means as long as the underlying balancing exercise is carried out.” If the Tribunal carried out the essential balancing exercise defined in MacCulloch – and see also para. 12 above – its decision should stand even if there are some flaws in the precise formulation, and we do in fact note that in para. 258 it did use some of the classic terminology deriving from Bilka-Kaufhaus . We do not however believe that it did properly carry out that exercise. Our reasons are as follows. 37. The essence of the Tribunal’s reasoning was that the Appellant had not demonstrated a “real need” to limit its spending on the Scheme to £12m – or, to put it another way, to limit its spending on all three schemes to £50m. It held that it had not done so because it had not shown that payment of the additional £19.7m was “unaffordable”. By that it evidently meant that the Appellant had not shown that the funds were absolutely unavailable, in the sense that they could not be paid without insolvency: it pointed out that the Appellant’s reserves far exceeded that amount (albeit that Treasury approval was needed to spend them [6] ) and that later in the same year, in the ATP, it contemplated spending a far greater figure. In our view, to apply a test of unaffordability in that sense is to fall into the error of treating the language of “real need”, or “reasonable needs”, as Balcombe LJ put it in Hampson , as connoting a requirement of absolute necessity. It is well established that that is not the case: see the judgments of the Court of Appeal in Barry ( [1999] ICR 319 , at p. 336 A-B) and in Cadman v Health and Safety Executive [2005] ICR 1546 , and of Elias P in this Tribunal in Blackburn (above), at paras. 17-21 (pp. 509-510). In Cadman Maurice Kay LJ said, at para. 31 (p. 1560 B-C):[31]“The test does not require the employer to establish that the measure complained of was "necessary" in the sense of being the only course open to him. That is plain from Barry . … The difference between "necessary" and "reasonably necessary" is a significant one …” The effect of that principle, applied to a case like the present, seems to us to be that an employer’s decision about how to allocate his resources, and specifically his financial resources, should constitute a “real need” – or, to revert to the language of aim and means, a “legitimate aim” – even if it is shown that he could have afforded to make a different allocation with a lesser impact on the class of employee in question. To say that an employer can only establish justification if he shows that he could not make the payment in question without insolvency is to adopt a test of absolute necessity. The task of the employment tribunal is to accept the employer’s legitimate decision as to the allocation of his resources as representing a genuine “need” but to balance it against the impact complained of. This is of course essentially the same point, adjusted to the different formulation of the test, as we make at para. 34 above. If the Tribunal had carried out that exercise it would, we believe, inevitably have come to the same conclusion as we have reached, on our approach, at para. 35. 38. We have not in reaching this conclusion lost sight of the fact that the cheapness criterion was indeed disproportionately unfavourable to employees in the Claimants’ age group, and we can well understand their disappointment at their non-selection. But it is fundamental that not all measures with a discriminatory impact are unlawful. Sometimes they have to be put up with. And we are bound to say, though this is not central to our reasoning, that in the present case the impact is not as severe as in many cases. Those affected did not lose their jobs or lose any other benefit which they had a positive expectation of receiving: rather, they lost out on the chance to take advantage of a benefit – admittedly a very substantial benefit – on which they had no right to count and which could indeed be described as a windfall [7] . It is also important to emphasise that a central element in our conclusion is the Tribunal’s finding that the Appellant had no real alternative, if it was to select at all, to using the cheapness criterion. W e are not sure that that was an inevitable conclusion, and in other similar cases the evidence might well produce a different result. It does not, accordingly, follow from this decision that the use of a similarly discriminatory criterion will necessarily be justified in other cases. 39. Although we have reached that conclusion by the route that we have set out, we do gain some support from the decisions in Barry and Blackburn relied on by Ms Wheeler (see para. 30 above). Those are both cases where the employees’ essential case, when analysed, was that the employers should have adopted a different scheme (in one case a voluntary redundancy scheme and in the other a scheme of payment for shift-working) than they did, on the basis that such a scheme would have been more favourable for women. In both cases it was held that the only question for the tribunal was whether the measures complained of were justifiable in the context of the scheme actually adopted: what scheme to adopt was a matter for the judgment of the employer. Although the situations with they were concerned, and the details of the reasoning, are not directly comparable, the reasoning is at least congruent with our belief that the Tribunal in the present case should have accepted the £12m (or £50m) limit as representing a legitimate aim, or real need, while weighing it against its discriminatory impact. 40. We do not believe that what we say about “unaffordability” is in any way inconsistent with the decision in Pulham , as Mr Brittenden submitted. The assessment of affordability discussed in Pulham (see in particular para. 43, at p. 355H) was clearly not concerned with “absolute” affordability: otherwise the claimants would necessarily have succeeded. As for our observation that the size of an employer’s budget for a particular purpose could not be decisive of the question of justification, that must be read in the context of the issue in that case. The employer was there seeking to justify the continuation of a directly age-discriminatory pay provision. Its only justification was that it was too expensive to remove it: that question could not be closed off by the employer’s own decision as to what budget to allocate for that purpose. By contrast, the “budget” that we are concerned with in the present case is the budget for a particular project – namely a redundancy programme – which is not directly discriminatory, but which, as it turned out, required a selection exercise which, as it turned out, could only practicably be done on a basis involving some indirect age discrimination. 