A McDermott v Sellafield Ltd & Ors.

I will refer to the parties as they were in the employment tribunal, as claimant and respondents. The broad background and context, which I draw from the tribunal’s decisions, are as follows.

The claimant is a consultant specialising in Equality, Diversity and Inclusion (EDI). The first respondent operates the Sellafield site in Cumbria. It has since 2016 been a wholly-owned subsidiary of the second respondent, which is a statutory body responsible for decommissioning nuclear sites. The third respondent was, at the relevant time, the first respondent’s HR Director. From 2016 the services of the claimant were supplied to the second respondent by Capita, initially as an agency worker, and then via her own company, Interim Diversity Limited (IDL). The first respondent then contracted with IDL for the provision of the claimant’s services to it, for an 18-month fixed term from 1 September 2018, subject to termination on notice “for convenience”. On 29 October the required notice was given to terminate that contract with effect on 27 November 2018.

The claimant complained that she had been subjected to detriments on the grounds of protected disclosures, contrary to section 47B Employment Rights Act 1996, (the PD complaints) and, relying on the same factual case, victimised contrary to section 27 Equality Act 2010 (the victimisation complaints). The treatment complained of included termination of IDL’s contract.

It was common ground: (a) that the claimant was a worker of the first respondent within section 43K of the 1996 Act and a contract worker supplied to the first respondent within section 41 of the 2010 Act: (b) that the second respondent could not be liable to the claimant in respect of her PD complaints under the 1996 Act; and (c) that the second respondent could be liable to her in respect of victimisation by the first respondent, if found, but only if the first respondent had acted as its agent, or it had been instructed, caused or induced so to act by the second respondent, or the second respondent had aided such victimisation (pursuant to sections 109, 111 or 112 of the 2010 Act).

The complaints came to a 13-day full-merits hearing before Employment Judge Lancaster, L Anderson-Coe and K Lannaman, sitting at Leeds. In a reserved decision the tribunal dismissed all of the complaints. In a further decision in respect of costs applications made by all three respondents, which were considered on paper, the tribunal ordered the claimant to pay £20,000 towards the costs of the first and third respondents and £20,000 towards the costs of the second respondent.

I heard the claimant’s appeals against the liability and costs decisions, by way of ten grounds relating to the former and three relating to the latter. Mr Allen KC appeared for her, James Arnold of counsel having appeared before the tribunal; as before the tribunal, Mr Panesar KC and Ms Levene of counsel appeared, respectively, for the first and third respondents, and for the second respondent.

Section 43B(1) Employment Rights Act 1996 provides:

Section 48(1) gives a worker the right not to be subject to a detriment by any act, or any deliberate failure to act, by his employer, done on the ground that the worker has made a protected disclosure. Section 48(2) provides that, on such a complaint, it is for the employer to show the ground on which any act, or deliberate failure to act was done. As noted, it was common ground that any qualifying disclosure made by the claimant to the first respondent would be a protected disclosure, and any detrimental treatment of her by the first respondent would be within scope of section 48(1).

Section 27 Equality Act 2010 includes:

Section 41 includes provision that a principal must not victimise a contract worker. As I have noted, it was common ground that the claimant was a contract worker of the first respondent.

The tribunal used the word “disclosure” to refer to each communication the claimant relied upon as amounting to a protected disclosure, a protected act, or both. Originally, there were eight claimed disclosures, but disclosure 5 was ultimately withdrawn and, for the purposes of this appeal, I am not concerned with disclosures 7 and 8. How the remaining claimed disclosures were described in the list of issues, was set out by the tribunal in the following paragraphs.

I interpose that the particulars of claim and list of issues identified that, in respect of all the claimed PDs, the “relevant wrongdoing” for the purposes of section 43B(1) was said to be actual or likely contravention of sections 13, 15, 26, 27 or 149 of the 2010 Act, and/or of common law duties of care to employees including the implied duty of trust and confidence, and/or that the health and safety of employees had been, was being, or was likely to be endangered, and/or that such matters were being, or likely to be, concealed by the failure of senior leadership and HR to address them.

The alleged conduct, said to amount to detrimental treatment for the purposes of the claimant’s complaints, was set out by the tribunal in the following paragraphs.

The issues raised by the liability appeal focus on the tribunal’s decision in relation to disclosures 1 and 6, and, in one respect, disclosure 4; and on its decision in relation to detriment 1 and those detriments that related to the termination of the contract with the claimant’s company, which, for shorthand, it referred to, as shall I, as the dismissal. In describing the facts found at this stage, I will therefore focus on these aspects of the decision.

The tribunal began with events relating to disclosure 1, in the following passage.

The tribunal went on to make findings about initial conversations the third respondent had with a number of colleagues who were identified in the Safecall Report as potentially having information about VC’s alleged behaviours. It concluded that the assertion in the particulars of claim, that she told the claimant on 12 September 2018 that she had “sat on this report for weeks” and taken no action was “demonstrably false”, as she had, at that point, only known about it for two days and had already taken steps in connection with it. The tribunal continued:

The claimant sent the third respondent her draft terms of reference and further draft questions she proposed to ask participants, which did “in part focus attention directly upon levels of satisfaction” with the HR lead team and the third respondent [68]. The focus groups “did in fact afford an opportunity for participants to have reported any express concerns about VC.” [70] During the period of the sessions the claimant reported to the third respondent that one participant – SP – had spoken of historic concerns regarding VC. Another – FS – told the claimant confidentially of “gossip” that VC had done something that had made others uncomfortable.

The tribunal also referred to an anonymous letter of 8 October 2018, ostensibly by members of the HR department, to senior managers in the first and second respondents, which referred to problems re sexual harassment being ignored. It alleged that the third respondent had taken no action in relation to someone having recently raised such an issue – which the tribunal considered that anyone aware of it would have connected to the Safecall report.

Upon her return from compassionate leave in mid-October the third respondent sought an update from the claimant on information about VC obtained from the focus groups, was told that only FS had expressed concerns, and then queried the situation regarding SP. The claimant sent her draft report to the third respondent on 16 October. The claimant noted that she (the claimant) had “not included the issue re VC” and did not want to put anything re FS in writing as their comments were made confidentially. On 18 October D Vineall, Group Chief People Officer of the second respondent, spoke to the third respondent about the anonymous letter. She told him that, on advice from Mr Carr, she was using the claimant to carry out focus groups. The tribunal accepted that she was, at that stage, awaiting any further information from the claimant regarding SP, and had yet herself to speak to FS.

In this passage the tribunal also noted that, subsequent to the period with which it was concerned, an executive director of the first respondent carried out 19 fact-finding interviews and on 21 December 2018 found that there was no case to answer in relation to the Safecall allegations.

The tribunal made detailed findings of fact in relation to disclosures 2 and 3. It suffices to note that disclosure 2 related to an alleged conversation between the claimant and the third respondent about an email alleging that a colleague had not been treated with respect by a then unnamed member of the HR team (identified to the tribunal). The tribunal went on to conclude in light of its findings of fact, that the claimant’s account of this was a “well-nigh total distortion” of what happened, and that there was no disclosure of information. Disclosure 3 related to an alleged conversation with an HR team member, about allegations of homophobic abuse that had been made by two employees. The tribunal found that that particular conversation with the team member did not happen.

Disclosure 4 related to an employee, CL, who had been on long-term sick leave since May 2018 with mental health issues, and who had previously had lengthy absences because of certain physical ailments, and who the tribunal said would almost certainly have met the definition of someone with a disability. The disclosure was said to arise from a colleague expressing a concern to the claimant about his well-being, leading to the claimant speaking to a member of the HR team, L Bowen, about him. The tribunal said of this:

After making some further findings relating to this incident, including as to communications between the claimant and another HR colleague, A Rankin, the tribunal continued:

The tribunal recorded that disclosure 5 was no longer relied upon, and stated that the email in question (the text of which it set out) could never have properly been construed as a disclosure that the third respondent had breached the confidentiality requirements of the Safecall procedure.

At the start of its findings about disclosure 6 the tribunal found that, following the away day on 5 September 2018, the claimant expressed to some members of the HR team, criticism of another member, A Thompson; and she also admitted in evidence that she had by that stage formed the view that the third respondent should not remain in post. This view was based upon two earlier interactions in respect of which it preferred the third respondent’s evidence as to what occurred over that of the claimant. In the focus groups she invited negative feedback in relation to A Thompson, and inappropriately directed her questioning to securing adverse criticism of the third respondent. The tribunal stated that the claimant had produced no notes of any of the feedback sessions.

The third respondent discussed the receipt of the anonymous letter with the lead team on 16 October. The claimant initially emailed a draft of her report to the third respondent that day. In further text exchanges between them the third respondent said she was a bit disappointed “however if you ask you have to be prepared for the answer.” The claimant wrote “I think there are some simple things that could be done to really demonstrate to people you (plural) have heard the issues.” The third respondent indicated that she would be discussing it with the HR lead team on 18 October and that “we will look at a plan of 2 – 3 things to do immediately and the option of a dept wide brief.”

As to the contents of the report, the tribunal said this.

The tribunal accepted the third respondent’s evidence that, when she shared the report with the lead team, they were shocked and horrified, and expressed concerns about the negative way it was written. Nevertheless it was clear that some action was warranted, and everyone agreed to take time to reflect. I need to set out the next passage in the tribunal’s findings of fact in full.

