United Kingdom Employment Appeal Tribunal

APPEARANCES

 

For Colt Technology Services Limited	MR CHRISTOPHER JEANS (One of Her Majesty's Counsel) Instructed by: Greenberg Traurig LLP 8th Floor, The Shard 32 London Bridge Street London SE1 9SG
For Mr J B Brown	MS AMANDA HART (of Counsel) Instructed by: Leigh Day Solicitors Priory House 25 St John's Lane London EC1M 4LB


 


SUMMARY

DISABILITY DISCRIMINATION - Compensation

The Claimant, who continues to be employed by the Respondent, succeeded in certain claims of harassment and discrimination and in relation to certain failures to make reasonable adjustments. The decision of the Tribunal following a remedies hearing was appealed by the Respondent employer and was also the subject of an appeal and cross-appeal by the Claimant.

On the various issues for determination:

(1) The law on the insurance exception to the principle of deducting avoided loss in the calculation of damages is as set out by the Court of Appeal in Gaca v Pirelli General plc [2004] 1 WLR 2683. There was evidence before the Tribunal that was sufficient to support a conclusion that the Claimant had contributed indirectly to the PHI policy taken out by the employer and the decision to deduct from loss of earnings only 50% of the payments he had received should not be interfered with;

(2) In adding to the Claimant's award sums representing the redundancy and notice payments the Claimant would have received had he been able to work and subject to redundancy in 2013, the Tribunal had erred by failing to take into account that these were sums that the Claimant has not lost and may still be entitled to. To that extent the Claimant's condition had resulted in his avoiding the loss for which the Tribunal had sought to compensate him and the appeal would be allowed to the extent of deducting those sums from the overall award;

(3) The Tribunal had erred in finding that the Claimant's receipt of PHI payments would cease for reasons other than his ability to work, but that finding could be deleted while leaving standing the Tribunal's general conclusions on his ability to recover and return to work;

(4) The Tribunal had been entitled on the available evidence to reach the conclusions it did on the anticipated period of the Claimant's recovery, the level of his future earnings and to take account of the risk of relapse as part of the exercise of making a broad estimate in relation to those matters; but

(5) The Tribunal had erred in approaching apportionment by reference to the division of events into those caused by the Respondent's discriminatory acts and those that were not so caused rather than by looking at the Claimant's condition and assessing whether it was divisible and if so how much of the harm suffered had been caused by the Respondent's discriminatory acts. The correct approach having been confirmed by the Court of Appeal in BAE Systems (Operations) Ltd v Konczak [2017] EWCA Civ 1188, the matter would be remitted back to the Tribunal for assessment of the issue of apportionment using the correct approach.

Appeal and cross-appeal both allowed in part.


 


THE HONOURABLE LADY WISE

Introduction

1. The Respondent is an organisation that provides data centre consultancy and storage management and services. The Claimant, Mr J B Brown, has been employed by the Respondent since 14 March 2005. Initially he worked as a manager in training, then as a senior manager from 2009 and as a senior specialist in training from January 2011. He suffered a breakdown at work on 9 March 2012 and a second breakdown on 19 April 2012, since when he has been absent from work due to incapacity. He presented a claim to the Employment Tribunal in July 2012 making various complaints under the Equality Act 2010 in respect of the protected characteristic of disability. In August 2014 the Employment Tribunal (chaired by Employment Judge Glennie) issued a Judgment ("the Liability Judgment") finding some but not all of the claims of harassment and discrimination arising from disability established and also finding three failures to make reasonable adjustments. The decision on one of those three failures was subsequently reversed following a reconsideration hearing.
2. The present appeals both relate to the Tribunal's Judgment of 18 October 2016 following a remedies hearing in September 2015. Both parties seek to appeal that Judgment, mostly on different grounds, although both are agreed that the Employment Tribunal erred in relation to what was likely to occur in relation to the payments of permanent health insurance ("PHI payments") to the Claimant in future. For convenience I will refer to the parties as Claimant and Respondent as they were in the Tribunal below. The Claimant was represented both before the Tribunal and in this appeal by Ms A Hart of counsel. Mr C Jeans QC represented the Respondent both before the Tribunal and on appeal. I regret that, as in the Tribunal below, it has taken a considerable period of time to produce a Judgment in this case.
3. Due to the number and complexity of the issues raised, I will set out the Tribunal's analysis on each of the issues raised in the appeals and cross-appeal in conjunction with the arguments made where necessary. It may be convenient, however, to note at this stage that the Tribunal identified at paragraph 4 of the Remedy Judgment 17 issues for determination, seven of which remain relevant to the arguments raised in the appeals and cross-appeal. These are:

"1. What is its assessment of the Claimant's medical condition including the prognosis and prospects of returning to work.

2. What, if any, contribution did the discriminatory acts make to that condition and consequently to the Claimant's sickness absence from work.

3. What apportionment should there therefore be in respect of any losses established.

4. Would or might the Claimant have been made redundant in about April 2013 in any event and, if applicable, how should this be reflected in any award of compensation.

…

6. For what period should the Claimant be compensated.

…

8. Should there be a 50% or a 75% deduction from loss of earnings in respect of PHI payments.

9. What is likely to occur concerning the PHI payments in the future."

Arguments Presented in the Respondent's Appeal

4. There were three grounds of appeal advanced in the Respondent's appeal and Mr Jeans presented each in turn. The first ground contends that the Tribunal erred in deciding that only payments representing 50% of the Claimant's salary should be deducted from the PHI benefit payments not the full 75% actually received by the Claimant. This first ground requires a little understanding of the Respondent's "flexible benefits" scheme. The scheme allows employees to construct a benefits package tailored to their lifestyle. It offers choice in relation to the level of annual leave, death in service life insurance, private medical insurance, permanent health insurance ("PHI") and the ability to make contributions to critical illness cover, childcare vouchers, dental cover and travel insurance. Some other benefits, such as pension contributions are compulsory. The Employment Tribunal dealt with the arguments in relation to this matter in the following way:

"Issue 8: Should there be a 50% or a 75% deduction from loss of earnings in respect of PHI payments.

52. The parties were agreed that the principle to be applied was that expressed in Atos Origin v Haddock [2005] ICR 277 whereby PHI for which the employer has paid the premiums is deductible in full from compensation. Conversely it was agreed that where a claimant has taken out insurance, any payments received would not be deductible because he should not be deprived of the benefits of insurance for which he has paid. The issue in the present case was whether on the correct interpretation of the operation of [the] Respondent's PHI scheme, the Claimant should be treated as having paid to increase the level of salary protection from 50%, which was automatic under the scheme, to 75%.

53. In Gaca v Pirelli General Plc [2004] 1 WLR [2683] the Court of Appeal held that where an employer pays the premiums under a group insurance policy, the employer rather than the employee should have the benefit, unless it can be shown that the employee had paid or contributed to the premiums "directly or indirectly". It was not sufficient that the employer's payments could be regarded as part of the "fruits" of the employee's labour. In Gaca the Court of Appeal cited with approval the approach of the Canadian judge Cory J in Cunningham v Wheeler that, to establish a payment or contribution, there should be some type of consideration given up by the employee in return for the benefit; and that an example was that of a payroll deduction in return for the benefits.

54. There are two elements to the Respondent's scheme: compulsory benefits and flexible benefits. The compulsory benefits included 50% salary protection. The flexible element included the possibility of enhancing the salary protection to 60% or 75%; taking other benefits; or taking the flexible element as additional salary. The default position, however, is that the salary protection is maintained at 75%: in other words, if the employee makes no changes to the flexible element, he will be credited with 75% salary protection.

55. The Claimant made changes to other benefits, within the parameters of the package, including to his annual leave and pension entitlements. He left the salary protection at the default position of 75%. Mr Jeans argued that, as the Respondent had paid all the premiums necessary to maintain this protection, and the default position had been undisturbed throughout the Claimant's employment, the payments should be deductible in full. He contended that it would be absurd, and would subvert the purpose of offering PHI if the Claimant could have chosen some other benefit, the additional 25% of salary received should not be credited against his losses.

56. The Tribunal preferred Ms Hart's submissions on this point. The benefits that the Claimant could have chosen instead of the additional 25% protection included taking additional salary. It was common ground that if the Claimant had paid, in the sense of handing over money, for the enhanced protection, the additional 25% would not be deductible. We could see no difference in principle between that scenario and the actual one in which the Claimant chose not to take additional salary but instead retained the enhanced protection. The Tribunal concluded that there was no material difference between an employee choosing to pay for enhanced protection out of salary that he had already received, and an employee choosing not to receive salary in order to obtain enhanced protection. In both cases there was consideration given in the Cunningham v Wheeler sense. It seemed to the Tribunal that in principle it should make no difference whether the default position was set at 50% or at 75% protection: in either case the employee exercised a choice between increased salary or increased protection.

57. The Tribunal concluded that the Claimant should be treated as having paid for the additional element of protection and that therefore the amount of credit against the loss of earnings that should be given to the Respondent was 50% of salary."

5. It was not in dispute that the Claimant had made other changes to tailor his benefits package but he had left the PHI payment at the default position and so had received benefits at the 75% level since 2012. Mr Jeans submitted that the Employment Tribunal had erred in law by failing to deduct all of the PHI payments because those payments had diminished the Claimant's loss. Reference was made to the leading authority on the insurance exemption of Hussain v New Taplow Paper Mills Ltd [1988] AC 514. In the employment context the EAT had also held in Atos Origin IT Services UK Ltd v Haddock [2005] ICR 277 that benefits resulting from PHI taken out and paid for by the employer are to be deducted from loss of income. As Mitting J put it in that case:

"29. … The basic rule is that avoided loss, save in the benevolence and insurance exception cases, is to be treated as avoided, so that a claimant cannot recover damages in respect of it."

Mr Jeans accepted that the position is different if it is the employee who has bought the insurance. There, as a matter of legal policy, the employee should not be deprived of the benefit of the insurance he has taken out by having it deducted from the compensation. This was clear from the decisions in Hussain at page 527, Gaca v Pirelli General plc [2004] 1 WLR 2683 at paragraphs 41 to 59 and Haddock at paragraphs 28 to 29. It was submitted that the question of who paid the premiums was a decisive one. If the employer pays the premiums then the PHI benefit is deductible. In Gaca v Pirelli cited above the Court of Appeal made clear that the exception to deduction of insurance benefits received was restricted to situations where the employee had paid or contributed to the premiums "directly or indirectly". This would include, for example, the employer deducting a sum from the employee's wages or salary which the employer then remitted to the insurance. In the present case the Employment Tribunal's decision on this point was based on the Claimant's election of the full 75% PHI protection when he could have taken additional salary and less protection although there was a minimum threshold of 50% protection. Counsel submitted that this reasoning was erroneous because making a choice from a menu of "flexible benefits" did not mean that the employee was paying for each benefit; he was simply choosing from the list of options all of which were provided by the employer. While it was accepted that there was a value ascribed to the protection, it was described as a "co-benefit" in the relevant schedule and so was a compulsory benefit from which the employee could not opt out. It was submitted that it would be absurd to penalise an employer for offering flexibility in the benefits package. To do so would constrain an employer such as the Respondent to remove flexibility in relation to PHI protection before the employer could derive benefit from providing it. The Tribunal's construct, that the Claimant had "bought" the 25% difference between the minimum available benefit and the one that he elected to take was particularly artificial given that the default allocation was 75% and the Claimant had never sought to alter it. A fundamental point was that the Claimant had never paid the premium. It had been paid by the Respondent who was then entitled to be credited with the full benefit of the insurance it had purchased. The choice made by the Claimant in this case did not detract from the determinative fact that the Respondent paid for all of the benefits. Mr Jeans explored what might have happened had the Claimant elected by positive choice to take only the 50% protection. The same argument would apply in that to say that he has paid for a different level of protection because he could have done something else was a misuse of language. Had he opted for 50% protection he would have received lower payments than he has in fact received and so he has benefited from the 75% he elected to receive in a situation such as the one that has arisen. The Tribunal had simply failed to follow the established authorities from this jurisdiction in, particularly Gaca v Pirelli.

