Unite the Union v Nailard

This is an appeal and cross-appeal against a decision of the Employment Appeal Tribunal (HH Judge Richardson, Mr Peter Gammon CBE and Mrs Gillian Smith MBE) handed down on 27 September 2016. By its decision the EAT allowed in part an appeal against the decision of an Employment Tribunal sitting at Watford (chaired by Employment Judge Manley) dated 1 July 2015. In order to explain the issues I need to summarise the essential factual background (which I can take almost verbatim from Judge Richardson’s excellent judgment) and the procedural history. The Claimant in the original proceedings is the Respondent before us, but it will be clearer if I refer to her as the Claimant.

The Claimant was employed by the Appellant trade union (“the Union”) as a regional officer with effect from 6 June 2012.  In January 2013 she transferred to its office at Heathrow.  She became a regional officer in respect of Heathrow Airports Ltd (“HAL”) within the Union’s London and Eastern Region.  Her immediate line manager was Mr Wayne King, the senior regional officer.  Senior to him were Mr Kavanagh, the regional secretary, and Mr Murray, the Union’s chief of staff.  These were all employees of the Union.

The Union’s rule book – in effect its constitution – makes provision for local branches, generally based upon a workplace.  They are to have elected officials, including a chair, a treasurer, an equality officer and a secretary.  The rule book also provides for shop stewards and workplace representatives to be elected within a workplace.  Two such officials within Heathrow were Mr Saini, a convenor, and Mr Coxhill, a branch chair: I will sometimes in this judgment refer to them for short as “the lay officials”.  By agreement between the Union and HAL they carried out union duties full-time while remaining employed by HAL.

There was an unfortunate history of conflict between the employed officers of the Union and the locally elected officers at Heathrow.  The Claimant was the fourth employed officer in respect of HAL in as many years.  Mr Saini and Mr Coxhill, as the ET found, treated the Claimant in a bullying and offensive manner which amounted to harassment related to her sex within the meaning of section 26 of the Equality Act 2010. I need not give the details.

The Claimant complained to the Union about the conduct of Mr Saini and Mr Coxhill, eventually presenting a formal grievance on 17 March 2014.  Mr Murray asked Mr Hughes, another employed official of the Union, to carry out an investigation.  The ET found that although the Union – in particular Mr Hughes, Mr Murray and later Mr Kavanagh (to whom I will sometimes refer for short as “the employed officials”) – acknowledged that the Claimant was, as Mr Kavanagh put it in a contemporary e-mail, subjected to “a sickening and orchestrated campaign of harassment … [including] bullying and even sexual harassment”, it failed to deal with it firmly or decisively. For convenience, I will refer to that failure as their “inaction”, though that label slightly overstates the position since the Union did take some steps, albeit inadequate.

By August 2014 Mr Kavanagh was involved.  He decided to transfer the Claimant away from Heathrow.  At a meeting on 1 August he offered her a transfer to Southampton or alternatively to offices in the London area.  She protested, and on 4 August she resigned with immediate effect.  I will return later to the ET’s findings about the reasons for Mr Kavanagh’s decision.

For the purpose of the analysis that follows I will refer to the original conduct on the part of Mr Saini and Mr Coxhill as “the lay officials’ conduct”, and to the subsequent conduct of Mr Hughes, Mr Murray and Mr Kavanagh – that is, their inaction followed by the decision to transfer her – as “the employed officials’ conduct”.

The Claimant brought proceedings against the Union in the ET. Not all her claims are now live. Those that are can be summarised as follows:

The claim was heard over a number of days in April and May 2015. The ET’s decision was, as I have said, handed down on 1 July 2015. The Reasons are full and carefully structured. I should note at this stage that the ET proceeded by reference to an agreed list of no fewer than 31 issues. At paras. 11-14 of the Reasons it found the facts. At paras. 15-37 it set out the applicable law, an exercise which to some extent incorporated the parties’ submissions. It completed its summary of the submissions at paras. 38-42. At paras. 43-132 it considered and reached conclusions on each of the issues. At paras. 133-139 it gave an overall summary of its conclusions.

The findings with which we are concerned can be summarised as follows:

The Union appealed to the EAT. It did not challenge the finding of unfair dismissal. As regards the findings of harassment related to sex:

The EAT dismissed the Union’s appeal as regards the lay officials’ conduct. It overturned the ET’s finding that Mr Saini and Mr Coxhill were employees but it upheld the alternative finding that the Union was liable for their conduct as its agents. However, it allowed the appeal in relation to the conduct of the employed officials: it found that the ET had misdirected itself as to the necessary ingredients for liability and remitted this part of the claim for a re-hearing.

The Union appealed to this Court against the EAT’s decision that it was liable for the conduct of the lay officials. Permission was granted by Lewison LJ. The Claimant does not seek to resurrect the finding that Mr Saini and Mr Coxhill were employees of the Union but she resists the appeal on the basis that they were its agents. The Claimant also appealed against the EAT’s decision as regards the employed officials. Again, permission was granted by Lewison LJ. (I should add that he refused the Union permission to appeal on a further ground – “ground (3)” – concerned with the ET’s finding that the dismissal was discriminatory.)

The Union was represented before us by Mr Oliver Segal QC and Ms Katharine Newton and the Claimant by Mr Bruce Carr QC leading Mr James Wynne. Mr Segal and Mr Wynne appeared at both stages below. The case was very well argued on both sides. The one-day listing proved tight, and both parties, with the permission of the Court, lodged short written submissions, referring to some further materials, following the hearing.

Section 109 of the 2010 Act provides as follows:

“(1)

Anything done by a person (A) in the course of A’s employment must be treated as also done by the employer.

(2)

Anything done by an agent for a principal, with the authority of the principal, must be treated as also done by the principal.

(3)

It does not matter whether that thing is done with the employer’s or principal’s knowledge or approval.

(4)

In proceedings against A’s employer (B) in respect of anything alleged to have been done by A in the course of A’s employment it is a defence for B to show that B took all reasonable steps to prevent A -

(a)

from doing that thing, or

(b)

from doing anything of that description.”

I would note two points at this stage about the drafting of section 109:

The decision of this Court in Kemeh v Ministry of Defence [2014] EWCA Civ 91, [2014] ICR 625, is the most recent authoritative decision on the liability of a principal for an agent’s discriminatory acts. It was in fact concerned not with section 109 but with section 32 of the Race Relations Act 1976. This performed essentially the same role as section 109 but the two sections are not identically drafted. Section 32 read:

“(1)

Anything done by a person in the course of his employment shall be treated for the purposes of this Act as done by his employer as well as by him, whether or not it was done with the employer's knowledge or approval.

(2)

Anything done by a person as agent for another person with the authority (whether express or implied, and whether precedent or subsequent) of that other person shall be treated for the purposes of this Act as done by that other person as well as by him.

(3)

In proceedings brought under this Act against any person in respect of an act alleged to have been done by an employee of his it shall be a defence for that person to prove that he took such steps as were reasonably practicable to prevent the employee from doing that act, or from doing in the course of his employment acts of that description."

It will be noted that there is no equivalent in section 32 (2) to the phrase in sub-section (1) “whether or not it was done with the employer's knowledge or approval”. However Elias LJ, who delivered the leading judgment, said, at paras. 11-12 (p. 631 A-E):

“11.

Read literally, subsection (2) might suggest that the principal must authorise the act of discrimination itself before liability arises. But I agree with the EAT in Lana v Positive Action in Training (Housing) Limited [2001] IRLR 501 para 32 (Mr Recorder Langstaff presiding) that this would virtually render the provision a dead letter. In my judgment, Parliament must have intended that the principal will be liable wherever the agent discriminates in the course of carrying out the functions he is authorised to do [emphasis supplied]. It is a moot point whether the common law would in any event impose liability in these circumstances. The scope of the principal's liability for an agent at common law is not entirely clear, although it seems likely that he will be liable for certain tortious acts of the agent, such as misrepresentations, provided they are sufficiently closely related to the agent's actual or apparent authority: see Bowstead and Reynolds on Agency, 19^th edition, para 8-182. Whether racial abuse would fall within that principle is problematic, but section 32(2) removes the uncertainty which might otherwise exist.

12.

If that analysis is right and the principal can be liable even though he has not authorised the act of discrimination itself, it follows that the act itself may be - and no doubt usually will be - without the principal's knowledge or approval. It is perhaps surprising that the draftsman did not make this plain in sub-section (2) as he did in subsection (1). It might be argued that the omission indicates a conscious decision by Parliament that in the case of a principal he should only be liable if he knows or approves of the discriminatory act itself, but I am not persuaded that this is a legitimate inference to draw from the difference in wording of the two subsections. Indeed, in the Equality Act 2010 section 109(3) expressly states that the principal will be liable irrespective of whether he knew or approved of the act of discrimination.”