The tribunal went on to find that there was no further contact from the claimant save to submit her final invoice, which was not disputed, and was paid. The email confirming payment “also confirmed that the reason for termination of the IDL contract was financial.”

The tribunal went on to make findings about claimed disclosures 7 and 8, to D Vineall of the second respondent, to which this appeal does not, as such, relate. However, this passage includes the following paragraphs of relevance to this appeal.

The tribunal also found that the third respondent informed the HR director of the second respondent that there was no hard evidence to justify investigating the Safecall report and that the reason for the termination of the IDL contract was financial. D Vineall was also told by the CEO of the first respondent that the reason was financial. They also agreed that the second respondent would pay for a diagnostic review of the HR function, which led to the commissioning of a report from Price Waterhouse Coopers (PWC). The second respondent also commissioned a third party investigation into the Safecall report following the identification of a potential witness who was on secondment.

The tribunal opened the section of the decision setting out its conclusions as follows.

In relation to disclosure 1, and the claimed detriments relating to it, I will set out the tribunal’s conclusions in full.

In relation to disclosure 2 the tribunal found the claimant’s case to be “a well-nigh total distortion of what actually happened.” It was only claimed to involve a protected disclosure, not a protected act, but it involved no disclosure of information. Failure to investigate it was no longer pursued as a detriment. The claimant was not subjected to a detriment because of it. The substantive matter raised was investigated. It did not have any bearing on the decision to dismiss.

Disclosure 3 was found not, factually, to have occurred; and even on the claimant’s case, the tribunal concluded, there was no disclosure of information or protected act. Even had the claimant, on her own case, exhorted L Bowen to take immediate action in relation to the matter, she was not subjected to any detriment because of having done so.

As to disclosure 4, the tribunal found that the passing comment asking L Bowen to explore if everything was being done did not disclose information and did not amount to the doing of a protected act and did not allege discrimination. It was not a significant conversation that registered in the mind of L Bowen. The claimant suffered no detriment. She was reassured by L Bowen, who said she would look into it.

As to disclosure 6 and the related complaints I need to set out the tribunal’s conclusion in full.

In the course of its conclusions regarding disclosures 7 and 8, the tribunal also concluded that there was no evidence that the second respondent had induced the termination of the IDL contract or any other alleged detriment, nor did it knowingly help the termination of the IDL contract to take effect after the service of notice. So, even had there been a basic contravention, the second respondent could not have been liable therefor under sections 111 or 112 of the 2010 Act. It continued:

The tribunal therefore concluded its decision as follows.

Grounds 1 to 4 all concern the tribunal’s conclusions about whether the claimant made protected disclosures, specifically, Mr Allen KC confirmed in oral submissions, by claimed disclosures 1 and 6. An overarching theme of this group of grounds is that the tribunal failed to follow the structured approach commended in Williams v Brown, UKEAT/0044/19/OO at [9] – [10], of working through each of the elements of the definition of a qualifying disclosure, and determining whether each in turn was satisfied. Strikingly, submitted Mr Allen KC, the tribunal did not, in its decision, cite the words of the statute, nor any authority or guiding principles. Whilst he accepted that a failure to state the law did not, of itself, show that the tribunal had not understood or applied it correctly, he submitted that this meant that the EAT lacked the tangible reassurance that the tribunal had done so; and the lack of a structured approach did lead the tribunal into error.

Ground 1 contends that the tribunal erred by failing to find that disclosure 1 and disclosure 6 amounted to disclosures of information.

As to disclosure 1, Mr Allen KC submitted that the information which the tribunal should have found was conveyed was that the claimant “considered that a formal investigation was called for (and not just ‘carrying out focus groups’) when allegations had been made of sexual harassment by a senior HR Lead Team member and a cover up.”

The expression of an opinion can convey information: Norbrook Laboratories (GB) Limited v Shaw [2014] ICR 540. Relevant context included: that – on the claimant’s case – there was a “toxic culture” at Sellafield, yet the tribunal took the view at [23], that that was not an issue; that the claimant had worked for the first respondent since 2016 and been praised for her work; that there had been recent reports, on EDI, HR Engagement and Employee Engagement at Sellafield, which the third respondent knew had been critical of the work environment and the HR function; that the third respondent’s evidence was that she herself had concerns about the leadership team; that Mr Rankin, the HR Transformation Lead, also had critical views of the HR Team; and that an effective HR function was critical to the safe running of a nuclear site.

Mr Panesar KC submitted that, while an expression of opinion may convey information, the tribunal properly concluded that, on this occasion, it did not. Further, there was an important difference between the factual case advanced, and the facts found. The claimant’s case was that the third respondent had told her that she had “sat on this for weeks” and the claimant had then responded that there should be an investigation conducted “immediately”. But the tribunal found that the third respondent had not so stated. It also found that the claimant understood that the focus groups would not be instead of a formal investigation, but by way of the next stage of the process. In the context found, the tribunal did not err by finding that the claimant had merely expressed an opinion.

Further, he submitted, the tribunal was plainly aware of all of the wider contextual features relied upon by the claimant, all of which were referred to in evidence and submissions, and some of which were referred to in its decision. It did not need to refer to them all; and it was not bound to conclude that any of them indicated that this was a disclosure of information in the requisite sense.

My conclusions in relation to this strand of ground 1 are as follows.

As is well-known to employment lawyers, in Cavendish Munro Limited v Geduld [2010] ICR 325 the EAT postulated that, in this context, there is a significant distinction between disclosing information and making an allegation. At [24] it said this:

In Kilraine v London Borough of Wandsworth, Geduld was considered by the Court of Appeal. Sales LJ, as he then was, (Kitchin LJ as he then was concurring) said this:

Sales LJ went on to say that the EAT in Geduld had been right to hold that the letter in that case was devoid of any sufficient specific factual content such as to bring it within section 43B(1). However, he agreed with Langstaff J in Kilraine, that it was wrong to treat a communication as being either “information”, and so within section 43B(1) or an “allegation”, and so not. He continued:

Accordingly, whether there has been a disclosure of information depends on whether the communication has “sufficient factual content and specificity” so as to be capable of showing one of the matters listed in sub-section (1) – that is, capable of tending to show relevant wrongdoing within section 43B(1). Just as some communications which may be described as communicating an allegation will pass that test, but some will not, so some communications which may be described as communicating an expression of opinion may pass that test, but others will not.

In the present case, it appears to me that the tribunal did not make the principled error of assuming that, just because the communication conveyed an expression of opinion, it could not also have communicated information sufficient to fulfil the concept of a qualifying disclosure. Rather, it concluded that it was “no more than” an expression of opinion [213] as well as not even being “an allegation such that in an appropriate context it might nonetheless qualify for protection” [214].

However, in so concluding, it appears to me that the tribunal was particularly influenced by its finding that the claimant “did not have any actual knowledge as to the substance of the [Safecall] report” (at [210]), and so what she said was therefore “no more than an expression of opinion” ([213]). It also relied on its findings that she was not told that the third respondent had been sitting on the matter for weeks, and [211] that she understood that the focus groups were “intended to form the first stage of any investigation”, there being nothing in the conversation to suggest that, had substantiated allegations come to light, they would not have been subjected to a formal process [216].

However, in focussing on these aspects the tribunal failed to address itself to the fact that the claimant did know the fact, as such, that an allegation of sexual harassment had been made against VC through the Safecall facility, including a suggestion of a cover up. She also knew that it was being proposed that there should be focus groups. I do not think the tribunal’s reading of the claimant’s use of “if” as denoting conjecture or uncertainty is tenable. As the tribunal found, she was told by the third respondent that this was what had happened and what was proposed. The plain sense of “if” in her response, is: “in that case”. The claimant was playing this information back to the third respondent, as part of what she was saying. As Mr Allen KC correctly submitted, it does not matter that the third respondent already knew it: see section 43L(3).

The features relied upon by the tribunal appear to be more pertinent to the question of whether the claimant reasonably believed that the information (if information it be) communicated by her tended to show relevant wrongdoing. As Sales LJ explained in Kilraine, the nature of the relevant wrongdoing contemplated by the statute may provide a reference point when considering whether the communication was of information in the relevant sense. But the tribunal still needed as a starting point to engage with the claimant’s case as to the nature of the information embedded in what she said, in order to be able to assess whether that was the sort of content that she could potentially regard as tending to show relevant wrongdoing, and, if so, whether she in fact so regarded it, and, if so, reasonably so regarded it.

Further, it appears that the tribunal’s conclusion was that the sequence of the conversation was that the third respondent told the claimant about the Safecall report, and asked her to carry out the focus group exercise, the claimant then responded with disclosure 1, and the third respondent then responded in turn by explaining that this approach was being adopted at this point, because of the difficulty created by the report having been anonymous, and in light of the advice she had received from Mr Carr. The tribunal needed to judge whether the claimant was disclosing information, at the point when she made the disclosure. It was of course entitled to take into account the sequence in which the conversation overall unfolded and concluded, when considering such matters as whether the claimant was subjected to detriment for this claimed disclosure, but that is a different matter.

I test the matter this way. Suppose the claimant had, instead of saying what the tribunal found she said to the third respondent, at that moment left and gone to see another manager, and informed them that the third respondent had told her that there was a Safecall report of an allegation of harassment against VC and of a possible cover up, and that the third respondent had asked her to carry out focus groups, but she, the claimant, considered that there should be a formal investigation. That would surely have been sufficient to amount to a disclosure of information. This is the equivalent, in terms of content and meaning, to what she said to the third respondent; she just did not need to spell out to the third respondent what she was, in substance, referring to.