6. It was anticipated that counsel for the Claimant would rely on the citation with approval by the Court of Appeal of the Canadian case of Cunningham v Wheeler [1994] 113 DLR (4^th) 1 in Gaca v Pirelli. Mr Jeans submitted that the Canadian case was easily distinguishable from the present one. In Cunningham v Wheeler it was made clear that what was required to constitute evidence of indirect payment or contribution by the employee was some type of consideration given up by the employee in return for the benefit. The Canadian court had given an example of a contribution to an insurance plan by the employee paid for directly or indirectly by a reduced hourly wage reflected in a collective bargaining agreement, which it indicated would be sufficient to avoid deduction of the benefits subsequently received. Accordingly, that case went no further than to support the contention that the employee had to contribute, directly or indirectly, to the premiums before the insurance payments would not be deducted. In the present case the Claimant did not take a reduced wage in return for the Respondent paying for the scheme. His overall remuneration package was the same regardless of the choice he made. He had taken no deduction in salary of any sort in order to receive the enhanced 75% protection. As all he had done was make a choice between a number of benefits all paid for by the employer. Accordingly, he would receive a windfall were only 50% of the PHI payments he continues to receive being deducted. It was clear that the Tribunal had been taken in by the examples given in the Canadian case and so gone beyond the established English authority on this point. It was wrong to equiparate making a choice with making a payment. While the facts of the present case might have thrown up a novel situation, the underlying principles in terms of the legal approach were clear. The Claimant had avoided loss by receiving 75% of his salary while unable to work and so deductions had to be made to that extent.
7. The second ground of appeal related to the inclusion of enhanced redundancy and notice payments as part of the loss suffered by the Claimant. The Tribunal expressed its decision on this matter as follows:

"Issue 4: Potential redundancy in April 2013

31. The Tribunal found that, had the Claimant remained in good health at least to the extent of being able to work or having returned to work by this stage, he would have been made redundant in April 2013 and would have receive a net redundancy payment agreed at £51,688.59. Although Ms Hart argued that the redundancy was not a certainty, there was nothing to contradict Mr Cross's evidence about this aspect. He states that, following a re-structure in early 2013, there was a reduced requirement for employees to carry out centralised technical training functions. The Claimant and a colleague Ms Walsh were put at risk of redundancy, and Ms Walsh was made redundant. The consultation with the Claimant was put on hold because of his ill-health. Mr Cross stated that by around April/ May 2013, the Claimant's role ceased to exist.

32. The Tribunal considered that the redundancy payment that the Claimant would have received should be taken into account in the calculation of his losses, subject always to the application of the 70% contribution. This was a sum of money that the Claimant would have received in the absence of his mental health condition, albeit that, as we shall explain, we find that he would have experienced a significant reduction in earnings thereafter."

The reference to the 70% contribution related to the Tribunal's conclusion that the discriminatory acts were more significant in the causation of the Claimant's illness than the non-discriminatory factors and that an apportionment of 70%/30% reflected the respective contributions of discriminatory and non-discriminatory factors. Accordingly each head of claim was awarded to the extent of 70%.

8. The Respondent accepted that if the Claimant had been made redundant in April 2013 he would have received those benefits. However, it was noteworthy that no claim for a redundancy payment was put forward in any of the Claimant's schedules of losses nor was any such claim asserted in submissions until, some months after the remedies hearing, the Tribunal requested written submissions on the point. It was also of interest that at paragraph 32 the Tribunal awarded these amounts stating that they were sums which the Claimant would have received "in the absence of his mental health condition". The proper starting point in assessment of losses in a case of this type was whether any losses claimed were caused by the acts of disability discrimination. In any event, the indisputable fact was that the Claimant has not "lost" a redundancy payment or payment in lieu of notice because he has retained his job. If he is made redundant in the future he will receive a redundancy payment at that point. The Tribunal's reasoning failed to address the question of what would occur if the Claimant is now dismissed, the Tribunal having already allowed a claim in respect of redundancy payment. It was inconsistent and wrong to award an employee who continues in employment a redundancy payment as part of their loss. On the Tribunal's own findings at paragraph 31 the Claimant had in effect been saved from being made redundant by his illness. Further, the value of the PHI payments he received after the date on which he would have been made redundant far exceeds the value of the redundancy payment and notice monies combined. As there was no real possibility of the Claimant being dismissed prior to his health recovering and as he will continue to receive PHI benefits until his recovery there is simply no identifiable loss of redundancy pay or pay in lieu of notice. As a fall-back position, if the Tribunal was right on the issue of principle, then at its highest the award would be for the possibility of redundancy on a loss of a chance quantification. In that event the Tribunal would have to make findings on how much less this future inchoate right was worth because of the chance that the Claimant would not be made redundant. The award in three months' pay in lieu of notice by the Tribunal was described as even more extraordinary. There was no possible basis on which the Claimant could be denied payment in lieu of notice if he is ever dismissed whether on the basis of redundancy or otherwise. It was a fundamental failure on the part of the Tribunal to fail to notice that the Claimant's right to redundancy and notice payment has not been lost due to discrimination. This could be attributable to the wording in paragraph 32 already mentioned. On the simple analysis of a redundancy payment being the quid pro quo for losing a job the Claimant could not possibly be awarded something of this nature when he is still in employment.
9. The third and final ground in the Respondent's appeal related to the potential cessation of the PHI payments. Mr Jeans explained that it is now common ground between the parties that the Tribunal erred in holding that PHI payments might be discontinued without the Claimant having sufficiently recovered his health to be disqualified from further payments. While little turned on this ground from the perspective of the Respondent's appeal it had a bearing on the Claimant's appeal and so some argument was presented in relation to it. The Tribunal addressed the issue of the period during which the Claimant should be compensated in some detail. The relevant passages are in the following terms:

"42. As indicated by its written question to the parties inviting further submissions, the Tribunal has found it difficult to assess what is likely to happen to the PHI payments in future and what impact they are likely to have on other aspects. At present it is in neither party's interests to disturb a situation in which the Claimant is receiving £65,000 per annum gross while unable to work. He is dependent on those payments financially, and while they are being made the Respondent is able to argue that he is suffering little or no financial loss.

43. Although there is no indication of the insurers threatening to stop the payments or to take steps that might lead to a cessation of them on some grounds, the Tribunal found it difficult to envisage that they would be content to go on paying such substantial amounts for another 20 years or more. Payments depend on the Claimant remaining in the Respondent's employment and it seemed to the Tribunal that, however this might be put under the terms of the policy, it was likely that there would come a point when the insurers would decline to continue paying. For example, they might say that as the Claimant has not attended work for several years and as the medical evidence is that, whatever other work he might be able to do, he will never return to work for the Respondent, that the Respondent should dismiss him, thus bringing the insurers' liability to an end. Alternatively they might assert that they should be entitled to treat both parties as if the Claimant had been dismissed.

44. Furthermore, if contrary to all of this the payments continue to be made, the Claimant might conclude that, as he is unlikely to receive more money as a result of returning to employment than he is receiving from the PHI payments, it would be in his best interests to leave the situation as it is. In saying this the Tribunal is not attributing to him any cynical motive: he might take the view that any return to work might risk further stress and anxiety, by which time he would have lost the security of the PHI payments. In that event, however, the Tribunal considered that it would be unjust to say that the Respondent should therefore bear the loss of earnings that would follow for the rest of the Claimant's working life. If the Claimant did take such a view, that would amount in the Tribunal's judgment to a new intervening cause of the loss which would break the chain of causation from the discriminatory acts.

…

46. The Tribunal considered it likely that, once the Claimant reaches the point when he can make a return to work, he will do so, even though this might mean the loss of the [PHI] payments when he makes that return. We so found for the following reasons:-

46.1. Until the events with which this case has been concerned occurred, the Claimant showed himself to be thoroughly committed to his work. He enjoyed it and was motivated to do well, as his appraisals illustrated. It seemed unlikely that in the long term he would be content to remain unemployed.

46.2. It is likely that once he is fit enough to contemplate a return to work, the medical advice will be that whatever the financial position might be, it would be beneficial for his health to do so.

46.3. Looking then at the financial considerations, as we have stated it seems unlikely that the Claimant will be able to rely on the PHI payments being made indefinitely.

…

49. The Tribunal also found that it should be assumed that the PHI payments will cease when the Claimant returns to work. This is on the basis that, if he takes up work with a different employer, it cannot realistically be said that he is still in the employment of the Respondent, which is a condition of the PHI payments being made."

10. Mr Jeans stated that there had been no evidence or submissions to support the Tribunal's third stated reason in paragraph 46.3. The relevant policy of insurance had been before the Tribunal and did not confer any discretion to withhold payments while the Claimant remained unable to perform "the material and substantial duties of [his] normal occupation". The insurer, Canada Life, had confirmed on more than one occasion in correspondence that was before the Tribunal that the PHI benefits would continue and that "… for as long as the medical evidence indicates that you [the Claimant] should remain on sick leave". The Respondent had also confirmed in correspondence that it would not progress redundancy consultation in respect of the Claimant whilst PHI was in payment. This was reiterated by the Respondent's witness at the remedy hearing. In addition to being against the self-interest of the Respondent to try to terminate the Claimant's employment while PHI payments were in force, there would have been implied contractual restrictions on the power to do so in the face of a PHI scheme. In Aspden v Webbs Poultry and Meat Group [1996] IRLR 521 it was held that where an employee benefits from a PHI scheme, the employer's contractual power to give notice is inapplicable. In these circumstances there was no foundation for the Tribunal's speculation that it was likely that there would come a point where the insurers would decline to continue paying under the scheme. The insurers had no legal right to do so for so long as the Claimant was incapable of working under the terms of the policy and in the absence of any evidence that the insurers would ever attempt to do so. The Tribunal had also speculated that the insurers might tell the Respondent to dismiss the Claimant to bring the liability to an end, another finding for which there was no basis in law or in fact. The Respondent had no difficulty with the finding that the Claimant would be likely to recover his health sufficiently to resume working by 2017 and that the PHI payments would continue until then however, if the appeal on that point was allowed the position would change and the unwarranted finding of the Tribunal in this respect would become relevant.
11. In addressing the Respondent's first ground of appeal in relation to the insurance exception Ms Hart confirmed that the legal principles were not in dispute in that she accepted that as a matter of law the Claimant cannot be compensated for a loss that will be avoided. However, the insurance exception to that applies where the employee has paid or contributed to the premiums. The rationale for the insurance exception is that a prudent man who lays by for a rainy day should get the benefit of it. The key case on this rationale was Gaca v Pirelli [2004] 1 WLR 2683 but in that case the employee was arguing that he should receive the benefit merely by reason of his labour and so the issue of fact in that case became the determination of who paid the premiums for the policy in question. While the particular form of the issue that arose in this case had not previously come before the courts of this jurisdiction, there was support for the Claimant's position in the Canadian case of Cunningham v Wheeler already referred to in the Respondent's submissions. In that case the Judge, Cory J had stated that in order to fall within the insurance exception what was required was:

"… evidence adduced of some type of consideration given up by the employee in return for the benefit. The method or means of payment of the consideration is not determinative. Evidence of a contribution to the plan by the employee, whether paid for directly or by a reduced hourly wage, reflected in a collective bargaining agreement, will be sufficient." (Gaca v Pirelli at paragraph 56 referring to Cunningham at page 407)

12. Ms Hart referred to a pay slip which had been lodged before the Tribunal and which identifies as a sum the flexible adjustment element of the Claimant's wages. The present case was distinguishable from Hussain v New Taplow Paper Mills where Kerr LJ had stated that the insurance exception did not apply because there was no evidence that the benefits had been paid for directly or indirectly including that the plaintiff in that case had not offered to show some detriment such as foregoing some other advantage which he would have had. In Page v Sheerness Steel Company plc [1996] PIQR Q26 at 33 to 34 Dyson LJ had stated that the central requirement of the insurance exception was that the cost of the insurance be borne "wholly or at least in part" by the Claimant. That too had been cited in Gaca. Ms Hart submitted that the employer in this case was entitled to the benefit only of the compulsory part of the scheme and a non-compulsory element was what was an issue here. The Claimant had made a positive choice in relation to a plan paid for by the employer. The various choices within the "flexi scheme" were qualitatively different. For example, the employee could opt to have the benefit of travel or medical insurance and if that choice was not taken they would receive a sum of money instead. The PHI protection was a deferred and contingent benefit. The Claimant made a prudent choice of opting for the maximum amount of benefit namely 75%. The three available options were PHI at 50% of scheme salary, 60% of scheme salary and 75% of scheme salary and as noted by the Tribunal the Claimant had left the PHI salary protection at the default of 75% but made changes to other core benefits including annual leave and pension. That was sufficient evidence that he had made positive choices in relation to each element. By foregoing the additional salary that he would have received had he made a different choice, he had paid or at least contributed indirectly to the insurance premiums, thus he should receive the benefit. As the Tribunal had noted there was no difference between an employee choosing to pay for enhanced protection out of salary he had already received and someone choosing not to receive a salary in order to obtain enhanced protection. In both situations there is consideration in the sense pointed out in Cunningham v Wheeler. While the Respondent had repeatedly emphasised that it was the company that paid the premium, that focus ignored that the issue was whether the Claimant had contributed directly or indirectly to that premium which he had by taking less salary in return for the maximum 75% benefit. There was no proper distinction between a deduction and payment in this context. A person such as the Claimant who opted for the 75% PHI protection has sacrificed the additional remuneration he would have received if he had chosen 50%. It could be inferred that the additional salary he would have received was the cost of the premiums. Accordingly, there was a direct correlation between the premiums paid by the employer and what is paid out to the employee. The Tribunal had been correct to conclude that it made no difference that 75% was the default position. The position taken by the Respondent in this case would penalise those who are more cautious and opt for the maximum protection. Any submission about eroding the availability of flexible schemes in general or punishing the employer were irrelevant. All that required to be decided was who had contributed to the premiums and the Tribunal had been correct to conclude that the Claimant had so contributed and so only 50% of the PHI payments should be deducted from damages.
13. In answer to the Respondent's second ground of appeal in relation to the redundancy payment and associated payment in lieu of notice as losses included by the Tribunal, Ms Hart relied on the Tribunal's reasoning at paragraphs 31 and 32 of the Remedy Judgment in resisting this ground of appeal. It was the Respondent that had urged on the Tribunal a finding that the Claimant would have been made redundant in April 2013 absent any discrimination. Accordingly, the Tribunal was entitled to take into account any consequential payments that he would have received as a result of that redundancy. The accepted task of the Tribunal in this context was to put the Claimant in the position he would have been had there been no discrimination. Under reference to HM Prison Service v Beart [2005] EWCA Civ 467 it was submitted that where a discriminatory act causes a personal injury which affects subsequent employment prospects, future losses are not terminated by a non-discriminatory event such as a dismissal. Rather the fact of the dismissal merely changes the nature of the comparator. Having found that the Respondent's discriminatory acts had caused the Claimant to suffer a deterioration in his health which affected his ability to work and that as a result he would suffer a loss of earnings until 2020, the Tribunal was correct to adopt an approach that first identified what the Claimant would have earned within the relevant time period but for the discriminatory act and then deduct from that sum what it considered would be the Claimant's likely earnings as a consequence of the discriminatory act (including any PHI deduction). The redundancy and notice pay were relevant to the first part of that calculation unlike the PHI deduction which is relevant only to the second part. When characterised that way it became apparent that it was nonsensical to suggest that the Tribunal should not have taken redundancy pay and notice pay into account when calculating the Claimant's future losses simply because that had not been pled by the Claimant or included in the schedule of loss. This was easily explicable by it never having been the Claimant's case that he would have been made redundant in April 2013. When the Respondent had put that argument forward, the Claimant had argued in response that if the Tribunal was to find that the Claimant would have been made redundant then that did not act as a cut off as the Respondent had claimed but merely changed the nature of the comparator. The Tribunal had been entitled to accept that argument and make the awards that it did.
14. The exercise of working out what someone would have earned but for a discriminatory act is by definition a hypothetical assessment. The corollary of the Claimant's situation would be an argument by someone that would have been promoted but for a discriminatory act but had not been. The calculation would involve quantification of the salary the person would have received in the promoted post. In this case the Tribunal required to assess what the Claimant's remuneration package would have been but for the discrimination and the redundancy would have been part of that picture and so changed the nature of the comparator. The argument was the same in relation to payment in lieu of notice as the Tribunal was looking at what would have happened had the Claimant not been discriminated against. It was clear that the expression "in the absence of his mental condition" in paragraph 32 should be read as "but for the discrimination" as it could not be said to be unclear what the Tribunal was doing. The Tribunal was not looking at the Claimant's current situation but at what would have happened to him if he had not been discriminated against. The position contended for by the Respondent was cherry-picking in that they sought to reduce the Claimant's losses by £65,000 but without taking into account the consequences of what would have happened to get to that point. This argument required to be considered in the context of the bigger exercise the Tribunal was embarked on in comparing what would have happened to the Claimant. The financial consequences of getting to the point the Tribunal found he would have been able to achieve post redundancy (a salary of £65,000) had to be taken into account. There was no error on the part of the Tribunal in including these payments in the calculation of losses.
15. Ms Hart addressed the Respondent's third ground of appeal which was the same as the point made in the Claimant's cross-appeal namely that the Tribunal erred in law in holding that the PHI payments might be ended. As indicated she was in agreement with Mr Jeans that the Tribunal's conclusion on this matter should not stand. The problem was that the Tribunal concluded that the PHI payments would stop before the Claimant had recovered and fully replaced it when both sides had suggested the contrary. Reference was made to the relevant submissions before the Tribunal. There was a relationship between the Tribunal's finding on the cessation of PHI payments and the Claimant's return to work. While the Respondent had focused on the Tribunal's findings that the payments might cease because the insurer may decline to continue payments at some point in the future, the Claimant's focus was on the finding that the payments might cease because remaining on PHI would constitute an intervening act by the Claimant. The specific finding challenged in this respect is in the following terms:

"44. Furthermore, if contrary to all of this the payments continue to be made, the Claimant might conclude that, as he is unlikely to receive more money as a result of returning to employment than he is receiving from the PHI payments, it would be in his best interests to leave the situation as it is. In saying this the Tribunal is not attributing to him any cynical motive: he might take the view that any return to work might risk further stress and anxiety, by which time he would have lost the security of the PHI payments. In that event, however, the Tribunal considered that it would be unjust to say that the Respondent should therefore bear the loss of earnings that would follow for the rest of the Claimant's working life. If the Claimant did take such a view, that would amount in the Tribunal's judgment to a new intervening cause of the loss which would break the chain of causation from the discriminatory acts."

16. Counsel submitted that the notion of an intervening event was not canvassed at all before the Tribunal. It was the Tribunal's own idea and was an erroneous one. An intervening act (novus actus interveniens) is "a new cause which disturbs the sequence of events something which can be described as either unreasonable or extraneous or extrinsic" - The Oropesa [1943] P.32 at 39. In the context of the acts of claimants, in McKew v Holland & Hannen & Cubitts (Scotland) Ltd [1969] 3 All ER 1621 Lord Reed confirmed that "if the injured man acts unreasonably he cannot hold the defender liable for injury caused by his own unreasonable conduct. His unreasonable conduct is novus actus interveniens". The Court of Appeal had examined what was meant by acting unreasonably in this context in the case of Spencer v Wincanton Holdings Ltd [2010] PIQR P8. On the basis of these authorities, counsel submitted that in order for an intervening act by a claimant to break the chain of causation it was required to be conduct which involved a high degree of unreasonableness in that it was conduct that was more unreasonable than an act which could otherwise have fallen within the ambit of a failure to mitigate to apportion damages. In the present Claimant's case, any refusal to terminate the PHI payments would be to his benefit and so inherently reasonable. In any event, the Claimant's receipt of PHI was part of the sequence of events that followed from the Respondent's conduct. It was a direct consequence of the original act and not an intervening event. The Claimant remained qualified to receive those payments and the Respondent benefits from that receipt. The Tribunal had found that the Claimant would initially return to work on a part time basis and at a lower level than would have been the case had the discriminatory events not occurred. No account had been taken of the fact that even if the Claimant was to reach earnings of £65,000 per annum gross (the amount received by the Claimant in PHI payments at the time of the Tribunal hearing) this would still be around £11,000 per annum less than his projected PHI payments by 2020, the point at which the Tribunal found he would be earning £65,000 per annum (paragraph 50). What the Tribunal had failed to take into account was that the PHI payments were increasing by average RPI. There was material before the Tribunal confirming a 2.1% increase in 2014 and a 2.2% increase in 2015. Projecting that forward the Claimant's salary in 2020, if still dependent on PHI payments, would be £86,103 gross (75% of total salary). The Claimant's schedule of loss had given the relevant figures at 50%.
17. It was contended that the Tribunal, having correctly accepted that it was the Claimant rather than the Respondent who had "paid" for the additional 25% PHI protection then perversely went on to penalise the Claimant for seeking to reap the benefit of his decision to take out such cover. In making a finding of novus actus interveniens the Tribunal had effectively concluded that the Claimant would be acting with a high degree of unreasonableness were he to take the view that it was in his best interests to remain on PHI due to the likely drop in his income and other uncertainties that may arise were he to leave that his actions would be so unreasonable to break the chain of causation. The Tribunal's finding in this respect flew in the face of fairness and so was perverse.

