England and Wales Court of Appeal (Civil Division) Decisions

IN THE SUPREME COURT OF JUDICATURE
COURT OF APPEAL (CIVIL DIVISION)
ON APPEAL FROM EMPLOYMENT
APPEAL TRIBUNAL

		Royal Courts of Justice Strand, London, WC2A 2LL Thursday 22nd March, 2001

LORD JUSTICE SEDLEY :

This is the judgment of the Court:

Events

1. The Appellant, Dr. Chinasa Anya, appeals by leave of this court from a decision of the Employment Appeal Tribunal dismissing his appeal against an adverse finding of the Industrial Tribunal (now renamed an Employment Tribunal) given, with extended reasons, on 26 March 1998 after a hearing spread over 11 days. His allegation against Oxford University and his postdoctoral supervisor Dr Roberts was of direct race discrimination in the award of a postdoctoral research assistant's post. The Appellant, who is black, is a Nigerian permanently resident in the UK. The post went to the other shortlisted candidate, Dr Lawrence, who is white.
2. Our understanding of the case which was put forward is that, as Dr Anya had clearly indicated in his own originating application, the justiciable act of discrimination for which he sought redress was his rejection at interview. His evidence for this allegation lay in his account of his personal and professional relationship with his supervisor, Dr Roberts. The Industrial Tribunal, which shared this understanding, broke the culminating event down into three elements (the application, the interview and the decision: paragraphs 19, 20 and 21), and separately tabulated the material episodes in the preceding relationship. We do not accept the submission of Mr Andrew Hochhauser QC on Dr Anya's behalf that each of these elements was a discrete complaint of race discrimination. They were correctly marshalled as evidence for the single complaint that the admitted act of discrimination in choosing Dr Lawrence rather than Dr Anya was racially motivated.
3. Oxford University's Department of Materials is an institution of world standing. The Appellant came to his post there as a Post-Doctoral Research Assistant (PDRA) with a doctorate in metallurgy from Strathclyde University, where had stayed on to do research sponsored by an EU-funded organisation, Brite-Euram. He came to Oxford University in 1994 on a 2-year project on ceramic nanocomposites funded by the Engineering and Physical Sciences Research Council (EPSRC). This represented a change of focus for him to a process known as sintering, but the Industrial Tribunal records that he achieved "considerable success" in it. Dr Anya's post-doctoral supervisor was a lecturer in the department, Dr Steven Roberts (the second respondent). As his project drew to an end, two new ones came up. One was an abortive one put to the EPSRC, with Dr Roberts as principal unit investigator. The other, a multi-institutional application for EU funding through Brite-Euram, succeeded, and a new postdoctoral research post in it was advertised. Of the 26 applicants, Dr Anya and Dr Lawrence were the shortlisted candidates. It follows that each was fully qualified, in c.v. terms, to hold the new post and that the choice might be a close one.
4. The interview was conducted by a panel of three: Mr Briant, the departmental administrator, who is not a scientist; Dr Roberts; and Dr Jan Czernuska, a lecturer in the department and an expert in ceramics, though not the particular aspect relevant to the project. One of the few factual findings of the Industrial Tribunal is that Dr Roberts had already formed an adverse view of the applicant's suitability for the post and that Mr Briant knew this. This is their account of the interview:

"20. Secondly, the applicant complained that when interviewed he was faced with a panel of three. The panel was chaired by Mr. Briant who had no scientific background at all. Dr. Jan Czernuska, a Lecturer within the Department, had been invited to join the panel as an expert in ceramics, albeit of a somewhat different specialisation, together with Dr. Roberts. Dr. Roberts had already formed an adverse view of the applicant's suitability for the post and had shared those views with Mr. Briant. Dr. Czernuska approached the interview with an entirely unbiased mind. With regard to Dr. Lawrence, neither Dr. Czernuska nor Mr. Briant had any prior information about him beyond that contained in his application form and CV. It was argued that the applicant was thus faced with an interview panel who already slanted away from him when they were at least neutral towards his competitor, Dr. Lawrence. This is an aspect which again the Tribunal considered; one member of the Tribunal took the view that in these particular circumstances prior knowledge of the apparent strengths and weaknesses of an internal candidate was not necessarily less favourable treatment. Criticism had been made on the process of why the shortlisting had been carried out. This had been done by Dr. Roberts alone in the first instance to produce a shortlist of four. The University's policies prescribed a shortlisting panel of two with the candidates to be shortlisted against a "person specification". No such specification had been drawn up at that stage, nor was any drawn up until minutes before the interview commenced. The evidence from the respondents was that the principal criteria were technical and scientific skill and ability, followed by organisational management and presentational skills, specific modelling abilities in analysis and computer simulation, and drive and enthusiasm. The policy specified that applicants should be told beforehand of the person specification whereas in the case of the two applicants there was no such communication other than the delivery to them of the Brite-Euram project application documentation the evening before. This was complex documentation from which it appears that they were expected to analyse the precise nature of the skills which would be required. It was acknowledged by the respondents that this was hardly satisfactory. It emerged during the course of the shortlisting process that the University had a practice or a policy of offering any internal applicant for a post an interview unless manifestly unsuitable. This did not accord with the formal policies of the University. The policies further indicated that the decision would be made following interview after consideration of the Curriculum Vitae and the taking up of references. No references were taken up in this case."