The reasoning in that passage addresses the point to which I drew attention at para. 16 (1) above. Elias LJ holds that the effect of section 67 (2) is that (in the words I have italicised in para. 11) “the principal will be liable wherever the agent discriminates in the course of carrying out the functions he is authorised to do”, whether the acts in question were specifically authorised or not. That interpretation must apply also to section 109 (2); and indeed Elias LJ points out that his reading of section 67 (2) is given express effect by section 109 (3), which had by then been enacted.

Turning to the particular issue in Kemeh, the claimant was a black soldier working as a cook in the Falklands Islands garrison. The butchery department in the mess where he worked was partly staffed by civilian employees of Sodexo under a sub-contract from Serco, who had a contract with the Ministry of Defence (“the MoD”) to provide facilities management services in the South Atlantic. He made a request to one of the Sodexo staff, a Ms Ausher, for a supply of meat. She refused to give him the quantity he asked for, giving a racially offensive reason. He brought proceedings against the MoD for racial discrimination, claiming that the remark had been made by her as agent for the Ministry and with its authority within the meaning of section 32 (2). The claim succeeded in the ET but that decision was reversed by the EAT.

This Court dismissed the claimant’s appeal. The essential question was, as I have said, whether Ms Ausher made her remark in the course of carrying out the functions which she was authorised by the MoD to do. On the face of it she was not authorised by the MoD to do anything: she was employed by Sodexo. But the claimant argued that in the context of section 32 the concepts of “agent” and “authority” should be given a meaning wider than at common law, so as to apply to any “relationship, other than an employment relationship, where the agent is (1) subject to a degree of direction from the principal; (2) where there is a degree of integration with the principal's employees; and (3) where there is a degree of proximity between the agent and principal” (see para. 27 of Elias LJ’s judgment (pp. 633-4)). Elias LJ observed that that submission bore no relationship to the concept of agency as normally understood. At para. 46 of his judgment (p. 639 A-B) he said:

“Whatever the precise scope of the agency concept in section 32 - and in my view it must at least reflect the essence of the legal concept - Ms Ausher does not fall within it.”

At para. 70 of his concurring judgment (p. 643 G-H) Lewison LJ said that the terms “agency” and “authority” must be “interpreted in accordance with ordinary legal parlance”. On that basis Ms Ausher could not be said, when doing her work in the butchery department, to have been acting as an agent of the MoD. At para. 41 of his judgment (pp. 637-8) Elias LJ said:

“Ms Ausher's contract with Sodexo is the source of any authority she has to make decisions relating to the butcher's department in the mess. … [There was no] authorisation by the MoD to allow Ms Ausher to act on its behalf with respect to third parties.”

And, at para. 44 (p. 638G):

“Ms Ausher may be said in a general sense to be working for the benefit of the MoD, but she is not acting on its behalf. She is not, as it were, standing in the shoes of the MoD in relation to independent third parties.”

The confirmation in Kemeh that section 109 (2) is rooted in “the legal concept of agency” led Mr Segal to place reliance on the definition of agency in article 1 of Bowstead and Reynolds. Rules (1) and (2) read (so far as material):

“(1)

Agency is the fiduciary relationship which exists between two persons, one of whom expressly or impliedly manifests assent that the other should act on his behalf so as to affect his relations with third parties, and the other of whom similarly manifests assent so to act or so acts pursuant to the manifestation. …

(2)

In respect of the acts to which the principal so assents, the agent is said to have authority to act; and this authority constitutes a power to affect the principal’s legal relations with third parties.”

The following commentary elaborates the propositions in those rules. Mr Segal referred us to a passage in para. 1-002 which states that “for the purposes of most legal usage” an essential element is that the agent has authority to affect the principal’s relations with third parties. He also referred us to the passages to the same effect in paras. 1-003 and 1-004.

It is right to note – though the point is not in fact central to the issue in this case – that in the commentary the authors make it clear that the definition in article 1 involves choices on their part which are not necessarily – if I may be forgiven for putting it this way – definitive; and that “there are many acceptable uses of the term [agency] which do not always coincide with each other” (see para. 1-002). They also observe, at para. 1-003, that “no one has the monopoly of the ‘correct’ use of this or any other term”.

In the present case we are concerned specifically with the authority of lay officials of a trade union – in older terminology, shop stewards. In that regard Mr Segal referred us to the decision of the House of Lords in Heatons Transport (St Helens) Ltd v Transport and General Workers’ Union [1973] AC 15. In that case TGWU shop stewards had been found to have engaged in an unfair industrial practice, within the meaning of the Industrial Relations Act 1971, by organising the blacking of lorries operated by the complainant. The issue was whether the union was liable for their actions. The House of Lords held that it was. Lord Wilberforce, who delivered the opinion of the appellate committee, said, at p. 98H:

“The essential issue is … one of authority; the authority of the shop stewards to take the action complained of on behalf of the union.”

He noted, at p. 100, that the issue only concerned liability under the 1971 Act, though he observed that liability for tortious acts might be “a closely connected subject”.

Lord Wilberforce started his exposition of the applicable principles by saying, at p. 99 C-D:

“No new development is involved in the law relating to the responsibility of a master or principal for the act of a servant or agent. In each case the test to be applied is the same: was the servant or agent acting on behalf of, and within the scope of the authority conferred by, the master or principal ?”

He noted that normally the authority of a servant – in more modern language, an employee – was wider than that of an agent, but he said that that was not invariably the case and that were cases where an agent is elected to a position which gives him “authority of considerable generality” (pp. 99-100). He then analysed the position of shop stewards under the rules of the TGWU and concluded that the acts in question were within their authority. His summary, at pp. 112-113, reads (so far as relevant):

“In accordance with the policy of devolution followed by the Transport and General Workers' Union, and consistently with its rules and practice, shop stewards of the union have a general implied authority to act in the interests of the members they represent and in particular to defend and improve their rates of pay and working conditions. They may do so by negotiation or by industrial action at the relevant place of work. They are not authorised to do any act outside union rules or policy.”

Both parties made various references to the Union’s rule book, and it is obvious – and confirmed by Heatons Transport – that in principle the express terms of the rules will be an important, even if not the only, source for establishing the scope of the authority of lay officials. In truth, however, it is common ground that it was part of the role of Mr Saini and Mr Coxhill as branch officials – more in fact by necessary implication than by reference to any express provision in the rules – to act as the Union’s representatives in its relations with HAL and also with its members. As will appear, the real issue in this appeal is about the legal consequences of that role, on which the rules have nothing to contribute. I do not therefore propose to attempt any summary of them here.

The ET’s reasoning on this aspect is at para. 90 of its Reasons. It had in the previous passage held that Mr Saini and Mr Coxhill were employees of the Union. It said:

“If we are wrong about that we must consider whether the agency provisions apply to these two individuals.  The case of Kemeh … is helpful in this case and also the extract provided to us by the claimant’s representative with respect to agency [this appears to be a reference to Bowstead and Reynolds].  We do find that they were acting as agents for the respondent.  They were carrying out work on behalf of the respondent in their dealings with local members, officers, other trade unions and the employers.  They had express authority to do so through the rule book and on the basis of credentials provided by the respondent.”

The EAT started its consideration of the issue by setting out the reasoning of the ET and summarising counsel’s submissions: see paras. 40-42 (pp. 130-1). The core of Mr Segal’s submissions for the Union was summarised at para. 41 as follows:

“He submitted that Mr Saini and Mr Coxhill (1) were not carrying out activities assigned to them by the Respondent; (2) were not under the control of the Respondent, which could not direct what they did; (3) were not in a fiduciary position; (4) did not have any express or implicit authorisation from Respondent to act as they did.  On this last question he submitted that the authority of shop stewards and elected officials was limited: they were not authorised to do any act outside union rules or policy - see Heatons Transport v Transport and General Workers’ Union [1972] ICR 308 HL at 405G.”

As will appear, most of those points were not pursued before us.

At para. 43 (p. 131 E-F) the EAT summarised the ET’s findings as to the circumstances in which the acts of harassment found against the lay officials took place, as follows:

“The words and acts found by the ET to constitute sexual harassment generally took place in the context of meetings.  Thus Mr Saini was found to have conducted himself in this way at a meeting with HAL management on 24 February 2014.  He attacked, abused and threatened her at a branch meeting so she felt physically unsafe.  At pay negotiations in December 2013 and on other occasions he informed HAL negotiators that they were not to contact her about certain issues.  He would cut across her, shout her down and accuse her of being ‘on the take’ at meetings.  He called her ‘headmistress’.  Mr Saini said at a meeting of HAL chairs and convenors that he ‘wanted that woman off the airport’.  He wrote to HAL informing them that the Claimant was no longer to act on behalf of the members of his branch in any capacity.  The findings against Mr Coxhill are less extensive; but they include a specific finding that he used overtly sexual language towards her at a meeting on 11 March 2014.”

