SOFTWARE 2000 LTD v 1) MR B ANDREWS 2) MR I TRINDER 3) MR R PROWSE 4) MR S LAWRENCE [2007] UKEAT 0533_06_2601

EMPLOYMENT APPEAL TRIBUNAL
BAILII case number: [2007] UKEAT 0533_06_2601Case No Appeal No. UKEAT/0533/06Venue 58 VICTORIA EMBANKMENT, LONDON EC4Y 0DS
Between
SOFTWARE 2000 LTDAPPELLANT1) MR B ANDREWS 2) MR I TRINDER 3) MR R PROWSE 4) MR S LAWRENCERESPONDENTS
Before
Mr Justice EliasDate 26 January 2007
[1]APPEARANCES For the Appellant Mr Daniel Tatton-Brown (of Counsel) Instructed by: Messrs Metson Cross Solicitors Quality House 5-9 Quality Court LONDON WC2A 1HP For the Respondent Ms Judy Stone (of Counsel) Instructed by: Messrs Lemon & Co Solicitors 34 Regent Circus SWINDON SN1 1PY SUMMARY Four employees successfully established before the Employment Tribunal that they had been unfairly dismissed for redundancy. The Tribunal found that there had been procedural defects. In particular the assessments in the redundancy exercise had been inadequate and subjective. The Tribunal considered whether the dismissals were fair under s.98A(2) of the Employment Rights Act 1996 , or whether a Polkey reduction should apply. They appear to have concluded that there was no reliable evidence to find either. The EAT held that the Tribunal were entitled to find that the dismissals had not been shown to be fair by virtue of s.98A(2), but that there was evidence which they ought to have considered in order to decide whether, and to what extent, a Polkey reduction was appropriate. Case law on the application of Polkey considered and certain principles summarised. THE HONOURABLE MR JUSTICE ELIAS (PRESIDENT) The Employment Tribunal sitting at Reading found that the four claimants, the respondents to this appeal, had been unfairly dismissed by reason of redundancy because of various procedural defects. The Tribunal rejected a submission that they should find that the dismissals were fair under s.98A of the Employment Rights Act 1996 on the basis that they would have occurred in any event even if fair procedures had been complied with. They also assessed compensation on the basis that there should be no reduction for the risk of dismissal, thereby rejecting a submission that Polkey required such a reduction in the circumstances of this case. It is that latter conclusion which is the subject of this appeal. Background 2. The four employees who are respondents to this appeal were all employed in different positions by the appellant company. Mr Andrews was group marketing manager, Mr Trinder was a project manager, Mr Prowse was group customs services manager, and Mr Lawrence was a product manager. They were all made redundant and they all claimed unfair dismissal.[3]The appellant specialises in printing and imaging software technologies. It employs some 96 people in its Oxford office. It is part of a group with subsidiaries in other jurisdictions in Japan and Canada and the group employs about 182 people.[4]The company faced economic difficulties and proposed to make redundancies. On 15 July 2005 all staff worldwide were warned of the need for job losses and voluntary redundancies. The procedure adopted was to identify a number of distinct pools embracing everyone in the company except the chairman, two cleaners and an administrator.[5]One pool, described as Business Services A, was comprised of the group HR manager and then senior persons such as directors, operations manager, and more senior IT staff. Business Services B comprised more junior staff. There was then a separate pool for product management and technical staff which was described as engineering/innovation A; less senior technical staff and engineering staff were located in engineering/innovation B. The final pool was called development support services, covering quality assurance and customer support.[6]The company decided to adopt a matrix system for selecting potentially redundant employees. Those responsible for implementing this were Ms Anderson, the group HR manager, and Mr Philip Walker, the group legal manager. They based the criteria on those suggested by ACAS and added certain further criteria of their own. The following criteria emerged: ability to relate to colleagues/customers; aptitude for work; commitment/reliability; relevant skills and experience; and standard of work performance.[7]The assessment was left to the managers. Each criterion then was to be scored against a maximum of 20 so that the overall potential total was 100. However, the assessors were not told how they should approach the assessment nor, indeed, were they given any fuller explanation of what these criteria involved. Ms Anderson told the Tribunal what her understanding was, but she had not communicated these views to the assessors as a group.[8]On 26 July a pool briefing meeting took place with employees from each pool. The employees were stunned by the news. They were told of the proposed selection criteria and it was confirmed that the lowest scoring employees would be most likely selected for redundancy. By this time some managers had filled out their assessments but others had not. Those who had not done so were given score sheets by Ms Anderson and asked to complete the sheets by the close of business that day. In all there were eighteen separate managers involved in scoring, and Miss Anderson spoke to just three of them.[9]These three, but none of the others, were told that ten marks would be the norm for somebody who was performing adequately in relation to the particular criterion in issue. Fifteen was intended to indicate exceptional performance; and five, serious performance problems of a disciplinary nature.[10]All these score sheets were collected on 26 July 2005. Mr Lismer, who had conducted the briefing meetings and was chief executive officer, reviewed them all to ensure that there had been a broadly consistent approach on grading and also to satisfy himself that he agreed with the scores for the individual employees.[11]He accepted that he did not know the quality of everyone's work but his impression was that the scoring was broadly in line with what he would have expected for the individuals in respect of whom he did have knowledge.