[2025] EWCA Civ 1466

It is common ground that section 47B was amended by the 2013 Act in response to the decision in Fecitt. The purpose of the amendments was to make employees and agents of an employer liable for their acts of victimisation against a co-worker, and to make the employer vicariously liable for such acts. Section 19 of the 2013 Act is headed ‘Worker subjected to detriment by co-worker or agent of employer’. Section 19(1) inserts subsections (1A)-(1E) in section 47B of the 1996 Act. Section 19(2) inserts words into section 48(5) of the 1996 Act (which deals with complaints to ETs) to make clear that in section 48 and in section 49, any reference to the employer includes, in the case of proceedings against a worker or agent under section 47B(1A), the worker or agent. Section 47B(2) was not amended by the 2013 Act, and has not been amended since.

The decision in Osipov

This recent decision of this court (Underhill LJ, with whom Rafferty LJ and Sales LJ agreed) directly concerned the reach of section 47B(2).

The claimant, Mr Osipov was employed as the CEO of IPL (R1). Mr Timis (R2) was the largest individual shareholder and a director. Mr Sage (R3) was the Chairman. R2 decided that the claimant should be summarily dismissed. R3 agreed, and dismissed the claimant by email. An ET found that the principal reason for the dismissal was that the claimant had made protected disclosures; so his claim under section 103A that his dismissal by R1 was automatically unfair succeeded. The ET also held that R2 and R3 had subjected the claimant to a detriment or detriments contrary to section 47B and that they were jointly and severally liable with R1 to compensate him for his losses, amounting to over £2m. The detriments included instructions or recommendations given by R2 and R3 which culminated in the dismissal (referred to as ‘detriment (m)’).

The respondents appealed to the EAT (Simler P), and the claimant cross-appealed. There were many issues. The EAT dismissed R1’s appeal (save in one very minor respect). It also dismissed the appeals of R2 and R3 on all issues, except that it held that they could not be liable for the basic award for unfair dismissal (see section 219 of the 1996 Act) as that award can only be made against “a party liable for the dismissal”.

The reasoning of the EAT was that the express purpose of the legislation was to protect whistleblowers. It was therefore appropriate to construe section 47B(2), so far as it can properly be construed, to provide protection rather than deny it. All the words, including those in brackets, must be construed in light of that intended purpose. It did not exclude all claims for detriment amounting to a dismissal. Instead, it was limited to detriments amounting to dismissal within the meaning of Part X, in other words to detriments amounting to unfair dismissal claims necessarily against the employer. The submissions for the respondents ignored the words in brackets. Section 47B(2) did not relieve a co-worker of liability for a detriment amounting to dismissal not within the meaning of Part X. This approach was coherent and did not strain the meaning of the legislation. It put employees in the same position as workers who never lose the right to make claims against individuals for detriments amounting to dismissal and ensured that employees are given the same protection as workers who are subjected to the most serious detriments and not put in a worse position than those workers. An employee would be unlikely to pursue a claim for a whistleblowing detriment amounting to a dismissal against an employee, rather than against the employer, but there was no principled reason for excluding it.

R1 tried to appeal to this court but its appeal was dismissed by consent after it became insolvent. R2 and R3 appealed on two grounds. The first was that section 47B(2) meant that they could not be liable to the claimant for an instruction to dismiss him, and could not be liable for losses flowing from his dismissal. The second was that R3 could not be liable to the claimant in any event because the ET had found that R2, not R3, had given the instruction to dismiss the claimant.

Underhill LJ gave the single reasoned judgment. It is detailed and multi-faceted and we only seek to draw out its essential reasoning.

At paragraph 1(5) he described the principal issue in the appeal as being whether it was open to the ET to award the claimant compensation against the directors, as individuals, for the losses occasioned by his dismissal. At paragraph 24, he noted that this turned on the meaning and effect of section 47B(2). He helpfully described the legislative history of the relevant provisions. At paragraph 32 he said that the effect of the amendment was that an employer could be liable by one of two routes: liability for its own act under section 47B(1), and vicarious liability under section 47B(1B). Which route was available would be important in cases in which an employer relies on a reasonable steps defence, as that defence is only available in claims under section 47B(1A). In paragraph 33 he adverted to the differences between this regime and the statutory provisions in the Equality Act 2010 which relate to discrimination.

In paragraph 59, Underhill LJ accepted that whether a detriment amounts to a dismissal is a question of substance, not form, and that it was artificial to treat detriment (m) as distinct from the claimant’s dismissal. That was only part of the answer, however, as the claimant’s case was that section 47B(2) did not apply to a claim as against the directors.