41. For those reasons we allow the Appellant’s appeal and dismiss the claim by the age discrimination Claimants. B. MRS McGLUE THE FACTS 42. Mrs McGlue was on maternity leave until 31 March 2008. She had previously applied for, and been granted, a six months’ career break, to follow immediately thereafter; but in February 2008 she was granted an extension to take the total period to one of five years, expiring on 30 March 2013. The Tribunal found that employees on career break enjoyed a unilateral right to return, prior to the expiry of the break, on giving reasonable notice. 43. Mrs McGlue was notified of the Merging Offices Scheme and applied for compulsory early severance under it. There had at that point been no decision on how to treat applicants who were on career breaks (or indeed absent for other reasons such as maternity leave or long-term sickness) and none of the materials available to employees addressed the question. It was only at the meeting of 25 March 2009 that Mr Simm and his team decided that those returning after 31 March 2010 should not be offered release in the initial tranche (though in theory they could have a second chance in the repêchage, in the highly unlikely scenario of a very large number of applicants refusing the formal offers made). That decision was made without any consultation with PCSU or any prior notice to the affected employees. Ms Walton, a “Senior HR Business Partner” with the Appellant, acknowledged in cross-examination that the failure to notify such employees was unfair because they might wish to consider whether to advance their return in order to qualify under the Scheme (see Reasons para. 172). The Tribunal evidently agreed with that concession. 44. There was, as we have noted, a repêchage. Because of the theoretical possibility that, if a very large number of applicants refused the formal offers, employees on career break might after all have their applications considered, Mrs McGlue received the same standard-form letter as other unsuccessful applicants, which referred to the possibility of a second-round offer. The Tribunal found that that was misleading in her case, since there was no realistic prospect of such an offer. 45. It was only in an FAQ document included in a letter sent following the meeting of 17 April that it was revealed that applicants on a career break with a return date after 31 March 2010 had been excluded. Mrs McGlue raised a grievance. She complained of indirect sex discrimination, but she also made the point that if she had been told of the importance of the date of 31 March 2010 she would have given notice to return to work before that date: her original request for a five-year break had been precautionary only, and she had been minded (though she had not yet told anyone this) to come back to work after her daughter’s third birthday, which was in February 2010. The grievance was in due course rejected, as was a subsequent appeal. Mrs McGlue did in fact give notice to return to work on 25 February 2010 and did so on 30 March. 46. Mrs McGlue’s case is of indirect sex discrimination. Section 6 (2) of the Sex Discrimination Act 1975 renders it unlawful to discriminate against an employee by, inter alia, subjecting her to a detriment. As we have already noted in relation to the age discrimination claims, the Appellant accepts that denial of the opportunity to apply for release under the Scheme constitutes a detriment within the meaning of the Act. By section 1 (2) of the Act a person (indirectly) discriminates against a woman if: (b) he applies to her a provision, criterion or practice which he applies or would apply equally to a man, but—(i) which puts or would put women at a particular disadvantage when compared with men,(ii) which puts her at that disadvantage, and(iii) which he cannot show to be a proportionate means of achieving a legitimate aim. ” 47. At para. 261 of the Reasons the Tribunal accepted Mr Whitcombe’s concession that the Appellant had applied to Mrs McGlue a PCP “that in order to be considered for [compulsory early severance] or [compulsory early retirement] employees currently on a career break should be due to return to work before 1 April 2010”. It found at paras. 262-264 that that criterion put women at a particular disadvantage when compared with men because women took proportionately longer career breaks than men. That finding was not challenged before us. 48. The next question under section 1 (2) was whether the application of that criterion to Mrs McGlue put her at a particular disadvantage. At para. 265 the Tribunal held that it did, because the Appellant’s failure to notify employees in her position of the criterion deprived them of the opportunity to choose to return to work before the essential date. We are not sure that we understand the relevance of that point to this particular issue (though it may be important to the reasoning more generally – see below); but the conclusion that Mrs McGlue had shown prima facie indirect discrimination was, again, not challenged before us. 49. The only issue was thus justification. The Appellant’s case as recorded at para. 266 of the Reasons was that to release employees on career breaks who were not due to return in the current year would not assist in achieving the aims of the Scheme. It would not help to achieve the reduction in headcount required to permit the merger of the offices because the employees in question were not working at any of the offices anyway; nor would it effect any reduction in payroll. 50. The Tribunal did not accept that case. It said, at the beginning of para. 267, that it “… [could] not accept that the decision to exclude people in Mrs McGlue’s position corresponded to a real need of the respondent and [was] appropriate with a view to achieving the objects pursued”. It proceeded to give three reasons for that conclusion, as follows:
“Firstly, those on a career break had, as we have found, a unilateral right to return to work on a reasonable notice. It was therefore a misplaced assumption upon the part of the respondent that they would not return to work before the end of the agreed career break period. They should therefore have been taken into account as there was always the potential of a return upon reasonably short notice and the consequent need for the respondent to resume paying salary. Secondly, there were only 15 employees who had career break return to work dates post 1 April 2010. Given the cash reserves available to the respondent the Tribunal cannot accept that it was proportionate to take a decision to exclude such a small cohort of employees. Thirdly, there was a range of numbers who could be released (paragraph 95). Including those career breakers and allowing their early release would not have materially affected the respondent’s organisational planning. There was simply no evidence to support Mrs Walton’s contention that to allow the career breakers to participate meaningfully in the process would open the floodgates to a significant liability upon the part of the respondent.”