The tribunal therefore in my judgment took an erroneous approach to its consideration of whether this disclosure was a disclosure of information. I reach this conclusion without needing to consider the background or contextual matters relied upon by Mr Allen KC. I agree with Mr Panesar KC that it is safe to assume that the tribunal was aware of all of them, nor was it obliged to make a finding of its own as to whether the claimant was right to consider that there was a general “toxic culture” at Sellafield. I am inclined to think that they are in any event, if relevant, more pertinent to the issue of reasonable belief. But for the reasons I have given, I uphold the first part of ground 1.

The second part of ground 1 relates to the tabling by the claimant of her HR Survey Report in October 2018. The tribunal is said to have erred by failing to address whether the report disclosed information; failing to conclude that it disclosed information that the HR Lead Team and Director were viewed by the vast majority of respondents to the survey as broken and dysfunctional; and failing to consider this in the same context as was referred to in respect of disclosure 1.

Mr Allen KC submitted that there plainly was much information in the report by way of the contents of the executive summary and its setting out a number of statements of interviewees. However, because the tribunal had failed to follow a structured approach, it was unclear whether it had found that this was not a qualifying disclosure because there was not a disclosure of information, or because some other essential element of the definition was in its view not satisfied.

Mr Panesar KC again submitted that there was no basis to assert that the tribunal was not aware of, and did not take into account, so far as relevant, the wider context. The main thrust of his skeleton was that the report did not reflect in a balanced and accurate way, the views of those that it professed to reflect, and was skewed so as to downplay and exclude the positive, and highlight and exaggerate the negative. His skeleton sought to make this good by reference to a number of the tribunal’s findings, and the evidence on which it drew, including that only 37 of 90 members of the department were interviewed, the framing of questions – see [144] – [146] – and the findings that four employees had themselves expressed concerns as to the content of the report – see [168] – [171].

In oral argument Mr Panesar KC accepted that the tribunal had treated the report as a disclosure of information, but submitted that it had also properly concluded that it was not a disclosure made in the reasonable belief that it tended to show relevant wrongdoing; and the PD complaints relating to it had in any event properly failed because causation was not made out. Mr Allen KC in reply said that the tribunal had still failed to find what particular information the report disclosed, and then consider the other elements of the definition in relation to that, in a structured way.

My conclusion in relation to this strand of ground 1 is as follows.

It is correct to say that there is no express or discrete finding by the tribunal as to whether the report contained a disclosure of information. Its conclusions go straight to the question of whether the claimant genuinely and reasonably believed that the quotations it contained from interviewees tended to show relevant wrongdoing. However, I think it can be inferred from this that the tribunal took it as a given that the contents amounted to a disclosure of information. The claim that this was a PD fell at the reasonable-belief-in-wrongdoing hurdle. Further, the particulars of claim and list of issues indicated that the claimant simply relied on the tabling of the whole report as a disclosure, without specifying any particular part. Although the tribunal does not cite passages from the report, it is clear that the tribunal read and considered the whole thing, including the executive summary. All of that being so, while it would have been better to spell out that the claimant was over the “information” hurdle in this case, the tribunal did not err by not addressing it separately, as such.

This second strand of ground 1 is therefore not upheld.

Ground 2 asserts that the tribunal erred by not addressing whether the claimant believed, and did so reasonably, that her disclosures were made in the public interest. Mr Allen KC submitted that there was simply no reference to these strands of the test at all. He acknowledged, however, that whether this ground had any traction depended on the fate of other grounds.

Mr Panesar KC noted that the tribunal referred at [6] to having been referred in the parties’ submissions to the relevant authorities, and to the law, as such, not being controversial. The tribunal had been fully addressed on this aspect in written submissions, including references to Chesterton Global Limited v Nurmohamed [2007] EWCA Civ 979; [2018] ICR 731, and to Blackbay Ventures Limited v Gahir [2014] ICR 747. It was not to be inferred that the tribunal had erred merely because it did not set the law out in its decision. In any event, where the tribunal had determined that a claimed disclosure was not protected because some other essential element of the definition was not present, it was not an error not to address these strands as well. There was no need.

I conclude as follows. The tribunal indeed did not address, in relation to either of these two disclosures, whether the claimant believed, and reasonably believed, that it was made in the public interest. Whether that was an error depends on whether it in any event properly disposed of the complaints relying on these claimed disclosures on other grounds, either because, for other reasons, they were not qualifying and protected disclosures, or because, even if they were, the detriment claims that relied upon them in any event failed.

Ground 3 also has two strands. The first asserts that the tribunal erred in respect of disclosure 1, by failing to address whether the claimant had a reasonable belief that the disclosure tended to show relevant wrongdoing; the second asserts that it erred in respect of disclosure 6, by appearing to treat it as a legal requirement that the claimant have stated her belief in relation to relevant wrongdoing, when making the disclosure, as opposed to merely actually and reasonably holding it.

As to disclosure 1, Mr Allen KC submitted that [213] to [216] did not address the particular question of reasonable belief in wrongdoing. In relation to disclosure 6 he cited various passages in the report, quoting statements by interviewees and/or the claimant’s comments on them. He said it could be inferred from these that she reasonably believed that the report contained information which tended to show a breach of legal obligations to employees, endangerment of employees’ health and safety or a cover-up. The only passage in the report which the tribunal specifically discussed in its conclusions was that relating to CL at [245]. But it wrongly failed to consider the context, from which it should have inferred that it related to CL’s disability. At [165] and [243] the tribunal appeared also to have wrongly required the claimant to have stated her belief in the report itself.

In relation to disclosure 1, Mr Panesar KC relied once again upon the fact that the tribunal was fully addressed on the law relating to this limb of the legal definition. He also submitted that, as the tribunal had found that this was not a disclosure of information, that was determinative. He also maintained that it could be inferred from paragraphs [213] to [216] that the tribunal considered that the claimant did not have a reasonable belief that this disclosure tended to show relevant wrongdoing.

In relation to disclosure 6 Mr Panesar KC submitted that [215] showed that the tribunal understood that there is no need for the employee to state their belief that the disclosed information tends to show relevant wrongdoing, so long as they do in fact reasonably hold it. Nor did [165] or [243] show that it misunderstood the law in this regard. Rather, the tribunal was making the point that the failure to state any such belief evidentially supported the conclusion that she did not actually hold such a belief. Further, the word “Nor” at the start of [165] showed that this was only one factor relied upon by the tribunal. In any event, the tribunal set out its own appraisal of the claimant’s methodology and the imbalanced contents of the report, from which it could be inferred that it considered that any such belief which the claimant might have had was not reasonably held.

My conclusions on this ground are as follows.

In relation to disclosure 1, it is, I think, clear from [58] and the discussion at [210] to [216], that the tribunal’s view was that the claimant had so little hard information at this point, about the substance of the allegation against VC, and whether there was any evidence to support it, that she was not in a position to form an informed opinion about the substance of that allegation. I think it can be inferred from this, that the tribunal considered that, if the claimant believed that some act of sexual harassment had occurred, that was not a reasonably-held belief.

However, the error properly identified by ground 1, means that the tribunal failed to consider whether the claimant believed, and if so, reasonably believed, that the fact that a Safecall allegation of harassment had been made against VC, together with the proposal to proceed by way of focus groups, tended to show a risk of relevant wrongdoing. Had it done so, and had it found that the claimant did so believe, it would have needed to consider whether the contextual matters relied upon by Mr Allen KC had a bearing on whether any such belief was reasonably held. However, it seems to me that the tribunal once again erred by failing to consider this claimed disclosure in that way.

I therefore uphold this ground in relation to disclosure 1.

In relation to disclosure 6, I agree with Mr Panesar KC that the tribunal did not make the error of thinking that it was a necessary requirement for the claimant to communicate (as opposed to reasonably holding) the requisite belief. Rather, it relied upon the contents of the report – what it did and what it did not say – in evidential support of its conclusion that she did not in fact reasonably hold such a belief. It was entitled to rely on the report as evidential material in that way.

Mr Allen KC submitted fairly that the report contained material from which the tribunal could have inferred that the claimant reasonably believed in relevant wrongdoing. However, the tribunal was not bound to regard the claimant’s assessment that morale was low, the lead team dysfunctional, or as to the lack of strategic direction and purpose, or any particular substantive passage cited from the interviewees, as necessarily pointing to the conclusion that the claimant reasonably believed that it disclosed relevant wrongdoing of the kinds on which the claimant relied. I note that Mr Allen KC specifically disavowed any perversity challenge. What weight to attach to this evidential feature of the report was a matter for the tribunal.

I do not agree with Mr Panesar KC, that the word “nor” at the start of [165] specifically signified that the absence of the claimant stating that she held the requisite belief in the report was only one factor relied upon by the tribunal. Following the discussion, at [163] and [164], of features of the content touching upon whether there was a protected act, the word “nor”, and what followed, indicated the tribunal moving on to features touching upon whether it was a protected disclosure.