Underhill LJ recorded in paragraph 6 the respondents’ argument that if the claimant’s construction were correct, it would undermine the “careful demarcation” between Part V and Part X. In practice, every claimant could make a claim against the employer “(on the back of the co-worker’s liability under subsection (1A)) as well as, or instead of, under section 103A”. There were two advantages for claimants; they could get damages for injury to feelings and rely on a less restrictive causation test. Section 103A would become a dead letter. There was no reason to think that Parliament was concerned about the possibility that an employer could become insolvent. That was a risk in all litigation and Parliament had addressed it in Part XII, to the extent which it considered appropriate. Nor did the respondents accept that it was more coherent for employees and workers to be treated in the same way; their circumstances are fundamentally different. Contrary to the view expressed by Simler P, it was not an unacceptable anomaly that a co-worker could not be liable for a whistleblower dismissal when he or she could be liable for a prior act which caused it. That was simply a consequence of the division of labour between Parts V and X.

At paragraph 67, Underhill LJ said that he had not found this point straightforward, but he had decided that the decision of the EAT was right. He agreed with Simler P that a construction of section 47B(2) which prevented a claimant from bringing a claim against a co-worker based on the detriment of dismissal would produce an incoherent and unsatisfactory result and was accordingly unlikely to conform to Parliament’s intention. If Parliament had decided to make co-workers liable for whistleblower detriment it was hard to see any reason in principle why they should, uniquely, not be so liable in a case where the detriment amounts to dismissal. That produced the obvious anomalies to which Simler P had referred.

Further, Parliament’s allocation of different rights to Parts V and X was not as significant as the respondents argued:

“70 … It is of course right that the two Parts constitute distinct and largely self-contained regimes. However, that simply reflects the separate and historically prior development of protection against unfair dismissal. As it came to be recognised that treatment on certain proscribed grounds required additional protection, the choice was made not to disturb the existing unfair dismissal regime but to create a complementary regime covering detriments other than dismissal. But the different regimes address different aspects of the same mischiefs; and, as I have shown, they mostly employ substantially the same drafting. There is no reason to believe that they reflect any great conceptual gulf perceived by Parliament between dismissal and other kinds of detriment. That being so, although anti-overlap provisions of the kind found in section 44(4) were required in the interests of good order, the policy behind them is unlikely to have been anything more than that a claimant should not claim under Part V where the identical right was available under Part X.”

(our emphasis)

Underhill LJ considered that there was no reason to suppose that that policy changed in 1998, or in 2013. There was no basis for thinking that the introduction of individual liability meant that, for the first time, section 47B(2), which had not been amended, was intended to exclude liability which would arise under the other provisions of the section but which was not provided for in section 103A. He said that he was initially troubled by two points: by the fact that compensation for injury to feelings can be recovered in detriment claims but not in unfair dismissal claims, and by the difference between the tests for causation in Part X and Part V. Those two points were anomalies but they were “particular wrinkles” in the statutory scheme without any wider significance. There was no indication that the draftsman had focused on compensation for injury to feelings at all. The difference in the tests for causation, too, “may have been inadvertent; or the draftsman may simply have taken the view that the difference would rarely matter in practice and that it would be too difficult to find a way of assimilating the two tests”. He acknowledged that Parliament could have changed the law about compensation for injury to feelings by amendment, if that did not reflect its intention, in the same way as it had filled the gap identified in Fecitt; but that was not always a safe assumption.

He continued:

“75 I accordingly approach the construction of the language of section 47B(2) on the basis that I would expect Parliament to have intended to exclude liability under the operative provisions of the section only where the identical remedy was available under section 103A; and thus that it would not exclude a co-worker’s individual liability for the detriment of dismissal under subsection (1A) (or, which follows, any vicarious liability of the employer under subsection (1B)). I do not believe that the statutory language compels a different construction.”

(our emphasis)

He therefore reached this conclusion:

“77 I would accordingly hold that section 47B(2) does not prevent the claimant proceeding against the appellant directors under Part V on the basis of their responsibility for the dismissal itself.

78 I accept that this approach to the meaning of section 47B(2) does not produce a particularly elegant result. It is clumsy that an employee dismissed on whistleblower grounds should be able to pursue distinct causes of action, with significant differences as regards the conditions of liability and (perhaps) compensation, against his or her employer. It may well be that Parliament did not really think through the technical challenges of inserting into the framework of the 1996 Act a scheme of individual liability largely borrowed from the discrimination legislation. But the resulting awkwardnesses are insufficient to justify a construction that would produce much more serious anomalies and seems contrary to the overall policy of these provisions”.