[12]There were some exceptions. For example, he observed in relation to one of the criteria relating to Mr Trinder that it was "surprising" that he had been marked as low as he had. But having so observed, he took no steps to alter the score.[13]The company had decided to abandon the idea of having a distinct and separate marketing department. It identified in particular that the posts of Mr Lawrence, product manager, and Mr Andrews, the marketing manager, would not continue. The Tribunal noted that they might have been fairly selected for redundancy on this basis. However, these gentlemen were not immediately made redundant. Rather, they were placed in their appropriate pool from which redundancies were drawn and therefore given the chance to remain in employment. (The original grounds of appeal suggested that this showed that the employees could have been dismissed fairly and that there should be no compensation on that basis. However, the point was not in the event pursued on appeal, rightly in our view. The issue was what would have happened, not what could have happened.)[14]On his score calculation, Mr Andrews scored between 51 and 54 out of 100 (this range was because he was given a mark of 9/12 for one of the criteria). Mr Lawrence's score was 52; Mr Trinder's overall score was 48 and Mr Prowse was quite low at 44. Each was either bottom, or second from bottom, of their respective pools. Mr Andrews and Mr Lawrence were in the same pool. They were the only two selected from their pool out of a total of nine employees. Mr Trinder and Mr Prowse were also in the same pool, although a different pool to the other two employees. They were the lowest two in a group of 22, in which three were selected for redundancy. Therefore, they were all at risk of redundancy.[15]Each was then contacted by letter on 27 July 2005. That falsely represented that the employees had agreed both the criteria and the pool, but it was conceded that that was not so. The letter informed them of their scores with respect to each criterion. Each was invited to a separate consultation meeting, though they were told that there were no vacancies suitable for them in terms of their skill, status and terms of employment. Each was on that same day escorted from the premises, their electronic passes were cancelled and network access was denied.[16]The so-called consultation meetings were to be held in August. Mr Andrews had a meeting on 1 August. Mr Trinder chose not to attend his, which had been scheduled for 2 August and was postponed to 8 August. Mr Prowse and Mr Lawrence also attended meetings on 1 August. They subsequently were sent letters on 9 August informing them that they had been made redundant with effect from 5 August. This backdating was, as the appellant company conceded, an error.[17]In all there were 16 compulsory and 2 voluntary redundancies. It is common ground that they were all selected from the 96 Oxford staff.[18]Each of the four employees appealed. In the case of Mr Andrews, Mr Prowse and Mr Lawrence, the appeals were rejected after a hearing. Mr Trinder was offered a post following his appeal, as customer support engineer, at the same salary and with the same conditions and benefits as he had previously received. He, however, rejected the offer on the grounds that the way the redundancies had been handled had destroyed his trust and confidence in the company.[19]During the course of the next year they were notified of vacancies arising in the company, but for various reasons did not follow them up.[20]The Tribunal considered and rejected a number of grounds on which it was submitted that the procedures were unfair. First, they rejected a submission that the dismissals were automatically unfair because of a failure to comply with s.98A(1) of the 1996 Act. Second, there was a contention that the redundancy selection procedure was a sham and that those to be made redundant had been pre-selected. That, too, was rejected. Third, they considered and rejected a contention that the pools had not been properly or sensibly identified. They did consider them to be curious because the company had not identified the roles or categories of roles that they would require, but held nonetheless that the selection did not betray any error of law because the employers genuinely considered these to be sensible pools and there was a rational basis for them.[21]Then it was said that it was unfair for managers to score employees in their own pool. There might be a tendency to mark them down because the assessors were in competition with the staff they were marking. Again, it was held that the employers were entitled to rely upon the integrity of those marking and it was a reasonable approach. Similarly the Tribunal rejected a contention that it was unfair to have a number of managers making the assessments in each pool. However, they did consider that given both these features, it was particularly important to have consistent and fair criteria with a proper review.[22]The Tribunal was, however, critical of the criteria on the grounds that they were unduly subjective. Such factors as commitment/reliability could not be measured objectively, and nor could ability to relate to colleagues/customers or aptitude for work. The Tribunal pointed out that there were regular appraisals of staff which could have formed a better and more objective basis for selection. Notwithstanding these criticisms the Tribunal was not prepared to say that the criteria adopted were such that no reasonable employer could have selected them.[23]However, as far as the application of the criteria was concerned, including the manner of the consultation, they were very critical of the company's approach. There was no explanation as to how the criteria were to be assessed and only some managers were given guidance on this, on a very ad-hoc basis. There was a significant risk, said the Tribunal, that not all the assessors would be "singing from the same hymn sheet". Accordingly, they could not be satisfied that the criteria were fairly applied. Whilst the subjective nature of the criteria themselves did not render the decision unfair, when coupled with the assessment by a large number of different managers, without proper guidance, they did render the dismissals unfair.