At paras. 44-51 the EAT summarises the effect of Kemeh and Heatons Transport. It states its conclusion, at paras. 52-53 (p. 133 E-H), as follows:

“52.

The ET found that Mr Saini and Mr Coxhill were acting as agents for the Respondent because they were carrying out work on behalf of the Respondent in their dealings with local members, officers, other trade unions and employers.  We consider that the ET was entitled to reach this conclusion.  It is important to keep in mind that a union is a contractual association of subscribing members; this association authorises its officers to act on its behalf in a variety of ways - conducting branch meetings, representing the union at meetings with other unions and employers, liaising with employed officers and holding internal meetings to prepare for external meetings.  Such matters are core union work.  It is well within the scope of authority of local officers to speak on the Respondent’s behalf at such meetings and to correspond with the employer about matters concerning the manner and scope of negotiations.  It is well within the scope of authority of local officers to liaise with employed officers.  As Lord Wilberforce observed [in Heatons Transport], an office holder may have to perform a variety of functions in the course of the duties of office; we have no doubt that this applied to the Respondent’s office holders at HAL.

53.

Contrary to Mr Segal’s submission we consider for these reasons that the ET was entitled to find that the branch officers were acting within the scope of their authority as officers of the Respondent when speaking at meetings concerning matters to be negotiated with HAL and when corresponding about such negotiations.”

In short, therefore, the EAT held that the ET was entitled to find that the acts of Mr Saini and Mr Coxhill towards the Claimant were to be treated as acts of the Union, in accordance with section 109 (2), because they were done in the course of the performance of their functions on behalf of the Union as prescribed by the rules – specifically, “conducting branch meetings, representing the union at meetings with other unions and employers, liaising with employed officers and holding internal meetings to prepare for external meetings”.

At paras. 54-59 the EAT considers other points made by Mr Segal, most of which are not pursued before us. But at para. 55 it addressed his point that Mr Saini and Mr Coxhill could not be agents because they owed no fiduciary duties to the Union. It said:

“We do not think this is the place for a detailed consideration of the fiduciary duties of a branch officer; but in principle we see no reason why these should not exist.  Take, for example, the extreme case of a branch officer who used his position to take bribes; we see no reason why he should not be accountable to the Respondent as owing a fiduciary duty.”

Mr Segal’s case before us was more focused than the points summarised by the EAT at para. 41 of its judgment and comes down essentially to a single point. He accepted that, as the EAT held, the acts complained of were indeed done by Mr Saini and Mr Coxhill in the course of functions which they were authorised by the Union to perform. But he submitted that that was not enough. The acts in question had to be acts towards third parties, because it was only in relation to such acts that they could be said to be acting (in the language of section 109 (2)) “with the authority of the principal” or (as glossed in Kemeh) “on [the Union’s] behalf”. The ET and the EAT had ignored the essential distinction between the lay officials’ dealings with HAL and their dealings with the Claimant (even if those dealings occurred in the course of meetings with HAL). As regards the former they were indeed acting on behalf of the Union, but as regards the latter it made no sense to say that they were acting towards her “on behalf of the Union”: she was not a third party but herself, like them, a representative of the Union towards whom they required no authority to act. In his post-hearing written submissions Mr Segal summarised his position as follows:

“A person (A) is appointed by another person (P) to a role/job in which he is inter alia authorised by P to perform certain acts as P’s agent so as to affect P’s legal relations with certain third parties (“the Authorised Acts”). Does A act as P’s agent, with the legal authority of P:

(i)

(Claimant’s case) whenever A does something in that role/job, regardless if it is one the Authorised Acts – so that P is liable to all persons, not just those third parties, for any act done by A in that role ? or

(ii)

(Union’s case) only when A does an Authorised Act in respect of one of those third parties?

The second is the correct legal position.”

In support of that submission Mr Segal referred to the passages from Bowstead and Reynolds which I have set out or referred to at para. 22 above and to Kemeh and Heatons Transport. Specifically:

Mr Segal also referred us to the decision of the EAT in Commissioner of Police of the Metropolis v Weeks [2011] UKEAT 013/11, but I do not think that that case advances the argument. Quite apart from the fact that it is not binding on us, the issue of agency arose in a highly fact-specific situation and the judgment of Judge McMullen QC contains no general discussion of the issue which arises in this appeal.

Mr Segal sought to reinforce his submission by reference to the absence of a “reasonable steps” defence for principals: see para. 16 (2) above. He submitted that the EAT’s approach in practice involved equating the positions of agents and employees – that is, that the principal was liable for anything done by the agent in the course of his or her performance of their role. But if so it could not be right that principals should have a co-extensive liability with employers and yet not enjoy the same defence. That anomaly would be avoided if it were recognised that the liability of principals under sub-section (2) was more limited in scope than that of employers under sub-section (1).

As regards the relevance of whether the lay officials owed fiduciary duties to the Union, he accepted the EAT’s point that they might do so. But the point remained of some relevance since it could hardly be said that they owed the Union such duties in their dealings with the Claimant. This seems to me simply another, and if I may so, somewhat marginal aspect of his central submission.

Mr Carr supported the reasoning of the EAT. He submitted that it was enough that Mr Saini and Mr Coxhill did the acts in question in the course of performing their roles “as the Union’s representatives in relation to its members and third parties”. He contended that the essential question under section 109 (2) was whether the person who did the act complained of was in an agency relationship with the principal: if they were, then liability attached to the principal for “anything done” in connection with the exercise of their role. The requirement that the agent be standing in the principal’s shoes as regards third parties was necessary to establish the agency relationship in the first place; but it did not follow that the act complained of had to be done to such a third party.

Mr Carr submitted that that result produced the right answer in policy terms. If both the Claimant and the lay officials had been employees of the Union it would unquestionably have been liable for harassment of her by them which occurred while they were working together in the performance of their duties on behalf of the Union. Parliament cannot have intended that the fact that lay officials were not employees would produce a different result. He sought to reinforce that submission by reference to section 57 of the 2010 Act. This proscribes discrimination, harassment and victimisation by a “trade organisation”, which would include a trade union, against its members. Most members of a trade union deal primarily not with its full-time (employed) officials but with the lay officials of their own branch. If Mr Segal’s submissions were correct it would mean that if in the course of such dealings an official discriminated against, or harassed, a member the Union would not be liable. Again, Parliament could not have intended such a result: the answer was to treat acts done by lay officials in the course of the performance of their roles as done with the union’s authority.

As regards Mr Segal’s point based on section 109 (4), Mr Carr submitted that if the denial of the reasonable steps defence to principals was indeed an anomaly it would not be removed by reducing the scope of section 109 (2) in the way contended for. It might be mitigated, in the sense that it would arise in fewer cases, but such mitigation would be logically accidental, and was not relevant to the correct construction of the sub-section.

I believe that the decision of the EAT was right. My reasons, which are essentially in line with Mr Carr’s submissions, are as follows.

The starting-point is the statement of Elias LJ at para. 11 of his judgment in Kemeh that the effect of what is now section 109 (2) is that “the principal will be liable wherever the agent discriminates in the course of carrying out the functions he is authorised to do” (see para. 19 above). That formulation effectively equates the circumstances in which a principal may be liable for the acts of an agent with the “course of employment” test governing the liability of employers for the acts of their employees. It may well extend the scope of the liability beyond what would apply at common law (Footnote: ²); but there is no reason why Parliament should not have chosen to effect such an extension in discrimination cases. (This approach is not of course in any way inconsistent with the Court’s insistence that the terms “agent” and “authority” should be construed in accordance with “ordinary legal parlance”: that point was being made in response to the argument about whether Ms Ausher was the MoD’s agent at all, not to the question of whether it would be liable for her acts if she were.)

If, therefore, the effect of the language of section 109 (2) is to render a principal liable for the acts of his or her agents done in the course of the performance of their authorised functions, I can see no justification for limiting that liability in the way proposed by Mr Segal. An agent may stand in the shoes of the principal in dealing with A, but if while wearing them he treads on B’s toes I see no good reason why he should not be liable to B just as much as if it had been A’s toes that were crushed: in both cases the wrong is done in the course of performing the authorised functions. The proposition based on Bowstead and Reynolds that it is inherent in the principal/agent relationship that the agent be in a position to affect the principal’s legal relationship with third parties is fine as far as it goes (Footnote: ³), but it misses the point that we are not here considering whether an agency relationship exists at all but with liability in tort for acts done in the course of it. The same goes for the passages which Mr Segal relies on from Kemeh. The question in that case was indeed whether there was any agency relationship between Ms Ausher and the MoD, and that is why it was relevant for Elias LJ to point out that she had no authority to act on its behalf as regards third parties. But that is not the question here. Heatons Transport is even less relevant. In truth its only value is to the Claimant, since it confirms – though in fact this is not contentious – that the kind of dealings in the context of which the harassment of the Claimant took place were functions which the lay officials were performing on behalf of the Union.