At the end of the next paragraph the Industrial Tribunal record their finding that it was Dr Czernuska who first expressed the conclusion that Dr Anya would not be suitable for the post. Mr Briant and then Dr Roberts expressed their concurrence.

5. The Appellant took the issue through the University's grievance procedures. Although the Grievance Panel did not reverse the interview panel's decision, it found shortcomings in the way the University's equal opportunities and recruitment policies had been operated.

Law

6. There is no need for the purposes of this appeal to recall more than that it is unlawful by virtue of s. 4 of the Race Relations Act 1976 to discriminate on grounds of race in deciding who is to be offered a particular job. By s.1(1)(a) discrimination for this purpose means treating one person less favourably than another on racial grounds. As the Industrial Tribunal correctly found, the award of the post to Dr Lawrence meant that the Appellant was treated less favourably. The single question was therefore whether the Appellant's race had played any significant role in the choice.
7. Deciding such questions is not easy. The problem was classically addressed in this court by Neill LJ in King v Great Britain-China Centre [1992] ICR 516. In a well-known passage which the IT clearly had in mind, he summarised the relevant principles in this way:

"From [the] authorities it is possible, I think, to extract the following principles and guidance. (1) It is for the applicant who complains of racial discrimination to make out his or her case. Thus if the applicant does not prove the case on the balance of probabilities he or she will fail. (2) It is important to bear in mind that it is unusual to find direct evidence of racial discrimination. Few employers will be prepared to admit such discrimination even to themselves. In some cases the discrimination will not be ill-intentioned but merely based on an assumption that "he or she would not have fitted in." (3) The outcome of the case will therefore usually depend on what inferences it is proper to draw from the primary facts found by the tribunal. These inferences can include, in appropriate cases, any inferences that it is just and equitable to draw in accordance with section 65(2)(b) of the Act of 1976 from an evasive or equivocal reply to a questionnaire. (4) Though there will be some cases where, for example, the non-selection of the applicant for a post or for promotion is clearly not on racial grounds, a finding of discrimination and a finding of a difference in race will often point to the possibility of racial discrimination. In such circumstances the tribunal will look to the employer for an explanation. If no explanation is then put forward or if the tribunal considers the explanation to be inadequate or unsatisfactory it will be legitimate for the tribunal to infer that the discrimination was on racial grounds. This is not a matter of law but, as May LJ. put it in North West Thames Regional Health Authority v. Noone [1988] ICR 813, 822, "almost common sense." (5) It is unnecessary and unhelpful to introduce the concept of a shifting evidential burden of proof. At the conclusion of all the evidence the tribunal should make findings as to the primary facts and draw such inferences as they consider proper from those facts. They should then reach a conclusion on the balance of probabilities, bearing in mind both the difficulties which face a person who complains of unlawful discrimination and the fact that it is for the complainant to prove his or her case."

8. It should be made clear that when Neill LJ refers under head (4) to discrimination, he means it in its literal, not its objectionable, sense: that is to say, he is referring simply to a choice or a process of selection such as occurred here. It is not unduly onerous, as has sometimes been suggested, to proceed from the simple fact of such a choice, if it is accompanied by difference in race, to a request for an explanation. In the allocation of jobs by any sensibly-run institution, the explanation will be straightforward: the candidates were interviewed by an unbiased panel on an equal footing, using common criteria which contained no obvious or latent elements capable of favouring one racial group over another; and the best one was chosen. By parity of reasoning, evidence that one or more members of the panel were not unbiased, or that equal opportunities procedures were not used when they should have been, may point to the possibility of conscious or unconscious racial bias having entered into the process. It will always be a matter for the Tribunal's conscientious judgment.
9. This reasoning has been valuably amplified by Mummery J in Qureshi v Victoria University of Manchester (EAT 21.6.96), a decision which Holland J in the present case in the Employment Appeal Tribunal understandably described as "mystifyingly unreported". It is therefore worth quoting at length from Mummery J's judgment.