However, while it would perhaps have been better for the tribunal to have said more in its decision than it did, about the substantive content of the report as a whole, its remarks about the methodology, and descriptions of the content, show that it actively engaged with it; and I cannot conclude that the tribunal failed to consider the substance of the evidence that it included. Further, although the tribunal did not set out, in its summary of the issues at the start of the decision, the kinds of relevant wrongdoing which the claimant relied upon, it plainly had them in mind, specifically referring to them at [165] and [243], and considering also the particular passage in relation to CL relied upon as raising a concern by reference to the 2010 Act. It was also entitled to rely on its findings as to how the claimant pitched the report, and what needed to be done to fix the issues raised by it, to the third respondent when she first tabled it; and it referred to her own evidence about it.

The decision did specifically discuss the substantive content relating to CL. As I will discuss, it properly concluded, for cogent reasons, at [245], that this did not amount to a protected act; so it did not err in not concluding that this part of the report contained information that the claimant reasonably believed tended to show a breach of legal obligation by way of disability discrimination.

This ground therefore fails in relation to disclosure 6.

Ground 4 contends that the tribunal erred by failing to consider whether the cumulative effect of disclosures 1 and 6 amounted to a disclosure of information, citing Simpson v Cantor Fitzgerald Europe [2020] EWCA Civ 1691; [2021] ICR 695. Mr Allen KC’s skeleton did not really develop this ground beyond referring to the fact that submissions were made to the tribunal in the claimant’s counsel’s closing skeleton on this same point. I cannot see that in this case it adds anything of substance to ground 1, nor to my conclusions upon it. I do not uphold it as a separate ground.

Ground 5 asserts that, in relation to detriments 3 and 5 – which both relate to the “dismissal” – the tribunal erred, because it did not apply the correct legal test when deciding whether the dismissal was “done on the ground that” the claimant had made disclosure 1 or disclosure 6.

Mr Allen KC, citing NHS Manchester v Fecitt [2011] EWCA Civ 1190; [2012] ICR 372 at [45], submitted that the correct test is whether a disclosure “materially influences (in the sense of being more than a trivial influence)” the impugned conduct. But, instead, at [22], the tribunal referred to whether it was a “material factor”; in its conclusions, at [249], it used the expression “not as a result of”, and at [251] it referred to the “reason why” the IDL contract was terminated. In the absence of citation by it of Fecitt, the EAT could not be confident that the tribunal had applied the correct test.

Mr Allen KC also referred to the finding at [254] that, in the context of victimisation, had there been a protected act, the failure to refer at the time to the full underlying reason for “financial constraints” being relied upon would have led to a shifting of the burden of proof. He noted that, in the context of protected-disclosure detriment, section 48(2) of the 1996 Act indicates that it is for the employer to show the ground on which the impugned conduct was done. He submitted that the tribunal had effectively found that the third respondent had lied to senior colleagues about the reasons for the dismissal. It was not, he submitted, apparent that the tribunal had properly considered the implications of this for whether the first and third respondents had discharged the burden on them to show that the dismissal was not because of disclosures made in the claimant’s report.

Mr Panesar KC noted that Fecitt at [45] was specifically cited in the claimant’s closing submissions to the tribunal; and he submitted that its reference to “material factor” at [22] showed that it had the substance of the test on board. To suggest that it did not, because of a difference between the words “material influence” and “material factor”, was precisely the sort of over-pedantic approach to reading a tribunal’s decision that had been repeatedly deprecated in the authorities.

My conclusions on this ground are as follows.

First, it is clear that Fecitt was cited to the tribunal, and it is in any case a very well-known authority. In his skeleton Mr Allen KC accepted that the words used at [22] may not themselves amount to an error of law. In my judgment they do not. The key point made by Elias LJ is that, in relation to detriment, it is sufficient that the disclosure be a material influence – it does not have to be the sole or principal reason, as applies in relation to unfair dismissal. Even though it might have been safer to cite the precise words used by Elias LJ, there are a number of analogous forms of words in which the same test may, in substance, be expressed. I consider that the tribunal was obviously referring at [22] to the Fecitt test. The reference, at [232], to whether disclosure 2 had “any material bearing” upon the decision to dismiss, again shows that it kept the substantive test in mind.

I do not think that the use of “not as a result of” at [249] bespeaks an error of law. The tribunal was there addressing why the four members of the lead team expressed serious doubts about the claimant’s report, and a lack of confidence in her. The tribunal found, as it was entitled to, that these were conclusions that they had genuinely and independently reached; and in stating that their views were formed “not as a result” of any disclosure, the tribunal was effectively ruling out any material causative link. It also specifically addressed whether their conduct “induced” the termination “because” of the reference to CL, but ruled it out, which it was entitled to do.

Mr Allen KC, however, fastened on the tribunal’s use of “because”, or “because of”, which he submitted was a “demonstrably incorrect test”. As to that, while section 47B of the 1996 Act continues to use the language “on the ground that”, the 2010 Act has adopted (in both section 13 and section 27) “because” – but they mean the same thing (see, e.g., the discussion in Amnesty International v Ahmed [2009] ICR 1450, at footnote 4 and associated text). It is therefore not hard to see how, in a case which involved the same alleged conduct being the subject of both PD and victimisation complaints, the tribunal used the compendious expression: “because of”. Further, particularly given that it is clear that the tribunal had the Fecitt point in mind, its use of “because of”, rather than “on the ground that”, does not show that it, in substance, applied the wrong test.

Nor do I think that the tribunal’s failure, at [254], to refer to section 48(2) shows that the tribunal erred in its approach to the burden of proof in relation to the PD detriment complaints. The tribunal was specifically making a point there about the victimisation complaints; and in its conclusions it made positive findings as to the reasons for the dismissal (and related detriments), and positively ruled out the (claimed) protected disclosures as a materially contributing cause.

Ground 5 therefore fails.

Ground 6 asserts that the tribunal erred in law, in failing to find that each of disclosures 1 and 4 amounted to protected acts. Reliance is placed specifically on the provision in section 27(2)(c) of the 2010 Act that a protected act includes “doing any other thing for the purpose of or in connection with” that Act.

In relation to disclosure 1, although he made a number of other points, it appears to me that the heart of Mr Allen KC’s submission was that the context in which the claimant was specifically told that the Safecall allegation was of sexual harassment, meant that the tribunal erred by not concluding that that disclosure was within scope of section 27(2)(c).

In relation to disclosure 4 Mr Allen KC highlighted the tribunal’s findings that CL was, in law, almost certainly a disabled employee, by reference to mental health issues, and that the corridor conversation included the claimant asking Ms Bowen to explore if everything appropriate was being done for him. He submitted that, while the tribunal found, at [239], that this did not involve any allegation of discrimination, it failed to consider whether it fell within section 27(2)(c).

In his skeleton Mr Allen KC also submitted that disclosure 6 was a protected act.

Mr Panesar KC submitted, in relation to disclosure 1, once again, that the tribunal had found that the conversation which occurred was materially different to the one alleged. The claimed protected act as factually alleged was therefore not made out; but in any event the tribunal did not err in concluding that this communication did not amount to a protected act within any part of section 27; or that, if it did, it made no difference to the outcome.

In relation to disclosure 4, once again the tribunal had found that the conversation which in fact occurred was very different from the one that the claimant had alleged. The tribunal found that in fact the respondent had taken extensive measures to support CL, who had mental health issues, and was off sick; that colleagues had expressed concerns about him in exchanges with the claimant; that this was a very brief corridor encounter in which the claimant did no more than convey a similar general expression of concern about his case; and that the claimant accepted that Ms Bowen cared about CL and said she would look into the matter. The tribunal had rejected the claimant’s original account to the effect that she had been critical of how the HR team had allowed the situation to develop. It was therefore entitled to find that the actual communication was not a protected act.

My conclusions on this ground are these.

In relation to disclosure 1, the tribunal concluded at [215] that “this expression of opinion” did not amount to the doing of anything “for the purposes of, or in connection with, the 2010 Act” because the claimant “was not actually aware of any facts which she was reporting as deserving of investigation.” It therefore referenced here the words of section 27(2)(c).

However, a disclosure of information is not a requirement of the section 27(2)(c) test, and the lack of knowledge of any hard facts about what was alleged would not preclude the claimant’s conduct being found to be “in connection with” the 2010 Act. Further, as I have already noted when considering ground 1, the tribunal found as a fact that the claimant had been told that there was an allegation of sexual harassment (which the 2010 Act of course proscribes), and she plainly implicitly referred to that general allegation in what she then said. In view of that, I conclude that the tribunal did err by relying on the reason that it gave, to conclude that a protected act was not made out.

In relation to disclosure 4, having regard to all the facts found about the actual context and content of this conversation, I do not think that the tribunal erred in finding that it did not amount to a protected act. It was not bound to find that the claimant’s expressing concerns of the kind, and in the way, that she did, amounted to something done by reference to the 2010 Act, whether because CL’s ill health probably amounted to a disability, or otherwise. There was no overt, or necessarily implicit, link to the 2010 Act in this exchange, as found by the tribunal to have taken place.

Disclosure 6, also mentioned by Mr Allen KC, was not strictly within scope of this ground; but, in any event the tribunal properly concluded at [245] that the content of the report relating to CL did not amount to a protected act. It was not bound to infer that the subject matter was connected to the legal implications of his (possibly) being a disabled person, as to opposed to simply his underlying health problems, their impact and management, and a general concern about his welfare.

Ground 6 is therefore upheld, but upheld solely, in relation to disclosure 1.