The Claimant's Cross-Appeal

18. Having already addressed the Claimant's cross-appeal in relation to the Tribunal's findings on cessation of the PHI payments, Ms Hart confirmed that three grounds of the Claimant's grounds of appeal would not have been necessary if the Tribunal had not made the finding that it did on termination of PHI. These aspects were the length of the Claimant's recovery period, the level of his future earnings and the risk of relapse. These three grounds were addressed first on the basis that they mattered only if the Claimant did not remain in receipt of PHI payments.
19. The first of these three challenges was to the Tribunal's finding that the Claimant's recovery would commence from the date of the remedy hearing. Counsel submitted that there was no evidence to support that finding, or in the alternative it was perverse one, on the available evidence. The starting point was the Tribunal's acceptance that the Claimant's condition at the remedy hearing was "clearly serious" and that he was not capable of any employment at that time. Indeed the Tribunal noted (at paragraph 16) that the Claimant's condition had worsened with the progress of the litigation. Notwithstanding those findings the Tribunal went on to make the following finding:

"17. Viewing the case as at September 2015, the Tribunal found it likely that the Claimant would make progress and that in about six months time, as suggested by Dr Isaac, he would be able to start thinking about returning to work. It is important to note that this is the point at which the Claimant could start thinking about matters, rather than taking action. We concluded that it would be realistic in terms of the timescale to find that the Claimant would return to some form of employment about 12 to 18 months from September 2015. Taking the mid point of around 15 months, this would be around the beginning of January 2017."

20. It was submitted that the findings in paragraph 17 were inconsistent with the Tribunal's earlier conclusion at paragraph 14 that it would assume that the Claimant will be satisfied with the outcome of the proceedings. The context of this is explained in that paragraph as follows:

"14. In submissions Ms Hart conceded that Dr Isaac's evidence on the prognosis was more convincing than Dr Owino's. She pointed out that this prognosis was predicated on the Claimant being satisfied with the outcome of these proceedings, and that in his oral evidence Dr Isaac had said that it was impossible to say whether the dispute would be resolved to the Claimant's satisfaction. The Tribunal concluded that it should assume that the Claimant will be satisfied with the outcome of the proceedings on the basis that if he is not that would amount to a separate cause from the acts of discrimination in respect of which he is entitled to compensation."

As the Tribunal decision was not handed down until October 2016 that would be the earliest date from which the Claimant could commence recovery. Taking a point six months from September 2015 presupposed that the Claimant would begin to recover before he knew of the outcome of the proceedings. Secondly, the Tribunal purported to rely on the evidence of Dr Isaac but that evidence was clearly to the effect that recovery would not begin until resolution of the whole dispute, not just the litigation. References were made to parts of Dr Isaac's evidence on which that submission was based. Ms Hart contended that there was no evidence from Dr Isaac that the Claimant would recover within three to six months of the remedy hearing in circumstances where the dispute had not been resolved or the litigation process completed. It was arguable that the litigation process was still ongoing at the time of appeal which would take the commencement of recovery to a much later date than six months after October 2016 when the Tribunal's Judgment was issued. The important distinction was between the end of the hearing in September 2015 and the resolution of the whole dispute. As the evidence only pointed to recovery commencing at the latter stage there was no evidence for the Tribunal's conclusion on the period of recovery which was in any event perverse.

21. Turning to the Tribunal's finding that the Claimant would earn £65,000 per annum by 2020, Ms Hart submitted that there was no evidence to support that finding or alternatively that it was perverse. Having found that the Claimant would have been made redundant in April 2013, the Tribunal had gone on to explore and make conclusions on the Claimant's likely earnings after that redundancy. It found (at paragraph 33) that a generic trainer would typically have a salary of around £35,000 and a training manager one of between £45,000 and £70,000 gross. It concluded that but for the discrimination suffered, the Claimant would have been given notice in April 2013 and would subsequently have found alternative employment at a gross salary of around £65,000 per annum soon after that - paragraph 37. The Tribunal's reasons for that conclusion included the following:

"41. … Ms Hart's approach involved the Tribunal finding that the Claimant would never recover from the effects of the discriminatory acts to an extent that would enable him to achieve the sort of earnings that he might have done had they not occurred. That also seemed to the Tribunal to be an unlikely scenario. We have found that the Claimant is likely to be able to return to some form of work as from around January 2017. The discriminatory acts have had a serious effect on him and it seems likely that, to some extent at least, the impact of those events will be with him for a very long time, if not for the rest of his life. However if, as we find he will, the Claimant moves into other employment, the likelihood is that doing so will enable him to come to terms with these events and at least so far as his earning capacity is concerned, put them behind him. This is not something that is specifically canvassed in the medical evidence but it seemed as a matter of logic to the Tribunal to follow from the prospect of a return to work."

It was noteworthy that the Tribunal concedes in its own Judgment that there was no medical evidence to support this conclusion. It was said to be a matter of logic but counsel pointed out a general conclusion that someone would return to work was very different from making findings on what work they would undertake and what they would earn. The specific issue of the Claimant's earning capacity on any return to work was not canvassed by the medical evidence. The two experts, Dr Isaac and Dr Owino, both said in evidence that the Claimant could not go back to his previous role. It was their evidence that if he worked it would be at a lower level of skill and remuneration than his previous employment. Dr Owino gave specific examples of the types of work that the Claimant would not be able to perform, including those that were client facing and involved achieving deadlines or other pressurised roles. Dr Isaac's evidence was similar and he expressed the view that the Claimant would be unlikely to thrive in a closely managed environment where he would be required to interact as an active team player. At its highest Dr Isaac's evidence was to the effect that the basic training elements of the Claimant's role, referred to as "generic things" were not closed to him but he said also that not only would the Claimant not work at his previous level but he would probably not work in the same industry as before. Further, he expressed the view that any form of employment would require to be a loose arrangement. On that basis, there was no foundation in the evidence for the figures adopted by the Tribunal at paragraph 50 where the conclusion was reached that the Claimant's likely earnings would be £20,000 per annum gross in 2017, £35,000 gross in 2018 and £50,000 gross in 2019 and rising to £65,000 in 2020 as previously indicated. The Claimant had made submissions in relation to the likely future level of any work that could be carried out by the Claimant and the Tribunal appeared to have misunderstood those. In addition to there being no foundation for the findings, they were perverse and illogical.

22. The third ground of appeal (ground 7) that was intertwined with the issue of the continuation or otherwise of the PHI payments related to the chance of a relapse. Ms Hart submitted that the Tribunal had failed to take into account the chance of a relapse in the Claimant's condition in determining future loss of earnings. The starting point for this challenge was the Tribunal's finding (at paragraphs 18 and 19) that, on the basis of medical evidence, the Claimant had a "relatively high risk of occurrence of 50% or more" of a relapse. It was specifically acknowledged in that paragraph that this risk would impact on the calculation of future loss. The Tribunal stated in terms that:

"19. … the best approach to the risk of relapse was to reflect this in the period of future loss which we will explain below. …"

Notwithstanding that paragraph 19 foreshadowed a future explanation, none was then forthcoming when the Tribunal addressed, at paragraphs 39 to 50, the whole issue of future loss of earnings. All the Tribunal did was identify the period of time that it considered the Claimant would take to return to his previous level of earnings. There were three factors that the Tribunal considered (at paragraph 19) might decrease the impact of a relapse on the period of future losses. These were expressed as follows:

"19. … the prospect that any relapse might be the result of some new intervening event, that treatment would be available, and that the risk is likely to diminish with the passage of time …"

Counsel submitted that a relapse by definition was connected to the sequence of events arising out of the discriminatory acts. Accordingly, the 50% risk of relapse was both foreseeable and intrinsic to the original cause. Had the Tribunal wanted to factor in unexpected events this would normally be dealt with by a way of discount for contingencies. Secondly, on the issue of available treatment, the finding of 50% or more risk of relapse was based on the opinion of a medical expert and there was no suggestion in evidence that a figure for risk did not take into account the possibility of available treatment. Thirdly, on the likelihood of risk diminishing over time, again the finding of a 50% or more chance of relapse encompassed that there was up to 50% or less chance of no relapse. Accordingly the chance of a relapse never taking place had already been taken into account and it would be double counting to reduce the 50% or more risk to take into account any diminution over a period of time. A specific finding having been made on the basis of medical evidence that there was a 50% chance at least of a relapse, there was no logical basis for taking anything other than the full 50% chance into account in determining future losses. The Tribunal had erred in failing to do so or alternatively had failed to explain how that risk was taken into account when determining future losses at paragraphs 39 to 50 of the Judgment.

23. Turning to the grounds of appeal that stood alone regardless of the issue with the continuation or otherwise of the PHI payments, the first two of these were taken together and amounted to a submission that the Tribunal was wrong to find that some or all of the identified non-work related elements and work related but non-discriminatory factors were causative of and/or contributed towards the Claimant's condition. It was accepted at the outset that there was evidence about these matters on which the Tribunal made findings and so it would normally be difficult to interfere with those on appeal. In any event, the argument would become academic if a finding in relation to apportionment (the 70%/30% distribution as between discriminatory and non-discriminatory factors) was perverse which was a separate ground of appeal. In any event, the argument relating to this first aspect was that the law made clear that the first step in establishing causation was to eliminate irrelevant causes (Clerk & Lindsell on Torts 21st Edition paragraph 2-09). The purpose of the exercise was not to identify all possible causes but make findings on the effective causes of the resulting damage in order to assign responsibilities. The Tribunal had failed to do this and so had expressed itself inappropriately by stating:

"29.1. Due allowance had to be made for the impact for the non-work related, and the work related but non-discriminatory elements that contributed to the Claimant's condition. The effect of these was not negligible."

Ms Hart submitted that it was difficult to glean from the Tribunal's reasoning why this conclusion had been reached in light of earlier findings on the impact of non-work related and work related elements. In summary, at paragraph 22 the Tribunal had identified three main non-work related elements of the Claimant's condition, namely bullying and developmental difficulties at school, history of alcohol dependence and the Claimant's daughter being diagnosed with autism together with the death of the Claimant's mother in 2009. The first of these should have been considered an example of a pre-existing vulnerability rather than a material cause. It was clear from the evidence of Dr Isaac that these historic difficulties were simply part of the background. On the history of alcohol dependence, the Tribunal had already found that both medical experts had agreed that this was "probably not a significant factor" in the development of the Claimant's condition (paragraph 12). Accordingly, it should not have been included in consideration of causation. So far as the third matter is concerned, the Tribunal had failed to explain how the daughter's diagnosis and/or the death of the Claimant's mother contributed to the Claimant's condition given that Dr Isaac had not considered that they had had a major emotional effect on the Claimant and had not precipitated undue anxiety or low mood (at paragraph 22). So far as the non-discriminatory work related elements were concerned the Tribunal had identified the restructuring of the workplace (paragraph 23), the liability findings on specific incidents (paragraph 24) and the liability findings on failure to make reasonable adjustments (paragraphs 24, 106, 107, 111 and 112). In essence it was submitted that none of the non-work related or work related but non-discriminatory elements identified by the Tribunal had caused or contributed to the development of the Claimant's condition. There was considerable overlap between these submissions and the ground of appeal challenging the specific apportionment of 70%/30% which Ms Hart contended was perverse. Accordingly, she focused on that important matter of apportionment.

24. Ms Hart submitted that the primary problem with the decision of the Tribunal on apportionment (at 70/30%) is that the Tribunal had failed first to determine whether the injury was divisible or not before apportioning the loss. This was the approach required as confirmed in Olayemi v Athena Medical Centre and Another [2016] ICR 1074 and cited with approval in BAE Systems (Operations) Ltd v Konczak [2017] EWCA Civ 1188. In Konczak the Court of Appeal had specifically addressed the question of whether psychiatric harm was in principle divisible and expressed the following view:

"71. What is therefore required in any case of this character is that the tribunal should try to identify a rational basis on which the harm suffered can be apportioned between a part caused by the employer's wrong and a part which is not so caused. I would emphasise, because the distinction is easily overlooked, that the exercise is concerned not with the divisibility of the causative contribution but with the divisibility of the harm. In other words, the question is whether the tribunal can identify, however broadly, a particular part of the suffering which is due to the wrong; not whether it can assess the degree to which the wrong caused the harm."