[24]As to the review process carried out by Mr Lismer, they were not satisfied that this was anything more than a rubber stamping exercise. Even where he found a surprising result, he did nothing about it, and the whole exercise was carried out in the course of a single evening.[25]The Tribunal also considered the scores in relation to the particular claimants. They recognised that it is not for the Tribunal to subject the graded assessment systems to a micro-analysis and they must not substitute their own view for that of the employer. The question was whether the scores were wholly unjustified so that a reasonable employer could not have allocated the relevant score to the employee. Whilst noting that certain criticisms of some marks had been made, the Tribunal was not prepared to say that any score was wholly inappropriate.[26]The question of alternative employment was also considered. The Tribunal here considered that in relation to two of the claimants, namely Mr Lawrence and Mr Andrews, there was a sales position which was given to a Mr Bartlett, the IT manager, and for which they were suitable. The Tribunal were not impressed by the reason given for failing to consider those two individuals for that post. They did not conclude that they would have got the job, but they should, said the Tribunal, have been given the chance to compete for it.[27]Finally, the Tribunal considered an argument advanced by the employers that even if the dismissals were unfair, nonetheless, had a fair procedure been adopted, these claimants would have been dismissed in any event. In the case of Mr Andrews and Mr Lawrence, it was submitted that the tribunal itself had recognised that they could have been fairly selected on the basis that their posts had ceased to exist. Hence it was said that in the circumstances, applying the Polkey principle as required by s.98A(2) of the Employment Rights Act 1996 the dismissal should be held to be fair.[28]In the alternative it was contended that the Tribunal should have reduced the compensation to take account of the chance that these employees would have been dismissed in any event. The Tribunal dealt with these arguments in the following terms (para 36):
"The Tribunal then considered whether or not a fairer procedure would have made any difference and if so to what extent either under Section 98A(2) of the ERA or Polkey . The Tribunal was of the view that the selection process adopted by the Respondent was fundamentally flawed. It is impossible for the Respondent to say whether any or all of the Claimants would have been dismissed in any event had the selection criteria been properly and fairly applied. In relation to this point the Respondent argued that the posts of Mr Andrews and Mr Lawrence were redundant and that they would have been made redundant in any event. The Tribunal did not accept this submission. It is right that these two Claimants could have been made redundant discreetly (in the absence of suitable alternative employment) as their roles were redundant. However, the Respondent chose not to approach the matter in this way choosing instead to put them in the pools for selection with everybody else. Accordingly, they required to be assessed in the same manner as everybody else and it is impossible therefore for the Tribunal to conclude that had the process been applied fairly that they would have been dismissed."
[29]The Tribunal then considered the loss flowing from the dismissals and fixed a sum in respect of each of these claimants. Relevant legal principles[30]Section 123 of the Employment Rights Act 1996 sets out how the compensatory award should be calculated:
"… the amount of the compensatory award shall be such amount as the tribunal considers just and equitable in all the circumstances having regard to the loss sustained by the complainant in consequence of the dismissal in so far as that loss is attributable to action taken by the employer."
[31]In determining the loss sustained, it is plainly material for a tribunal to consider what would have happened had no dismissal occurred. Sometimes that might be a matter of fact, such as where the workplace closed shortly after the dismissal making everyone redundant (see e.g. James Cook and Co.(Wivenhoe) Ltd v Tipper [1990] ICR 716.) In most cases, however, it involves a prediction by the Tribunal as to what would be likely to have occurred had employment continued.[32]A Tribunal may also have to make predictions as to what would have happened had a dismissal which is subject to procedural defects been conducted in a procedurally fair way. In some cases this may require the Tribunal to find that the dismissal was fair notwithstanding the procedural errors. This is the effect of s.98A(2) of the Employment Rights Act which is as follows: "Subject to subsection (1) failure by an employer to follow a procedure in relation to the dismissal of an employee shall not be regarded for the purposes of subsection 98(4)(a) as by itself making the employer's action unreasonable if he shows that he would have decided to dismiss the employee if he had followed the procedure." (The reference to subsection (1) is to the statutory procedures, breach of which renders a dismissal automatically unfair. The concept of procedure in s.98A(2) is by contrast a broader one, as the EAT made clear in Kelly Madden v Manor Surgery [2007] IRLR 17 ).[33]I set out the context in which that section was passed in Kelly Madden (para 35):
"Prior to the coming into force of this section, the House of Lords had established that if a dismissal was found to be unfair for procedural defects then the fact that the employer would have been dismissed in any event, even had he complied with all the proper procedures, would not generally render a dismissal fair. It could only do so in the very exceptional circumstances where a reasonable employer could dispense with all such procedural safeguards. The relevance of the fact that the employee might have been dismissed in any event went to the question of remedy and not liability: see Polkey vAE Dayton Services Limited [1987] IRLR 503 HL overruling the earlier Court of Appeal decision in British Labour Pump Co Ltd v Byrne [1979] ICR 347 IRLR 94."