I turn next to these three grounds, which respectively postulate that the tribunal erred by: failing to engage with the claimant’s case, taking a fragmented approach, and/or producing a decision which was not Meek-compliant. There is some overlap among them, and with earlier grounds.

In support of ground 8 Mr Allen KC cited Brightman v Tiaa Limited, UKEAT/0381/19 at [52] where the EAT said: “An employment tribunal cannot, and is not expected to, refer to every disputed issue of fact or argument, but it does have to engage with the key points advanced by the parties.” The ground lists some eight features of the claimant’s case (to which I will come) with which the tribunal is said to have failed to engage, including a list of matters which her counsel had relied upon as supporting adverse inferences. The ground also cites Patel v Surrey County Council, UKEAT/0178/16 at [101] where it was said that a tribunal should not uncritically accept a reason given by a respondent for detrimental action, and that the fact that the tribunal may consider the reason given to be reasonable does not absolve it from further enquiry.

Mr Panesar KC once again submitted that this ground was an impermissible attempt to reopen findings of fact. A number of points simply reiterated other grounds. Some points referred to broad-brush submissions made by the claimant; but the tribunal had properly considered and focussed on the factual matters and arguments that it needed to consider, in order to resolve the particular issues relating to each of the claimant’s claimed disclosures and the claimed detriments. He reminded me of the well-established guidance, reiterated in cases such as ASLEF v Brady [2006] IRLR 576 and DPP v Greenberg [2021] EWCA Civ 672; [2021] IRLR 1016, that a tribunal’s reasons should not be subjected to over-pedantic scrutiny, and that a tribunal is not required to address every aspect of the evidence, make findings of fact about every feature canvassed before it, nor address every submission made to it; and that failure to refer to particular evidence should not mean that it was not considered. The present tribunal had, he submitted, made considered, reasoned findings as to the true reasons for the dismissal and all other matters that it was required to determine.

Ms Levene made similar submissions, to the extent relevant to the complaints against the second respondent. She also contended that ground 8 had no bearing on the alleged protected acts (as opposed to protected disclosures) so that the dismissal of the victimisation complaints as against her client must in any event stand.

My conclusions in relation to ground 8 are as follows. Point 1 criticises the tribunal’s statement, at [23], that the case “is not about whether there was a so-called toxic culture at Sellafield.” That statement followed the accurate summary overview given by the tribunal of the particular issues to which the PD and victimisation complaints gave rise. The tribunal was right that it was not required to make a stand-alone generalised finding (either way) of that sort. To the extent that this was relied upon as context, the point repeats strands of other grounds. Points 2, 3 and 4 also reiterate points raised by other grounds. Point 5 contends that the tribunal failed to “engage properly” with the contents of the Safecall report, but again it seems to me adds nothing of substance to ground 1.

Point 6 contends that the tribunal failed properly to address 18 points relied upon by the claimant’s counsel in his closing argument, as supporting adverse inferences against the respondents. As to this, the EAT once again has to strike a careful balance in appraising the sufficiency of the tribunal’s reasons. The tribunal was not obliged to mention, or address, every feature of the evidence, or submission, raised with it; but it was obliged to engage with the key elements of each side’s case, and to produce a decision which is Meek-compliant.

The bulk of the 18 points raise matters that were said to have supported the claimant’s case as to the true reason for the dismissal. In summary these were: (a) that differing accounts were given at different times for the reason for dismissal, and reliance on the financial constraints explanation, which it was said simply did not hold water, was, by degrees, largely abandoned; (b) though no longer challenging their authenticity, what was said to be the “odd way” in which the concerns of the three HR members following the claimant’s report were committed to writing; (c) issues about the credibility of the assertion that in-house EDI capability was considered now to be sufficiently developed that the claimant was no longer needed; (d) evidence that the third respondent kept the substantive contents of the claimant’s report even from her own CEO; and (e) the extreme rapidity and suddenness of the reversal of the claimant’s fortunes, from the point of delivery of her report, to the decision to terminate, and despite months’ more work having recently been scoped out for her.

Other strands referred to evidence that there had been two previous investigations relating to VC, the failure to hold a formal investigation of the VC report despite the claimant’s advice, and to certain other particular features of the evidence. Reliance was also placed on the failure of the respondents to call certain particular individuals as witnesses.

I will take with this, points 7 and 8 of this ground, which complain of a failure by the tribunal to engage with the fact that managers at the second respondent were informed that the termination was on financial grounds; and with the findings in the subsequent PWC report.

My conclusions on these further points 6 to 8 of ground 8 are as follows.

First, the essence of the difference between the claimant’s case and what the tribunal found, was that the claimant’s case was that the third respondent decided to dismiss her because of the substance of the (claimed) disclosures made in her report. The tribunal, however, found that it was not, but was principally because of the concerns raised by members of the HR lead team about the lack of balance in the report, and about the way in which the claimant went about conducting the interviews, as well as because of the information given to the third respondent about the costs of the exercise, and her appraisal as to whether the lead team could manage the EDI function going forward.

On the tribunal’s findings, it was, in that sense, not purely coincidental that the dismissal followed hard on the heels of the tabling of the report. But, on the tribunal’s findings, there was a material change from how the third respondent reacted to the report when she first read the draft sent to her, which change came about as a result of the criticisms and concerns subsequently raised by lead team members. It appears to me, overall, that the tribunal’s findings do sufficiently convey its position in relation to the claimant’s points about the significance of the timing.

Secondly, the tribunal made specific findings of fact that the written complaints were genuine, and were the result of the writers independently raising their concerns, and that their concerns were about lack of balance, not including their own positive comments, and the claimant having conducted the interviews in a way that suggested that she was looking for negative material – rather than about the substantive (claimed) disclosures in the report. The claimant’s case was that these criticisms were self-serving, or defensive of colleagues – and the tribunal was invited to infer that the complaints had been orchestrated. But that was for the tribunal to consider and evaluate. Again, this aspect of the claimant’s case appears to me to have been considered and addressed by the tribunal.

Thirdly, the tribunal was entitled to conclude that financial considerations did play some part in the decision, because of genuine concerns about the costs incurred in the exercise, and whether it represented value for money, having regard to the issues raised about methodology, balance, and the overall quality of the report, and the absence of any specific content relating to the Safecall allegations against VC. The tribunal’s finding, which it was entitled to make, was also that it was the reference “only” to financial constraints as the explanation for the dismissal, that was misleading. The tribunal also referred, at [254], to the failure to inform the claimant “or indeed anyone else who asked at the time” of the full underlying reasons. This plainly refers to the claimant’s case, that it was not just her, but also senior managers at the first and second respondents who were kept in the dark.

The tribunal also clearly found that it was following the reactions of colleagues, and because she was in light of them contemplating terminating the IDL contract, that the third respondent asked Ms Bowen what the implications would be for the EDI work of doing so. This was, implicitly, the tribunal’s answer to the contention that the respondents’ case was contradicted by the fact that a significant programme of work going forward for the claimant had only recently been agreed.

More generally, the tribunal plainly had on board, that the third respondent, as it put it at [24], did not enjoy the confidence of Mr Vineall or Mr Vallance (another senior manager) of the second respondent. It also made findings about the periodic enquiries made of the third respondent as to what was being done about the Safecall report, and about discussions in relation to the anonymous letter. It must plainly have had on board the suggestion that the third respondent was concerned about how the report might be seen by senior managers at the first or second respondents as reflecting badly on her and the HR function. The tribunal also made findings about the aftermath, including the steps taken to commission a further investigation into the Safecall report and the commissioning of the PWC report. I therefore do not think that it can be said to have wilfully ignored these features.

Overall, the claimant’s case as to what happened and why, and her essential contentions as to why the main planks of the respondents’ case did not hold water, were plainly understood, and, I conclude, considered, and sufficiently addressed by the tribunal. Further, in light of all the foregoing features, and passionately though the claimant maintains that the tribunal got it wrong (though short of arguing perversity), I do not agree that the respondents’ case was uncritically accepted by it.

I do not think the tribunal was obliged to address separately the point which relied on the respondents’ failure to call certain witnesses, given that the third respondent and five other witnesses gave evidence for the first respondent, and three senior managers gave evidence for the second respondent. Other aspects of the eighteen points refer to details of the evidence or how it unfolded, that I do not think the tribunal was obliged to address separately in the liability decision.

Finally, I note that the burden of this challenge was “failure to engage” with key aspects of the claimant’s case. It was not perversity. But nor was it any part of this ground – or any of the grounds – to assert that the tribunal had impermissibly distinguished between the substantive contents of the report, and the reasons that it found for the dismissal – that is, the type of argument that was discussed in Kong v Gulf International Bank (UK) Ltd [2022] EWCA Civ 941; [2022] ICR 513.

Ground 8 therefore fails.

Ground 9 contends that the tribunal erred by taking a fragmented approach to each of the disclosures, and not having regard to the overall picture. As to that, the absence of a statement that the tribunal did so does not, by itself, make good the criticism. The claimant’s case, as originally advanced, was that there was a steady series of disclosures by her, building up to the report, and indeed following it. But, in light of the tribunal’s findings, a number of these fell factually by the wayside; and this appeal has focussed on disclosures 1 and 6. Further, the tribunal plainly recognised that there was an overall chain of events that unfolded, with the Safecall report leading to the focus groups, leading to the claimant’s report, leading to the reaction, leading to the dismissal; and it specifically addressed at [225] whether disclosure 1 factually influenced the decision to dismiss.

