Stockport Metropolitan Borough Council v K Bonsu

I will refer to the parties as they were in the employment tribunal, as claimant and respondent. The claimant was, during a period from 2005 to 2019, engaged by the respondent to work shifts as a causal support worker through what was called the Support Worker Bureau. Following the decision of the respondent to cease offering him shifts, he presented a claim form complaining of unfair dismissal and direct race discrimination in a number of respects. He identified as of African/black origin and described his ethnicity as black Ghanian.

At a preliminary hearing it was determined that the claimant had not been an employee and his complaint of unfair dismissal was dismissed. The complaints of direct race discrimination went forward to a full merits hearing, by way of a hybrid hearing at Manchester in August 2021 before EJ Phil Allen, Mrs Mr T Dowling and Dr B Tirohl. Both parties were represented by counsel.

As framed in the list of issues for that hearing, the complaints of direct race discrimination related to (1) the decision to instigate an investigation; (2) the conduct of the investigation; (3) the outcome of the investigation; and (4) the decision to cease offering shifts to the claimant. The tribunal unanimously dismissed the complaint in respect of (1). By a majority (the judge dissenting), it upheld the complaint in respect of (2). By a majority (Dr Tirohl dissenting), it dismissed the complaints in respect of (3) and (4). An oral decision was given and written reasons were subsequently provided. Today I have heard the respondent’s appeal in respect of the outcome of complaint (2).

On 28 January 2019 the claimant was engaged to provide support to a particular service user at the residence where the user lived, on a particular shift. The tribunal described the user as a disabled, vulnerable adult, explained what his disabilities were and described his condition, behaviours and care plan.

When the claimant arrived, the user was in the communal lounge of the residence. There was a factual dispute as to how events then unfolded. In summary, the claimant’s account was that he tried to entice the service user to go to his room, so that the claimant could carry out certain mandatory checks. Giving him crisps and Coke, which the claimant fetched from his room, did not work; but fetching and showing the service user his coat and promising to take him out after the checks had been done, did work. However, by the time the checks were completed, according to the claimant, it was raining and they could not go out.

Three support workers – Mr Hoy, Ms Lomas and Mr Chandler – gave evidence that they were all in the lounge when the claimant arrived. According to them, he began by saying “Come”, to the service user, but Mr Hoy told the claimant that the service user was enjoying his tea and was fine. The claimant then went to the service user’s room and returned with the crisps, which he offered, but then pulled away. Mr Hoy said that he challenged the claimant’s behaviour. The claimant then went and got the coat, partially putting it on the service user, before walking him out of the communal area.

All three of these witnesses told the tribunal that they believed the claimant’s behaviour to be concerning and inappropriate. The claimant’s evidence was that none of them had been present at the start of his shift and they were lying. The tribunal observed that where the accounts differed most significantly, was as to the manner of the claimant’s approach to the service user, and the way the crisps and coat were used, which the respondent’s witnesses considered to be coercive and undermining of the service user’s dignity. However, it was not necessary for the tribunal to determine what factually occurred in this incident.

The tribunal’s account of the aftermath began, at [30]:

Mr Hoy provided a statement on 13 February 2019. The tribunal observed that there was delay on the part of both Mr Hoy and Mr Fenna, which it found reflected in part the lack of seriousness with which the issue was viewed. However, the claimant agreed in evidence that, once the report had been made, the respondent had a duty to investigate it.

The manager of the Bureau, Emma Crewdson, called the claimant about a week after the shift in question, requesting a chat. She told him that it was not a safeguarding issue and there was nothing to worry about. The claimant was then interviewed by Mr Fenna and Ms Crewdson on 15 February. Before continuing the narrative, I interpose that Ms McNulty is a union official who gave evidence for the respondent under a witness order. Mr Maphosa is a support worker who gave evidence for the claimant. The tribunal continued:

Following the meeting on 15 February, the claimant was told by a colleague that it was Ms Chandler and Ms Lomas who had raised the complaint against him. Mr Maphosa told him that white workers were trying to find a way to remove him.