"On the basis of (a) those authorities, (b) the experience of the members of this Tribunal and (c) the experience of the parties, the advisers and the Tribunal in this case, we tentatively add the following observations and thoughts to the guidance in Neill LJ's judgment in King v. Great Britain China Centre –

The complainant

The Industrial Tribunal only has jurisdiction to consider and rule upon the act or acts of which complaint is made to it. If the applicant fails to prove that the act of which complaint is made occurred, that is the end of the case. The Industrial Tribunal has no jurisdiction to consider and rule upon other acts of racial discrimination not included in the complaints in the Originating Application. See Chapman v. Simon [1994] IRLR 273 at paragraph 33(2) (Balcombe LJ) and paragraph 42 (Peter Gibson LJ). In this case the principal complaints made by Dr Qureshi were decision of the FRC not to support a recommendation for his promotion to the post of senior lecturer in October 1992 and the decision of the Dean of the Law Faculty in October 1993 not to put his name forward to the APC with a favourable recommendation for promotion to senior lecturer. The considerations of the Tribunal and their decision should, therefore, focus on those complaints and on the issues of fact and law which have to be resolved in order to decide whether the complaints are well founded or not.

The issues

As the Industrial Tribunal have to resolve disputes of fact about what happened and why it happened, it is always important to identify clearly and arrange in proper order the main issues for decision eg,

(a) Did the act complained of actually occur? In some cases there will be a conflict of direct oral evidence. The Tribunal will have to decide who to believe. If it does not believe the applicant and his witnesses, the applicant has failed to discharge the burden of proving the act complained of and the case will fail at that point. If the applicant is believed, has he brought his application in time and, if not, is it just and equitable to extend the time?

(b) If the act complained of occurred in time, was there a difference in race involving the applicant?

(c) If a difference in race was involved, was the applicant treated less favourably than the alleged discriminator treated or would treat other persons of a different racial group in the same, or not materially different, relevant circumstances?

(d) If there was difference in treatment involving persons of a different race, was that treatment "on racial grounds"? Were racial grounds an effective cause of the difference in treatment? What explanation of the less favourable treatment is given by the respondent?

In answer to each of these questions the Tribunal must make findings of primary fact, either on the basis of direct (or positive) evidence or by inference from circumstantial evidence.

The evidence

As frequently observed in race discrimination cases, the applicant is often faced with the difficulty of discharging the burden of proof in the absence of direct evidence on the issue of racial grounds for the alleged discriminatory actions and decisions. The Applicant faces special difficulties in a case of alleged institutional discrimination which, if it exists, may be inadvertent and unintentional. The Tribunal must consider the direct oral and documentary evidence available, including the answers to the statutory questionnaire. It must also consider what inferences may be drawn from all the primary facts. Those primary facts may include not only the acts which form the subject matter of the complaint but also other acts alleged by the applicant to constitute evidence pointing to a racial ground for the alleged discriminatory act or decision. It is this aspect of the evidence in race relations cases that seems to cause the greatest difficulties. Circumstantial evidence presents a serious practical problem for the Tribunal of fact. How can it be kept within reasonable limits? This case is an illustration of the problem. The complaint of racial discrimination is usually sparked by a core concern of the applicant: in this case his failure to obtain support and recommendations for his promotion to a senior lecturer in the Faculty of Law. Dr. Qureshi relied extensively on circumstantial evidence that there was a racial ground for the acts and decisions he complained about. The circumstantial evidence included incidents ranging over a period of nearly six years, from 1988 to 1994. The incidents relied on by him ante-date, accompany and post-date the alleged acts of racial discrimination and victimisation particularised in his 1993 and 1994 applications. It was necessary for the Tribunal to find the facts relating to those incidents. They are facts (evidentiary facts) relied upon as evidence relevant to a crucial fact in issue namely, whether the acts and decisions complained of in the proceedings were discriminatory "on racial grounds". The function of the Tribunal in relation to that evidence was therefore two-fold: first, to establish what the facts were on the various incidents alleged by Dr Qureshi and, secondly, whether the Tribunal might legitimately infer from all those facts, as well as from all the other circumstances of the case, that there was a racial ground for the acts of discrimination complained of. The temptation for the complainant and his advisers, in these circumstances, is to introduce into the case as many items as possible as material from which the Industrial Tribunal might make an inference that "racial grounds" are established. The respondent has to respond to the introduction of those items. He may dispute some of them as factually incorrect. He may seek to introduce other evidence to negative any possible inference of racial grounds eg, non-racial explanations for his acts and decisions. The result of this exercise is that the parties and their advisers may confuse each other (and the Tribunal) as to what the Tribunal really has to decide; as to what is directly relevant to the decision which it has to make and as to what is only marginally relevant or background. It is a legitimate comment that in some cases of race discrimination so much background material of marginal relevance is introduced that focus on the foreground is obscured, even eclipsed. In practical terms this may lead the case to run on and on for many days or weeks. In the experience of this Tribunal, the longest cases heard in the Industrial Tribunals are cases of racial discrimination.