[34]In Kelly Madden the EAT held that the effect of s.98A (2) was to reverse Polkey and to reinstate the position as it had been under British Labour Pump v Byrne (save where the statutory procedures have been breached.) This means that if the employer satisfies the Tribunal on the balance of probabilities that the employee would have been dismissed even had fair procedures been adopted, then the dismissal must be held to be fair. If the Tribunal finds that there would have been a chance of such a dismissal but falling short of 50%, then the Tribunal must find that the dismissal was unfair but reduce the compensation accordingly in line with the Polkey decision itself.[35]There are other circumstances when a Tribunal may have to make a prediction as to what would have happened had the employee not been dismissed. The evidence might show that the employee would have been dismissed in the relatively near future in any event, and moreover this might be for a reason wholly unrelated to the circumstances of the dismissal. In O'Donoghue v Redcar and Cleveland Borough Council [2001] IRLR 615 the Court of Appeal held that a Tribunal was entitled to find on the evidence that an employee unfairly dismissed by reason of sex would have been fairly dismissed for misconduct six months later in any event because of her antagonistic and intransigent attitude.[36]There have been a number of decisions which have considered the question whether a Tribunal is obliged to carry out this exercise in all cases, or whether it is entitled to say that it is too speculative because the evidence is too unreliable. In King v Eaton (no.2) [1998] IRLR 681 the Inner House of the Court of Session held that it would sometimes be impossible for a court to engage in this exercise. That was a case where the employee was found to have been unfairly dismissed in circumstances where there was no consultation over the selection criteria, and where the criteria had not been fairly applied. The employers sought to adduce evidence at the remedies hearing to establish that dismissal would have occurred in any event even if fair procedures had been adopted. The Tribunal refused to allow such evidence to be adduced.[37]The Employment Appeal Tribunal rejected the appeal, and the Inner House dismissed the further appeal. Lord Prosser said this (para 22):
"…it seems to us that in this case the absence of consultation at the stage when a method was being adopted, and criteria for assessment selected, has the effect of putting that method and those criteria in doubt. To say that this 'goes to the heart of the matter' appears to us to be entirely justified. To ask whether the same method and criteria would have been adopted, if there had been consultation beforehand, or to try to show what method and criteria would have been adopted, in the light of consultation, is in our opinion to embark upon a sea of speculation, where the opinions of witnesses could have no reliable factual starting point. In such a situation, a tribunal is in our opinion well justified in refusing to allow evidence as to whether the unfair act or omission 'made a difference."
[38]It is important to note that the Court is not making a finding that employment would have continued indefinitely; it is simply saying that the nature of the exercise is too speculative to enable the court to take any sensible view about that. It is not practical that it should try to do so. The consequence of that approach, however, is that the Tribunal must thereafter make the assessment on precisely the same basis as if it had found that employment would continue indefinitely. That may work an injustice to the employer. That fact alone cannot of course dictate that a tribunal should undertake an exercise which it justifiably believes is simply too speculative, but it does suggest that the Tribunal should not be unduly reluctant to engage in the process.[39]In King v Eaton the Court sought to draw a distinction between mere procedural errors and what it termed "errors of substantive importance" which go to the heart of the case (para 19): "If there has been a 'merely' procedural lapse or omission, it may be relatively straightforward to envisage what the course of events would have been if procedures had stayed on track, rather than briefly leaving the track in this way. If, on the other hand, what went wrong was more fundamental, or 'substantive', and seems to have gone 'to the heart of the matter', it may well be difficult to envisage what track one would be on; in the hypothetical situation of the unfairness not having occurred. It seems to us that the matter will be one of impression and judgment, so that a tribunal will have to decide whether the unfair departure from what should have happened was of a kind which makes it possible to say, with more or less confidence, that the failure made no difference, or whether the failure was such that one simply cannot sensibly reconstruct the world as it might have been." ... Italics added. )[40]The distinction between procedural and substantive matters had not earlier found favour with the Court of Appeal. In O'Dea v ISC Chemicals Ltd [1996] ICR 222, 234-5 Gibson LJ had in terms stated that he found the distinction "unhelpful", although the concept of "procedure" is now one which will need to be defined since it occurs in s.98A(2).[41]In Lambe v 186K Ltd [2005] ICR 307 , para 59, Wall LJ, with whose judgment Butler-Sloss and Laws LJJ agreed, whilst sharing the view that it was not helpful to distinguish between substantial and procedural failings, nonetheless approved the passage from Lord Prosser's judgment which we have italicised, saying that it provided a "straightforward and sensible yardstick" for Tribunals.[42]It is pertinent to note that in the Lambe case there were both defects in the process of selection and an absence of consultation, but the Court of Appeal held that the Tribunal was entitled to conclude that the dismissal would have taken place at the end of the consultation period in any event.