I agree with Mr Carr that that result makes sense in policy terms. The distinction on which Mr Segal relies would plainly have no application in the case of an employee: indeed in very many, if not most, cases where an employer is liable for an employee’s discriminatory acts by virtue of section 109 (1) the claimant is a fellow-employee rather than a “third party”. It would be surprising and unsatisfactory if the position were different where the discriminator was acting on behalf of the organisation in question but under some different arrangement: the conduct complained of would still be being done by the agent in the course of the principal’s business. The anomalies which would result are particularly obvious in the case of an organisation where persons who are not employees (even in the extended sense) perform important functions, as in the case of trade unions. Mr Carr’s example based on liability under section 57 is apt.

As regards Mr Segal’s point based on section 109 (4), I acknowledge that it is an apparent anomaly that the reasonable steps defence is not available to principals as regards the acts of their agents when it is available to employers as regards the acts of their employees (Footnote: ⁴). But that anomaly would be present whichever construction of section 109 (2) is preferred; and I accept Mr Carr’s submission that it cannot be right in principle to adopt what would otherwise be the wrong construction simply so that the anomaly would arise in fewer cases.

Section 13 (1) of the 2010 Act defines direct discrimination. It reads:

“A person (A) discriminates against another (B) if, because of a protected characteristic, A treats B less favourably than A treats or would treat others.”

The Act does not treat harassment as a form of discrimination but as conceptually distinct. It is defined in section 26. I need only set out sub-section (1), which reads (so far as material) as follows:

“A person (A) harasses another (B) if—

(a)

A engages in unwanted conduct related to a relevant protected characteristic, and

(b)

the conduct has the purpose or effect of—

(i)

violating B's dignity, or

(ii)

creating an intimidating, hostile, degrading, humiliating or offensive environment for B.”

One of the protected characteristics is sex: see sub-section (5). I will refer to the consequences specified at (i) and (ii) under (b) as “the proscribed consequences” and to an environment of the kind described at (b) (ii) as a “hostile environment”. The key phrase for the purpose of the issues which we have to decide is “related to [sex]” under (a).

I should mention a problem about language. It is natural to describe harassment under section 26 in a case where the protected characteristic is sex as “sexual harassment”. Both the ET and the EAT in this case sometimes use that shorthand and I will do the same. But in ordinary language “sexual harassment” is often used more specifically to refer to conduct of an overtly sexual character. It should be borne in mind that section 26 may cover other kinds of conduct which has the purpose or effect of producing one of the proscribed consequences, provided always that it is “related to sex”: in the present case there were elements of both.

Discrimination against, and harassment of, employees (and applicants for employment) are proscribed by sections 39 and 40 of the Act respectively. I need not set out section 39. Section 40 reads:

“(1)

An employer (A) must not, in relation to employment by A, harass a person (B)—

(a)

who is an employee of A's;

(b)

who has applied to A for employment.”

(Despite the “(1)”, there are no other sub-sections: this is a result of the legislative history explained below.)

Section 136 of the 2010 Act contains provisions about the burden of proof in claims under the 2010 Act. So far as material, it reads:

“(1)

This section applies to any proceedings relating to a contravention of this Act.

(2)

If there are facts from which the court could decide, in the absence of any other explanation, that a person (A) contravened the provision concerned, the court must hold that the contravention occurred.

(3)

But subsection (2) does not apply if A shows that A did not contravene the provision.

(4)-(6) …”

It is necessary to say something about the legislative history of sections 13, 26 and 40 of the 2010 Act.

In the various provisions defining direct discrimination in the predecessor legislation the phrase “on the grounds of” (or a cognate) was used, rather than the formulation “because of” which appears in the 2010 Act. It is now well established that the change was not intended to make any substantive difference: see, e.g., para. 19 of my judgment in Onu v Akwiwu [2014] EWCA Civ 279, [2014] ICR 571 (p. 581 C-E).

In the earliest versions of the discrimination legislation there was no distinct proscription of harassment. Cases of what we would now regard as harassment were brought as cases of ordinary direct discrimination. The fit with the legislative language was awkward, and some difficult case-law was generated. However, an amended version of the EU Equal Treatment Directive (EU/2002/73 EC), promulgated in 2002, required member states to proscribe “harassment”, which was defined in the Directive as “where unwanted conduct related to the sex of a person occurs with the purpose or effect of violating the dignity of a person, and of creating an intimidating, hostile, degrading, humiliating or offensive environment [my emphasis]”.

That requirement was sought to be implemented in 2005 by secondary legislation which inserted an express prohibition on harassment – section 4A – into the Sex Discrimination Act 1975. (Similar amendments were made to the legislation relating to other protected characteristics.) Section 4A essentially tracked the Directive, save that it used the formulation “on the ground of her sex” – that is, the same language as in the definition of direct discrimination – rather than “related to sex”.

The Equal Opportunities Commission believed that the amendment legislation failed in that respect – and in several others – to conform to the requirements of the Directive. It brought judicial review proceedings. In Equal Opportunities Commission v Secretary of State for Trade & Industry [2007] EWHC 483 (Admin), [2007] ICR 1234, (“the EOC case”) Burton J upheld the Commission’s challenge. We are only concerned with two of the grounds of challenge, which I take in turn.

First, the Commission argued, and Burton J accepted, that the Directive’s formulation of “related to … sex” proscribed not only harassment which was “caused by” the claimant’s sex but also harassment which was “associated with” it: see paras. 6-28 of his judgment. Burton J illustrated the distinction between the two types of case, at paras. 10-11 (p. 1242-3), by accepting three examples taken from the case-law by counsel for the Commission (Dinah Rose QC), namely:

Secondly, the Commission contended that the Directive required member states to provide for employers to be liable in cases where they failed to take reasonable steps to protect employees against harassment by third parties – labelled as “third party liability” (Footnote: ⁵). However in her oral submissions Ms Rose retreated from that position. She is recorded by Burton J at para. 40 of his judgment (p. 1250H) as acknowledging that:

“… there is nothing explicit, or even arguably implicit, in any of the Articles requiring a Member State to impose vicarious liability on an employer, or indeed liability for negligent failure to take steps, such as were expressly found by the House of Lords in Pearce not to exist in the present discrimination legislation”. 

(The reference to Pearce is to the decision of the House of Lords which I consider at para. 88 below.) Her submissions proceeded on the basis that, even if there was no such requirement in the Directive, the Secretary of State himself had accepted that in some circumstances third party liability might be appropriate – see paras. 36 and 37 of the judgment (pp. 1249-50). Her argument was that the language of “on the grounds that” precluded such an outcome, but that the problem could be resolved by “adopting an associative rather than causative approach” which she was already contending for on other grounds. As I read his judgment, Burton J accepted that that was so, and the effect of his decision to require the adoption of “associative” language gave Ms Rose what she was asking for. But that was not in itself the reason for adopting that construction, and in para. 63 of his judgment, where he gives his final decision, he does not expressly refer to the third party liability point (see p. 1258B).

In response to the decision in the EOC case the Secretary of State, exercising his powers under the European Communities Act 1972, made the Sex Discrimination (Amendment of Legislation) Regulations 2008, which took effect from 6 April 2008. So far as relevant for present purposes the Regulations did two things:

As originally enacted, section 40 had three further sub-sections, as follows:

“(2)

The circumstances in which A is to be treated as harassing B under subsection (1) include those where—

(a)

a third party harasses B in the course of B's employment, and

(b)

A failed to take such steps as would have been reasonably practicable to prevent the third party from doing so.

(3)

Subsection (2) does not apply unless A knows that B has been harassed in the course of B's employment on at least two other occasions by a third party; and it does not matter whether the third party is the same or a different person on each occasion.

(4)

A third party is a person other than—

(a)

A, or

(b)

an employee of A's.”

As will be seen, those sub-sections were designed to impose liability on employers who had failed to take reasonable steps to prevent their employees being harassed by third parties.

However by section 65 of the Enterprise and Regulatory Reform Act 2013 sub-sections (2)-(4) were repealed, and there is now no explicit liability on an employer for failing to prevent third party harassment.