Inferences

The process of making inferences or deductions from primary facts is itself a demanding task, often more difficult than deciding a conflict of direct oral evidence. In Chapman v. Simon (supra) at paragraph 43 Peter Gibson LJ gave a timely reminder of the importance of having a factual basis for making inferences. He said –

"...Racial discrimination may be established as a matter of direct primary fact. For example, if the allegation made by Ms Simon of racially abusive language by the headteacher had been accepted, there would have been such a fact. But that allegation was unanimously rejected by the Tribunal. More often racial discrimination will have to be established, if at all, as a matter of inference. It is of the greatest importance that the primary facts from which such inference is drawn are set out with clarity by the Tribunal in its fact-finding role, so that the validity of the inference can be examined. Either the facts justifying such inference exist or they do not, but only the Tribunal can say what those facts are. A mere intuitive hunch, for example, that there has been unlawful discrimination is insufficient without facts being found to support that conclusion." (See also Balcombe LJ at para. 33(3))

In the present case, it was necessary for the Tribunal to examine all the allegations made by Dr Qureshi of other incidents relied upon by him as evidentiary facts of race discrimination in the matters complained of. There is a tendency, however, where many evidentiary incidents or items are introduced, to be carried away by them and to treat each of the allegations, incidents or items as if they were themselves the subject of a complaint. In the present case it was necessary for the Tribunal to find the primary facts about those allegations. It was not, however, necessary for the Tribunal to ask itself, in relation to each such incident or item, whether it was itself explicable on "racial grounds" or on other grounds. That is a misapprehension about the nature and purpose of evidentiary facts. The function of the Tribunal is to find the primary facts from which they will be asked to [draw] inferences and then for the Tribunal to look at the totality of those facts (including the respondent's explanations) in order to see whether it is legitimate to infer that the acts or decisions complained of in the originating applications were on "racial grounds". The fragmented approach adopted by the Tribunal in this case would inevitably have the effect of diminishing any eloquence that the cumulative effect of the primary facts might have on the issue of racial grounds. The process of inference is itself a matter of applying common sense and judgment to the facts, and assessing the probabilities on the issue whether racial grounds were an effective cause of the acts complained of or were not. The assessment of the parties and their witnesses when they give evidence also form an important part of the process of inference. The Tribunal may find that the force of the primary facts is insufficient to justify an inference of racial grounds. It may find that any inference that it might have made is negated by a satisfactory explanation from the respondent of non-racial grounds of action or decision.

Conclusion

The additional comments are intended to provide some assistance to the Tribunal to whom this case is remitted (and to other Tribunals) in deciding what are, in our view, the most difficult kind of case which Industrial Tribunals have to decide. The legal and evidential difficulties are increased by the emotional content of the cases. Feelings run high. The complainant alleges that he has been unfairly and unlawfully treated in an important respect affecting his employment, his livelihood, his integrity as a person. The person against whom an accusation of discrimination is made feels that his acts and decisions have been misunderstood, that he has been unfairly, even falsely, accused of serious wrongdoing. The accusations may not only be hurtful to him as a person but may also be damaging to his employment, his prospects and his relationships with others. In our experience, the Industrial Tribunals discharge this delicate, difficult function conscientiously and carefully. It should not be regarded as a criticism on the Chairman and members of this Tribunal that we have found their lengthy, conscientious and detailed decision, taken after many hearing days and many hours of deliberation, to contain an error of law."

10. Running through this guidance, and the guidance cited in it, is the ubiquitous need to make the findings of primary fact without which it is impossible to consider the drawing of relevant inferences. It can be found again in this court's judgment in Marks and Spencer plc v Martins [1998] ICR 1005, which cites Lord Browne-Wilkinson's reminder in Glasgow City Council v Zafar [1998] ICR 120, 125 that

"Claims [of race and sex discrimination] present special problems of proof for complainants since those who discriminate on grounds of race or gender do not in general advertise their prejudices"

Once again, King was highlighted by Lord Browne-Wilkinson as the leading source of guidance. It may be added that, as King and other cases demonstrate, the breakdown of issues need not always be as elaborate as was necessary in Qureshi.