An injury which is indivisible cannot be apportioned and so it was wrong for the Tribunal in this case simply to refer back to its findings on contribution in assessing apportionment. In essence the Tribunal had failed to address the matter required by Konczak namely whether there was a "rational basis for distinguishing between a part of the illness which is due to the employer's wrong and a part of the illness which is due to other causes". Ms Hart accepted that there should be apportionment unless the Tribunal found that such a division was not possible. The Tribunal had erred in its assumption that the causative contribution of 70% automatically resulted in harm being divided into 70% attributable to the employer and 30% to other causes. Whether any division was possible would always depend on the facts identified in evidence.

25. Counsel went on to submit that the findings of fact made by the Tribunal supported by the medical evidence of both experts in the case illustrated that the injury to the Claimant was in fact indivisible. Reliance was placed on the fact that the Claimant suffered a breakdown at work on 19 April 2012 since when he has not been able to return to work; Dr Owino stated that the various discriminatory acts were "significant precipitating cause" of the development of the Claimant's symptoms; Dr Isaac gave evidence that it was not possible to ascribe cause and effect to the individual instances that the effects on the Claimant were cumulative; the Tribunal had found (at paragraph 22) that it was not possible to attribute any particular element of the Claimant's condition to any particular event and that the Tribunal had also found in that paragraph that the Claimant's condition was a "single" one, the totality of which had been contributed to by various factors. Accordingly, there was a sufficient basis to substitute a finding that the Claimant's injury was indivisible. It was acknowledged, however, if that there was more than one possible conclusion on this from the evidence there would require to be a remit back to the Tribunal.
26. In anticipation of the arguments to be presented in response on this point, Ms Hart submitted that it would it be wrong for the Respondent to suggest that there is a presumption in favour of apportionment. The current state of the authorities made clear that a two stage process was required in terms of which the first question was whether or not the harm was divisible and the tone they felt was answered in the affirmative was the second stage of apportionment required. While the Tribunal had relied on the decision in Thaine v London School of Economics [2010] ICR 1422, the recent decision in Konczak adopted a more nuanced approach. In Konczak the Court of Appeal approved Thaine only insofar as it preferred the position of Hale LJ (as she then was) to that of Smith LJ on whether psychiatric injury was divisible. That dispute was no longer relevant as it had been resolved by Konczak which stated clearly what the correct approach now is. Insofar as the Respondent was likely to argue that there were different chronological stages to the Claimant's injury, the Claimant pointed out that the Tribunal's reconsideration decision on liability addressed this at paragraphs 24, 26 and 28. The Claimant's family history was not relevant to the issue of the cause of his current mental health condition. There was no evidence of any pre-existing mental health problem or of the difficulties the Claimant had suffered in childhood making any difference to his mental state at all. There had been clear evidence from Dr Isaac in particular that the brief episode during his childhood that had been disclosed would have required no more than a few days sick leave.
27. As a preliminary to his submissions in relation to the Claimant's appeal Mr Jeans QC emphasised that, while the Respondent had limited its appeal to issues of law notwithstanding disagreement it might have with the Tribunal's Judgment on issues such as the level of apportionment the findings on future earnings, the Claimant had taken a different approach. Central to the Claimant's various arguments about the conclusions reached by the Tribunal were that they ignored that the assessment of compensation was necessarily "rough and ready" and that assessments on future loss are necessarily speculative. It was well established that it was for the fact finder to make such assessment having regard to the primary facts of the Judgment and the Tribunal's experience.
28. While it was common ground that the Tribunal had raised matters that had not been anticipated by the parties in relation to the Claimant not being able to rely on the PHI payments, that did not automatically mean that the conclusions in paragraph 46 were wrong. The parties were agreed that paragraph 46.3 was wrong in that the Tribunal should not have found that there was a date after which the Claimant could not rely on PHI payments. However, the finding that once well enough it would be beneficial for the Claimant's health to return to work coupled with the findings that he was committed to working and well-motivated were perfectly consistent with the evidence. Unless it could be said that on the available evidence the Tribunal was not entitled to reach the views that it had in relation to when and at what level the Claimant would return to work, the decision on these aspects could stand. The specific figures and likelihoods expressed by the Tribunal might seem high or low but, as the Court of Appeal had confirmed in Vento v Chief Constable of West Yorkshire [2003] ICR 318 at paragraph 39 a broad and impressionistic prediction made by the Tribunal on the basis of the evidence is likely to stand. The premise in this case that the Claimant will not be able to rely on PHI payments was an unnecessary one for the overall conclusions the Tribunal reached about when the Claimant would return to work and at what level. It was well-established that the test for perversity was a high one and the Claimant in this case had transgressed the boundary between identifying errors of law on the one hand and selecting lines of evidence on which to mount a critique, ignoring that the Tribunal would have taken all of the evidence into account.
29. Turning to the specific grounds of the Claimant's appeal, Mr Jeans submitted that the Claimant's arguments against the Tribunal's finding as to the timing of the Claimant's recovery were fallacious because they constituted an unfair reading of the Tribunal's reasons, involved a mechanistic reading of the evidence and ignored the importance of the Tribunal's fact finding powers. The Tribunal's narration of Dr Isaac's evidence as being that in three to six months the Claimant's condition would progress such that he would be able to think about the future was deliberately and consciously vague. The Tribunal could not say with certainty when the Claimant would return to work. The end of the oral hearing in September 2015 was on any view an important stage for the Claimant. It could not seriously be contended now that that it had not been stressful for the Claimant to attend the hearing. Special arrangements on timing and other matters had been made to alleviate that stress and the Claimant had required a break in proceedings when other witness's evidence caused him further stress. Within the bundle of medical notes that had been provided for the appeal there was a record by his community psychiatric nurse about the Claimant's distress and suffering during his giving evidence at the liability hearing in 2014 and that he had required to take Diazepam. It was clear to the Tribunal at the remedies hearing that attendance there caused the Claimant stress. That was what Dr Isaac's evidence was about when he said that the problem for the Claimant was the process and everything to do with it. The oral hearing was clearly a big part of that. It was noteworthy also that Dr Isaac was not asked to explain what was meant by the whole dispute and so nothing could be taken from that beyond the difficulties for the Claimant with the oral hearing. Dr Isaac had been specifically projecting from the time of the hearing not any anticipated Judgment. It was important also that the Claimant had unsuccessfully sought reconsideration on the basis that the recovery date should start from the end of the whole dispute. It was the end of the active litigation process that was important. Once the Tribunal had accepted that recovery was a staged process, it was a matter of fact and degree for the Tribunal to decide whether a stage had been reached. As there was evidence to support the finding made by the Tribunal on period of recovery and accordingly no basis on which it should be interfered with.
30. On the issue of the level of salary that the Claimant would eventually obtain, senior counsel contended that the Claimant's suggestion that the Tribunal had ignored the relevant evidence on this was unfair. While the Tribunal does rely on a logical progression following from the Claimant's return to work at paragraph 41, properly read the Judgment was not indicating that there was no medical evidence in relation to a return to work, it was more that a common sense view had to be taken on how the Claimant would gradually improve following that return. Ms Hart had referred to Dr Owino's evidence in support of this argument but it was common ground that Dr Isaac's evidence on the prognosis was more convincing than Dr Owino's - specifically conceded by Ms Hart and recorded at paragraph 14 of the Judgment. On that basis it was meaningless now to refer to Dr Owino's evidence. The Tribunal had, using the very wide latitude available to it, made a prediction on how long it would take the Claimant to return to a certain level of work. At paragraph 35 there is a specific acknowledgement that the Claimant would struggle with certain roles. It had evidence from Mr Cross as to the range of functions and responsibilities which could command an income around the £65,000 per annum level. There was specific evidence that such a salary could be obtained that did not involve management responsibilities, possibly even with the Respondent. The expertise of the Tribunal members was important as their industrial experience and insight were able to be applied. They could add something that medical professionals not expert in the market were unable to bring to this exercise of legitimate guesswork. The Tribunal had before it the Claimant's impressive curriculum vitae which also had to be fed into the equation. The Claimant's assertion was that the conclusion of the Tribunal was wrong but the possibility of a different view being taken could not equate to error on the part of the Tribunal. There was no question of the Tribunal's prediction being irrational. The exercise that had been conducted was a carefully calibrated set of conclusions taking everything into account, including the current salary of the job presently held by the Claimant.
31. In response to the Claimant's arguments in relation to the risk of relapse, Mr Jeans accepted that while the Tribunal had stated at paragraph 19 that it would factor the 50% risk of relapse into future loss, the Judgment did not specifically state that again at paragraph 50. However, there was no reason to assume that the risk had not been factored into the overall assessment of the period of future loss. The approach taken by the Tribunal was in line with that indicated by the Court of Appeal in Wardle v Credit Agricole [2011] IRLR 604. There the Court of Appeal had specifically reversed a Tribunal compensation decision which had made separate allowance for a contingent risk that the Claimant might not ultimately recover his career (in banking) from which a discriminatory dismissal had diverted him. Elias LJ had referred to the "usual approach" of:

"51. … assessing the loss up to the point where the employee would be likely to obtain an equivalent job, does fairly assess the loss in cases - and they are likely to be the vast majority - where it is at least possible to conclude that the employee will in time find such a job. …"

The Court of Appeal found that the Tribunal in that case had been wrong to make a discount from the anticipated earnings to reflect the risk that the Claimant would not obtain a banking job in the time anticipated. This was because an assessment by the Tribunal that a Claimant is likely to secure an equivalent job by a specific date will normally encompass the possibility that he might be fortunate and secure such a post earlier or unlucky and find the job later than predicted. The Tribunal's best estimate of what was likely ought to provide appropriate compensation without making separate percentage allowances for risk. In Wardle, the Court of Appeal had reassessed compensation taking a median date for alternative employment which reflected contingencies. That was essentially what the Tribunal had done in this case, although adding the prediction of a staged recovery. Such an approach contains no error of law. Indeed it would have been arguably wrong for the Tribunal to adopt a more complex formulation and to make a mathematically separate computation to reflect the possibility of relapse.