Referring to the decision in Melia, Underhill LJ said that his analysis of the legislative policy was substantially the same as that of Chadwick LJ, but that Melia concerned a different issue and was decided before the 2013 Act. At that time the phrases “the same loss or detriment” and “the identical claim”necessarily meant the same thing. A more refined approach was now required.

Underhill LJ finally summarised his construction of section 47B(2):

“91 …I can summarise my essential conclusions as follows.

(1)

It is open to an employee to bring a claim under section 47B(1A) against an individual co-worker for subjecting him or her to the detriment of dismissal, i.e. for being a party to the decision to dismiss; and to bring a claim of vicarious liability for that act against the employer under section 47B(1B). All that section 47B(2) excludes is a claim against the employer in respect of its own act of dismissal.

(2)

As regards a claim based on a distinct prior detrimental act done by a co-worker which results in the claimant’s dismissal, section 47B(2) does not preclude recovery in respect of losses flowing from the dismissal, though the usual rules about remoteness and the quantification of such losses will apply.”

Our construction of the legislation

For convenience, we reproduce section 47B(2). Referring to section 47B as a whole, it provides that:

“(2)

This section does not apply where—

(a)

the worker is an employee, and

(b)

the detriment in question amounts to dismissal (within the meaning of Part X).”

Section 47B(2) is part of a careful and detailed scheme which gives employees, and in some cases workers, rights not to be subjected to detriments on specified grounds. The scheme has many common features, such as the formula we describe in paragraph 33, above, which concerns the relationship between the relevant substantive sections in Part V and Part X. Those common features have been in Part V since its enactment, and have not changed; they have just been applied to an increasing number of proscribed grounds. In particular, those features were not changed by the amendments to Part V in 1998, or in 2013.

The formula in each substantive provision is Parliament’s choice about the remedial scheme which is available to an employee who suffers a detriment on a relevant ground. Whether that scheme is rational, or could be better, is not a matter for the court. Parliament has decided that where the detriment in question “amounts to a dismissal (within the meaning of Part X)”, the employee cannot make a claim under Part V. That provision is not ambiguous. It does not use the phrase ‘unfair dismissal’, so its scope is not restricted to cases in which a claim of unfair dismissal is or might be available. ‘Dismissal’ is defined in section 95 (see paragraph 29 above). The scope of the exclusion is defined by the phrase “amounts to dismissal (within the meaning of Part X)”. The phrase ‘amounts to’ makes it clear that whether the exclusion applies in a particular case is a question of substance, not form. So, if an employee’s complaint about a detriment is, in substance, a complaint about his dismissal, he cannot bring a complaint about that detriment under Part V; he is necessarily limited to making a claim under Part X. The regimes in the two Parts are parallel and complementary. There is no warrant for the view that employees should, as well as being able to claim under Part X, be given the windfall of a claim under Part V also based on their dismissal, when Parliament has unambiguously decided that the extent of their protection from dismissal should be a claim under Part X.

The inquiry whether the exclusion applies, that is, whether the detriment amounts, in substance, to a dismissal (‘within the meaning of Part X’) has three main linked elements. First, is the detriment which is the subject of the complaint in substance a dismissal? Second, is the loss for which compensation is sought in substance a loss consequential on the dismissal, or compensation for a loss suffered before the dismissal? That is the approach of this court in Melia – see paragraphs 40-44 above – and we agree with it. Third, did the employer dismiss the employee? This third element might be thought to be redundant, since it might be thought that only an employer can dismiss an employee, but it needs to be stated, because of the suggestion in some of the authorities that someone other than the employer can, in law, dismiss an employee.

There is no exclusion provision in Part V which applies to workers who are not employees. The reason for that is, on our analysis, obvious: a worker who is not an employee cannot claim that he has been unfairly dismissed. There is therefore no need for an exclusion in their case. Such a worker can therefore complain about any detriment to which he has been subjected on a proscribed ground. It does not matter whether it is in substance a dismissal by the employer, or not.

There are two further important indicators. The first is that the draftsman has, in some cases, created an exclusion by reference to the whole section, and in two cases, an exclusion which disapplies only some subsections of the relevant substantive section (see paragraph 38, above). That technique was therefore available to, and has been used by, the draftsman. When section 47B was amended by the 2013 Act, the draftsman could, if that had been Parliament’s intention, have used that technique, and have amended section 47B(2) so as to disapply section 47B(2) to the new subsections (1A)-(1E). The draftsman did not do that. Section 47B(2) therefore unarguably applies to the whole of section 47B, as amended, and not just to section 47B(1). There are no textual grounds for supposing that this was a drafting oversight, still less for concluding that it is an ‘obvious’ drafting error. The criteria which enable the court to intervene to correct even an obvious statutory drafting error are very strict: Inco Europe Limited v First Choice Distribution [2000] 1 WLR 586 (HL (E)) (‘Inco’). There has been no suggestion that they are met in this case.