Although in this part of the appeal we are concerned with the conduct of the employed officials, the starting-point must be the conduct of the lay officials with which they were found to have dealt inadequately and which was found to have contributed to the eventual decision to transfer the Claimant. It is unnecessary to give details of the conduct itself, but it is important to note that it included an incident in which Mr Coxhill made an explicitly sexual “joke” about her – the so-called “dropped pen incident” – and referred to “the time of the month”; and that Mr Saini was found to have used derogatory gender-specific language about her such as “the headmistress” or “that woman”. The Tribunal found at sub-para. 13.31 of the Reasons that “Mr Saini’s treatment of the claimant was influenced by discriminatory motives”.

The way that the employed officials dealt with the Claimant’s complaints of how she was treated by Mr Saini and Mr Coxhill is the subject of detailed findings in para. 13 of the Reasons. I need not set them out: all that matters for this appeal is that they were found, as I have said, to have acted insufficiently firmly or decisively. But I should note that it was found that the employed officials believed that the Claimant’s complaints were justified and were aware that they included an element of “sexual harassment”. There are no explicit findings about the reason for their inaction.

Mr Kavanagh’s decision to transfer the Claimant from Heathrow is dealt with at para. 14 of the Reasons. Again, I need not set out the details. But at sub-para. 14.15 the Tribunal made findings as to the reasons for that decision. The sub-paragraph starts:

“Mr Kavanagh had mixed reasons for his decision to move the claimant. It was to remove her from the bullying and harassment. That included an element of harassment related to sex carried out by the elected officials, Mr Saini and Mr Coxhill, as well as more generic bullying.”

It then says that his decision was “triggered” by the emergence of a DVD of a television interview which the Claimant had given many years previously in which she appeared to be critical of industrial action being taken by some members, and which was thus potentially embarrassing. However, it continues:

“Other matters were of greater concern. He was aware, of course, of the sexual harassment carried out by Mr Coxhill. He was also aware that the claimant had stated that the reasons for Mr Saini’s treatment of her were because she was a woman. He must have understood that the environment she was working in was hostile and intimidating. Those factors played a material part in his decision to transfer the claimant. It was therefore tainted by discrimination.”

At sub-para. 14.16 the Tribunal goes on to say – effectively by way of anticipation of its dispositive reasoning:

“We say here for completeness and will confirm in our conclusions that the decision to transfer the claimant itself amounted to unwanted conduct that was related to sex (because of the background of harassment related to sex) and had the effect of violating her dignity and of creating a hostile and intimidating environment. Although Mr Kavanagh was not guilty of any discriminatory motive, it cannot be said that the decision to transfer which was made, against the wishes of the employee, part of which was because of sexual harassment, was itself free of any discrimination. The decision to transfer the claimant was tainted by discrimination. It was also unwanted conduct related to sex which had the effect of violating her dignity.”

The ET discussed in general terms the effect of the relevant statutory provisions at paras. 17-21 of the Reasons. Although the discussion is careful it does not shed any real light on the Tribunal’s reasoning on the disputed issues, and I do not believe that I need set it out.

I start with the Tribunal’s reasoning on the discrimination claim, which appears at paras. 100-104 of the Reasons. At para. 101 it confirmed, in accordance with its previous findings, that the purely factual elements of the conduct alleged against both the lay and the employed officials were made out. At para. 102 it found that the lay officials’ conduct was “because she was a woman”: it referred to both the dropped pen incident and the use of gender-specific language. As regards the employed officials it said:

“The failures by the respondent to protect her, take appropriate steps in investigating and disciplining those responsible and the decision to transfer her without her consent arose from those difficulties she faced because she is a woman. The burden therefore shifts to the respondent to explain.”

At para. 104 the Tribunal goes on to find that the Union had not discharged that burden: the details of the reasoning relate to a specific point on which the Union had sought to rely and I need not explain it here. That two-stage approach of course reflects the structure of the burden of proof provisions.

I turn to the harassment claim. At para. 106 the Tribunal held that the conduct which it had found on the part of both the lay and the employed officials was “unwanted”, and at para. 107 that it had the effect of creating a hostile environment for the Claimant. There is no issue about those findings and I need not set them out. Paras. 108-109 deal with the issue whether that unwanted conduct was, in accordance with section 26 (1) (a), “related to” the Appellant’s sex. They read:

“108.

Issue 13: “Are there facts from which the tribunal could properly decide that any such conduct that is established was related to the claimant’s sex?”, taken together with Issue 14: “If so has the respondent shown that it was not so related?”

It must be clear from what we have said that the conduct was related to sex. We find this to be so not only for the more obvious dropped pen incident but also for Mr Saini’s treatment of the claimant. The respondent has not shown that it was not related to her sex.

109.

Although we accept that the conduct of those investigating matters and taking other decisions affecting the claimant is less obviously related to sex, we have found that it is so related. The claimant’s complaints were clear and unambiguous and the respondent’s failure to act appropriately with respect to those matters must therefore also be related to sex. The final decision to transfer the claimant, who was a victim of bullying and harassment which included an element of sex discrimination, was related to sex. If we are wrong about this, as we say above, we find it was an act of sex discrimination rather than conduct which amounts to harassment.”

The reasoning in para. 109 is succinct, but I understand the phrase “we have found” to be intended to refer back to the earlier passages which I have set out or summarised.

As already noted, the Tribunal concluded the Reasons with a summary of what it had decided and why. Most of it – as indeed one would expect – does not add to the more detailed preceding reasoning. But Mr Carr drew our attention to para. 138, which is directed to the question whether the Claimant’s constructive dismissal could itself be held to be an act of sex discrimination. In the course of that paragraph the Tribunal said:

“It cannot be said that the decision to transfer and the consequent resignation was ‘in no way’ tainted by unlawful discrimination. It cannot be said that the protected characteristic in question, that of sex, did not influence Mr Kavanagh when he decided to transfer the claimant, as he was under no illusion that there had been sexual harassment of the claimant.”

I need to identify the reasoning in those paragraphs with some care. The Tribunal’s reason for finding that the employed officials acted as they did “because of” the Claimant’s sex was that their inaction, and Mr Kavanagh’s decision to transfer her, “arose from those difficulties she faced because she is a woman”. It is for that reason that it was “tainted by discrimination” (a phrase also used in para. 14.16). It did not base its conclusion on any finding that they were themselves (consciously or sub-consciously) motivated in their inaction, or (in the case of Mr Kavanagh) in the decision to transfer the Claimant, by the fact that she was a woman or that her complaint was a complaint about sexual harassment. Likewise, it is clear from the central two sentences of para. 109 that the finding that their conduct was “related to” the Claimant’s sex for the purpose of the harassment claim was based on the nature of her complaints against the lay officials. I will repeat them for ease of reference:

“The claimant’s complaints were clear and unambiguous and the respondent’s failure to act appropriately with respect to those matters must therefore also be related to sex. The final decision to transfer the claimant, who was a victim of bullying and harassment which included an element of sex discrimination, was related to sex.”

The “therefore” which I have italicised is explicit. It is also clear that the same reasoning is implicit in the second sentence; but in any event para. 14.16 explicitly acquits Mr Kavanagh of any “discriminatory motive” but says that the transfer is nevertheless “unwanted conduct that was related to sex (because of the background of harassment related to sex) [emphasis supplied]”. The same approach can be seen, though not quite so clearly articulated, in para. 138.

I should say, without at this stage expressing any view on the substance of the issues, that the phrase “tainted by discrimination” is dangerous. If it is only a shorthand intended to paraphrase the statutory language, well and good; and I am aware that it is sometimes so used in the case-law. But there is always a risk that such metaphorical paraphrases can distract attention from the statutory language, and the key authorities in which that language is expounded.

The EAT, as I have said, held that the ET had misdirected itself as to the necessary ingredients of liability for harassment under sections 26 and 40. It came to the same conclusion about the ET’s fallback finding of liability for direct discrimination under sections 13 and 39. Its reasoning is to be found at paras. 81-105 of its judgment. I can summarise it as follows.

The EAT starts with liability for direct discrimination. It refers at para. 83 to the well-known authorities which establish that, in a case where the act complained of is not inherently discriminatory, liability will depend on whether the protected characteristic formed part of the “mental processes” or “motivation” (which in this context is not the same as “motive”) of the putative discriminator, that being “the reason why” they did it. It refers specifically to the speech of Lord Nicholls in Nagarajan v London Regional Transport [2000] 1 AC 501, at p. 511, and to my own judgment in the EAT in Amnesty International v Ahmed [2009] UKEAT 0447/08/1308, [2009] ICR 1450, at para. 34 (p. 1470 A-C). It also, at paras. 86-87, refers to the decision of this Court in CLFIS (UK) Ltd v Reynolds [2015] EWCA Civ 439, [2015] ICR 1010, which points out that the scheme of the Act is that wherever an employer is liable for discrimination the employee who did the actual act complained of will be individually liable as well.