In light of all of that, I therefore do not think that the tribunal erred by taking a fragmented approach. Ground 9 fails.

Under the heading of “not Meek-compliant” ground 10 asserts, in a similar vein, that the tribunal should have stood back and seen that this was a case of an expert and respected external consultant coming into a troubled EDI and safety-critical environment, whose raising of concerns about the actions and inactions of the HR lead team, led to the termination of her company’s contract. Apart from that, the ground itself simply refers to grounds 1 to 7.

Though she profoundly disagrees with them, the tribunal’s reasons convey to the claimant why she lost (and to the respondents why they won). I do not think this ground adds anything of substance to the other grounds. I do not uphold it.

I now turn to ground 7, which relates to the tribunal’s conclusion at [275] to [278] that the first respondent was not the agent of the second respondent within the meaning of section 109 of the 2010 Act. There is an overlapping strand of ground 8, but it adds nothing to ground 7.

The tribunal is said to have erred by failing to consider, and conclude, that the source of the first respondent’s authority was section 10(2)(k) of the Energy Act 2004, which conferred power on the second respondent to “carry out its functions through subsidiaries”. The first respondent was such a subsidiary, which was authorised to carry out a range of such functions, under the agreement between it and the second respondent (the Services Agreement). The tribunal is said, erroneously, to have instead focussed on the first respondent’s contract with IDL.

Citing Wijesundera v Heathrow 3PL Logistics Limited [2014] ICR 523, it is said that a broad view of the concept of the scope of the agency should be taken. Mr Allen KC also noted that the tribunal had evidence that it was the second respondent which commissioned the subsequent PWC report into the first respondent’s HR function; and he referred to other features of the evidence which he said touched upon this issue, but which the tribunal had failed to address.

Ms Levene highlighted that, as set out by the tribunal at [189], the Services Agreement expressly provided that it did not establish any partnership or joint venture, and that it did not authorise the first respondent to make or enter into any commitments on behalf of the second respondent. The tribunal had correctly identified that Ministry of Defence v Kemeh [2014] EWCA Civ 191; [2014] ICR 625 confirms that common law principles of agency apply in the context of section 109, and found as a fact that the second respondent had a limited degree of control, by advising, challenging or supporting the first respondent. It had made a finding of fact, at [188], accepting evidence from the second respondent’s witness, that it could not directly intervene in the affairs of the first respondent. It properly applied Kemeh to the facts found.

The tribunal properly found that the first respondent had no authority to act on behalf of the second respondent with respect to third parties. Neither the 2004 Act nor the Services Agreement provided the necessary consent for the first respondent to act as the second respondent’s agent, in relation to the claimant, or at all. Nor could the second respondent have done any acts in relation to the claimant as principal, as the contract with the claimant’s company was not made by it. The first respondent could not, as its agent, do that which it could not do itself. The tribunal’s reasons were adequate to address the issue.

Mr Allen KC and Ms Levene also traded extensive arguments as to whether the claimant’s case on this point had been sufficiently pleaded or supported by evidence. Mr Allen KC also submitted that the provisions of the Service Agreement were not decisive, as the tribunal was concerned with the position in tort (that is, the statutory tort of discrimination), not contract.

Though I heard very extensive argument on this ground, of which I only have only given brief highlights, I can state my conclusions on it relatively shortly.

In Kemeh the MoD contracted with Serco, which in turn contracted with Sodexho, for the provision of certain services. Sodexho’s employee racially abused Mr Kemeh. Elias LJ appears to have inclined to the view that a person can act as an agent for another person in certain ways, even though their authority does not extend to the power to affect the principal’s relations with third parties; but without deciding that point he held that, in any event, the putative agent must be acting on behalf of the putative principal and with its authority. The fact (in that case) that an employee of a contractor performed work for the benefit of the third party that (indirectly) hired the contractor, was not sufficient to make her the agent of the hirer. Her contract with her own employer was the source of her authority to do what she did. Nor did the limited control which the hirer had over her suffice to make her its agent.

I do not think that the present tribunal erred in concluding that the present case was essentially conceptually similar to Kemeh. The claimant was complaining about the acts of individuals employed by the first respondent. Their authority derived from that employment relationship, and so the first respondent was liable to the claimant for their actions. The first respondent, for its part, was acting as a contractor retained by the second respondent to perform various tasks pursuant to the Services Agreement. The tribunal did not err by failing to conclude that this, as such, made the employees concerned, or the first respondent itself, the agent of the second respondent.

I do not think that the subsidiary relationship or the provisions of the 2004 Act should have pointed the tribunal to a different analysis. Mr Allen KC acknowledged, as he was bound to, that the mere fact that the first respondent was a subsidiary of the second respondent was not enough. The tribunal plainly had on board – see [10] – that the second respondent is a creature of statute, with particular powers assigned to it by Parliament. But one of the things that Parliament expressly authorised it to do was set up separate subsidiaries to carry out some of those tasks that it would otherwise have to perform itself. I ask, rhetorically, what the purpose of that provision would be, if the subsidiary were to be deemed in any event to be performing the tasks as agent of the parent.

But in any event I consider that the tribunal properly regarded the provisions of the Service Agreement, cited at [189], as relevant to this issue. Although, doctrinally, as Elias LJ at least contemplated in Kemeh, the fact that the first respondent could not act as the agent of the second respondent in contract, would not be bound to preclude it from being able to do other things on the second respondent’s behalf, that provision was properly regarded as a strong pointer towards the intention being that the relationship established should, in principle, be no different to that established between a party and a contractor which it chooses to hire to perform tasks for it, that it could otherwise have performed itself.

The claimant does not appear to have pointed to any other positive provision in the Service Agreement that should have led the tribunal to the conclusion that the first respondent was authorised by the second respondent to act as its agent in any capacity. The fact that it may be said that the carrying out of HR functions was ancillary to, or even in practice necessary to, the carrying out of the tasks given to the first respondent under the Service Agreement takes the argument no further, if, as I conclude, it was not wrong to conclude that the assignment of those tasks did not itself in any event involve the conferring of any authority sufficient to establish an agency.

Whether the other evidence to which Mr Allen KC referred me, as to the degree to which the second respondent interested itself in such issues, and its interactions with the first respondent in relation to them, established a sufficient measure of control so as to give rise to implied authority amounting to agency, was a matter for the appreciation of the tribunal. Given the tenor of the dicta in Kemeh on this point, he rightly accepted that he could not go behind the tribunal’s findings of fact. Nor do I think that the reasons are inadequate on this point. The finding at [188] (for which the tribunal also cited an evidential source), that the second respondent could not directly intervene in the affairs of the first respondent, sufficiently supported its conclusion.

The fact that managers at the second respondent plainly had their own concerns about HR matters at Sellafield, and that the tribunal had evidence that it itself subsequently commissioned the PWC report in relation to them, are not facts that the tribunal erred by not treating as pointing to the conclusion that, in relation to the matters complained of, the first respondent was its agent for the purposes of section 109.

It may be said that the tribunal’s finding that the first respondent’s contract with the claimant’s company was the source of its authority over her was not, and could not have been, decisive. But nor was it wholly irrelevant that it was the first respondent that, on this occasion, contracted for the provision of the claimant’s services, not the second respondent (which had done so in the past). But in any event, for reasons I have given, the tribunal did not err by concluding that the first respondent did not, in respect of the conduct of its human agents of which the claimant complained, itself act as the agent for, and with the authority of, the second respondent within scope of section 109.

Ground 7 is dismissed.

The success of the claimant in relation to grounds 1 and 3 relating to disclosure 1, means that its findings on whether she thereby disclosed information, and on the reasonable-belief questions to which ground 3 related, cannot provide a sound and complete answer to the criticism of its failure to address the reasonable-belief questions to which ground 2 related. I have also found, in response to ground 6, that the tribunal erred by not concluding that disclosure 1 was a protected act.

But the tribunal made specific, reasoned, findings, that the claimant was not subjected to the detriment of being pressurised to carry out the focus groups, by reason of the fact of what she said in disclosure 1, nor dismissed because of it. The outcome of this appeal has not disturbed those findings, nor the findings that the contents of the claimant’s report did not amount to protected disclosures or protected acts, nor was she dismissed because of those contents. Further, in relation to the second respondent, in any event ground 7, relating to the agency point, has also failed on its own merits.

The tribunal should have set out an account of the essential elements of the law, rather than just referring to counsel’s submissions; and the errors that it did make might have been avoided, had it broken down its reasoning more fully by reference to the distinct components of each cause of action. Nevertheless, the essential conclusions that were fatal to the complaints emerge intact. The EAT does not hear the evidence, find the primary facts, or decide what inferences or conclusions should be drawn from them. It can only intervene in a tribunal’s decision where there has been an error of law which means that the substantive outcome cannot stand. Notwithstanding the claimant’s partial success, that challenge has not been made good. The liability appeal is therefore dismissed.

Rule 76(1) Employment Tribunals Rules of Procedure 2013 includes provision that a tribunal shall consider whether to make a costs order where it considers (a) that a party has acted vexatiously or otherwise unreasonably in bringing the proceedings, or part, or in the way that the proceedings, or part, have been conducted, or (b) that a claim had no reasonable prospect of success.

Rule 78 enables the tribunal to order a party to pay another party a specified amount up to £20,000, or an amount determined by detailed assessment. Rule 84 provides that in deciding whether, or in what amount, to award costs, the tribunal may have regard to the paying party’s ability to pay.