[43]In Gover v Property Care Limited [2006] EWCA Civ 286 ; [2006] 4 All ER 69 the Employment Tribunal found that there were fundamental failings in the way in which the employers sought to bring about changes in employment contracts. This led to constructive dismissals and a finding of unfair dismissal by the Tribunal. But it was held that even if proper consultation had occurred, the employees would not have accepted the fundamental changes which the employers were seeking to introduce into their contracts, and that all they had lost was the period during which consultation would have occurred, which the Tribunal found to be four months.[44]In the course of giving judgment Buxton LJ in turn cited with approval the italicised passage in King , and added that the fact that the issue was a matter of impression and judgment for the Tribunal "indicates very strongly that an appellate court should tread very warily when it is being asked to substitute its own impression and judgement for that of the Tribunal…." (para 22). He also observed that the Polkey approach - assessing what would have happened had the dismissal been fair - was wholly consistent with the principle of assessing loss flowing from the dismissal on a just and equitable basis, which is the principle underlying section 123. These should be approached as "a matter for the common sense, practical experience and sense of justice of the Employment Tribunal sitting as an industrial jury" (para 14). He also approved the way in which HH Judge McMullen QC had described the process in the EAT in that case (para 26) where the judge had said that the Employment Tribunal's task was "to construct, from evidence not speculation, a framework which is a working hypothesis about what would have occurred had the [employer] behaved differently and fairly."[45]The most recent decision of the Court of Appeal on this topic is Scope v Dr Carol Thornett [2006] EWCA Civ 1600 , a case strongly relied upon by the appellants. In Scope the employee was a highly qualified engineer employed by Scope (formerly The Spastics Society) which provides aides for disabled persons. The job involved spending three or four days a week carrying out field assessments and the rest of the time taking part in the manufacture and adaptation of equipment. The appellant was suspended for alleged bullying and harassment of a colleague and following disciplinary proceedings she was given a final written warning and it was proposed that she should be relocated from the office in Exeter to one near Cardiff. She refused to transfer and was told that she was regarded as having resigned.[46]The Tribunal found that there was a dismissal which was unfair because a reasonable employer would not have insisted on transferring her in that way. It found, however, that there was 25% contributory fault. They went on to conclude that if the parties had been encouraged to try and make the arrangement work in Exeter then it would have lasted only about six months. Compensation was assessed on that basis.[47]The employee appealed to the Employment Appeal Tribunal. The EAT held that the Employment Tribunal had no proper basis for concluding that the employment would be limited to six months in the future. The Employment Tribunal itself had stated that it was "a highly speculative matter", and the EAT concluded that this demonstrated that it had "launched upon a sea of speculation" which is precisely what King v Eaton and subsequent authorities show that it should not do. Accordingly the compensation should have been assessed on the assumption that the post would last indefinitely.[48]The Court of Appeal upheld the appeal and restored the decision of the employment tribunal. The Court noted that there was a body of evidence capable of justifying the Tribunal's conclusion and held that they were in principle right to assess the loss in the way they had done.[49]The leading judgment in the Court of Appeal was given by Pill LJ. In time honoured fashion he also referred to the much cited passage from Lord Prosser's judgment, italicised above, and approved the observations of Buxton LJ that the appellate courts should be wary of interfering with the Employment Tribunal's conclusions on this matter of judgment. He then dealt with the criticism made by the EAT that the Tribunal had been improperly speculating in the following terms (paras 36-38):
"The EAT appear to regard the presence of a need to speculate as disqualifying an employment tribunal from carrying out its statutory duty to assess what is just and equitable by way of compensatory award. Any assessment of a future loss, including one that the employment will continue indefinitely, is by way of prediction and inevitably involves a speculative element. Judges and tribunals are very familiar with making predictions based on the evidence they have heard. The tribunal's statutory duty may involve making such predictions and tribunals cannot be expected, or even allowed, to opt out of that duty because their task is a difficult one and may involve speculation. Giving judgment in the leading case on loss of earning capacity, Moeliker v A. Reyrolle & Co. Ltd [1977] 1 WLR 132, an important head of damage in personal injury cases, Stephenson LJ when seeking words to define the correct approach to be followed stated, at page 144: "