Mr Segal had argued that those authorities showed that the approach taken by the ET, as I have analysed it at para. 69 above, was wrong in law. At para. 88 the EAT accepted that submission. It said:

“In order to decide whether the treatment of the Claimant by the paid officers was less favourable treatment on the grounds of sex the ET was required in each case to focus upon their mental processes.  It was not correct to say that their decisions were because of sex simply on the ground that they were ‘tainted’ by the conduct of Mr Coxhill and Mr Saini.  This seems to have been a key part of the reasoning of the ET: see especially paragraphs 14.15, 14.16 and 102.”

It proceeded at para. 89 to consider whether its conclusion required that the case should be remitted to the ET. It held that it did. It said:

“If the ET meant, in paragraph 42 and elsewhere in its Reasons, to find that the paid officers, in the failures to act or the decision to transfer, were not themselves influenced by the protected characteristic of sex, that would seem to indicate that the Respondent discharged any burden of proof on the reason why question.  But we are not sure whether this is what the ET meant.  There is, as we have seen, a difference between motive and ‘reason why’; and it is possible that the ET used the word only on the question of motive.  The question of direct discrimination will therefore have to be remitted.”

(The reference to para. 42 of the Reasons seems to be a slip, and I have not identified the particular passage which the EAT had in mind. But that does not affect the reasoning.)

The EAT then turned to the question of harassment. It made two points by way of preliminary:

At paras. 93-94 it considered Conteh v Parking Partners Ltd [2010] UKEAT 0288/10/1712, [2011] ICR 341. In that case the respondent worked as a contractor for another company. The claimant was one of its employees who had been subjected to racial abuse by an employee of the client. She brought proceedings against the respondent for direct discrimination and harassment on the basis that it had failed to take effective action to deal with the problem. Langstaff P in the EAT held that inaction of that kind could, depending on the facts, be found to have created a hostile environment within the meaning of section 3A of the Race Relations Act 1976 (which corresponded to section 4A of the 1975 Act). However the section required that the conduct complained of be “on grounds of” race. There was no finding in the ET that the employee’s race was the ground for the employer’s inaction. Langstaff P held that that was fatal to her claim. He said, at paras. 30-31 (p. 1349 B-E):

“30.

The ‘unwanted conduct’ … can (but not necessarily will) include inaction: but that conduct has to be taken on the grounds of race or ethnic or national origins if it is to create the hostile environment and thereby come within the heading of harassment. Thus, if inaction occurs because, for instance, the relevant person in the employment of the employer is ill, or for instance because the office is so completely inefficient as to fail to deal with something, or for various other reasons which can easily be imagined which have nothing to do in themselves with race or ethnic or national origin, then the inaction, however regrettable it may be, is not on the grounds of race or ethnic or national origin.

31.

[The claimant’s] argument, as it seems to us, places too much weight upon the nature of the conduct of the third parties. Assuming that third party conduct is to be taken as inherently racist, which must depend on the particular facts of any particular situation, does that mean that a failure to deal with it is itself inherently racist and therefore must be taken to have been itself on the grounds of race? The question nearly answers itself despite [counsel’s] persistent submissions, but in any event she too drew back from the conclusion which would inevitably follow from her argument that if, for instance in this case, [the manager] had had a heart attack which hospitalised him immediately after having had the complaint from the Claimant, and that that had been in truth the reason for his not dealing with the complaint, nonetheless he would have to be condemned and his employer with him as having acted in a racially discriminatory manner by subjecting the Claimant to racial harassment. That would be too far, but she acknowledged it was the result of her primary submission. It demonstrates to us that that submission was in error.”

It will be noted that Langstaff P in that passage regarded it as unacceptable that a manager should be individually liable for discrimination simply because some omission on his part – whether or not culpable – had led to an employee being subjected to harassment by someone else.

At paras. 96-99 the EAT considered the effect of the replacement, following the decision in the EOC case, of the phrase “on the ground of” by “related to”. It referred to the examples given by Burton J of the difference between a “causative” and an “associative” link (see para. 56 above). It then said, at para. 98:

“We have no doubt that the re-cast definition of harassment was intended to encompass cases such as these.  We note, however, that these are all cases where the association is between the conduct of the alleged perpetrator and the protected characteristic.  They are cases where it is just to impose liability on the perpetrator.”

It referred at para. 99 to the Commission’s complaints in the EOC case about “third party liability”; but it described the arguments on the point as “tentative” and said that Burton J had reached no definite conclusion. It also referred to the decision of the EAT in Sheffield City Council v Norouzi [2011] UKEAT 049710/1406, [2011] IRLR 897, which had proceeded on the basis that the effect of the EOC decision was that the Directive did indeed require the proscription of third party liability and could be enforced directly against a public-sector employer; but it pointed out that that had been conceded by the employer. (It is convenient to say at this stage that the EAT was right to make that point: see para. 7 of my judgment in Norouzi (p. 899), where we recorded some unease about the concession).

The EAT’s ultimate decision and reasoning appear in paras. 100-104, which I should reproduce in full:

“100.

In our judgment section 26 requires the ET to focus upon the conduct of the individual or individuals concerned and ask whether their conduct is associated with the protected characteristic - for example, sex as in this case.  It is not enough that an individual has failed to deal with sexual harassment by a third party unless there is something about his own conduct which is related to sex.  We reach this conclusion for the following principal reasons.

101.

Firstly, this approach seems to us to accord with the natural meaning of the words in the European and domestic legislation.  The first task is to identify the conduct (in which, as in Conteh, we would include a settled course of inaction); the next to ask whether that conduct is related to the protected characteristic.  It is not sufficient to ask whether some other, prior, conduct by someone else is related to the protected characteristic.

102.

Secondly, this approach caters for the kind of case which Langstaff J identified in paragraphs 31 and 32 of Conteh.  If inaction is due to illness or incompetence or some real non-discriminatory constraint upon action one would not naturally say that it was ‘related to sex’; but if inaction or a cold shoulder is really indicative of silently taking sides with the perpetrator - even without encouraging the perpetrator - one might well say that it was related to sex.  The focus will be on the person against whom the allegation of harassment is made and his conduct or inaction; it will only be if his conduct is related to sex that he will be liable under section 26.  So long as the ET focuses upon the conduct of the alleged perpetrator himself it will be a matter of fact whether the conduct is related to the protected characteristic. 

103.

Thirdly, there is, as far as we can see, no other mechanism in any Directive, or in UK domestic law (other than the provisions now repealed in section 40) for distinguishing those cases where liability ought to be imposed in relation to third party harassment and those where it ought not to be.  In the EOC case and in Norouzi there was discussion of steps or practices which an employer might be required to undertake in order to prevent or mitigate the effect of third party harassment; but there is no basis - no legal test - for such measures to be found in the primary instruments of legislation. 

104.

In our judgment the ET did not apply the correct approach.  The error is at its clearest in paragraph 4.16 where the ET says that the decision to transfer the Claimant was related to sex ‘because of the background of harassment related to sex’.  This does not follow: as we have seen, it will depend on an assessment of the conduct of Mr Kavanagh rather than that of the perpetrators.  Similarly in paragraph 109 the ET thought that because the complaints were plainly related to sex the inaction of the decision makers must also be related to sex.  Again this does not follow.”

In short, the EAT allowed the appeal as regards the finding of harassment based on the conduct of the employed officials because the ET’s finding that their conduct, as opposed to the lay officials’, was “related to” the Claimant’s sex was based on a misunderstanding of the nature of the exercise required by the statute. The necessary relationship between the conduct complained of and the Claimant’s gender was not created simply by the fact that the complaints with which they failed to deal were complaints about sexual harassment – or, in the case of Mr Kavanagh, that part of the situation that led him to decide to transfer the Claimant was caused by such harassment.

That reasoning is not identical to the EAT’s reasoning as regards discrimination, because of the difference in the statutory language, but it depends on the same basic point, namely that it was necessary to focus on the conduct of the employed officials, who are the actual actors (or – to coin a word – inactors), and not on the nature of the complaints with which they were dealing.

At para. 105 the EAT held that this aspect too would have to be remitted to the ET because the answer to the issue as it should have been understood was not clear from its reasoning. It said:

“There was a lengthy period of inaction following the Claimant’s complaints about sexual harassment.  The decision to transfer the Claimant was made in the knowledge of the conduct aimed at her and without addressing it.  The ET’s finding about the motives of the paid officers may point in their favour but it does not directly address the statutory question in the way we have indicated.”

As will have appeared, the EAT regarded the present case as engaging the law about third party liability. It will be convenient to deal with this first before turning directly to the Claimant’s challenge to the EAT’s reasoning, although my analysis will to some extent answer that challenge. It is necessary for clarity to deal separately with the position under section 13 and under section 26.