11. The present case is a textbook example of a race discrimination claim. It makes it possible to see with some clarity how the principles established by authority ought to work out in practice. Here we have a shortlist of two candidates, one black, one white, both by definition qualified by training and experience for a specialised post. Whichever is to be chosen, good administration requires that he be chosen fairly; and to this the law has now added for a quarter of a century that the choice must not be affected in any way by his race. If it is, the unsuccessful candidate will have been treated less favourably on racial grounds and the university will be liable for direct discrimination. This was as true for Dr Lawrence as it was for Dr Anya: both were entitled to the protection of the Race Relations Act 1976. Very little direct discrimination is today overt or even deliberate. What King and Qureshi tell tribunals and courts to look for, in order to give effect to the legislation, are indicators from a time before or after the particular decision which may demonstrate that an ostensibly fair-minded decision was, or equally was not, affected by racial bias.
12. Mr Nicholas Underhill QC, for the respondents, has submitted that Qureshi is distinguishable from the present case inasmuch as there the Industrial Tribunal's reasons, far from overlooking the factual issues, had practically sunk under the weight of them. But the importance of the Employment Appeal Tribunal's guidance in that case in no way depends on this. It is of general relevance, among other things, in demonstrating why an inadequately reasoned judgment denies the parties, especially but not solely the losing party, the materials necessary to know why the outcome has been what it has and whether it is appealable - a principle reiterated by this court in Flannery v Halifax Estate Agencies Ltd [2000] 1 All ER 373, where the want of adequate reasons was held to be a freestanding ground of appeal. To this can now be added the jurisprudence of Article 6 of the European Convention on Human Rights, which the Human Rights Act 1998, by ss. 2 and 6, requires courts and judicial tribunals to take into account, to the effect that adequate and intelligible reasons must be given for judicial decisions. Extended reasons, such as have by law to be given in race relations cases, are designed among other things to meet precisely this need.

The Industrial Tribunal's decision

13. Here the Industrial Tribunal were satisfied, on balance, that despite inconsistencies which emerged under cross-examination Dr Roberts was essentially a truthful witness. Dr Roberts had explained his reasons, which had to do entirely with Dr Anya's qualities as a scientist, for not choosing him for the new post. It followed, in the Industrial Tribunal's judgment, that the inevitably less favourable treatment of Dr Anya had nothing to do with his race.
14. Such a conclusion was without doubt open to them, but only provided it was arrived at after proper consideration of the indicators which Dr Anya relied on as pointing to an opposite conclusion. His case was that the evidence showed two critical things. One was a preconceived hostility to him: this depended on matters of fact which it was for the Industrial Tribunal to ascertain or refute on the evidence placed before them. The other was a racial bias against him evinced by such hostility: this was a matter of inference for the Industrial Tribunal if and insofar as it found the hostility established. Experience shows that the relationship between the two may be subtle. For example, a tribunal of fact may be readier to infer a racial motive for hostility which has been denied but which it finds established than for hostility which has been admitted but acceptably explained. The Industrial Tribunal in paragraph 5 of its reasons directed itself correctly in law about this, with one arguable exception: it concluded the paragraph with this remark:

"If an employer behaves unreasonably towards a black employee it is not to be inferred, without more, that the reason for this is attributable to the employee's colour; the employer might very well behave in a similarly unreasonable fashion to a white employee."

As Neill LJ pointed out in King, such hostility may justify an inference of racial bias if there is nothing else to explain it: whether there is such an explanation as the Industrial Tribunal posit here will depend not on a theoretical possibility that the employer behaves equally badly to employees of all races but on evidence that he does.

15. In the present case the Industrial Tribunal embarked in exemplary fashion on the methodical approach which this court has said is essential. In paragraph 16 they tabulated five prior events put in evidence by Dr Anya as evidence of hostility on the part of Dr Roberts and denied or explained, as in each case the Industrial Tribunal records, by Dr Roberts himself and by other witnesses. They were, in brief, that Dr Roberts had not given Dr Anya adequate research guidance; had not given him the opportunity he gave others to co-propose research projects; had discouraged or blocked him from publishing research papers; had not given him the opportunity to present his work at scientific conferences; and had allocated him no research students in contrast to other PDRAs. These allegations were followed, in paragraph 17(b), by another which had emerged on disclosure of documents: that in the unsuccessful funding application to the ESPRC Dr Lawrence had been named as the intended research assistant. Even more directly related to the material job application is the other allegation set out in paragraph 17: that Dr Anya was notified by Mr Briant, the eventual panel chairman, of the coming opportunity in terms markedly less encouraging than those in which Dr Lawrence was notified of it. On none of these issues, from first to last, did the Industrial Tribunal record any conclusion as to where the truth lay and what, if anything, it indicated in terms of racial bias.
16. There then follows in paragraph 19 a summary of the rival accounts of the run-up to the advertisement of the Brite-Euram post. As to this the Industrial Tribunal say:

"… The Tribunal, whilst preferring the factual basis contended for by Dr Roberts and so finding as to the date and the content of the conversation, acknowledged that there is scope for both points of view as contended for by the parties."