The first and third respondents made a costs application in September 2020. This followed an application by the claimant to strike out their responses having been refused at a hearing in July 2020. A reconsideration application was subsequently refused at a hearing in January 2021. The first costs application was held in abeyance pending the trial. Following the liability decision, the first and third respondents made a second costs application. The second respondent also made a costs application of its own following the liability decision.

It was ultimately agreed that the costs applications could be determined on paper, on the basis of written submissions. They were determined by the same panel that made the liability decision.

In its costs judgment the tribunal ordered the claimant to pay £20,000 towards the costs of the first and third respondents and £20,000 towards the costs of the second respondent.

In relation to the application by the second respondent the tribunal noted that, following the initiation of the claim, the second respondent had applied to be removed from the proceedings. A judge had concluded that the merits of the claim could not be determined without hearing evidence; but that was not an expression of opinion on the merits. The same was true of a later refusal to list for a separate hearing, a further strike-out application made by the second respondent.

On the agency point, the tribunal noted that judges had observed at two preliminary hearings that the claimant’s contentions remained unclear. The tribunal noted that on 11 June 2021 (therefore, I interpose, on the eve of the trial) the second respondent had made the claimant an offer of £160,000 in full and final settlement of her claims against all three respondents, which letter also functioned as a costs warning in relation to the claims against the second respondent itself. That offer was declined.

At [22] to [27] the tribunal opined that the claimant had “never properly addressed her mind to whether and how she might prove the essential elements of her case.” She had originally asserted six protected acts, but two were later rightly abandoned as they never had any reasonable prospect of success. One was found not to have happened, as alleged, which the claimant would have known. Two were limited in scope in light of the claimant’s evidence, so that they did not then amount to protected acts (disclosures 1 and 4). In relation to the claimant’s report (disclosure 6) the particular disclosure within it relied upon (relating to CL) was not identified until closing submissions and did not bear scrutiny.

The claimant had then, said the tribunal at [28] to [31], adduced “no evidence whatsoever to establish a causal connection between these matters and any detriment to which she was subjected.” She had therefore failed to establish a basic contravention by the first respondent for which the second respondent might be co-liable. Nor did she establish any possible basis for concluding that it had induced, aided or helped the dismissal, or any other detriment. Nor had she “ever set out any proper basis in fact or law” for attributing liability to the second respondent under section 109.

The tribunal concluded, at [32] to [33], that while the claimant had been given the benefit of the doubt in permitting these matters to go to trial, on the facts actually put before the tribunal she “comprehensively failed” to make out the elements of her claim against the second respondent; and this claim “had, in fact, no reasonable prospect of success.”

There was, said the tribunal at [34], a strong suspicion that the claimant had an ulterior motive in bringing the claim against the second respondent “related to her desire to position herself as the champion of equality within the nuclear industry and to court publicity accordingly”. But it did not need to go further so as to conclude that the claim was actually vexatious. However, it continued at [35], she had been on notice throughout the entirety of the case of the strong defence of the second respondent, culminating in the costs warning, and had chosen to pursue it without “ever properly addressing her mind to the essential issues.” That was unreasonable conduct of the proceedings.

At [36] – [39] the tribunal concluded that the preconditions under rules 76(1)(a) and (b) were both made out. It then decided, in all the circumstances, “especially given the extent to which … a publicly funded body” had been exposed to the cost of defending an unmeritorious claim, to exercise its discretion to make an award. It noted that it was not required directly to apportion the costs to the unreasonable conduct, and that the amount sought was in any event limited to £20,000, though the total costs said to be incurred were £197,867.50 excluding VAT. It noted that it may, but was not obliged, to take into account the claimant’s means, and that, while she asserted that she had health issues and had not worked since the termination of her company’s contract with the first respondent, she had provided “no actual information as to her means and assets”. Having regard to that, and the proportion the amount claimed represented of the unassessed costs incurred, it was “not appropriate to take any further account of any unquantified alleged inability to pay.” [39]

Turning to the first and third respondents’ applications, the tribunal noted that the first of these related to the claimant’s unsuccessful strike-out application. The basis of that had included serious allegations of fabricating evidence, which were only withdrawn on the eve of the strike-out hearing. That related to the letters provided by three members of the HR leadership team, discussed by the tribunal at [168] to [171] of the liability decision (set out by me above). At [45] the tribunal said:

In relation to the second costs application, the tribunal referred again to its earlier observations about the substantive victimisation complaints, stating that similar criticisms applied of the claimant’s failure to engage with the essential evidential requirements of the PD complaints. The claimant was an experienced HR adviser who had had intermittent legal advice throughout, but came to the final hearing with her evidence still in a state of flux following last-minute amendments and withdrawals, and a lack of clear articulation of her own case on the core factual allegations.

The claimant had withdrawn two matters as victimisation complaints, and one of these also as a protected disclosure complaint. Of the remaining five, one did not happen, one was so limited on the actual evidence that it did not amount to a disclosure of information, and one was found to be a well-nigh total distortion of what actually happened. In relation to disclosure 1, the tribunal stated at [52] to [54], that the claimant’s “opinion did not carry, and could not have carried, the weight necessary to make it an allegation that could qualify for protection” and the factual case put, as to what the third respondent had said, until the claimant sought to put a different gloss on it at the hearing, was “demonstrably false”. Although the claimant was permitted, at a late stage, to amend the alleged detriment relating to this to include the allegation of attempting to put pressure on her, causing her anxiety and distress, that was “entirely at odds” with her own evidence.

As to disclosure 6, the tribunal observed at [55] – [57] that, despite repeated opportunities the claimant had never properly identified any specific part of the contents of her report relied upon by her. The tribunal went on to refer to what it had said at [207] of the liability decision (cited by me earlier). The claimant had not established any facts from which a prima facie causal link could be established between any of the putative disclosures and an alleged detriment.

At [61] to [63] the tribunal stated: “Objectively this claim as presented simply had no reasonable prospect of success.” It was not reasonable for her to have brought and continued proceedings which “on the actual facts that were known to her are, and were, misconceived.” The tribunal did not have to decide whether her legal advisers must share some responsibility. The unreasonable conduct lay in her failure to provide an accurate statement of her essential case which actually accorded with the facts as known to her, which would have made it clear why, objectively, it had no reasonable prospects. The tribunal took into account that the claimant had made a number of misleading statements “which she has never corrected in a timely manner.”

Once again the tribunal concluded that the preconditions for an award were made out and decided to exercise its discretion to make a costs award. Once again it noted that direct apportionment was not required, and that the costs sought were limited to £20,000. The unassessed costs, excluding VAT, actually incurred, were said to be £285,654.18, of which the costs of resisting the strike out application were £12,312.12 and resisting the reconsideration application £8,430.88. The tribunal made similar observations about ability to pay.

The grounds of appeal, and what seem to me to have been Mr Allen KC’s main points in support of them, are as follows.

Ground 1 relates to both costs awards. It contends that the tribunal erred in considering whether to make a costs order, by failing to take into account the following relevant factors: (a) the principle that in the employment tribunal awards of costs against the losing party are the exception, not the rule – Mr Allen KC cited Barnsley MBC v Yerrakalva [2011] EWCA Civ 1255; [2012] ICR 420 at [7]; (b) the fact-sensitive nature of whistleblowing complaints and importance of examining them at trial. Mr Allen KC submitted that the discussion in Anyanwu v South Bank Student Union [2001] UKHL 14; [2001] ICR 391 at [24] in relation to discrimination claims applied equally to PD complaints; (c) the difficulty of proving whistle-blowing claims and the burden of proof – reference was made again here to the various adverse inferences which the claimant had invited the tribunal to draw, which are said not to have been adequately addressed in either judgment; and (d) the significance of the £160,000 settlement offer which had been made by the second respondent, despite which the tribunal held that the complaints had had no reasonable prospect of success. Mr Allen KC submitted that the tribunal could and should have taken it into account at the discretion stage.

The lack of a balanced and measured approach to the exercise of the discretion is said to be emphasised by the “speculative and gratuitously hurtful” passage at [34] of the costs reasons.

Ground 2 relates to the award in favour of the second respondent. It contends that the tribunal erroneously relied, at [31], upon the proposition that the claimant had never set out any basis for attributing liability to the second respondent under section 109, as her counsel had in fact made detailed written submissions on the point in closing submissions; these are reproduced in an Appendix to the grounds of appeal. It also failed to take account of the fact that the second respondent’s settlement offer was made on behalf of all respondents, while claiming not to be a principal.

Ground 3 asserts that the tribunal failed to apply the overriding objective of dealing with cases fairly and justly. In particular, it failed to balance a number of unreasonable features of the first and third respondents’ conduct of the litigation against its appraisal as unreasonable, of the claimant’s conduct in withdrawing allegations of fabrication of documents only at a late stage in the proceedings. These were: (a) waiting until the day before the start of the liability hearing to make a settlement offer; (b) the failure of the third respondent to disclose a timeline document on which she relied until cross-examination; (c) the changing reasons given for why the claimant was dismissed; and (d) the failure of Ms Bowen to remember a key conversation in relation to disclosure 4 until cross-examined.

The tribunal should have also properly weighed into the balance, when deciding whether to make costs awards, that the claimant was at all times attempting to raise concerns about the culture that she saw and experienced while working at the first respondent, and the welfare of its employees. The £160,000 settlement offer reflected that the respondents were clearly concerned about her case.