The second indicator is that the relevant terms of the exclusion are identical, or virtually identical, in many different provisions in Part V. Those provisions must all have the same meaning throughout Part V, and must have had that meaning from the date of Royal Assent. We cannot accept Ms Jolly’s submission that the same words could mean one thing in section 47B(2) and something else in other provisions. Nor, given that the draftsman could, when section 47B was amended in 2013, have used the technique we have referred to in the previous paragraph, yet chose not to, can section 47B(2) have had one meaning before the amendments in 2013, and then have acquired a different meaning after those amendments were made. It means the same now as it did when it was considered by this court in Melia.

The upshot is that, whether or not an employee relies on section 47B(1), or on section 47B(1A)-(1E), he cannot bring a detriment claim if, in substance, the detriment about which he wishes to complain ‘amounts to’ his dismissal by his employer, and he is claiming compensation for losses consequential on his dismissal. The words are clear. It is just not possible for such an employee to rely on section 47B(1A)-(1E), because, if his complaint is in substance about a detriment which amounts to his dismissal by his employer, the whole of section 47B is deliberately disapplied, including section 47B(1A)-(1E).

We should nevertheless consider further one aspect of the analysis which was pressed on us by some of the parties. In summary, the point is that Part X is said only to apply to a dismissal by an ‘employer’. The new provisions are said to create the possibility of a dismissal by a co-worker, for which the employer would only be vicariously liable. That would mean that the detriment which is the subject of the complaint would not amount to a dismissal ‘(within Part X)’ because the claimant has not been dismissed by his employer, but by a co-worker. We cannot accept that analysis for three reasons.

First, by definition a dismissal is always an act of the employer. A dismissal ends the contract of employment between the employer and employee. It is also obvious that, in cases like these, in which the employer is not a sole trader, but a limited company, dismissal can only be effected by a person who is a co-worker of the claimant, as the company can only act through a human being. We cannot accept that there is a relevant legal distinction in these cases between a dismissal ‘by the employer’ and a dismissal ‘by a co-worker’. The effect of both, in substance (which is what section 47B(2) is aimed at), is that they amount ‘to a dismissal (within Part X)’. In other words, there is no such thing as a dismissal of an employee with sufficient length of service which does not fall within Part X.

Second, and in any event, this argument, is, on its own terms and in these cases, circular. Section 47B(1B) provides that anything done by a co-worker ‘is treated as also done by the employer’. If there were any doubt (and there is none) this means that the dismissal ‘by the co-worker’ is ‘treated as also done by the employer’. Its legal effect, therefore, is that the employee is dismissed by the employer, and that act ‘amounts to a dismissal (within the meaning of Part X)’.

Third, the question (if, contrary to our clear view, it arises) is not whether the liability for an act is primary or vicarious or secondary, or any other type of liability. The question is what the employer or co-worker is said to be liable for. In the terms of section 47B(2), the question, therefore, is what the act ‘amounts to’. If it amounts to a dismissal (within the meaning of Part X), the employer is liable for it and the employee cannot therefore make a complaint under Part V.

We next seek to explain why we respectfully differ from the analysis in Osipov.

It can be seen that our construction of section 47B(2) is based on a straightforward reading of the words of the sub-section, considered in the context of the statutory scheme. We do not consider the sub-section to be ambiguous or to require a remedial or purposive interpretation.

In contrast, we understand the approach to construction in Osipov in this court and in the EAT to be based on three themes. The first two themes are linked: in various respects, Parliament and the draftsman have not thought things through, or have overlooked things; and that the relevant provisions have a purpose which, in several respects, is contradicted by the statutory provisions. The third theme is that in this legislation dismissal by a co-worker is a distinct cause of action to dismissal by an employer. We have just explained at paragraphs 74-77 why we do not accept that proposition.

In regard to the first two themes, we have explained above why we consider that the statutory language is clear and not in need of interpretation, and we have referred to Inco at paragraph 71 above. Here, there are no circumstances which permit a court to ignore statutory language, or to downplay it, on the grounds that the draftsman might have overlooked something, or on the grounds that the words do not accord with a general purpose which is contradicted by those words or by their clear effect. In fact, we do not find the provision to conflict with the statutory purpose or share the view that applying it directly would produce “an incoherent and unsatisfactory result”. In our view, the different tests for causation between Part X (‘the reason or principal reason for the dismissal’) and in Part V (‘on the ground that’) are not anomalous. Nor is the fact that an employee cannot get compensation for injury to feelings under Part X, but can under Part V.