There was a further meeting on 21 February 2019 between the claimant, accompanied by Ms McNulty, and Mr Fenna and Ms Crewdson. Ms Crewdson explained that the outcome of the investigation was that the respondent would not be offering the claimant any more work. The claimant’s initial response was to say that this was race discrimination and that he knew which two people had complained. Mr Fenna said he was not questioning the integrity of the complainants. Ms McNulty said that no-one at that location was racist. In ensuing exchanges, the claimant shouted and said that they were all planting something. Ms McNulty sought to assure him that this was not the case and said that in her experience contracted staff had been removed for less. The claimant said that his accusers had not been present, but that another colleague, Ms Mathuthu, was there.

The tribunal observed that, as the outcome was confirmed at the start of this meeting, what was said during the course of the meeting did not have any bearing on why the decision to cease using the claimant’s services had been made. Following the meeting, no-one spoke to Ms Mathuthu and, as the tribunal put it, nothing whatsoever was done in response to the claimant’s allegations of race discrimination. The tribunal observed at [62]:

In the next section, the tribunal commented on the evidence of some individual witnesses and wider contextual and background issues. It accepted Mr Fenna’s evidence that he believed that there were two other people who corroborated Mr Hoy’s account and who they were, while noting also the absence of any attempt to take statements from them. It referred to Ms Crewdson’s evidence that the complaint had come from a very experienced member of staff, Mr Hoy, who she believed, and she emphasised what she perceived to be the claimant’s lack of understanding or awareness of what he had done wrong. The tribunal observed that this was consistent with the evidence of Mr Fenna.

The claimant’s named comparators were Ms Ballance, Mr Clayton, Mr Rushton and Mr Trimble. In the next passage, the tribunal made some findings of fact about what happened to each of them, but also about the case of another individual who was not a pleaded comparator, but to whose case the claimant had referred in his witness statement, Mr Ellis.

The claimant contrasted the conduct of Mr Fenna in meetings with him, with his conduct towards white workers such as Mr Ellis, a casual support worker who was removed after he had left his shift to go home and sleep. The notes it had of the fact-finding meeting in that case, conducted by Mr Fenna and Ms Crewdson, recorded that it began with open questions and what appeared to be an even-handed approach to obtaining Mr Ellis’s account. However, later in the meeting, Mr Fenna was recorded as saying that Mr Ellis’s judgment was totally off, that he had accepted responsibility and that it was pretty obvious that his actions should have been hugely different. The tribunal observed that Mr Fenna made his views clear in this way during the course of the meeting.

After a self-direction as to the law which is not criticised as such by either side, the tribunal set out its conclusions. It concluded that Ms Ballance and Mr Trimble were not relevant comparators because their circumstances were materially different. Mr Clayton was a casual worker relied upon only in relation to the claimant’s dismissal, but the respondent had also decided to cease using his services, and so he was not treated less favourably than the claimant. Mr Rushton was a casual support worker who was investigated in relation to two matters but remained on the Bureau. The tribunal concluding that the circumstances in relation to both those matters were materially different.

The tribunal continued:

The tribunal unanimously concluded that, once the complaint had been made, the respondent had a duty to investigate it and would have done so irrespective of the race of the subject. It therefore dismissed that particular complaint. Up until this point, the whole decision was unanimous. However, the remaining complaints were decided by majorities. In relation to the outcome of the internal investigation, Dr Tirohl was in the minority. The majority’s reasoning was as follows:

At [108] the tribunal analysed the further complaints before it in this way:

However, the tribunal stated that the material evidence and considerations were exactly the same in respect of the outcome of the investigation and the dismissal decision, because the latter, whether made collectively or by Mr Bentley, was entirely informed by, and reliant upon, the former. So, the majority reached the same decision in relation to the investigation outcome and the dismissal, as did the minority. In view of that I will from here on refer to the outcome of the investigation and the decision not to offer the claimant any more shifts compendiously together as “the dismissal”.

As to the conduct of the investigation, I will set out the relevant passage in full, including the parts that were unanimous and then the further reasoning of both the majority and the minority:

I pause to note that both the majority and the minority therefore treated the complaint about the “conduct of the investigation” as embracing both what happened following Mr Hoy’s report, including the investigation meeting on 15 February, and also the failure to take any further follow-up action on the claimant’s allegation of race discrimination made at the second meeting on 21 February.