17. In paragraph 20 the Industrial Tribunal deal with the crucial interview, conscientiously setting out the key elements and hinting, but no more, at the members' individual thinking about it. They do, however, note two things of potential significance. One is that, in breach of the University's own equal opportunities policy, no person specification had been drawn up until minutes before the interview. As any such tribunal will be aware, a person specification is an important aid to ensuring that candidates are considered on a basis of parity and without criteria which have an unjustifiably discriminatory effect. We do not know what the Industrial Tribunal made of the consequent inability of both the candidates and the panel to have adequate notice of the person specification. The second is that, in breach of University policy, no references had been taken up on the candidates. But again the Industrial Tribunal record no inferences or conclusions from these facts.
18. The last point, addressed in paragraph 21, that the Appellant was not appointed and that a white candidate was, is of course not a fresh issue but the culmination of the preceding ones. Again the Industrial Tribunal carefully set out the two sides' positions. In the next paragraph they describe the range of material which they have considered. They comment importantly:


"… In many respects inconsistencies have been exposed in these documents between their contents and the evidence given subsequently and to this Tribunal. Those are not without substance and we bear them in mind in the task we have to address, namely whether there has been less favourable treatment and, if so, whether it is attributable to the applicant's race. In this task we are entitled to rely on our assessments of the individuals who have come to us and given evidence. …"

They go on to point out, correctly, that they are unequipped to make judgments on the relative scientific merits of two highly qualified candidates in a leading-edge research field. They summarise Dr Roberts' view of Dr Anya as being not that he was a poor scientist but that he had, along with his strengths, weaknesses which placed him second to Dr Lawrence. They conclude:

"23…. It would be inappropriate for us to characterise any of the witnesses coming before us in this hearing as being untruthful, but we have to say that we regarded Dr Roberts and Professor Cantor as being essentially witnesses of truth despite the inconsistencies that were exposed under skilful cross-examination.

24. The unanimous view of the Tribunal is that we are satisfied that the applicant received less favourable treatment in that Dr Lawrence was appointed when he was not. We are invited to draw the inference [that this] was because of his race and not, as the respondents claim, on a genuine assessment of his scientific strengths and weaknesses. We are disposed to accept the respondents' explanation and in our view the evidence is not sufficient to justify us in drawing the inference of discrimination."

19. One has only to ask whether this would necessarily have been the outcome if Dr Anya's allegations of prior hostility had been made out in order to see what is lacking in this decision. The Industrial Tribunal, with all respect to them, have started at the far end of the process of reasoning and have never returned to base. Having concluded that Dr Roberts was essentially truthful (we do not know whether or not they thought the same of Dr Anya), they have abandoned the remainder of the inquiry. They make no findings about the substance, much less the significance, of the inconsistencies which they have noted in his and Professor Cantor's evidence; no evaluation of the further inconsistencies which they have found between the University's documents and its evidence; no findings as to whether the Appellant had, as he contended, been repeatedly sidelined by Dr Roberts in the course of his research work; no finding about the significance of the differences in the way he and Dr Lawrence were treated; and no indication of the significance, if any, of the interview panel's departures from university policy or good practice. Nor therefore have they been in a position to decide whether any hostility they might have found was in any significant degree racially motivated.