Mr Allen KC also made an overarching submission that, if the claimant succeeded on at least a significant number of liability points, then the costs awards would be unsafe.

Mr Panesar KC submitted that the tribunal had properly exercised its discretion to award costs in favour of his clients, having regard in particular to its findings that in respect of a number of the factual bases on which she advanced her claims the claimant was found to have lied, that she had withdrawn a number of allegations only during the course of the hearing, and its proper conclusion that she knew that her claims had no reasonable prospect of success.

Ms Levene made some similar submissions. She also submitted that it was not an error not to state that costs were the exception and not the rule. The tribunal properly found that the costs threshold under rule 76 was crossed. The tribunal did not err by not inferring that the settlement offered in itself signified that the claims had merit, nor by holding that the claimant’s rejection of that offer was unreasonable conduct. The tribunal also properly appraised the weakness of the case advanced by the claimant under section 109, or otherwise for asserting that the second respondent could be liable in respect of her complaints against the other respondents.

My conclusions on the costs appeal are as follows.

First, there was no error, as such, by the tribunal not stating that costs in the employment tribunal are the exception and not the rule. This general point was in fact acknowledged by the first and third respondents in their applications. This is also extremely familiar territory and the tribunal clearly had well in mind that no award could be made unless one of the threshold conditions in rules 76(1)(a) or (b) were satisfied. It also expressly recognised that, having so found, it then had to take a further decision as to whether to exercise its discretion to make an award.

Secondly, I do not think that the tribunal erred as alleged by failing to attach significance to the £160,000 settlement offer. The tribunal was not bound to infer that this offer, or the fact that it was made on behalf of all three respondents, was something that could be relied upon as indicating that they accepted that there was some potential merit in the claimant’s claims. Nor did it err by not regarding the timing of the offer as unreasonable conduct on the part of the respondents. Further, while it referred to the fact that the claimant did not accept the offer, its underlying (overlapping) conclusions were that her complaints as a whole had no reasonable prospect of success, and that she also unreasonably pursued them, because she should have known that from the outset.

As to what the tribunal said at [34], I was told that the second respondent’s costs application was not specifically confined to the grounds of lack of reasonable prospects, or unreasonable conduct; the tribunal’s conclusion was also that it did not find that the claimant acted vexatiously. Nevertheless, the tone of the remarks is troubling, given that the tribunal was avowedly speculating.

However that may be, I have to consider the remaining points raised by the grounds, which principally challenge the tribunal’s assessment of the overall prospects of success of the complaints, as failing to take fair account of their fact-sensitive nature, importance, and the significance of matters such as where the burden of proof lay, and the inferences on which the claimant sought to rely. In addition, as I have found, there were aspects of the liability decision in respect of which the tribunal erred. While these errors were not sufficient to disturb the liability outcome, I have to consider whether they contribute to the conclusion that the tribunal’s costs decision is unsafe.

In the course of its costs decision, the tribunal: relied on disclosures 1 and 6 not having been established; relied on disclosure 6 resting on the content relating to CL, which it said “did not bear scrutiny”; took the view that there was “no evidence whatsoever” to establish a causal connection between the disclosures and the dismissal or other detrimental conduct complained of; and referred repeatedly to issues relating to the claimant’s credibility and changing factual case. Overall, it seems to have considered that her case was fundamentally misconceived from the outset in almost every respect, because it was based on factual contentions that she must have known were not true.

In light of my conclusions on aspects raised by the liability appeal, however, the tribunal was wrong to say in the costs decision that the claimant did not have an arguable case that disclosure 1 was a protected disclosure; or that it was a protected act. While the tribunal did not err in finding in the liability decision that the disclosures relating to CL were not protected disclosures or protected acts, nevertheless, given the finding that he was probably a disabled person, and the nature of the concerns voiced by the interviewee quoted as referring to him in the report, it seems to me also to overstate matters to say that these parts of the claimant’s case “did not bear scrutiny”.

Further, while other claimed disclosures and detriments fell by the wayside, it seems to me that the central and most potentially significant part of the claimant’s case was always that her report amounted to a protected disclosure, and also a protected act, and that her company’s contract was terminated because of its contents. The facts as to the tabling and content of that report were not in dispute, nor when, how, and the gist of what, she was told of the decision to terminate her contract, nor the contents of the letter of confirmation. The tribunal correctly accepted that the report disclosed information. Whilst it did not err in concluding in the liability decision that the claimant did not reasonably believe that it tended to show relevant wrongdoing, that was a fact-sensitive determination. As to causation, that depended on the precise sequence of events, and what influenced the mind of the third respondent, and again was fact sensitive. Had the tribunal found in the liability decision that the report was a protected disclosure, or (as the tribunal itself said) a protected act, then the burden to disprove causation would have fallen on the first and third respondents.

All of that being so, and in view of the contents of the claimant’s report, and the nature of the relevant wrongdoing on which she relied, and while the tribunal did not err in concluding that she did not have a reasonable belief that it tended to show relevant wrongdoing, I think it went too far to say that that was not reasonably arguable. As to causation, while the tribunal ultimately was not bound to accept that any of the matters relied upon by the claimant in support of adverse inferences made good her case, again it is another thing to say that no feature of her case in relation to inferences, or as to causation generally, was reasonably arguable, bearing in mind also that, on the tribunal’s own findings, there was a connection between the reasons for dismissal and the report. I do not think that the conclusion that there was “no evidence whatsoever” to establish causation can stand.

While the tribunal also considered that the claimant’s conduct in relation to the fabrication allegation was unreasonable, and noted the costs attributed to defending her strike-out, and associated reconsideration, applications, it did not rely upon the late withdrawal of that allegation as such. It also at [45] relied, in this regard, upon its findings that the claimant had formed a view that the third respondent was not up to the job, and took advantage of the focus groups to elicit negative responses, and that her draft report was “anodyne”. Reading the decision overall, it is not clear to me whether, or if so, on what basis or to what extent, it would have made an award arising from her having made the fabrication allegation alone.

I have therefore come to the conclusion, in light of the considerations I have mentioned thus far, that the costs award in favour of the first and third respondents is not safe, and must be quashed.

What of the costs award in favour of the second respondent? From the material that I was shown, it does appear that the claimant did not articulate prior to the trial her detailed legal case as to how the facts were sufficient in law to make the second respondent co-liable for any unlawful victimisation on the part of the first respondent under section 109. But she did plead reliance on that section and, it appears to me, the essential facts which she asserted as supporting that conclusion. Further, I was shown that a reasoned legal analysis was put forward in her counsel’s closing submission. As the discussion in Kemeh shows, what constellation of facts would be sufficient in law to make good section 109 liability is open to some areas of argument; and the claimant’s failure on this point rested also in part on the tribunal having (properly) made findings of fact that went against her, in particular as to the limits of the second respondent’s control over events.

The tribunal also, it appears to me, heavily relied, in its decision to award costs in favour of the second respondent, once again on its view that the underlying complaints as a whole were misconceived from the outset, and that the claimant should have realised that. Although the claimant also relied on complaints of procuring, aiding or abetting which the tribunal found were not made out on the facts, once again that is not the same as a conclusion that they had no reasonable prospects. In all events it is not clear to me whether, on what basis, or to what extent, the tribunal would have awarded costs to the second respondent had that been the sole feature on which it relied.

I have therefore come to the conclusion that the costs award in favour of the second respondent also cannot stand.

The costs appeal is therefore, in respect of both awards, allowed.

For all the foregoing reasons, the liability appeal is dismissed. The costs appeal is allowed. This decision was sent in draft to counsel under embargo; and I at that stage invited, and then received, written submissions on behalf of all parties as to the further order that I should make, in particular in relation to the costs aspect. Mr Panesar KC and Ms Levene both invited me to remit fresh determination of the costs applications to the same tribunal panel; Mr Allen KC contended that remission should be to a freshly constituted panel. I have considered their respective submissions and applied the well-known guidance in Sinclair Roche & Temperley v Heard [2004] IRLR 763.

There is force in the respondents’ submissions that the previous panel would have the advantage of its familiarity with the matter including the relevant litigation history. However, this can be overstated. It is also rightly not suggested that it would not be possible for a new panel to deal with the matter. They will be able to read the liability decision, and, of course either the existing or a new panel would need to take on board the ramifications of my decision. If the matter returns to a new panel, counsel will be able to draw their attention to other relevant documents or matters of record relating to the course and conduct of the litigation, including how any of the parties put, or revised, their cases, so far as said to be relevant. In either case, general case management of the process leading to fresh determination of the costs applications will be a matter for the tribunal.

Importantly, as I have described, the broad premise of the first costs decision was that the tribunal considered that the claimant’s case was fundamentally misconceived from the outset in almost every respect. This, together with the language used at various points – such as at [34], and in phrases such as “did not bear scrutiny” and “no evidence whatsoever”, suggest that the tribunal, first time around, determined the costs applications on the basis of a very strong and pervasive overall view. As I have described, this appears to have affected the outcomes of all three respondents’ applications. Though I am sure that they would do their conscientious best were I to remit the matter to them, it would be difficult to expect the same panel to come to the matter entirely afresh. It is also important that, whatever the outcome next time, the parties are able to have confidence in it.

I have therefore concluded that I should remit the costs applications for determination by a fresh tribunal panel.