There are six grounds of appeal before me. The first three were permitted to proceed on the sift by HHJ Wayne Beard and the last three at a Rule 3(10) hearing before HHJ James Tayler. An answer was put in in respect of the first three but not the second three, but Mr Lassey took no point about that. I note also that the answer raised a cross-appeal in respect of the tribunal’s dismissal of the complaints relating to what I compendiously call the dismissal, but this was considered not to be arguable by HHJ Martyn Barklem and a Rule 3(10) hearing was not then sought in that regard.

Ground one challenges the majority’s conclusion that there were sufficient facts to cause the burden of proof to shift to the respondent. In substance, this ground has two strands. The first strand contends that the tribunal erred in treating the respondent’s failure to speak to appropriate witnesses, check CCTV footage or follow up on the claimant’s complaint of race discrimination as capable of shifting the burden of proof, as these were themselves features of the very conduct of which the claimant complained. The second strand contends that the tribunal erred by treating its conclusion that the respondent had acted unreasonably or unfairly as sufficient to shift the burden of proof. Reliance is placed on Bahl v The Law Society [2004] IRLR 799 and Zafar v Glasgow City Council [1998] IRLR 36 HL. A further strand of this ground, in paragraph 8, refers to failure to have regard to the attributes found of the hypothetical comparator, but it seems to me that this overlaps with a point raised, and which I will address, in relation to ground two.

Ground two challenges the majority’s decision at the second stage, on the assumption, contrary to ground one, that it was not wrong to find that the burden of proof had shifted. Again, it has two strands. First, it contends that the tribunal failed to engage with whether the respondent had provided a non-discriminatory explanation for the treatment. The majority had concluded that the failure to follow up on the race discrimination allegation was indicative of closed minds. They should have considered whether that was in turn a reflection of the fact that, from the very outset, Mr Hoy had been believed, and whether that provided a non-discriminatory explanation for the manner of conduct of the first meeting and the failure to follow up after the second meeting. This ground submits that they should have concluded, as the minority did, that it did provide such an explanation.

The second strand contends that, the tribunal having found, albeit by a majority, that consideration of the hypothetical comparator led to the conclusion that the dismissal was decided upon for a non-discriminatory reason, the majority who determined the complaint relating to the conduct of the investigation, then provided no explanation as to why their assessment of the same factors did not lead them to the same conclusion. Ground three contends that the reasoning of the majority is, in the same respect, not compliant with Meek v Birmingham DC [1987] IRLR 250.

Ground four contends that the reasoning of the majority is perverse. Again, this has two strands. The first is to the effect that there was no evidence that the investigation was conducted in the way that it was, because of anything other than the fact that Mr Hoy was believed from the outset. Indeed, it is said that the majority implicitly accepted that this was the reason, in its reference to the explanation for the lack of follow-up on the second meeting being closed minds. The second strand is, once again, that the majority’s reasoning in relation to this complaint was inconsistent with the reasoning and conclusion of the majority that decided the complaints relating to the dismissal.

Grounds five and six contend that the majority erred with respect to the reliance placed by it upon what happened in the cases of Mr Rushton and Mr Ellis. In relation to both matters of which Mr Rushton was accused, the tribunal had found material differences from the claimant’s case. The majority then erred by nevertheless attaching significance to the manner of conduct of Mr Rushton’s investigation meeting. The tribunal also erred because, although he referred to him in his witness statement, the claimant did not himself rely on Mr Ellis as an identified comparator, nor did his counsel in closing submissions. So the respondent’s counsel was not alerted to the feature of his case to which the tribunal, in its decision, attached significance. Had he been forewarned, he would have sought to rely on the evidence of Ms Crewdson, as to the even-handed manner in which the claimant had himself been treated in earlier fact-find investigations affecting him. But the tribunal held at [21] that it did not need to consider what it described as these historic issues at all.

In his skeleton and in oral argument, Mr Lassey relied on the reasoning of the employment judge in the minority, which he said did not exhibit the same inconsistency with the tribunal’s conclusions on the dismissal complaints that the reasoning of the majority did.