Some further points are made in paras. 268-269, but these seem to us essentially to re-state the points made in para. 267. 51. We cannot accept the entirety of the Tribunal’s reasoning in paras. 267-269. In particular, the fact that the inclusion in the Scheme of employees on career breaks extending beyond 31 March 2010 might not cause any significant expense or disruption does not seem to us to address the essence of the justification advanced. The point is that the object of the Scheme was to reduce headcount at the six offices, and if the inclusion of these employees would not reduce headcount the fact that it would do no particular harm is beside the point. However, “headcount” is not a black-and-white concept. Staff on career breaks are still (at least under the arrangements in place in the Civil Service) employees; and if, as the Tribunal found, they had the right to return on reasonable notice they could not be treated as definitively absent in the year under consideration. We have no difficulty in accepting that the Appellant was entitled for the purpose of its exercise to adopt reasonable criteria for deciding who did and did not count as part of the present strength of the offices in question, and that such criteria might well be justifiable even if they had some discriminatory effect: absolute equality of impact as regards groups with protected characteristics is not always practically achievable in a complex situation. But the Tribunal found that the criterion adopted here was not fair (or, therefore, reasonable), at least in the way that it was applied, because it did not allow employees on career breaks the chance to advance their return: see paras 47 and 51 above. That assessment of fairness was a matter for the Tribunal; and in our view once it was held that the practice adopted was unfair it could not be regarded as proportionate or relied on by way of justification. This is, we accept, not quite the way the Tribunal put it in paras. 267-269, but we are inclined to think – particularly in the light of para. 265, albeit that this is strictly in the wrong part of the argument – that it formed a central part of its reasons for finding in Mrs McGlue’s favour. Even if it is not, the point was squarely put that way by Mr Apthorpe in his submissions before us. 52. Accordingly we believe that the finding of indirect sex discrimination against Mrs McGlue should be upheld. CONCLUSION 53. We allow the Appellant’s appeal in the age discrimination claims, and those claims are dismissed. We dismiss its appeal in the case of Mrs McGlue. 54. We regret the time taken to promulgate this decision, which is principally the result of the pressure of other work in this Tribunal.[21][1] See, e.g., Webb v Emo Air Cargo (UK) Ltd [ 1993] ICR 175 , per Lord Keith at pp. 182-3; Allonby v Accrington and Rossendale College [2001] ICR 1189 , per Sedley LJ at para. 24 (pp. 1200 G-H); and Hardy & Hanson plc v Lax [2005] ICR 1565 , per Pill LJ, at para. 22 (pp. 1574-5). [2] It is not the case that any employee aged, say, 50, will necessarily be entitled to greater benefits than a colleague of 49 or 55, since both early severance benefits and pension entitlement reflect length of service, so that a 49-year-old or 55-year-old with long service would receive entitlements worth more than those of a 50-year-old in the same grade who had (albeit that this would be untypical) short service. [3] The assessment of the cost of release is not in fact straightforward, and there were some issues about it in the Tribunal; but the broad point made here applies irrespective of the various complexities. [4] In principle there could also have been difficulties about maintaining a proper service if too many RE2s were released, but the Tribunal found that that would not have been a problem on these numbers. [5] There might have to be some constraints or safeguards to retain proportions between grades and other desiderata, but that qualification is an unnecessary refinement for the present point . [6] We should record that Ms Wheeler did not seek to rely on the need for Treasury consent as a justifying factor in its own right: that being so, no distinction need, in the present context, be drawn between the Appellant and the Treasury. [7] That is for at least two reasons. Staff can have had no legitimate expectation of a redundancy exercise occurring at all, still less of one occurring in the comparatively short window that would make it particularly attractive in their case. Nor can they have any certainty that any exercise would not be over-subscribed or that, if it is, they would be among those selected (even if a non-discriminatory criterion were employed).