The arguments

20. Mr Underhill, in an impressive argument, has contended that a careful analysis shows that the Industrial Tribunal has found and stated all that was necessary for a rounded and defensible decision. He points, first, to the last part of the final sentence of the decision ("… in our view the evidence is not sufficient to justify us in drawing the inference of discrimination.") and submits that this deals adequately, albeit briefly, with Dr Anya's entire account of manifestations of hostility. In a plain enough case, he submits, where there is no evidence of race discrimination at the actual point of complaint (here, the interview), it is legitimate for the tribunal of fact to decline to travel in detail through the prior history. Accordingly, so long as the Industrial Tribunal bore everything in mind but found Dr Roberts, as they did, to be an honest witness in his repudiation of any racial bias, they had no obligation to make any further findings.
21. This may be so in an extreme case - but an extreme case in this context is one which is capable of being dismissed in limine on the ground that the facts relied on, even if proved, could not justify an inference of race discrimination. The present is not such a case, and the experienced solicitor who conducted it for the respondents before the Industrial Tribunal did not submit that it was. It is closer, in fact, to the opposite end of the scale. The choice between these two comparably well qualified candidates depended entirely on how the panel viewed their personal and professional qualities. Such a judgment is notoriously capable of being influenced, often not consciously, by idiosyncratic factors, especially where proper equal opportunity procedures have not been followed. If these are to any significant extent racial factors, it will in general be only from the surrounding circumstances and the previous history, not from the act of discrimination itself, that they will emerge. This court and the Employment Appeal Tribunal have said so repeatedly and have required tribunals to inquire and reason accordingly.
22. If nevertheless the Industrial Tribunal thought that this was the class of case in which they could take the short cut, the least they were obliged to do was to say so. The permissibility of taking it would then be open to scrutiny. Mr Underhill's invitation to us to infer that this is what they were doing has two serious vices. First, it is an invitation to an appellate court to speculate where it was the tribunal's duty to explain itself. Secondly, and more importantly, it requires us to assume as decided against Dr Anya the very issues which might for all we know have been determined wholly or in part in his favour.
23. Mr Underhill seeks to meet this objection by submitting that none of the historic issues set out by the Industrial Tribunal in paragraph 16 and 17 of their reasons was capable, even if resolved factually in Dr Anya's favour, of founding an inference of racial bias. They were simply, he argues, the kind of thing which can occur in any academic relationship of this kind. By not making explicit findings on them the Industrial Tribunal was, he says, taking an understandably emollient approach in what was a fraught case. Both these things may be true, but they do not answer the problem. It is precisely because a witness who by himself comes across as essentially truthful may be shown by documentary evidence or by inconsistency to be less reliable than he seems that the totality of the evidence in a case like this has to be evaluated; and there was in this case no useful way of approaching the totality except through its parts. If, to take one instance from the foregoing, the Industrial Tribunal had accepted to any significant extent Dr Anya's evidence that Dr Roberts had not given him the academic and professional support that he was entitled to expect between 1994 and 1996, they would have had to ask why this was so. In answering that question they might or might not have derived help from the conversations that Dr Roberts had had with Mr Briant about Dr Anya and the differing terms in which Mr Briant wrote about the new post to Dr Anya and Dr Lawrence respectively. They might then have needed to consider whether these were isolated events or part of a pattern; if part of a pattern, whether academic differences sufficiently explained it; if they did not, whether it was justifiable to infer the presence of a racial element. To the extent that Dr Czernuska had been the first to express his preference for Dr Lawrence, the Industrial Tribunal might have been concerned to establish what had been said in discussion before he did so. To the extent that Dr Roberts' own evidence was crucial, they will have needed to test the good impression he made against any contra-indications in the evidence. No appellate court can now do this job in place of the tribunal of fact; and no tribunal of fact could properly come to a conclusion without dealing with these issues.
24. The difficulty is not answered by the decisions of this court relied on by Mr Underhill (Martin v Glynwed Distribution Ltd [1983] ICR 511, and Meek v Birmingham City Council [1987] IRLR 250) to the effect that tribunals are not required to do more than make findings of fact and answer a question of law. In the race relations field this principle does no more than beg the questions: what findings, what law? It is elsewhere, above all in King, that the answers lie. In Tchoula v Netto Foodstores Ltd (Employment Appeal Tribunal, 6 March 1998) Morison J spelt out what this means in practice:

"A bald statement saying that X's evidence was preferred to Y's is, we think, both implausible and unreasoned and therefore unacceptable; and it might appear to have been included simply to try and prevent any appeal. It seems to us likely that there will be a great deal of background material which is non-controversial. There is no need to recite at length in the decision the evidence which has been received. What a tribunal should do is state their findings of fact in a sensible order (often chronological), indicating in relation to any significant finding the nature of the conflicting evidence and the reason why one version has been preferred to another.

It is always unacceptable for a tribunal to assert its conclusion in a decision without giving reasons."

25. To assert this is not to demand, as Mr Underhill sought to suggest it did, an infinite combing by the Industrial Tribunal through endless asserted facts or an over-nice appraisal of them. It is simply that it is the job of the tribunal of first instance not simply to set out the relevant evidential issues, as this Industrial Tribunal conscientiously and lucidly did, but to follow them through to a reasoned conclusion except to the extent that they become otiose; and if they do become otiose, the tribunal needs to say why. But the single finding of the Industrial Tribunal in this case on Dr Roberts' honesty as a witness, while important, does not make the other issues otiose: on the contrary, it begs all the questions they pose. Mr Underhill's reliance on it as effectively dispositive overlooks what Lord Goff said in The Ocean Frost [1985] I Ll.LR 1, 57:

"It is frequently very difficult to tell whether a witness is telling the truth or not; and where there is a conflict of evidence ….reference to the objective facts and documents, to the witnesses' motives and to the overall probabilities can be of very great assistance to a judge in ascertaining the truth."