In the answer and in response, in summary, to ground 1, Mr Welch submitted that the tribunal had properly relied upon all the facts found as providing the something more that shifted the burden of proof to the respondent. In relation to ground two, it could not be inferred that the tribunal had failed to engage with the respondent’s reasons for the treatment of the claimant. Other features of the evidence were said to be relevant, including that Mr Hoy had only made the complaint when asked to put it forward and that there had been a failure to take into account that English was not the claimant’s first language, when judging his behaviour. He submitted that it was understandable that the minority had not referred specifically to the characteristics of the hypothetical comparator, bearing in mind that these had been fully set out earlier. Bearing in mind the guidance not only in Meek but in DPP Law Limited v Greenberg [2021] IRLR 1016, the answer submitted that there was sufficient reasoning given on this point and the decision was not non-Meek compliant.

In relation to grounds four, five and six, Mr Welch frankly acknowledged that there were some inconsistencies in the overall decision, no doubt attributable, he said, to the fact that different complaints had been decided by different compositions. But this did not necessarily render the decision overall perverse, which was a high hurdle. The tribunal was entitled to attach weight to the facts found regarding the conduct of the meetings with Messrs Rushton and Ellis, and the parties had been free to make submissions, if they wished, about the reliance that should or should not be placed, in particular on the evidence about the treatment of Mr Ellis.

In relation to ground one, I do not agree that, within the four walls of its consideration of the conduct of the investigation, the majority erred by concluding that the burden of proof shifted. I do not need, on this occasion, to go over the discussions in Zafar, Bahl and Igen v Wong [2005] ICR 931, as to what significance may or may not be attached to a finding that a respondent has acted unreasonably or unfairly. There is a wider and more general point with regard to the approach to be taken by a tribunal to deciding whether there are sufficient facts to cause a shifting of the burden of proof for the purposes of section 136 Equality Act 2010. This relates to the frequently-cited dictum of Mummery LJ in Madarassy v Nomura International plc [2007] IRLR 246 CA, that something more than a difference in characteristic and a difference in treatment is required.

As has been said on a number of occasions, including recently in Jaleel v Southern University Hospital NHS Foundation Trust [2023] EAT 10 at paragraph 38, that dictum is not a rule of law. It was by way of a response to a submission that the burden automatically in all cases is shifted, merely by a difference in statement and in treatment. But in no case do such features, if present, occur in a factual vacuum. The process of considering whether or what inference to draw and whether the burden shifts is situation- and fact-specific. The task of the tribunal is always to apply the law, including section 136, to the particular facts and features of the case before it.

In this case, the tribunal relied on a number of features, including not interviewing other witnesses who were mentioned by Mr Hoy, and not looking at the CCTV, as well as the conduct of the initial investigation meeting itself and the failure to follow up on the claimant’s later race discrimination allegation. I do not accept Mr Lassey’s submission that these matters could not be relied upon because they all were, one way or another, features of the conduct complained of itself.

In some cases the conduct complained of may not be intrinsically such as to raise eyebrows if viewed in isolation. In such cases, the tribunal may, however, find some further feature which forms part of the wider factual matrix, shifting the burden. At the other end of the spectrum are cases where the conduct complained of, if factually found, is inherently or overtly discriminatory, such as the use of unambiguously racist language. In such cases, the tribunal does not need to consider the shifting of the burden or to look further than the fact of the conduct itself.

But there may be cases in between, where there are features of the factual conduct complained of which are such that the tribunal can properly consider that they could support an inference of discrimination in the absence of a non-discriminatory explanation. I do not think it is attractive or necessary to require a tribunal, in cases that appear to be of that type, to try to work out and separate out what is an aspect or feature of the conduct complained of from what is not. The tribunal’s task is simply to have regard to the overall factual circumstances of the case, and to consider whether those facts provide a sufficient basis for it to conclude that they could support an inference of discrimination in the absence of a non-discriminatory explanation. In the present case, the range of factual features relied upon by the majority were sufficient to support their conclusion that the burden had shifted.