The starting-point must be the requirement of section 13 that an act of direct discrimination be “because of” the proscribed factor. The guidance as to the effect of this requirement mostly concerns the previous language of “on the ground of” (or its cognates); but, as I have said, it is settled law that the substitution of “because of” in the 2010 Act was not intended to effect any substantive change. At the most general level, it can obviously be said that the question posed by the statutory language is whether the act complained of was “caused by” the protected characteristic. However that is too general to be helpful, and in fact in Chief Constable of West Yorkshire Police v Khan [2001] UKHL 48, [2001] ICR 1065, Lord Nicholls observed that “causation is a slippery word” and cautioned against its use in this context – see para. 29, at p. 1072 C-E. Courts have often preferred to ask what was “the reason why” the act was done; but that involves the same ambiguities.

The EAT, as we have seen, sought guidance from my judgment in Amnesty v Ahmed, where I made the point that there was more than one way in which the ground, or reason, for an act might be discriminatory. I said, at paras. 33-34 (pp. 1469-70):

“33.

In some cases the ground, or the reason, for the treatment complained of is inherent in the act itself. If an owner of premises puts up a sign saying ‘no blacks admitted’, race is, necessarily, the ground on which (or the reason why) a black person is excluded.  James v Eastleigh [James v Eastleigh Borough Council [1990] 2 AC 751] is a case of this kind. There is a superficial complication, in that the rule which was claimed to be unlawful – namely that pensioners were entitled to free entry to the Council's swimming-pools – was not explicitly discriminatory. But it nevertheless necessarily discriminated against men because men and women had different pensionable ages: the rule could entirely accurately have been stated as ‘free entry for women at 60 and men at 65’. The Council was therefore applying a criterion which was of its nature discriminatory: it was, as Lord Goff put it (at p. 772 C-D), ‘gender based’. In cases of this kind what was going on inside the head of the putative discriminator – whether described as his intention, his motive, his reason or his purpose – will be irrelevant. The ‘ground’ of his action being inherent in the act itself, no further inquiry is needed. It follows that, as the majority in James v Eastleigh decided, a respondent who has treated a claimant less favourably on the grounds of his or her sex or race cannot escape liability because he had a benign motive.

34.

But that is not the only kind of case. In other cases – of which Nagarajan is an example - the act complained of is not in itself discriminatory but is rendered so by a discriminatory motivation, i.e. by the ‘mental processes’ (whether conscious or unconscious) which led the putative discriminator to do the act. Establishing what those processes were is not always an easy inquiry, but tribunals are trusted to be able to draw appropriate inferences from the conduct of the putative discriminator and the surrounding circumstances (with the assistance where necessary of the burden of proof provisions). …”

I have set that reasoning out in full because of the EAT’s reliance on it. It is not of course binding on us. However I believe it to be to the same effect as the approach of the majority in the decision of the Supreme Court in R (E) v Governors of the JFS [2009] UKSC 15, [2010] 2 AC 728 – see in particular paras. 62-64 of the judgment of Lady Hale (pp. 759-760). That is unsurprising because both I and the Supreme Court based ourselves on the reasoning in James v Eastleigh and Nagarajan. In any event neither party before us sought to argue that Amnesty v Ahmed mis-stated the law.

I turn to how those principles are applied in the case of acts of discrimination or (more usually) harassment by a third party.

In Burton v De Vere Hotels Ltd [1997] ICR 1 two black waitresses, clearing tables in the banqueting hall of a hotel, were the butt of racist and sexist jibes made by a guest speaker entertaining the assembled all-male company at a private dinner party. Although the label “harassment” was used, the legislation at that time did not, as I have said, proscribe harassment as such; but the speaker’s conduct was held to constitute direct discrimination. The ET held that the respondent was not liable under the Act for the conduct of the entertainer or the guests. An appeal was allowed by the EAT, which held that an employer subjected an employee to an unlawful detriment under the 1976 Act if it permitted racial harassment to occur in circumstances in which it could control whether the harassment happened or not.

Burton was disapproved by the House of Lords in Pearce v Governing Body of Mayfield School, which was reported together with MacDonald v Advocate General for Scotland [2003] UKHL 34, [2003] ICR 937. The case concerned the liability of a school for the harassment of a lesbian teacher by pupils, which (again) was said to constitute direct discrimination on the grounds of sex. The claim was held to fail on the basis that the conduct of the pupils did not constitute unlawful harassment in any event; but Lord Nicholls, with whose speech the other members of the House agreed, held that even if it had the school could not be held liable for it. He said, at para. 29 (p. 946 C-F):

“Viewed in the broadest terms, the Burton decision has much to commend it. There is, surely, everything to be said in favour of a conclusion which requires employers to take reasonable steps to protect employees from racial or sexual abuse by third parties. But is a failure to do so 'discrimination' by the employer? Where the Burton decision is, indeed, vulnerable is that it treats an employer's inadvertent failure to take such steps as discrimination even though the failure had nothing to do with the sex or race of the employees. In this crucially important respect the decision gives insufficient heed to the statutory discrimination provisions. An essential element of 'direct' sex discrimination by an employer is that, on the grounds of sex, the employer treats the employee less favourably than he treats or would treat an employee of the opposite sex. Similarly with 'direct' racial discrimination: the 'less favourable treatment' comparison is an essential ingredient of the statutory wrong: see section 1(1)(a) of the Race Relations Act 1976. Unless the employer's conduct satisfies this 'less favourable treatment' test, the employer is not guilty of direct sex or racial discrimination. In making this comparison acts of persons for whose conduct an employer is vicariously responsible are to be attributed to the employer. It is otherwise in respect of acts of third parties for whose conduct the employer is not vicariously liable.”

He observed at para. 31 (p. 946H) that “the harassment in Burton was committed by third parties for whose conduct the employer was not vicariously responsible”. In those circumstances discrimination within the meaning of the Act could not be established.

The essence of that ratio is that in such a case the protected characteristic has to be the “ground of” – or reason for – the employer’s failure to protect the employee against the harassment by the third party, and that that is not established by showing simply that what he failed to protect the employee against was unlawful discrimination: the focus is on the grounds for the employer’s action, not the third party’s. That ratio does not of course preclude the employee from showing that on the facts of a particular case (and with the benefit of section 136 if necessary) a protected characteristic did form part of the mental processes that caused the employer’s failure – if, for example (to take a crude case) the putative discriminator had said in terms “I won’t intervene because the complainant is black”, or (more sophisticatedly) if he or she could be shown to have reacted differently to complaints of third party sexual harassment brought by white victims than by black victims. But that is the only route by which the employer can be liable.

The ratio in Pearce must apply to any provision using the terminology of “on the grounds of”. Thus it will have applied to the 2005 legislation which first outlawed harassment eo nomine, which used that terminology: I thus regard the case of Conteh which was relied on by the EAT as correctly decided. (Footnote: ⁶) It will also continue to apply to section 13 of the 2010 Act. I turn to consider whether the position is any different under section 26.

Given my conclusion in the foregoing paragraphs about the effect of the phrase “on the ground of”, the question is what, if any, change was effected by the substitution of the language of “related to”.

As to that, it is clear that, as the EAT held, the change was made in response to the decision in the EOC case. It must at least have been the intention both of the Secretary of State in 2008 and of Parliament in 2010 to ensure that the legislation applied in cases where Burton J had held that the current language did not (or arguably did not) reflect the requirements of the Directive. It follows that the change must have been intended to ensure that the statutory definition covered cases where the acts complained of were “associated with” the proscribed factor as well as those where they were “caused by” it, as illustrated by the examples that he gave (see para. 56 above).

I should like to make two points in passing before proceeding further:

The question then is whether the change in the language of section 4A was intended to go further than the kinds of case identified in Burton J's examples and to change the position about liability for third party harassment as established by the House of Lords in Pearce. Mr Carr contended that it was. He relied on the consideration of third party liability in the EOC case and submitted that the Secretary of State in proceeding to amend the 1975 Act so as to adopt the language of “related to”, and then Parliament in adopting it generally in 2010, must have been taken to have accepted Ms Rose’s argument, accepted by Burton J, that that language would open the door to third party liability.

I do not accept that submission. My reasons are as follows.

First, the Commission did not in the EOC case persist in its contention that the Directive imposed a requirement to make provision for third party liability, and Burton J proceeded on the basis that it did not: see para. 57 above. Ms Rose’s eventual case was based only on the fact that the Government itself appeared to agree with the Commission in regarding it as desirable that third party liability should be available. But that could not support the argument that there was an obligation to change the statutory language: if it was simply a matter of Government policy, it was a matter for the Government and not the Court to decide what should be done to implement it. The fact that the introduction of the “related to” formula would – as Ms Rose submitted and Burton J appeared to accept – open the door to third party liability was, as a matter of analysis, a desirable by-product of the change and not a reason why it was required.