As to the ability to recover losses from co-workers in respect of dismissal, Osipov was, perhaps unusually, a case where recovery was only likely to be possible against co-workers, and not against the employer. However, Parliament has considered, and provided protection for, the rights of employees in an insolvency to the extent that it has in Part XII, and no further.

These matters are the consequences of a deliberate choice by Parliament to give a remedy under Part X which differs in those respects from the remedy it has provided under Part V. Rather than being anomalous, they show that Parliament has created two distinct remedial regimes.

In three places, Underhill LJ describes his expectation that section 47B(2) could only exclude liability where “an identical remedy” was available under Part X. However, the statutory scheme unmistakeably does not provide for identical remedies as between Part X and Part V, and it is therefore difficult to see how this expectation could ever be met.

A further problem with the analysis in Osipov is that it gives no weight to the fact that provisions like section 47B(2) have been in the statutory scheme since enactment. The relationship between the two Parts of the 1996 Act has remained the same throughout. Its meaning and purpose were clearly explained in Melia, before the 2013 amendments. There is nothing to support the view that its meaning changed in 2013. To distinguish Melia on the basis that the issue was “wholly different” from the issue in Osipov is in our view problematic: Chadwick LJ was undertaking a high level review of the Parts V and X that remains convincing. Further, when section 47B was amended in 2013, the draftsman could also have amended section 47B(2), but chose not to. That is a clear indication that no part of section 47B applies where the worker is an employee and can make a claim under Part X.

For these reasons, our analysis of section 47B respectfully differs from the analysis found in Osipov.

Is Osipov binding on us, and what are the consequences for the appeals?

The issue in these two appeals is whether a claimant who has made a whistleblowing claim against his employer under section 103A can later amend his claim so as to add a claim against his employer, complaining of whistleblowing detriment by a co-worker who, he alleges, was responsible for his dismissal, without joining the co-worker to the claim. Such a claim is expressly based on part of the 2013 amendments to section 47B, that is, section 47B(1A).

The core of the reasoning in Osipov is that section 47B(2) does not mean what it says. The consequence for Mr Osipov was that section 47B(2) did not prevent him, an employee who complained about his dismissal under Part X, from also relying on section 47B(1A), one of the provisions introduced by the 2013 Act. Section 47B(2) is, therefore, no bar to a claim by an employee for unfair dismissal against his employer, and a simultaneous detriment claim against a co-worker, also based on his dismissal. It follows from Osipov that an employee who initially makes a claim for unfair dismissal may later apply to amend that claim to include a detriment claim against a co-worker, which is based on his dismissal. Subject to any discretionary factors, such as the expiry of any relevant time limit, an ET faced with an application for such an amendment would be bound by Osipov to grant it.

The question raised by these appeals is whether the decision in Osipov binds us to decide that an employee may apply to amend an unfair dismissal claim to add such a detriment claim without joining the co-worker to the claim. An employee in such a case is in a relevantly similar position to that of the employee in Osipov and to that of the employee posited in the last two sentences of the previous paragraph. This employee also wishes to rely on a provision introduced by the 2013 amendments, section 47B(1B), despite the apparent statutory bar in section 47B(2). We are bound by the doctrine of precedent to give section 47B(2) the same interpretation in this slightly different context. If section 47B(2), as interpreted in Osipov, does not bar reliance on section 47B(1A), it cannot, at the same time, bar reliance on section 47B(1B). This means that we are bound to hold that, contrary to our own construction, section 47B(2) does not bar the amendments proposed in the present cases.

We note, further, that the EAT in Wicked Vision expressly held that whether or not the co-worker was added to the claim was irrelevant. There is no Respondent’s Notice challenging that reasoning. The EAT was right to make this point: it is supported by what we have said in the previous three paragraphs.

We therefore consider that Osipov was binding on the ETs in these cases, and also on us. There was some discussion in argument about whether Underhill LJ accurately encapsulated the ratio of Osipov in the first clause of the first sentence of paragraph 91(1) of his judgment, or whether he expressed the ratio in the whole paragraph (see paragraph 65, above). The first clause of the first sentence describes the actual decision in that case. But the rest of the paragraph flows inevitably from the first clause, because it is based on the essential reasoning in Osipov, as we have just described it.