However, at the heart of this appeal is the proposition that, one way or another, the outcome and supporting reasoning in respect of the complaint relating to the conduct of the investigation, is simply inconsistent with the outcome and supporting reasoning in respect of the complaints relating to the dismissal. That essential proposition is foreshadowed in paragraph 8 of ground one, on the basis that the assessment of how the hypothetical comparator would have been treated in the conduct of the investigation, should have been informed by the assessment of how the hypothetical comparator would have been treated in relation to the dismissal. It is advanced by ground two on the basis that the majority failed to engage with the non-discriminatory explanation provided by the respondent, which was said to be essentially the same as the non-discriminatory explanation that had been accepted in relation to the dismissal. Essentially the same point is put as a Meek challenge by ground three and a perversity challenge by ground four. I have come to the conclusion that this feature of the tribunal’s overall decision does mean that this appeal must be allowed. My reasons are as follows.

First, I do not think it would inevitably have been inherently inconsistent, had the tribunal concluded that the respondent did not discriminate in coming to the decision to cease to use the claimant’s services, but then did subsequently discriminate in the way that it reacted or failed to react to his alleging race discrimination. But that still leaves the challenge to the conclusion that the respondent did not discriminate in relation to the conduct of the investigation prior to the decision to dismiss. As to that, I observe that what the tribunal majority said about the outcome of the investigation and the decision to dismiss, at [104] and [106] respectively, although presented as alternatives, was in reality, in my view, all part of the same reasoning process. Either the hypothetical comparator would not have been treated differently and the reason for the treatment was not race, which is what that majority concluded; or it might have concluded, but did not, that the hypothetical comparator would have been treated differently and the reason for the treatment was race.

What the majority concluded overall was that these particular decisions were taken wholly because the factual allegation, coming as it did from the trusted Mr Hoy, was believed, because the claimant disputed it at the 15 February meeting rather than showing the insight, contrition and willingness to learn that Mr Fenna expected of him, and because it was considered that, as the claimant was a casual support worker, the respondent could decide to stop using his services without the need to follow the same processes as it would, had he been an employee, whether having regard to its own procedures or because he would in that case have had the right to claim unfair dismissal.

The respondent’s case was essentially that one or more of those same reasons explained the features of the conduct of the respondent prior to and at the 15 February meeting of which the claimant complained. Both Mr Fenna and Ms Crewdson went into that meeting having already decided that they believed Mr Hoy’s account, and that it was not necessary to follow the more thorough process that they would have to follow in relation to an employee. This also explained why it did not follow up on the claimant’s allegation of race discrimination. Its processes did not require it to do so, and Ms Fenna and Ms Crewdson, accepting Mr Hoy’s account as they did, did not believe there was anything in that allegation.

Of course, it was for the tribunal to evaluate the respondent’s overall case, which it was not bound to accept. However, as the tribunal accepted as a fact that both Mr Fenna and Ms Crewdson believed Mr Hoy’s account, so that they did not feel the need to hear from the claimant or investigate further, and it found that this fact, together with the claimant’s attitude at the 15 February meeting, and his status as a casual worker, provided the whole non-discriminatory explanation for the dismissal decision, it must be inferred that neither what it considered to be the unreasonable conduct of the 15 February meeting, nor what it considered to be the unreasonable failure to follow up the claimant’s allegations of race discrimination made at the 21 February meeting, were regarded by the majority which decided the complaints relating to the dismissal, as sufficient to undermine the non-discriminatory reasons put forward for the dismissal. I cannot see how that can then be reconciled with the conclusion that these same features did undermine not only the non-discriminatory explanation put forward for the lack of proactive response to the claimant’s race allegation, but also, by casting light back on it, for the earlier conduct of the investigation.

I note that it appears to me that the claimant’s case before the tribunal appears to have been that the reason that Mr Hoy was believed was itself, consciously or not, also influenced by race. But, as the majority which decided the complaints relating to the dismissal did find that a factor contributing to the decision to dismiss was that Mr Hoy’s account was believed, but also regarded that as a non-discriminatory explanation, it cannot have considered that the acceptance of Mr Hoy’s account as true, was itself to any degree influenced by race. Had it thought so, it would surely have said so and its substantive decision on that complaint would then have been different.

I am therefore of the view that the reasoning leading to these two decisions simply cannot be reconciled. Mr Welch candidly surmised that the real world explanation for how this came about was that these two parts of the decision were reached by different majorities. But I do not need to come to any view about that. As Mr Welch properly accepted, it does not assist his client’s resistance of this appeal that these two decisions were reached by different majorities. A decision on a given complaint, reached by a majority, is then the tribunal’s decision on that complaint, and the majority’s reasoning is then the tribunal’s dispositive reasoning. The information that those decisions were reached by different compositions may explain, but does not resolve, their inherent irreconcilability.