Secondly, it is in my view highly significant that when the Government did in fact introduce provision for third party liability in 2008, and when Parliament continued it in 2010, it was done explicitly, first by the new section 6 (2B) and then by sub-sections (2)-(4) of section 40: see paras. 58 (2) and 59 above. That is hard to reconcile with third party liability being provided for separately as a result of the language of “related to” in (first) the amended section 4A of the 1975 Act and (then) section 26 of the 2010 Act.

Thirdly, even if we ignore the context considered above, I do not believe – here respectfully parting company with what appears to have been Burton J’s view in the EOC case – that the mere use of the formula “related to” is sufficient to convey an intention that employers who are themselves innocent of any discriminatory motivation should be liable for the discriminatory acts of third parties, even if they could have prevented them. In my view the “associative” effect of the phrase “related to” is more naturally applied only to the case where the discriminatory conduct is the employer’s own, as it was in each of Burton J’s examples. There are clearly good policy arguments for imposing liability in such a case, but negligent failure to prevent another’s discriminatory acts is a very different kind of animal from liability for one’s own: it requires careful definition, and I would expect it to be covered by explicit provision. I would also be uneasy about a situation where such liability was incurred not only by the employer but also by any individual employee who might be implicated in the failure to afford adequate protection. These are essentially the points made by the EAT at paras. 100-103 of its judgment (and also at para. 92), with which I respectfully agree.

It follows from the foregoing that the repeal in 2013 of sub-sections (2)-(4) of section 40 means that the 2010 Act, for better or for worse, no longer contains any provision making employers liable for failing to protect employees against third party harassment as such, though they may of course remain liable if the proscribed factor forms part of the motivation for their inaction: in substance the position remains as analysed in Conteh.

Mr Carr disputed that conclusion. He submitted that the repeal of sub-sections (2)-(4) was a red herring. In his post-hearing submissions he referred us to an article by Dr James Hand of the University of Portsmouth in the Industrial Law Journal (2013, vol. 1, p. 75) which argued that the sub-sections were superfluous and that third party liability could be established by reference to the language of section 26 itself. For the reasons which I have given I do not accept that. I do not, as I have said, believe that Parliament can be treated as having introduced a careful and explicit scheme providing for third party liability notwithstanding that such liability was already implicitly provided for elsewhere. It is, further, hard to see what would be the point of Parliament repealing that scheme if the effect were to leave in place an implied liability which was broadly equivalent but of uncertain scope, particularly since the purpose of enacting section 65 of the 2013 Act, as expressed in the Explanatory Notes, was that the sub-sections (2)-(4) were “considered to impose an unnecessary burden on business”. (Footnote: ⁸)

I appreciate that that conclusion means that employees do not have a protection which both the EAT in Burton and Lord Nicholls in Pearce thought it desirable that they should enjoy, and which indeed they did enjoy between 2008 and 2013. But the availability of third party liability is a matter for Parliament, and the policy decision effected by the 2013 Act must be respected.

I can at last turn to the Claimant’s grounds of appeal against the decision of the EAT. These were only pleaded in very general terms and I will focus on how they were developed by Mr Carr in his oral submissions.

Mr Carr took us in some detail through the factual findings of the ET. He emphasised that the employed officials were well aware of the conduct of the lay officials and that it was related to the Claimant’s gender. As for Mr Kavanagh’s decision to transfer the Claimant, at least partly because of her complaints about Mr Coxhill and Mr Saini, this was, he said, a classic case of penalising the victim instead of the perpetrators. He did not challenge the EAT’s analysis of the ET’s reasoning, which corresponds to my own at para. 69 above; specifically, he did not seek to argue that the ET had in fact based its decision on the motivations of the employed officials. But he said that it did not need to do so. Focusing in particular on the transfer decision, he contended that it was enough that the situation that led to her being moved arose because she had suffered sexual harassment: it was, applying the Amnesty analysis, a case where her sex was inherent in the reason for her treatment and accordingly where her treatment could be said to be “because of” it within the meaning of section 13.

I take first the inaction of the employed officials – that is, the conduct other than Mr Kavanagh’s transfer decision. Subject to the potential distinction which I discuss below, the present case is on all fours with Conteh, which I have already said that I regard as correctly decided and as reflecting the reasoning of the House of Lords in Pearce. A failure to protect an employee against discrimination by someone else does not by itself constitute an act (or omission) “on the grounds of” – or, therefore, “because of” – the relevant protected characteristic. If the employed officials are to be held liable for discrimination it must be, as Langstaff P held in Conteh, on the basis of their own discriminatory motivation.

The only difference on the facts between the two cases is that in Conteh the discrimination against which the employer failed to protect the claimant was on the part of the employer’s client whereas in the present case it was on the part of persons who were part of the employer’s own organisation. However, that is immaterial for the purposes of section 13. As the EAT rightly appreciated, under the scheme of the 2010 Act the liability of an employer for discrimination always depends on an individual employee or agent having done a discriminatory act, for which he or she will also be individually liable (and indeed sometimes solely liable, if the employer can invoke the reasonable steps defence in section 109 (4)): see the decision in CLFIS to which it referred. Thus in the present case, just as much as in Conteh, what matters is that A is being sought to be made liable in respect of the discriminatory acts of B, even though the Claimant has chosen not to pursue the employed officials individually.

Mr Carr’s reference to the Claimant’s sex being “inherent” in her treatment picked up on a reference at para. 34 of my judgment in Amnesty to cases where the discriminatory ground is “inherent in the act itself”. However, it is necessary to read that in context. In the passage in question I was explaining why it was unnecessary to consider the mental processes of the putative discriminator in cases where what Lord Goff in James v Eastleigh called a “gender-based criterion” had been applied. It is that concept of a “criterion” that is central to this type of case; and I note that that is the way it is put by Lady Hale in her judgment in the JFS case. In failing to deal adequately with the Claimant’s complaints the employed officials were clearly not acting “because of” her sex in that sense: it was, so to speak, simply the subject matter of their inaction.

The position of Mr Kavanagh at first sight seems different because he did take a positive decision to transfer the Claimant; but in the end I think the analysis must be the same. His decision to transfer the Claimant was (in part) because the relationship between her and the lay officials had broken down, and the reason for that breakdown was their conduct towards her, which was (in part) related to her sex. But I do not believe that those facts by themselves mean that his decision can be treated as being because of her sex in the relevant sense. It was not the “criterion” for her transfer. If he is to be liable for discrimination it must be because the Claimant’s gender formed part of his own mental processes.

Mr Carr submitted that, even if the employed officials’ conduct could not be said to be “because of” the Claimant’s sex, it was on any view “related to” it within the meaning of section 26. I have already explained at paras. 96-98 above why that language does not cover cases of third party liability; and for the reasons given at para. 104, the present claim is, on the ET’s reasoning, in substance such a case. If the employed officials, and through them the Union, are to be liable for harassing the Claimant because of their failure to protect her from the harassment of the lay officials, and (in the case of Mr Kavanagh) for transferring her, that can only be because of their own motivation, as to which the Tribunal made no finding.

Mr Segal sought in his post-hearing submissions to distinguish between a situation where an employer was “culpably inactive knowing that an employee is subjected to continuing harassment (as on the facts of Burton)” and one where he was culpably inactive without [any such knowledge]”; and to show that the ET’s findings established that the case was in the latter category. I am not sure of the relevance of the distinction; but since we did not hear oral submissions on it I prefer to say no more than that on the law as I believe it to be the employer will not be automatically liable in either situation. I repeat, to avoid any possible misunderstanding, that the key word is “automatically”: it will of course be liable if the mental processes of the individual decision-taker(s) are found (with the assistance of section 136 if necessary) to have been significantly influenced, consciously or unconsciously, by the relevant protected characteristic.

For those reasons I agree with the EAT that the reasoning of the ET was flawed. It found the Union liable on the basis of the acts and omissions of the employed officials without making any finding as to whether the Claimant’s sex formed part of their motivation. The case will in principle, as it held, have to go back to the ET for the necessary findings. (Footnote: ⁹) I have to say, however, that I hope that exercise can be avoided. Although we have heard no argument on the point, I find it hard to see how a finding of liability against the Union on the basis of the conduct of the employed officials could significantly add to the compensation to which the Claimant will be entitled to as a result of the findings made against it on the basis of the conduct of the lay officials and of her unfair, and discriminatory, constructive dismissal; and my assumption is that the Claimant’s appeal was pursued only in case she was unable to resist the Respondent’s appeal on the agency issue.

I agree.