I avoid "speculation" because this head of damage can really be nothing else". When Lord Prosser, in King , spoke of making decisions with more or less confidence" he was not saying that a prediction could only be made when confidence was complete; on the contrary, he contemplated prediction when there was "less" confidence. As already stated, however, there may come a point, at which evidence of countervailing factors is so slight that an indefinite continuation of the employment may be held to be an appropriate prediction. The parties should of course take care to place before the tribunal the material on which they seek to rely to establish how long the employment would have continued but for the unfair dismissal and tribunals can be relied on to scrutinise that evidence carefully when performing their duty to make a just and equitable compensatory award. They should adopt the flexible approach recommended by Buxton LJ in Gover ."[50]His Lordship also observed that whilst the evidence may be so sparse that the conclusion is that employment would have continued indefinitely, "where there is evidence that it may not have been so, that evidence must be taken into account." (para.34)[51]On the facts of this case the court held that the reasoning of the Tribunal was opaque and it was impossible to be clear why they had fixed on a period of 6 months. That was in principle a justifiable conclusion, but it was not clear how that figure had been arrived at. Accordingly, the case was remitted.[52]This case emphasises that the task is for the Tribunal to identify and consider any evidence which it can with some confidence deploy to predict what would have happened had there been no unfair dismissal. To fail to do this could lead to over compensating the employee, which would not be a just outcome. In this context we would caution against taking the phrase "constructing the world as it might have been" too literally.[53]The question is not whether the Tribunal can predict with confidence all that would have occurred; rather it is whether it can make any assessment with sufficient confidence about what is likely to have happened, using its common sense, experience and sense of justice. It may not be able to complete the jigsaw but may have sufficient pieces for some conclusions to be drawn as to how the picture would have developed. For example, there may be insufficient evidence, or it may be too unreliable, to enable a Tribunal to say with any precision whether an employee would, on the balance of probabilities, have been dismissed, and yet sufficient evidence for the Tribunal to conclude that on any view there must have been some realistic chance that he would have been. Some assessment must be made of that risk when calculating the compensation even though it will be a difficult and to some extent speculative exercise. Summary.[54]The following principles emerge from these cases:(1) In assessing compensation the task of the Tribunal is to assess the loss flowing from the dismissal, using its common sense, experience and sense of justice. In the normal case that requires it to assess for how long the employee would have been employed but for the dismissal.(2) If the employer seeks to contend that the employee would or might have ceased to be employed in any event had fair procedures been followed, or alternatively would not have continued in employment indefinitely, it is for him to adduce any relevant evidence on which he wishes to rely. However, the Tribunal must have regard to all the evidence when making that assessment, including any evidence from the employee himself. (He might, for example, have given evidence that he had intended to retire in the near future).(3) However, there will be circumstances where the nature of the evidence which the employer wishes to adduce, or on which he seeks to rely, is so unreliable that the tribunal may take the view that the whole exercise of seeking to reconstruct what might have been is so riddled with uncertainty that no sensible prediction based on that evidence can properly be made.(4) Whether that is the position is a matter of impression and judgment for the Tribunal. But in reaching that decision the Tribunal must direct itself properly. It must recognise that it should have regard to any material and reliable evidence which might assist it in fixing just compensation, even if there are limits to the extent to which it can confidently predict what might have been; and it must appreciate that a degree of uncertainty is an inevitable feature of the exercise. The mere fact that an element of speculation is involved is not a reason for refusing to have regard to the evidence.(5) An appellate court must be wary about interfering with the Tribunal's assessment that the exercise is too speculative. However, it must interfere if the Tribunal has not directed itself properly and has taken too narrow a view of its role.(6) The s.98A(2) and Polkey exercises run in parallel and will often involve consideration of the same evidence, but they must not be conflated. It follows that even if a Tribunal considers that some of the evidence or potential evidence to be too speculative to form any sensible view as to whether dismissal would have occurred on the balance of probabilities, it must nevertheless take into account any evidence on which it considers it can properly rely and from which it could in principle conclude that the employment may have come to an end when it did, or alternatively would not have continued indefinitely.(7) Having considered the evidence, the Tribunal may determine (a) That if fair procedures had been complied with, the employer has satisfied it - the onus being firmly on the employer - that on the balance of probabilities the dismissal would have occurred when it did in any event. The dismissal is then fair by virtue of s.98A(2). (b) That there was a chance of dismissal but less than 50%, in which case compensation should be reduced accordingly. (c) That employment would have continued but only for a limited fixed period. The evidence demonstrating that may be wholly unrelated to the circumstances relating to the dismissal itself, as in the O'Donoghue case. (d) Employment would have continued indefinitely. However, this last finding should be reached only where the evidence that it might have been terminated earlier is so scant that it can effectively be ignored. The grounds of appeal[55]The grounds of appeal are all directed to the Tribunal's conclusion in paragraph 36, set out above.