The Industrial Tribunal has not given any ground, and none is evident, for departing from this classic mode of reasoning in a case where every one of the ingredients mentioned by Lord Goff was present. The citation from The Ocean Frost in fact features in the transcript of this court's decision in Heffer v Tiffin Green (17 December 1998) where Henry LJ concluded, relevantly to the present case:

"Nor were the crucial contemporary documents given proper, detailed and dispassionate consideration, In my judgment they cannot be explained away … by an uncritical belief in Mr Heffer's credibility …"

Credibility, in other words, is not necessarily the end of the road: a witness may be credible, honest and mistaken, and never more so than when his evidence concerns things of which he himself may not be conscious.

26. There is at least one further obstacle to Mr Underhill's stalwart defence of the Industrial Tribunal's decision. The courts have repeatedly told appellants that it is not acceptable to comb through a set of reasons for hints of error and fragments of mistake, and to try to assemble these into a case for oversetting the decision. No more is it acceptable to comb through a patently deficient decision for signs of the missing elements, and to try to amplify these by argument into an adequate set of reasons. Just as the courts will not interfere with a decision, whatever its incidental flaws, which has covered the correct ground and answered the right questions, so they should not uphold a decision which has failed in this basic task, whatever its other virtues.

The Employment Appeal Tribunal

27. Although the parties agree that this is in practical terms an appeal against the Industrial Tribunal's decision, it would be wrong to part from the Employment Appeal Tribunal's judgment without saying a word about it. The Employment Appeal Tribunal recognised the lacunae in the Industrial Tribunal's reasons but concluded, in effect, that the case was bound to fail. They said:

"… So soon as we turn to the prospects for proving racial discrimination, we are struck by the following. First, no evidence of overt racial discrimination has ever been identified. Second, so far the only proven act of discrimination with possible racial overtones is the rejection of Dr. Anya by the interviewers in favour of the white Dr. Lawrence – or, to be more accurate, the failure of Dr. Roberts, as abetted by Mr. Briant, to overrule Dr. Czernuska's admittedly non-discriminatory rejection of Dr. Anya on academic grounds. We write 'so far' out of deference to the complaint of inadequate fact finding but it is difficult to see what other facts once found could serve to make Dr. Anya's case additional to the proven act more effective.

… Nobody in the context of a complaint of racial discrimination could have listened to evidence over so many days without a growing and legitimate realisation that Dr. Anya's task of proving such was speculative to the point of being hopeless. We think that the Tribunal obtained such an overview from the totality of that which was put before it and gave expression to it in the relatively brief general terms of the concluding paragraphs of the Extended Reasons. We think further that any such overview must consciously or unconsciously have affected the Tribunal's fact finding initiative: why find the specific facts that are sought when it is obvious that none such are going to provide those compelling grounds that will turn that which we have identified as possible into that which has been proved as probable?

We reiterate our deference to the authorities. We emphasise that this case features a complaint with such inherent improbabilities as in our judgment can properly curtail the fact finding process in favour of the mature, simple overview."

28. With the greatest possible respect, this reasoning is not acceptable. The first passage reproduces all the errors of approach which this court in King and the Employment Appeal Tribunal in Qureshi were at pains to correct. Evidence of racial discrimination does not have to be overt. Most commonly it is not. The only proven act of potential racial discrimination is not the final allocation of the research post: it is, in Dr Anya's contention, that event in the context of the series of prior events which, as the Employment Appeal Tribunal acknowledges, have been neither proven nor disproven. There is no difficulty in seeing what facts, if they were found, could make out the Appellant's case. Experience of other cases indicates, speaking generally, that the allegations made by Dr. Anya are not inherently improbable; nor, if his factual allegations are made out, are the reasons for them necessarily speculative. What were lacking were the Industrial Tribunal's conclusions on the factual issues essential to its conclusion and, in consequence, a proper and rounded determination of the single legal matter of complaint, the selection of Dr Lawrence in preference to the Appellant.

Conclusion

29. In spite of the daunting consequences of doing so, the only just outcome is to allow this appeal. The claim must be remitted for rehearing to a new Employment Tribunal. If and when it is reheard, the views of the Employment Appeal Tribunal on its intrinsic merits should play no part in either the argument or the decision.
ORDER: Appeal allowed; remitted to new Employment Tribunal for re-hearing. Appellants costs to be paid by Respondents. Community Legal Services Assessment; Application to appeal to the House of Lords refused. (Order does not form part of approved Judgment)