32. On the issue of the divisibility or otherwise of an injury, Mr Jeans submitted that it was plain from the authorities that only in an exceptional case is an injury likely to be truly indivisible and that the normal course was to attempt a "rough and ready" apportionment. Particular reference was made to Thaine v London School of Economics [2010] ICR 1422 at paragraphs 17 to 23, an EAT decision confirming that an employer should only be liable to the extent that the discrimination has contributed to the harm suffered, in this case the Claimant's ill health. In Thaine, the employer's liability was discounted by 60% as the discrimination had contributed to the Claimant's ill health only to the extent of 40%. The case of Sutherland v Hatton [2002] ICR 613 was authority for the proposition that sensible attempt had to be made to apportion liability where a constellation of symptoms suffered by a Claimant stemmed from a number of different extrinsic causes. Sutherland was cited with approval by the Court of Appeal in Konczak (at paragraph 59). The case of Konczak did not alter the basic proposition that apportionment was normally appropriate unless there was no basis for it at all on the evidence. The Court of Appeal had specifically stated (at paragraph 56) that injury would only be treated as indivisible "where there is simply no rational basis for an objective apportionment of causative responsibility". The chronology was important in this case. The 30% apportioned to non-discriminatory factors related to events prior to any acts of discrimination and had caused harm. It would have been perverse for the Tribunal not to make some apportionment findings. In a situation where the early stages of the Claimant's injury predated any discriminatory acts they could easily be separated out and so not considered indivisible. In a case such as this where anxiety is part of the Claimant's condition, the causes are more easily attributed to individual events as the Claimant had recited a number of things that had happened and said how he was affected by them. The more difficult case might be one of a psychotic disorder when problems of divisibility of injury can more readily be seen. In a situation where the Respondent is only liable for the extent of its contribution against the background of a requirement to apportion unless the harm is truly indivisible, the Claimant in this case could not show that there was no rational basis for the apportionment. Whether formally approached as a one stage or two stage enquiry, the Tribunal could only be said to have erred if its conclusion was inconsistent with the existence of separate harm. The summary of the events of 2009 (death of the Claimant's mother and his daughter being diagnosed with autism) and the restructuring in 2011 that had marked the beginning of the Claimant's anxiety and depressive symptoms set out in paragraph 8 of the Tribunal's Judgment provides reasonably clearly the basis for the apportionment. It was not in dispute now that the first discriminatory act took place in about November 2011 in the form of a failure to make a reasonable adjustment. The Claimant's symptoms insofar as arising from non-discriminatory events had crystallised before then. In fact in the grievance procedure, the Claimant had been keen to point out that his mental health decline had started in November 2010, as evidenced by an additional document provided at the appeal hearing. There was ample evidence to support the division of the different symptoms suffered by the Claimant and attribute them to different stressors. It was simply unanswerable that there was an element of harm for which the Respondent could not be held responsible in this case, the specific apportionment being a matter for the Tribunal. While it might be argued that paragraph 29 of the Tribunal Judgment and its reference to assessing the contribution of the discriminatory acts and to the Claimant's overall condition did not sit happily with paragraph 71 of the case of Konczak, counsel submitted that so long as there was sufficient evidence for the overall result achieved by the Tribunal to be permissible on that evidence there would still be no basis for interfering with it. Reading the Judgment as a whole it was apparent that the damage or harm to the Claimant was only partially caused by the discrimination. That is clear from paragraphs 8 to 10 of the Tribunal Judgment which forms the basis for the ultimate apportionment. The outcome of apportionment was wholly consistent with the decisions in Thaine and in Konczak and it would have been perverse to make no such apportionment. The Tribunal had been entitled to take a broad percentage view in determining the specific apportionment as between harm caused by the discrimination and otherwise. While the Respondent might also take issue with the specific percentage that was not an appropriate matter to be raised on appeal.

Discussion

The Respondent's Appeal

33. The first of the issues for consideration in the Respondent's appeal raises the important question of whether there should be a 50% or 75% deduction from the Claimant's loss of earnings in respect of PHI payments. This was identified as issue 8 in the Decision, dealt with at paragraphs 52 to 57 inclusive and set out earlier in paragraph 4 of this Judgment. The general law in relation to the principle of mitigation of loss including how it applies to insurance payments of this type, was not in dispute. A claimant cannot recover damages in respect of avoided loss with two clear exceptions, described as first the benevolence and secondly the insurance exceptions - Atos Origin v Haddock [2005] ICR 277 per Mitting J at paragraph 29. The insurance exception would apply to a case where the employee has taken out insurance and paid the premiums for it as he should not be deprived of the benefits of the insurance for which he has paid. The most recent authoritative statement of the general rule on insurance exception is that provided by the Court of Appeal in Gaca v Pirelli General plc [2004] 1 WLR 2683. It may be helpful to set out the relevant passages in full as they formed the basis for both sides' submissions and are in the following terms:

"54. Mr Foy's second submission is that the claimant should be treated as having paid or contributed to the premium simply by virtue of the provision of labour pursuant to his contract of employment. A similar argument was rejected by this court in Hussain [1987] 1 WLR 336, 344G-H. Lloyd LJ said, at p 345, that the evidence did not support the conclusion that the plaintiff would have received more pay but for the insurance. He continued:

"In truth the judge was, I think, resting his conclusion on a broader ground. Even if the plaintiff's wage would have been the same, he has nevertheless earned the benefits payable under the scheme by working for the defendants. As [counsel for the plaintiff] put it, in language adopted by the judge, the benefits are part of the wage structure. The difficulty I feel with that argument is that it would apply equally to sickness or injury benefit paid during the first 13 weeks of incapacity. It was never suggested that this payment should be left out of account. Yet those payments were 'earned' in exactly the same way as the subsequent payments."

55. Kerr LJ dealt with the point at p 351D-E. He too concluded that the moneys were to be deducted on the grounds that there was no evidence that the plaintiff's wages would have been higher but for the insurance scheme. Lloyd LJ's way of dealing with the broader ground was endorsed by Lord Bridge in Hussain [1988] AC 514, 529-530. Mr Foy's submission is also inconsistent with what I said in Page: see para 47 above.

56. It follows that an employee is not to be treated as having paid for, or contributed to the cost of, insurance merely because the insurance has been arranged by his employer for the benefit of his employees. The insurance moneys must be deducted unless it is shown that the claimant paid or contributed to the insurance premium directly or indirectly. Payment or contribution will not be inferred simply from the fact that the claimant is an employee for whose benefit the insurance has been arranged. There is little guidance in the English cases as to what is sufficient to constitute evidence of indirect payment or contribution. The issue has, however, been discussed in a number of Canadian cases, most notably in Cunningham v Wheeler [1994] 113 DLR (4th) 1. The majority decision was given by Cory J, who said, at p I5, that what was required was:

"that there be evidence adduced of some type of consideration given up by the employee in return for the benefit. The method or means of payment of the consideration is not determinative. Evidence of a contribution to the plan by the employee, whether paid for directly or by a reduced hourly wage, reflected in a collective bargaining agreement, will be sufficient."

57. He then gave a non-exhaustive list of possible examples of the sort of evidence that could well be sufficient to establish that the employee had paid for the benefit. In her dissenting judgment, McLachlin J said, at pp 38-39, that she regarded this approach as likely to to generate uncertainty. She preferred to hold that there was a general rule of deduction, subject only to exceptions for charitable payments, and non-indemnity insurance and pensions. If the payment was in the nature of an indemnity, then it should be deducted to prevent double recovery, regardless of whether the claimant had contributed to the cost, unless it was established that a right of subrogation would be exercised.

58. The approach adopted by Cory J is similar to that which Lloyd and Kerr LJJ had in mind when they concluded that there was no evidence that the wages benefits in Hussain had been paid for directly or indirectly by the plaintiff on the facts of that case. Similarly, my own judgment in Page.

59. Turning to the facts of the present case, Mr Foy cannot identify any evidence which shows that the claimant paid or contributed to the cost of the insurance policy. All he can point to is the fact that the fruits of the claimant's labour enabled the defendants to pay for the insurance. But for the reasons that I have given, that is not enough to avoid the deduction of the benefits from his damages."