[56]There has been some debate, instigated as a result of an intervention from one of the Employment Appeal Tribunal members, Mr Bleiman, as to precisely what the Tribunal did decide in that paragraph. Broadly, there are two possible interpretations of the Tribunal's analysis. One is that the Tribunal had merely focussed on s.98A(2) and were only reaching the conclusion that the employers had not satisfied them that dismissal would have occurred in any event, so that the dismissals were not thereby rendered fair. That would leave the question whether there was nonetheless a chance that the dismissals would have occurred which, if established, would have reduced the compensation in accordance with the Polkey principle.[57]The alternative view was that the Tribunal was making a finding that the situation was as in King v Eaton ; the evidence was so flawed that it was impossible to make any sensible reconstruction at all of what might have occurred had a fair procedure been adopted. This both precluded it from finding that the dismissals or any of them were fair, and that there was a risk of dismissal so as to justify a Polkey reduction. Both parties accept that the Tribunal was not making a positive finding that dismissal would not have occurred.[58]Mr Tatton-Brown, counsel for the appellant company, on reflection preferred the first interpretation. He accepts that if that interpretation is correct then that was a conclusion which the tribunal could properly reach on the evidence before it and he does not seek to challenge it. However, he submits that in that case the Tribunal ought to have gone on to consider the question whether there was nonetheless a chance that dismissals would have occurred. In the alternative, he submitted that if the latter interpretation was correct, the Tribunal erred in law because there plainly was evidence which it could and should have taken into account when considering making a Polkey assessment.[59]Ms Stone, counsel for the respondent employees, submits that the latter interpretation is the better one. The Tribunal was concluding that the whole exercise was too flawed and the evidence too unreliable to make any sensible prediction at all about what would have occurred. That was a conclusion they were entitled to reach in the exercise of their judgment. It could not be said to be perverse.[60]She made a further submission to the effect that the Tribunal was also at the same time making a finding that even if it did have reliable evidence enabling it to reconstruct the world as it might have been, its conclusion was that dismissal would have occurred in any event. We confess to finding it impossible to say that the Tribunal was reaching both those mutually incompatible conclusions at the same time. Construing paragraph 36.[61]There are pointers to both interpretations. Mr Tatton-Brown points out that the Tribunal say it is impossible to conclude whether the claimants would have been dismissed; but in respect of the compensation point the question is whether there is evidence from which it might be inferred that there was a chance that they might. The language used seems to reflect that in s.98A(2). Moreover, the rest of that paragraph is prefaced by the phrase "in relation to this point" and is directed specifically to the question whether Mr Andrews and Mr Lawrence would have been dismissed in any event.[62]Ms Stone says that the Tribunal plainly appreciated that it had to consider both s.98A(2) and Polkey and it should be assumed that the Tribunal was answering both those questions in that paragraph. Moreover, the reference to the process being "fundamentally flawed" points in the same direction. The errors were of the kind identified in the King case and the Tribunal was concluding that no sensible reconstruction of events was possible.[63]On balance we think that the Tribunal were intending to deal with both aspects, as Ms Stone submits. However, in our judgment it did so on the assumption that since the evidence was too unreliable to allow the Tribunal to determine whether dismissal would have occurred, then it was too unreliable to be taken into account at all. It is striking that the Tribunal did not make any reference to the evidence which, it is common ground, was identified by Mr Tatton-Brown as being potentially relevant to the Polkey assessment and which we consider below. Nor is there any indication in the judgment that it recognised that the lack of reliable evidence to show that dismissal would have occurred (i.e. a more than 50% probability, with the onus on the employer to show it would have occurred) did not necessarily determine whether there was reliable evidence that it might have occurred (i.e. a less than 50% chance but one which could properly lead to a percentage reduction in compensation).[64]In our view, therefore, the Tribunal misdirected itself on this matter. As the judgment of Pill LJ in Scope makes clear, the Tribunal must consider any evidence which is reliable and might justify a Polkey reduction in compensation. We accept that the Tribunal was fully entitled to conclude that the evidence was too unreliable to enable them to reach the view with any confidence that the dismissals would on the balance of probabilities have occurred, and Mr Tatton-Brown agrees. He is not therefore seeking now to reopen the finding of unfair dismissal with respect to any of these employees. His submission is simply that the Tribunal could not properly ignore patently potentially relevant evidence when assessing compensation.