Further, and in any event, I consider that the majority decision relating to the conduct of the investigation was not Meek-compliant. Given the decision of the majority on the complaint relating to the dismissal, and its supporting reasoning, if the different majority who reached the decision on the conduct of the investigation had different reasoning explaining why those factual features did not lead to the same conclusion, it needed to provide more of an explanation for that than it did at [115].

As I have noted, all of the live grounds of appeal that are before me relate to the decision of the majority in respect of the conduct of the investigation. The decision and reasoning in respect of the complaints regarding the dismissal (in both its stages), are not part of any live challenge before me. So, if these two parts of the overall decision cannot, as I have concluded, stand together, it is the decision in relation to the complaint relating to the conduct of the investigation that falls away.

As I am therefore allowing the appeal, I can deal more briefly with the remaining matters raised by grounds five and six. Having found that the circumstances in relation to the two matters relating to Mr Rushton were materially different in relation to both the conduct and outcome of the investigation in his case, I agree with Mr Lassey that the tribunal was wrong then to place weight on the evidence that it had relating to the conduct of his interview. At any rate, the tribunal would have needed to make more nuanced fact-finding if it considered that some aspects of the treatment of Mr Rushton were not materially comparable to the claimant’s case, but yet other aspects still were.

In relation to Mr Ellis, again the majority did not make a nuanced finding distinguishing aspects of the early conduct of the investigation meeting with him from the later conduct of it after Mr Fenna had heard his explanation. In addition, I see force in the submission that the respondent’s counsel should have been alerted to the particular feature of his case to which, in the event, the tribunal attached significance, as this was not a feature on which the claimant had relied.

In all events, for the reasons I have already given, this appeal must be allowed. What has happened in this case is singularly unfortunate. The tribunal’s decision overall shows a conscientious and insightful approach being taken by all the tribunal members, no doubt drawing on their combined experience and wisdom, to the complaints in this case. Whilst tribunals should always strive for unanimity, this is not always possible and the system allows for majority decisions, though they are relatively rare. But as I have said, where any element of a decision, whether as to fact-finding or conclusions, is reached by a majority, that stands as the decision of the tribunal on that point, and the reasoning and decision of the tribunal on all points before it on the same occasion must be consistent.

The appeal must, therefore, for the reasons I have given, be allowed.

I have heard further submissions as to what order I should make. Mr Lassey submits that, in light of the parts of the tribunal’s decision that survive this appeal, and my reasons for allowing the appeal, there is only one possible correct outcome on the question of whether there was direct race discrimination by the respondent’s conduct of the investigation, being that there was not. He therefore submits that I can and should substitute a decision to that effect. Mr Welch submits that I cannot and should not substitute, but should remit.

I have come to the conclusion that Mr Welch is right. I have found that the decision appealed was erroneous because it was inconsistent with the other decisions reached and/or that the reasons given were not Meek-compliant. My finding of inconsistency causes me to wonder whether, given the other findings and conclusions that still stand, the tribunal could on remission reach a consistent decision that the conduct of the investigation amounted to direct race discrimination. But the inconsistency lies in part in the particular reasoning, and I cannot say that that is logically impossible.

For that reason I will quash the decision under appeal, that the claimant was treated less favourably on the grounds of race by the respondent in the conduct of the investigation, but I will remit that matter for fresh determination by the employment tribunal. Both counsel agreed that in that case I should direct that upon remission the matter should be determined afresh by a differently constituted panel. I agree and will so direct. It would be hard to expect the previous panel to tackle this question afresh, given how their previous decision came out.

Both counsel also agree, as do I, that I should not give any further prescriptive directions as to case management upon remission. Finally, I note that both counsel have agreed that what will be a given, on remission to the fresh tribunal, will be the Allen tribunal’s unanimous findings of fact, and the reasoning and conclusions dismissing the complaints of direct race discrimination in relation to the decision to instigate an investigation and in relation to the outcome of the investigation and the decision to cease offering the claimant shifts on the Bureau, as these are undisturbed by this appeal.