34. It is accepted in this case that the Respondent employer paid all the premiums for the relevant insurance. It is also accepted that there is no contract between the Claimant and the insurance company. The sole question for the Tribunal was whether, on the evidence before it, the Claimant should be characterised as someone who had made an indirect contribution to the cost of the insurance policy premiums by electing to have more than the minimum 50% salary protection and opting for the full 75% cover. In short, the Respondent contends that in categorising the Claimant as someone who had made an indirect contribution the Tribunal went beyond the Court of Appeal authority in Gaca v Pirelli and relied on decision from the Canadian courts in Cunningham v Wheeler which could be distinguished on the facts of the present case. The Claimant on the other hand contends that the Tribunal did not innovate on the current authorities and was correct to categorise this as a situation of the employee making indirect contributions to the premiums.
35. It seems to me that the rationale used by the Tribunal for its conclusions is both rational and consistent with the leading authority in Gaca. On the available evidence, had the Claimant elected to take either 50% salary protection or 60% salary protection he would have received additional salary. Although the Respondent's flexible benefit scheme involved a choice on a range of other matters such as annual leave and type of pension provision and so on, there was no dispute that, in order to receive 75% cover rather than 50% cover, the Claimant was paid less than he would otherwise have received. The Tribunal had before it an example of the salary and flex fund information sheet relating to the Claimant from a period in 2011-2012 together with a sample pay slip and these documents were also available to me on appeal. What they illustrate is that by making choices in both core benefits and voluntary benefits an employee such as the Claimant could make a number of deductions from the "flex fund" provided by the employer. The value of the various benefits chosen was deducted from the total benefit allowance such that a final "flex fund balance" figure is given for the relevant period. On the Claimant's payslip, this translates into a payment from the employer which is described as "flex adjustment". The level of the flex adjustment payment is directly related to the choices made by the employee which are all expressed in pounds sterling on the salary and flex fund information document. The amount to be paid to the Claimant as a flex adjustment in his payslip is directly related to the choices he made in connection with the available benefits. That was the evidential basis for the Tribunal's statement that the Claimant could have chosen to take additional salary instead of the additional 25% PHI protection. The question then becomes whether, it was some sort of impermissible leap to go from that to categorisation of the Claimant's choice to take the maximum cover of 75% as tantamount to him making an indirect contribution to the premiums himself. Senior counsel for the Respondent relied on the fact that the Claimant had at all times kept the choice on permanent health insurance to the default position of 75% of base salary and so had effectively done nothing to choose something different to that arranged and paid for by the employers. In my view the Tribunal was correct to conclude that it makes no difference in principle whether the default position was set out at 50% or at 75% protection. The undisputed facts illustrated that the Claimant chose to receive lower salary in return for increased protection. I conclude that even if the examples given in the Canadian decision of Cunningham v Wheeler are ignored and the approach is simply that confirmed in Gaca v Pirelli as being whether the Claimant has proved an indirect payment or contribution to the premiums, then on the basis of the material before it, the Tribunal was entitled to regard the evidence in this case as sufficient to constitute such an indirect payment. It is not the case that, as was contended on behalf of the Respondent, the Claimant did nothing other than choose from a menu of benefits. He chose to receive less salary than he otherwise would have been paid so that he could secure the maximum 75% insurance cover. On the basis of Gaca v Pirelli, I conclude that it was wrong for the Respondent to contend that the payment of the premiums by the employer is effectively determinative as that ignores the indirect contribution aspect of the exception. While the particular facts that the Tribunal had to grapple with might have been novel, the legal principles were already established and I reject the argument that the Tribunal innovated on or deviated from them. That disposes of the first of the Respondent's grounds of appeal.
36. Turning to the issue of the Claimant's potential redundancy in April 2013 this is listed as issue 4 for the Tribunal's decision and its decision is summarised at paragraphs 31 and 32 of the Remedies Judgment. The issue relates to both redundancy and notice payments. The Tribunal awarded 70% (consistent with the general apportionment of 70%/30% throughout) of the redundancy payment the Claimant would have received in April 2013 together with a payment in lieu of notice. There was undoubtedly evidence that, had he not been absent from work through illness, the Claimant would have been in a small pool of candidates to be made redundant. The Respondent offered evidence that there was a near certainty that that would have occurred in April 2013. The central issue on this point is whether he can be regarded as having lost a sum representing the redundancy payment he would have received together with payment for the requisite period of notice. The argument about this arose only after the remedies hearing in the sense that the Claimant had not included it in his schedule of loss but the Respondent had presented the argument that he would have been made redundant at the hearing with a view to using that as a cut-off point. It then became the Claimant's contention that the redundancy and notice payments related to the issue of what the Claimant would have earned in the period prior to 2015 which was the first stage of a two stage process in which the deduction of the Claimant's likely earnings in consequence of the discrimination (including any PHI deductions) would be made. It seems to me that, while it is indisputably correct that the Tribunal required to approach the question of loss as one in which the Claimant should be put in the position he would have been had there been no discrimination, that approach is tempered by the general rule already referred to that the Claimant is not entitled to be compensated for avoided loss. On the accepted facts, the Claimant has thus far avoided redundancy. The evidence was that the Respondent would not contemplate redundancy while he was absent from work through illness and that remained the position at the time of the Tribunal's determination. Self-evidently, the Claimant has not lost a redundancy payment or even the chance of securing one. If he is made redundant in future by the Respondent he would require to receive redundancy and notice payments in accordance with his contract and the applicable law. That conclusion is not inconsistent with the evidence that had the Claimant not been absent from employment in April 2013 he would probably have been made redundant. The net redundancy payment he would have received would have compensated him to some extent for the loss of his job. The Tribunal found that the Respondent was at fault and that as a result (ignoring the apportionment issue for present purposes) the Claimant suffers from a mental health condition. He has incurred and continues to incur losses as a result. One loss that he has not incurred is the termination of his employment with the Respondent, through redundancy or otherwise. While the argument that, but for the discrimination, he would have been made redundant (and so should benefit from the payment that would have been made to him) is superficially attractive, it ignores the nature of redundancy and notice payments which are claims arising from the Claimant's employment with the Respondent that he still has. To award the Claimant a sum for a loss that has not yet happened but could still happen would be double counting to the extent that the Respondent would have to pay damages under this head while continuing to be obliged to make the appropriate payments for redundancy and notice in the future. Accordingly, I consider that the Respondent's submissions in relation to the second ground of appeal are to be preferred. The Tribunal erred in concluding that a logical consequence of accepting the evidence that the Claimant would almost certainly have been made redundant in April 2013 was to award him sums representing the net redundancy payment he would have received together with three months' pay in lieu of notice. These are sums that are not lost to him and which he may yet receive. This case is to be distinguished from one of discriminatory dismissal where, but for the dismissal, the Claimant would have been made redundant. The Claimant continues to be employed by the Respondent with all the rights and obligations that flow from that. Had I considered that the Tribunal was correct in taking into account the Claimant's likely redundancy had he been fit for work in April 2013, I would in any event have accepted the Respondent's submission that the Tribunal could not properly have awarded more than a figure for the loss of a chance of receiving redundancy and notice payments under this heading as that would be the most the Claimant could be regarded as having lost in the particular circumstances.
37. While the Respondent's third and final ground of appeal was uncontentious as between the parties but for different reasons, it merits some consideration. As indicated, the Tribunal addressed the issue of the period during which the Claimant should be compensated in some detail at paragraphs 42 to 49 of its Judgment. As part of that it concluded that the Claimant would not be able to rely on the PHI payments being made indefinitely. The reasons given for that conclusion amount, in my view, to speculation on the part of the Tribunal for which it is agreed there was no basis in the evidence. In fact there was direct evidence to the contrary. In particular, it was an error for the Tribunal to conclude that the insurer would somehow be in a position to insist that the Respondent employ the Claimant during any period for which the medical evidence indicated that he should remain on sick leave. The insurers have no contractual power to do that or to withhold payments having accepted the claim on the basis of the Claimant's inability to perform the duties of his employment. There is also authority in support of the proposition that when an employee is benefiting from an employer's permanent health insurance scheme that is dependent on the continuance of the employment relationship, it is an implied term of the contract of employment that the employers will not terminate the contract while the employee is incapacitated for work - Aspden v Webbs Poultry and Meat Group (Holdings) Ltd [1996] IRLR 521. Of course the Tribunal was required to assess separately, as best as it could, when the Claimant would recover and return to some form of employment. The error was in linking that issue to the cessation of employment payments beyond the obvious logical step of recording simply that the PHI payments would cease after the Claimant recovered sufficiently to work either with the Respondent or with any other employer. In other words, it was wrong to conclude that some knowledge of not being able to rely on the PHI payments indefinitely would be a precipitating factor for the Claimant's return to work as there was no evidence on which that conclusion could be based. Further, there was simply no basis in the evidence for the Tribunal's statements about a new intervening cause of loss and the established law on novus actus interveniens was not relevant to the Claimant's situation. What the Tribunal required to do was assess, on the assumption that the PHI payments would be in play for so long as the Claimant was unfit to work, assess how long he might take to recover and what he would do in future. There is no proper basis for the Tribunal's reasoning on the cessation of the PHI payments and the specific finding, recorded at paragraph 46.3 that it is unlikely that the Claimant will be able to rely on those payments being made indefinitely cannot stand. The primary finding that when the Claimant is able to make a return to work he will do so is, however, unaffected by the Tribunal's error in this respect and all but subparagraph 46.3 of paragraph 46 can remain.

The Claimant's Appeal

38. The first three challenges in the Claimant's appeal relate to the Tribunal's findings (i) on the time at which the Claimant's recovery would commence, (ii) that he would earn £65,000 per annum by 2020, and thirdly on whether the 50% chance of the Claimant relapsing had properly been taken into account. There is, as indicated, a relationship between those findings and conclusions and the view of the Tribunal that the Claimant would not be able to rely on the PHI payments being made indefinitely. However, the Respondent contends that removal of that aspect of the Tribunal's reasoning does not negate the whole decision on the issues of when and on what terms the Claimant will return to work. The basis for timing of any return to work required to be made on the medical evidence. The Tribunal relied on the evidence of Dr Isaac (at paragraph 17) that, from the date of the Tribunal hearing in September 2015, it would be about six months before the Claimant could start thinking about a return to work. Taking the various factors into account the Tribunal then concluded that it would be realistic in terms of the timescale for the Claimant to return to some form of employment about 12 to 18 months after that and took a mid-point such that it calculated matters on the basis that he would start working again in January 2017. There is then the graduated earnings and findings following from that in paragraph 50 where the Tribunal makes a broad assessment of what might happen in terms of the Claimant's earnings during the period 2017 to 2020.
39. On the issue of the medical evidence, it seems to me that there was a basis for the Tribunal's finding about when the Claimant would start thinking about returning to work and when he would be well enough to do so. There was no suggestion from the Claimant's side that the Tribunal was wrong to find that the Claimant was someone who had been a committed and industrious employee such that he would return to work when fit enough to do so. On that basis, the specific findings in relation to when he might return to work were a matter for the Tribunal on the evidence. While there does appear to have been a reference to the "whole dispute" in the evidence of Dr Isaac, it seems to me that the argument of the Claimant on this point rather takes that out of context. As with so many cases of this type, broad assessments require to be made. The Tribunal had to look at matters at the point on which the evidence was given because Dr Isaac was predicting forward from September 2015. There does not appear to have been any clarity about whether the process being referred to included the time to wait for a decision and possibly even appeal process. I do not consider it appropriate for an appellate Tribunal to seek to interpret an expression used by a witness where the parties disagree as to its meaning. In this case the Respondent's position is that the context of Dr Isaac's evidence was the stress suffered by the Claimant in having to participate in the oral hearing. It is sufficient for disposal of this first point that the Tribunal had, as at September 2015, medical evidence about the timescale for the Claimant's recovery and possible return to work and made a decision open to it on the evidence. I do not consider that any interference with that is justified on appeal.
40. Similarly, in relation to the specific finding that the Claimant would return to work and gradually increase his earnings such that he would be earning £65,000 per annum by 2020, Ms Hart contended that paragraph 41 of the Judgment records an effective concession by the Tribunal that there was no medical evidence to support its conclusion. I think that is an unfair characterisation, in light of the medical evidence about a return to work to which I have already referred. There was dispute in the medical evidence on the Claimant's prognosis as recorded at paragraph 14 of the Tribunal's Judgment, with Ms Hart conceding that Dr Isaac's evidence was more convincing than Dr Owino's. Having made an assessment of when the Claimant would be fit enough for work on the basis of the acceptable evidence the Tribunal then required to form a view on whether the Claimant would be able to return to a senior role immediately or whether that would take time and what impact that would have on his general earning capacity. There was evidence from a witness of the Respondent (Mr Cross) in relation to possible roles for the Claimant that would secure an income at the level that the Tribunal concluded the Claimant would ultimately attain, namely employment with a salary of about £65,000 per annum by 2020. It does not seem to me to be a question of the Tribunal misunderstanding any submission made on behalf of the Claimant; rather there was alternative evidence that the Tribunal accepted upon which it appears to have based this finding. It cannot be an error on the part of the Tribunal to make a broad assessment of what the Claimant would ultimately achieve having regard to the various possible scenarios put in evidence. The Tribunal was entitled to reject the approach predicated upon the Claimant never recovering that had been put forward by his counsel.
41. The issue of a chance of a relapse is related to this whole question of whether the Tribunal was entitled to reach the view that it did on the Claimant's future earnings. The starting point was a finding by the Tribunal that the Claimant had a relatively high risk of relapsing in the order of 50% or more (paragraphs 18 and 19). Counsel for the Claimant's argument was that the Tribunal had failed to explain how the risk of relapse was reflected when it came to assessing future loss despite having indicated that it would do so. However, it is important to note that actual decision of the Tribunal (in paragraph 19) on this issue was that it would reflect the risk of relapse in the period of future loss. It is not entirely clear to me whether the words "which we will explain below" simply foreshadow that the period of future loss will be explained later in the Judgment or, as Ms Hart contended, that there was a subsequent failure to explain how the risk of relapse was factored in. What is clear enough, however, is that the Tribunal was foreshadowing that the period over which compensation would be awarded would take into account that risk. The issue then becomes whether the Tribunal erred in its approach to the period of future loss. The relevant passage of the Judgment is in succinct terms and is expressed at paragraph 50 as follows:

"50. Making the best estimation that it could, the Tribunal found it likely that it will take a further three years from 2017 for the Claimant to reach the point where he is earning £65,000 per annum gross as he would have been but for the breakdown in his health. There is little by way of evidence to show what the Claimant's earning might be over this period of 3 years; but using the information available about the salaries available for training roles, and assuming a pattern of gradually increasing working hours, the Tribunal estimated the Claimant's likely earnings over those three years as follows:-

1. 2017: £20,000 per annum gross, giving approximately £16,000 net.

2. 2018: £35,000 per annum gross, giving approximately £27,000 net.

3. 2019: £50,000 per annum gross, giving approximately £37,000 net."

42. It is apparent from the above quoted part of the Tribunal's reasoning that it was engaged in the type of necessary speculation routinely undertaken in a case of this sort. Having accepted the medical evidence on risk of relapse and indicated that it would be taken into account in the period of future loss the Tribunal has then set out, at paragraph 50, what it considers that period to be. It was in my view unnecessary for the Tribunal to state how exactly the risk of relapse had been factored in; it said that it would do so and the estimate of the period of loss given in paragraph 50 must be taken to include this as a factor. The usual approach is not to make separate allowance for a contingent risk that a Claimant might not ultimately recover to the extent of being able to resume employment at broadly the previous level. As Elias LJ put it in Wardle v Credit Agricole [2011] IRLR 604 at paragraph 52:

"52. … In the normal case if a tribunal assesses that the employee is likely to get an equivalent job by a specific date, that will encompass the possibility that he might be lucky and secure the job earlier, in which case he will receive more in compensation than his actual loss, or he might be unlucky and find the job later than predicted, in which case he will receive less than his actual loss. The Tribunal's best estimate ought in principle to provide the appropriate compensation. The various outcomes are factored into the conclusion. In practice the speculative nature of the exercise means that the Tribunal's prediction will rarely be accurate. But it is the best solution which the law, seeking finality at the point where the court awards compensation, can provide."