[65]He submits that having identified that the criteria were in principle acceptable, that the choice of line managers to carry out the assessment was appropriate, and since it was not possible to say that any individual score was plainly wrong, then the exercise was not unduly difficult or speculative. If the Tribunal intended to say that it was, that was a plain error of law. 66. He identified a number of specific features of the case which he submitted strongly indicated that there was a real risk of dismissal occurring for each of these employees.[67]There were a total of 9 persons in the pool from which Mr Lawrence and Mr Andrews were selected as the lowest two. That of itself created a realistic risk of redundancy. Moreover, whilst there were three others around the same mark as Mr Andrews - at least assuming that he was at the top end of his flexible mark - (Mr Lawrence being two marks lower), the other four in that pool were marked substantially higher. Moreover, the manager who assessed Mr Andrews' mark was the same as the person who gave a Mr Heys, who was top of the pool, a very high total of 91/93. Since it was accepted that these were both bona fide marks, that strongly suggested that Mr Andrews at least had not been unduly harshly assessed.[68]He submitted that the cases of Mr Trinder and Mr Prowse are even more striking. In their pool there were a total of 3 redundancies out of 22. However, they were well down the field. The break mark in their case was 60. Mr Prowse was 44 and Mr Trinder 48. The next lowest was59. There had been a comment in relation to Mr Trinder that one of his marks was somewhat low. Even so, the overall mark would still have been significantly lower than the break point of60. They were so far adrift from the break point, that given the finding that they were assessed bona fide and on the basis of common criteria applied to all - albeit not satisfactorily clarified - this was plainly evidence which the Tribunal had to consider when assessing overall compensation. 69. He contended that it was a significant injustice to the employers for the assessment exercise, much of which was accepted as one which a reasonable employer could adopt effectively, to be ignored altogether. That led the Tribunal to adopt the unrealistic positive assumption when assessing compensation that Mr Andrews and Mr Lawrence would not have been in the bottom two in their relatively small group, and Mr Trinder and Mr Prowse would not have been in the bottom three of theirs.[70]Ms Stone, in an attractive submission, emphasised that the cases to which we have referred were all cases where the decision of the Employment Tribunal had been upheld. The appeal court should not lightly interfere with the judgment of the Employment Tribunal on the question of the reliability of the evidence. She also emphasised in particular the fact that there had been no common standard identified for how the assessments should be conducted. The Tribunal should be taken to have concluded that this did indeed make the tables so unreliable that nothing could properly be inferred from them.[71]She also suggested that had there been further consultation then other or different criteria might have emerged. However, that does not seem to have been a finding of the Tribunal, and we do not think we can properly make that assumption. It is not clear that the lack of collective, as opposed to individual, consultation ever emerged as a significant issue. Conclusion[72]We think that the Tribunal, having properly applied s.98A(2) did not in fact go on to specifically address the question whether there was any reliable evidence relevant to the Polkey issue, as they were required to do so when assessing compensation under s.123. If they did indeed address that question then we are satisfied that they erred in law in concluding that there was no such potentially relevant evidence. Conversely, if they did not then it is a matter they should have considered.[73]We are persuaded that the matters which were identified by Mr Tatton-Brown both before the Employment Tribunal and before us were potentially material factors which could not in justice be simply ignored or treated as too unreliable.[74]We see the force of Ms Stone's submission that there was no consistent approach laid down for how the marking should be assessed. Plainly that would lead to some variations (although in the practical world we would be very surprised if assessors did not have some discussions with each other as to how the task should be approached). However, the maximum mark for any criterion was 20 and there would be unlikely to be any striking difference as to what the mark for average performance would be.[75]There was in addition the review by Mr Lismer which, whilst largely a rubber stamp as the Tribunal found, did identify what he considered to be some of the more surprising assessments. The criticism was that having done so, he did nothing about it.[76]We believe that it would be an injustice for the Tribunal simply to ignore these factors and act on the assumption that these employees would not have faced any risk of dismissal. That seems to each of us to be a wholly unrealistic basis for assessing compensation. There was evidence of sufficient reliability to require the Tribunal to engage in the Polkey exercise, albeit that there was insufficient reliable evidence to satisfy the Tribunal that any of these employees probably would have been dismissed. What weight should be given to the evidence is of course a matter for the Tribunal. Disposal[77]It follows that this case should now be remitted to the same Tribunal to consider the extent, if any, to which the compensatory award for each of these employees should be reduced in the light of the principles we have set out. We have every confidence that they will do so. Apart from the error we have identified, their decision was an impressively clear and careful analysis of the issues in the case. We would leave it to the Tribunal to determine whether it thinks it appropriate to hear further evidence or not. No doubt it will want to hear further submissions.

Cited in 17 later judgments

and 8 more of the most senior