In the present case, having acknowledged the 50% risk of relapse, the Tribunal balanced that against the factors which militated for and against an increase in the impact of that risk on the period of future loss. Ms Hart was critical of the Tribunal for taking into account the factors that might lessen the relevant impact such as new treatment, the likely diminution of risk of relapse over a period of time and the chance that a relapse might be the result of some new intervening event, but she did not acknowledge that the Tribunal also had in mind factors that might increase the impact of the risk of relapse on the period of loss such as potentially serious consequences of it happening and the fact that 50% risk of recurrence was a high one. In my view it was both acceptable and appropriate for the Tribunal to look at those factors in the round when it came to make the broad assessment that it did of the period of loss for which the Claimant should be compensated and in relation to the levels of earnings he was likely to achieve gradually over that period. I conclude on this point that there is nothing in the Tribunal's Judgment to indicate that it did anything other than take a number of relevant factors into account in assessing, as best as it could, the Claimant's likely loss. It explained its conclusions succinctly in relation to that period of loss and was not required to factor in the risk of relapse in some sort of specific percentage when it came to that later stage. Accordingly this ground of the appeal also fails. For the avoidance of doubt, I conclude also that the very high test for perversity is not met on any of these grounds of appeal. The Tribunal's conclusions were ones it was entitled to reach on the evidence, albeit that a different Tribunal may have concluded differently on issues such as the timescale for the Claimant's return to work standing the accepted risk of relapse.

43. The main thrust of the remainder of the appeal relates to the manner in which the Tribunal approached the identification of non-work related elements and work related but non-discriminatory factors in apportioning compensation on a 70%/30% basis such that the Claimant is to receive 70% of each award. I am in no doubt that the Tribunal was entitled to make the findings it did in relation to the division of the various events into those that emanated from the Respondent's discriminatory acts and those that did not. The central question in this part of the appeal, however, is whether the Tribunal failed to follow the approach now approved by the Court of Appeal in BAE Systems (Operations) Ltd v Konczak [2017] EWCA Civ 1188 by focusing on the Respondent's positive contribution in relation to the Claimant's mental health condition rather than the divisibility of the harm. It is now accepted that the question is whether the Tribunal can identify, however broadly, a particular part of the Claimant's suffering which is due to the Respondent's wrong and not whether it can assess the degree to which the wrong caused the harm. Counsel for the Claimant accepted that there should be apportionment unless it was found that such a division was not possible. She contended that the Tribunal had erred not just in its approach but also in finding that any division was possible.
44. The law in this area at the time the Tribunal made its decision was best expressed in the case of Thaine v London School of Economics [2010] ICR 1422 and it is on that decision that the Tribunal relied. The law as now properly articulated is contained in the more recent decision in Konczak by the Court of Appeal the judgment of which was handed down at the end of July 2017. Prior to the Court of Appeal's decision a difference in views had been expressed in relation to the divisibility of psychiatric injury. On that issue the Court of Appeal determined that what Keith J said in Thaine v London School of Economics was to be preferred, such that where an employee's psychiatric ill health is caused by a combination of factors, some of which amount to unlawful discrimination but others which are not the legal responsibility of the employer, the Tribunal can discount the employee's compensation by such percentage as reflects its apportionment of that responsibility. In relation to correct approach to be taken in assessing apportionment in this area, however, the Court of Appeal in Konczak went on to express the following view:

"71. What is therefore required in any case of this character is that the tribunal should try to identify a rational basis on which the harm suffered can be apportioned between a part caused by the employer's wrong and a part which is not so caused. I would emphasise, because the distinction is easily overlooked, that the exercise is concerned not with the divisibility of the causative contribution but with the divisibility of the harm. In other words, the question is whether the tribunal can identify, however broadly, a particular part of the suffering which is due to the wrong; not whether it can assess the degree to which the wrong caused the harm.

72. That distinction is easy enough to apply in the case of a straightforward physical injury. A broken leg is "indivisible": if it was suffered as a result of two torts, each tortfeasor is liable for the whole, and any question of the relative degree of "causative potency" (or culpability) is relevant only to contribution under the 1978 Act. It is less easy in the case of psychiatric harm. The message of Hatton is that such harm may well be divisible. In Rahman the exercise was made easier by the fact (see para. 57 above) that the medical evidence distinguished between different elements in the claimant's overall condition, and their causes, though even there it must be recognised that the attributions were both partial and approximate. In many, I suspect most, cases the tribunal will not have that degree of assistance. But it does not follow that no apportionment will be possible. It may, for example, be possible to conclude that a pre-existing illness, for which the employer is not responsible, has been materially aggravated by the wrong (in terms of severity of symptoms and/or duration), and to award compensation reflecting the extent of the aggravation. The most difficult type of case is that posited by Smith LJ in her article, and which she indeed treats, rightly or wrongly, as the most typical: that is where "the claimant will have cracked up quite suddenly, tipped over from being under stress into being ill". On my understanding of Rahman and Hatton, even in that case the tribunal should seek to find a rational basis for distinguishing between a part of the illness which is due to the employer's wrong and a part which is due to other causes; but whether that is possible will depend on the facts and the evidence. If there is not such basis, then the injury will indeed be, in Hale LJ's words, "truly indivisible", and principle requires that the claimant is compensated for the whole of the injury - though, importantly, if (as Smith LJ says will be typically the case) the claimant has a vulnerable personality, a discount may be required in accordance with proposition 16."

45. The question in this case, then, is whether the Tribunal's decision on this point is illustrative of material error by failing first to address the question of whether the harm suffered by the Claimant was divisible or not. It seems to me that the exercise the Tribunal required to carry out was to look at the evidence about the Claimant's constellation of symptoms and if some of these clearly did not arise from the Respondent's fault then to make a sensible attempt to apportion liability. That remains the correct approach discussed in Sutherland v Hatton [2002] ICR 613 and approved by the Court of Appeal in Konczak at paragraph 59. That said, the approach to divisibility of psychiatric injury in cases of this sort having been clarified recently by the Court of Appeal, the Tribunal did not approach the matter in accordance with the requirement to consider the issue of divisibility of harm (and that in two stages) rather than divisibility of the events into wrongs of the Respondent on the one hand and fault neutral events on the other. There was sufficient material before the Tribunal to enable it to carry out that exercise. If the Tribunal erred, consideration has to be given to whether it can be concluded with confidence that the outcome would have been the same.
46. Of course apportionment is often a "rough and ready" assessment, but the issue of whether it should be approached from the perspective of divisibility of harm (and if so apportionment of the harm caused by discriminatory factors) was clearly put in issue on behalf of the Claimant at the remedy hearing. At that time, the decision of this Tribunal in Konczak was available (UKEAT/0243/13) and counsel for the Claimant contended on the basis of that decision that the issue of whether the Claimant's mental state was divisible had to be determined prior to considering apportionment. She did also submit that there was a strong case for arguing that psychiatric injury is in fact indivisible, citing what at that time were unresolved conflicting views expressed on that matter by Smith LJ and Hale LJ. As already indicated, the Court of Appeal has decided that in cases of this sort the injury may well be divisible as between that caused by the employer's wrong and that which was pre-existing or otherwise not so caused. The specific approval of the dicta of Keith J in Thaine was in relation to the need for the Tribunal to attempt to apportion harm if possible, to avoid an outcome that saw the employer paying in full when it is known that only part of the harm is their fault. To that extent Thaine remains good law. What the Court of Appeal in Konczak does not support is an approach that does not ask first whether the harm is divisible, on the evidence led about the particular condition in question. Senior counsel for the Respondent pointed out that the medical evidence before the Tribunal pointed to harm caused to the Claimant by non-discriminatory factors by reference to the chronology of events. At paragraph 8 the Tribunal summarises the events of 2009 when the Claimant's mother died and his daughter was diagnosed with autism and the restructuring within the organisation in 2011 as specific factors that had marked the beginning of the Claimant's anxiety and depressive symptoms. On any view these predated the discriminatory acts which began in November 2011. That said, there is no doubt that the Tribunal's statement at paragraph 29 that what it was doing was assessing the "contribution of the discriminatory acts to the Claimant's overall condition" is indicative of a very different approach to that identified by the Court of Appeal as required in cases of this sort. The reference at the outset of that paragraph to the Claimant's "overall condition" suggests either that no consideration was given to the divisibility of harm or even that the condition was not regarded as divisible. In assessing apportionment the Tribunal then divided into four subparagraphs the specific discriminatory acts that contributed to the Claimant's condition. The failure to remove the Claimant from his people management role is described (in paragraph 29.2) as "the most important single factor", which serves to emphasise that it was the Respondent's acts that were being divided, not the harm itself.
47. While I do not rule out that the divisibility of harm approach would achieve the same result, it is not for this Tribunal to conduct that assessment, which is a contentious one between the parties on the basis of the evidence. I am not in a position to conclude that, had the Tribunal approached the matter as now required following the Court of Appeal decision in Konczak, it would have reached the same conclusion on the issue of apportionment. The Tribunal's error in law, as the law is now authoritatively stated, is sufficiently material to allow the appeal such that the issue of apportionment can be considered of new. Accordingly, the case will require to be remitted back to the Tribunal for a decision on whether, on the facts already found, the harm suffered by the Claimant was divisible and if so, how much of that harm is attributable to the Respondent's discriminatory acts.

Decision

48. For the reasons given, my decision in relation to each of the issues raised in the appeals and cross-appeal is as follows;

(i) The appeal in relation to the deduction of 50% of the PHI payments from loss of earnings does not succeed;

(ii) The appeal in relation to the addition to the award of the sum the Claimant would have received by way of redundancy and notice payments had he been made redundant succeeds;

(iii) The appeal and cross-appeal in relation to the Tribunal's finding (at paragraph 46.3) that the time during which the Claimant can rely on the PHI payments being made is limited by circumstances other than his ability to work succeeds;

(iv) The appeal in relation to the Tribunal's assessment of the anticipated length of the Claimant's recovery period, the level of his future earnings and the risk of relapse fails;

and

(v) The appeal in relation to the 70/30 apportionment succeeds, to the extent that the Tribunal's approach is not in accordance with that now required in line with the Court of Appeal's dicta in BAE Systems (Operations) Ltd v Konczak [2017] EWCA Civ 1188.

49. Counsel agreed that it would be appropriate for a decision in principle to be given on each of the issues with an opportunity being given for any recalculation to be carried out by agreement. However, as the outcome of the Claimant's appeal being allowed in part is that the case will require to be remitted back to the same Tribunal on the issue of apportionment, it would clearly not be possible for recalculation to be carried out at this stage. Accordingly, I will remit the case back to the same Tribunal to determine, on the basis of this Judgment and on the evidence already before it: (1) the effect of the removal of the award of redundancy and notice payments, (2) the apportionment of loss in light of the conclusions I have reached on the correct approach to that, and (3) the recalculation of the total award of compensation and interest in light of (1) and (2) above and the payments already made by the Respondent.